LIBRARY
Michigan State
7 3 University

This is to certify that the
thesis entitled

COMPETITIVENESS OF COWPEA-BASED PROCESSED
PRODUCTS: A CASE STUDY IN GHANA

presented by
TOMOKAZU NAGAI

has been accepted towards fulfillment
of the requirements for the

MS. degree in Agricultural, Food, and
Resource Economics

/ M//P:{ a jig
ajor ro essor’ ignature
May 9-7; 2008

 

Date

MSU is an affinnative-action, equal-opportunity employer

 

PLACE IN RETURN BOX to remove this checkout from your record.

TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE DATE DUE DATE DUE

 

 

mam:

grill"

 

 

COMPETITIVENESS OF COWPEA-BASED PROCESSED PRODUCTS:
A CASE STUDY IN GHANA

By

Tomokazu Nagai

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE

Department of Agricultural, Food, and Resource Economics

2008

 

     
 

, ABSTRACT
riff-I: ‘ ETITIVENESS OF COWPEA-BASED PROCESSED PRODUCTS:
‘ A CASE STUDY IN GHANA
By
Tomokazu Nagai

In West and Central Africa, cowpeas (blackeye peas) are an important source of
MB for low-income consumers. While they are used for a variety of dishes, industrial
processing of cowpeas is still negligible. This study examines the competitiveness of two
selected processed products using cowpeas in Ghana. Using the data collected through
fieldwork, enterprise budgeting and sensitivity analyses are conducted. One of the
selected products is dry cowpea meal for preparation of kosei (cowpea fritters), which is
expected to help street vendors who currently prepare kosei from cowpea grain using
lime-consuming methods. However, the study shows that under current conditions in
A0018, industrially-processed dry meal would be too expensive for the majority of kosei
vendors to use as the substitute for cowpea grain. The other selected product is a weaning
food called Weanimix, which is a fortified product of the traditional roasted-maize-based
mining food, called Tom Brown. In addition to groundnuts, either cowpeas or soybeans
«the used as a fortifier in Weanimix. The study shows that: (1) cowpeas could be price-

. metifive with soybeans as an ingredient in Weanimix, unless customers prefer

  
  

~ (2) typical difference in prices currently observed between Weanimix and Tom

 

    
   
  
 

A -, . W to be much larger than the difference in the cost of production between

~ , ‘y-pmcessed Weanimix, comparedto Weanimixthatcoum
‘ mmwmmm

   

    
   
   
     
 
 
   

grunt»

r

3;" .. In 3.1.x".-

.9 '. “laid lv.:‘ '.'

I.

“‘1‘ l

, ,

V. .Iu’o \‘-.|Ir.v.',

.m-

.‘,r . 0': t". 'l'

m' m- 1 ‘

0;". I‘ M“ eri' l '4 . l
.1

. I am .4 t . To my parents
. 'T
. ~ cu an"

’1’ (M El 3 1”, -‘

6L 1-

  
  

“ - 9%}: A.

r I

t.. m, it would in?!

—'2 -

'~'. .

,-

t
.

l i

‘

gr;

   
  
   

Without

 

" ' of it , ACKNOWLEDGMENTS

   
 
   

' 'W deepest gratitude goes to Dr. John Staatz, who has been my academic advisor
Lane to MSU in 2004. Without his extreme patience and accurate direction, I
z. I 7' ltnr.

‘ ’ could not have survived the program, let alone finished this thesis.

I

lalso would like to thank Dr. Richard Bemsten, my assistantship supervisor, for
Mme the great opportunity to work on this study and also for reviewing the draft of
lilii'slhesis very carefully.

m It I am also grateful to Dr. Irvin Widders and Dr. Mywish Maredia, both my
eommittee members, for reading this extensive thesis promptly and giving me feedback
With insightful comments.

My special thanks go to Dr. Eric Crawford for his help in my search for the
Midgefing methods and Dr. James Hilker for his help in my search for data on the prices
of soybeans.

. Iwould like to sincerely thank Dr. Suzanne Thornsbury. If she had not hired me as

aresearch assistant, 1 could not have continued my second year of study at MSU. When I

 

 
  
 

WIS struggling, l was mentally supported by her, Mollie Woods, and other members of
t!

Idem-nu. A--.” -

   

I owe everything that I obtained during my fieldwork in Ghana to Dr. Esther
I ,

   

-~.’ ":3 - i , 5» . t . George Annor, and their colleagues in the Department of Nutrition and
lit}

 
 

at the University of Ghana-Legon. Dr. Sakyi-Dawson kindly and

   

- me with all the support I needed to conduct the fieldwork. Without
. u d 'i.

>.. ‘
:xd, .

  

B ~£

;_ gmchmitwouldhavebeenabaolmelyimposaiblete

     

    

.
‘

 
 

  
   
  
   
 

of Agricultural Economics and Agribusiness also provided a variety of input
.1 My. Also, I greatly appreciate all the respondents who willingly participated 1n
IA}:

1,!er sacrificing their time. I would also like to thank the Ministry of Food and
. ‘0 l

W for generously providing the price data that they collected.

gill inn! truly thankful to Dr. Robert Dixon Phillips, Dr. James Lowenberg-DeBoer,
and Dr. Joan Fulton for their very helpful guidance and input at the early stage of this
Study. I am grateful to the Bean/Cowpea Collaborative Research Support Program and
through it, the United States Agency for International Development for allowing me to
conduct this study and gain invaluable experiences.

My thanks also go to Dr. John Holtzman for kindly sharing the report of the
reletrant study that he and his colleagues previously conducted.

I Want to express my extreme gratitude to my friends from West Africa, Kwami
Menu, Doe Adovor, Marthe Diallo, Nyadia Goita, Clarice Mensah, Patrick Ofori, and
Enyo Quiet. They were not only kind in spending their precious time for helping me to
m the questionnaires, but they also facilitated my stay in Ghana in a variety of ways.

‘ lam also really grateful to Subu Kumarappan, a Ph.D. student in the Department.
Hissharp eyes needed only one single glance to point out an important error that I had in

myeadier model. That was amazing.

 

lvmuld also like to acknowledge Dr. Ayumi Onuma, Mr. Francois Roussel, Mr.
J .

 

 

    

TABLE OF CONTENTS

  

1 i ................................................................................. I
no... .
r WMURES ................................................................................ 1
.WODUCTION ................................................................................... 1
1.1 Problem Statement ......................................................................... 1
1.1.1 Background ........................................................................ 1
4 . 1.1.2 Justification ........................................................................ 4
1.1.3 Target Products .................................................................... 6
rs; llObjectives ................................................................................. 10
M ‘ 1.2.1 General Objective ............................................................... 10
1.2.2 Specific Objectives .............................................................. 10
1.2.3 Research Questions .............................................................. 11
1.3 Organization of Thesis .................................................................. 13
CHAPTER 2
LITERATURE REVIEW .......................................................................... 14
2.1 Cowpea Subsector in Ghana and West and Central Africa ......................... 14
2.2 Development of Processed Cowpea Products in Ghana and West Africa 17
2.2.1 Kosei and Dry Cowpea Meal ................................................... 18
2.2.2 Cowpea-Fortified Weaning Foods ............................................. 21
2.2.3 Soybeans in Ghana .............................................................. 25
2.3 Enterprise Budgets—Purpose and Procedure ........................................ 25
CHAPTER 3
METHODS AND DATA .......................................................................... 28
‘ 3.1 Approach .................................................................................. 28
3.2 Survey Respondents and Instrument Design ......................................... 29
3.2.1 Street Vendors of Kosei ......................................................... 31
n t . 3.2.2 Custom Millers ................................................................... 32
3.2.3 Retailers ........................................................................... 34
i. j 31.2.4 Weaning Mothers ................................................................ 38
3 2.5 Industrial Grain Flour and/or Weaning Food Processors ................... 39
3.3. Secondary Data ........................................................................... 42
‘ - 3&412fl Cowpea Meal and Kosei Preparation Experiments ............................ 43
J. Weight of Difi'erent Crops ................................................ 44

Methods ...................................................................... 46
Cowpea Meal as an Ingredient 1n Kosei ............................. ....47
Weanimix

 

 

 

 
 
 

    

   
   

‘ . 1 ‘MEAL FOR PREPARATION OF KOSEI
.» 111.1 ANALYSIS ..................................................................... 51
' 4.1 Ovarview of Kosei Business and Grain Flour Production in Accra ............... 51

4.2 Dmpfion and Implication of Data ................................................... 55
4.2.1 Street Vendors of Kosei and Agawu .......................................... 55

“4.2.2 Custom Millers ................................................................... 78

4.2.3 Industrial Cowpea/Soybean Flour Processors ............................... 83

4.2.4 Retailers ........................................................................... 90

4.2.5 Price of Cowpeas ................................................................ 95
4.2.6 Dry Cowpea Meal and Kosei Preparation Experiments ................... 100

4.3 Summary ................................................................................. 111

CHAPTER 5
DRY COWPEA MEAL FOR PREPARATION OF KOSEI

_ -BUDGETING AND SENSITIVITY ANALYSIS ........................................... l 15
5.1 Budgeting Analysis—Model .......................................................... 115
5.1.1 Returns to Meal Processors ................................................... 115

5.1.2 Returns to Kosei Vendors ...................................................... 118

5.2 Budgeting Analysis—Results ......................................................... 120
5.2.1 Budgets for Producing Cowpea Meal ....................................... 120

5.2.2 Budgets for Preparing Kosei from Wet-Milled Cowpeas ................. 124

5.2.3 Budgets for Preparing Kosei Using Dry Cowpea Meal ................... 126

5.3 Sensitivity Analysis ..................................................................... 129
5.3.] Change in the Technical Efficiency of Processing Dry Cowpea Meal .. 131

5.3.2 Change in the Volume of Production ........................................ 134

5.3.3 Bulk Purchase of Dry Cowpea Meal by Kosei Vendors .................. 140

5.3.4 Change in the Price of Cowpeas ............................................. 143

5.3.5 Change in the Retail Margin .................................................. 146

5.3.6 Combination of Different Scenarios ......................................... 147

5.4 Summary ................................................................................. 152

CHAPTER 6

WEANIMDi—DESCRIPTIVE ANALYSIS .................................................. 154
6.1 Overview of Ghanaian Weaning Foods ............................................. 154

6.2 Description and Implication of Data ................................................. 156

' 6.2.1 Prices of Different Weaning Foods .......................................... 156

6. 2. 2 Weaning Mothers ............................................................... 158
., . . 6.2.3 Industrial Weaning Food Processors ......................................... 167
1 ' a ' 1‘16. 2. 4 Cost of Custom Milling ....................................................... 175
.1 . ’ 1m Retailers ......................................................................... 176

.. .6 Representative Prices of Maize, Cowpeas, Groundnuts, and Soybeans

 

  
  
  

“1 +—£UDGETING AND SENSITIVITY ANALYSIS ......................... 186

“ I1 ADRIYSlS—MOdel .......................................................... 186
7.1.1 Estimated Prices of Industrially-Processed Tom Brown and Weanimix
on" 1.1.2 Estimated Prices of Self-Prepared Tom Brown and Weanimix .......... 188
7.2 Budgeting Analysis—Results ......................................................... 190
NW1 “N 7.2.1 Processor Price Estimates of Tom Brown, Cowpea-Weanimix, and
Mrv- :1 ~ . Soybean-Weanimix ............................................................ 190
7.2.2 Representative Budgets and Retail Price Estimates of Tom Brown,
*1" '1 1 Cowpea-Weanimix, and Soybean-Weanimix ............................... 194
CI «1 1.1.3 Sensitivity Analysis ..................................................................... 200
7.3.1 Change in the Technical Efficiency of Industrial Processing ............ 201
7.3.2 Change in the Volume of Industrial Production ............................ 202
7.3.3 Change in the Prices of Raw Materials ...................................... 205
7.3.4 Change in the Retail Margin .................................................. 210
7.3.5 Combination of Different Scenarios ......................................... 213
7.4 Summary ................................................................................. 217
CHAPTER 8
CONCLUSIONS .................................................................................. 220
8.1 Summary ................................................................................. 220
8.2 Policy Implications ..................................................................... 224
8.2.1 Dry Cowpea Meal for Preparation of Kosei ................................ 224
8.2.2 Cowpea-Weanimix ............................................................. 225
8.3 Limitations .............................................................................. 227
8.4 Future Research ........................................................................ 230
8.4.1 Dry Cowpea Meal for Preparation of Kosei ................................ 230
8.4.2 Cowpea-Weanimix .............................................................. 232
APPENDIX 1
Miniatures ..................................................................................... 235
A.1.l Questionnaire for Street Vendors of Kosei ....................................... 235
A.l.2 Questionnaire for Processors of Cowpea or Soybean Flour, Weaning
I Foods and/or Gari ................................................................... 245
A13 Questionnaire for Custom Millers ................................................. 273
11.1.4 Questionnaire for Retailers ......................................................... 278

 

3.1.5 Questiomiaire for Weaning Mothers .............................................. 289

 

 

[-1 urn-Inn

 

II! 1X11 ‘
Mm Calculation ‘10

       
   
  

. . .. ' .MothersasConsmners of Kosei 312
4.4 1 .

‘ 21: “I II 8

ortifiedGari ........................................................................... 314

mamas .................................................................................... 317

.TIth'

Table ‘1 :3

lelr. 1 ‘ " .

labia -‘.

15bit»: . . 1

 

 

    
 

4.111

mists

Table 4.4
Me 4.5
Table 4.6

Table 4.7
Table 4.8

Table 4.9
Table 4.10
'litble 4.11

Table 4.12
.151,104.13

W414

 

\ ,

LIST OF TABLES

» Nutrient content of cowpeas and soybeans (grains, dried) ..................... 25
Characteristics of kosei and agawu vendor respondents ........................ 56
Types and sources of cost components to prepare kosei/agawu
among the respondents .............................................................. 58
Product price and daily sales of kosei/agawu among the respondents ....... 61
Daily shortage and leftover of kosei/agawu among the respondents ......... 68
Seasonality in kosei/agawu business among the respondents .................. 72

Experience and interests of the respondents in commercial dry
cowpea meal/flour ................................................................... 74

Characteristics of custom miller respondents .................................... 78

Milling charges for dry and wet cowpeas among the 15 custom

miller respondents ................................................................... 80
Representative custom milling/dehulling charges for cowpeas ............... 81
Characteristics of cowpea/soybean flour processor respondents .............. 84

Equipment for producing dry cowpea meal and possession by
the respondents ...................................................................... 87

Mean retail margin for milled grain products among small shop
and supermarket respondents (%) ................................................. 92

Mean retail margin for milled grain products combined with the
VAT/N HIS among small shop and supermarket respondents (%) ............ 95

Representative cowpea price paid by processor respondents in February
2007 and estimated range in price fluctuation during the year ................ 99

Ram 11t80fthe wet- and dry-milled kosei preparation experiments .......... 102
-. 11,-. ..
‘ W of the kosei preparation experiment (wet-milling) 1117

Wmmfthekoaeipreparationemperimuns

 

.. use . w... »
-. 1... ' 1,,
‘0 I ' V

 

    
  

Brew-output conversion rate across kosei vendor respondents
using the wet-milling method .................................................... 109
' - ' ‘ .... 111

4 19 Dry equivalent rates of the kosei preparation expenment (dry milling)

Me 5.1 Estimated processor price, cost, return per kg of cowpea meal .............. 121
me 5 2 Representative budgets for preparing 1 kg of wet-milled kOSCI (¢) ......... 124
Ibblc 5 3 Budgets for preparing 1 kg of kosei using dry cowpea meal ................. 127

Table 5 4 Ratio of the total difference in returns between wet- and dry-milled

kosei to the opportunity cost of labor per day .................................. 128

Table 5 5 Sensmvity analysis (1 ): Technical efliciency—budgets for cowpea

meal processors (¢ per kg of meal) .............................................. 132

Table 5 6 Sensitivity analysis (1): Technical efficiency—budgets for kosei

vendors (per kg of kosei) ......................................................... 133

Table 5 7 Sensrtrvrty analysis (1): Technical efficiency—Ram of the total

difl'erence in returns between wet- and dry-milled kosei to the
opportunity cost of labor per day ................................................ 134

Table 5 8 Sensrtivity analysis (2): Increase in the volume of production

—-budgets for cowpea meal processors (1! per kg of meal) ................... 137

Table 5 9 Sensrtivity analysis (2): Increase in the volume of production

—budgets for kosei vendors (per kg of kosei) ................................. 138

Table5 10 Sensitivity analysis (2): Increase in the volume of product1on

—-Ratio of the total difference in returns between wet- and
dry-milled kosei to the opportunity cost of labor per day ..................... 139
Table 5 11 Sensitivity analysis (3): Bulk purchase of meal by kosei vendors
—budgets for cowpea meal processors or per kg of meal)” .. . . 141
I.
“5.12 Sensitivity analysis (3). Bulk purchase of meal by kosei vendors
—budgets for kosei vendors (per kg of kosei) ................................. 142

1111-. i3

 

Wandysis (3): Bulk purchase of meal by kosei vendors
mo of the total difl‘erence' 1n returns between wet- and

.1 gang‘s-1‘1 m 1.

  

edkoseitotheopportumty cost oflaborperday ..................... 143
. -"."Vf<-"-b)¢Ur" ‘ .

 

 

Table 5.17

Table 5.18

Table 5.19

Table 5.20

Table 6.1

Table 6.2
Table 6.3

Table 6.4

Table 6.5

Semitivity analysis (4): Change in the price of cowpea grain
w-bridgets for cowpea meal processors (¢ per kg of meal) ................... 144

Sensitivity analysis (4): Change in the price of cowpea grain
-——budgets for kosei vendors (per kg of kosei) ................................. 145

Sensitivity analysis (4): Change in the price of cowpea grain
~—Ratio of the total difference in returns between wet- and
dry-milled kosei to the opportunity cost of labor per day ..................... 146

Sensitivity analysis (5): Change in the retail margin—estimated
retail price of dry cowpea meal ................................................... 147

Sensitivity analysis (6): combination of different scenarios
—-budgets for cowpea meal processors (¢ per kg of meal) ................... 148

Sensitivity analysis (6): combination of different scenarios
—budgets for kosei vendors (per kg of kosei) ................................. 149

Sensitivity analysis (6): combination of different scenarios
—Ratio of the total difference in returns between wet- and

dry-milled kosei to the opportunity cost of labor per day ..................... 150
Mean prices of different types of weaning foods observed in

Accra during February and March 2007 ........................................ 157
Weaning habits of the respondents ............................................... 159
Most often purchased weaning foods selected up to three by

the respondents ..................................................................... 163
Willingness to pay for cowpea- and soybean-fortified gari

among the respondents .............................................................. 164
Characteristics of weaning food processor respondents ...................... 168

Equipment for producing Weanimix and possession by the
respondents ......................................................................... 172

Representative maize, cowpea, groundnut, and soybean prices

' paid by processor respondents in February 2007 and estimated

‘ . range in price fluctuation during the year ....................................... 181

by cost component for industrially produced 1 kg of
Wwemhmmandsoybean-Weanimix .................... 191 . , ‘1

 

Representative budgets to produce 1 kg of Tom Brown, cowpea-
Weanimix, and soybean-Weanimix and estimated retail prices .............. 196

Increase in the volume of production and change in the budgets of
industrially-produced weaning foods ........................................... 204

 

Price fluctuations of raw materials and estimated difference in retail
Ftp; ~- prices of industrially-produced cowpea-Weanimix and Tom Brown ....... 208

[$97.5 Estimated changes in the price of cowpea-Weanimix self-prepared
by grain/flour-type product vendors and the estimated price of
cowpea—Weanimix industrially-produced when the cost of raw
materials are the lowest ........................................................... 210

Table 7.6 Estimated smallest and largest differences in retail prices of
industrially-produced cowpea- and soybean-Weanimix ...................... 214

Table 7.7 Estimated smallest and largest differences in retail prices of

industrially-produced cowpea-Weanimix and Tom Brown .................. 215
Table 7.8 Estimated lowest and highest retail prices of industrially

-produced cowpea-Weanimix .................................................... 216
Table A.2.1 Olonka-kg conversion rates ...................................................... 300
Table A.2.2 Rubber cup-kg conversion rate .................................................... 300
Table All Representative price per kg of raw materials .................................. 301
ThbleA.5.l Cost and sale per day of kosei business ......................................... 305
TableA.5.2 Cost and sale per kg and representative budgets of kosei business ......... 309

TableA. 7. 1 Frequency of eating kosei, knowledge on how to prepare kosei,
. and experience in purchasing cowpea flour among weaning
mother respondents ................................................................ 312
“I 'Z. .
MAZZ Home preparation of kosei and interests in dry cowpea
new flour/meal among weaning mother respondents ............................... 313

 

m: 1 Budgets for producing 1 kg of cowpea-fortified gari (¢) ..................... 314

 

 

    
 

LIST OF FIGURES

Cowpea grain .......................................................................... 2
. Dish with cowpeas 1 .................................................................. 2
Dish with cowpeas 2 .................................................................. 2
1 Figure 1.4 Cowpea fritters ........................................................................ 2
It. .'
Figure 1.5 Map of Ghana ......................................................................... 5
Figure 1.6 Competition between wet plate-/hand-milled cowpea paste and
dry cowpea meal ...................................................................... 8
Figure 1.7 Competition among Tom Brown, cowpea-Weanimix, and
alternative fortifier-Weanimix ...................................................... 10
Figure 2.1 Map of regions in Ghana ............................................................ 15
Figure 2.2 Procedure for making a functional (suitable for making kosei)
cowpea meal .......................................................................... 19
Figure 2.3 Recipe for making Tom Brown and Weanimix .................................. 24
Figure 3.1 Districts of the Greater Accra Region ............................................. 30
Figure 3.2 Milling machine (plate mill 2-A type) used by a custom miller in
Ga East District ...................................................................... 33
Plane 3.3 A grain/flour-type product vendor stall at a market in Accra
Metropolis District .................................................................. 35
Figure 3.4 A small shop located in the Accra Metropolis District ......................... 36
Blue 3.5 A supermarket located in the Ga East District ................................... 38

Olonka container (left) and margarine tin (right) ............................... 45

 

 

 

Piano 4.8

Figure 5.1

Figure 5.2

Figure 5.3

Figure 6.1
Figure 6.2
Figure 6.3
Figure 6.4
Figure 6.5

Figure 6.6

  
 

. Hendsealer ........................................................................... 88

Collection of VAT/N HIS ............................................................ 93

Reported/observed cowpea prices in Accra in March 2007 .................... 96

5 Real monthly cowpea wholesale and retail price at urban markets

of the Greater Accra Region, Ghana (average of Accra, Ga, & Tema
Districts) (2002-2006) ............................................................... 98

Seasonal index of cowpea wholesale prices at urban markets
of the Greater Accra Region, Ghana (average of Accra, Ga, & Tema
Districts) (2002-2006) ............................................................... 99

Estimated processor price, cost, return per kg of cowpea meal
(Resps 1, 2, and 3) .................................................................. 122

Relationship between the purchase price of dry cowpea meal and
the return for kosei vendors ...................................................... 130

Relationship between the price of dry cowpea meal and the

difference in returns between wet- and dry-milled kosei ..................... 151
Fermented maize dough sold in a market ....................................... 155
Tom Brown sold in a market ...................................................... 155

Ingredients that respondents use to make self-prepared weaning foods 160
Roasting pans ....................................................................... 172
Roaster .............................................................................. 172

Real maize, cowpea, and groundnut monthly wholesale prices at

urban markets of the Greater Accra Region, Ghana (average of Accra,

Ga, & Tema Districts) (2002-2006) and prices of soybeans reported

by a wholesaler in Nima market (inflation adjusted) .......................... 178

Seasonal index of maize, cowpea, and groundnut wholesale prices
at urban markets of the Greater Accra Region, Ghana (average of Accra,
Ga, & Tema Districts) (2002-2006) ............................................. 180

I A ' . ~ Relationship between the high price of maize and the prices of

., t and groundnuts in the same month .................................... 182

 

  
   
  
  
  

 

2. 7' Estimated retail prices of industrially-produced (with BB budget) and
.‘ , selfiprepared weaning foods for difierent levels of input-output ratio

" . ’1 W retail prices of cowpea-Weanimix produced by LB

processors, HB processors, and grain/flour-type product vendors for

202

1.1 triflerent levels of retail margins ................................................. 212

g ' . ~ _ 1:815di for producing 1 kg of cowpea-fortified gari ........................ 315

the mou m M. m

V . ‘ V -

 

 

ilill

 

 

 
      

CHAPTER 1

INTRODUCTION

 

“a Miriam Statement
r 1-1-1 m

Cowpea (Wgna unguiculata L. Walp.) is an indigenous African annual legume
which is also called southern pea, blackeye pea, crowder pea, lubia, niebe, coupe or
ii‘ijole (Davis et al., 1991; Langyintuo et al., 2003). Its tender green leaves, immature
pods and green seeds, as well as dry mature seeds, serve as human food, and after the
harvest of the pods the rest of the plant can be used as animal fodder (Davis et al., 1991;
Mutational Institute of Tropical Agriculture [IITA], 2006). The grain contains around
24% protein and about 62% carbohydrates (Lambot, 2002). The plant is more drought
tolerant than common bean (Davis et al., 1991) and can be grown on and improve poor
soils (TITA, 2006). Even though the plant is very susceptible to pests and diseases (II'I‘A,
2006) and the grain to weevils (Lambot, 2002), cowpeas are important in many parts of
West and Central Africa as a source of protein for those who cannot afford meats or fish
(ETA. 2006), as “a food security crop” (Lambot, 2002, p. 368) for populations that
commie cowpeas as a traditional staple food, and as “a major cash crop” (Langyintuo et
$30011). 215).
. ‘ thiWest and Central Africa, cowpea grain is used for a variety of dishes; the whole
a w eaten with cereals or used as an ingredient of soups or stews, while milled
ll M igmake fritters or steamed cakes (Langyintuo et alum

-. "

com fritters are “the most common cowpea

 

i .. IN.‘
~

 

l

3 . -. .. - t t- .
.r._-. - t n-..m “k I Irvin. filHVnhhrtflrblc

 

throughout Africa” (Dovlo, Williams, & Zoaka, 1976, p. 30). In Nigeria, for example, “in
addition to home preparation, akara [cowpea fritters are called akara in Nigeria] is
prepared and sold by street vendors in the marketplace and by small-scale processors for
home delivery and catering services” (McWatters, 1986, pp. 13-15), while Fulton (2004,
p. 10) reported that in Dakar, Senegal, “most households do not prepare akara at home,

but rather purchase it from street vendors when they want the product for family

consumption, or for entertaining.”

Figure 1.1 Cowpea grain Figure 1.2 Dish with cowpeas 1

Source: Author.

 

 

 

   
 

'. ' mm COWpeas are popularly consumed at home and as a street food,
processing of cowpeas is negligible. Only a very limited use of cowpeas for
m crackers, flours, and weaning foods has been observed in Senegal and Ghana
”Cowpea Collaborative Research Support Program [B/C CRSP], 2006a; Langyintuo
$11., 2003). For the goals of enhancing cowpea consumption, utilization, and food
acuity, many studies have been conducted by B/C CRSP food scientists at the
University of Ghana-Legon (UGL) and University of Georgia (UGA) to develop
limitious and afl'ordable cowpea-based processed productsl (B/C CRSP, 2006a, 2006b;
Phillips et al., 2003). Cowpea-based processed products seem to have high potential
became currently home- or artisanally-prepared cowpea products have tastes preferred by
West African consumers, such as the light texture of fritters, which is difficult to produce
With other grain legumes. Furthermore, scientists working on the B/C CRSP have
hypothesized that the growing urbanization and the increase in opportunity cost of
women’s time in West Africa will increase the demand for processed products in the
region (B/C CRSP, 2006a, 2006b).

However, various constraints to creating and promoting such processed COWpea
Mutts have been identified (B/C CRSP, 2006a). Constraints identified by the UGL and
UGAbam, as well as Lambot (2002) include higher prices of cowpea grain compared to
lfimbstimtes, lower protein content than soybeans (which are substitutes for cowpeas in
‘. .hwducflon of some of the processed products), fluctuations in price and quality, lack

" ility, and possibly poor functionality of processed products.

~ mmwhfihrfifrmmmg_ , ‘

 

 

the above-mentioned constraints to creating and promoting cowpea-based

   

products, price-related constraints seem to have received less attention than
out. However, if the price of cowpea-based products is high compared to their
”helium and potential customers are not willing to pay enough of a premium for the
m of cowpea-based products to let these products overcome that price difference,
then We products will never attract enough customers in the market to become
profitable. The importance of price-related constraints seems to be becoming even greater
today because of the movement in coastal countries of West Africa to promote
production and increase consumption of soybeans, a potential substitute for cowpeas in
processed products.

Another issue with regard to price-related constraints is how the difference in
techniques of processing cowpeas affects the cost of production. When alternative ways
exist to process cowpeas into the same end product, we need to examine which way is the
most cost-effective.

With regard to non-price-related constraints, extensive studies have been
conducted on the functionality of different cowpea-based processed products and the
assessment of consumer acceptance and potential demand for such products by the BIG

CRSP food scientists and economists (for recent studies, see B/C CRSP, 2006b).

fibwev" er, fewer studies have been conducted on non-131309430“5t?aims at the processors’

 

 

 

 

Figure 1.5 Map of Ghana

a i r . ' ‘ . ‘7 ' ,~
,4 ._. , ; ‘ i ." . I
80, ._ , 1 :\ 339mg:
;. - fl. ‘~.‘ ‘.

"‘- .4 .Wu i j/mur/

I/YenriTI. ,1, ‘ ,
lkf-j' 'ijale {I H? “BJE IN

Hmong ..,,, ,3) _ . ,
AS5139» .2313}:
. , C )I‘EHAI‘NHA gala/gill! ; I ..

“Ow/V?”
a:

 

 

 

 

 

 

   

 

 

' t ,a r as An ' a Seknnal- (I 150 mi
m
0 998. Eneyelo audio Britannica, Inc. ' “I 0 200 km

 

 

 

 

Source: htt ttp://www.britannica.com.
consumers in many other West African countries. Ghana was a major consumer of
cowpeas (the second biggest importer of cowpeas after Nigeria, recording the largest
percentage deficit [67%] during 19905 [Langyintuo et al., 2003]). The Ghanaian per
capita income level was in the middle in the West African region, with an annual average
of about US$350 during 1995-2004 (United Nations Statistics Division [UNSD], 2005),
and Ghana experienced a strong growth rate of 2.2% per year during 2000-2004 (World
Bank, 2006). This suggests that the opportunity cost of women’s time, especially in urban
areas, is higher than that in lower-income countries, thus opening the potential scope for
economically viable commercial processing. Also, a relatively high level of income

Wine necessary for consumers to adopt more nutritious but at the same time more

 

infl‘

 

   
  

' cowpea-based processed products. In addition, a higher education level in
Macy rate of 70.7% among the population of 15-24 year-olds during 2000-2004
Cm, 2006]) is expected to have a positive effect on women’s decisions to purchase
Wis that have greater nutritional value. Finally, B/C CRSP scientists have conducted
my studies in Ghana on cowpea-based processed products, including weaning foods

(Phillips et al., 2003). As a result of these previous studies, there exist data related to

processed cowpea products in Ghana.

1. 1.3 Target Products

The purpose of this study is to analyze the competitiveness of a few selected
industrially-processed2 cowpea products in Ghana, focusing on both price-related and
non-prioe-related factors. With few cowpea-based processed products currently available
in the Ghanaian market, ready-to~use dry meal and cowpea-fortified weaning foods were
selected3 as promising products, based on personal communications with agricultural
economists and food scientists associated with the B/C CRSP’s West Africa regional
project team (J. Lowenberg-DeBoer and J. Fulton, personal communication, June 15,
2006; R. D. Phillips, personal communication, August 22, 2006).

Dry cowpea flour/meal is a primary processed product that can be used as an

Went in further processed products. The difference between flour and meal is that

f“ ’11» term “industrial processing” in this thesis does not necessarily mean processing activities by
. <W With large scale personnel and/or capital. Rather, it includes processing activities by companies
‘ _ with (Le, minimum size is one single person) meeting the following criteria. (1) outputs have a
“ ll! quality and unit quantity; (2) outputs have a relatively long shelf life; and (3) output: are
a Wholesale basis to buyers who m tum resell them to others.
' V W (grated, fermented, and roasted cassava) was also among the suggesmdwgu
Mr -However,duringthefieldsm'veyin0hana,itwasfoundoutthntgari
. . “mouths
,ammhfihm Hm _ _ , _p
' ;. in» am _ - .27".

 
  

 

 
 

 

 
   

" we ._

are

I

mm. mathemtfl H, :~.

.1“-

  

  
  
   

   
 
  
  

 

milled than meal. Studies by the BIO CRSP’s West Africa regional

1‘“ . '

 

: ' ' have found that dry cowpea flour/meal can be used for fortification of other

‘ , . and that cowpea meal processed in a certain way is suitable for making

'fim (called kosei in Ghana“). If dry cowpea meal is used to make kosei, it
. , ' ' g _with wet plate- or hand-milled cowpea paste that housewives and street vendors

” .. currently use (Figure 1.6).

.h m phen- “WM-lee. archer m .I We?
, ',., hm: ,. . '. ”r“ 3“!“ fi" 7 ‘

 

 

   
 

V ‘dryeowpeameal

Kosei vendors/Housewives

Preparation

'- iGompetition between wet plate-lhand-milled cowpea paste and

Industrial
processors

 

 

cowpea
grain

 

processing

 

 

 

 
 

 

Qfle—"I Using on
figuration Meal
cowpea
meal

 

 

 

 

cowpea
meal

 

 

 

 

 

 
 

add water

preparation

 

 

  
   

 

 

 

 

 

 

 

——V

  
 

Weaning foods are among the products that could be fortified with cowpeas
because most of the traditional home-prepared weaning foods popular in Ghana are
maize-based porridges that are relatively low in protein (E. Sakyi-Dawson, personal
communication, December 19, 2006). Fortification could be done by simply adding
cowpea flour to traditional weaning foods while preparing them, or by developing
processed weaning foods containing cowpeas as an ingredient so that the product
achieves a target nutrient content. One such weaning food is Weanimix, introduced in
Ghana in 1987 by the Ministry of Health Nutrition Division and the United Nations
Children’s Fund (UNICEF) (Lartey, Manu, Brown, Peerson, & Dewey, 1999). Weanimix
is a fortified product of Tom Brown, a Ghanaian traditional weaning food made from
roasted maize, and “composed of 10—15% soybeans or cowpeas, 10% groundnuts
(peanuts) and 75—80% maize (corn)” (Lartey et al., 1999, p. 391). Using this formula,
weaning mothers can prepare Weanimix at home, while its industrial processing is also
possible. Difi'erent types of commercial Weanimixes (including “quasi-Weanimix”
Pmducts made with ingredients or composition of ingredients different from the standard
formula) were found during fieldwork conducted in Accra during February and March
2007. As reported later, the majority of those products contained soybeans rather than
W. In order for cowpeas to be successfully used as an ingredient in Weanimix,
°°WP°83 would likely have to compete with maize (the major ingredient of traditional
Tom Brown, a part of which cowpeas would replace) and with alternative sources of

- ammo“ likely soybeans, in terms of different criteria such as price of grains, tastes,

 

 

 
 
 

, m 1.7 Competition among Tom Brown, cowpea-Weanimix, and

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    
 
   

alternative fortifier-Weanimix
Cowgea- Alternative F ortifier—
m Brown Weanimix Weanimix
‘ e maize o maize o maize
o cowpeas 0 alternative
0 groundnuts fortifier
(e.g., soybeans)
o groundnuts
processing processing processing
Tom Cowpea— Alternative
Brown Weanimix Fortifier-Weanimix
1.2 Objectives
1.2.1 General Obiective
The general objective of this study is to determine under what conditions cowpeas
W01!“ be an economically competitive ingredient in selected industrially-processed food
products in Ghana.
1.2.2 fiecfl Obiectives

 

 

 

.aliitmnpecific Objectives are to:
, 1’ “ representative budgets for industrially-processed ready-to-use dry

:13; ingredient in kosei) and cowpea-

- 4 no AC1“ - ..:.. “u‘ L

\.

 

 

  
  

item the price gaps that those products would need to overcome to be

competitive with their substitutes in Ghana;

b) Document current/potential non-price-related factors that would aflect the
competitiveness of ready-to-use dry cowpea meal and cowpea-Weanimix in
Ghana, with a special focus on the processing level (e.g., regular availability of

grain, difference in processing procedures, and packing and storage).

1.2.3 Research Questions

To achieve those specific objectives, the following research questions were

explored:

0.1) Price competitiveness 0t dug cowpea meal as an ingredient in kosei

i. What are the cost components and how much does each component cost for

preparing a unit of kosei using the wet-and-plate-milled5 cowpeas?

ii. What are the cost components and how much does each component cost for

industrially processing a unit of ready-to-use dry cowpea meal suitable for

making kosei?

 

 

 

fi

    
 

the current production and marketing costs, how competitive would

1 Wally-processed ready-to-use dry cowpea meal be, compared to wet-and-

plnte-milled cowpeas?

v. How does the competitiveness of ready-to-use dry cowpea meal change when

the value of cost components changes?

a-Z) firige competitiveness at cowgea— Weanimix

i. What are the alternative processing procedures (including technologies and
scales) and cost components for producing and marketing (e.g., ingredients,
labor, equipment, and transportation) cowpea-Weanimix, Tom Brown, and

soybean-Weanimix?

ii. How much does each component cost for processing a unit of cowpea-

Weanimix, Tom Brown, and soybean-Weanimix?

Given the current production and marketing costs, how competitive is

y...

ii .
industrially-processed cowpea-Weanimix, compared to its substitutes?

\t.

t 1‘? “in. How does the competitiveness of industrially-processed cowpea-Weanimix

 

. ‘ "Mileage when the value of cost components changes?

 

fi

  

W competitiveness

i. What non-price-related factors, especially at the processing level, affect the
competitiveness of industrially-processed ready-to-use dry cowpea meal and
cowpea-Weanimix in Ghana (e.g., regular availability of grain, difference in

processing procedures, and packing and storage)?

1.3 Organization of Thesis

The thesis organization is as follows: Chapter 2 reviews the literature on the
ODWpca subsector and development of processed cowpea products in Ghana and West
Africa, as well as the use of budgeting methods to analyze the profitability of productive
activities. Chapter 3 describes the approach, instruments, and analytical methods used for
this study, as well as field data collection conducted in Ghana. Chapter 4 presents
descriptive analysis of the data collected on dry cowpea meal, discusses the findings
about the non-price-related constraints, and describes the steps taken to prepare the data
for the budget analysis. Chapter 5 presents the model and results of the budgeting and
sensitivity analysis carried out to assess the economic viability of industrially producing
dry 00pr meal. Chapter 6 presents descriptive analysis of the data collected on
”ma-Weanimix, discusses the findings about the non-price-related constraints, and
describes the Steps taken to prepare the data for the budget analysis. Chapter 7 presents

‘10 MOdel and results of the budgeting and sensitivity analysis carried out to assess the .

     
 
    

WC Viability of industrially producing cowpea-Weanimix. Finally Chapter 8,

 

- H w : of the thesis, identifies policy implications and limitations of this

 

 

 

,.

 

 

 

-
,.
. H;
.
- ‘
ucsub
<1
, ,
It
k
' .
u-..
u
--

 

a
-
'- ._
.
‘ a
.-
u.\
r
1
~.
-
'.
',
b.
.5.
1
.-
‘.

 

CHAPTER 2

LITERATURE REVIEW

 

2.1 Cowpea Subsector in Ghana and West and Central Africa

Today cowpeas are grown worldwide, with Afi’ica producing the largest share.
Langyintuo et al. (2003) estimated that 3.7 million mt of cowpeas were annually
produced on about 9.7 million hectares during the 1990S in Afi’ica, the Americas, Europe
and Asia, of which West and Central Africa accounted for 2.6 million mt (69% of the
world’s production) on 7.8 million hectares (80% of the world’s harvested area). As
individual countries, Nigeria, the largest cowpea grower in the world, accounted for 45%
of the world production, followed by Brazil (17%), and Niger (10%) during the same
period. Although “cowpea is a warm-season crop well adapted to many areas of the
humid tropics and temperate zones” (Davis et al., 1991), the drier areas in West Africa
have a comparative advantage in producing legumes, compared to humid areas, because
pests and diseases make production in the humid areas more difficult (Langyintuo et al.,
2003). The primary cowpea production zone in West and Central Africa lies between the
300 and 1,000 mm annual rainfall isohyets or approximately between 10 and 15 degrees
north latitude (Langyintuo et al., 2003). In Ghana, cowpeas are mainly produced in the
Northem and Upper West Regions (see Figure 2.1) (Nimoh, 2004).

A special characteristic of the cowpea trade in Afiica is that the trade occurs

. $516!. that Sahelian countries producing a larger amount of My , = '.

.' ‘ ‘1'; Mufti“ .U‘ 'l 5861-. at C: 4 “’np‘fil“ W
, 'N'. ‘1" _. - . , _ ‘ n ' 4 _ . .7 V ’11

. .1 x.
.1 _u -

 

 

 

Source. h ://www ' ' mm "’ " ' manaif
tend to be net exporters, while more humid coastal countries tend to be net importers
(Langyintuo et al., 2003). An important exception is Nigeria, which is both the largest
producer and the largest importer of cowpeas in the region. Langyintuo et a1. (2003)
identified two largely independent cowpea trade zones. The first is the “Nigerian cowpea
smashed,” composed of Niger, Burkina Faso, Mali, Cameroon, Chad, Benin (net
exporters, listed in descending order by the average annual amount exported during
19903) and Nigeria, Ghana, C6te d’Ivoire, Togo, and Gabon (net importers, listed in
Mg order, except for Gabon, for which data are unavailable). The second is the

We“: cowpea grainshed, ” composed of Senegal (a net exporter during 19903) and
'1

“,4..." .1... , Gambia and Guinea Bissau (net importers). Per capita consumption of
“~‘ 1) ,)

«Quoted in Langyintuo et al. (2003) showed enormous differences among

lapeoplewereesumatedtohaveconsumedmkgofcowpeapen

59.1... l.’s'lr._tq

 

 

 

 

Ghana and Togo; and between 1.5 to 2.5 kg in Burkina Faso, Cameroon, Chad,

bd’l‘vnire, Mali, Mauritania, Niger, and Senegal.

Both publicly-managed and informal market places for cowpea grain exist in West
Id Central Africa (Langyintuo et al., 2003). Langyintuo et al. (2003) described typical
market channels for cowpea grain observed in Benin, Togo, Burkina Faso, and Ghana:
rural assemblers buy cowpea surpluses from farmers and sell them to small or large urban
molwalers or through commission agents; retailers buy from wholesalers or commission
agents or sometimes at harvest time directly from farmers and sell to consumers. The
price of cowpea grain is influenced by grain quality, time of selling, transport costs,
storage costs, market tolls and taxes, and sales location (Langyintuo et al., 2003). Storage
is important for maintaining grain quality, protecting them from pests and weevils
(Langyintuo et al., 2003). Price declines during and soon after the harvest season (i.e.,
October through January in Niger, Benin, Nigeria and Ghana) and increases in the
followulg months when grain is scarce; and geographically, as one would expect, lower
prices have been observed in production sites and in exporting countries, compared to
cities and importing countries (Langyintuo et al., 2003).

Cowpea varieties have different attributes, such as texture of the testa (skin), grain
Size, and skin and eye color (Langyintuo et al., 2003). The share of different varieties sold
inmate“ varies by country and even by market within the same country. This variation
ild'fected by the types of foods prepared and storage conditions (smooth skin texture

1.; ., Who preferred in humid areas with poor storage conditions) (Langyintuo et al.,

. w- , Ntoukam, Murdock, Lowenberg-DeBoer, & Miller, 2004). Consumer:

’ “mamas in most markets of Cameroon and Norm can; "

on '_: -
.‘ “- . ~ _.lk

. I~ .'.‘

v
1,»

 

    

rby Langyintuo et al. (2004), most markets of Southern Ghana studied by Mishili

wand all the markets of Senegal studied by Faye (2005).

Ali-i-.. .

ml;
an Development of Processed Cowpea Products in Ghana and West Africa

According to FAO (1997), food processing has a variety of positive impacts on
individual human nutrition and society as a whole—including: (l) improving digestibility
of foods; (2) diversifying diet; (3) transforming raw materials into products that are more
convenient to use; (4) extending the availability of foods beyond the area and season of
production, which increases food security; and (5) creating employment, thus improving
income of those who are involved. More than a decade ago, Smith (1995) reported that
there was a trend among African women to prefer convenience foods because of women’s
clunging role and participation in the workforce. However, today popular but
Wptible-to-weevil cowpeas are still mainly sold as grain and cooked at home in West
Africa (BIC CRSP, 2006a).

Through the B/C CRSP project, a variety of processed products using cowpeas as
an ingredient have been developed. Phillips et al. (2003) has summarized studies on these
Meta, focusing on different quality aspects of the developed products, including

Mica! composition, nutritional quality, physical and functional behavior, sensory

     
    
 

. m and consumer acceptance, and costs and impacts of technology adoption. The

 

 
   

' “developed and tested included: (1) baked products and snack chips (e.g., yeast-

' f .-.‘ '- muffins, buttermilk doughnuts, tortillas, and tortilla and corn chips) in

 

musedtoreplacetheirmainingredients; mutate-tn

'FL l; --—--¢-»'

‘1‘“). >-;_ (if; 4-

 
  

 

 

 

 

. .
.a
_. --
a
.I"
Q Q
0‘.
l
u
‘ I
.I
.
'.-.. .
-.-
‘ .-
u
. r.,
.. _
x _ ‘
.
. 4
.. _v
a
A.
's
5‘.
. '>
a_‘

 

 

    

   

' i 7 and (5) cowpea-fortified traditional foods including hausa koko (a spiced millet-
, -M drink) and gari (grated, fermented and roasted cassava). Among these products,
“alto-use dry cowpea meal and cowpea-fortified weaning foods were selected as the

Wt products of this case study, as mentioned in Chapter 1.

2.2.1 Kosei and D9; Cowpea Meal
Kosei are traditional cowpea fritters, which are popular as a breakfast or snack
food in West Africa (McWatters, Resurreccion, & Fletcher, 1990). The traditional method
of preparation described by Dovlo et al. (1976) involves (l) dehulling (meaning “remove
wed coats of’) cowpea grain by either the “wet method” (soak in water and dehull
GOV/peas by hand) or “combined dry and wet method” (break cowpeas into smaller bits
using a grinding stone, blowing off seed coats, and then soaking); (2) grinding into a
paste; (3) whipping to incorporate air; (4) adding chopped onions, peppers and salt; and
(5) frying as small balls in vegetable oil. These dehulling and grinding procedures are
time-consuming and labor-intensive (McWatters et al., 1990). Development of ready-to-
use dry 00me flour/meal, which can be turned into a paste by simply adding water, has
been proposed by B/C CRSP food scientists to help lighten the burden of kosei
Mutation. Many tests have been conducted to create a good quality flour/meal,
:- Wally suitable to make kosei, by examining how the quality of end products was
multilateral such'as the difference in the variety of cowpeas used, particle size of
, , dehulling method (wet or dry), presoaking time, and storage condition of

MeWattersg Oguntunde, & Phillips, 1996; KW

 

I Lll‘ “hm-J. .
V. i it} 3r "~‘1I(

gag; ..

 

fit

sonar.

 

This research led to the development of a mechanical procedure to produce a dry

mes meal, from which kosei very similar in taste to traditional kosei can be made

(Phillips et al., 2003). The detail of the processing procedure is shown in Figure 2.2. It is

important to note that the developed product is a meal, not flour. Another point to note is

tint the processing procedure for making this newly-developed meal requires drying the

h‘.

. Nimitz

Figure 2.2 Procedure for making a functional (suitable for making kosei)
cowpea meal

 

1. Briefly soak cowpea grain (dipping for about 1 minute is usually
sufficient).

2. Dry the grain at a temperature less than 60°C until the moisture
content is less than 10%, preferably 7-8%.

3. Loosen seed coats using a plate mill with the gap adjusted to crack
(not grind) the grain. This operation can be done using a mortar
and pestle by those ‘skilled in the art’.

4. Using either a readjusted plate mill or hammer mill with an about 3
mm screen, grind the cotyledons so that about 50% of particles are
larger than 400 pm in diameter. Particles larger than about 800 um
require longer soaking and possibly some wet grinding to produce
a smooth batter.

5. This meal can be hydrated for 30-60 min, whipped to incorporate
air, spiced with chopped or ground peppers/onions/garlic, and fried
in hot oil.

 

 

 

Source: R. D. Phillips, personal communication, August 22, 2656.

perspective theproduction'anduseofdrycowpeeflourdoesmteeemtobem
Ammvloeuwmimsmwmmpxmwwmmm .
. u-jdmeofchme, that r we: I”, ., -.
.. was: w communism jif '

 

 

     
  
 

which implies that this meal can only be produced by processors who have
‘ ' ' drying equipment—if the meal is to be produced on a regular basis without
*3 on solar drying.
1!" ’ In Senegal, Fulton (2004) and Faye (2005) conducted studies of the industrial
guessing of cowpeas into flour. In 1999, they found five small-scale processors who
fled cowpeas to make flour and weaning foods. However, in 2004, only one of these
processors was still producing cowpea flour, and none was still using cowpeas to produce
weaning foods. The processors reported several constraints with regard to cowpea flour,
including: (1) low demand (possibly and partly because of lack of promotion) and (2)
deterioration during storage (related to poor packaging). Also, a food processor
interviewed, who had not tried cowpea flour, mentioned the lack of drying machines and
high transportation costs (for selling products to vendors) as problems. These constraints
may also apply to any processors producing cowpea flour/meal. As for the use of cowpea
flour/meal for preparing kosei, a vendor interviewed reported that by adding baking
m, it was possible to make kosei from the commercial cowpea flour that tasted the
same as traditional kosei. However, while Fulton noted that there is a potential market for
mm flour as an ingredient in kosei, she noted that the cost of the flour must be
Wendy low and the taste of kosei acceptable. Finally, she noted that because making
. -hlnsing the traditional method was time-consuming and labor-intensive, the kosei

Rm becoming “extinct” in Dakar, where the Opportunity cost of WOmenis time

i.” Nimclh (2004) med, “0118 Other t0pics, aw . at 7
MW ‘ :3: '4‘ iii";- 7*' .‘ ’e‘, " if i

'39". with «new; r‘ ,2. -_ 5 _‘

 

 

     

, . margin analysis, using data collected on the daily revenue and costs of cowpea
- m businesses including kosei vending. The results showed that the average profit
fight—calculated as {[revenue] — [cost]} / [revenue]—in the six studied communities
fist-kosei business was 26.3%, ranging from a high of 35.8% in two communities in the
Greater Accra Region to a low of 18.6% in a community in the Volta Region. The present

thesis conducts a similar analysis and also analyzes the effect of the use of cowpea meal

(‘mstead of grain) on the profit.

2.2.2 Cowpea-Fortified Weaning Foods

Weaning foods are foods prepared for children in transition from breast milk to
solid foods (6 to 18 months old) (Phillips et al., 2003). Such foods have to be carefully
prepared to promote young children’s healthy growth. According to FAQ/World Health
Organization (WHO) and Heimendinger, Zeitlin, and Austin (as cited in Mensa-Wilmot,
Phillips, Lee, & Eitenmiller, 2003), a good weaning food should be: ( 1) nutrient dense;
(2) easily digestible; (3) of suitable consistency; and (4) affordable to the target market.
In developing countries, children are often given weaning foods that do not meet these
criteria Until the age of four to six months, the growth rates of fully breast-fed infants do
Int differ much between developing and developed countries, while unsatisfactory
M among children in developing countries commonly begin after this period. The
main reason for this difl‘erence is considered to be the lack of access to nutrient-dense

\ . foods, together with frequent infections (Lartey et al., 1999).

, .

. .\~‘.'. .
--)k"“p " ‘ i...
' ‘J

. i. H NiinthOM)imputesthevatueofdreveudctsm.m. 7 ~'..';~. ..
Hwy web/relay unnamedwhahucenmu its):
hum Sell-i .

 

  

 

 

Cereals are the main ingredients in most of the traditional weaning foods in West

'w. However, it is almost impossible for small children to meet their needs for
duties and proteins from the amount of these low-nutrient-density weaning foods that
by can digest (Nti & Plahar, 1995; Onofiok & Nnanyelugo, 1998). Many studies have
been conducted to improve the nutritional quality of traditional weaning foods while
maintaining adequate sensory and functional (e.g., digestibility) quality (Afoakwa, Sefa-
Dcdeh, & Sakyi-Dawson, 2004; Annan & Plahar, 1995; Bentley et al., 1991; Egounlety,
Aworh, Akingbala, Houben, & Nago, 2002; El-Habashy et al., 1995, 1997; Lartey et al.,
1999; Mensa-Wilmot, et al., 2003; Nti & Plahar, 1995; Plahar et al., 2003). In these
Studies, cereal-based weaning foods were often fortified with locally-available low-cost
ingredients in order to make them widely affordable. Legumc crops such as cowpeas and
soybeans have been frequently used as a source of protein. In its early stage, research by
food scientists to develop such weaning foods usually used linear programming methods
(0 identify a least-cost combination of potential ingredients that meet the target nutrient
content (El-Habashy et al., 1995; Hayes, Mwale, Tembo, & Wadsworth, 1995; Mensa-
Wilmot, Phillips, & Sefa-Dedeh, 2001b). This ensured that the new formulas were not
”biliary combinations of nutritious ingredients, but passed a systematic cost test.
HOWwer, the price of ingredients used in such analyses seems to be limited to a single set

fiflcptesentative prices (either average of the prices during a certain period or price at the
‘ “of purchase). Further economic analyses, such as the impacts of seasonal price
' r' not ingredients on the cost of preparation, would provide information needed to

" ‘ ' Walsets ofing‘redientsthatmightdifl‘er—dependingonthemnaof

     

 

    

   

ln'Ghana, cereals such as maize, millet and sorghum are traditionally used as the

flufin‘gredients of weaning foods (Plahar, Onuma Okezie, & Gyato, 2003). Among
withere are two types of maize-based popular weaning foods: one is a porridge made
from inlet fermented maize, and the other is a porridge made from roasted and dried maize
(E. Sakyi-Dawson, personal communication, December 19, 2006). The former is called
koko and the latter is popularly called Tom Brown. Since the main ingredient in both
porridges is maize, they tend to be low in protein. Also, since they are porridges with
high moisture content, their overall nutrient density is low as well.

In order to enhance the nutrients of these traditional weaning foods, new formulas
have been developed. The suggested method for fortifying koko is to add dehulled
cowpeas or soybeans to soaked maize to make a cowpea-fortified fermented maize dough
(CFMD) or soybean-fortified fermented maize dough (SFMD), from which fortified koko
can be prepared. As is the case with traditional fermented maize dough, both CF MD and
SFMD can be used not only for koko but also for various popular Ghanaian staple dishes
(Afoakwa et al., 2004; Plahar, Nti, & Annan, 1997; Sefa—Dedeh, 2005).

The method suggested for fortifying Tom Brown is to add roasted cowpeas or
soybeans and roasted groundnuts to roasted maize before milling. As introduced in
Chapter 1, fortified Tom Brown, originally developed by the Ghanaian Ministry of Health
in collaboration with UNICEF, was named Weanimix. This thesis focuses on the cost
lilllysis of Weanimix because the fieldwork conducted in Ghana (see Chapter 3) revealed

- . iflci ' '
“i ‘ .infinmation on the cost of production was more readily available for Weanimix than

 

 

Figure 2.3 Recipe for making Tom Brown and Weanimix

 

A. Tom Brown
1. Roast the maize grain
2. Dry-mill the grain to produce a flour
3. Cook the flour into porridge and add sugar

B. Fortified Tom Brown (Weanimix)

1. Roast maize, cowpea/soybean, groundnuts, and mix them
together (maize:cowpea/soybean:g-nuts = 75:15:] or
8:1:1)

2. Dry-mill the mixture to produce a flour

3. Cook the flour into porridge and add sugar

 

 

 

Source: Lartey et al. (1999); (1 A. Annor, personal communication, February 16, 2007.

Before Weanimix was first introduced in 1987, the Ministry of Health tried to
teach weaning mothers to add groundnut paste to porridge that they fed their children (J.
G Amarh, personal communication, April 2, 2007). However, mothers often did not
follow the instructions for using groundnut paste. Consequently, the Ministry decided to
make flour that already contained groundnuts, as well as another nutritious high-protein
legume (i.e., cowpea/soybean). According to Sefa—Dedeh (as cited in Bernsten, 1993, p.
29), mothers were first advised to prepare Weanimix by themselves. However, this
attempt also failed—mothers often did not follow the right formula. These experiences
led to the development of commercialized Weanimix, which had three major objectives:
(1) ifnprovement of the nutritional status of children; (2) income generation for the

producers of the products; and (3) reduction of drudgery for preparation at a household

_ ‘ $510.6 Amarh, personal communication, April 2, 2007).

 

.v. {2.1.3 Sagan: in Ghana
u w ,‘ Soybean is a legume crop that has a higher protein and fat content, but lower in
mollydrates than cowpeas. The nutrient content of cowpeas and soybeans is presented

infl'able 2.1 for comparison.
Table 2.1 Nutrient content of cowpeas and soybeans (grains, dried)

- hydrate,
M253“ Protein (g) Fat (g) total Ash (g)

(incl. fiber)

(calories)

 

As soybean is not an indigenous legume in Ghana, there are no Ghanaian '
traditional dishes that use soybeans as an ingredient (Mensa-Wilmot et al., 2001b).
However, the use of soybeans as an ingredient in weaning foods has been promoted by
the Ministry of Health (Mensa-Wilmot, Phillips, & Hargrove, 2001a). Mensa-\Vrlmot et
al. (2001b) reported that among 133 mothers interviewed in the Greater Accra Region,
65% used soybean flour for weaning their children, while they could not identify whether

soybeans were familiar to women in villages in the Central Region because there was no

word in the local language that meant soybeans.

 

‘ Mmhudgefingasabasicmolmanalyzetheprofitabilityofm.

 

 
     
 

.., mating, " issauicandmd Irv-alarm, “.1 ,:

  

 

     
   
 

2002, p. 1). It is a method that is usually taught in farm management courses,

 

‘ was general courses in business management. Harsh, Connor, and Schwab (1981, p.
HI) divided the analyses of farm business into three categories: (1) “descriptive
mlysis” (“What is ...?”); (2) “diagnostic analysis” (“What is wrong ...?”); and (3)
Native analysis” (“What would happen if ...?”). Budgeting is a technique to conduct
this predictive analysis, building on the first two analyses. “It is a method of estimating
on paper the dollars-and-cents efiects of changes in farm organization or production
practices before resources are actually allocated to such changes” (Crawford, 2002, p. 1).
Among different budgeting techniques, enterprise budgets state the income,
expense, and resource needs of one particular productive activity on a per unit basis (e.g.,
income, expense, and resource needs of corn production per acre) (Harsh et al., 1981, p.
190). The purposes of enterprise budgets include: (1) assessment of the profitability of
that enterprise (targeted productive activity) and (2) comparison of the relative
profitability of difl'erent enterprises (Crawford, 2002). Although budgeting in general is a
tool mainly used for forward planning, enterprise budgets can also be summaries of
actual costs and returns (Crawford, 2002). In this thesis, the tool is applied to assess the
profitability of: (l) industrially-processed dry cowpea meal per se; (2) kosei made from
industrially-processed dry cowpea meal, as opposed to the kosei made from wet-milled
WWpea grain; and (3) industrially-processed cowpea-Weanimix, as opposed to its
Substitutes.
. i To prepare enterprise budgets, a variety of information has to be gathered,
, " r the input/output relationships, that is, “the quantity and quality of inputs needed
ma or cum" (Hush at al.. 1981. p. 185). were as;

~_ c ..,
._‘~\. . ~ _. ,

". I.
' . .‘r' " ‘ I
~ ,1 ' ~ \
_s, a _ ,
.— A1.
»
r

 

\ n.» "
. v
. .
K ‘V.
. —
.
fl. ...
.
-‘ - '
, ..
I - ‘
- .—
— ‘ 1
' I
_ .
n -
-- .
(nu, .
. . g _,
a
I... .-
-.
‘ .
‘ -

--
‘iv

 

I S

   
   
 

‘ “ variables such as materials, labor, land, and equipment (Debertin, 1986, p. 295),
‘ A . variables to include depends on the purpose of analysis. Input/output relationship
~ ehange if a new production technology is applied. Among the sources of data on
W relationships are accounting records and experimental data. Another set of
information needed is the prices of inputs and output. Once all the information is
Med, enterprise budgets can be prepared, which are usually presented in a table
hm where calculated variables are shown in categories such as the value of output,
variable costs, and fixed costs (Crawford, 2002; Harsh et al., 1981, p. 177-193). In
Interpreting the results, caution is needed when accounting records are used as a source of
costs. For example, family labor is often not included in accounting records. In such a

case, the calculated profit figure of the enterprise includes a return to the family labor and

management (Debertin, 1986, p. 296).

 

 

 

CHAPTER 3

METHODS AND DATA

 

M Approach

First, B/C CRSP researchers at Purdue University, the University of Georgia
(UGA), and the University of Ghana-Legon (UGL) were consulted to obtain key
infomation needed to decide which products to analyze. Based on these discussions,
ready-to-use dry cowpea meal and cowpea-fortified weaning foods were selected as
target products for this case study. Once the decision was made regarding product choice,
additional information was obtained including the common procedures used for kosei
preparation by street vendors in Ghana, types of weaning foods that are currently
available in the market, and the scale of existing companies producing dried flour-type
products in Accra, Ghana. Finally drafi questionnaires for conducting a field survey were
prepared.

Second, during February and March 2007, fieldwork was conducted in the
Greater Accra Region, with support of the Department of Nutrition and Food Science,
UGL. The goal of the fieldwork was to obtain both quantitative and qualitative data to
Claim the competitiveness of the two target cowpea-based processed products.

Wired questionnaires were prepared for collecting data from five difi'erent categories

01mm); respondents: street vendors of kosei, custom millers, retailers, weaning mothers,

 

- _Ls-»n_

 

 

J. -_._.

-.-—u

“spur

 

.-
T I'.
u ,
\ N
I. ~

.
'I

m.
-
‘ v
5

'Q.

- n
.

.

. I
k‘ I
\,_

.

.

.,

.

 

 

    
 

«conducted to estimate the input-output ratios. Finally, weights of difi‘erent crops
W (a standard measuring container used in Ghana) were measured from samples
W at a market in order to convert prices recorded in these standard volumetric
measures to prices per kg.

Third, using these data, budgeting and sensitivity analyses were conducted to
calculate the price-competitiveness of industrially-processed dry cowpea meal and

cowpea-Weanimix under different scenarios.

32 Survey Respondents and Instrument Design

Respondents interviewed during the field survey were identified and selected with
the help of UGL researchers. The purpose of this case study is to use data gathered from
survey respondents to develop typical budgets to assess the profitability of various
processes and to use sensitivity analysis to estimate the range in the values of cost
variables over which the target products are price-competitive. Therefore, “purposive
(non-probability) sampling method” (Bemsten, 2005) rather than random sampling was
used to select respondents for each of the five groups.

The interviews were conducted in four out of six districts in the Greater Accra
Region: Accra Metropolis (the capital city of Ghana), Ga West, Ga East, and Tema
Municipal. The Greater Accra Region is bounded by the Gulf of Guinea on the south, the
W Region on the west, Eastern Region on the north, and Volta Region on the East

.,,\§lgme 2.1). Within the Greater Accra Region, Accra Metropolis District is bounded

" ‘7'»? '. 5; n of Guinea on the south, Ga West District on the west, Ga East District orrthe

    
   

   
  

are located to the east of Tema Municipal District (Figure 3.1). While the
Accra Region is the smallest administrative region of Ghana with a land area of
. 3,345 km”: (Briggs, 2004), it has the second largest population (2.9 million), following the
ma region (3.6 million), according to the 2000 census (as cited in Briggs, 2004).

hiru'

:9: Figure 3.1 Districts of the Greater Accra Region

 

 

 

 

 

Source: http://ww answers “ ' ’J‘ ‘ icrs-of-gmna.

All respondents were visited by the author and his assistant, George A. Armor
Who was a research assistant of the Department of Nutrition and Food Science, UGL.
When necessary, he translated the questions written in English into local languages (most
“the time, Twi) and translated the respondents’ answers from the local languages into

': Wish. The translation was done question-by-question, and the author wrote those

 

 

 

 

 

 

 

 

 

 

 

for the interviews, see Appendix 1).

3.2.1 Street Venglm of Kosei

Although no list of kosei vendors doing business in Accra exists, there are so
many vendors that it is easy to find one when someone wants to buy kosei. During the
fieldwork, vendors were observed along the main roads, relatively small roads in
populated residential areas, and on and around the UGL campus. The author and the
assistant targeted mainly the residential and market areas of the Accra Metropolis and Ga
East Districts, stopped when kosei vendors were found, and made sure that those vendors
were potential respondents by observation and sometimes by asking the vendors brief
questions.

Two criteria were used to screen the kosei vendors. First, only kosei vendors who
had at least five kosei balls remaining to sell (at the beginning of the interview) were
selected because we needed to buy those balls in order to weigh them and estimate the
average weight of one kosei ball made by each respondent (these data were later used to
estimate the number of balls the respondents sold during the day before the interview was
conducted). Because the size of kosei balls fluctuated from one to another, the number “at
least five” was decided for obtaining an accurate average weight, although there was no
scientific reason to choose the five kosei balls. Second, the vendors who were selling
food products that shared ingredients with kosei were avoided to the extent possible

because, depending on their way of preparing those products, it would be difficult to

31

 

.-s

fl.

.-

| .-
~

-.-r

,4

hi

.. ,

.v~.

 

 

accurately estimate the share of kosei in the payment for such ingredients'. Only vendors
who passed these two screening criteria were asked to participate in the survey.

A total of 20 kosei vendors were interviewed (Accra Metropolis: 9; Ga East: 11).
All of the vendors were interviewed while they were preparing or selling kosei or right
after finishing selling (therefore, to obtain daily data needed for the budgeting analysis,
the questions were asked about the business of the day before the interview was
conducted). All of the vendors were interviewed by the author’s assistant because they
did not speak English or felt more comfortable being interviewed in the local language.

The respondents were asked questions related to: (a) the composition of the costs
for preparing kosei; (b) processing procedures; (c) daily sales; (d) seasonal variations in
operation, costs, and sales; and (e) willingness to use commercially-processed cowpea

flour/meal and constraints associated with it.

3.2.2 Custom Millers

The term “custom miller” refers to a professional miller who owns one or more
milling machines (set up in a small building), mills food items that customers bring, and
charges a fee, which varies depending by the types and conditions (wet or dry) of the
items milled. Custom millers play an important role in the preparation of Ghanaian
cuisine because many Ghanaian staple dishes are made from milled grains such as

fermented maize dough. Kosei vendors are among the regular customers of these millers.

___

' In reality, as reported in Chapter 4, it was very rare to find a vendor who sold only kosei—~typically, the
vendors used at least some elements of the cost components of kosei preparation for preparation of other
food products. Also, some respondents mentioned that they used some of the kosei ingredients for family
cooking. Therefore, in order to secure enough interviews, we had to accept some vendors who used a part
of ingredients for preparing other products or for family cooking. In such a case, we asked them to report
the share of ingredients used in kosei preparation. However, we did not interview vendors from whom we
thought it would be very difficult or impossible to obtain accurate information on the cost of kosei
Preparation because of their business style.

32

 

 

Figure 3.2 Milling machine (plate mill 2-A type) used by a custom miller
in Ga East District

 

Cowpeas, which were traditionally hand-milled using a mortar and pestle, are today
custom-milled in most West African cities including Accra (J. Lowenberg-DeBoer,
personal communication, October 25, 2006', E. Sakyi-Dawson, personal communication,
December 19, 2006). In Accra, custom millers are located inside the markets (there are
often areas in the markets where many custom millers are placed together and work as a
group) and in residential areas. During the field survey, we learned that there is an
association of custom millers called the Greater Accra Co-Operative Food Crop
Processors Union.

A total of 15 custom millers, located in the Accra Metropolis, Ga East, and Ga
West Districts, were interviewed. Interviews were conducted in the local language, except
with one respondent who spoke English fluently.

The respondents were asked questions related to: (a) types of products milled,

frequency milled, and fees charged for food items that they milled; (b) how they milled

33

 

cowpeas for kosei vendors; and (c) characteristics of the machines they used.

3.2.3 Retailers

Retailers in Accra who currently or could potentially sell ready-to-use dry cowpea
meal and Weanimix can be categorized into three groups: (1) grain/flour-type product
vendors in the markets; (2) small shops; and (3) supermarkets.

The main purpose of interviewing retailers was to obtain data on the retail margin
of industrially-processed cowpea flour/meal and Weanimix, where available. Therefore,
only retailers were selected as respondents who sold either dry cowpea flour/meal or
Weanimix, or similar flour-type products such as soybean flour/meal and Tom Brown,
from which we could obtain reasonable estimates of retail margins for the target products
(although this sampling approach did not work with small shops. See Section 3.2.3.2).
The respondents were asked the purchasing and selling prices of these products, as well
as the seasonal variation in prices. Also, the respondents were asked who were the major
customers of COWpea/soybean flour/meal (and of weaning foods, when these products

were sold by grain/flour—type product vendors).

3. 2. 3. 1 Grain/flour-tvpe product vendors in the market

Grain/flour-type product vendors in the markets sell different types of grains such
as maize, millet, sorghum, cowpeas, soybeans, and groundnuts, and ground or primary
processed products such as Tom Brown, fermented maize dough, and gari (i.e., grated,
fermented, and roasted cassava). Market vendors sell their commodities by volume using

a set of standard containers, rather than by weight.

34

Figure 3.3 A grain/flour-type product vendor stall at a market in Accra
Metropolis District

  

1" ..

Source: uthr . i :1

While pre-testing the questionnaire, it was learned that those primary processed
products were very often prepared by the sellers themselves. This meant that these sellers
were also potential/current processors of the targeted products. Therefore, we decided to
ask these vendors an additional set of questions regarding the cost of production
(current/seasonal price of ingredients and roasting and/or milling costs), when the
respondents indicated that the target products were self-preparedz.

A total of seven grain/flour-type product vendors located in five different major
markets were interviewed (2 at Madina market, Ga East District, 1 at Mallam-Atta, 3 at
Kaneshie, and 1 at Agbogbloshie markets, Accra Metropolis District). All the interviews

were conducted in the local language.

M—
2 ' ' u a - A n
The questlonnalre was not modified to formally include these additional questions.

35

 

 

 

3.2. 3. 2 Small shogs

Small shops, located inside the markets or along the roads, were observed
everywhere within the area visited during the fieldwork. However, the size of these shops
varied to a certain degree. Also, the commodities sold were diverse, ranging from food
products such as beverages, oil, rice, and weaning foods to goods for daily use such as
detergent, polyethylene bags, and insect repellent. Most of the commodities sold by these
retailers were commercially produced, well-packaged products including imported goods.

After observing several small shops, it was found that only few of them sold
industrially-processed target products (cowpea flour/meal, Weanimix, soybean flour/meal,
or Tom Brown), while most sold Cerelac, one of the most popular commercial weaning
foods in Ghana, produced by the company Nestlé. After making this observation, it was

decided to ask about the purchasing and selling price of Cerelac when interviewing small

Figure 3.4 A small shop located in the Accra Metropolis District

 

36

 

 

 

 

 

shops, although the retail margin of Cerelac (produced by a large scale company) might
not provide useful information to estimate the retail margin of our target products. Also,
to obtain an idea of the potential retail margin of dry cowpea flour/meal—a product
which most small shops were not selling—the respondents were shown a sample bag of
cowpea flour (purchased from one of the weaning food processor respondents) and asked
a question designed to explore its potential selling price. The question asked was “If you
buy this product for ¢7,OOO/10,000/12,OOO3, at what price would you sell it?”
Interviews were conducted at a total of seven small shops (1 in Ga East District
and 6 in Accra Metropolis District). All the interviews were conducted in the local

language.

3.2.3.3 S_upermarkets
Compared to the other two categories of retailers, there were far fewer
supermarkets in Accra, and they seemed to be patronized only by the rich. Supermarkets
varied in size, ranging from about the same size as a convenience store in the US to about
one-half the size of a typical US supermarket. While a large share of the commodities
sold in the Ghanaian supermarkets seemed to be imported goods, local products were also
on the shelf—including industrially-processed flours and weaning foods.
Staff of a total of four supermarkets were interviewed (2 in Accra Metropolis
District and Ga East District, respectively). The local language was used for one

interview, and the remaining three interviews were conducted in English.

 

 

3 , . . .
These three different purchasing prices were proposed to each respondent. This additional question does
not appear in the questionnaire in Appendix 1; ¢: Ghanaian cedis (US$1 = approximately 569,200 at the time
of the survey. In July 200?, Ghana denominated its currency, and ¢l0,000 became GH¢1. However, the
currency of the time when the survey was conducted is used throughout this thesis.)

37

 

Figure 3.5 A supermarket located in the Ga East District

 

Source: Author.

3.2.4 Weaning Mothers

To interview weaning mothers, four healthcare institutions were selected, which
were located in Ga East and Accra Metropolis Districts: Adenta Clinic (Ga East),
Matemal/Child Health/Family Planning Clinic in Madina (Ga East), Princess Marie
Louise Hospital (also called Children’s Hospital) (Accra Metropolis), and Mamobi
Polyclinic (Accra Metropolis). These locations were selected because of the convenience

for finding potential respondents: each clinic/hospital had a day of the week when

 

mothers could have their small children weighed and keep the record (such a day was
called “weighing day”). During the opening hours of weighing day (it was often during
the morning and early in the afternoon), the clinics/hospital were crowded with mothers

and their small children. Permission to conduct a survey was obtained from the authority

38

 

 

 

 

, J.. .

rt: 3..

 

 

 

 

 

 

of each clinic/hospital4. Then, a nurse in charge of weighing introduced us to mothers and
asked for cooperation before the interviews were conducted.

A total of 30 interviews were conducted (Adenta: 6; Madina: 9, Children‘s
Hospital: 7; and Mamobi: 8). Most interviews were conducted in the local language,
except in a few instances when the respondent was fluent in English.

The respondents were asked questions related to their weaning practices such as:
(1) frequency of giving weaning foods to their children; (2) types and the recipe of
weaning foods that they prepared at home; (3) types and costs of weaning foods that they
purchased; and (4) relative frequency of using self-prepared weaning foods compared to
commercial ones. The mothers were also asked some additional questions about their
consumption of cowpea grain and kosei at home. The intention was to obtain information
from the consumers’ perspective that would supplement the analysis of ready-to-use dry
cowpea meal. Respondents were asked questions such as: (1) whether they prepared

kosei at home and (2) if they had ever purchased commercial cowpea flour.

3.2.5 [rt—dyad Grgin Flouwmm

The number of industrial grain flour and/or weaning food processors and their
scale of production is small in Ghana (E. Sakyi-Dawson, personal communication,
December, 19, 2006). To identify potential respondents, retail shops (mostly
supermarkets) were visited to look for the target or similar products that were on the shelf,

and the processors’ information (mainly telephone numbers) of the products found was

¥

4 To obtain permission to conduct these interviews, we were required to submit an ofi'icial letter from the
Department of Nutrition and Food Science, UGL. Dr. Esther Sakyi-Dawson prepared this letter for us. We
used it in many situations, such as for obtaining the permission to conduct interviews with supermarket
t’nanagers, for building a friendly relationship with grain flour/weaning food processors, and for obtaining
information from the Ghana Standards Board, Ministry of Food and Agriculture, etc.

39

 

 

 

 

 

 

 

noted. The Food Research Institute (FRI) and Ghana Standards Board (GSB) were also
visited to collect additional information needed to identify more processors. A total of 24
companies were identified.

Out of these 24 potential respondents, 16 companies were contacted by the UGL
assistant, who briefly explained the purpose of the survey, and asked for an appointment
to visit. Several companies had to be excluded from the list at this stage for several
reasons, such as the company was not producing the target product anymore. Twelve
companies were visited. During the first visit, the potential respondents were informed
that the goal of the survey was to conduct an economic analysis and that we would like to
collect accounting data. The refusal rate was surprisingly small (only one company
refused). The respondents were also asked to name all the products that they produced.
From the list of products obtained, we found out that there was a significant overlap
among grain flour processors and weaning food processors, that is, many of the
respondents produced both some type of grain flour products and weaning foods5 .
Another finding was that only two of the companies were producing cowpea flour.
Therefore, we decided to ask soybean flour processors the same set of questions prepared
for COWpea flour processorsé. At the end of the first visit, the appointment for a second
visit was made, and a full interview was conducted during the second visit. Although this
strategy was taken at the beginning of the survey, it turned out that this method took too
much time (the respondents, who were most of the time the owner of the company, were
very busy, and it was difficult to meet with them twice). Therefore, we tried to conduct a

full interview during the first visit. It worked to a certain degree, especially when the

 

 

s . . . . . .
The questionnaire was reVised so that it could be used for the respondent producmg either of the target
Products and those producing both of them.
Again, the questionnaire was revised accordingly.

4O

 

 

 

 

respondents had available the company’s accounting record7.

Out of the 11 processors who agreed to participate in the survey, one processor
was interviewed for pre-testing of the questionnaire. All the remaining 10 respondents
were producing at least one of the following products: Weanimix (whether COWpea- or
soybean-), “quasi-Weanimix” (similar product to Weanimix but not exactly the same),
and Tom Brown. Among the IO respondents, 1 produced both cowpea and soybean flour.
This respondent was asked about cowpea flour production (in addition to Weanimix).
Another respondent who produced cowpea flour (but not soybean flour) was asked
questions about its production as well. Soybean flour was produced by four other
respondents, who were also asked about its production. To summarize, interviews were
conducted with processors of Weanimix/quasi-Weanimix/Tom Brown (10), cowpea flour
(2), and soybean flour (4). The locations where the interviews were conducted were as
follows: Accra Metropolis: 2; Tema: 1; Ga East: 4; and Ga West: 3.

The respondents were asked questions related to: (1) general information about
their facility (e.g., operation time, number of workers, and products they produce); (2)
quantity of each product produced during the past month and past year (when the
information was available); (3) price of the target products; (4) total sales of the company
during the past month and past year (when the information was available); (5) equipment
they used to produce the target products (e.g., types, size, year of purchase, price when
purchased, and maintenance costs); (6) costs of operating the facility (e.g., labor,
electricity, fiiel, water, rent, transportation, printing, stationery, and telecommunication);

and (7) information specific to the target products (e.g., input-output ratio, processing

 

 

7 . . - . ' I ' .
InterViews With two respondents were completed in one ViSit, while most of the remaining respondents
were visited twice.

41

 

, .i.
,i'.‘
u .-

‘4;
1,;

5..

procedures, seasonal change in production quantity and price, main customers, and their

experience of and perceptions about using coneas as an ingredient).

3.3 Secondary Data

Historical price data on COWpeas as well as other grain crops were collected from

the Statistics Research & Information Directorate (SRID) of the Ghanaian Ministry of
Food and Agriculture (MoFA). The data set included monthly retail and wholesale prices
of cowpeas, maize, millet, and groundnuts for five years (from 2002 to 2006; but with
many missing months) at “urban markets” of the Greater Accra Region (average of Accra
Metropolis, Ga [Ga East and Ga West combined], and Tema Municipal Districts). Since
no data on the price of soybeans were collected by the MoFA, a wholesaler in Nima
market (Accra Metropolis District) was consulted to obtain a seasonal trend in prices of
soybeans. To estimate a representative price of each crop in February 2007, prices
observed during the fieldwork in different markets were noted. Also, data sets were
collected from Tradenet (httgflwww.tradenetbiz), a website providing detailed up-to-
date price data of different commodities in different countries mainly in West Africag.

The monthly Consumer Price Index (CPI) for Ghana, covering the period from
January 1990 to February 2007, was obtained from the International Monetary Fund’s
(IMF) website (http://www.imforg/external/data.htm). These data were used to convert
the prices and monetary values at different points in time to the current prices and values.

Ghana’s minimum daily wage rate during February and March 2007 was obtained

from the website of the Bank of Ghana (http://www.bog.gov.gh/indexl.php?linkid=l 74).

——¥

3 . . . ' '
Derived representative prices of each crop as well as the methods used to derive these prices are
summarized in Appendix 3.

42

The data were used to estimate the Opportunity cost of labor when analyzing the

profitability of kosei prepared using dry COWpea meal.

3.4 Dry Cowpea Meal and Kosei Preparation Experiments

Two types of experiments were conducted——a laboratory experiment at the
Department of Nutrition and Food Science, UGL and a field experiment with a kosei
vendor. The laboratory experiment was conducted to estimate the input-output ratio for
preparing the B/C CRSP-deveIOped dry cowpea meal suitable for making kosei (i.e., how
many kg of cowpea grain is needed to make 1 kg of B/C CRSP dry cowpea meal).
COWpea grain (type called “Togo”) was purchased from a grain vendor at a market,
weighed, washed with water, soaked for about five minutes, and machine-dried overnight
at 50°C. The next day, the dried cowpea grain was dehulled using a plate mill, machine-
winnowed, milled into meal using a hammer mill, and weighed.

Then, the kosei preparation experiment was conducted to estimate the input-
output ratio for two different kosei preparation processes: (I) wet-milled kosei, prepared
using cowpea grain; and (2) dry-milled kosei, prepared using B/C CRSP dry COWpea
meal. These input-output ratios were needed because, first, for the wet-milling procedure,
many kosei vendor respondents could not estimate their daily revenue from kosei (see
Chapter 4), and second, none of the respondents used the dry meal developed by the B/C
CRSP because it was not commercially available. These facts made it impossible to
calculate the total weight of the kosei that vendors prepared in a given day by the

formula:

 

[[revenue from kosei]

. . x [average wt. of kosei balls purchased during interview] .
[price of 1 kosei ball]

43

 

u. eke

...o

I.’ .

l".‘ '
.

alt I’J'

u' \.
i

M ‘
y.

- |
o . .

I -\ ‘,
Ir.

A I
“~.' _I
1

“J .
i. . .

I»
rud-
~i..

"J
’J-

However, the total weight of prepared kosei is an indispensable piece of information
required to calculate the unit cost of kosei (i.e., cost per kg). Thus, an alternative formula
based on the weight of ingredients used by the vendors was used to obtain the total
weight of prepared kosei, for those respondents who could not provide the revenue of the
day before the interview was conducted. This formula is explained in detail in Section
4.2.6.3.

Among the 20 kosei vendors interviewed, one vendor with five years of
experience who prepared kosei using the “standard” method (i.e., the method used by
many of the respondents) was selected and asked to participate in this experiment. After
she agreed, she was provided two sets of ingredients for preparing kosei: (1) cowpea
grain, onion, pepper, and oil and (2) laboratory-prepared B/C CRSP cowpea meal, onion,
pepper, and oil. With the first set of ingredients, she was asked to make kosei following
her normal procedures, using all the grain provided and the corresponding amount of the
other ingredients. For the second set of ingredients, she was asked to make kosei by first
soaking all the dry cowpea meal provided and then using the corresponding amount of
the other ingredients. The ingredients were weighed before, duringg, and after the
experiment. Then, the kosei balls were counted and weighed.

This series of experiments were conducted twice with the same kosei vendor (one

in March and another in May 2007).

3.5 Per Olonka Weight of Different Craps

Much of the quantity data collected during the fieldwork were reported by the

 

 

9 o u a u o . . . ,
Although the experiments were conducted mice, for the first replication, there is missmg information on
the weight of ingredients during the experiment (see Table 4.15).

44

 

 

7 ‘.
...nb. ..- \

.-
_. 1';
, ..uh
.v-l'v

 

l .
'45 .i \,
'I g

I
I‘ .V

k ’1 a
J » ...
.3“' 0
“~45. .
'n '-

1 .

' «4. |
”PI.
. l:

respondent in volume units rather than weight units. The units regularly used in Ghana to
measure grains and flour-type products are the olonka container and margarine tin
(Figure 3.6). The volume of one olonka is six times the volume of one margarine tin.
During the fieldwork, a sample of each container was purchased at a market, and the
circumference and height were measured. The olonka container had a circumference of
51.0 cm and a height of 17.2 cm, making its volume 3,560 cm}, while the margarine tin
had a circumference of 28.0 cm and a height of 10.2 cm, making its volume 636 cm3 . The
ratio of the calculated volumes is 5.6, which is not exactly six. Since the ratio is known to
be exactly sixlo, this discrepancy implies that some errors occurred when measuring
either the olonka container, the margarine tin, or both. Therefore, these volumetric values
obtained from the purchased container and tin samples have to be considered as
approximation.

Nima market was visited in late March and one olonka of different crops was

Figure 3.6 Olonka container (left) and margarine tin (right)

 

Source: Author.

y

w ' . . _ _ . . . .
This was verified by pouring water from Six full margarine tins into one olonka container, which exactly
filled the container.

45

 

J" ‘
u.» _- e

 

 

2-. ~,,
.

““w

'Wu‘.

 

.‘ ‘u’
f

 

 

purchased, including COWpeas (type called “Niger”), maize, millet, groundnuts, and
soybeans. The weight of each crop was measured in a laboratory at the Department of
Nutrition and Food Science, UGL, using an electronic scale. The representative olonka-
kg conversion rates obtained for different crops are presented in Appendix 2. These rates
were used throughout this study when unit conversions were necessary between olonka
and kg.

However, a potential problem of this method was found. When the samples of
grains were purchased or when markets were visited for interviews, we observed that
“one” olonka or “one” margarine tin does not equal the volume of commodities that fits
into the measuring container (i.e., flat at the ceiling level of the container), because the
grain is conventionally heaped up to form a cone shape. This implies that the volume of
one olonka and margarine tin is arbitrary to a certain degree because the vendors can, for
example, reduce the amount of the heaped-up part when the purchasing price of the
commodity is high, or add one more handful of grains to the cone if her friend is the
customer. Moreover, this means that the weight of the one-olonka samples that we
purchased is just one observation out of a whole population of olonka volumes, and
might be inaccurate if applied to the same commodities but purchased at different times
and/or locations. If this occurred, the estimated prices per kg of corresponding
commodities, derived using the “representative” olonka—kg conversion rates, were

inaccurate as well (see Section 4.2.5.3 for an additional discussion).

3.6 Analytical Methods

The data collected from five different categories of respondents were entered and

46

 

 

...Q. “I -'
i5.“ . .r'
I

..h 5“ U

c
4..)

 

 

 

 

g

 

cleaned in separate spreadsheets of Microsoft Excel. When necessary, monetary values

were converted to current prices using the CPI obtained from the IMF.

3.6.1 Dry Cowpea Meal as an Ingredient in Kosei

First, a descriptive analysis was conducted, using the data collected from kosei
vendors, custom millers, cowpea/soybean flour processors, retailers, and weaning
mothers (as consumers of kosei). The analysis included both a description of quantitative
data and a discussion of qualitative data with regard to non-price competitiveness of dry
cowpea meal. The results of this analysis are reported in Chapter 4.

Second, to estimate the price of industrially-processed dry cowpea meal, an
enterprise budget for industrially producing 1 kg of dry cowpea meal was prepared for
each of the cowpea/soybean flour processors interviewed. However, because of missing
information, the budgets could be constructed only for four out of six respondents.

Third, an enterprise budget for producing 1 kg of kosei was prepared for each of
the kosei vendors interviewed (a total of 13”). Based on the 13 budgets derived, three
representative budgets, for different levels of profitability, were constructed.

Fourth, by replacing the cost of cowpea grain in each of the three representative
budgets with the estimated cost of dry cowpea meal, representative budgets were
prepared for producing 1 kg of kosei from dry cowpea meal.

Fifth, the representative budgets for preparing 1 kg of kosei using cowpea grain
and dry meal were compared to each other to examine the price-competitiveness of dry

COWpea meal, as an ingredient in kosei.

 

 

H As discussed in Section 4.1, among 20 kosei vendors interviewed, 7 vendors were selling a different, if
strictly speaking, product from kosei. For the budgeting analysis, it was judged appropriate to not include
these vendors.

47

 

Finally, sensitivity analysis was conducted by assuming changes in different
variables in the budgets, which would potentially change the price-competitiveness of

industrially-processed dry cowpea meal.

The mathematical method (model) used to prepare the budgets and the results of

analysis are discussed in Chapter 5.

3.6.2 Cowpeg- Weanimix

First, a descriptive analysis was conducted using the data collected from custom
millers, weaning food processors, retailers, and weaning mothers. The analysis included
both a description of the quantitative data and a discussion of the qualitative data with
regard to non-price competitiveness of industrially-processed cowpea-Weanimix. The
results of this analysis are reported in Chapter 6.

Second, to estimate the processor prices of industrially-processed Tom Brown,
cowpea-Weanimix, and soybean-Weanimix, enterprise budgets for producing 1 kg of
these products were prepared for each of the weaning food processors interviewed.
However, because of missing information, the budgets could be constructed only for five
out of 10 respondents. Based on the five budgets derived, two representative budgets, for
different levels of profitability, were constructed. Then, to estimate the retail price of the
three weaning food products, a representative retail margin (derived using the
information collected from the retailer respondents) was added to the estimated processor
prices of the products.

Third, using the data collected from the grain/flour-type product vendors who

were self-preparing Tom Brown (7 observations), a representative enterprise budget for

48

 

J.._

 

 

‘ I! 7'—

.
.,..... a»
u

— 6‘

 

‘u.
A ."

‘u.

 

 

 

 

self—preparing 1 kg of Tom Brown was constructed. In turn, this budget was used to
construct enterprise budgets for self-preparing 1 kg of cowpea-Weanimix and soybean-
Weanimix, by changing the raw material costs from only maize to maize +
cowpeas/soybeans + groundnuts.

Fourth, to examine the price-competitiveness of industrially-processed cowpea-
Weanimix, its estimated retail price was compared to the estimated retail price of: (l)
industrially-processed Tom Brown; (2) industrially-processed soybean-Weanimix; and (3)
self-prepared cowpea-Weanimix.

Finally, sensitivity analysis was conducted by assuming changes in different
variables in the budgets, which would potentially change the price-competitiveness of

industrially-processed cowpea-Weanimix.

The mathematical method (model) used to prepare these budgets and the results of

analysis are discussed in Chapter 7.

3.7 Summary

In this chapter, the sources of data used in this study were described, and the
analytical methods used were summarized.

Fieldwork was conducted in the Greater Accra Region during February and
March 2007. Interviews were conducted with 20 kosei vendors, 15 custom millers, 18
retailers, 3O weaning mothers, and 10 industrial grain flour/weaning food processors.
Price data for different commodities were collected from different sources, including the
Ministry of Food and Agriculture. Laboratory and field experiments were conducted to

estimate input-output ratios of dry cowpea meal and kosei.

49

 

;

Using the qualitative data collected, descriptive analysis was carried out to
examine non-price-related factors that would affect the competitiveness of industrially-
processed dry COWpea meal and cowpea-Weanimix. Quantitative data were used to
prepare enterprise budgets to analyze the price-competitiveness of the target products.

Sensitivity analysis was conducted to analyze the change in price-competitiveness under

different scenarios.

50

 

 

i -
... .
..s ....
i
he. _
~—-
.
b..-
t \_
a,“
:—

 

 

 

CHAPTER 4
DRY COWPEA MEAL FOR PREPARATION OF KOSEI

—DESCRIPTIVE ANALYSIS

4.1 Overview of Kosei Business and Grain Flour Production in Accra

As described in Section 2.2.1, kosei is a deep-fried cowpea fritter that usually
contains dehulled and milled cowpeas, water, onions, peppers, and salt. However, during
the fieldwork in Accra, some street vendors were observed not using pepper, while others
used additional ingredients such as spices, eggs, green leaves, and soybeans. Some
vendors only sold kosei during the morning, while others also sold it in the afternoon‘.
Kosei was rarely sold alone but along with other food products such as hausa koko
(millet-based porridge with spice and sugar), bread, groundnuts, and other fried foods
such as pinkaso (small-doughnut-looking fried wheat flour containing onions and eggs
with sprinkled sugar). Customers could buy kosei in individual balls or in bulk. Most
vendors sold kosei in polyethylene bags: a thin inner transparent bag and a thicker-outer
black bag.

Two types of “kosei” were sold in Accra—one type was made from wet-milled
COWpeas, and the other type was made from dry-milled cowpeasz. Street vendors called
the wet-milled type kosei and the dry-milled type agawu. However, the majority of
consumers in Accra seemed to call both types kosei. Although they did not distinguish

between kosei and agawu by name, these products were very different in appearance and

 

 

' Some vendors may only sell kosei in the afternoon. However, since most of the kosei vendor interviews
were conducted during the morning, we did not observe such vendors.

The difference is whether cowpeas are wet or dry when milled. Therefore, both “wet method” and
“combined dry and wet method” described by Dovlo et al. (1976) are considered as wet-milled.

51

taste. In general, kosei has a rounder shape, while agawu has a more jagged shape (see
Figures 4.1 and 4.2), and kosei has a light spongy texture, while agawu has a harder and
drier texture". Therefore, it was expected that consumers preferred one over the other, and
patronized the vendors who sold the type that they preferred. In Accra, kosei seemed to

be more popular than agawu (E. Sakyi-Dawson, personal communication, March 21,

2007).

Figure 4.1 Kosei Figure 4.2 Agawu

     

”We

'''''

    

»~ r a
Source: Author.

To make agawu, cowpeas are first dry dehulled and milled into flour. Since it is

dry, this flour can be kept for a longer time—compared to wet-milled cowpea paste that
spoils quickly. Vendors decide on the amount of flour needed for one day of business, add
water, whip, add other ingredients, and fry. This procedure is the same as when using dry
Cowpea meal developed by the B/C CRSP. However, it is important to note that the B/C
CRSP-developed meal is for preparing kosei, not agawu. As described in Section 2.2.1,
this meal can be used to make kosei that tastes similar to the one prepared from wet-

milled COWpeas.

 

3 The color of the inside is largely affected by the amount of pepper used. Therefore, it is not clear whether
the difference in color is attributed to the difference in processing procedures (i.e., wet or dry mill).

52

.‘1.

_...§~-
l

n' n'

\

. as
n — .
H ..-

u

an,

up...

f.-

N..

 

i“

 

If it becomes commercially available, dry cowpea meal suitable for making kosei
is expected to have various advantages. These advantages include“:

(1) it would reduce preparation time because kosei vendors could skip time-
consuming parts of processing procedures such as sorting out bad grains, dehulling, and
milling;

(2) since the meal can be turned into paste simply by adding water, it might allow
kosei vendors to make a fine adjustment in the quantity of kosei to prepare, depending on
the sales of the day. This is not possible with the current procedure because if vendors run
out of paste and want to sell more, they would have to start from wet-milling cowpea
grain. By the time vendors have prepared another batch of paste, all customers would
have left. Likewise, for kosei vendors who often have leftover paste or balls when they
want to finish for the day, dry meal might allow them to start with a smaller amount of
paste and prepare additional paste little-by-little, as the sales go so that they do not end
the day with leftovers;

(3) if packaged in an air-tight container, the meal could be kept longer than when
it is self-prepared and stored in an open-to-air condition5 . This would enable kosei
vendors to purchase the main ingredient at a stable price, compared to the current
situation where vendors have to buy cowpea grain regularly, which exposes them to price
fluctuations; and

(4) it might stimulate the preparation of kosei at home, thereby increasing the
demand for coneas—anecdotal evidence suggests that some potential kosei consumers

do not currently buy kosei from street vendors because they dislike the ways the vendors

 

 

4 Collected through personal communications with B/C CRSP fOOd SCientiStS and economists.
Dry flour/meal is not totally insect-free if not stored in an air-tight container (G A. Annor, personal
communication, February 16.2007)-

53

 

 

i. -v<

vva.>

”A.
"4..

M.

“"LI

...v

by

.1,

 

 

 

prepare kosei. If ready-to-use commercial cowpea meal becomes available, these
potential consumers might start preparing kosei at home, tailoring it to their taste.

These potential benefits of commercial dry cowpea meal are analyzed in the rest
of this chapter, based on the data collected during the fieldwork.

In today’s Ghana, a wide range exists in the types and scales of grain flour
processors. At the most commercial level are multinational companies such as Nestle’
Ghana. The company produces a variety of weaning foods that are made from grain
flours mixed with other ingredients. Producing weaning foods and meal require similar
equipment and processing procedures. Therefore, a large multinational company
currently producing weaning foods is a potential processor of dry cowpea meal. If
companies of such a scale produce dry cowpea meal, unit costs would most likely be
lower than when it is produced by smaller companies due to economies of scale.

The smallest scale processors of grain flour are individuals. Anybody who has
access to custom millers can self-prepare grain flour. Custom millers also make it
possible for grain vendors in the market to produce and sell grain flour. In fact, anyone
can be a potential dry cowpea meal processor, if he/she has access to a dryer and if a
custom miller can mill cowpeas into the appropriate particle size with the mill that they
are currently using.

Between the above-mentioned largest- and smallest-scale grain flour processors,
there exist small- to medium-scale local companies that either specialize in grain-flour
production or produce a wide range of food products. There exists a variety in scale
within these companies: some have more than 50 workers including managers and

accountants, while others are owned and operated by just a person or run as a family

54

_,', ..nw

 

st,“

 

 

 

 

business, possibly hiring temporary workers when busy. Some own all or most of the
equipment they need for processing, while others outsource some or many parts of
processing procedures (e.g., having a hand sealer for packaging but using custom millers
for milling). Some have a constant demand for their products and operate throughout the
year, while others produce with interruption depending on the demand. All the 10
companies interviewed during the field survey were small- to medium-scale locally

owned processors.

4.2 Description and Implication of Data
This section presents descriptive analyses of the data collected during the

fieldwork and describes the steps taken to prepare these data for the budget analysis.

4. 2.] Street Vendors of Kosei and A gawu
4. 2. 1.] Characteristics ofkosei and agawu vendor reswdents

Of the 20 street vendors interviewed, 13 were kosei and 7 were agawu vendors.
Characteristics of these respondents are shown in Table 4.].

While all kosei and agawu vendors were women, the number of years they were
in business varied widely across the respondents. Many respondents were assisted by
members of their family such as daughters and sisters or, with a lower frequency, hired
workers. Among kosei vendors, only two respondents worked without their family
members, although one of these two respondents worked with two hired workers. Among
agawu vendors, three respondents had no family assistance, although one of these three

respondents worked with one hired worker. The majority of respondents worked everyday,

55

 

 

 

Table 4.1 Characteristics of kosei and agawu vendor respondents

 

 

 

 

 

 

 

 

F kosei agawu
Characteristics vendor vendor

L: 131 (n = 7)

Male 0 0

Sex Female 13 7
Fewer than 1 year 2 0

Experience in l — less than 3 years 2 3
kosei/agawu business 3 — less than 10 years 6 3
10 years or more 3 1

Number of family (I) i ?
members assisting with 2 4 2
the busmess 3 2 l
0 12 6

Number of hired workers 1 0 l
2 l 0

Number of working days 6 5 2
per week 7 8 5
None 3 0

Hausa koko 9 7

Other food products Bread 5 7
selling along with kosei Groundnuts 2 3
Pinkaso 2 O

Boflot 0 1

Have other source of No 12 3
income Yes I 4

 

 

 

 

 

 

Note: Hausa koko: a spiced millet-based drink; Pinkaso: small-doughnut-looking fried
wheat flour containing onions and eggs with sprinkled sugar; Boflot: a fried wheat

flourbaH
Source: Field survey in Accra, February and March 2007.

while the remaining respondents worked six days a week.

As briefly mentioned earlier, most respondents, as well as other kosei/agawu
vendors observed during the fieldwork, sold other food products along with kosei or
agawu. Among kosei vendors, three respondents were specialized kosei vendors, while
the other 10 respondents sold one or several other products. Among such foods sold with
kosei, the most popular was hausa koko, a spiced millet-based drink, which was served
with sugar in a bowl for customers who wanted to eat it at the place of purchase or in a
transparent polyethylene bag for takeout. Among the 10 multiple-product~selling kosei

respondents, 9 sold hausa koko. While the other one did not sell hausa koko herself, she

56

 

 

. - ~b
... ...w
q.
,'
"4 l
r-ID K I
.-' l'
.
.. ..
.
... .
.
I‘- ~
\J
J. ._
.
w.

 

‘r
-.\.

‘-».I .
.

u». _
., ‘.
A»

.‘C __

.,_‘_

-1
t3,

 

 

 

worked with another person who sold hausa koko right next to her. The second most
popular product among kosei vendors was bread. When customers requested, the vendors
cut the bread, put kosei balls into it to make a sandwich, and sold it wrapped in a paper.
The other food products that kosei respondents sold were groundnuts and pinkaso.
Among agawu vendors, all seven respondents sold both hausa koko and bread. In
addition, three respondents sold groundnuts, and another respondent had just started
selling boflot, which is a fried wheat flour ball. There was no agawu respondent who sold
pinkaso. While the sample size is too small to make any generalization, based on the
observation during the field survey and discussion with food scientists at UGL, it seems
that one could safely say hausa koko and bread are often sold with both kosei and agawu
in Accra.

Of the 13 kosei respondents, only 1 had another source of income (selling clothes),
while four of the seven agawu respondents had another source of income (selling fruits,
selling soaps, selling charcoal, and working as a seamstress). When asked to estimate the
share of their income from selling kosei/agawu, two of them, including the clothes-selling
kosei vendor, did not know the answer, while the remaining three reported that agawu
and the products sold along with agawu made up the majority of their total incomeb.
Again, given the small sample size, no generalization can be made. However, it is
interesting to see such a large difference between the kosei and agawu vendor
respondents in terms of the percentage who had another source of income. A possible
hypothesis is that agawu preparation requires less labor (than kosei) because agawu is

prepared from dry-milled cowpea flour. Therefore, agawu vendors have more time that

 

 

6 . . .

The agawu vendor who was also a seamstress implied that she would have her shop in the future. Whether
kosei/agawu business creates enough return for vendors to save money and eventually allow them to invest
in a potentially more profitable job is an interesting topic to study.

57

 

 

 

 

[J
..lv‘ ‘“
.
i
‘F'; .3;-
_JDA'
l
9' ‘
nu' i\.
‘Qi
c
n . .
;"“I<

cu -.

 

 

 

 

 

they can allocate for other activities. If a survey were conducted to test this hypothesis,
the results would provide useful information to help to assess the benefits of developing

dry cowpea meal for kosei preparation.

4.2.1.2 Types and sources ofcost components to premre kosei and agawu
Data collected to estimate the cost components for preparing kosei and agawu are

shown in Table 4.2.

Table 4.2 Types and sources of cost components to prepare kosei/agawu
among the respondents

 

kosei agawu
Cost components vendor vendor
(n = l3) (n = 7)

 

O

Burkina

Niger

Nigeria

Togo

Vendor in the market
Wholesaler
Onion

Pepper

Soybean

Ginger

Garlic

Green leaves
Salt

Spice

Water vendor
Source of water Use house water
Do not buy water
T pe of oil Vegetable
Charcoal

Type of fuel Firewood

LP-gas

To dehull

To mill

Source: Field survey in Accra, February and March 2007.

Origin of cowpeas used

 

Source of cowpeas

 

Other ingredients

_‘

 

 

 

—w—Noww——mowwo\iwwoo

——‘
_—

 

C

Use of custom miller

—
L3)

 

 

 

 

 

\JONOOQNNOMOQOOONWQMN——A~

 

58

 

 

 

0‘.) s‘.
...i. 5'. .

..T"

.4... ...

.
f1“.

".

.0“

.
uh.

:1

an.

There are a large number of different COWpea varieties (see for example Dovlo et
al., 1976, frontispiece, for a picture of 42 varieties of cowpea grains, which differ in color
and size). In Accra, grain vendors and customers mainly distinguish varieties based on
the country of origin. The varieties observed during the fieldwork were referred to as
“Burkina,” “Niger” (there were two different types of “Niger”), “Nigeria,” “Togo,” and
another was called “small bean.” The eight kosei vendor respondents who used Niger
preferred it because of its: (1) better taste (mentioned by 5 respondents); (2) better
swelling capacity (2 respondents); and (3) better look of kosei (2 respondents). The three
respondents who used Nigeria preferred this type because: (1) of its better taste (2
respondents); (2) of its better swelling capacity (2 respondents); (3) kosei does not break
easily after fried (1 respondent); and (4) the respondent thought customers liked Nigeria
and felt that she made more money by using it (1 respondent). The two vendors who used
Togo preferred it because of its: (1) better taste (1 respondent); and (2) better look of
kosei (1 respondent). It was interesting to find that while the respondents used different
types, they mentioned the same reasons for choosing the type.

Among the agawu vendor respondents, four used Niger, which they preferred
because of its: (1) better taste (mentioned by 3 respondents); and (2) better swelling
capacity (1 respondent). Burkina, Nigeria, and Togo were each used by one respondent,
respectively, which was preferred because of its better swelling capacity (Burkina),
softness and better swelling capacity (Nigeria), and better taste (Togo).

All the respondents, including both kosei and agawu vendors, answered that they
purchased COWpeas from either vendors in the market or wholesalers. However, since

wholesalers sold cowpeas in the market, the respondents who answered “vendor in the

59

 

 

 

 

\

 

I!

 

 

 

 

 

market” might have meant “wholesaler in the market.” As a result, it is not clear from the
data how many of the vendors purchased cowpeas from retailers and how many
purchased from wholesalers. No respondent bought cowpeas from a retail store outside of
the market or directly from farmers.

Onion, pepper, and salt were the most common ingredients added to kosei/agawu.
All 13 kosei vendor respondents used these three ingredients. In addition, some of the
respondents added other ingredients such as ginger, garlic, green leaves, and spices. All
seven agawu vendor respondents added onion and salt, but only three added pepper (one
of these 3 respondents added pepper powder to already fried agawu, but only when
customers requested). The only other ingredient added by two out of seven agawu
respondents was soybeans; one of them reported that she added soybeans because it made
agawu taste better and more nutritious.

The majority of respondents bought water from water vendors, mostly on a daily
basis. All the respondents, both kosei and agawu vendors, used vegetable oil to fry their
products. Almost all the respondents used charcoal to heat the oil.

For dehulling cowpeas, two kosei respondents reported that they used mortar and
pestle while another kosei respondent simply mentioned wet-dehulling without
specifying the tool she used. While one agawu vendor made agawu without dehulling
COWpeas, all the other kosei and agawu vendors custom-dehulled cowpeas. For milling

COWpeas, all the kosei and agawu respondents used custom millers.

60

 

 

 

 

4.2.1.3 Prices and sales

Data related to the price of kosei/agawu and daily sales are shown in Table 4.3.

Table 4.3 Product price and daily sales of kosei/agawu among the respondents

 

 

 

 

 

 

 

 

 

 

 

 

kosei agawu
Product price and daily sales vendor vendor
(n = 13) (n = 7)
Price per ball“ 95500 13 7
20-30 g 5 1
30-40 g 8 2
. 40-50 g 0 1
Weight per ball 50-60 g 0 3
Mean (g) 32 43
Standard deviation (g) 5 13
Size of ball changes during the day No 12 7
Yes I 0
O < revenue 3 50,000 3
Scale of business: revenue from 50,000 < revenue 5 100,000 3
kosei (¢) the day before the l00,000 < revenue 5 200,000 4
interview was conducted” 200,000 < revenue 5 500,000 2
500,000 < revenue 1
Scale of business: quantity of g : cowpea grain E i 3 2
COWpea grain (kg) used the day cowpea grain ’ 2
before the interview was 4 < cowpea gram 5 8 4 3
conducted*** 8 < cowpea grain S 16 2 0
16 < cowpea gram 1 o

 

* Exchange rate: USSI = approximately ¢9,200 (Ghanaian cedis) at the time of the
survey. In July 2007, Ghana denominated its currency, and ¢l0,000 became GH¢l.
However, the currency of the time when the survey was conducted is used
throughout this thesis.

** Includes estimates (see Section 4.2.6.3); leftover balls, if any, were assumed to be
sold; the variable is not available for agawu vendors because of missing information.

*** Includes estimates (the methods used are available from the author upon request); for
two agawu respondents who used soybeans, the values were the sum of cowpeas and
soybeans.

Source: Field survey in Accra, February and March 2007.

All respondents sold kosei and agawu for ¢500 (Ghanaian cedis) per ball
(approximately US$0.054 at US$1 z ¢9,200). However, the weight per ball showed a
wide range, implying that the unit price of kosei/agawu was different across the
respondents. The mean weight per ball of five or six kosei balls measured after the

interview averaged 32 g with a standard deviation of 5 g. The mean weight per ball of

61

 

 

«pet

_‘ippl.
-

In"

l..».~.

’;:‘t F

.\.. y.

“L:-

“MD.

 

 

five or six agawu balls averaged 43 g with a standard deviation of 13 g.

All the respondents except one kosei vendor stated that they usually kept balls the
same size throughout the day. This means that most vendors did not make smaller balls
during the peak time, although they could possibly make more money by doing so.
Therefore, the unit price of kosei/agawu within a day seemed to be constant per
respondent, although it varied across respondents7.

The scale of business (measured in terms of daily revenue and quantity of
cowpeas used) varied widely across the respondents: for kosei, three respondents
earned ¢0 — ¢50,000 (US$0 — US$5.44) from kosei alone (i.e., not including the revenues
from other food products) from the 0-2 kg of cowpea grain used the day before the
interview was conducted; three respondents earned ¢50,000 - ¢100,000 (US$5.44 —
US$10.88) using 2-4 kg; four respondents earned ¢100,000 — ¢200,000 (US$10.88 -
US$21.76) using 4-8 kg; two respondents earned ¢200,000 -— ¢500,000 (US$21.76 —
US$54.34) using 8-16 kg; and one respondent earned more than ¢500,000 (over
US$54.34) using more than 16 kg. For agawu, two respondents used 0-2 kg of cowpea
grain the day before the interview was conducted (sum of the weights of cowpeas and
soybeans for one of them); two respondents used 2-4 kg; and three respondents used 4-8

kg (sum of the weights of cowpeas and soybeans for one of them).

4. 2. 1.4 Processing procedure for kosei and agawu
Each respondent had her own way of preparing kosei or agawu. These procedures

might have been developed based on the vendors’ experience, or they might have been

 

 

7 One might doubt if the respondents honestly answered this question. However, as is shown later in
Section 4.2.1.7, many respondents mentioned that they changed the size of balls when COWpeas were scarce.
Therefore, their reply that they kept the size of balls throughout the day seemed to be trustworthy.

62

 

 

".4
nth

'1'

”vii
...s.

4‘”
\sbu

the methods passed down from generation to generation. However, although their
procedures were not exactly the same, it still is possible to derive a representative
processing procedure that was typically followed by the majority of respondents. Figure

4.3 describes the typical procedure for preparing kosei and agawu.

Figure 4.3 Representative processing procedures for kosei and agawu

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Kosei A awn
Custom Custom
dehull (dry) dehull (dry)
Manual Manual
winnow winnow
‘ Custom mill
Soak 7 Wash ‘ (dry)
l Just wash r
Wash Soak
Add other
ingredients
1 .
Custom mill Add water
(wet) and whip
_ Add other
Wh'p ingredients
Fry Fry

 

 

 

63

 

 

 

 

 

 

 

 

 

\
Ln...

\
“on...

"u
n-i h

‘si

9-

Whether the vendor prepared kosei or agawu, most respondents first “custom-
dehulled” dry cowpea grain, that is, they brought cowpeas to a custom miller, who
adjusted the plates so that the machine did not mill but just cracked the cowpeas. After
the cowpeas were passed through the machine, they broke into smaller pieces, and most
hulls were separated from the endosperms. Then, the vendors manually winnowed the
hulls. Kosei vendors brought home these dry-cracked cowpeas, while agawu vendors had
the custom millers put these cowpeas back into the machine to mill them into flour. Using
these procedures, the vendors could have dehulled/milled a large quantity of COWpeas
(e.g., l sack of 40 olonka), if they could afford to buy such a large amount at once. By
doing so, these vendors could save money because of the discount associated with bulk
purchase and bulk custom dehulling/milling.

For kosei vendors, the next step was to select the quantity of cracked cowpeas that
they needed for one day and do one of the following: (1) wash and then soak; (2) just
wash; or (3) soak and then wash. There was no typical soaking time among the 13
respondentss: 8 respondents reported they soaked for 0-1 hour; 2 respondents for 2-3
hours; and 3 respondents overnight. No relation was found between the length of soaking
time and the vendor’s experience in kosei business. Then, the vendors added other
ingredients to the cowpeas, had them custom milled into paste, and finally whipped and
fried them. Whipping time also varied among the respondents): two respondents reported
that they whipped for less than 10 minutes; three respondents for 30 minutes; and another
three respondents for 60 minutes (5 respondents could not recall exactly how long they

whipped). From the data, no obvious relationship was found between the whipping time

 

 

8 Including the three respondents who did not custom-dehull cowpeas.
Again, including the three respondents who did not custom—dehull COWpeas.

64

 

 

 

and vendor’s experience in kosei business. On the other hand, a negative relationship was
found between soaking and whipping time.

To prepare agawu, most respondents added water to the flour and whipped it to
form a paste (no respondent mentioned that she soaked the flour before whipping). Then,
they added the other ingredients to the paste and fried the batter. Whipping time, which
was reported by five respondents, varied from 10 to 40 minutes").

Even though the use of custom millers for dehulling cowpeas allows kosei
vendors to skip the tedious procedure of traditional wet dehulling, kosei preparation still
seems to be more time and labor consuming than agawu preparation. In the morning,
kosei vendors began with soaking or washing cowpeas, and then brought their cowpeas
and other ingredients to a custom miller. Even when vendors soaked cowpeas overnight,
they still needed to go to a custom miller. On the other hand, agawu vendors could start
whipping after simply adding water to the flour. Preparation of other ingredients is the
only more involved step in agawu preparation, compared to kosei preparation: while
kosei vendors can have other ingredients ground with cowpeas by custom miller, agawu
vendors have to grind or chop other ingredients separately. The difference in time needed
to prepare kosei and agawu seemed to affect the difference in time when vendors started
preparation: on average, kosei vendors started preparation at 3:54 am”, while agawu

vendors started at 5:38 am.

 

 

|0 . .

A kosei vendor respondent who apparently knew how to prepare agawu mentioned that dry~milled
cowpeas take more time for whipping because they take more time to swell.

Mean of 10 respondents: excluded are one vendor who dehulled cowpeas in the morning using a mortar
and pestle and two vendors who reported information about their processing procedures for their evening
sales.

65

 

 

 

 

i)¢'

's .‘u

4.. .
O~5J

...u.

..y vb

 

'3
u.

4.2.1.5 Time and labor savinggfiects ofdry cowpea meal
The use of commercial dry cowpea meal to prepare kosei means that the
processing procedure becomes closer to that of agawu, and perhaps even simpler because
vendors would not need to bring cowpeas to a custom miller to prepare dry meal by
themselves. However, there are at least four elements that must be considered to
accurately estimate the potential advantage for kosei vendors of using dry cowpea meal
in terms of time and labor saving effects:

(1) Although the use of dry meal would most likely allow vendors to start
preparation later in the morning, it does not necessarily mean the vendors could wake up
later than their current wake-up time. In fact, four agawu vendors'2 reported that they
woke up, on average, 1 hour 45 minutes earlier than they started preparation of agawu.
This was because they had to prepare hausa koko, which they sold along with agawu.
This finding implies that if kosei vendors also sell hausa koko and if hausa koko takes
more time to prepare than kosei, the time and labor saving effect of using dry cowpea
meal would have to be valued based on what kosei vendors could do during the spare
time created by the use of cowpea meal while preparing hausa koko. To estimate the
value of saved time and labor, we would need to know how each member of the
respondents’ family is currently involved in the business (e.g., just helping sell kosei or

helping prepare both kosei and hausa koko), and how that involvement would change if
the respondent switched to using dry meal;
(2) Soaking dry meal would be an additional task, and the time needed for this

procedure should be considered. While agawu vendors did not report that they soaked

 

 

12 . .
Wake-up time was not obtained from the other three agawu respondents.

66

 

 

 

“ILA“.
‘, I 5 I

J‘.‘

~urt.

W1»

1...
v

 

to:

..q

in

'i.‘

q.

 

flour, dry cowpea meal deveIOped by the WC CRSP needs to be soaked for 30 to 60
minutes (see Figure 2.2);

(3) Whipping time might change. As discussed later in the section for the kosei
preparation experiment, too short whipping time negatively affected the quality of kosei.
There was a wide range in the current whipping time among the kosei vendor respondents.
Therefore, depending on the current practice of whipping, a kosei vendor might need
either a shorter or longer whipping time after switching to dry cowpea meal;

(4) Grinding onion, pepper, and other ingredients, if any, would be an additional
task. As mentioned earlier, kosei vendors currently have custom millers mill these
ingredients together with cowpeas. If dry cowpea meal is used, these ingredients would
have to be ground separately. Whether they are chopped manually by the vendors or
ground using a mortar and pestle or a home blender, this new step would require
additional time and labor for the kosei vendors, although the amount of time and labor
needed would be much smaller than the overall amount of time and labor saved by using
dry meal. The vendors could bring onion, pepper, and other ingredients to a custom miller,
if the quantity is large enough. However, it would reduce the benefit of using dry meal
because one of the big advantages of dry cowpea meal is that vendors would no longer
have to go to a custom miller early in the morning.

The use of dry cowpea meal would definitely save overall time and labor of kosei
vendors. However, the above mentioned elements would affect how much is saved, and

therefore affect the premium that kosei vendors would be willing to pay for dry COWpea

meal.

67

 

 

 

 

I".

u

4
i‘

4.2.1.6 Fine gdmstment in the quantity ofikosei to prgpare

Among the potential advantages of dry cowpea meal discussed earlier in this
chapter was that the meal might allow kosei vendors to make a fine adjustment in the
quantity of kosei prepared, according to the sale of each day. To examine this potential,
data related to the respondents’ experience in daily shortages and leftovers of

kosei/agawu balls were collected. The results are shown in Table 4.4.

Table 4.4 Daily shortage and leftover of kosei/agawu among the respondents

 

 

 

 

 

 

 

 

 

 

kosei agawu
Shortage and leftover vendor vendor
(n = 13) (n = 7)
Experienced daily shortages No 0 0
in the past month Yes 13 7
Feeling when kosei/agawu ran Unhappy 7 3
out Happy 4 2
Both 2 2
Experienced daily leftovers in No 3 0
the past month Yes 10 7
1-20 4 5
Typical amount of leftover 2"40 0 I
(number of balls) 41-60 . 2 I
It depends (With max 40 or 50) 3 0
Don’t know 1 0
Threw away 0 0
Gave to family 7 6
What to do with leftover Gave to friends 5 4
Gave to children 2 0
Re-fried and sold (when a lot) 1 0
More often than lefiovers 6 7
In the past month, shortages Less often than lefiovers 0 O
happened: As often as leftovers l 0
Don’t know 6 0

 

 

 

Source: Field survey in Accra, February and March 2007.

All the respondents reported that in the past month they experienced running out
of kosei/agawu when there were still customers who wanted to buy their products. When
kosei/agawu ran out, seven kosei and three agawu respondents felt they should have

Prepared more kosei/agawu to make more money and that they would prepare more the

68

 

w- ‘
a
.uik

,-

‘3‘

.b. t
.

next day; four kosei and two agawu respondents felt happy that the kosei/agawu sold out
and that they would prepare the same amount of kosei/agawu the next day; and two kosei
and two agawu respondents answered they had both of these feelings‘3 . For those
respondents who answered that they felt they should have prepared more, methods to
avoid daily shortages would be helpful.

All except three kosei vendors reported that they had experienced leftover paste or
balls during the past month. The quantity of leftovers that they usually had varied across
the respondents, with a maximum of 60 balls (measuring the leftover in terms of the
number of balls). The methods utilized by those three kosei respondents who did not have
leftover were as follows: always bring the amount she thinks she can sell; when there is
leftover in the morning, re-fry and sell it in the evening; and when the end of the day
approaches and if there are remaining balls, give customers more balls than they paid for.

No respondent threw away the leftovers'4 .,Rather she gave the balls to family,
friends, or children in Muslim schools. When there was a lot of leftover, one kosei
respondent reported that she put them in the freezer, re-fried, and sold them at another
time.

Except six kosei vendors who did not answer the question, most respondents ran

out kosei more often than they had leftovers.

 

'3 Obviously, a vendor cannot feel at the same time that she will prepare “more” and “the same amount” the
next day. Those vendors might have answered “both” because they were actually asked two different
questions at the same time (i.e., whether or not they felt happy for running out of kosei/agawu and whether
or not they felt they would prepare more the next day). They could have thought “unhappy for running out,
but will prepare the same amount the next day,” or “happy for running out, but will prepare more the next
day” Another possible explanation would be that their feelings were not always the same. In fact, one of
the two kosei vendors answered “it depends.” It probably depends on what else she has to do on that day

14If the respondents had leftover in the form of paste, they most likely could not give it to somebody and
therefore threw it away. The result of zero “throw away’ ’implies that the respondents fried all the paste
they prepared, whether or not they could sell all the balls.

69

 

YA .

 

I‘ll

Al

It was unexpected that all the agawu respondents experienced daily shortages and
the majority of them felt they should have prepared more. Rather, it was expected that
agawu vendors, who prepare their product from dry COWpea flour, could easily make
additional balls by quickly adding water to flour, whipping and seasoning the paste with
other ingredients, and frying it. One of the agawu respondents mentioned that she wanted
to fry all the paste at once and did not want to start preparation over again (whip the paste,
put charcoal on fire, and fry). In fact, she did not bring extra flour to the place where she
was selling agawu. Her statement matches the observation made during the fieldwork: no
vendors were found frying agawu based on customers’ orders. Rather, they prepared a
relatively large amount of paste at once, fried it, kept agawu balls in a container, and
waited for customers (perhaps preparing another batch if she was a large-scale vendor).

Under the current conditions, kosei vendors are unlikely to be able to make
additional balls when they run out of kosei because they most likely do not have time to
run to a custom miller and prepare kosei before the customers leave. If they used dry
cowpea meal, they would not need to go to a custom miller anymore. However, as field
observations and this story from the agawu vendor indicate, using dry meal would not be
a sufficient condition for kosei vendors to be able to make additional balls when they run
out of kosei. They would have to be diligent enough to be willing to start the procedure
from the beginning after frying a whole batch of paste.

Soaking time might be another and perhaps more significant constraint for using
dry COWpea meal to respond to an unexpected demand. Following the above discussion,
one might argue that dry cowpea meal is still useful for making quantity adjustments in

situations as follows: a relatively large-scale kosei vendor prepares three batches of paste

70

 

DI)"
5‘51

w»:
155\

'H
.ui

' 'v'
13..

. .1

f )
’71

3'-
Di.

fr

everyday. One day, when she finishes frying the second batch, she realizes that she is
receiving many more customers than usual, and she is sure that she will run out of kosei.
If she used dry cowpea meal, she could make the third batch a larger volume to capture
the demand from additional customers. In such a case, dry cowpea meal would help the
vendor. However, as mentioned earlier, to use dry cowpea meal developed by the B/C
CRSP, 30 to 60 minutes of soaking time is necessary to prepare good quality kosei.
Therefore, the vendor using this meal would need to predict whether she would run out of
kosei and how much more to prepare at an early point in the day, so that she would have
enough time for soaking the meal.

Both of the above mentioned constraints (i.e., cumbersomeness of starting over
the preparation and soaking time) would apply when discussing the usefulness of dry
COWpea meal for avoiding leftovers. The idea was to begin with a smaller amount of paste
and prepare more paste as the vendors sell kosei. To succeed in this method, vendors
would have to accept the less convenient not-all-at-once preparation, and they would also

have to be able to accurately predict the amount of the day’s sale.

4. 2. 1. 7 Seasonality

Data were collected to examine how much dry cowpea meal would help kosei
vendors to cope with the seasonal fluctuation in the availability of cowpea grain. This
data set is shown in Table 4.5.

All the 20 respondents answered they sold kosei/agawu throughout the year, using
the same ingredients, without changing the amount of kosei/agawu they prepared (in

terms of the amount of ingredients, not the number of balls) regardless of the seasonal

71

 

 

 

’J
w.

4

.l-

 

{H
Isl

Table 4.5 Seasonality in kosei/agawu business among the respondents

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

kosei agawu
Seasonality vendor vendor
(n = 13) (n j)
Sell kosei/agawu throughout the year NO 0 0
Yes 13 7
. . No 0 0
th t th h
Use e same ingredien s roug out the year Yes 1 3 7
Change the amount of kosei/agawu, depending on No 13 7
the seasonal availability of cowpeas Yes 0 0
Change the size of kosei/agawu, depending on the No 4 4
seasonal availability of cowpeas Yes 9 3
Daily sale levels change, depending on the 13:5 3 (I)
l 'l b'l't f
seasona avai a l i y o cowpeas Don't know / Can’t tell 3 6 i

 

 

 

 

Source: Field survey in Accra, February and March 2007.

availability of cowpea grain”. On the other hand, nine kosei and three agawu16
respondents stated that they changed the size of kosei/agawu, and five kosei vendors
stated that the daily sales from kosei/agawu varied, depending on the seasonal availability
of cowpeas. This data set needs to be interpreted with caution because many
inconsistencies were found in the respondents’ answers”. However, the finding that many
respondents changed the size of kosei/agawu at different times of the year seems to be
relatively trustworthy (because it was a straightforward question and easy to answer

unless they had incentives to lie) and an important factor to consider for estimating the

benefits of dry cowpea meal.

 

'5 Including two kosei vendors who were in business for seven and eight months, respectively (these
respondents provided information for the time they had been in business). This applies to all the data
presented in this section.

One of the agawu respondents mentioned that she kept the same size of balls because the price of
COWpeas had not changed for two years at the place where she purchased COWpeas, implying that she may
change the size if the cowpea price fluctuates.

Some respondents increased the size of balls when cowpeas were more available while, according to
them, the sales were either constant or even increased. On the other hand, another respondent mentioned
that her sales increased while keeping the same size of balls. These answers are not possible if they always
used the same amount of ingredients and sold kosei at a constant price (assuming that the input-output ratio
does not change. One of the kosei respondents stated that fresh cowpeas swelled less. If this is true, when
COWpCaS are more available, less swelling would lower the sales of those vendors who use the same amount
of ingredients. This would make the respondents’ answers even more inconsistent.)

72

 

 

54.:

1;,9'
‘5.

l..

.1.”

Since the price per ball of kosei/agawu in Accra does not change depending on the
season of the year (G. A. Annor, personal communication, February 16, 2007), change in
the size of balls means change in the unit price of kosei/agawu. The degree to which
customers allow kosei/agawu vendors to reduce product size when cowpeas are scarce
would determine how easily the vendors could handle the seasonality by adjusting the
ball size. The less tolerant the customers are, the more difficult it would be for the
vendors to reduce the ball size, and therefore, the more valuable the dry cowpea meal
would be because dry cowpea meal is an input which would have little or no price
fluctuation. One of the agawu respondents who always sold the same size balls
mentioned that customers would complain if she reduced the size after she had increased
it, implying her customers were not tolerant of the change in size. On the other hand, one
of the kosei respondents who changed the ball size mentioned that customers knew about
the availability of cowpeas and would complain if the ball size was too small when
COWpeas were cheap, implying that her customers were tolerant to a certain degree of the
change in ball size. Unfortunately, the data are not extensive enough to estimate how
tolerant general customers are of the change in size of kosei/agawu balls. However, even
if customers are sufficiently tolerant to let the vendors change the ball size to offset the
change in cowpeas’ cost, dry cowpea meal would still help the vendors, because as long
as the price of meal is constant, it would allow them to constantly provide large size balls,
which would probably attract more customers. Discussions with industrial processor
respondents indicated that they sold their products at a constant price over a certain
period of time and raised prices only occasionally and in a stepwise fashion as their input

cost increased.

73

i-Ԥ\
he.

For the purpose of this study, the discussion has been limited to how useful dry
cowpea meal would be for kosei vendors. However, if the use of commercial cowpea
flour/meal in general (e.g., for fortification of traditional foods at home) is considered,
the benefit from reduced storage losses (due to insect damage) could be substantial——
assuming that cowpea grain is processed into flour/meal soon after harvest and
subsequent storage cost is less than for grain. The total demand for dry cowpea flour/meal

would determine how big the benefit would be.

4.2.1.8 Experience and interests in commercial dry cowpea meal/flour
The respondents’ experience and interests in using commercial dry cowpea

flour/meal to prepare kosei/agawu are shown in Table 4.6.

Table 4.6 Experience and interests of the respondents in commercial dry
cowpea meal/ flour

 

 

 

 

 

 

 

 

kosei agawu

Experience/Opinion vendor vendor

(n = 13) (n = 7)

Ever used commercial cowpea No 13 7
flour to prepare kosei/fiagawu Yes 0 0
Interested in using commercial No 7 l
flour/meal to prepare kosei/agawu Yes 6 6
Vendor in the market I 0

Would want to buy commercial Supermarket 0 l
flour/meal from: Wholesaler 0 1
(multiple choices) Processor 4 4
Other I l

 

 

Source: Field survey in Accra, February and March 2007.

No respondent had ever used commercial dry cowpea flour to prepare her
products. Most of the vendors had never used it because they had never seen it.

Among kosei vendors, six respondents were interested in using commercial

flour/meal. This number should be interpreted with caution because at the early stage of

74

~wr,’

kilns

ii’v‘

J}\\

t V'
I;

.J
A»

51.;

Yr-

ti

conducting interviews, it was not emphasized that good quality kosei could be made from
dry meal. Some respondents were not interested because they thought dry flour/meal was
used for making agawu and not kosei. After we started to clarify the respondents that we
were describing a special meal developed for preparation of kosei, all the kosei
respondents were interested in the product. In terms of the quality concern, one kosei (as
well as 1 agawu) respondent mentioned that the processor of the flour/meal might
contaminate the flour/meal by mixing maize into the COWpeas.

Those respondents who were interested in the meal were asked how much they
would be willing to pay. However, none of them could answer the question; they told us
they would need to first try making kosei using the meal. After samples of commercial
cowpea flour (400 g, packaged in a polyethylene bag) were obtained from processor
respondents, we showed a sample bag to the kosei respondents after the interviews were
conducted. All three kosei respondents who were shown the sample and told the price
(price at the processors’ level) mentioned that it was too expensive.

These results indicate that kosei vendors would be interested in using dry cowpea
meal if: (1) they are informed that the meal is not for preparing agawu but for kosei; (2)
they are assured that the meal is of a good quality (no contamination by other milled
grains); and (3) the price is attractive.

The agawu respondents, who use dry flour everyday for preparing their products,
showed a strong interest in the commercial flour'8 (6 out of 7 respondents answered that
they were interested). One of the respondents stated that she would only be interested if

the COWpea variety was the one she wanted. Another respondent stated she would be

 

'8 Of course, the B/C CRSP-developed meal was for preparing kosei, but the same questions were asked to

agawu vendor respondents as well without revealing to them that our focus was dry meal for preparing
kosei.

75

 

~Alv

-\

interested only if the quality of output would not change. The respondent who was not
interested gave the quality concern (maize contamination) as a reason. Vendors’ answers
to the question about acceptable price of flour and reactions to the sample varied across
the respondents: three respondents answered that they would have to try it first; two
respondents answered that they would pay a higher price per olonka than the price they
were paying for cowpea grain (33% and 50% higher, respectively); and one respondent
answered she would pay less than the price of her COWpea grain (14% less). The price of
sample bag was too high for all the respondents except one, who was very interested and
said she would accept the price (price at the processor’s level).

The benefits and potential of commercial dry flour to agawu vendors might be
very different from those of commercial dry meal to kosei vendors because of the
difference in the processing procedures for these two products. Therefore, to understand
the competitiveness of dry cowpea flour for agawu preparation, another study would be
needed.

The respondents’ most preferred outlets for cowpea meal/flour were meal/flour
processors, while vendors in the market, supermarkets, wholesalers, anywhere, and a
place close to the house were each mentioned by one respondent. No respondent

mentioned a small retail store. Many respondents preferred to directly purchase from a

processor because they thought the price would be lower.

4.2. 1. 9 Home Preparation ofKosei

Among the potential benefits of the new dry cowpea meal is that its availability

could increase the demand for cowpeas by increasing home preparation of kosei. Indeed,

76

 

1’7.
'..

“"’

rub

n-

'r
.5.

‘ii

hLi‘

as reported later, two cowpea flour processors who were interviewed mainly targeted
sales towards housewives rather than kosei vendors.

It should be noted that an increase in home preparation of kosei would not
automatically increase the demand for COWpeas: if home kosei preparation using
commercial dry COWpea meal results in replacing street-vended kosei by the same amount
of home-prepared kosei, it would simply reduce the sales of kosei vendors without
increasing the total demand for cowpeas. However, the demand for cowpeas would
increase if: (1) women who currently neither prepare kosei at home nor buy kosei from
street vendors—because they dislike the ways the vendors prepare it, or for any other
reasons—start preparing kosei at home using dry cowpea meal; (2) women who do not
currently prepare kosei at home but buy kosei from street vendors increase their
consumption of kosei by preparing an additional amount at home using dry cowpea meal;
or (3) women who currently prepare kosei at home from COWpea grain increase the
amount they prepare by using dry cowpea meal.

To assess how much incremental demand for cowpeas, if any, would be created
through these channels, a full survey on kosei consumption among targeted population
would be required. Weaning mothers who were interviewed to obtain information on their
weaning habits (data to be analyzed in Chapter 6) were also asked a brief set of questions
related to their consumption and home preparation of kosei to provide supplemental
information on this topic. Among the 30 respondents, 10 knew how to prepare kosei. Of
these 10 respondents, 8 were interested in dry COWpea flour/meal (details of the results

are reported in Appendix 7).

77

4.2.2 Custom Millers
4. 2. 2.1 Characteristicsggstom miller respondents

Characteristics of the 15 custom millers interviewed are shown in Table 4.7.

Table 4.7 Characteristics of custom miller respondents

 

 

 

 

 

 

 

 

 

 

 

 

 

 

custom
Characteristics miller

n = 15)

0-3 I

Number of years the 3'10 0
facility is operating 10-20 4
More than 20 6

Don’t know 4

Tpe of mill Plate mill 15
Size of mill 2'6 . . . 14
Missmg information I

Source of power Electricity 15
10 5

Horsepower of the 15 6
motor 20 1
25 3

Type of customers Mainly kosei/agawu vendors 13
(for cowpeas) Don’t know / Missing info. 2

 

Source: Field survey in Accra, February and March 2007.
The majority of facilities had been in business for more than 10 years. All the

respondents used a plate mill, 2-A type19 (except 1 respondent for whom the information

 

19 During the fieldwork in Accra, different types (or sizes) of plate mill were observed. The most custom
miller and grain flour/weaning food processor respondents used a 2-A type. According to the machine
dealer we met at the 11th Ghana International Trade Fair (Feb. 21 -— Mar. 8, 2007, Accra), there are l-A and
2-A types for grinding mill, and these are an international standard (Machine dealer, personal
communication, March 2, 2007). Basic product information on each type of the grinding mills sold by the
seller is shown below.

Product information on l-A and 2-A type grinding mill sold by a machine dealer at the
11th Ghana lntemational Trade Fair (Feb. 21 - Mar. 8, 2007, Accra)

 

 

 

 

 

 

 

 

 

Brake Approximate Approximate Diameter of
Type RPM Horsepower Output for Dry Output for Wet Plate
Required Material Material
l-A 550-650 6HP 180 kg/hour lOO kg/hour 10 inches
2-A 550-650 8HP 275 Mom 140 lglhour 12 inches

 

Source: Fieldwork in Accra, February and March 2007.

78

 

 

1., ‘.
dbix

 

 

is missing), which was powered by an electric motor”. The horsepower of motor varied
across the respondents, ranging from 10 to 25 HP. Most respondents mentioned that

kosei/agawu vendors were their main customers for processing COWpeas.

4.2.2.2 Milling Charge
Respondents reported that the milling charge for each food product was set by the
Greater Accra Co-Operative Food Crop Processors Union (the Union, hereafter) and that
all custom millers operating in the region were obliged (in theory) to follow those
common guidelines. Many respondents had a one-page price list that was printed by the
Union. According to one respondent, the Union reviewed the prices and issued a new
price list every three years. Therefore, the milling charge was expected to be the same
across the respondents. However, this was not the case: respondents reported a range in
the milling charges. One of the reasons seemed to be that not all the respondents had the
latest price list. The latest list was dated August I, 2005, while another list dated
September 2, 2002 was also observed being used (a wide range in the charges was found
between the two lists). Moreover, not all the respondents strictly followed the listed
charges. As a result, the milling charges for dry and wet cowpeas collected from the
respondents diverted from the charges on the list. This is shown in Table 4.8.
While the list noted the same charge for dry and wet cowpeas, respondents
generally charged more for milling dry cowpeas than for wet cowpeas. Compared to the
Charges on the current price list, the respondents’ charges were generally much lower.

Among the respondents, there was a difference in preference for dry and wet

 

 

2°.Severa1 respondents had more than one milling machine. The data reported in the table is for their
principal mill.

79

 

x

Table 4.8 Milling charges for dry and wet cowpeas among the 15 custom
miller respondents

 

Charge er olonka (¢)

 

 

 

 

 

 

 

 

Charge Dry Wet
ckopeas cowpeas
Current price list (Aug. 2005) 4,500 4,500
Older price list (Sep. 2002) 3,264 3,264
Respondents mean 3,400 2,550
(standard deviation) (890) (956)
Respondents median 3.000 2,000
Respondents mode 3,000 2,000

 

 

Source: Field survey in Accra, February and March 2007.

cowpeas. Out of the 15 millers, 8 charged more for milling dry than for wet cowpeas
because they needed to mill dry cowpeas several times to obtain a fine output, while wet
cowpeas required only one milling; 2 charged the same price; 2 charged more for milling
wet than for dry cowpeas and another 1 only accepted wet cowpeas when the customer
had a large amount because cleaning after wet-milling cowpeas was tediousm; and the
other 2 did not mill wet cowpeas (one respondent had never milled wet cowpeas, and the
other asked customers who brought wet cowpeas to go to his neighbor miller who had a
small mill [probably specialized in wet milling]).

All the respondents stated they offered a discount when customers brought a large
volume to mill. This was in accordance with the price list in which the charges for certain
commodities were listed for two or three different units (olonka, pan, and bag)—
apparently 22 with discounts for the larger units. However, it was found that many
respondents were offering a larger discount rate than those noted on the official price list.

Since data on the cost of custom dehulling/milling were not reported by many

 

 

2' The majority of respondents opened and cleaned the mill after milling, whether they milled dry or wet
cowpeas.
“Apparently” because it is not clear how many olonka of each commodity one pan and one bag contain.
The information collected from different respondents indicates that one bag usually contains between 40
and 60 olonka of commodities. If this is the case, the price list offered discounts for larger volumes.

8O

 

 

\
um .'

 

 

 

kosei vendor respondents, representative charges to be used for the budgeting analysis of
kosei preparation were estimated, based on the data collected from custom millers and
those kosei respondents who reported the charge 23. These estimated representative

charges are shown in Table 4.9.

Table 4.9 Representative custom milling/dehulling charges for cowpeas

 

 

 

 

 

‘MGIW cowpeas) Mill injfiwet co eas)
Charge Charge
Olonka (¢) Olonka (¢)
Up to 1 1,500 Less than 2 2,000
> I up to 2* 2,000 2 — less than 3 4,000
> 2 up to 3* 2,500 3 — less than 4 6,000
> 3 up to 4* 3,000 4 -— less than 5 8,000
> 4 up to 5* 3,500 5 —- less than 6 10,000
> 5 up to 6* 4,000 6 — less than 7 12,000
> 6 up to 7* 4,500 7 — less than 8 14,000
> 7 up to 8* 5,000 8 —— less than 9 16,000
> 8 up to 9* 5,500 9 — less than 10 18,000
> 9 up to 10 6,000
20 15,000
40* 37,500
60 60,000

 

 

 

 

"‘ Assumed value from other data points.
Source: Fieldwork in Accra, February and March 2007.

Interestingly, the data shows increasing unit costs for dehulling, as volume increased after
10 olonka. This might be due to the too small sample size (i.e., 5), or millers might
charge less for customers with small volumes in order to attract patronage (as opposed to

hand dehulling).

4.2.2.3 Custom millers and dry comea meal
As mentioned earlier in this chapter, there is a potential for any individuals having

access to a dryer and custom miller to produce dry cowpea meal if custom millers can

k

23 . . . - '

Since no information was obtained from custom miller respondents about the charge for custom
dehulling, representative dehulling charges were estimated based on the information obtained from five
kosei vendor respondents.

81

 

H"

1i
“.1.-
I
o~, 1'
r~ ‘vt

 

~,.-

can.

“1'

1.

b‘.

 

u,

 

 

mill cowpeas into the appropriate particle size. To explore this point, data were collected
to determine how easy it is for custom millers to mill COWpeas into different particle sizes.

The miller respondents reported that they adjusted the particle size in two ways:
by tightening or loosening the plates (i.e., changing the distance between the two plates)
and by changing the number of times the grain is milled. The tighter the plates are set,
and the more times the grain is milled, the finer the flour will be.

Although the ways that respondents adjusted the plates to obtain a specific
particle size seemed to be technically the same across the respondents, their perceptions
about the process were different: 11 respondents stated that it was easy for them to mill
COWpeas into different particle sizes, while 4 respondents stated that it was not easy.
Previous studies in food science have found that it is very important to mill dry cowpea
meal to a specific particle size range, if the meal is to be used for kosei preparation (50%
of particles must be larger than 400 pm in diameter; see Figure 2.2). Therefore, it is also
important for custom millers to be able to consistently achieve this particle size range, if
custom millers prepare dry cowpea meal. Success in consistently milling COWpeas to the
correct particle size would depend on the skill of each miller. However, generally
speaking, it would not be easy to mill cowpeas into the exact particle sizes using a plate
mill alone—the miller would have to also use a sieve to separate out large particles for
further milling (E. Sakyi-Dawson, personal communication, March 21, 2007). If the
miller used a hammer mill, the work could be done much more easily“. However, no

respondent used a hammer mill. A further survey of custom millers would be needed to

—¥

2’ Hammer mills, which can only be used for dry products, are more expensive than plate mills. However, it
is easier to obtain a desired particle size with a hammer mill because it has a built-in sieve. Also, with a
hammer mill, fewer millings may be required; cleaning of a hammer mill is easier, and the hammer lasts for
a long time, although the sieve has to be replaced more frequently (G. A. Annor, personal communication,
March 21, 2007).

82

 

 

 

 

find out how successfully custom millers can mill cowpeas into the right particle size

range for dry cowpea meal (with or without a sieve) and whether they are willing to do

this when asked by customers.

4.2.3 Industrial Cowpea/Soybean Flour Processors
4. 2. 3. 1 C haracteristicgficowfpea/mbean flouLprocessors

Among 10 grain flour/weaning food processor respondents, 6 were producing
either cowpea or soybean flour. The characteristics of these respondents are shown in
Table 4.10.

As described earlier in this chapter, while all the respondents were small- to
medium-scale companies, there was a wide range in scale among them. One-half of the
companies were managed by women and two by men (the other company appeared to be
co-managed by a woman and a man), and these companies ranged from a pure family
business operating in their house to a structured enterprise which hired personnel and
operated in an isolated production facility. The majority had been in existence for more
than 10 years. All of the respondents utilized a total of fewer than 15 workers, including
family members (information is missing for 1 respondent). The value of the products that
the respondents produced during February 2007 ranged from less than ¢50 million to
more than ¢100 million (information is missing for 1 respondent).

None of the respondents specialized in flour products—they all produced eight or
more different commodities25 . Commodities other than processed grain products included
honey, oil, spices, groundnut paste, and fruit jam. In fact, some respondents started their
business with other products such as spices and later began producing flour products.

_E

25 . . .
Not all of the products were necessarily produced when the interView was conducted.

83

 

Table 4.10 Characteristics of cowpea/soybean flour processor respondents

 

Characteristics

flour
processor
(11 = 6)

 

Sex of the manager

Female
Male

 

Missing information

 

Fewer than 5

 

Number of years the company 5—10
has been in business 11-20
More than 20
1-5
Total number of workers 6-10
(including family members) “-15

 

Missing information

 

Value of all products produced in
Feb. 2007* (¢; USS] “=- ¢9,200)

Less than 50 million
50-100 million
More than 100 million

 

Missing information

 

Total number of products
(including cowpea and/or
soybean flour)

1-5

6-10

11-15

More than 15

 

Whether producing cowpea
and/or soybean flour

Cowpea flour only
Soybean flour only
Both

 

Number of years the company
has produced cowpea/soybean
flour

Fewer than 2
2-5
6-10

 

Missing information

 

Share of cowpea/soybean flour
in the total revenue (Feb. 2007)

Less than 5%

 

Missing information

 

 

Major source of raw materials

 

Vendors in the market
Wholesalers
Owned farm
Middlemen/Sflipliers

 

u—.—-——ww—-——MN-—A—-NNNo————W—-Nwo—-w—_._~m

 

 

 

 

* Include one respondent from whom only the annual revenue in the past year was
available—the figure was divided by 12 and used as approximate monthly revenue.
Source: Field survey in Accra, February and March 2007.

COWpea flour was produced by two respondents, while soybean flour was
produced by five respondents. Although the respondents had been in business for a long
time, they only started producing these products in recent years, and the share of these

products in the total sales seemed to be small. These findings indicate that the production

84

l}, y\

‘ :-

blun'

in.

0:

IA.

(I!

('13——

of cowpea/soybean flour was just a small and relatively new part of the respondents’
entire business.

The respondents procured their raw materials from different sources: one
respondent purchased mainly from vendors in the market, one respondent from
wholesalers, one respondent had his own farm where he grew crops for his products, and
the other three respondents obtained their raw materials from “middlemen” or
“suppliers”—people who brought crops to the respondents’ facility, apparently from rural
production zones. Although the respondents were not asked how much they paid for raw
materials, it is expected that those respondents who owned a farm or who had access to

middlemen or suppliers paid lower prices for their inputs than the other two respondentszf’.

4.2.3.2 Processing procedure and equipment use

If the processing procedure for B/C CRSP-developed dry cowpea meal is different
from the procedure currently used by the respondents to mill cowpeas, whether that
difference causes any difference in the costs of production must be considered before
making cost calculations using the data collected from the respondents. Furthermore,
since the data collected from the four respondents were for soybean flour production, the
difference between the processing procedure for cowpea flour and that for soybean flour
also needs to be examined to justify the use of their data for estimating the cost of
production of dry cowpea meal. For this purpose, information was collected on how the

respondents produced cowpea/ soybean flour.

 

 

2’ In terms of the type of cowpeas, one of the two respondents who produced COWpea flour used Niger
because this respondent felt that Niger tasted better and kosei vendors also used Niger. The other
respondent used both domestically produced cowpeas and Togo. According to this respondent (one of the
larger processors), domestically produced cowpeas were hard and difficult to dehull, but easier to trace
back to their point of origin than imported cowpeas, an important consideration for assuring food safety.

85

L_

 

 

 

Data collected during fieldwork indicated that the procedure that the respondents
used to make cowpea flour was almost the same as the agawu vendors prepared their
flour, that is, dry-dehulling by plate mill followed by manual winnowing and dry-milling.
Therefore, the major differences compared to the processing procedure for the B/C CRSP
dry cowpea meal were that: (1) the respondents’ process did not include soaking and
drying; and (2) they milled COWpeas into flour rather than coarser meal. In the budgeting
analysis, the first difference was addressed by assuming that the respondents used a dryer
to produce cowpea flour and adding into the budget the assumed share of cowpea flour in
the cost of dryer use2 F or the second difference, it was assumed that processors
currently producing flour using a plate mill could also produce meal if they used a sieve.
Based on this assumption, the share of cowpea meal in the cost of sieve use was assumed
and added into the budget’28

The differences between the processing procedures for COWpea flour vs. soybean
flour were that the majority of respondents cooked soybeans before milling29 and some
respondents did not dehull soybeans. Thus, strictly speaking, to estimate the cost of
producing dry cowpea meal, data on the cost of production of soybean flour would have
to be adjusted to handle these differences in processing procedures. However, it would be

very difficult to accurately estimate the difference in costs of production associated with

27 The additional cost associated with soaking was ignored because such a cost seemed to be minor. For the
cost of dryer, monthly share of cowpea flour in the purchase price of the dryer was assumed for each
respondent based on the data collected and added to the cost of equipment 0f their budgets. Although ‘he
use of dryer would also increase other costs such as labor and fuel, such additional costs were very difficult
to estimate and therefore were not included into the budget

”This adjustment was made for those respondents who did not currently use a sieve or hammer mill to
produce cowpea/soybean flour. As done for the dryer, only the assumed monthly share of COWpea meal in
the purchase price of sieve was added to the budgets (i.e., potential increase in the costs other than the
egquipment cost was ignored).

9Respondents cooked soybeans: (1) to improve the taste; (2) to extend shelf life; and (3) because they
expected consumers to just add the soybean flour to other foods and consume without further cooking,
while cowpea flour was expected to be further COOked-

86

these identified differences in processing procedures. Therefore, the difference in the cost
of production between cowpea and soybean flour was assumed to be only the difference
in the cost of raw materials (i.e., cowpea and soybean grains).

Major equipment that processors would need to produce dry cowpea meal
(following the B/C CRSP-developed instructions) and sell outputs on a commercial basis
are: (1) a dryer (not limited to machines exclusively designed for a drying function, as
some respondents used a gas oven or roaster to dry their ingredients); (2) a mill (plate
mill [for dehulling] and hammer mill [for milling] or plate mill and sieve); and (3) a
packaging machine. Current possession of such equipment by the respondents is

summarized in Table 4.11.

Table 4.11 Equipment for producing dry cowpea meal and possession by
the respondents

 

flour
Equipment inventory processor
(n = 6)

 

Own any equipment that No

can serve as a dryer Yes

Own plate mill and sieve

M' Own only hammer mill
ill .

Own only plate mill

De not own mill

No

Yes

Source: Field survey in Accra, February and March 2007.

 

 

QO—‘N—NQO

Own sealing machine

 

 

 

 

 

All the respondents owned some type of equipment that could serve as a dryer.
With regard to mills, two respondents owned both a plate mill and a sieve. One of these
two respondents owned a hammer mill, as well, but was not using it for some reason.
Also, one of them described the sieve as a “sieving machine,” implying that it was a

mechanical sieve. The one respondent, who owned only a hammer mill, custom-dehulled

87

raw materials. Two of the respondents owned only a plate mill. The remaining respondent,
who did not own a mill, used a custom miller. The plate mills that we had the opportunity
to observe at the respondents’ facility were the same size as the mills owned by custom
millers (i.e., 2-A type) or slightly smaller. All the respondents packaged outputs at their
facility into polyethylene bags for sale. The most commonly used packaging machine was

a hand sealer (Figure 4.4). One of the respondents also used a foot sealer.

Figure 4.4 Hand sealer

 

Source: Author.

To summarize, most respondents processed cowpea/soybean flour using the same
technology as custom millers (i.e., plate mill), and most of them looked technically
capable of producing dry cowpea meal with their current equipment or with a small

investment (those who do not have a sieve would need to purchase one)”.

 

30 Originally, this study intended to identify optimal technologies to produce our target products (i.e., dry
cowpea meal and cowpea-Weanimix). However, since the technologies used by the industrial grain
flour/weaning food processor respondents and custom millers were very similar to each other, such an
analysis could not be conducted. Although large-scale multinational companies most likely used different
technologies to produce similar type products, the information was missing. It would be interesting to
explore why there were few differences in technologies among local processors—Le, were they constrained
in the types of technologies available to them, or was the technology observed the optimal one under the
current circumstances?

88

4.2.3.3 Demand for COWpea flour and its shelflifie

One of the two industrial processor respondents who produced COWpea flour sold
it mostly to retailers and exported a small amount. The other respondent mainly sold
cowpea flour to retailers (supermarkets), but also sold to restaurants, hotels, and
individuals, and exported a small amount. Both respondents mainly targeted housewives,
who prepared kosei“ at home. The cowpea flour produced by one of the processors
contained onion as an ingredient, while the other respondent produced 100% COWpea
floun

Of the two respondents, one was not producing COWpea flour when the interview
was conducted because of an equipment problem. Therefore, the questions about recent
demand were skipped. The other respondent mentioned that because the demand for
COWpea flour was irregular, the quantity of production varied from month-to-month. This
respondent, who stored cowpea flour at her facility, reported that shelf life of the flour
was one year. This answer confirms a potential advantage of dry cowpea meal—a more
stable input price for kosei vendors. Processors could purchase cowpeas and produce the
meal when the grain is inexpensive, and store the outputs for sale later during the
COWpea-scarce season, fixing the price at a season-average price.

Among the four respondents who produced soybean flour, rather than COWpea
flour, three mentioned that they either planned or had thought about producing COWpea
flour. One of the respondents commented that advantages of cowpea flour were that

consumers were more knowledgeable of COWpeas than soybeans and that the COWpea

 

 

3: The label of their products had the word “kose” or “koose,” both different spellings of kosei. However,
Since both products were flour, fritters made from their products would be more like agawu than kosei.
Therefore, these products might not gain popularity among consumers who prefer kosei over agawu.

89

flour could be used for preparing a greater variety of dishes than soybean flour. As a

constraint, another respondent observed that launching a new product would be costly.

4. 2.4 Retailers

Before discussing the data collected from retailers, it is important to discuss
whether or not retailers would have a role to play, if commercial dry cowpea meal
becomes available to kosei vendors. As reported earlier, the majority of kosei vendor
respondents said that they would prefer to buy cowpea meal directly from processors,
expecting that the price would be cheaper. Their expectation is definitely correct, if
processors would allow vendors to come to their facility and buy cowpea meal at the
same price as they sold it to other outlets such as retailers. Among the six
COWpea/soybean flour processors interviewed, three sold their products to individuals.
Therefore, there is no reason to assume that those processors would refuse to sell cowpea
meal directly to kosei vendors. Following these observations, the budgeting analysis
assumed that the kosei vendors could purchase dry cowpea meal directly from meal
processors. Therefore, a retail margin was not added to the estimated processor price of
dry COWpea meal when calculating the price that kosei vendors would have to pay for the
meal.

However, although it seemed to be less likely, it would still be useful to examine
the case in which kosei vendors purchased cowpea meal from intermediaries for two
reasons: (1) the analysis would provide information regarding how much the retail
margin would negatively contribute to the price-competitiveness of dry COWpea meal; and

(2) the potential customers of dry cowpea meal are not only kosei vendors. If housewives

9O

t") '
~ l

5‘. ti.

5,1

r

.i,

..n.
.

[u

are to buy dry cowpea meal to prepare kosei at home, they would probably prefer buying
it where they usually shop, rather than from the processors. If so, the price-
competitiveness of cowpea meal for housewives has to be examined at the retail level.
For these reasons, estimates of the retail margin for cowpea meal were derived from the
collected data, and the case in which the kosei vendors (and housewives) purchased meal
from retailers was examined in a sensitivity analysis.

As described in Chapter 3, during the fieldwork, three different types of retailers
were identified in Accra: grain/flour-type product vendors in the market, small shops, and
supermarkets. Among them, grain/flour-type product vendors seemed less likely to
become an outlet for industrially-processed dry cowpea meal because they were rarely
observed selling packaged products. Although few small shops were observed selling
locally-produced-and-packaged milled grain products, their retail margins were still of
interest because this type of shop was common in residential areas, seemingly serving
ordinary-level income customers. Therefore, if dry cowpea meal becomes available and
popular among housewives, these small shops would be a high potential outlet for the
product. As reported in the previous section, supermarkets were the major outlet for
industrially-processed cowpea flour. Therefore, their retail margins were included in the
analysis.

Since the data obtained from one of the four supermarket respondents were
inconsistent, these data were excluded from the analysis. The mean retail margins for

milled grain products32 were estimated for the remaining respondents (3 supermarkets

E

32 Including both cowpea/soybean flour and weaning foods. Since the sample size was limited for
cowpea/soybean flour and there was no reason to assume that the retail margins were different between
these two types of products, they were treated together. Therefore, the same retail margins obtained here are
used for the analysis of weaning foods in Chapters 6 and 7.

91

 

 

and 7 small shops)”. The results are shown in Table 4.12. Throughout this study, a retail
margin is defined in percentage terms, as the mark-up the retailers charge over the price

paid to their suppliers.

Table 4.12 Mean retail margin for milled grain products among small shop
and supermarket respondents (%)

 

 

 

 

 

 

 

Resgdents Min. Mean Max. S.D.*
Small shop (it = 7) 3 13 31 9
Supermarket (n = 3) 20 28 37 8

 

 

* 8.0.: standard deviation.
Source: Field survey in Accra, February and March 2007.

The data varied widely across the respondents: the mean retail margin among
small shops ranged from 3% to 31%, while that of supermarkets ranged from 20% to
37%. The mean margins of supermarkets were all higher than those of small shops,
except for the highest-margin small shop respondent.

The consumer price is not equal to the processor price plus retail margin. In
Ghana, a 12.5% of Value Added Tax (VAT) and a 2.5% of National Health Insurance
Scheme (NHIS)—total of 15%—are levied on all goods and services. While “locally
produced foods,” as opposed to “imported foods,” are among the goods exempted,
products such as flour are considered to be “processed” or “refined” products and are not
exempted (VAT officer, personal communication, March 29, 2007). The VAT and NHIS
are supposed to be collected at each stage of supply chain to the value added by each

actor. Figure 4.5 describes how the VAT/NHIS are collected using a numerical example.

 

 

33 . . . .
Details ofthe calculation to obtain mean values are reported in Appendix 6.

92

Figure 4.5 Collection of VAT/NHIS

Before VAT/N HIS was introduced

 

 

Processor Retailer Consumer

 

 

 

 

 

 

¢10,000 margin ¢12,000
20%

After VAT/N HIS was introduced

 

 

 

Processor Retailer

 

 

 

 

Consumer

 

 

 

¢ll,500

 

 

 

margin ¢13,800
I 20% I
¢10,000

¢13,500 (11,500 + 2,000)

¢1,500 (15% 0f¢10,000) ¢300 (15% Of¢2,000)
Government ‘—J l

 

L
‘fi

 

 

 

Source: VAT officer, personal communication, March 29, 2007.

If the demand were perfectly price-inelastic, when processors sold their products
to retailers, they would sell them at a 15% higher price than the original price (¢11,500 in
the above example). This extra revenue (¢1,500) would be paid to the government as
VAT/NHIS. Then, the retailers would set the margin 15% higher than their original
margin (2,000 + 300 = ¢2,300). Again, this extra revenue (¢300) would be paid to the
government. In the end, both processors and retailers would receive the same revenue as
they did before the VAT/NHIS was introduced (¢10,000 and 13,500 — 11,500 = 912,000,
respectively). However, consumers would pay a 15% higher price than before (¢13,800),
and the government would receive that 15% extra surcharge paid by the consumers
(¢1,800). Therefore, the formula for obtaining the consumer price is given by Pc = 1.15

(1 + M) Pp, where Pc is consumer price, m is retail margin, and Pp is processor price.

93

However, if processors increase the price of a product, it is expected that the

demand for the product will decrease (i.e., demand is not perfectly price-inelastic),
leading to a new equilibrium point where the quantity of the product produced and sold is
smaller, the consumer price is higher, and the processor price is lower than the original
values. Although the consumer price should be 15% higher than the processor price (due
to the VAT/NHIS; under the assumption of no retail margin), the difference between the
new consumer or processor price and the original price depends on the elasticity of
demand and supply for the product. In particular, the more price elastic the demand for
the product, the more the processor would have to absorb the cost of the tax. We would
expect that the demand for dry cowpea meal would be price elastic because of the
presence of a close substitute, namely wet-milled cowpea paste.

However, the purpose of this study is not to examine the effect of VAT/NHIS on
the price of dry cowpea meal. Rather, it is to examine the price-competitiveness of dry
COWpea meal as an ingredient in kosei. Moreover, as discussed later in Section 5.2.3, the
VAT/NHIS may not be collected when kosei vendors purchase dry cowpea meal directly
from processors. Therefore, for the budgeting analysis, two extreme cases were assumed:
(I) the VAT/NHIS is not collected; and (2) the VAT/NHIS is paid entirely by the
customers (i.e., the consumer price increases by 15%, while the processor price does not
change). In reality, the consumer price would likely not increase that much, while the
processor price would decrease, unless the supply of dry cowpea meal is perfectly price-
elastic. Although a lower processor price would decrease the profitability of dry COWpea

meal for industrial processors, this effect was not analyzed in the budgeting analysis.

The mean retail margins for milled grain products among small shop and

94

supeMarket respondents reported in Table 4.12 were modified to include VAT/NHIS

(15%). The results are shown in Table 4.13.

Table 4.13 Mean retail margin for milled grain products combined with
the VAT/NHIS among small shop and supermarket respondents

 

 

 

(%)
Respondents Min. Mean Max. S.D.*
Small shop (it = 7) 19 29 50 11
Supermarket (n = 3) 38 47 57 10

 

 

 

 

 

 

* S.D.z standard deviation.
Source: Field survey in Accra, February and March 2007.

Since the mean retail margins vary without showing any general tendency, a
representative value for the retail margin, combined with the VAT/NHIS, could not be

determined. Therefore, for the sensitivity analysis, different values from 20% to 60%,

with a 10% range (i.e., 20, 30, 40, 50, and 60), were used.

4. 2.5 Price 0[ Cowpeas

As reported earlier, six cowpea/soybean flour processor respondents purchased
raw materials from different sources, implying that the price they paid for
COWpeas/soybeans most likely varied. However, to conduct the budgeting analysis, a
representative price of cowpeas for a bulk purchase, which was derived from the prices
reported by kosei/agawu vendor respondents, as well as the prices observed in the Nima
market during the fieldwork, was used as an approximate price that the processor
respondents paid for cowpeas“. For the sensitivity analysis, the range of seasonal

fluctuation in cowpea prices was estimated using historical data obtained from the

_

34 . . . '
To save interView time, the processor respondents were not asked about the cost of raw materials. The
same representative cowpea price derived here is used across the respondents in the budgeting analysis.

95

Ghanaian Ministry of Food and Agriculture. Described below is how the values used in

the analyses were derived from the collected data.

4. 2. 5.1 Comma prices in Accra in March 2007
Prices of four different types of cowpeas (for a bulk purchase”) in March 200736,

which were reported by kosei/agawu vendor respondents”, as well as observed in the

Nima market, are shown in Figure 4.6.

Figure 4.6 Reported/observed cowpea prices in Accra in March 2007

 

6,000 4 xx
52000 — .
Atooo « '
:rooo « , , 3‘ --- .
:Looo < » 1
1.000 4 5 1

price (cedis/ kg)
x x a >
x

 

_ l

l 0 I l I ‘ ’ ' ‘1 l

l 0 20 4o 60 80 100 120 140 160 i
1 purchase quantity (kg) \
l

 

 

- 1' Burkina x Niger A Nigeria 0 TQng

Source: Fieldwork in Accra, February and March 2007.

 

 

As shown above, the prices varied within the same cowpea type. Although the
data are not presented here, price variation within the same type was also observed for

small quantity purchases. Unfortunately, the COWpea price data obtained from the MoFA

 

 

5 Defined for this study as equal to or more than 10 olonka.
The data obtained for February were too few to derive a representative price for February.

The majority of kosei/agawu vendor respondents providing these data also purchased COWpeas at the
Nima market.

96

 

"at:
A“. _

x 1'0
A”

5A-

iu i

at
+5.

7')!
ihr

v—fi

(It

were not separated by cowpea type. Rather, the data reported a single price for “COWpeas”
for each month. Therefore, it was impossible without additional data to say which types
were more expensive than others”.

Among kosei respondents, Niger was the most commonly used cowpea type.
Therefore, the prices of Niger were used to derive the representative cowpea price used in
the budgeting analysis. The mean price of Niger (for a bulk purchase in March, as shown
in Figure 4.6) was ¢5,635 per kg (sample size: 8; standard deviation: 1,225). Since most
processor respondents provided us with their cost of production in February, this
representative cowpea price for March was then converted into a representative price for
February, using the historical data obtained from the MoFA. This calculation is explained

below.

4.2.5.2 Historical data obtainflom the MoFA

The price data of cowpeas obtained from the Ghanaian Ministry of Food and
Agriculture are summarized in Figure 4.7.

Figure 4.7 shows the 5-year trend (2002-2006) in wholesale and retail prices of
COWpeas in real terms (base month = February 2007) at urban markets of the Greater
Accra Region. Both wholesale and retail prices were relatively stable over this period,
except in 2005. In 2004 and 2005, drought and locust infestation reduced production and
increased the price, particularly in Sahelian areas that exported COWpeas to Ghana.

Seasonal indices were constructed to examine the monthly fluctuation in the real

wholesale price: first, the mean real wholesale price over five years was calculated for

’8 For this very small sample size, it appears that Nigeria cost more than the other types. However, the
small sample size prevents us from making a definite conclusion. Casual observation during the study did
seem to indicate that Nigeria was priced higher than the other types.

97

Figure 4.7 Real monthly COWpea wholesale and retail price at urban
markets of the Greater Accra Region, Ghana (average of Accra, Ga,
& Tema Districts) (2002-2006)

 

 

 

 

 

 

10,000 ~
8,000a
£2
\ 6,000-
.22 .
B . '
0 4,000 ~~ .. ~+~
1
2,000+ l
0 fl : t ‘4 If A. L i l rel
N m m ~<r in in \o \O
3. a: e a 5'? C: O. 9 c? 9
g 5! .53 2 5; 3. :3 3 E 3.

 

P retail + wholesale] 1

Note: base month for real price = Feb. 2007; retail prices in Oct. and Nov. 2005 were the
same, not only for COWpeas but for all the commodities for which the data were
obtained. It is possible that either the data for Oct. or Nov. was mishandled and
recorded as the data for the other month.

Source: (for the price series) SRID, MoFA; (for the CPI) IMF.

 

 

 

each month; second, the mean real wholesale price of all the data points (from January
2002 to December 2006) was calculated; and third, the derived monthly mean prices were
indexed based on the mean price of all the data points as 100. The result is shown in
Figure 4.8.

The index for March was 101 while the index for February was 98. The minimum
index was 92 (November), and the maximum index was 111 (September). Using these
indices and the representative price of cowpeas in March 2007, as derived in the previous
section (¢5,635 per kg), the representative price of cowpeas in February 2007 and the
range in price fluctuation to be used for the sensitivity analysis were determined, as

shown in Table 4.14.

98

 

Figure 4.8 Seasonal index of COWpea wholesale prices at urban markets of
the Greater Accra Region, Ghana (average of Accra, Ga, & Tema
Districts) (2002-2006)

 

 

seasonal index

110 ~

 

 

 

 

80

l
I t 1' . F I I I 7

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

 

Source: (for the price series) SRID, MoFA; (for the CPI) IMF.

Table 4.14 Representative cowpea price paid by processor respondents in
February 2007 and estimated range in price fluctuation during the

 

 

 

 

 

 

 

year
Lowest February Highest
price 2007 price
Price (¢ / kg) 5,155 H 5,474 6,221

 

 

 

Source: Fieldwork in Accra, February and March 2007.

4. 2. 5.3 C onversionfrom olonka to kg and potential _problem with this method

To convert one olonka (volume unit) of cowpeas to kg (weight unit), a conversion
rate of 2.5094 was used throughout this study. The rate was obtained from the sample
purchased during the fieldwork, as described in Chapter 3. As price data obtained from
the MoFA were reported per kg, they did not need to be converted. However, it is not
clear whether the commodities were weighed at each purchase date or if a representative
conversion rate was used to calculate prices per kg. The price data obtained from

kosei/agawu vendor respondents were all per olonka (or containers containing a certain

99

 

number of olonka); therefore, these data were converted to prices per kg using the
representative conversion rate of 2.5094.

As mentioned in Chapter 3, the potential problem of this method for converting
olonka to kg is that the volume of one olonka varies because the amount that is heaped on
the olonka cup is subjectively decided by vendors. Langyintuo et al. (2004, p. 209)
reported “As a result of the non-standardized volume measures, the weight of a bowl of
COWpea on a given market day on the Tamale, Bolgatanga and Wa markets in Ghana,
ranges from 2.25 to 3.2 kg, 2.2 to 3.5 kg and 2.1 to 3.1 kg, respectively.” Therefore,
one olonka of cowpeas purchased by a kosei vendor respondent might have not weighed
the same as one olonka of cowpeas purchased by another respondent. Moreover, if the
MoFA’s per-kg price series were obtained using a representative conversion rate
throughout the year, then the data likely understated the seasonal variation in prices, as
one olonka probably contained less in terms of weight when cowpeas were scarce and
more when cowpeas were abundant. In addition, the difference in varieties and moisture
contents of grain may make the weight of one olonka of cowpeas vary significantly.
Although these potential problems were recognized, it was not possible to obtain better

estimates of the unit price of cowpea grain using the available data.

4.2.6 Dry CowpetLMeM Kosei Prepargtion Experiments

As described in Section 3.4, a set of experiments for kosei preparation using the
wet-milling method (i.e., field experiment with a kosei vendor, using her usual
procedures) and dry-milling method (i.e., laboratory experiment for preparing dry

COWpea meal and field experiment with a kosei vendor, using that meal) was conducted

100

twice. The major objectives of conducting these experiments were to obtain a
representative input-output ratio for converting: (I) COWpea grain to dry COWpea meal, (2)
cowpea grain and other ingredients to kosei, using the wet-milling method, and (3) dry
cowpea meal and other ingredients to kosei. The procedures to derive these ratios and the

results of the experiments are described below.

4. 2. 6.1 Results of the experiments

The results of the experiments are presented in Table 4.15.

Before analyzing the data, it should be noted that the first dry-milling method
replication (i.e., the laboratory experiment for making dry meal and the field experiment
for preparing kosei using that meal) failed: the kosei had a heavier texture and coarser
particles than wet-milled kosei, and the difference in taste was obvious. Both the kosei
vendor who conducted the experiment and UGL food scientists suspected that this failure
was due to inadequate whipping time, which did not allow cowpeas to swell sufficiently”.
Therefore, it was decided not to use the data obtained from the first dry-milling method
replication.

The second dry-milling method replication successfully produced kosei that had a
texture close to that of wet-milled kosei. This time, the dry meal was soaked for 116
minutes and then whipped for 55 minutes with occasional breaks. While the taste
improved, compared to the first replication, this long soaking and whipping time would

not allow kosei vendors to use dry meal for making a fine adjustment in the quantity of

k

39 For the first replication, dry meal and ground onion and pepper were given to the vendor. She started
Preparing the paste while she was frying wet-milled kosei. The paste (seasoned with the other ingredients)
was whipped with interruptions. The total soaking and whipping time was 97 minutes (i.e., from the time
the ingredients were mixed to the time she started frying), although the exact whipping time was not
measured.

101

Table 4.15 Results of the wet- and dry-milled kosei preparation experiments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Wet-Millino Method DiLMilling Method
lst 2nd lst 2nd
replication* replication* replication* repfiatiom
Cowpeas Raw (g) 2,543 2,497 2,540 2,438
Before milling (g)** no data 2,100 no data 2,091
After milling (g) no data 1,862a 1,965 2,066
Onion Raw (g) 1,033 1,550 1,680 1,700
[_ Before milling (g)** no data 1,400 no data 1,500
2 After milling (g) no data 1,242 a no data 1,164 b
E Pepper Raw (g) 182 40 3i 40
After milling (g) no data 35 a no data 31b
Salt (g) 56 c 53 c no data no data
Water (g)*** no data 1,700 no data 2,094
Oil (g)**** 1,242 1,159 463 723
t— t— Kosei (g) 5,223 5,000 4,300 4,800
8 2 Number ofballs 157 128 104 117
Weight per ball (g) 33 39 41 41
Notes:

* Italicized figures are estimated values. See the other notes for how they were estimated.

** The weight of cowpeas “before milling” refers to dehulled and winnowed cowpeas;
and the weight of onion “before milling” refers to peeled onion.

*** For the 2nd wet-milling method replication, the weight of water reported refers to the
water absorbed by cowpea grain during soaking. For the 2nd dry-milling method
replication, the weight of water reported consists of: (1) the water added to onion and
pepper before custom-milling these ingredients; and (2) the water added to cowpea
meal for whipping it.

**** The weight of oil reported consists of the oil absorbed by kosei and the oil
evaporated during frying (the amount evaporated would most likely be minor,
compared to the amount absorbed by kosei). To convert the unit of oil from liters to
grams, the ratio of 0.93 was used. The ratio was calculated from data in the United
States Department of Agriculture Commodity Requirements Vegetable Oil Products
(VP7 and VOIO) (Retrieved August 6, 2007, from

h_ttp://www.fsa.usda.gov/FSAtnebapp?area=home&subiect=coop&topic:p§.)

a. For the wet-milling method, cowpeas, onion, and pepper were custom-milled together.
The total weight of these mixed ingredients before and after milling was measured
during the 2nd replication. The loss rate was 0.1 13. This rate was used to estimate the
weight of each ingredient after milling.

b. For the dry-milling method, onion and pepper were custom-milled together with added
water during the 2nd replication. The loss rate was 0.224. This rate was used to
estimate the weight of each ingredient after milling.

c. For the 2nd wet-milling method replication, the weight of salt was estimated by the
weight of paste before frying, minus the weight of all the other ingredients (including
added water). Then, the share of salt in the total weight of kosei (53/5000) was used
to estimate the weight of salt in kosei in the lst wet-milling method replication.

Source: (lst replication) fieldwork in Accra, February and March 2007; (2nd replication)

G. A. Annor.

102

 

kosei they prepare, depending on the sales of the day. A further study to determine the
soaking and whipping time necessary to prepare a good quality kosei using dry cowpea
meal would have to be conducted.

Comparing the results of the first and second wet-milling method replications
with that of the second dry-milling method replication, the wet-milled kosei absorbed
more oil per gram of kosei ([1,242/5,223 = 0.24 g] and [1,159/5,000 = 0.23 g] vs.
[723/4,800 = 0.15 g])40.

With regard to the taste, the kosei prepared from dry cowpea meal (the second
replication) still did not have the exact same taste as the wet-milled kosei. Originally, this
study intended to include the analysis of consumers’ willingness to pay for kosei prepared
from dry meal, if its taste is different from kosei prepared from wet-milled cowpeas.
However, to conduct such an analysis, a whole food scientific study would be needed to
test the consumers’ acceptance, which was beyond the scope of this study. Therefore, for
this study, it was assumed that the difference in taste of the two types of kosei is

undetectable to general consumers.

4.2.6.2 [_nput-output ratio calculation 1: dry cowpea meal
During the second replication, 2,066 g of dry COWpea meal was produced from
2,438 g of COWpea grain. The input-output ratio was thus derived as 2,066/2,438 = 0.8474.

The loss of about 15% consisted of bad grains that were sorted out, husks that were

’0 The comparison between the number of kosei balls obtained from the wet-milling and dry-milling
method replications of this experiment does not provide a rigorous argument because the weight of
ingredients used in each replication differed (the purpose of this experiment was to obtain input-output
ratios within each replication, rather than to compare the results across the replications). However, how
many kosei balls can be made from a unit of cowpea grain is a crucial question for kosei vendors because it
directly affects the profitability of their business.

103

dehulled“, and other losses during the washing, dehulling, and milling procedures.

Since the first dry-milling method replication failed, this is the only estimate
available to use as a representative input—output ratio for dry cowpea meal. However,
since the figure was derived from data obtained by a laboratory experiment, it might not
reflect the reality: Harsh et a1. (1981, p. 186) warned that experimental data obtained
from farm research centers commonly achieve better outcome than under actual farm
conditions. This may apply to the input-output ratio derived here. Thus, the input-output
ratio that industrial processors would experience, if they produce dry cowpea meal at
their facility, might be lower because the laboratory experiment was conducted using well
maintained equipment, and the researchers paid attention to having as little waste as
possible. In contrast, small- to medium-scale industrial processors might use older
equipment and might not pay as much attention to minimizing losses because they mill in
bulk. During the interview, only a few processor respondents could provide the quantities
of input and output of cowpea/soybean flour”. The input-output ratio calculated using the
data obtained from those limited number of respondents ranged from 0.7488 to 0.8469.
Therefore, for the budgeting analysis, it was assumed that the input-output ratio of the
respondents does not change between processing flour and meal, while the
“representative” input-output ratio obtained from the laboratory experiment (i.e., 0.8474)

was only used as the maximum input-output ratio in the sensitivity analysis.

4' According to a UGL food scientist, about 6% of the weight of COWpeas are hulls (E. Amonsou, personal
Ezommunication, March 29, 2007).

Some respondents provided the quantities of both input (in olonka) and output (in number of containers
and kg per container). However, when the unit of input was later converted into kg, it turned out that their
outputs were heavier than their inputs. This result did not make sense because there must have been some
waslte as a result of the processing. Therefore, these respondents’ input-output ratios were not used in the
ana ysis.

104

4.2.6.3 lnput-ouiput ratio calculation 2: kosei

A representative kosei input-output ratio was needed for: (1) the wet-milling
method of kosei preparation, to estimate the weight of kosei prepared by kosei
respondents who could not or did not want to report their revenue on the day before the
interview was conducted; and (2) the dry-milling method of kosei preparation (i.e.,
preparation using dry COWpea meal), to calculate the weight of dry COWpea meal that each
respondent would need to prepare the same weight of kosei as they prepared by the wet-
milling method. First, the representative kosei input-output ratio for the wet-milling
method was calculated based on the data reported in Table 4.15, using the procedure
described below.

To prepare kosei, the main ingredients—cowpea grain, onions, and peppers—are
generally dehulled or peeled, mixed with each other, and milled into paste. Both
dehulling/peeling and milling involves waste. The waste rate would be different across
kosei vendors, depending on how they prepared the ingredients. Water is added to form
paste, but all three main ingredients originally contain different percentages of moisture.
To consider the water content in kosei as a whole, it is necessary to separate the moisture
content out from the other ingredients. The ratio of the weight of ingredients that are
dehulled, milled, and dehydrated to the weight of raw ingredients—called “dry equivalent

rate” hereafter—can be mathematically expressed as follows:

105

ry.:ll—dyll—finl—hi) (1)

where,
r = dry equivalent rate for the wet-milling method,
i = raw material (COWpea grain, onion, and pepper),
j = kosei vendor, d = dehulling/peeling waste rate,
f== milling waste rate, h = moisture content

Using this expression, the weight of kosei can be decomposed into the weight of each

ingredient as follows:

919' = ’4'ch + ’01 on + *2/ij + ij + Q!) + Qsj (2)

where,
Q = quantity in kg, k = kosei, c = cowpea grain, o = onion
p=pepper, v=water, l=oil, s=salt
It is a reasonable assumption that the ratio of water and oil to the dry equivalent

ingredients does not fluctuate much when kosei is made by experienced vendors. This

assumption is expressed as follows:

Q.)- = 81' (rchcj + 52'on + rijpji (3)
Q0 = (01(2ch! + r0} on + er’Qp/‘j (4)
where,

e = ratio of water to dry equivalent ingredients in terms of weight
(p = ratio of oil to dry equivalent ingredients in terms of weight

By substituting ij in equation (2) with the expression (3) and Q11. in equation (2) with the

expression (4), and manipulating the new equation algebraically, a kosei input-output

106

equation (based on weight, rather than volume) can be derived as:

.=a~r. .+r. .‘i'r. .+ .
Q"! J i e Q61 0] Q01 p1 Qt?!) Q51

where, aj =I+ej+goj

Based on the data obtained from the second wet-milling method replications and

the moisture content of each ingredient reported by Leung (1968), the values of dry

equivalent rates and aj—called “input-output conversion rate” hereafter—were calculated,

as shown in Tables 4.16 and 4.17.

Table 4.16 Dry equivalent rates of the kosei preparation experiment (wet-

 

 

 

 

 

 

milling)

rc = (1 - dc) (1 -fi-) (1 - he)
2,100

0.67 — 1—0.113 1-0.108
2,497

ro = (1 ~40) (1 —fo) (1 -ho)
1,400

0.09 — l — 0.113 1 — 0.885
1,550

rp : (1 _dp) (1 ".fp) (1 Thp)

0.23 1—0 1—0.113 l—O.742

 

 

Note: since the dehulling/peeling waste and milling waste were not measured during the
first replication, the rates were calculated based on the data collected during the

second replication.

Source: (experiment) G. A. Annor; (the values of ho, ho, hp) Leung (1968).

107

 

Table 4.17 Input-output conversion rates of the kosei preparation
experiments (wet-milling)

 

 

 

 

1 Qk = a( re QC + ro Q0 + rp Qp )+Qs

lst rep . 5.223 2.83 2.543 1.033 0.182 0.056
7

2nd repl. 5.000 2.73 0'67 2.497 0'09 1.550 0'3 0.040 0.053

mean a 2.78

 

 

 

Note: since the dry equivalent rates could not be calculated for the first replication, the
rates calculated for the second replication were used as an approximation.

Source: (lst replication) fieldwork in Accra, February and March 2007; (2nd replication)
G. A. Annor.

Initially, it was expected that the values of dry equivalent rates and input-output
conversion rates would not vary much across the respondents; if dry equivalent rates vary,
this means that the vendors’ rates of wastage of ingredients (during processing) vary; and
if input-output conversion rates vary, this means that the weight ratios of dry equivalent
ingredients to oil and water vary, which implies that the taste of the kosei would likely
vary. Such variations might be wide among less experienced vendors, but less among
experienced vendors. To examine this hypothesis, the value of input-output conversion
rates was calculated for 6 out of 13 kosei vendor respondents, who could provide enough
information for making this calculation, assuming that the dry equivalent rates were the
same as the rates obtained from the experiment across the respondents. The results are
shown in Table 4.18.

The results indicated that the input-output conversion rates, estimated from data
provided by the six kosei vendor respondents and the experiment, ranged from 1.78 to
4.28, with a mean of 2.54, standard deviation of 0.84, and median of 2.42. The variation
among the respondents was wider than expected, and no clear relation was found between
the level of the input-output conversion rates and the experience of respondents.
Therefore, to estimate the weight of kosei produced by the other kosei respondents, it was

decided to use the median value of 2.42. Using this value and equation (5), the weight of

108

Table 4.18 Input-output conversion rate across kosei vendor respondents,
usrng the wet-milling method

 

 

 

 

 

R Expe-

esp. .

number rience Qk = a (r c QC + r 0 Q0 + rp Qp) + Qs

(year)

3 15 7.59 4.28 2.51 0.71 0.09 0.07
4 3 3.48 1.95 2.51 0.71 0.03 0.08
7 0.6 2.54 2.12 1.67 0.57 0.06 0.04
9 9 7.42 1.78 0.67 6.02 0.09 0.48 0.23 0.10 0.18
10 >20 10.81 2.42 6.24 1.78 0.23 0.25
13 >20 1.82 2.47 1.04 0.20 0.04 0.04
Experiment 2.78

Meana 2.54

(standard deviation) (0.84)

Mediana 2.42

 

 

 

Note: the value of Qk was calculated as [revenue] / [price per ball] X [weight per ball];
the calculation to derive QC for Respondents 10 and 13 as well as Q0, Qp, and Q, for
Respondent 9 included assumptions due to missing information; the values of Q0 for

all the respondents as well as Qp for Respondents 7 and 13 were calculated as [price

of onion/pepper for the amount used] / [representative price per kg of onion/pepper
(see Appendix 3)]. Details for calculation of these values are available from the

author upon request.
Source: Fieldwork in Accra, February and March 2007; Table 4.16.

kosei prepared the day before the interview was conducted was calculated43 for those
respondents who either could not or did not want to report their revenue.

Second, the input-output equation for the dry-milling method was derived based
on the same logic that was followed for the wet-milling method:

In this formula, the main input is dry cowpea meal. When dry meal is used,
onions and peppers have to be peeled and milled separately from cowpeas. Therefore, the

dry equivalent rates are modified as follows:

¥

’3 Using: (1) the weight of cowpea grain, onion, pepper, and salt used by each respondent (estimated using
the data obtained from each respondent) and (2) the same dry equivalent rates derived from the experiment
across the respondents. Then, the revenue was also calculated for each of these respondents, usmg the
following formula: [revenue] = [price per ball of kosei] X [estimated weight (kg) of kosei prepared the day

before the interview was conducted] / [average weight (kg) per ha" 0f kosei].

109

um =(I—hm) (6)

uoj =(1—dchI—f5j11—h0) (7)
“pj=(1_dij1—f;jxl’hp) (8)
where,

u = dry equivalent rate for the dry-milling method,

m = dry COWpea meal,

j“ = milling waste rate: this rate can be different from f because
onions and peppers are now milled separately from
cowpeas, perhaps involving a different milling method (e.g.,
they can be ground using a blender at home)

Based on the same assumption of a constant ratio of water and oil to the dry equivalent

ingredients, the kosei input-output weight equation for the dry-milling method can be

expressed as:

5]

ij=a,-(u Q +u0jQ ijj+e <9)

. ,+u .
m 'm/ 01 a

Note that aj does not change, whether vendors wet-mill cowpeas or use dry cowpea meal.
By moving Qmj in equation (9) to the left hand side, the weight of dry cowpea meal

needed to produce Q19 is expressed as:

Q Q

' 0:.qu 0-—aj",,j pj
S
Qmjz J J J (10)

an
jm

110

Finally, by dividing both sides of equation (10) by ij~, the formula to calculate the weight

of dry con/pea meal needed to prepare 1 kg of kosei is expressed as follows:

 

. .— .—a.u. .—a.u. .
QM] = Q0 Qty 1 01 Q0) 1 p) Q]?! (11)
ij jqukj

To use equation (11), the values of dry equivalent rate for each ingredient are
needed. Representative dry equivalent rates were derived using the data obtained from

the second dry-milling method replication. The results are shown in Table 4.19.

Table 4.19 Dry equivalent rates of the kosei preparation experiment (dry-

 

 

 

 

 

 

 

 

milling)
um = (l “ hm)
0.93 1 —- 0.075
“0 = (1_do) (1 'f‘ko) (1 “170)
1,500
0.08 —-- 1 - 0.224 1 — 0.885
1,700
up = (1 — dp) (1 —f‘*p) (1 - hp)
0.20 1 — 0 1 - 0.224 1 — 0.742

 

Source: (experiment) G. A. Annor; (value ofhm: 7%-8°/o) R. D. Phillips.
The values obtained in this section were used to calculate budgets for the kosei

vendors. The model used and the results are reported in detail in the following chapter.

4.3 Summary

Interviews with street vendors of kosei revealed that there were actually two types

of fried cowpea ball products: one made from wet-milled cowpeas was called kosei, and

111

the other made from dry-milled COWpeas was called agawu. Among 20 respondents, 13
were kosei and 7 were agawu vendors. The scale of business varied across the
respondents. Most respondents used custom millers to dehull cowpeas, and all of them
used custom millers to mill cowpeas.

Potential benefits of dry cowpea meal included: (1) reduction in preparation time
of kosei; (2) fine adjustment of the quantity of kosei to prepare depending on the sales of
the day; (3) more stable price of input (i.e., meal) compared to cowpea grain; and (4)
increase in consumption of COWpeas through increasing home preparation of kosei.

With regard to the reduction in preparation time, it was found that it would not be
easy to accurately estimate how much kosei vendors would value the time that they could
save by switching from the use of cowpea grain to dry cowpea meal. The amount of time
saved would depend on: (1) how they prepare other food products that they sell along
with kosei; (2) change in soaking and whipping time; and (3) how easily they can grind
ingredients other than cowpeas.

The use of dry cowpea meal to make fine adjustments of the quantity of kosei to
prepare seems to have two potential constraints: (1) vendors who currently fry a whole
batch of cowpea paste at once would have to accept less convenient not-all-at-once
preparation; and (2) they would also have to be able to accurately predict the amount of
the sales of the day, because soaking dry meal takes time.

The statement of kosei vendors indicated that they would be interested in using
dry COWpea meal if: (1) they are informed that the meal is not for preparing agawu but
kosei; (2) they are assured that the meal is of a good quality (no contamination by other

milled grains); and (3) the price is attractive.

112

All custom millers used 2-A type plate mills, and none of them used hammer
mills. Among the respondents, there was a difference in preference for milling dry and
wet cowpeas. Whether dry cowpea meal for preparation of kosei could be produced by
custom millers would depend on how easily they could mill cowpeas into the correct
particle sizes. Although the ways to adjust particle sizes seemed to be technically the
same across the respondents, some stated that it was easy for them to mill COWpeas into
different particle sizes, while others stated that it was not easy.

While potential processors of dry cowpea meal would range from individuals to
large-scale multinational companies, all the processor respondents interviewed were
small- to medium-scale local companies. The respondents produced different types of
food products along with grain products. Most of the respondents producing
cowpea/soybean flour used the same technology as custom millers (i.e., plate mill). The
production of cowpea/soybean flour seemed to be just a small and relatively new part of
their entire business. Housewives were the major target of COWpea flour.

There was no new major finding about the non-price-related constraints for the
industrial production of dry cowpea meal. The processor respondents producing cowpea
and/or soybean flour seemed to be capable of producing dry cowpea meal with their
current equipment or with a small investment.

Among the retailer respondents, a wide variation was found among their retail
margins. However, based on the information obtained from kosei vendor and processor
respondents, it was assumed for the budgeting analysis that kosei vendors would
purchase dry cowpea meal directly from processors, without paying retail margins. On

the other hand, it was found that the meal would have to be sold with the Value Added

113

Tax/National Health Insurance Scheme, with a rate of 15%.
Also reported in this chapter were the representative prices of covvpea grain
derived from the price data collected during the fieldwork, as well as the results of dry

cowpea meal and kosei preparation experiments.

114

CHAPTER 5
DRY COWPEA MEAL FOR PREPARATION OF KOSEI

—-BUDGETING AND SENSITIVITY ANALYSIS

5.1 Budgeting Analysis—Model
A budgeting analysis was conducted to assess both the profitability of: (1)
industrial grain flour processors to produce dry COWpea meal; and (2) kosei vendors to

use commercial COWpea meal to make kosei.

5.1.1 Returns to Meal Processors
The unit return to meal processors—retums to processors’ management and

investment for 1 kg of dry cowpea meal—can be expressed by the following equation:

17'." c’."
fl 2P”! _ ft (1’))

m j t at
Q): Q)!

 

 

where,
17 = return, m = dry cowpea meal, j = meal processor,
t = time period, Q = quantity in kg, P = price per kg,
C = total cost payment
The total cost payment consists of the components noted in equation (13) below. Since
the processor respondents produced cowpea/soybean flour and other products, most of
the cost components were joint costs, which were not paid separately for each product.

Incorporating this fact into equation (12), the cost of producing dry cowpea meal is

expressed as follows:

115

m __ m
C]! — PC? MC!" + 6]" gWMy't + ZygtEejt + 21h?! ant (13)
e n

where,

c = cowpea grain, M = raw material in kg, 5 = share in wage,

W = wage per worker, w = worker, e = piece of equipment,

)2 = share in equipment use, E = equipment payment,

n == other cost component
(electricity, fuel [excluding fuel for vehicle], water, rent,
transportation [including fuel for vehicle], printing and
stationery, telecommunication, packaging material, and
miscellaneous)

2. = share in other cost component, X = other cost payment

Although desirable, it is virtually impossible to accurately measure the share of cowpea
meal in each cost component (i.e., cost share associated with producing cowpea meal).
For example, to estimate the wage cost share would require watching all the workers to
measure the time each worker spent on producing cowpea/soybean flour. To accurately
estimate the transportation cost share would require measuring the amount of each
product that the respondents put into their car each time they sell their products.
Therefore, approximated values were used to estimate the share of cowpea meal in each
cost component, unless the respondent was able to provide costs on a product-by-product

basis. The values used in this case study are as follows:

97}
291g.
8

(1) : share in the quantity (weight) produced

 

where, g = product produced by the respondent

The formula above was used to estimate the cost of wage, electricity, fuel, water,
and transportation.

116

PQO

jt .
(2) -——Pg—g. share in the value of products produced
2 Q]!
g
The formula above was used to estimate the cost of rent and telecommunication.
q e . .
(3) 2g QQ/m /jg :share in the equrpment use time
qg
gee ej
where,

q = hourly output
(Therefore, qgej denotes “hourly output of product g by equipment
e for processor j”)

ge e = products produced using equipment e

The formula above was supposed to be used to estimate the cost of equipment.
However, information on the hourly output of each product for each piece
of equipment was almost never available. Therefore, an approximated
value was used most of the time (e.g., for a plate mill, the weight of each
product processed using the mill during the month was multiplied by the
number of milling necessary to process that product [reported by the
respondent], and the proportion of cowpea/soybean flour in the total
production for this variable was calculated and used as an approximation).

”7
(4) fl : share in the number of packages

24%

g

 

where, A = number of packages

The formula above was used to estimate the cost of printing and stationery.

Using the model described above, the cost of producing dry cowpea meal was
first estimated for each respondent, using the data that they reported on the cost of
Producing cowpea/soybean flour. The potential processor price per kg of dry cowpea

meal was then estimated, assuming that the respondents would set the price so they could

117

11k

receive the same level of returns as they were receiving for producing cowpea/soybean

floun

5.1.2 Returns to Kosei Vendors

Unit returns to kosei vendors (i.e., return for 1 kg of kosei) can be expressed by

the following equation:

fl=l_Z£/_Q_’J (14)
Q19 )8] l ij

where,

j = kosei vendor, k = kosei, Pb = price per ball of kosei,
[3’ = weight in kg of one ball of kosei,
i = cost component

(cowpea grain or dry cowpea meal, onion, pepper, salt,
water, oil, fuel, custom dehulling and/or milling)l

Using this model and the data collected from 13 kosei vendor respondents, three
representative budgets for preparing 1 kg of kosei were constructed as follows:

(1) the payment for each cost component (per kg of kosei) was calculated for each

of the kosei vendor respondents;

(2) from the 13 observations, the median values of the following variables were

selected, excluding assumed values and outliers)? (i) calculated payment for

' Note that in equation (2) (see Section 4.2.6.3) the notations of ij (weight of water used) and Q1} (weight
of oil used) represented only the quantity of water and oil absorbed in kosei and did not include any waste.

In equation (14), the notation PijQij was used for the convenience, but the costs of water and oil that were

actually collected from the respondents and used for calculating budgets were the total payment for water
and oil (i.e., including waste)-

Assumed values refer to those values that respondents could not provide and therefore were assumed,

based on the information reported by others; outliers were defined as values outside mean i 2 standard
deviations (calculated excluding assumed values).

118

each cost component, (ii) weight per ball of kosei (obtained from the samples
of kosei purchased during the interview), and (iii) weight of dry cowpea meal
needed to prepare 1 kg of kosei (calculated using equation [11])3;

(3) for each of the above variables, the values plus one standard deviation
(recalculated excluding outliers found in the previous step) and the values
minus one standard deviation were calculated, respectively;

(4) the budget using the median was defined as “median representative budget
(M8),” the budget with the median plus one standard deviation was defined as
“least profitable representative budget (LB),” and the budget with the median
minus one standard deviation was defined as “most profitable representative
budget (PB);”4 and

(5) the total cost of each budget was calculated as the sum of the cost components
in the same budget; the revenue was calculated as the price per ball (i.e., 500)
divided by the weight per ball (i.e., kg/ball) of the same budget; and the return
was calculated as the revenue minus the total cost of the same budget.

Then, to estimate how the returns would change if dry meal were substituted for

COWpea grain, each representative budget was modified as follows: (1) the cost of COWpea
grain was replaced by the estimated cost of dry cowpea meal5 ; and (2) the cost of custom

dehulling and milling, which vendors would not have to pay anymore if they used dry

3 Since this value is highly influenced by the value of input-output conversion rate, the values derived using
the assumed input-output conversion rate were also treated as assumed values and excluded from the
4calculation of representative budgets.

One might be suspicious about this method because a vendor could pay a higher or lower unit cost than
other vendors depending on the cost components. However, a positive correlation was found between the
respondents’ per-kg-of-kosei payment for cowpea grain and payment for oil, which were the two major cost
components. This finding supports the method used to construct the LB and PB.

Calculated as the estimated price of dry meal (per kg) times the weight (kg) of dry meal needed to prepare
1 kg of kosei (calculated for each representative budget).

119

 

u‘i.

meal, was set to zero.

5.2 Budgeting Analysis—Results

5.2.1 gadgets for PMinLCowmegMeal

Budgets for industrially producing dry cowpea meal were constructed using the
model described in the previous section. Unfortunately, not enough information for this
analysis could be obtained from two out of six cowpea/soybean flour processor
respondents. Also, due to many missing and apparently erroneous data values for key
variables reported by the remaining four respondents, it was necessary to utilize various
assumptions in the analysis“ The results of the analysis are presented below.

For all respondents, the cost component having the largest share in the total cost
was the raw material (i.e., cowpea grain)—the ratio of the payment for cowpea grain to
the revenue from cowpea meal (i.e., processor price per kg of cowpea meal) was 22% to
39%, equal to ¢6,510 to 939,639 per kg of cowpea meal (Table 5.1. See also Figure 5.1).
However, the same representative cowpea grain price derived in Section 4.2.5 (Table
4.14) was used across the respondents except for Respondent 4, from whom the unit cost
of production was obtained from the respondent’s accountant. In reality, all four
respondents had their own procurement systems for raw materials (see Section 4.2.3.1).
Therefore, the price that the respondents actually paid for cowpea grain might have been
higher or lower than the representative price. If this was the case, the actual unit cost, and
therefore the share in the total cost, of cowpea grain would be higher or lower than

presented in Table 5.1.

 

 

6 Details of the problems encountered and the methods used to handle those problems are available from
the author upon request. Some of them are reported in the rest of this chapter as well as in Appendix 4.

120

 

Table 5.1 Estimated processor price, cost, return per kg of COWpea meal

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Resp 1 Resp2 Resp3 [ Resp4 *Aj
Processor price (¢; VAT/NHIS exclusive) 29,705 21,768 27,035 E 24,639]
Raw material (i.e., cowpea grain) 6,510 6,887 7,310 l 9,639
Wage 3,925 431 6,044 2,921
Equipment 1,075 189 742 N.Av.
Electricity 491 127 204
Fuel (excl. fuel for vehicles) 589 736 954 2,500
Water 0 68 51
Rent 0 0 196 0
Transportation (incl. fuel for vehicles) 3,739 712 3,344 2,599
Printing & stationery 4,861 3,000 1,206 800
Telecommunication 1,508 239 491 N.Av.
Packaging material 550 1,100 733 1,200
Miscellaneous 1,789 260 636 N.Av.
Total cost (¢) 25,037 13,750 21,911 N.Av.
Return (¢) 4,668 8,018 5,123 N.Av.
Total revenue of the month (million it; approximate) 40 100 50 N.Av.
Share of covmea/soybean flour in total revenue 0.02 0.04 0.04 N.Av.

 

 

 

 

 

 

 

Notes on assumed values (italicized figures): (1) raw material cost of Resp2 was
calculated based on the input-output ratios obtained from other respondents; (2) cost
of a sieve (included in equipment cost) of Resp3 was assumed, based on the values
obtained from other respondents; (3) cost of fuel of Resp2 was assumed based on the
values obtained from other respondents; (4) transportation cost of Resp2 was
calculated using information obtained from other respondents; (5) transportation cost
of Resp4 is the mean of the values of the other three respondents; and (6) costs of
packaging material of Resps 1 and 3 were assumed, based on the values obtained
from other respondents. Details of the methods used for calculation are available
from the author upon request.

*Resp4: the values were not derived using the model described in Section 5.1.1 but
obtained directly from the respondent’s accounting record for cowpea flour that was
produced sometime in 2006 (not in February 2007. Since the information was
missing for which month of 2006 the record was, the data could not be adjusted for
inflation); the respondent gave a total of ¢2,500 for electricity, fuel (excluding fuel

for vehicle), and water, rather than individual amounts; N.Av.: not available.
Source: Field survey in Accra, February and March 2007; secondary data used in other

parts of this thesis.

121

Figure 5.1 Estimated processor price, cost, return per kg of cowpea meal
(Resps 1, 2, and 3)

 

 

 

 

 

 

 

30,000
I return
25,000 x miscellaneous
I packaging material
20’000 _ 72 telecormunication
:15 printing & stationery
.fl ,-. transp. (incl. car fuel)
'8 15,000 -
o I rent
tr. water
IO’OOO l as fire] (excl. car fuel)
ii electricity
5900 ‘ 6 equipment
a wage
0 J I; raw materials
Respl Resp2 Res p3

 

 

Source: Table 5.1.

The wage paid per kg of cowpea meal varied across respondents, from ¢43l
to ¢6,044 in absolute values, and from 2% to 22% in terms of the ratio to the revenue.
Due to a smaller number of employees and a lower wage rate with a larger scale in
production, Respondent 2 paid a much lower unit wage, compared to Respondents 1 and
3.

The unit transportation cost7 varied widely across the respondents, ranging
from ¢7l2 (3% of the revenue) to ¢3,739 (13% of the revenue). As mentioned in Section
4.2.4, for the analysis of this study, kosei vendors were assumed to come to the
processors’ facility to purchase dry cowpea meal. Therefore, kosei vendors, rather than
processors, were assumed to pay the transportation cost. However, it was difficult to
x;—

7 Composed of: (1) payment for fuel; (2) estimated monthly share of purchase price of owned vehicle(s);
and (3) maintenance cost of owned vehicle(s).

122

estimate such a cost that would potentially be paid by kosei vendors. Therefore, it was
further assumed that the unit transportation cost that processors currently paid to deliver
their products to their outlets were close to the unit transportation cost that kosei vendors
would have to pay to visit processors for purchasing meal. Based on this assumption, the
transportation costs were left in the budgets—these values represent the additional cost
for kosei vendors to use dry cowpea mealg.

Printing and stationery was another important cost component for Respondents l
and 2 (¢4,861 per kg of cowpea meal [16% of the revenue] and ¢3,000 [14%],
respectively) while not as important for Respondents 3 and 4 (¢1,206 [4%] and ¢800
[3%], respectively). It was expected that a larger scale of production would drive down
the unit cost of printing and stationery. However, such a trend was not observed among
the respondents.

The equipment, electricity, fuel (excluding fuel for vehicles), water, rent,
telecommunication, packaging materialg, and miscellaneous costs did not contribute
much to the total cost.

The estimated processor price of cowpea meal ranged from ¢21,768 to ¢29,705
per kg.

The returns to the capital, management, and perhaps labor of the respondents and

8 Respondent 4 actually did not pay for transportation at all because the suppliers brought raw materials to
the respondent’s facility and the customers also came to the facility to buy outputs. However, for the same
reason that kosei vendors would have to pay transportation costs, the mean unit transportation cost of the
Sother three respondents was added to the budget of this respondent.

One of the respondents stated that a high packaging cost was a constraint to introduce a new product. She
most likely meant that the cost of creating a new package (as. (13518111118 a label) was high rather than the
cost of packaging material.

123

their family membersIO (excluding Respondent 4 whose total cost was not available)
ranged from ¢4,668 to ¢8,018 per kg of COWpea meal or from 16% to 37% of the revenue
(i.e., processor price per kg of cowpea meal).

Since the sample size (i.e., 4) seemed to be too small to justify the use of mean
value as a representative price of cowpea meal, the lowest and highest estimated prices
were used in the following sections of this chapter to analyze the price-competitiveness

of cowpea meal as an ingredient in kosei.

5. 2.2 Budgets [or Preparing K osei from Wet-Milled C owgeas

Representative budgets for preparing 1 kg of kosei by the wet-milling method
were constructed based on the collected data. As was the case with the budgets for
producing dry cowpea meal, the calculation required modifications, estimations, and

assumptions of various data values (see Appendix 5). The results are shown in Table 5.2.

Table 5.2 Representative budgets for preparing 1 kg of wet-milled kosei (¢)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3 i: 33 ... h ._ E E g “5 m 7“
o O 4 Us a 32 a: 9'
'0

LB 4,305 1,487 745 99 511 5,309 1,145 951 14,553 13,594 -959 .53
(32) (11) (5) (1) (4) (39) (8) (7) (107) (100) (-7)

MB 3,427 1,042 457 72 264 3,391 913 669 10,233 15,142 4,909 .42
(23) (7) (3) (0) (2) (22) (6) (4) (68) (100) (32)

PB 2,549 597 I68 45 16 1,472 682 387 5,914 17,089 11,175 .31
(15) (3) (1) (0) (0) (fl (4) (2) (35) (1001 (65)

 

*

Notes: LB: least profitable representative budget; MB: median representative budget; PB:

most profitable representative budget; the figures in parentheses are the ratio to the
value of revenue; for details of the calculation, see Appendix 5.

'0 The wage paid by Respondent 2 (and reported in Table 5.1) does not include the payment to the
respondent and family members, while it is not clear whether the wage paid by the other three respondents
included the payment to the respondent and family members or not.

124

 

The total costs to prepare 1 kg of kosei ranged from 515,914 to ¢14,553 (US$0.64
— US$1.58), while the revenue ranged from ¢l3,594 to ¢17,089 (US$1.48 — US$1.86),
making returns range from ¢-959 to 113-11,175 (USES-0.10 —- US$1.21). Note that the return
includes the return to all the labor (i.e., the labor of respondents, family members, and/or
hired assistants) involved in preparing 1 kg of kosei as well as the return to management
and capital. The negative return in LB is not an unrealistic result; if a vendor sells other
food products such as hausa koko along with kosei, and if those products are more
profitable, the overall return could be positive1 '.

The most important cost components were cowpea grain (with a share of between
15% and 32% in the revenue) and oil (with a share of between 9% and 39% in the
revenue).

The weight of dry cowpea meal needed to prepare 1 kg of kosei ranged from 0.31
to 0.53 kg. This value is important because it determines the change in the difference in
returns between wet-milled kosei (i.e., prepared from cowpea grain) and dry-milled kosei
(i.e., prepared using dry cowpea meal), when the price per kg of dry meal changes (e.g.,
if this value is 0.31, a ¢1 increase in the price of dry meal leads to a ¢0.31 increase in the

difference in returns)12 . Obviously, the smaller the difference in returns is, the more

 

” Vendors would continue selling kosei, even though having negative returns from kosei, only if the use of
kosei as a “loss leader” to attract customers to buy more profitable products leads to a higher profit than
when they sell only those profitable products (without kosei).

One can see this relationship by using equation (14): subtracting the unit return for dry-milled kosei from
the unit return for wet-milled kosei ends in the form:

t d
1] we _ 17 ry _ 1)QO _ (Pch + Pcustomd/chustom d/m)

Qk Qk Qk Qk Qk
where, custom d/m = custom dehulling/milling.
Therefore, the effect of change in the price of meal on the lefi-hand side of the equation is derived as:

wet 0'
At” —” W)=-QflAPm-
Qk Qk Qk

 

 

 

125

 

attractive to kosei vendors the meal would be.

5.2.3 1_3u_dg_e_§f0r Prepgrigg Kosei Using Drv Cowper; Meg!

Finally, the weight of dry cowpea meal needed to prepare 1 kg of kosei in the
three representative budgets was multiplied by the lowest and highest estimated prices
per kg of dry cowpea meal, generating 3 X 2 = 6 case scenarios. As discussed in Section
4.2.4, it was assumed that the kosei vendors purchased cowpea meal directly from the
processors. It was not clear whether the processors would collect the VAT/NHIS for such
a transaction”. Therefore, both cases—processors do collect (and kosei vendors pay the
entire 15% of tax; see discussion in Section 4.2.4) and do not collect the VAT/NHIS—
were considered, making a total of 12 different case scenarios. As explained earlier in this
chapter, the calculated costs of dry cowpea meal were entered in the budgets for kosei
preparation, replacing the cost of cowpea grain, and the costs of custom dehulling and
milling were set to zero. The results are shown in Table 5.3

The results indicate that, for all of the four potential prices of dry COWpea meal,
the meal would lead to a negative return if used by a vendor operating with LB (with
returns ranging from ¢-13,830 to ¢-7,254 per kg of kosei) or MB (from ¢-5,293 to ¢-106),
while a positive return if used by a vendor operating with PB (from ¢3,642 to ¢7,440).

The difference in returns between wet- and dry-milled kosei ranged from ¢6,295

t0 ¢12,871 per kg of kosei for LB vendors, ¢5,015 to ¢10,202 for MB vendors,

 

 

'3 The guide book obtained from a VAT officer during the fieldwork indicates that the VAT/NHIS has to be
collected from everybody (except the Ghanaian President and some organizations such as foreign
embassies) unless the commodity is VAT/NHIS exempted. Therefore, cowpea meal processors would be
supposed to collect the VAT/NHIS whether they sell products to retailers or to kosei vendors. However,
considering the finding that not all the processor respondents were collecting the VAT/N HIS from retailers
regardless of the regulation, it would be possible that they do not collect VAT/N HIS from kosei vendors.

126

Table 5.3 Budgets for preparing 1 kg of kosei using dry COWpea meal

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Price per Cowpea Other Total Sales Returns B1fference
kg of meal Budget meal costs (¢) cost (¢) (¢) (¢) in returns"

(¢) (¢) 11¢)
Processors 25,033 LB 13,283 9,297 22,580 13,594 -8,986 8,027
collect [21,768 x MB 10,477 6,138 16,615 15,142 -l,473 6,382
VAT/NHIS 1.15] PB 7,671 2,979 10,650 17,089 6,439 4,736
from kosei 34,161 LB 18,127 9,297 27,423 13,594 —13,830 12,871
vendors [29,705 X MB 14,298 6.138 20,435 15,142 -5,293 10,202
1 . 15] PB 10,469 2,979 13,447 17,089 3,642 7,533
Processors LB 11,551 9,297 20,847 13,594 -7,254 6,295
do NOT 21,768 MB 9,111 6,138 15,248 15,142 —106 5,015
collect PB 6,671 2,979 9,650 17,089 7,440 3,735
VAT/NHIS LB 15,762 9,297 25,059 13,594 -11,465 10,506
from kosei 29,705 MB 12,433 6,138 18,570 15,142 -3,428 8,337
vendors PB 9,103 2,979 12,082 17,089 5,007 6,168

 

* Difference in returns between wet- and dry-milled kosei.
Source: Tables 5.1 and 5.2.

and ¢3,735 to ¢7,533 for PB vendors.

 

Whether kosei vendors would adopt dry cowpea meal would depend on whether
the value of saved time and labor exceeds the increasing cost of preparation by switching
from the wet-milling method to the use of dry cowpea meal (i.e., the difference in returns
between wet- and dry-milled kosei). However, for the reasons discussed in Section
4.2.1.5, it would be difficult to predict from the available data the amount of time and
labor that could be saved by using dry cowpea meal. As an alternative, the use of the
daily minimum wage as a proxy for the opportunity cost of the vendor’s time and labor
can provide some insights. According to the Bank of Ghana (mvwboggovgh), the
Ghanaian daily minimum wage (during February and March 2007) was ¢19,000. In
addition, ¢10,000l4 and ¢3 8,000 were selected as alternative minimum wages to examine
the effects of change in the opportunity cost on the profitability of dry cowpea meal.

Since the difference in returns widens as the volume of production of kosei increases, the

14A kosei vendor respondent actually paid ¢10,000 per day to each of her two hired young assistants.
(Note: this payment was removed fi'om the budget of this respondent so that the returns in the budgets of all
the 13 respondents included the return to all the labor needed to prepare kosei.)

127

 

volumes of 5, 10, and 15 kg were selected as examples”. The differences in returns per
kg of kosei with the lowest and highest estimated prices of dry meal (see Table 5.3) were
multiplied by 5, 10, and 15 for each of the three budgets to derive the total difference in
returns. Then, the ratios of these figures to the three different opportunity costs of labor

were calculated. The results are shown in Table 5.4.

Table 5.4 Ratio of the total difference in returns between wet- and dry-
milled kosei to the opportunity cost of labor per day

 

 

 

 

 

 

 

 

 

 

_ Volume of Total Ratio of total difference in returns to

Budget 3:19:12); production difference opportunity cost of labor per day (%)
per kg (¢) of kosei 1n returns OPP- 903‘ 0f labor per day (¢)

(kg) (¢) 10,000 19,000 38,000

5 31,473 315 166 83

21,768 10 62,947 629 331 166

LB 15 94,420 944 497 248

5 64,354 644 339 169

34,161 10 128,707 1,287 677 339

15 193,061 1,931 1,016 508

5 25,075 251 132 66

21,768 10 50,150 502 264 132

MB 15 75,226 752 396 198

5 51,010 510 268 134

34,161 10 102,020 1,020 537 268

15 153,029 1,530 805 403

5 18,677 187 98 49

21,768 10 37,354 374 197 98

PB 15 56,031 560 295 147

5 37,666 377 198 99

34,161 10 75,332 753 396 198

15 112,998 1,130 595 297

 

 

 

 

 

 

 

Source: Table 5.3; (opportunity cost of labor) Bank of Ghana.
Table 5.4 shows that, with the lowest estimated price of dry COWpea meal, for LB
vendors preparing 5 kg of kosei per day to adopt dry meal, they would have to be able to

save an amount of labor that is equivalent of 166% of the current daily minimum wage

¥

‘5 Among 13 respondents, there were 4 whose estimated volume of kosei production of the day before the
interview was conducted fen between 2.5 and 7.4 kg, another 4 whose estimated volume fell between 7.5
and 12.4 kg, and 2 whose estimated volume fell between 12.5 and 17.4 kg (the estimated volume of the
other 3 respondents were outside of this range).

128

(i.e., 1 and 2/3 person days), by switching from the wet-milling method to the use of dry
meal. MB vendors would have to be able to save an amount of labor that is equivalent of
132% of the daily minimum wage, and PB vendors would have to be able to save an
amount of labor that is equivalent of 98% of the daily minimum wage. With regard to the
change in the volume of production, the amount of labor that vendors would have to be
able to save doubles for 10kg and triples for 15kg, because of the linear relations between
these variables. With regard to the change in the opportunity cost of labor per day, if the
opportunity cost doubles, the amount of labor that vendors would have to be bale to save
falls to one-half of the original amount, because of the inversely proportional relations
between these variables.

It would be unlikely that many kosei vendors could save the amount of labor
derived above by switching from cowpea grain to dry meal. Therefore, under the
conditions observed in Accra during February and March 2007, dry cowpea meal would

not be a price-competitive ingredient in kosei for the majority of vendors.

5.3 Sensitivity Analysis

Before conducting sensitivity analyses for different scenarios, breakeven prices of
dry cowpea meal were calculated for each of the three representative budgets for kosei
preparation. The relationship between the purchase price of dry cowpea meal and the
return for kosei vendors is plotted in Figure 5.2.

The breakeven prices of dry cowpea meal (i.e., price at which returns for kosei
vendors equal zero) were ¢8,098 per kg for LB vendors, ¢21,5 14 per kg for MB vendors,

and ¢46,045 per kg for PB vendors. The processor price of cowpea flour produced by the

129

 

Figure 5.2 Relationship between the purchase price of dry cowpea meal
and the return for kosei vendors

 

 

15,000 7'”
10,000 1
5,000 1
0
-5,000 -
-10,000 7
-15,000 4
~20,000

25,000 4— - ——————~————— --———- ————-—
0 10,000 20,000 30,000 40,000 50,000

 

 

return per kg ofkosei (cedis)

 

price of meal (cedis / kg)

 

 

_—LB

 

 

 

Manurfl

 

 

Source: Calculated by the author.

two industrial processor respondents interviewed averaged ¢25,870 per kg (VAT/NHIS
exclusive; with a standard deviation of ¢5,841). Thus, only the PB vendors could expect
to break even using dry meal priced at this level.

Sensitivity analyses were then conducted for the following six scenarios that
would potentially change the price-competitiveness of dry cowpea meal: (1) change in
the technical efficiency of processing dry cowpea meal; (2) change in the volume of
production; (3) bulk purchase of dry COWpea meal by kosei vendors; (4) change in the
price of cowpeas (seasonality analysis); (5) change in the retail margin; and (6) a
combination of different scenarios.

In these sensitivity analyses, except for a special assumption made for scenarios
(2) and (6), the respondents were assumed to modify the price at which they would sell

COWpea meal so the returns remained constant, whenever the cost of production changed.

130

While these assumptions were made to simplify the analysis, assuming constant returns is
the equivalent of assuming a perfectly competitive market for the products analyzed. The

appropriateness of this assumption is discussed in Section 8.3.

5.3.1 Change in the Technical Efficiency of Processing Dix Cowgea Meal

As reported in Section 4.2.6.2, the input-output ratio (or [1 — waste rate]) of dry

 

COWpea meal obtained by the experiment was 0.8474, while the input-output ratio of
cowpea/soybean flour obtained from the respondents ranged from 0.7488 to 0.8469. A
higher ratio means a higher technical efficiency because less waste is generated from the
same amount of raw material. Therefore, a higher technical efficiency leads to a lower
raw material cost. The first sensitivity analysis was conducted on the technical efficiency
of processing: considering the ratio obtained from the laboratory experiment as the
maximum value, the input-output ratio of each respondent was set to 0.8474. The results

are shown in Tables 5.5, 5.6, and 5.7.

131

 

Table 5.5 Sensitivity analysis (1): Technical efficiency—budgets for
cowpea meal processors (¢ per kg of meal)

 

 

 

 

 

 

 

 

 

 

Resp 1 Resp2 Resp3 Resp4

Processor price (VAT/NHIS 29,655 21,341 26,184 24,634
exclusive) (-50) (-427) (-850) (-6)
Raw material 6,459 6,459 6,459 9,633
(-50) (427) (—850) (-6)

Other costs 18,528 6,863 14,602 N.Av.
Total cost 24,987 13,322 21,061 N.Av.
(-50) (-427) (-850) (-6)

Return 4,668 8,018 5,123 N.Av.

 

 

Note: the numbers in parentheses are the difference from the original values (in Table
5.1); N.Av.: not available.
Source: Calculated by the author.

In this scenario, the cost of raw material (i.e., cowpea grain) for cowpea meal
processors declined by {56 to ¢850 per kg of meal, depending on the initial technical
efficiency and the price of cowpea grainl6 (Table 5.5). Assuming no associated change in
the costs of equipment, electricity, fuel, and water occurred”, both the total cost and the
processor price declined by the same value for each processor respondent. As a result, the
estimated price of the least expensive dry cowpea meal changed from ¢21,768 to ¢21,341,
and the estimated price of the most expensive meal changed from ¢29,705 to ¢29,655 per

kg.

‘6 As explained earlier, the same price of cowpea grain was used for Respondent 1 through 3, which
resulted in the same cost of cowpea grain (¢6,459 per kg of meal) across these three respondents.
'7 Strictly speaking, these variables must change because the input amount changed.

132

Table 5.6 Sensitivity analysis (1): Technical efficiency—budgets for kosei
vendors (per kg of kosei)

 

 

 

 

 

 

Price per Cowpea T Difference
otal Returns .
kg of meal Budget meal cost (¢) (¢) 1n returns“
(¢) (¢) (¢)

Processors 24,542 LB 13,022 22,319 -8,726 7,767 (-261)
collect [21,341 x MB 10,272 16,409 -1,267 6,176 (-206)
VAT/NHIS 1.15] PB 7,521 10,500 6,590 4,585 (-151)
from kosei 34,103 LB 18,096 27,393 -13,799 12,840 (-31)
vendors [29,655 x MB 14,273 20,411 -5,269 10,178 (-24)
1.15] PB 10,451 13,430 3,660 7,515 (-18)
Processors LB 11,324 20,621 -7,027 6,068 (-227)
do NOT 21,341 MB 8,932 15,070 73 4,836 (-179)
collect PB 6,540 9,519 7,571 3,604 (-131)
VAT/NHIS LB 15,735 25,032 -11,439 10,480 (-27)
from kosei 29,655 MB 12,412 18,549 -3,407 8,316 (-21)
vendors PB 9,088 12,067 5,023 6,152 (-15)

 

 

 

 

 

 

 

* Difference in returns between wet- and dry-milled kosei.

Note: the numbers in parentheses are the difference from the original values (in Table 5.3).
Source: Calculated by the author.

 

The new lowest estimated price of cowpea meal (i.e., ¢21,341) was slightly lower
than the breakeven price for MB vendors (i.e., ¢21,514). Therefore, at this price, the
returns to MB vendors turned to a small positive number (i.e., ¢73) (Table 5.6).

The estimated decline in the price of cowpea meal only increased the return for
kosei vendors by a minimal amount, from ¢15 (when the meal costs ¢29,655 per kg and
is used by a PB vendor) to ¢261 (when the meal costs ¢24,542 per kg and is used by a LB

vendor) per kg of kosei.

133

 

Table 5.7 Sensitivity analysis (1): Technical efficiency—Ratio of the total
difference in returns between wet- and dry-milled kosei to the
opportunity cost of labor per day

 

 

 

 

 

 

 

P , f Volume of Total Ratio of total difference in retumsoto
Bud et d ”$80“ production difference Opflmflt)’ C051 oflabor per dill/A 0
g p erry kg (¢) of kosei in returns Opp. cost of labor per day (¢)
(kg) (¢) 10,000 19,000 38,000
LB 60,679 607 319 160
(2,267) (23) (42) 1-6)
48 362 484 255 127
21,34 1 ’

MB 1 0 ('l .789) (—18) (-9) (-5)
[)3 36,044 360 190 95
(4,310) (-13) (7) (-3)

 

 

 

 

 

 

 

 

 

Note: the numbers in parentheses are the difference from the original values, calculated
for the lowest meal price of ¢2 1 ,768 per kg (in Table 5.4).
Source: Tables 5.4 and 5.6; (opportunity cost of labor) Bank of Ghana.

These increases in returns were not enough to make any major change in the
results derived earlier about the profitability of using meal: with the modified lowest
price of meal and for 10 kg of kosei, the ratio of the total difference in returns to the
opportunity cost of labor per day (¢19,000) decreased only by 7 percentage points (when
the meal is used by a PB vendor) to 12 percentage points (when the meal is used by a LB

vendor) (Table 5.7).

5.3.2 Change in the Volume at Production

When processors have their own equipment, the purchase cost is a fixed cost.
Therefore, the average equipment cost becomes lower as the volume of production
increases up to the level of full capacity utilization. In the same way, if processors pay
fixed monthly wages to their employees, the average wage becomes lower as the volume
of production increases. However, if the facility is operating at its full capacity, it cannot
increase production without purchasing new equipment or hiring additional workers.

Therefore, whether the processor respondents operated at their full or under capacity was

134

first examined to identify if it was viable to conduct a sensitivity analysis with regard to
the change in the volume of production.

The weekly operation time of the facility obtained from Respondents 1, 2, and 3
were between 40 and 48 hours. None of them operated in shifts. Although the information
is missing on the standard weekly working time in Ghana, with the current working time
of 40 hours or longer per employee, it would likely be difficult for the respondents to
increase the volume of production by extending their operation time.

The next question was whether their equipment and employees were constantly
working throughout the facility’s operation time. If not, there would be a potential for an
increase in production by making them “work harder,” while causing no or minimal
increase in the cost of equipment and labor. Ideally, the full capacity of a facility has to be
estimated from the hourly capacity and labor requirement of each piece of equipment for
producing each product, and the estimated production under the full capacity has to be
compared with the actual production of the facility. However, such detailed data were not
available. For this study, a simple method was used to obtain a rough idea about the
degree of capacity utilization of the respondents: the total weight of products produced
during February was divided by the number of total working hours (sum of hours worked
by each employee) during the same period. The result turned out that although all three
respondents used similar processing technologies, Respondent 2 produced at least 4.3
times more per worker per hour than Respondent 1 and 4.8 times more than Respondent 3.
Differences of more than four times seemed to be large enough to judge that Respondent
2 produced more intensely, or in other words, more closely to the facility’s full capacity

than Respondents 1 and 3, even taking into consideration the difference in the types of

135

products produced by each respondent and the difference in the degrees of equipment and
labor involvement for different products.

Based on the finding, the following assumptions were made: (1) Respondents 1
and 3 can produce 4.3 and 4.8 times more products”, respectively, without an increase in
the fixed costs”; (2) Respondent 2 can produce 10% more products without increase in
the fixed costs; and (3) there is no change in the average variable costs (i.e., assumption
of constant returns to scale). The 10% in the second assumption was arbitrarily selected.
Respondent 2 might have been already operating at the full capacity. Therefore, this
assumption was just experimental.

In addition, when processors increase their volume of production, they could
possibly reduce their per-unit margin to make their products more competitive with other
companies’ products. By doing so, they could still increase their total returns since their
total sales volume would increase, if the demand is sufficiently price-elastic. Therefore,
for the budgets of Respondents 1 and 3, two alternative assumptions were made such that

when the unit cost of production decreases due to the increase in the volume of

 

'8 4.3 or 4.8 more of each product (not only cowpea meal). Thus, the share of cowpea meal in the total
production (be weight) remained the same. An alternative sensitivity analysis could be conducted by
increasing the share of cowpea meal while keeping the total quantity of production constant. However, as
described in Section 4.3, for many of the cost variables, an accurate share of each product was impossible
to obtain. Therefore, the weight share was used as an approximation. When modeled in this way, the cost
per kg of cowpea meal of these variables are actually not affected by the weight share of cowpea meal;
what matters is the total weight1 of all products (for example, wage per kg of COWpea meal is expressed as

 

g WWI: ———ZW th [see equations (12) and (13), note that Qm j, cancel out]).
Q]! 7:5th W élel
Therefore, the change in the share of cowpea meal in the total production (by weight), while keeping the
level of total production, would not make much difference to the original budget. For this reason, this
sensitivity analysis was not conducted.
9The following cost components were considered to be fixed costs: wage, purchase prices of equipment,
rent, telecommunication, and miscellaneous; while the following cost components were considered to be
variable costs: raw materials, maintenance costs of equipment, electricity, fuel, water, transportation,
printing and stationery, and packaging material. In reality, Respondent 1 did not pay fixed monthly wages
but paid the wages depending on the volume of work of each month. Therefore, the assumption of fixed
wages for this respondent was experimental.

136

production, the respondents modify the selling price of cowpea meal: (1) so the per-unit
returns remain constant (the same assumption made for the other scenarios); and (2) so
the per-unit returns become one-half of the original value. The result of the sensitivity
analysis is shown in Tables 5.8, 5.9, and 5.10.

As expected, the unit cost paid by Respondents l and 3 declined markedly due to
the decreasing average fixed costs (Table 5.8). The 10% increase in production did not
have a significant effect on the total cost paid by Respondent 2 (only ¢99 less per kg of
meal). Under the assumption that the processors set the selling price so the per-unit
returns remain constant, only the lowest estimated price of cowpea meal (i.e., ¢21,011,
for Respondent 3) was slightly lower than the breakeven price for MB vendors

(i.e., ¢21,514). Under the assumption that the processors set the selling price so the per-

Table 5.8 Sensitivity analysis (2): Increase in the volume of production—
budgets for cowpea meal processors (¢ per kg of meal)

 

 

 

 

 

 

 

 

 

 

 

Resp 1 Resp2 Resp3
Processor price With original return 23,840 21,668 21,011
(VAT/NHIS . . . ('5’866) ('99) (’6’024)
exclusive) With 1/2 of origlnal 21,506 18,449
return (-8,199) (-8,5 86)
Wage 913 391 1,259
(-3,013) (~39) (-4,785)
Equipment 753 174 550
(~323) (-15) (-l92)
Rent 0 0 41
(0) (0) (-155)
Telecommunication 351 218 102
(-1,157) (-22) (-389)
Miscellaneous 416 236 132
(-1,373) (~24) (-503)
Other costs 16,740 12,631 13,802
Total cost 19,172 13,650 15,887
-5,866) (-99) (-6,024)
Return OrLginal return 4,668 8,018 5,123
1/2 of original return 2,334 2,562

 

Note: the numbers in parentheses are the difference from the original values (in Table 5.1).

Source: Calculated by the author.

137

 

 

unit returns become one-half of the original value, the estimated prices of cowpea meal

were ¢21,506 (for Respondent 1) and ¢18,449 (for Respondent 3), both of which were

lower than the breakeven price for MB vendors.

Under the assumption that the processors set the selling price so the per-unit

returns remain constant, the decline in the lowest estimated price of meal increased the

returns to MB vendors by only ¢317 per kg of kosei (if the VAT/NHIS was not collected),

which is less than the value of 1 ball of kosei (i.e., ¢500) (Table 5.9). Under the

assumption that the processors set the selling price so the per-unit returns become one-

half of the original values, the decline in the lowest estimated price of meal increased the

returns to MB vendors by ¢1,389 per kg of kosei (if the VAT/N HIS was not collected).

Table 5.9 Sensitivity analysis (2): Increase in the volume of production——
budgets for kosei vendors (per kg of kosei)

 

Cowpea

 

 

 

 

 

 

 

 

 

 

 

 

 

kzrgcfe "‘13:; Budget meal Total Returns Difference in
cost returns"

(¢) (¢) (¢) (¢) (¢)
With 24,162 LB 12,821 22,118 -8,524 7,565 (-462)
. 21,011 X MB 10,113 16,251 -1,108 6017 (-365)

Wt VAT/NHIS [ ’

oggllnal 1.15] PB 7,404 10,383 6,706 4,469 (~267)
return 5 Without LB 11,149 20,445 -6,852 5,893 (-402)
VAT/NHIS 21,011 MB 8,794 14,931 211 4,698 (-317)
PB 6,439 9,418 7,672 3,503 (-232)
With 21,216 LB 11,258 20,555 -6,961 6,002 (-2,025)
With VAT/NHIS [18,449 x MB 8,880 15,018 125 4,784 (-1,597)
1/2 of 1.15] PB 6,502 9,481 7,609 3,566 (-1,170)
original Witho t LB 9,789 19,086 -5,493 4,534 (-1,761)
returns V AT /1\llHIS 18,449 MB 7,722 13,859 1,283 3,626 (4,389)
PB 5,654 8,633 8,457 2,718 (-1,017)

 

* Difference in returns between wet- and dry-milled kosei.

Note: the numbers in parentheses are the difference from the original values, calculated
for the lowest meal price of ¢2 1 ,768 per kg (without VAT/NHIS) and 925,033 per kg

(with VAT/NHIS) (in Table 5.3).

Source: Calculated by the author.

138

 

 

Under the assumption that the processors set the selling price so the per-unit

returns remain constant, the increases in returns were not enough to make any major

change in the profitability of using meal: with the modified lowest price of meal and for

10 kg of kosei, the ratio of the difference in returns to the opportunity cost of labor per

day (¢19,000) declined only by 12 percentage points (when the meal is used by PB

vendors) to 21 percentage points (when used by LB vendors) (Table 5.10).

Under the assumption that the processors set the selling price so the per-unit

returns become one-half of the original value, MB vendors who prepare 10 kg of kosei

per day might adopt dry meal if they could save at least an amount of labor that is

equivalent of 191% of the daily minimum wage, using the lowest estimated price of meal.

Table 5.10 Sensitivity analysis (2): Increase in the volume of production
—Ratio of the total difference in returns between wet- and dry-

milled kosei to the opportunity cost of labor per day

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

P . f Volume of Total Ratio of total difference in returns to
B d t d mneigal production difference (mportumty cost of labor per day (%)
u ge p erry kg (¢) of kosei in returns Opp. cost of labor per day (¢)
(kg) (¢) 10,000 19,000 1 38,000
LB 58,928 589 310 155
With (4401‘?) (-40) (-21) (-11)

. . 46 981 470 247 124
on ma] MB 21,01 1 10 ’ J
,6ij (3,170) (-32) (47) (8)

PB 35,033 350 184 92

(-2,321) (-23) (- l 2) (6)

LB 45,336 453 239 119

Withfl/Z (-17,611) (-176) (-93) (-46)

0 36 260 363 191 95

. . IO ’
orlgmal MB 18'4” (-13,891) (439) (-73) (-37)
returns PB 27,183 272 143 72

(-l0,l7l) (402) Q54) l (~27)

 

 

 

 

 

 

 

 

Note: the numbers in parentheses are the difference from the original values, calculated

for the lowest meal price of ¢21,768 per kg (without VAT/NHIS) (in Table 5.4).
Source: Tables 5.4 and 5 .9; (Opportunity cost of labor) Bank of Ghana.

139

 

These results indicate that: (1) if the processors set the selling price of their meal
so the per-unit returns remain constant, when the cost of production changes, the increase
in the volume of production alone would not be enough to improve significantly the
price-competitiveness of dry cowpea meal as an ingredient in kosei; and (2) if the
processors set the selling price of their meal so their per-unit returns become one-half of
the original value, when they increase the volume of production, the price-
competitiveness of dry meal would improve greatly, compared to under the per-unit-
retums-remain-constant assumption. However, it is not clear whether the dry meal would
become attractive enough for MB vendors to adopt it.

Finally, it should be noted that whether the processors can sell all the increased
outputs, whether holding the per-unit margin (i.e., per-unit returns) constant or cutting it
to one-half of the original value (while increasing the total returns), depends on how

price-elastic the demand for the product is.

5.3.3 Bulk Purchase oLDrv Cownegfligbv Kosei VM

It might be possible for kosei vendors to negotiate with dry COWpea meal
processors and sign a contract for regular bulk purchases. For the processors, the benefits
of such a contract would be a secure constant demand for the meal as well as the cost
savings for printing and stationery and packaging material, as they would not need to
individually package small quantities of meal in a container with a printed label. Because
of these benefits, the processors might be willing to offer a discount. This scenario was
examined in the following sensitivity analysis, assuming that (1) the cost of printing and

stationery becomes zero for the contracted bulk purchase; and (2) the cost of packaging

140

material becomes ¢73 per kg of meal”. The results are shown in Tables 5.11, 5.12, and
5.13.

The results indicated that the bulk purchase of COWpea meal would decrease the
price of meal by ¢1,866 to 955,338 per kg (Table 5.11). As a result, the lowest price among
the four processor respondents was estimated to be ¢17,741 per kg of meal (for
Respondent 2), which is lower than the breakeven price for MB vendors (i.e., ¢21,514).
However, the processor prices estimated for the other three respondents stayed above the

breakeven price.

Table 5.11 Sensitivity analysis (3): Bulk purchase of meal by kosei
vendors—budgets for cowpea meal processors (¢ per kg of meal)

 

 

 

 

 

 

 

 

 

 

Resp 1 Resp2 Resp3 Resp4
Processor price (VAT/NHIS 24,367 17,741 25,168 22,712
exclusive) (-5,338) (-4,027) (-1,866) (-1,927)
Printing and stationery 0 0 0 0
(-4,861) (-3,000) (-1,206) (—800)
Packaging material 73 73 73 73
(-477) (-1,027) (660) (-1,127)
Other costs 19,626 9,650 19,972 N.Av.
Total cost 19,699 9,723 20,045 N.Av.
(-5,33 8) (-4,027) (-1,866) (-1,927)
Return 4,668 8,018 5,123 N.Av.

 

Note: the numbers in parentheses are the difference from the original values (in Table

5.1); N.Av.: not available.
Source: Calculated by the author.

 

20 During the field survey, a gari processor was found selling gari in bulk. The cost of packaging material
(two different types of inner and outer polyethylene bags) was ¢9,000 for 62 olonka (estimated to be 123
kg). Therefore, the unit cost was 9,000/123 = ¢73 per kg. It was assumed that the material suitable for gari
can be used for dry cowpea meal as well.

141

 

With the new lowest price, the return for MB vendors would be ¢l,579 per kg of
kosei, if processors did not collect the VAT/NHIS from the vendors, and 56466 per kg of
kosei, if processors collected the VAT/NHIS (Table 5.12). If processors did not collect the
VAT/NHIS, the decline in the price of meal increased the returns to kosei vendors

by ¢l,234 (when the meal was used by PB vendors) to ¢2,137 (when used by LB

vendors).

Table 5.12 Sensitivity analysis (3): Bulk purchase of meal by kosei

vendors—budgets for kosei vendors (per kg of kosei)

 

 

 

 

 

 

 

 

 

 

Price per Cowpea Total Returns Difference
kg of meal Budget meal cost (¢) (¢) 1n returns“
(¢) (¢) (¢)

Processors collect 20,402 LB 10,826 20,123 -6,529 5,570 (-2,457)
VAT/NHIS from [17,741 X MB 8,539 14,677 466 4,443 (-1,938)
kosei vendors 1.15] PB 6,252 9,231 7,858 3,317 (-1,4l9)
Processors do NOT LB 9,414 18,711 -5,117 4,158 (-2,137)
collect VAT/NHIS 17,741 MB 7,425 13,563 1,579 3,330 (-1,685)
from kosei vendors PB 5,437 8,416 8,674 2,501 (-1,234)

 

* Difference in returns between wet- and dry-milled kosei.

Note: the numbers in parentheses are the difference from the original values, calculated
for the lowest meal price of ¢21,768 per kg (without VAT/NHIS) and ¢25,033 per kg
(with VAT/NHIS) (in Table 5.3).

Source: Calculated by the author.

142

 

 

If preparing 10 kg of kosei, using the lowest priced meal (without the VAT/N HIS
being charged), MB vendors would have to be able to save at least 1.75 days of one
person’s labor for them to adopt the meal (or 0.88 days for 5 kg of kosei), while PB

vendors would have to be able to save at least 1.32 days (or 0.66 days for 5 kg of kosei)

(Table 5.13).

Table 5.13 Sensitivity analysis (3): Bulk purchase of meal by kosei
vendors—Ratio of the total difference in returns between wet-
and dry-milled kosei to the opportunity cost of labor per day

 

 

 

 

 

 

 

 

_ f Volume of Total Ratio of total difference in returns to

B d :nce 0 1 production difference opportumty cost of labor per day (%)
u get eff/k2?) of kosei in returns Opp. cost of labor per day (¢)

p (kg) (¢) 10,000 19,000 38,000

LB 41,579 416 219 109

(-21,368) (-214) (-112) Q56)

33,296 333 175 88

MB '7’741 10 (-16,854) (-169) (-89) (-44)

PB 25,013 250 132 66

(-12,341) (-123) (-65) (-32)

 

 

 

 

 

 

 

 

Note: the numbers in parentheses are the difference from the original values, calculated
for the lowest meal price of ¢2 1 ,768 per kg (without VAT/NHIS) (in Table 5.4).
Source: Tables 5.4 and 5.12; (opportunity cost of labor) Bank of Ghana.

5.3.4 Change in the Price of Cowpeas

The potential benefit of dry cowpea meal for kosei vendors, in terms of input
price stabilization, was examined by the following sensitivity analysis. As reported in
Section 4.2.5.2 (Table 4.14), the estimated lowest price of cowpea grain in 2007

was ¢5,155 per kg, while the highest price was estimated to be 1.10 times the price in

143

March2 1. Using these estimates, it was assumed that cowpea meal processors purchased
cowpea grain and produced meal when the price of cowpea grain was the lowest, stored
the meal, and sold it when the price of grain was the highest, without changing the price
of meal. Based on this assumption, the budget for kosei vendors to prepare 1 kg of kosei
by the wet-milling method was also modified. However, it was assumed that the prices of
all cost components to prepare kosei stayed constant, except cowpea grain. The results
are shown in Tables 5.14, 5.15, and 5.16.

The lower price of cowpea grain drove down the cost of cowpea meal production,
and therefore the price of meal, by ¢38O to ¢426 per kg of meal (Table 5.14). The lowest

estimated price of meal was ¢21,366 per kg for Respondent 2.

Table 5.14 Sensitivity analysis (4): Change in the price of cowpea grain—
budgets for cowpea meal processors (¢ per kg of meal)

 

 

 

 

 

Resp 1 Resp2 Resp3

Processor price (VAT/NHIS 29,326 21,366 26,608
exclusive) (-380) (-402) (-426)
Raw material 6,130 6,485 6,884
(-380) (-402) (-426)

Other costs 18,528 6,863 14,602
Total cost 24,658 13,348 21,485
(-380) (-402) 4426)

Return 4,668 8,018 5,123

 

 

 

 

 

 

Note: the numbers in parentheses are the difference from the original values (in Table
5.1); Respondent 4 was excluded from this analysis because the information was
missing on which month of the year 2006 the data were collected.

Source: Calculated by the author.

 

2' Since the majority of kosei vendor respondents seemed to report the cost of cowpeas that they purchased
during March, the highest cost of cowpea grain per kg of kosei during 2007 was estimated to be the cost of
COWpeas in the original budgets (Table 5.2) times 1.10, which was derived by dividing the maximum price
index of COWpea grain (i.e., 111) by the price index for March (i.e., 101) (see Section 4.2.5.2).

144

 

The increase in the price of cowpea grain drove down the returns to kosei
prepared by the wet-milling method by only ¢265 to (£447 per kg of kosei (Table 5.15). In
contrast, the decrease in the estimated price of the least expensive COWpea meal slightly
drove up the returns to the kosei prepared from dry meal. As a result, the difference in
returns between wet- and dry—milled kosei declined by only (13388 (when the meal was
used by PB vendors) to ¢660 (when used by LB vendors) per kg of kosei, if the processor

did not collect the VAT/NHIS.

Table 5.15 Sensitivity analysis (4): Change in the price of cowpea grain—
budgets for kosei vendors Qaer kg of kosei)

 

 

 

 

 

 

 

 

 

 

 

 

. Cowpea .
Pr1ce per meal/ Total Returns Difference
kg of meal Budget . 1n returns“
gram cost (¢) (¢)
(¢) (¢) (¢)
Processors collect 24,571 LB 13,038 22,335 -8,741 7,335 (-692)
VAT/NHIS from [21,366 x MB 10,284 16,422 -1,279 5,832 (-549)
kosei vendors 1.15] PB 7,530 10,509 6,581 4,330 (-406)
Processors do NOT LB 11,337 20,634 -7,041 5,634 (-660)
collect VAT/N HIS 21,366 MB 8,943 15,080 62 4,491 (-524)
from kosei vendors PB 6,548 9,527 7,563 3,348 (-388)
_ LB 4,752 15,000 -1,406
Budget for kosei vendors by the (447) (447) (_447)
wet-millmg method wrth 1.10 . MB 3,783 10,589 4,553
times higher pr1ce of cowpea gram 6
than the price in the original (356) (3 5 ) ('356)
budget” PB 2,814 6,179 10,910
4265) (265 (-265)

 

"‘ Difference in returns between wet- and dry-milled kosei; the numbers in parentheses
are the difference from the original values, calculated for the lowest meal price
of ¢21,768 per kg (without VAT/NHIS) and ¢25,033 per kg (with VAT/NHIS) (in

Table 5.3)
** The numbers in parentheses are the difference from the original values (in Tables 5.2).

Source: Calculated by the author.

145

 

Since the changes in the price of cowpea grain resulted in only minor changes in
the difference in returns between wet- and dry-milled kosei, the ratio of the difference in

returns to the opportunity cost of labor per day did not improve much (Table 5.16).

Table 5.16 Sensitivity analysis (4): Change in the price of cowpea grain—
Ratio of the total difference in returns between wet~ and dry~
milled kosei to the opportunity cost of labor per day

 

 

 

 

 

 

 

 

P , f Volume of Total Ratio of total difference in returns to
Bud at d megs“ production difference opportunlty cost of labor per day (%)

g pet's/kg (¢) of kosei in returns Opp. cost of labor per day (¢)
(kg) (¢) 10,000 19,000 38,000
LB 56,344 563 297 148
(6.603) (-66) {-35) Q7)
44 910 449 236 118

21 ,366 10 ’

MB (5.241) (-52) (~28) (.14)
PB 33,475 335 176 88
(~3,878) (~39) (~20) (~10)

 

 

 

 

 

 

 

 

Note: the numbers in parentheses are the difference from the original values, calculated
for the lowest meal price of ¢2 1 ,768 per kg (without VAT/NHIS) (in Table 5.4).
Source: Tables 5.4 and 5.15; (opportunity cost of labor) Bank of Ghana.

To conclude, the analysis showed that a reasonable range of seasonal fluctuation
in the price of cowpea grain alone would not make dry cowpea meal much more

attractive to kosei vendors.

5.3.5 Change in the Retail Margin

It has been assumed so far that the kosei vendors buy cowpea meal directly from
meal processors. If vendors have to buy meal from retailers, the price will be higher
because of the retail margin. Also, for most housewives, retailers would be the major
outlet if dry COWpea meal becomes available. The retail prices of cowpea meal were
estimated for the different retail margins combined with the VAT/NHIS, derived in

Section 4.2.4. The result is presented in Table 5.17.

146

The estimated price of dry COWpea meal ranged from ¢26,121 to ¢47,528 per kg.
This range almost falls between the breakeven price for MB vendors (¢2l,514) and the

breakeven prices for LB vendors ((646,045).

Table 5.17 Sensitivity analysis (5): Change in the retail margin—estimated
retail price of dry cowpea meal

 

 

Retail margin + Lowest price Highest price
VAT/NHIS (%) (¢) (¢)
20 26,121 35,646
30 28,298 38,617
40 30,475 41,587
50 32,652 44,558
60 34,829 47,528

 

 

 

 

 

Source: Calculated by the author.

5.3.6 Combinwn of Dflerent ScenaLios

Finally, the most favorable scenario for a higher price-competitiveness of dry
cowpea meal was analyzed. It was assumed that the scenarios 1 through 4 analyzed above
happened at the same time (i.e., technical efficiency improved, volume of production
increased, while the processors set the selling price of meal so their per-unit returns
become one-half of the original value, discount was offered for bulk purchase, and
COWpea price fluctuated in favor of dry meal), while keeping the assumption of direct
purchase of meal by kosei vendors from meal processors (i.e., no retail margin was added
to the processor price). The results of this sensitivity analysis are shown in Tables 5.18,

5.19, and 5.20.

147

 

As expected, the estimated prices of meal declined sharply (Table 5.18). The
difference, compared to the originally estimated prices, ranged from ¢4,933 to (£13,964
per kg, which led to the lowest meal price of ¢15,3 54 per kg and highest price of ¢16,835

per kg.

Table 5.18 Sensitivity analysis (6): combination of different scenarios——
budgets for cowpea meal processors (¢ per kg of meal)

 

 

 

 

 

 

 

 

 

Respl Resp2 Resp3
Processor price (VAT/N HIS exclusive) 15,741 16,835 15,354
(43,964) (~4,933) (~11,681)
Raw material 6,082 6,082 6,082
(~428) (~805) (~1,228)
Wage 913 391 1,259
(~3,012) (~40) (4,785)
Equipment 753 174 550
(~322) (~15) (~ 192)
Rent 0 0 41
(0) (0) (-155)
Printing & stationery 0 0 0
(~4,861) (~3,000) (~1,206)
Telecommunication 351 218 102
(~1,157) (~21) (~389)
Packaging material 73 73 73
(~477) (~ 1027) (~660)
Miscellaneous 416 236 132
(~ 1 ,3 73) (~24) (~504)
Other costs 4,819 1,643 4,553
Total cost 13,407 8,817 12,792
(11,630) (~4,933) (9,119)
Return 2,334 8,0 1 8 2,562
62.32) (0) (2,5621

 

Note: the numbers in parentheses are the difference from the original values (in Table 5.1).

Source: Calculated by the author.

148

 

The large decline in the prices of dry COWpea meal reduced the difference in
returns between wet- and dry-milled kosei to between ¢1,505 (minimum difference)
and ¢4,570 (maximum difference) per kg of kosei (Table 5.19). Nonetheless, even with
the lowest estimated price of dry meal (¢ 1 5,3 54 per kg), wet-milling still yielded a higher

net margin than use of the dry meal.

Table 5.19 Sensitivity analysis (6): combination of different scenarios—
budgets for kosei vendors (per kg of kosei)

 

 

 

 

 

 

 

 

 

 

 

 

 

. Cow ea .
Pr1ce per mezil/ Total Returns D1fference
kg of meal Budget . 1n returns“
r gram cost (¢) (¢)
(¢) (¢) (¢)

17,657 LB 9,369 18,666 -5,072 3,666 (-4,36 l)
Processors collect [15,354 X MB 7,390 13,528 1,614 2,939 (-3,443)
VAT/NHIS from 1.15] PB 5,411 8,390 8,699 2,211 (-2,525)
kosei vendors 19,360 LB 10,273 19,570 -5,976 4,570 (-8,301)
[16,835 X MB 8,103 14,241 902 3,651 (-6,551)
1.15] PB 5,933 8,912 8,178 2,733 (-4,800)
LB 8,147 17,444 -3,850 2,444 (-3,851)
15,354 MB 6,426 12,564 2,578 1,975 (-3,040)
Processors do NOT PB 4,705 7,684 9,405 1,505 (-2230)
gone"; VAT/N215 LB 8,933 18,230 ~4,636 3,230 (~7,276)
0'" 056' ve" 0’5 16,835 MB 7,046 13,184 1,958 2,594 (5,743)
PB 5,159 8,138 8,951 1,959 (4,209)

Budget for wet-milled kosei with LB 4,752 15,000 '15406

1.10 times higher price of cowpea MB 3,783 10,589 4,553

grain than the original budget p3 2314 6,179 10,910

 

* Difference in returns between wet- and dry-milled kosei. .
Note: the numbers in parentheses are the difference from the original values (1n Table 5.3).

Source: Calculated by the author.

149

 

With the lowest estimated price of meal without the VAT/NHIS being charged
(¢15,354 per kg), PB vendors preparing 10 kg of kosei per day would have to be able to
save at least 0.79 days of one person’s labor by using meal for them to adopt the meal
(based on the opportunity cost of labor per day of ¢19,000) (Table 5.20). MB vendors

would have to be able to save 1.04 days of one person’s labor under the same conditions.

Table 5.20 Sensitivity analysis (6): combination of different scenarios—
Ratio of the total difference in returns between wet- and dry-

milled kosei to the opportunity cost of labor per day

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

13.1.. of Volume.“ .Tma' f"?.§.§i°”l.‘lf?€i§§§.‘15:.”‘20)")
Budget dry meal PFOdUCtIQD dlfference PP L! p y 0
per kg (¢) of kose1 1n returns Opp. cost of labor per day (¢)
0(8) (¢) 10,000 19,000 33,000
15 354 24.441 244 129 64
LB ’ (-38,506) (~385) (-203) (-101 )
19 360 45,699 457 241 120
’ ('834008) (830) (437) (~218)
15 354 19,746 197 104 52
MB ’ ,0 (30,405) (-304) (~160) (-30)
19 360 36.513 365 192 96
, ('651506) (655) (345) (.172)
15,051 151 79 40
PB ”354 (22.303) (-223) (-1 17) (-59)
19 360 27,328 273 144 72
’ 443,004) (-430) (253) (-126)

 

 

Note: the numbers in parentheses are the difference from the original values (in Table 5.4).
Source: Tables 5.4 and 5.19; (opportunity cost of labor) Bank of Ghana.

The relationship between the price of dry COWpea meal and the difference in
returns between wet- and dry—milled kosei is linear, as shown in Footnote 12. The s10pe
and y-intercept, calculated using the highest estimated costs of cowpea grain for kosei
vendors during 2007 (i.e., the values used in the sensitivity analyses [4] and [6]), are 0.53
and ~5,703 for LB vendors, 0.42 and ~4,452 for MB vendors, and 0.31 and ~3,200 for PB

vendors. These relations are plotted in Figure 5.3.

150

Figure 5.3 Relationship between the price of dry cowpea meal and the
difference in returns between wet- and dry-milled kosei

 

 

 

 

 

 

 

 

 

 

 

"52

0 7,600

.2:

(4—

0

on

:4 5,700 7’7 ’

L-

0)

a. A

w 2

E 8 3,800 1

H 0

8 v

.E

:1) 1,900 ‘

U

C

15’

e. o . 1

.0 o o o O O o O o
o o o O O o O o
o o O O o. o o o
o N <1- 50 °° o N ‘7
—- — _- .— v-I N N N

price ofmeal (cedis / kg)
— — LB MB ' ' ' 'PB

 

 

 

Source: Calculated by the author.

The difference in returns between wet- and dry-milled kosei reaches ¢1,900 per
kg of kosei (i.e., ¢19,000——the minimum daily wage—per 10 kg of kosei) when the price
of meal is about ¢14,300 per kg for LB vendors, ¢15,200 for MB vendors, and ¢16,600
for PB vendors. At these prices, vendors would be interested in adopting the meal if the
use of meal allows them to save 1 day of one person’s labor, or ¢19,000 in monetary term,
for 10 kg of kosei (or 1/2 day of one person’s labor for 5 kg of kosei). If the opportunity
cost of labor doubles, or if the use of dry meal saves 2 days of one person’s labor for 10
kg of kosei (or 1 day of one person’s labor for 5 kg of kosei), the acceptable price of meal
would become about ¢17,900 for LB vendors, ¢19,700 for MB vendors, and ¢22,800 for
PB vendors.

Even under the most favorable scenario, the estimated price of meal did not
decline below ¢15,200. On the other hand, the estimated price of meal declined

below ¢19,700 only when: (1) Respondent 3 increased the volume of production while

151

setting the price of meal so the per-unit return becomes one-half of the original value and
without collecting the VAT/NHIS (sensitivity analysis [2]); (2) Respondent 2 offered a
discount for bulk purchase of meal, without collecting the VAT/NHIS (sensitivity analysis
[3]); and (3) Respondents 1 through 3 produced and sold the meal under the most
favorable scenario’s conditions, whether or not collecting the VAT/NHIS (sensitivity
analysis [6]). Thus, these results suggest that, for the majority of kosei vendors, dry
cowpea meal could only be a price—competitive ingredient in kosei under a combination

of very favorable conditions.

5.4 Summary

In this chapter, enterprise budgets were constructed for: (1) industrially-processed
dry cowpea meal; (2) kosei prepared by the wet-milling method (i.e., using COWpea
grain); and (3) kosei prepared by the dry-milling method (i.e., using dry cowpea meal).

Of the six processor respondents producing cowpea/soybean flour, four provided
enough information for constructing the budgets. The raw material (i.e., cowpea grain)
was estimated to be the cost component that accounted for the largest share of total cost.
For cost variables such as wage, transportation, and printing and stationery, the estimated
payment (per kg of output) varied widely across the respondents. The estimated price of
COWpea meal ranged from about ¢2l,800 to ¢29,700 per kg, while the estimated returns
to the processors ranged from about (£4,700 to ¢8,000 per kg.

With regard to the budgets for the wet—milled kosei, the total cost ranged from
about ¢5,900 to ¢14,600 per kg of kosei, while the revenue ranged from about (113,600

to ¢17,100 per kg, making returns range from about ¢~1,000 to ¢11,200 per kg. The most

152

important cost components were cowpea grain and oil.

The budget constructed for the dry-milled kosei found that, under the conditions
observed in Accra during February and March 2007, dry COWpea meal would not be
price-competitive with cowpea grain for the majority of kosei vendors. The difference in
returns between wet- and dry-milled kosei was estimated to range from about ¢3,700
to ¢12,900 per kg of kosei. This means that the meal would only be attractive to those
vendors who could save an amount of labor that is equivalent of 197% to 677% of the
minimum daily wage (¢19,000) for every 10 kg of kosei they prepare (or 98% to 339% of
the daily minimum wage for 5 kg of kosei), by switching from the wet-milling method to
the dry-milling method (i.e., meal), which is a very unlikely condition.

The sensitivity analysis showed that none of the suggested scenarios to improve
the price-competitiveness of meal—improvement in the technical efficiency, increase in
the volume of production (with or without reduction in the processing margin), bulk
purchase of meal, and fluctuation in cowpea price—would greatly change the results of
the original analysis, if these scenarios occurred individually. If all of these scenarios
occurred simultaneously, the price of dry cowpea meal would decline significantly.
However, to adopt the meal, kosei vendors would still have to be able to save an amount
of labor that is equivalent of 79% to 241% of the current minimum daily wage for every
10 kg of kosei they prepare (or 40% to 120% of the daily minimum wage for 5 kg of
kosei). The results suggest that, for the majority of kosei vendors, dry cowpea meal could
only be a price-competitive ingredient in kosei under a combination of very favorable

conditions.

153

CHAPTER 6

WEANIMIX—DESCRIPTIVE ANALYSIS

6.1 Overview of Ghanaian Weaning Foods

As described in Chapters 1 and 2, traditional Ghanaian maize-based weaning
foods are porridges made from fermented maize, called koko, and from roasted maize,
called Tom Brown. Since they are maize-based, both koko and Tom Brown are
inexpensive to prepare but low in protein. At home, they can be either prepared from the
raw material—maize grain—using custom millers for milling maize or from ready-to-use
semi-processed products: fermented maize dough (for koko) and roasted maize flour (for
Tom Brown; the flour is also called Tom Brown). During the fieldwork in Accra,
grain/flour-type product vendors in the market were found to be the major sellers of these
semi-processed products. They often prepared these products by themselves (using
custom millers). Industrially produced and individually packaged Tom Brown was also
available and mostly sold in supermarkets.

As mentioned in Section 2.2.2, Weanimix was developed to enhance the nutrients
of Tom Brown by fortifying it with cowpeas or soybeans and groundnuts. However,
home preparation of Weanimix has not been widely adopted by women who were taught
how to prepare it—possibly due to the higher cost and longer time required to prepare
Weanimix, compared to Tom Brown (G. A. Armor, personal communication, March 26,
2007). Preparation of Weanimix is indeed more costly than preparation of Tom Brown
because the additional ingredients are more expensive than maize (see Section 6.2.6); and

it is more tedious because each ingredient has to be roasted separately. Plahar et al.

154

Figure 6.1 Fermented maize dough Figure 6.2 Tom Brown sold in a market
sold in a market

      

Source: Authbr.
(2003) also pointed out the inconvenience and possible cost ineffectiveness of small-scale
production at the household level, and suggested small-scale enterprise production as a
solution. During the fieldwork in Accra, such Weanimixes/quasi-Weanimixes produced
by small- to medium-scale local companies were observed in supermarkets.

In addition to the above—mentioned traditional and fortified traditional weaning
foods, several brands of weaning foods produced by large-scale multinational companies
were also observed in supermarkets. Among them was Cerelac, produced by Nestle, and
sold in various forms such as “Maize & Milk,” “Rice & Milk,” and “Wheat & Milk.” In
addition, the company launched “Nestlé Wheat & Beans” in 2003
(http://wwwnestleghana.com), a product containing cowpeas as a protein source
(cowpeas are called beans in Ghana). Cerelac was found to be widely sold in small shops
as well. Also observed were a variety of formula produced for weaning-age infants such
as Lactogen (produced by Nestle), Nursie (Blédina), and SMA (SMA Nutrition). These

multinational companies’ weaning foods and formula were in general much more

155

expensive than traditional weaning foods (see Section 6.2.1).

This and next chapters analyze the competitiveness of cowpeas as an ingredient in

Weanimix.

6.2 Description and Implication of Data

This section reports descriptive analyses of the data collected during the fieldwork,
examines non-price-related factors affecting the competitiveness of cowpea-Weanimix,

and processes these data to prepare for the budgeting analysis.

6.2.1 Prices of Different Weaning Foods

A wide range of prices was observed among different weaning foods described
above (Table 6.1).

Fermented maize dough, the ingredient of koko, was the cheapest among the
identified commercial weaning foods, with a mean price ranging from ¢3,211 to ¢5,976
per kg, depending on the season of the year. Tom Brown sold in the market was more
expensive than fermented maize dough, with a mean price ranging from ¢12,295
to ¢15,390 per kg. However, the difference in prices between fermented maize dough and
the other products (in Table 6.1) should be interpreted with caution. Fermented maize

dough sold in the market is moist (see Figure 6.1), while the other products are sold in a
dried form. Therefore, the price per kg of dry-equivalent fermented maize dough would
not be as low as presented in Table 6.1.

Tom Brown produced by small- to medium-scale local companies was more

expensive than Tom Brown sold at the traditional market, costing ¢17,000 per kg.

156

Table 6.1 Mean prices of different types of weaning foods observed in
Accra during February and March 2007

 

 

 

 

 

 

 

 

Low price Price in High price
. during the Feb/Mar during the .
Weaning foods year“ 2007 year“ Mam outlet
(¢/kg) (st/kg) (it/kg) B
Fermented maize dough (produced by 33:11]) 2:23) (1539178 Markets
'n/ fl - d t d ’
gral our type pro uc ven ors) n = 4 n = 4 n = 3
Tom Brown (produced by grain/flour- 12’295 14’295 15’390 Markets
type product vendors) (4’394) (3’630) (4’668)
n=7 n=6 n=7
Tom Brown (Produced by small- to 110(2):; Supermarkets
medium-scale local companies) 11 = 2
Weanimix and quasi-Weanimix“ 23,854 Supermarkets
(produced by small- to medium-scale (7,719)
local companies) 11 = 10
69,167 Small shops
Cerelac (6,221 ) Supermarkets
n = 6
Multinational companies’ weaning 147,183 Supermarkets
foods/ infant formula (other than (34,610)
Cerelac) n = 7

 

 

 

 

 

Note: numbers in parentheses are standard deviation;

“n” denotes the number of

respondents for fermented maize dough and Tom Brown produced by grain/flour-
type product vendors and the number of the variety of products for the other

categories.

* Low/High prices during the year are not prices observed during the fieldwork but prices

reported by the vendors.

** Include products that were named Tom Brown but included ingredients other than
maize and therefore seemed to be appropriate to be called Weanimix.
Source: Fieldwork in Accra, February and March 2007.

Weanimix produced by those companies cost ¢23,854 per kg on average. Weaning

foods/infant formula produced by multinational companies were far more expensive than

weaning foods produced by local companies. Cerelac, the cheapest among them, still

cost ¢69,l67 per kg on average, while the mean price of the other brands observed during

the fieldwork was ¢147,l83 per kg.

These observations clarified a question that needs to be explored. Weanimixes

that were industrially produced by local companies were more expensive than Tom

Brown produced by local companies, and much more expensive than Tom Brown

157

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

produced by grain/flour—type product vendors. In addition to the price difference,
industrially-processed Weanimixes were mostly sold in supermarkets. This implied that
those products were not targeted at mothers in low-income families because supermarkets
in Ghana appeared to be frequented by higher-income consumers ’ . Thus, while
Weanimixes were already commercialized in Accra, they were not yet available to
mothers whose children would most benefit from Weanimix. The question is whether it is
possible to lower the price of Weanimix to a level where low-income families can afford

to buy it, while processors still earn enough returns to motivate them to produce it.

6.2.2 Weaning Mothers
This sub-section summarizes the data collected from 30 weaning mothers. As
described in Section 3.2.4, all the respondents were selected from mothers who visited a

clinic or hospital on a “weighing day.”

6. 2. 2. I Weaning habits

The age of the respondents’ weaned children ranged from 6 to 20 months, with a
mean of 9.6 months. The majority of the respondents gave weaning foods to their child
three or more times a day (Table 6.2).

Among the 30 respondents, 3 always prepared weaning foods by themselves.

' Reardon and Timmer (2007) reported that although a rapid increase in the demand for and supply of
supermarket services, beyond high-income consumers in the capital cities, has been observed since 19905
in developing countries, “sub-Saharan Africa presents a very diverse picture, (p. 2831)” Ghana seems to
be a part of “the great majority of Africa, [that] can be classified as not yet entering a substantial
‘takeoff’ of supermarket diffusion. (p. 2831)”

158

Table 6.2 Weaning habits of the respondents

 

 

 

 

weaning
Weaning habits mother
(n = 30)
How many times per 1 1
day give weaning foods 2* 8
to the child 3 or more 21
Always self-prepare 3
Self-prepare more often than buy 12
Self-prepare or buy Self-prepare as often as buy 4
' . Self-prepare less often than buy 4
weaning foods
Always buy 6
Both self-prepare and buy but don’t
. . 1
know whlch is more often

 

 

 

 

 

* Includes two respondents who answered “twice or three times a day.”
Source: Field survey in Accra, February and March 2007.

Their reasons for not buying commercial weaning foods2 were: (1) her child did not like
the commercial weaning food that she fed the child; (2) her child likes self-prepared
weaning foods very much (but she is thinking of starting to use commercial ones); and
(3) she has never fed her child a commercial weaning food, but she thinks her child might
not like their taste.

A total of 21 respondents fed their child both self-prepared and commercial
weaning foods. The majority of them prepared weaning foods by themselves more often
than they used commercial ones.

Six respondents always purchased weaning foods. Their reasons for not preparing
weaning foods by themselves were’: (1) self-preparing weaning foods is time consuming
(mentioned by 2 respondents); (2) her child prefers the taste of commercial weaning

foods (2 respondents); (3) commercial weaning foods are more nutritious (2

2 In this section (6.2.2), “self-prepared” weaning foods refer to those weaning foods that are prepared from
raw materials (e.g., grains), while “commercial” weaning foods refer to those weaning foods that are sold
in a semi-processed form and purchased by weaning mothers. Therefore, koko, for example, was
considered to be self-prepared if respondents prepared it from maize grain, while it was considered to be
pommercial if respondents prepared it from fermented maize dough that they purchased.

Asked as a multiple choice question.

159

respondents); and (4) her child is still too young to eat self-prepared weaning foods (1

respondent).

6.2.2.2 Selt-gregared weaning Zoods

Information was collected on: (1) whether the respondents used any of the 14
selected ingredients (see Figure 6.3) to prepare weaning foods by themselves; and (2)
three kinds of weaning foods that the respondents prepared most often4, as well as the
ingredients of each of those weaning foods. Data were obtained from 23 respondents—
excluded were 6 respondents who always purchased weaning foods and l respondent

who had just started giving weaning foods to her child and had not yet tried different

types.

Figure 6.3 Ingredients that respondents use to make self-prepared weaning foods

1

1121 Used for any of3 most often prepared w. foods I Used for other w. foom

 

Number ofrespondents

 

F

Fish I
Egg ::-

Rice
Groundnuts

Millet

o
.5

an
E

Cassava
Sorghum
Soybeans
Cowpeas
Meat“
Bambara ‘
beans

 

 

 

 

 

 

 

 

* Excluding two respondents who mentioned that they tried but their child could not chew.
Source: Field survey in Accra, February and March 2007.

 

4 At most three kinds: five respondents mentioned only one kind and another five respondents only two
klnds of weaning foods that they prepared most often.

160

Among the six staples selected, the largest number of respondents used yam
(Dioscorea rotundata) (22 out of 23 respondents), followed by maize (21), rice (21),
millet (18), cassava (14), and sorghum (6). This order and the associated number of
respondents change significantly when the criterion is limited to the ingredients used for
the three most-ofien-prepared weaning foods: maize becomes the most commonly used
(21 respondents), followed by rice (12), yam (10), cassava (10), millet (1), and sorghum
(0).

Among the eight non-staples selected, fish was used by the largest number of
respondents (all 23 respondents), followed by milk (20), soybeans (19), egg (18),
groundnuts (17), cowpeas (17), meat (14), and bambara beans (3). COWpeas and soybeans
were used by about the same number of respondents. However, when the criterion was
limited to the ingredients used for the three most-often-prepared weaning foods, cowpeas
were used by only 2 respondents, while soybeans were used by 15 respondents. It is also
noteworthy that such a high proportion of respondents used nutritious but relatively
expensive food materials such as fish and meat as ingredients in their weaning foods.

The weaning foods that were frequently self-prepared by the largest number of
respondents were koko and/or banku. Banku is prepared from fermented maize dough
like koko, but with less water added so it becomes a solid, which can be formed into
balls’. A total of 20 of 23 respondents mentioned either koko or banku as one of their
three most-often-prepared weaning fooddeoko was always prepared with different
combinations of additional food materials such as milk, fish, and egg, depending on the

respondents. Among the 13 respondents for whom koko was one of the three most-often-

 

5 Fermented cassava flour is often added to fermented maize dough when banku is prepared. Banku is
enjoyed with stew or soup by all Ghanaians (i.e., adults as well as children) in their daily meals, rather than
only used as a weaning food.

161

prepared weaning foods, 10 added soybeans, while only 1 added COWpeas.

Only two respondents mentioned Tom Brown as one of their three most-often-
prepared weaning foods, while no respondents mentioned Weanimix. However, one of
these two respondents fortified Tom Brown with soybeans, and the other fortified it with
soybeans and groundnuts. Therefore, their “Tom Brown” would probably have to be
called Weanimix. In any case, Tom Brown/Weanimix were clearly less popular than
koko/banku among the respondents.

The other three most-often-prepared weaning foods mentioned by the respondents
included rice cooked in different ways such as rice porridge and rice balls (12
respondents), mashed yam (9), and Tuo Zafi (or commonly called T-Z: a food similar to

banku, but made from non-fermented maize and fermented cassava) (6).

6.2.2.3 Commercial weaninfigg foods

Respondents were asked to name up to three weaning foods that they purchased
most often. Data were collected from 26 respondents—excluding 3 respondents who
always prepared weaning foods by themselves and 1 respondent who had just started
giving weaning foods to her child and had not tried different varieties yet.

The overwhelmingly preferred commercial weaning food was Cerelac, which was
used by 23 respondents. Ten respondents mentioned Lactogen and/or SMA, both of
which were mainly used by those respondents to fortify traditional weaning foods that
they prepared by themselves. Eight respondents mentioned maize dough, which they used
to prepare koko or banku". Only three respondents mentioned Tom Brown and/or
Weanimix (Table 6.3).

k

6 Some of them both self-prepared and bought maize dough.

162

Table 6.3 Most often purchased weaning foods selected up to three by the

 

 

respondents

weaning mother

Purchased weaning foods who purchases

weanlng foods

(n = 26)

Cerelac 23
Lactogen and/or SMA* 10
Maize dough 8
Tom Brown and/or Weanimix 3
Other 3

 

 

 

 

* Mainly used for fortification.
Source: Field survey in Accra, February and March 2007.

Among the 26 respondents who purchased weaning foods, 21 were asked whether
they had ever bought Weanimix (i.e., information is missing for 5 respondents). Besides
two respondents who mentioned Weanimix as one of their three most-often-purchased
weaning foods, three had bought Weanimix: one bought it from the hospital once a week
when she came to weigh her child; one bought it from the hospital, but her child did not
like it, and therefore, she discontinued to use it; and one mentioned she once bought Tom
Brown fortified with soybean and groundnut, which most likely was soybean-Weanimix,
but her child did not eat it. The remaining 16 respondents had never bought Weanimix.
Except for the three respondents who had never bought it because they prepared it by
themselves, most of them had never seen it or they did not know what it was. It was
unexpected that such a high proportion of the respondents were not aware of Weanimix,
because weaning mothers visiting a clinic/hospital for weighing their child would most
likely have more opportunities to learn about nutritious weaning foods than weaning
mothers in general.

While most respondents purchased weaning foods produced by multinational
large-scale companies (i.e., Cerelac, Lactogen, and SMA) at small shops, a few

respondents purchased at supermarkets. Most respondents purchased maize dough either

163

 

at the market or from maize dough vendors. Tom Brown was purchased at the market,
and Weanimix at either the market or the hospital.

One of the clinics where the interviews were conducted sold: (1) soybean flour;
(2) soybean-Weanimix; (3) cowpea-Weanimix; (4) rice & soybean flour; and (5) dry fish
powder, packaged in small polyethylene bags7. A nurse said soybeans had been promoted

more than cowpeas.

6.2.2.4 C owpeas versus soybeans as a fortifier

The respondents were asked if they would be willing to buy cowpea- and
soybean-fortified gari8 if they found such products on the shelf in the store. Although the
products for which the question was asked are different from Weanimix, the respondents’
answers were expected to reflect their perception about cowpeas and soybeans as
fortifiers of traditional foods in general, and thereby provide insights on their willingness

to pay for COWpea~ and soybean-Weanimix. The results are shown in Table 6.4.

Table 6.4 Willingness to pay for cowpea~ and soybean-fortified gari
among the respondents

 

 

 

 

 

Willingn_esst£pay (n = 28)*
Would pay more for cowpea- than soybean-fortified gari 3
Would pay more for soybean- than cowpea-fortified gari" 13
Would pay equally more for cowpea- and soybean-fortified gari 9
Would not buy either c- or s-fortified gari if pLices are higher than normal gari 3

 

* weaning mother respondents who would buy either COWpea~ or soybean-fortified gari
** Includes two respondents who would buy only soybean-fortified gari.
Source: Field survey in Accra, February and March 2007.

 

7 Each of them was for 132,000. On the day when the interview was conducted, only soybean flour, soybean-
Weanimix, and rice & soybean flour were available for purchase. We bought and weighed one of each
formulation. The weight was between 60 and 80 g.

As mentioned in Chapter 1, cowpea-fortified gari (grated, fermented, and roasted cassava) was initially
among the target products for this case study. The questionnaire for weaning mothers was designed to
examine the difference in their willingness to pay for cowpea- and soybean-fortified gari rather than
cowpea- and soybean-Weanimix.

164

 

Among the 28 respondents who showed their interest in buying either cowpea- or
soybean-fortified gari, 3 were willing to pay more for cowpea- than soybean-fortified gari,
while 13 were willing to pay more for soybean- than for cowpea-fortified gari.

The result showed a higher willingness among the respondents to pay more for
soybean- than for cowpea-fortified gari. While no generalization can be made from these
data because of the limited sample size, this result suggests a need to conduct a
comprehensive consumer survey. Interviewed mothers preferred soybean- over COWpea~
fortified gari, which suggest that consumers may prefer soybean- over cowpea-Weanimix.
If such a preference is commonly held by weaning mothers in Accra, it implies that
cowpeas face a significant challenge as an ingredient in Weanimix~especially because

soybeans are less expensive than cowpeas (see Section 6.2.6).

6.2.2.5 Comparison with findings of_qnother sum

The descriptive analysis reported above suggests a different picture about
weaning mothers in Accra, compared to the results of a study conducted by Mensa-
Wilmot et al. (2001b). In their study, 133 weaning mothers were randomly selected and
interviewed at “residential areas, open-air markets and Maternal and Child Health (MCH)
centers in the city suburbs and rural areas in the Greater Accra Region” (pp. 84-85). Their
findings, among others, were that: (1) 50% of the respondents had never bought
“proprietary food” (the term seems to refer to multinational companies’ weaning foods);
and (2) 87% used cowpeas in the preparation of weaning foods, while 65% used

soybeans.

165

Mensa-Wilmot et al. (2001b) stated cost constraints as a reason why one-half of
the respondents chose cereal-based traditional porridges over multinational companies’
weaning foods. In contrast, in this study, 26 out of 30 respondents (87%) indicated that
they had bought multinational companies’ weaning foods at least once. Moreover, for
most of the remaining four respondents, cost did not seem to be the reason why they did
not buy those weaning foods. It is not clear whether this discrepancy in results is caused
by an increase in the wealth of weaning mothers in Accra since the previous study was
conducted or due to a potential bias in our data because of: (1) the limited sample size; or
(2) the population from which the sample was drawn—i.e., weaning mothers visiting a
clinic/hospital to weigh their child. This population would most likely include a higher
proportion of mothers who are aware of, or concerned about, child nutritional issues than
exists in the general population. However, it should be noted that our sample of weaning
mothers, who were rich enough to purchase expensive multinational companies’ weaning
foods, were most likely not representative of mothers whose children were malnourished
and therefore would benefit from Weanimixq. Therefore, it would most likely be
inappropriate to extrapolate the answers from our respondents to the low—income weaning
mothers for whom Weanimix was originally targeted.

The data collected by Mensa-Wilmot et al. (2001b) indicated that cowpeas were
more popular than soybeans for the ingredients in weaning foods. However, Mensa-
Wilmot et al. (2001a, p. 850) also observed that soybeans were “gaining popularity as a
nutritious ingredient for infant food supplementation due to the efforts of public health

workers for the Ministry of Health.” In our study, the respondents preferred soybeans to

9As far as the author could observe during the interview, all the weaned child of the respondents looked
healthy, and no apparently malnourished children were observed.

166

 

cowpeas (see Figure 6.3). Again, it is not clear whether the discrepancy is caused by the
actual change in the popularity of soybeans among weaning mothers in Accra over time
or by a potential bias in our data. Another interesting finding by Mensa-Wilmot et al.
(2001b) was that 23 focus group interviews conducted in traditional villages in the
Central Region revealed that the respondents were not familiar with aroma of soybeans
used in the sample of weaning foods that they prepared—as mentioned in Section 2.2.3,
the researchers could not determine if the respondents were familiar with soybeans
because there was no word for soybean in the local language. The implication of these
findings is that the popularity of soybeans would most likely vary across time and regions
in Ghana. If so, it would be important to accurately assess soybean popularity among the
target population of Weanimix, because the knowledge and acceptance of soybeans
among the target population would greatly affect the competitiveness of cowpea~

Weanimix with soybean-Weanimix.

6.2.3 Industrial Wearflgflood Processors

Including the 6 COWpea/soybean flour processors analyzed in Chapter 4, a total of
10 industrial grain flour/weaning food processors were interviewed to collect information
on their production of weaning foods. This section presents a descriptive analysis of the

data collected from them.

6. 2. 3.] Characteristics Qth’dt’litLflfOOQt‘OCéSSOl‘S

The characteristics of 10 respondents are summarized in Table 6.5.

167

Table 6.5 Characteristics of weaning food processor respondents

 

 

 

 

 

Characteristics processor
(n — 10
Female
Sex of the manager Male
Missingiwnformation
Less than 5
Number of years the company 5-10
has been in business 11-20
More than 20
1~5
Total number of workers (15; {(1)5

(including family members)

More than 15

 

Missing information

 

Monthly revenue from all
products* (¢; 1 US$ :: ¢9,200)

Less than 50 million
50-100 million
More than 100 million

 

Missing information

 

 

1~5
Number of products (including 6-10
weaning foods) 11-15
More than 15
Cowpeas (not soybeans)

Weanimix or quasi-Weanimix
includes as an ingredient:

Soybeans (not COWpeas)**
Both cowpeas and sybeans

 

N.A. (do not produce Weanimix but Tom Brown)

 

Number of years the company is
producing Weanimix or quasi-
Weanimix

Less than 2
2-5

6-10 ,

10-15

 

Missing information & N.A.

 

Less than 10%

 

 

 

 

Share of Weanimix or quasi- 10-20%

Weanimix in the total revenue* More than 20%
MissLng information & N.A.
Vendor in the market

Major source of raw materials Wholesalers
Owned farm
Middlemen/Sup1iiers

 

W—-—MAN-—-WNNOAN——\I—N-bNN—N—O\——NAN—J>NW~WO\

 

* Revenue of the month for which the data were collected (mostly January or February
2007); approximate values were used for those respondents who reported, instead of
the monthly revenue from all products, (1) the quantity and price of each product
produced—the value of products was used as the approximate revenue, assuming
that all the products were sold out in the month they were produced; and (2) the
annual revenue in the past year—the figures were divided by 12 and used as
approximate monthly revenue.

** Includes one respondent who was using soybeans at the time the interview was
conducted, but reported also using cowpeas instead of soybeans depending on their

availability.

Source: Field survey in Accra, February and March 2007.

1.68

 

The majority of the companies interviewed (6 out of 10) were managed by a
woman. The number of years that the companies had been in business varied across the
respondents; all except 1 respondent had been in business for fewer than 20 years. The
total number of workers (including family members) also varied; most respondents had
fewer than 15 workers. Monthly revenue from all products was obtained from nine
respondents, including assumed values (see notes for Table 6.5): the revenues were less
than ¢50 million (US$5,435) for the majority of respondents. The number of products
produced by the company also varied across the respondents; the majority produced more
than 10 products.

Among the nine respondents'0 who produced Weanimix or quasi-WeanimixI ', six
used soybeans rather than COWpeas as an ingredient in their Weanimix; only one used
cowpeas rather than soybeans; one used either cOWpeas or soybeans, depending on the
availability at the owned farm and was using soybeans when the interview was
conducted; and one used both cowpeas and soybeans. The majority of the respondents
had produced Weanimix for fewer than five years. There was variation across the
respondents in the share of Weanimix in the total revenue.

The major sources of raw materials were vendors in the market for five
respondents, wholesalers for one respondent, owned farm for one respondent, and

middlemen or suppliers for three respondents.

 

:2) Among the 10 respondents, 1 did not produce Weanimix/quasi-Weanimix but produced Tom Brown.

In this study, quasi-Weanimix refers to a product that (1) is composed of three roasted ingredients of
standard Weanimix—maize, either cowpeas or soybeans, and groundnuts—but the share of each ingredient
is very different from that of standard Weanimix (i.e., 7.5 or 8 : 1.5 or 1 : 1) or (2) contains roasted
ingredient(s) in addition to the three standard ingredients. Hereafter, when it is unnecessary to distinguish
between these different formulations, the term Weanimix is used to refer to both Weanimix and quasi-
Weanimix.

169

 

6.2.3.2 Tom Brown and Weanimix

Among the nine respondents who produced Weanimix, four also produced Tom
Brown, while the remaining five did not. The majority of Weanimixes produced by the
respondents, as well as other local processors who were not interviewed but identified
during the fieldwork, were actually named Tom Brown. Most of the few respondents who
“correctly” named their product Weanimix also produced Tom Brown, and perhaps
needed to use the term Weanimix to differentiate the product from Tom Brown.

The reason the respondents called their product Tom Brown, rather than
Weanimix, was not asked during the interview. A hypothesis is that Weanimix was not
well known among consumers—which was the case with weaning mother respondents,
and therefore the products sold better by being marketed under the name of well-known
Tom Brown. However, a disadvantage of this naming practice is that their Weanimix,
which was more nutritious and more costly to produce, was recognized by consumers as
the same product as Tom Brown, unless they carefully read the label and found the
difference in ingredients.

Among the four respondents who produced both Weanimix and Tom Brown, one
stated that if the price of Weanimix was higher than Tom Brown, people would not buy it.
So, this respondent actually set the price of Weanimix the same as Tom Brown, while
three other respondents set the price of Weanimix 1.2 to 1.9 times the price of their Tom
Brown.

Respondents’ Weanimix and quasi-Weanimix, whether they called it Weanimix or
Tom Brown, were mostly composed of the same ingredients as standard Weanimix.

However, the share of each ingredient varied across the respondents’ formulations. The

170

 

only ingredient other than the standard three was millet used by two respondents. Among
the nine Weanimixes/quasi-Weanimixes produced by the respondents, six had a higher
share of cowpeas and/or soybeans than the standard share of 10~15% (used in the original
formula for Weanimix), two had the standard share, and one had a lower share. In terms
of groundnuts, four had a higher share than the standard share of 10%, while five had a

lower share.

6.2.3.3 Processing procedure and equipment use

 

There were no major differences between the model processing procedure to
prepare Weanimix (see Figure 2.3) and the processing procedures that the respondents
used to make their Weanimix. The only difference worth noting is that some respondents
stated that they sieved their product. To analyze the cost of producing the same quality of
Weanimix across the respondents, in the budgeting analysis reported in the next chapter,
all the respondents were assumed to use a sieve, and the purchase price of a sieve was
added to the cost of equipment.

Major equipment needed to produce and sell Weanimix on a commercial basis
are: (1) a roaster (or other equipment that can serve as a roaster, such as oven); (2) a mill;
and (3) a packaging machine. However, a processor would not necessarily have to own a
roaster or mill because custom roasters and custom millers were available. Also, a toaster
could be just a roasting pan (see Figure 6.4). Therefore, the capital seemed to be a
minimal constraint to enter the business of producing and selling Weanimix. The

equipment ownership among the respondents is shown in Table 6.6.

171

 

Figure 6.4 Roasting pans Figure 6.5 Roaster

   

Source: Author

Table 6.6 Equipment for producing Weanimix and possession by the
respondents

food
processor

Equipment inventory

roast

Own a mill

No

   

Own a sealing machine

Source: survey

Two respondents custom-roasted ingredients, while eight respondents roasted
ingredients at their facility using equipment such as roasting pan, roaster, or gas oven.
The type of roaster shown in Figure 6.5 was fueled by gas for heat and electricity for the
motor (to stir), and was the most expensive equipment among those three types:
about 11311 million (US$1,196) in 2007 cedis (mean of 3 observations, with a standard
deviation of 161.5 million). The cost of a gas oven, the second most expensive piece of
equipment, varied widely across the respondents, ranging from about ¢1.4 million
(US$153) to about ¢7.1 million (US$768) (sample size of 3). A roasting pan, the cheapest

piece of equipment, cost about ¢90,000 (US$10) per pan.

172

Four respondents custom-milled ingredients, while six respondents owned a mill
and milled ingredients at their facility. All the respondents owned sealing machines. Most

of the sealing machines observed during the interview were hand sealers.

6.2.3. 4 Demand for Weanimix and its .917le life

For most respondents, retailers were the only or the major customer of their
Weanimix. None of them mentioned selling their products to hospitals or schools. For one
respondent, export was the major outlet.

No respondent produced the same amount of Weanimix throughout the year.
Rather, the majority changed the amount produced depending on the demand. The
majority of the respondents sold their Weanimix at the same price throughout the year.
One respondent reported increasing the price when it was impossible to keep it constant,
taking different factors into consideration such as changes in raw material and labor costs.

Among the five respondents who had been in business for more than three years
and had regularly sold Weanimix mainly to domestic customers, four mentioned that their
production of Weanimix increased over the past three years. Among these processors,
three (1 processor of cowpea-Weanimix [named Tom Brown] and 2 processors of
soybean-Weanimix [both named Weanimix]) mentioned they increased production
because customers became more aware of or patronized their product, while another
soybean-Weanimix processor attributed the production increase to his business expansion
in general. Another respondent stated that the amount of production fluctuated over the
past three years.

Among the nine respondents who produced Weanimix, eight stored their product

173

at their facility. The maximum storage time ranged from 6 to 36 months”. Based on this
information, it was assumed in the sensitivity analysis that the respondents could produce
Weanimix when the price of raw materials was the lowest during the year and sell outputs

for up to one year, avoiding the price fluctuation of inputs.

6.2.3.5 Cowpeas versus soybeans as an ingredient in weaning foods

While the respondent who produced cowpea-Weanimix had never used soybeans
as an ingredient in weaning foods, this respondent had no special reason for using
cowpeas rather than soybeans.

Among the seven respondents who produced soybean-Weanimix and the one
respondent who produced Tom Brown and another kind of weaning food, five had never
used cowpeas as an ingredient in their weaning foods. Of these five processors, three
mentioned that soybeans were more nutritious; one mentioned that she might try cowpeas
later but was used to using soybeans; and the other one mentioned that cowpeas would
cause gas in the stomach and therefore would not be good for weaning foods. Another
respondent used to produce a maize-based weaning food containing both cowpeas and
soybeans as ingredients, but dropped cowpeas and added groundnuts. The reason given
for staying with soybeans was that the respondent heard that people were getting to know
about soybeans. Another respondent used to produce COWpea-Weanimix, but switched to
soybean-Weanimix because cowpeas became too expensive. As mentioned earlier, one
respondent used either cowpeas or soybeans, depending on the availability at his own

farm.

¥

'2 It was not clear whether the respondents answered the actual storage time or shelf life of their product.
The latter can be longer than the former.

174

 

One respondent preferred the taste of soybeans over that of cowpeas, and another
respondent mentioned that soybeans had a nice aroma. Easier storage of soybeans
compared to cowpeas, which easily gets weevil infestation, was mentioned by one
respondent as an advantage of soybeans. As for the advantage of cowpeas, two
respondents mentioned that cowpeas were nutritious, and another respondent mentioned
that cowpeas were better known than soybeans. However, one respondent mentioned that
the difficulty of obtaining a constant supply of some varieties of COWpeas as a constraint
to using cowpeas as an ingredient in weaning foods. Another respondent was concerned
that some people might not like the taste of weaning foods containing cowpeas13 .

An unexpected result was that only one respondent mentioned the higher price of

cowpeas, compared to the price of soybeans, as a constraint to using cowpeas.

6.2.4 Cost of Custom Milling

The information on the charge for custom-milling roasted maize (to prepare Tom
Brown) and a mixture of roasted maize, cowpeas/soybeans, and groundnuts (to prepare
Weanimix) was collected from the 15 custom miller respondents, 7 retailer respondents
who self-prepared Tom Brown, and 4 weaning food processor respondents.

Among the custom miller respondents, groundnuts were an unpopular crop to mill

because, according to them, it is tedious to clean the machine after milling groundnuts.

 

'3 There was also one respondent who mentioned as a disadvantage of COWpeas, that cowpeas had to be
dehulled because people did not like to have black-eyes in foods, which would add an extra step in the
processing procedures. However, this respondent had never used cowpeas as an ingredient in weaning
foods, and this statement might not reflect reality. Among the respondents who had experience using both
cowpeas and soybeans as an ingredient in their Weanimix, one mentioned that black-eyes were removed
while sieving (all the respondents who stated they sieved their Weanimix were processors of soybean-
Weanimix. Therefore, it would not be a task only for cowpea-Weanimix), and another respondent said that
there was no difference between the processing procedures for cowpea- and soybean-Weanimix. Therefore,
in this study, the processing procedures for cowpea- and soybean-Weanimix were assumed to be identical,
and no difference in the cost of production associated with processing procedures was considered.

175

However, most custom miller respondents answered that they would charge the same
price for milling roasted maize and a mixture of roasted maize, cowpeas/soybeans, and
groundnuts, as long as the mixture was maize-based”. Also, all the respondents indicated
that the charge would not change whether cowpeas or soybeans were added to the maize.
These answers imply that there is little or no difference in the labor needed to mill the
grains for preparing Tom Brown, cowpea-Weanimix, and soybean-Weanimix.

A representative charge for custom-milling to prepare Tom Brown/Weanimix was
derived from a total of eight retailer and weaning food processor respondents who
custom-milled the ingredients in bulk (i.e., mill 10 olonka or more at once). The charge
was ¢1,857 per olonka of raw materials (with a standard deviation of ¢764). This derived
value was used in the budgeting analysis to estimate the cost of custom-milling that

grain/flour-type product vendors paid to self-prepare Tom Brown/Weanimix.

6.2.5 Retailers
6. 2. 5.1 Retail margins

To estimate representative retail prices of industrially-processed Tom Brown,
COWpea—Weanimix, and soybean~Weanimix, the mean retail margin of 29% for small
shops (potential outlets, as Opposed to the current outlet, which is supermarkets), which
was derived in Section 4.2.4, was used as the representative retail margin in the budgeting
analysis”.

Although industrially produced Tom Brown and Weanimix are subject to the

 

 

'4 Since it was expected that not all custom millers had previously milled a mixture of roasted maize,
cowpeas/soybeans, and groundnuts, the question was asked “How much would you charge if a customer
brings a mixture of...?”

Since industrially-processed Tom Brown and Weanimix are subject to the VAT/NHIS (see discussion in
Section 4.2.4), the representative margin used included the VAT/NHIS.

176

 

VAT/NHIS, some processor respondents did not collect the VAT/NHIS, and as mentioned
in Appendix 6, it was not clear whether small shops collected the VAT/NHIS. Therefore,
in the sensitivity analysis, three different ranges in retail margins were assumed, based on
the data reported in Tables 4.12 and 4.13: (1) small shops without VAT/NHIS: 3%-3 1%;
(2) small shops with VAT/NHIS: 19%-50%; and (3) supermarkets with VAT/NHIS: 38%-

57%.

6.2.5.2 Grain/flour-tiipegroclztct vendors as processors of Tom Brown/Weanimix

Interviews were conducted with seven grain/flour—type product vendors who self-
prepared Tom Brown and sold it in the market. One of them also self-prepared soybean-
Weanimix. Only two of them bought maize from markets in Accra, while the remaining
five bought maize from different regions. Among these five respondents, two bought
maize from farmers, while another two from wholesalers (information is missing for l
respondent). There were two respondents who custom-roasted maize, while the other five
roasted maize by themselves using their own pan(s).

Among the seven respondents, four changed the selling price of Tom Brown
depending on the season of the year, while three did not change the price (for the range in
price change, see Table 6.1). However, as discussed in Section 4.2.5.3, the actual amount
of commodities sold as “1 olonka” can vary because of the vendors’ subjective measuring
practices. Therefore, unit prices of Tom Brown might actually fluctuate more than
reported in Table 6.1.

Although many respondents mentioned that their customers could be anybody,

their main customers seemed to be weaning mothers and students.

177

 

6. 2. 6 Mesentative PriceMMaizeLCowpeasL Groundn uts, and Soybeans

As was done for cowpeas in Chapter 4, representative prices of maize, groundnuts,
and soybeans to be used in the budgeting and sensitivity analyses were derived from the
secondary data.

To obtain a general picture of the differences in prices of these crops, historical
wholesale real prices (base month = Feb. 2007) of maize, cowpeas, and groundnuts at
urban markets of the Greater Accra Region, obtained from the MoFA, were plotted in
Figure 6.6, along with prices of soybeans reported by a wholesaler in Nima market (as

mentioned in Section 3.3, the MoFA does not collect soybean prices).

Figure 6.6 Real maize, cowpea, and groundnut monthly wholesale prices
at urban markets of the Greater Accra Region, Ghana (average of
Accra, Ga, & Tema Districts) (2002-2006) and prices of soybeans
reported by a wholesaler in Nima market (inflation adjusted)

 

 

1 12,000 ~ — - -- ~ 2-222252, ~-- - “

 

 

 

 

 

 

“—— ~- --—“1
1 10.000 l , ~ .
|
. l
a, 8,000“ , ~ ' . '
x r
\ Q
.2 6,0009 ,
'0 1
Q)
U
4,0001 ~
2.000“ - i
O T I r T— I v—T 7"— T— I I j
N N N ('1 t") M V V V W V5 V) \O \O \O
9 C? <~7 <.= 9? <.> 9 s» 9 e O. 9 e C? <.>
C: >\ D. C >1 0- : >5 D. C >\ G. ‘1' >5 0.
(U {U Q) (3 ('5 U (U (U 0) (U (U Q) N (U Q)
“r 2 m “3 2 W " 2 W "‘ 2 W _. 2 m
I L:—0—maize —B-—-cowpeas +groundnuts +soybeans]

 

 

 

 

Note: base month = Feb. 2007.

Source: (maize, cowpeas, groundnuts) SRID, MoFA; (soybeans) a wholesaler in Nima
market; (CPI) IMF.

178

As shown in Figure 6.6, the prices of maize, cowpeas, and groundnuts have
followed a similar trend during the last five years”. Maize has always been cheaper than
cowpeas, and groundnuts were almost always more expensive than cowpeas. The
soybean prices plotted in the figure were derived from the soybean prices recalled by a
wholesaler in the Nima market. The lowest adjusted pricel7 was 953,889 per kg, and the
highest price was ¢4,548 per kg. Assuming that the price of soybeans fluctuated within
this range during recent years, soybeans have been a cheaper ingredient in Weanimix than
cowpeas.

To obtain representative prices of maize and groundnuts in February 2007, data
sets were collected from Tradenet (http://www.tradenet.biz). For both maize and
groundnuts, the price series in Makola market (Accra Metropolis) and Tema (Tema
Municipal) were very close to the price series of the MoFA until December 2006 (when
the MoFA data ended). Therefore, it was decided to use the mean price in the Makola
market and Tema in February 2007 as representative prices. The mean price of maize
was (33,000 per kg, while the mean price of groundnuts was ¢7,650 per kg.

The representative price of cowpeas that was derived for the analysis of dry
cowpea meal (¢5,474 per kg; see Section 4.2.5.2) was also used in the analysis of
Weanimix.

Since no data on prices of soybeans could be obtained from either the MoFA or
Tradenet, the representative price in February 2007 was approximated by the mean of the
four wholesale prices observed by the author at the Nima market in March 2007. The

derived price was (254,069 per kg (with a standard deviation of ¢298).

’6 The price of cowpeas in Figure 6.6 is the same as reported in Figure 4.7.
The prices provided by the wholesaler were converted to real prices per kg (base month = Feb. 2007),
using the olonka-kg conversion rate for soybeans (see Appendix 2) and the CPI.

179

 

Then, the seasonal fluctuation in the prices of maize and groundnuts was
examined to derive the lowest and highest representative price estimates during the year
2007. As done in Chapter 4, seasonal indices were constructed for maize and groundnuts,

using the data presented in Figure 6.6. The result is shown in Figure 6.7.

Figure 6.7 Seasonal index of maize, cowpea, and groundnut wholesale
prices at urban markets of the Greater Accra Region, Ghana
(average of Accra, Ga, & Tema Districts) (2002-2006)

 

 

seasonal index

 

 

 

70 T 7 f , e , , . 5 r T
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

 

 

 

1—0— maize —B— cowpeas + groundnuts

 

 

Note: the indices for COWpeas are the same as reported in Figure 4.8.
Source: (for the price series) SRID, MoFA; (for the CPI) IMF.

Figure 6.7 shows that among the three crops, maize has had the widest price
fluctuation over the last five years, followed by groundnuts. Cowpea prices have been the
most stable. Using the lowest and highest seasonal index, the seasonal index for February,
and the representative prices derived above for maize and groundnuts, the lowest and
highest representative prices for these two crops were estimated. The lowest price of
maize was ¢2,541 per kg; the highest price of maize was ¢3,809 per kg; the lowest price

of groundnuts was ¢6,240 per kg; and the highest price of groundnuts was ¢8,486 per kg.

180

To summarize, the representative prices of maize, COWpeas, groundnuts, and

soybeans are presented in Table 6.7.

Table 6.7 Representative maize, cowpea, groundnut, and soybean prices
paid by processor respondents in February 2007 and estimated
range in price fluctuation during the year

 

 

 

 

Crop Lowest February Highest

(¢ / kg) (9‘ / kg) (¢ / he)
Maize 2,541 3,000 3,809
Cowpeas 5,155 5,474 6,221
Groundnuts 6,240 7,650 8,486
Soybeans 3,889 4,069 4,548

 

 

 

 

Source: Calculated by the author based on the data collected from: (maize and
groundnuts) Tradenet (http://www.tradenet.biz) and SRID, MoFA; (COWpeas) Table
4.14; (soybeans) wholesalers in Nima market.

In the sensitivity analysis, the effects of changes in the price of raw materials on
the price-competitiveness of cowpea-Weanimix will be examined. As will be shown later,
cowpea-Weanimix becomes more competitive with Tom Brown as the price of maize
increases and the prices of cowpeas and groundnuts decrease. However, the combination
of the above-derived highest representative price of maize and the lowest representative
prices of cowpeas and groundnuts could be an unrealistic assumption. As shown in Figure
6.7, the prices of maize, cowpeas, and groundnuts have followed a similar seasonal trend,
although the peak month seemed to slightly differ across the crops. This implies that
when the price of maize is high, the prices of cowpeas and groundnuts are also expected
to be high, rather than low.

Whether the combination of the highest representative price of maize and the
lowest representative price of cowpeas and groundnuts is reasonable was examined as
follows: (1) from the MoFA data, the maize prices that fell within the lowest

representative price of maize :t ¢250 per kg were selected; (2) the relationship between

181

the price of cowpeas and groundnuts in the same month from when the prices of maize
were selected was plotted in a graph; and (3) the relationship between the lowest
representative price of cowpeas and groundnuts was plotted in the same graph. If the
assumption was unrealistic, the combination of the lowest representative price of
cowpeas and groundnuts has to be found as an outlier in the graph. The result is shown in

Figure 6.8.

Figure 6.8 Relationship between the high price of maize and the prices of
cowpeas and groundnuts in the same month

 

 

 

 

 

 

 

 

 

 

I-..
‘ 10,000 - -
o 1 -_
91000 2 ’ 5 0 prices of cowpeas &
37, 8,000 1 . t ‘ groundnuts corresponding
1‘ 7,000 - . j to maize prices from original
cn . .
§ 6,000 1 : I 1 data w1thln.HlGI-IEST.rep.
3 , pr1ce ofmalze d: 250 cedis/kg
5’ 5’000 1 l n 1 t ' f
:3 . 7 owes .rep. pnces o
'E 4’000 1 ., cowpeas & groundnuts 11
g 3,000 .. ‘ '11
‘59 2,000 . 9 1
1900 1 l 2* 11
O 1 1 1 1 1 1
0 2,000 4,000 6,000 8,000 1 0,000 1
L COWpeas (cedis / kg) 1

 

 

Source: Derived by the author based on the data collected from: (for the price series)
Tradenet (http://www.tradenet.biz) and SRID, MoFA. ; (for the CPI) IMF.

The position of the combination of the lowest representative price of cowpeas and
groundnuts in Figure 6.8, relative to the other points, suggests that the assumption is
viable.

In the same way, the viability of using the combination of the lowest
representative price of maize and the highest representative price of cowpeas and

groundnuts was examined as the most unfavorable price sets for cowpea-Weanimix. The

182

 

result is shown in Figure 6.9.

Figure 6.9 Relationship between the low price of maize and the prices of
cowpeas and groundnuts in the same month

 

 

 

 

 

 

 

 

 

10,000 71—7—--7777- “7777—fr—fl- ~~v
9’000 g ‘ 2 If, I 9 prices ofcowpeas & 1
33 8,000 1 .0 8 #—‘5 1 groundnuts corresponding 1
'1‘ 7,000 ~ .9. 1 to maize prices from original '
g 6,000 1 . 1 data within LOWEST. rep.
3 , 1 price ofmaize i 250 cedis/kg?
g} 5,000 +77 7 ’ 1 . _
2 4 000 , 1 A hlghest. rep. pnces of
“g ’ I cowpeas & groundnuts
'3 3,000 1
2 1
W 2,000 7 7 7 77 77 7 1
1,0001 77 1 2
l 0 n 1 1 1 fl
1 0 2,000 4,000 6,000 8,000 10,000
1 cowpeas (cedis / kg) .
l

 

Source: Derived by the author based on the data collected from: (for the price series)
Tradenet (11ttp:wawtradenet.biz) and SRID, MoFA. ; (for the CPI) IMF.

Again, the position of the combination of the highest representative price of
cowpeas and groundnuts in Figure 6.9, relative to the other points, suggests that the
assumption is viable.

Based on these results, the sensitivity analysis (with respect to the price change in
raw materials) was conducted using the combination of the representative price of maize,

COWpeas, and groundnuts as described above.

6.3 Summary
A wide difference was found among the prices of existing weaning foods
available in Accra. Industrially produced Weanimix/quasi-Weanimix were mainly sold at

supermarkets and therefore seemed to target high-income families.

183

 

All the 10 processor respondents interviewed managed small- to medium-scale
local companies, most of which had fewer than 15 workers (including family members).
The majority of these companies produced more than 10 different products, and had
produced Weanimix/quasi-Weanimix for fewer than five years. The share of Weanimix in
total revenue varied across the respondents. Demand for Weanimix was not constant
throughout the year. For most respondents, retailers were the only or major customers for
their Weanimix.

These companies’ Weanimixes were mostly composed of the same ingredients as
standard Weanimix (i.e., maize, cowpeas/soybeans, and groundnuts). However, the shares
of each ingredient differed across the respondents. Many of these products were actually
named Tom Brown. While the processor respondents indicated that consumers’
awareness and the demand for Weanimix had increased in recent years, answers from
weaning mother respondents suggested that the term Weanimix was not yet widely
known among consumers.

Among the processor respondents, soybeans were more popular than cowpeas as
an ingredient in industrially-processed weaning foods. The advantages of soybeans
mentioned included higher nutritional value, better taste, lower price, and easier storage
of grain. However, the lower price of soybeans did not seem to be the major reason for
most of the respondents to choose soybeans over cowpeas.

With regard to processing procedures, there were few differences between the
procedures used by the respondents to make Weanimix and the model procedure
(presented in Section 2.2.2). Most of the 15 custom miller respondents answered that they

would charge customers the same price for milling roasted maize and a mixture of

184

 

roasted maize, COWpeas/soybeans, and groundnuts, as long as the mixture was maize-
based. Their answers imply that there is little or no difference in the labor needed to mill
the grains for preparing Tom Brown, cowpea—Weanimix, and soybean-Weanimix.

Of the seven grain/flour—type product vendors who self-prepared Tom Brown, two
custom-roasted maize, while the other five roasted maize by themselves using their own
pan(s). The main customers for their Tom Brown seemed to be weaning mothers and
students.

Of the 30 weaning mother respondents, all of the 23 who were asked about their
self-prepared weaning foods could afford to use fish as an ingredient; and 23 out of the
26 respondents who were asked about their purchased weaning foods could afford to buy
Cerelac. These facts indicated that our sample of weaning mothers did not include
weaning mothers from low-income families whose children were malnourished and
therefore would benefit most from Weanimix.

Also, this chapter discussed how representative prices of maize, cowpeas,
soybeans, and groundnuts were derived from prices observed during the fieldwork and

secondary data.

185

CHAPTER 7

WEANIMIX—BUDGETING AND SENSITIVITY ANALYSIS

7.1 Budgeting Analysis—Model

A model similar to the one used to construct the budgets for industrially-processed
dry COWpea meal (in Chapter 5) was used in this chapter to estimate retail prices of Tom
Brown, COWpea-Weanimix, and soybean-Weanimix. The prices were estimated for each
of these weaning foods that are: (1) produced by industrial processors and sold at small
shops (potential outlets, as opposed to the current outlet of supermarkets); and (2) self-
prepared and sold by grain/flour-type product vendors in the market. During the
fieldwork, no grain/fiour-type product vendors were found selling cowpea-Weanimix and
only one vendor was selling soybean-Weanimix. Therefore, Weanimix self-prepared by

grain/fiour-type product vendors is considered to be a potential new product.

7.1.1 Estimated Prices of Industrially-Processed TomfiBlown and Weanimix
Processor prices of industrially-processed Tom Brown, COWpea-Weanimix, and
soybean-Weanimix were estimated for each processor respondent using the following

equation:

186

 

,V Z
_fle+_fl
j! - Q}! Q); (15)

Jr 1’

where,
P = price per kg,
2 = target weaning food
(Tom Brown, cowpea-Weanimix, and soybean-Weanimix)
j = weaning food processor, t = time period, 17 = return,
y = currently produced weaning food product
(Tom Brown or cowpea- or soybean-Weanimix/quasi-
Weanimix)
Q = quantity in kg, C = total cost payment
As described in Section 3.6.2, a budget for each respondent was first constructed for the
weaning food product that the respondent currently produced. When respondents
produced different weaning foods, the same product or one closest in characteristics to
our target weaning foods was selected (denoted as “y” in equation [15]). Then, holding
the returns to the processor and quantity of production constant, processor prices of our
target weaning foods were estimated by replacing the cost of production of the selected
currently-produced weaning food with the estimated cost of production of each of our
target weaning foods (i.e., Tom Brown, cowpea-Weanimix, and soybean-Weanimix;
denoted as “2” in equation [15]).

The cost of production was calculated in the same way as the cost of production

of dry COWpea meal was calculated, using the following equation:

187

Cit = Zpathgjt +6jtZijt +Z7Sthejz +211:thnd (16)
a w e n

where,
a = type of raw material
(maize, COWpeas, soybeans, and groundnuts)
M = raw material in kg, 6 = share in wage, W = wage per worker,
w = worker, e = piece of equipment, y = share in equipment use,
E = equipment payment,
n = other cost component
(electricity, fuel [excluding fuel for vehicle], water, rent,
transportation [including fuel for vehicle], printing and
stationery, telecommunication, packaging material, and
miscellaneous)
/1 = share in other cost component, X = other cost payment
The shares of the target weaning food in different cost variables were approximated in the
same way as the shares of dry cowpea meal were calculated in Chapter 5. In making
these calculations, the following shares of raw materials were used for Weanimix: maize
75%, cowpeas/soybeans 15%, and groundnuts 10%. Also, due to the lack of information,
the waste rate during processing (in terms of weight) was assumed to be the same across
maize, cowpeas, soybeans, and groundnuts. In other words, input—output ratio was
assumed to be the same across these four raw materials.
Finally, retail prices were estimated by adding representative retail margins

reported in Section 6.2.5.1 to the estimated processor price of each of the target weaning

foods.

7.1.2 Estimated Prices of M- regared Tom Brown and Weanimix
Compared to industrial processors, grain/flour-type product vendors most likely
pay for a smaller number of cost components to produce Tom Brown or Weanimix. Their

major cost components would be: (1) raw materials; (2) fuel for roasting; and (3) charge

188

for custom milling. Returns to the vendors are equal to the difference between the selling
price and the sum of these costs, and include the returns to: (1) their labor; (2) investment
in buying roasting pans, which were most likely a small amount (see Section 6.2.3.3); (3)
transportation, if they carried the product from the place of production to the market; (4)
polyethylene bags to contain the product; and (5) the pro-rata share of the product in the
membership fee required to do business in the market, if they pay such a fee.

Since data on the quantity of Tom Brown produced were not collected from the
vendor respondents ‘ , budgets for producing Tom Brown were constructed using
representative values for each cost variable, as follows:

(1) cost of maize grain to produce 1 kg of output was estimated by dividing the
representative price per kg of maize (derived in Section 6.2.6) by the
representative input-output ratio (mean of input-output ratios obtained from
two industrial processor respondents);

(2) cost of fuel for roasting maize to produce 1 kg of output was estimated by
dividing the average payment for fuel for roasting 1 kg of maize (calculated
using data reported by five grain/flour-type product vendor respondents who
roasted maize by themselves) by the representative input-output ratio; and

(3) cost of custom milling to produce 1 kg of output was estimated by dividing the
representative charge for custom-milling 1 kg of maize2 by the representative
input-output ratio.

lAlthough it was found during the pre-testing of the questionnaire that grain/flour-type product vendors
self-prepared Tom Brown, it would have been too time-consuming to ask them both the original questions
designed for them (as sellers of the product) and a similar set of questions asked to industrial processor
respondents (as processors of the product). Therefore, as processors of Tom Brown, grain/flour—type
roduct vendors were only asked about the costs of raw material (i.e., maize), roasting, and milling.
Calculated using the representative charge for custom-milling l olonka of maize (derived in Section 6.2.4)
and representative olonka-kg conversion rate of maize (see Appendix 2).

189

The representative return was calculated by subtracting the sum of these costs
from the mean price (per kg) of Tom Brown reported by the vendor respondents.

Finally, representative prices of cowpea- and soybean-Weanimix were estimated
by replacing the cost of maize in the budget for producing Tom Brown with the cost of
maize, cowpeas/soybeans, and groundnuts3, holding the values of other variables (i.e.,

cost of fuel, custom milling, and returns) constant.

7.2 Budgeting Analysis—Results

7.2.1 Processor Price Estimates of Tom Brown, Cowpea- Weanimix, and
Soybean-Weanimix

Derived budgets for industrially producing 1 kg of Tom Brown, cowpea-
Weanimix, and soybean-Weanimix and estimated processor prices of these products are
summarized in Table 7.1. Note that not all of the 10 respondents could provide enough
information to estimate their payment for every cost component. As a result, the number
of observations for each cost component varied from two to eight.

There were only two respondents for whom the cost of raw materials could be
calculated. This is because the other respondents could not recall the information required
to estimate input-output ratio or the information that they provided was apparently
erroneous4. Raw materials were a major cost component for all three products, with a
mean of ¢3,876 per kg of Tom Brown, ¢4,956 per kg of cowpea-Weanimix, and ¢4,683

per kg of soybean-Weanimix. Note that the prices of raw materials were not collected

 

3 Cost of maize, cowpeas/soybeans, and groundnuts to produce 1 kg of Weanimix were estimated by {0.75
X [representative price per kg of maize] + 0.15 X [representative price per kg of COWpeas/soybeans] + 0.10
X [representative price per kg of groundnuts]} divided by [representative input-output ratio].

4 Several respondents reported the weight of output that was higher than the weight of inputs. This does not
make sense because there should be waste during the processing. Inputs other than the main ingredients,
such as spices, were not included in the calculation. However, such ingredients would not have contributed
much to the weight of the output.

190

Table 7.1 Budgets by cost component for industrially produced 1 kg of
Tom Brown, COWpea-Weanimix, and soybean-Weanimix

 

 

 

 

 

 

 

 

 

 

 

 

Number Mean“ Standard
of obser— Min. (¢) Max. (¢) deviation

vations (¢) J)
TB 5 10,878 14,295 16,332 2,631
Processor price c-WM 5 11,957 15,375 17,436 2,632
s—WM 5 11,685 15,102 17,158 2,63]
TB 2 3,789 3,876 3,962 123
Raw material c-WM 2 4,845 4,956 5,067 157
s-WM 2 4,578 4,683 4,788 149
Wage 6 0 2,644 6,044 2,273
Equipment 5 84 637 1,338 457
Electricity 7 0 73 491 176
Fuel (excl. fuel for vehicle) 5 514 736 1,080 261
Water 8 0 25 78 35
Rent 8 0 13 413 145
Transportation (incl. fuel for vehicle) 3 712 2,599 3,739 1,646
Printing & Stationery 6 487 1,706 3,889 1,413
Telecommunication 4 144 383 855 321
Packaging material 7 367 767 4,737 1,535
Miscellaneous 5 13 5 438 1,036 362
TB 5 10,41 1 14,645 19,656 4,031
Total cost c-WM 5 1 1,491 15,724 20,761 4,045
s-WM 5 11,219 15,452 20,482 4,041
Return 5 -4,461 -350 5,522 3,907

 

 

* Mean values were calculated excluding outliers (defined as values outside of the mean
of all observations 1 2 standard deviations). The maximum value of the following
cost variables were found to be outliers, and therefore excluded from the mean
calculation: electricity, rent, and packaging materials.

Notes: TB: Tom Brown; c-WM: cowpea-Weanimix; s-WM: soybean-Weanimix; the sum
of cost components in each column does not equal the total cost in the same column
because the figures for each cost component, as well as total cost in the table, were
based on data obtained from a different number of respondents.

Source: Fieldwork in Accra, February and March 2007.

from each respondent, but the representative prices (derived in Section 6.2.6) were used
across the respondents.

Wage was a cost component whose value varied widely across the respondents,
ranging from ¢0 to ¢6,044 per kg of output, with a mean of 92,644 per kg. The
respondent who reported zero wage was the smallest processor who did not have any
employees.

Transportation cost also varied across the respondents, ranging from ¢712

191

to ¢3,739 per kg of output. In fact, there were respondents who reported zero
transportation costs. These respondents had intermediaries who came to their facility to
sell raw materials and to buy outputs. Since we did not have an opportunity to interview
intermediaries, for those respondents paying zero transportation costs, the mean value for
transportation cost (¢2,599) was added to their processor price, so that the estimated
processor prices became prices at the outlet level across the respondents.

The importance of printing and stationery also varied across the respondents,
ranging from ¢487 to ¢3,889 per kg of output, with a mean of ¢1,706 per kg.

Finally, packaging material costs also varied across the respondents, ranging
from ¢367 to ¢4,737 per kg of output. However, the maximum payment of ¢4,737 was an
outlier. The respondent used both a polyethylene bag and paper box to package the
product, while many other respondents used only a polyethylene bag. The mean cost of
packaging material (excluding this outlier) was ¢767 per kg.

Equipment, electricity, fuel for processing, water, rent, telecommunication, and
miscellaneous costs were minor cost components.

Due to missing information, the total cost of production could be estimated for
only 5 of the 10 respondents. Because of the wide variation in the payment for each cost
component across the respondents, total cost also varied widely, ranging from ¢10,411
to ¢19,656 per kg of Tom Brown, ¢11,491 to ¢20,76l per kg of COWpea-Weanimix,
and ¢11,219 to ¢20,482 per kg of soybean-Weanimix. Since it was expected that raw
material cost would be the only cost component that would change depending on the
choice of output (i.e., Tom Brown, cowpea-Weanimix, or soybean-Weanimix), the

difference in total cost was due to differences in raw material costs. The results indicate

192

that 1 kg of COWpea-Weanimix would be about ¢1,100 more expensive to produce than
Tom Brown and ¢300 more expensive to produce than soybean-Weanimix.

Returns ranged from ¢-4,461 to ¢5,522 per kg of output. As explained in Section
7.1.1, these are estimated returns to the processors for their current products, which are
similar to or the same as our target products. Therefore, the result indicates that some of
the respondents earned negative returns for their currently-produced weaning foods—two
of the five respondents had negative returns. There could be at least three potential
explanations for this outcome: (1) the respondents were actually losing money by
producing those weaning foods, but the returns of other products were large enough to
cover the loss. The respondents either continued producing those weaning foods without
noticing that the products were not profitable, or knew that the products were not
profitable but did not stop production for some reasons (for example, those non-profitable
weaning foods might be a “loss-leader,” promoting customers to buy other more
profitable products of the company); (2) the respondents were actually earning positive
returns. However, the approximated shares of those weaning foods in each cost
component in the budget calculation were not close enough to the real values, which led
to inaccurate results; and (3) due to seasonal variations in costs, the respondents were
losing money during the month for which the data were collected, but the respondents
earned positive returns in other months. Since the analysis of this study was conducted
using monthly data (collected for only one month of the year, rather than a series), it
could not capture the seasonal variations in costs.

Processor prices were estimated to be ¢10,878 to ¢16,332 per kg of Tom

Brown, ¢11,957 to ¢17,436 per kg of cowpea-Weanimix, and ¢11,685 to ¢17,158 per kg

193

 

of soybean-Weanimix. From the structure of the model used (i.e., assumption of fixed
returns, rather than returns based on a relative [percentage] processing margin to the cost
of production), the difference in estimated processor prices among the three products
equals the difference in total costs among the products. Therefore, the processor price of
COWpea-Weanimix was estimated to be about ¢1,100 higher than the processor price of

Tom Brown and ¢300 higher than the processor price of soybean-Weanimix.

7.2.2 Representative Budgets and Retail Price Estimates of Tom yam
C owpea- Weanimix, and Soybean- Wean imix

Of the 10 industrial processor respondents, 5 could provide information with a
relatively few number of missing and apparently erroneous values across the cost
componentss. Using the data obtained from these five respondents, two representative
budgets for producing 1 kg of Tom Brown, cowpea-Weanimix, and soybean-Weanimix
were constructed as follows: first, two out of the five respondents with the lowest values
of returns were selected, and the mean value was calculated for each component of their
budgets (this new budget is hereafter called the “low profit representative budget” [LB])6.
Second, the two respondents with the highest values of returns were selected, and the

mean values were calculated in the same way (this budget is hereafter called the “high

 

5 Details of the problems encountered analyzing the data collected from these five respondents and the
methods used to handle those problems are available from the author upon request. Some of them are
geported in the rest of this chapter as well as in Appendix 4.

When there were missing values for one of the two respondents, the value from the other respondent’s
budget was used in the representative budget.

194

profit representative budget” [HB]) 7.

Assuming that industrially-processed Tom Brown and Weanimix were sold in
small shops rather than supermarkets, a retail margin of 29% (VAT/NHIS included) was
added to the estimated processor price in the LB and HB. The results are shown in Table
7.2, along with the representative budget for grain/flour-type product vendors to prepare
Tom Brown, cowpea-Weanimix, and soybean-Weanimix, which were prepared as
described in Section 7.1.2.

The estimated retail prices of weaning foods self-prepared by grain/flour-type
product vendors were ¢14,295 per kg of Tom Brown, ¢15,375 per kg of COWpea-
Weanimix, and ¢15,103 per kg of soybean-Weanimix. These prices were derived as the
sum of raw material costs, cost of custom milling (¢946 per kg of output), fuel for
roasting (¢37O per kg of output), and returns to labor and other costs not included
elsewhere (¢9,106 per kg of output).

On the other hand, the retail prices of industrially-processed weaning foods sold
at small shops were estimated to be ¢17,488 to ¢18,341 per kg of Tom Brown, ¢18,870
to ¢19,753 per kg of cowpea-Weanimix, and ¢18,521 to ¢19,397 per kg of soybean-

Weanimix.

 

7 One might be skeptical about this method of constructing representative budgets, considering that the
value of returns would not be directly related to the level of each cost component. For example, a processor
paying a higher total cost than the others can set a very high processor price. Then, if the value of this
processor’s returns is among the top two, she will be considered as a HB processor. If the other processor
selected for the HB pays a lower total cost than the others, could the mean values of each cost component
of these two processors be called “representative?” Therefore, to examine the viability of the method used,
the values of each cost component were compared across the respondents. The result showed that the
values of wage and transportation costs of the two processors selected for the LB were relatively close to
each other, compared to the values of these costs of the other three respondents. In the same way, the values
of these cost variables of the two processors selected for the HB were relatively close to each other
compared to the values of the other three respondents. Therefore, the method used developed two
reasonable representative budgets by taking account of the widely varying wage and transportation costs,
which were major determinants of the variation among the total budgets.

195

Table 7.2 Representative budgets to produce 1 kg of Tom Brown, COWpea-
Weanimix, and soybean-Weanimix and estimated retail prices

 

 

 

 

 

 

 

 

 

 

 

 

Grain/flour- Industrial Industrial

ty e product rocessors — rocessors —

Processor (type of budget) zendors p LB p HB

(¢) (¢) (¢)
TB 14,295 18,341 17,488
Retail price c-WM 15,375 19,753 18,870
s-WM 15,103 19,397 18,521
TB 4,165 3,971
Retail margin (29%) c-WM 4,486 4,285
s-WM 4,405 4,206
TB 14,176 13,517
Processor price c-WM 15,268 14,585
s-WM 14,992 14,315
TB 3.874 3,918 3,831
Raw material c-WM 4,953 5,010 4,899
s-WM 4,681 4,735 4,630
Wage 0 4,985 1,281
Equipment“ 946 904 402
Electricity 0 347 1 16
Fuel (excl. fuel for vehicle) 370 772 545
Water 0 25 73
Rent 0 47 206
Transportation (inc1. fuel for vehicle) 0 3,542 712
Printing & Stationery 0 2,188 1,815
Telecommunication 0 545 220
Packaging material" 0 368 733
Miscellaneous 0 426 586
TB 5,190 18,068 10,522
Total cost c-WM 6,269 19,160 11,590
s-WM 5,997 18,885 11,321
Return 9,106 -3,892 2,995

 

 

Note: LB: low profit representative budget; 118: high profit representative budget; TB:
Tom Brown; c-WM: COWpea-Weanimix; s-WM; soybean-Weanimix.

* Equipment cost for grain/flour-type product vendors is the cost of custom milling.

** The cost of packaging material was missing for both of the LB respondents. An
assumed value of ¢368 was derived using the data provided by other respondents.

Source: Fieldwork in Accra, February and March 2007.

The difference between the estimated retail prices of industrially-processed
cowpea-Weanimix and soybean-Weanimix was only about ¢350 per kg. This result
suggests that, with the current difference between prices of COWpea and soybean grains,
cowpeas are price-competitive with soybeans as an ingredient in Weanimix. Since ¢350
per kg is a minor difference, the two products could easily be sold at the same price,

depending on the pricing practice used by processors and/or retailers. Then, the choice by

196

customers would entirely depend on their preference between cowpeas and soybeans.

The difference between the estimated retail prices of industrially-processed
cowpea-Weanimix and Tom Brown was about ¢1,400 per kgg—cowpea-Weanimix was
estimated to be about 8% more expensive than Tom Brown. Whether a weaning mother
would be willing to pay this premium would depend on both her budget for weaning
foods and her awareness of the difference in nutritional value of these two products, as
well as how much more she would be willing to pay for the difference in nutrients that
she is aware of.

The difference between the retail prices of industrially-processed weaning foods
and those self-prepared by grain/flour-type product vendors were estimated to be
about ¢3,2OO to ¢4,400 per kg. In other words, the industrially—processed products were
estimated to be about 22% to 28% more expensive than the self-prepared ones. The
sources of the difference in these prices are, as shown in Table 7.2, the difference in the
cost of production and the retail margin that applied only to industrially-processed
weaning foods.

The estimated returns to grain/flour-type product vendors were much higher than

the estimated returns to industrial processors. From the data available, it is not clear

 

3 This difference is much smaller than the difference observed during the fieldwork between the prices of
Tom Brown and Weanimix/quasi-Weanimix sold at supermarkets (about ¢6a900; see Table 6-1)- The
possible reasons for this discrepancy include: (1) retail margins at supermarkets were generally higher than
the retail margins at small shops. Therefore, the difference in retail prices of two products becomes larger at
supermarkets than at small shops, if retailers charge pr0portiona1 margins to the purchase prices, as
assumed in this study; (2) many industrial processors used a higher share of fortifier ingredients (i.e.,
COWpeas/soybeans and groundnuts) in their weaning food products than the standard share (see Section
6.2.3.2). Therefore, the difference in the costs of raw materials between Tom Brown and their
Weanimix/quasi-Weanimix was larger than between Tom Brown and standard Weanimix; and (3) industrial
processors might set the price of Tom Brown and Weanimix in a way that the price difference is larger than
the difference in the cost of production. This would make sense if industrial processors assume that the
customers for Weanimix are higher income and willing to pay more for a more nutritionally rich food for
their children (i.e., the sellers would be practicing price discrimination based on relative elasticities of
demand for closely related products).

197

whether there are other major cost components that vendors actually paid but were
omitted from the budget, or if vendors in fact charged a much higher processing margin
than industrial processors.

In terms of product quality, the difference between industrially-processed weaning
foods and those self-prepared by grain/flour-type product vendors would likely be
negligible, because no major difference in processing procedures was found. The only
major difference, other than prices, between these weaning foods would be that the
industrially-processed products are sold in a sealed polyethylene bag, with a label
displaying the Ghana Standards Board certified logo9, while the self-prepared products
are sold under open—air conditions, without a GSB logo. Because the sealed bag and the
label serve as a signal to consumers about cleanliness and consistency/quality of the
product, some consumers may prefer to pay more for the industrially-processed weaning
foods to get these implicit quality guarantees. Therefore, if a mother who currently buys
Tom Brown in the market values a better package with quality certification more than the
difference in prices, she is expected to buy industrially-processed weaning foods at small
shops.

Masters and Sanogo (2002) studied weaning mothers’ willingness-to—pay a
premium in a similar setting in Bamako, Mali. Their experiments found that respondents
were willing to pay about US$1.75 per kg more (on average) for a sealed-and-authority-

certified locally-produced weaning food sample than for an open-bagged-and-non-labeled

 

9 Note that not all the weaning foods produced by the respondents were GSB certified. However, the fees
for the GSB certification and the Food and Drugs Board’s registration were included in miscellaneous costs
for all the respondents, whether or not the respondents mentioned these costs during the interview (see
Appendix 4). Therefore, the estimated prices of industrially-processed weaning foods derived in this study
reflect these costs, and all the products are assumed to be GSB certified and FDB registered. (Note that
miscellaneous costs, which included the fees for the GSB certification and the FDB registration, were a
minor cost component in the budget [Tables 7.1 and 7.2]. Therefore, the assumption made here did not
affect the results of budgeting analysis very much.)

198

locally-produced weaning food sample. The value of US$1.75 in 1999 (when the
experiment seems to have been conducted) is equivalent to about ¢24,000 in February
2007”). This value is more than five times the difference in the estimated prices of
industrially-processed and self-prepared weaning foods in our budgets (i.e., ¢3,2OO
to ¢4,400 per kg). While the results of Masters and Sanogo’s study are not directly
comparable to our study, their findings suggest the need to conduct a similar consumer
survey in Ghana. If a similarly high willingness-to-pay for a better package and quality
certification is found among Ghanaian mothers, it would indicate that there is a high
potential for industrial processors to expand their sales of weaning foods through small
shop channels, which they currently do not use extensively.

Table 7.2 also shows that grain/flour-type product vendors could produce Tom
Brown, cowpea-Weanimix, and soybean-Weanimix for costs relatively similar to each
other. Since vendors prepare a large amount of these products, compared to households,
separate roasting of three major ingredients to prepare Weanimix would cause little extra
work compared to the preparation of Tom Brown. Therefore, vendors could likely sell
these three products at prices similar to each other, as shown in Table 7.2. However,
during the fieldwork, almost no vendors could be found who sold either self-prepared
COWpea- or soybean-Weanimix. Since no major constraints were found on the supply side,
the reason for the lack of Weanimix in the market is suspected to be found on the demand
side. A hypothesis is that the majority of weaning mothers who purchase Tom Brown in

the market either do not know about Weanimix or cannot tell from its appearance whether

 

‘0 The original figure in the Malian currency was FCFA455.24/400g (Masters and Sanogo [2002], Table 2).
This value was first converted to the value per kg, and then inflation-adjusted by using the CPI (obtained
from the IMF website). Finally, it was converted to cedis by using the average exchange rate between
FCFA and cedis in February 2007 (obtained from the website of OANDA at
h_ttp://www.oanda.com/convert/thistory).

199

the product is truly Weanimix (i.e., fortified Tom Brown), rather than plain Tom Brown.
Therefore, they would not pay a higher price for Weanimix, compared to Tom Brown,

which discourages vendors to prepare and sell Weanimix.

7.3 Sensitivity Analysis

Sensitivity analyses were conducted for different scenarios that would potentially
change the price-competitiveness of industrially-processed COWpea—Weanimix with: (1)
industrially-processed soybean-Weanimix; (2) industrially-processed Tom Brown; and (3)
COWpea-Weanimix self-prepared by grain/flour-type product vendors. The analyzed
scenarios are: (1) change in the technical efficiency of industrial processing; (2) change in
the volume of industrial production; (3) change in the price of raw materials (seasonality
analysis); (4) change in the retail margin; and (5) a combination of different scenarios.

Throughout these sensitivity analyses, different assumptions were made for
processors with HB (i.e., high profit representative budgets) and those with LB (i.e., low
profit representative budgets) about their ways to modify the prices of their products.
Whenever the cost of production changed, HB processors were assumed to modify the
price so the returns remained constant, except for a special assumption made for
scenarios (2) and (5). On the other hand, for LB processors, it was assumed that: (1)
when the modified total cost is larger than the total cost in the original budget, they
increase the price so the value of returns remains unchanged; and (2) when the modified
total cost is smaller than the total cost in the original budget, they do not change the price
until the returns reach zero, and change the price only when the returns would potentially

turn positive, by subtracting the value of potential returns from the price so the returns

200

 

remains zero. While these assumptions were made to simplify the analysis, assuming
constant returns is the equivalent of assuming a perfectly competitive market for the

products analyzed. The appropriateness of this assumption is discussed in Section 8.3.

7.3.1 Changein the Technical Etliciencv of Industrial Processing

The input-output ratio of HB processors (calculated from the original data) was
0.78, while that of LB processors was 0.77. The first sensitivity analysis was conducted
on the effects of a change in the input-output ratio of industrial processing on the
estimated retail pricesl 1.

First, the estimated prices of weaning foods produced by LB processors did not
decrease because the returns did not reach zero, even with the input-output ratio of 1.00.
The relationship between different input-output ratios and the estimated retail prices of
weaning foods produced by HB processors are plotted in Figure 7.1. For comparison,
Figure 7.1 also plots the estimated retail prices of weaning foods self-prepared by
grain/flour-type product vendors, for whom the input-output ratios were assumed not to
change.

The difference in the prices between industrially-processed cowpea-Weanimix
and industrially-processed Tom Brown or soybean-Weanimix slightly decreased as the
input-output ratio improved.

The difference in the prices between industrially-processed and self-prepared

cowpea-Weanimix decreased from about ¢3,500 per kg in the original budgets to

about ¢3,400 per kg, assuming an input-output ratio of 0.80; and about ¢3,000 per kg,

 

” Note that due to the lack of information, the waste rate (or [1 — input-output ratio]) was assumed to be the
same across different raw materials (i.e., maize, cowpeas/soybeans, and groundnuts).

201

Figure 7.1 Estimated retail prices of industrially-processed (with HB
budget) and self-prepared weaning foods for different levels of
input-output ratio

 

 

 

 

 

 

 

20,000 17 77-7 777~*——~——- -——-—-——— -——*_—-----__-.-.----2 _H ,
t I
5,0 19,000 ~ ‘“:\\:':\\ 1
\m w 1
{5 18,0001 ~ 1 1
3’1 .
0 17,000 ‘ i 1
2
3' 16.000 1 1
as ---------------------- _
935,000“ _7 ''''' _—-—---—_.,-,—-_ 1
14,000 T ,— fi-
0.7 0.75 0.8 0.85

 

 

 

 

 

. input-output ratio

 

1 _o— TB (118) ——-— c-WM (HB) —+— s-WM (HB)
TB(G) - - - -c-WM(G) -- - -s-WM(G) 1

 

 

 

 

Note: TB: Tom Brown; c-WM: cowpea-Weanimix; s-WM: soybean-Weanimix; HB: high
profit representative budget; G: grain/flour-type product vendors
Source: Calculated by the author.

assuming an input-output ratio of 0.85. As reported in Section 4.2.6.2, the input-output
ratio of dry cowpea meal prepared in the laboratory was 0.847. Assuming that a ratio of
similar magnitude is the maximum value for industrially-processed cowpea-Weanimix, it
could be concluded that the increase in the technical efficiency of industrial processing

would certainly improve the price-competitiveness of the product, but not dramatically.

7.3.2 Change in the Volume of Industrial Production
The second sensitivity analysis was conducted with regard to the change in the
volume of production of industrial processors. First, the same method used in Section

5.3.2 was used to estimate how much each of the LB and HB processors could increase

202

their volume of production without requiring an increase in fixed costs”. Second, the
budget of each respondent was recalculated using the derived multipliers. As done in
Section 5.3.2, it was assumed that there is no change in the average variable costs (i.e.,
assumption of constant returns to scale). Finally, the LB and HB were reconstructed with
the same combinations of respondents that were selected to construct the original LB and
HB.

To estimate the processor prices for HB, the two alternative assumptions made in
Section 5.3.2 were applied. When the unit cost of production declined, HB processors
were assumed to modify the selling price of their products: (1) so the per-unit returns
remain constant; or (2) so the per-unit returns become one-half of the original value (to
make their products more competitive with other companies’ products, while increasing
their total returns due to the increase in the volume of production). It was assumed that
the demand for the products is price-elastic enough for the increased outputs to be sold
out, whether the processors kept the same per-unit processing margin (i.e., per-unit
returns) or cut it to one-half of the original value (while increasing the total returns). The
results are shown in Table 7.3.

As a result of increased production, it was estimated that LB processors could
save a total of ¢4,905 per kg of output, due to the decline in fixed costs per unit of output.
Under the assumption noted previously about the way that LB processors modify the

price of their products, ¢3,892 were deducted to make the returns zero, and the processor

 

'2 Based on the highest volume of production per worker per hour estimated for one of the HB processors,
each LB processor was estimated to be able to produce at least 4.3 and 4.8 times more outputs, while the
other HB processor was estimated to be able to produce at least 2.5 times more outputs. In Section 5.3.2,
the respondent with the highest estimated volume of production per worker per hour was assumed to be
able to produce 10% more outputs. In this section, this assumption was not made (i.e., the budget of the HB
processor with the highest volume of production per worker per hour remained unchanged).

203

Table 7.3 Increase in the volume of production and change in the budgets
of industrially-processed weaning foods

 

Grain/flour-

Industrial

Industrial processors - HB

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fine-1)
Processor (type of budget) tyEIZESZSSUCt processors origiitrhal W(i)trliigli {/11 1of
(it/kg) (¢/kg) returns returns
Retail price TB 14,295 17,031 15,912 13,974
(retail margin for c-WM 15,375 18,444 17,293 15,356
industrially-processed s-WM 15,103 18,088 16,945 15,008
mducts: 29%) @310) (-1,577) (-3,514)
TB 13,163 12,298 10,801
Processor price c-WM "14,255 13,366 11,869
s-WM 13,980 13,097 11,600
(-1,012) (-1,219) (-2,7l6)
Wage 1,086 642
(-3,899) (-640)
Equipment 689 347
(-215) (-55)
Rent 10 83
(-3 7) (~124)
Telecommunication 124 131
(-421) (-89)
Miscellaneous 94 275
(-332) (-31 1)
TB 11,160 7,826
Other costs c-WM 12,252 8,894
s-WM 11,977 8,625
TB 5,190 13,163 9,304
Total cost c-WM 6,269 14,255 10,371
s-WM 5,997 13,980 10,102
(-4,905) (-1,219)
Return 9,106 0 2,995 1,497
(3,892) (0) (-l,497)

 

 

Note: the numbers in parentheses are the difference between the original values (in Table
7.2) and the values reported in Table 7.3; LB: low profit representative budget; HB:
high profit representative budget; TB: Tom Brown; c-WM: cowpea-Weanimix; 5-
WM: soybean-Weanimix.

Source: Calculated by the author.

price decreased by the remaining ¢1,012. As a result, the retail price was estimated to

decrease by ¢1,3 10 per kg for all three products.

It was estimated that HB processors could save a total of ¢1,219 per kg of output.

Under the assumption that the processors set the price so the per-unit returns remain

constant, the processor price was estimated to decrease by the same amount, and the retail

price was estimated to decrease by ¢1,577 per kg for all three products. Under the

204

 

assumption that the processors set the price so the per—unit returns become one-half of the
original value, the processor price was estimated to decrease by ¢2,716 per kg, and the
retail price was estimated to decrease by ¢3,514 per kg for all three products.

The difference between the estimated retail prices of industrially-processed and
self-prepared cowpea-Weanimix shrank from about ¢4,400 to ¢3,100 per kg for the LB.
Under the assumption that HB processors set the price so the per-unit returns remain
constant, the difference shrank from about ¢3,500 to ¢1,900 per kg. Again, whether these
new prices would be accepted by customers would depend on the premium that weaning
mothers would be willing to pay for the better package and quality certification that come
with industrially-processed products. Under the assumption that HB processors set the
price so the per-unit returns become one-half of the original value, the retail price of their
cowpea-Weanimix was estimated to be slightly lower (by 9519) than the retail price of
self-prepared cowpea-Weanimix. This result indicates that the increase in the volume of
production by industrial processors, associated with a decrease in the processing margin,
could make the price of their cowpea-Weanimix low enough to be competitive with self-

prepared cowpea-Weanimix.

7.3.3 Change in the Prices ol Raw Algerials

Third, sensitivity analysis with regard to the change in the prices of raw materials
was conducted to examine the change in the price-competitiveness of industrially-
processed COWpea-Weanimix, compared to industrially-processed soybean-Weanimix,
industrially-processed Tom Brown, and COWpea-Weanimix self-prepared by grain/flour-

type product vendors.

205

7.3.3.1 Industrially-processed cowfipea- W )animix vs. ina’ustrialll-processed soybean-
Weanimix

Using equations (15) and (16), as well as the relationship between processor and
retail prices, the difference between the retail price of industrially-processed cowpea-

Weanimix and that of soybean-Weanimix can be expressed as follows:

Pc-WM _ Ps—WM

rem” retail = (1 + retail margin)(share of cowpeas/soybeans in raw materials)

1
input - output ratio

 

)(P

COW/780 _ PSOerean)

Substituting each term by the figures used in the budget calculation”, the equation

becomes:

—WM —WM 1
PC _ P5 = (l + 0.29)(0. ll 5)(6J_7)(P0011pea — Pso't.’bean)

retail retail

2 0'25(Pc0wpea '— P

soybean )

 

'3 Mathematical note: starting from the equation:

 

 

c-—WM s—WM - - c—WM s—WM
P retail - retail = (1 + I' etall mar glnXP processor — P processor )1
canceling out the same terms in the right hand side of the equation leads to the following expression:
c—WM /s—WM
c—WM s—WM __ . . _ cowpea/ soybean j!
retail _ retail *- (1+ reta”margmxPOOH-77661j! Psoybean jt) y
Q j,
where,
c~WM /s—WM
cow / b , ’t
p eogjoy can] (i.e., quantity of cowpea/soybean needed to produce 1 kg of c-WM/s-WM)
jt

= [share of cowpeas/soybeans in raw materials] X l / [input-output ratio].
As noted in Section 7.1.1, the input-output ratio was assumed to be the same for maize, COWpeas, soybeans,

and groundnuts.
'4 For the input-output ratio, the mean of the ratio of LB and HB was used.

206

The equation shows that the difference between the retail price of cowpea-
Weanimix and soybean-Weanimix is equal to the difference between the price of cowpea
and soybean grains multiplied by 0.25. Therefore, for example, if cowpeas are more
expensive than soybeans by ¢1,000 per kg, the retail price of cowpea-Weanimix would
be ¢250 higher than that of soybean-Weanimix. Even if cowpeas were more expensive
by ¢4,000 (an unrealistic assumption, because at that point, soybeans would be cheaper
than maize [see Section 6.2.6]), cowpea-Weanimix would only be ¢1,000 more expensive
than soybean-Weanimix. This indicates that cowpeas could be price-competitive with
soybeans, as an ingredient in Weanimix, as long as the price of cowpeas and soybeans

fluctuate within a reasonable range.

7. 3. 3.2 lndzistria[Iv-processed cowpca-Wmnimix vs. industrial]y-processea’ Tom Brown
Applying the same mathematical procedure used in the previous section, the
difference between the retail prices of industrially-processed cowpea-Weanimix and

industrially-processed Tom Brown can be expressed as follows:

 

_ 1
PC—WM — PTB . = 1+ retail mar in
retail retail ( g )(input - output ratio
{(share of maize in raw materials of c - WM — 1) Pmm-ze (18)

+ (share of cowpeas in raw materials of c - WM)PCO,,.pea

+ (share of groundnuts in raw materials of c - WM)Pgmundnm}

Substituting each term by the figures used in the budget calculation”, the equation

becomes as follows:

 

'5 For the input-output ratio, the mean of the ratio of LB and HB was used.

207

c—WM _ TB _ 1 _
P P —(l+0.29)(—0'77)( 0.25Pm

retail retail

+ 0.15wapea + 0.1P

aize groundnut )

roundnut

= —0.42P +0.25P +0.17)"g

maize COW/360
The equation shows that the price-competitiveness of cowpea-Weanimix increases as the
price of maize increases, while its competitiveness decreases as the prices of cowpeas and
groundnuts increase.

Using the above equation and the estimated range in the price fluctuation of maize,
cowpeas, and groundnuts derived in Section 6.2.6 (Table 6.7), the difference in the retail
prices of industrially-processed cowpea-Weanimix and Tom Brown were estimated for

the most favorable and unfavorable cases for cowpea-Weanimix. The result is shown in

Table 7.4.

Table 7.4 Price fluctuations of raw materials and estimated difference in
retail prices of industrially-processed COWpea-Weanimix and Tom

 

 

Brown

Price of Price of Price of Diffeience m

Combination of prices of raw materials maize cowpeas g-nuts retai prices
(it/kg) (¢fkg) (st/kg) be“ °'WM and

TB (st/kg)

Most favorable for cowpea-Weanimix 3,809 5,155 6,240 743
Current 3,000 5,474 7,650 1,397
Most unfavorable for cowpea-Weanimix 2,541 6,221 8,486 1,915

 

 

 

 

 

 

 

Source: Calculated by the author.

The difference in the retail prices of industrially-processed cowpea-Weanimix and
Tom Brown was estimated to range from ¢743 to ¢1,915 per kg. This result indicates that
as long as the price of raw materials fluctuate within a reasonable range, the difference in
the cost of production between cowpea-Weanimix and Tom Brown does not change

dramatically, and therefore, their retail prices do not change significantly.

208

7.3.3.3 Industrially-processed vs. self-prepared cowpea- Weanimix

As reported in Section 6.2.] (Table 6.1), the price of Tom Brown self-prepared by
grain/flour-type product vendors varied over the year due to the seasonal change in the
cost of raw materials. In this section, the range of the seasonal change in the price of self-
prepared cowpea-Weanimix was estimated by modifying the original budget to reflect the
estimated changes in the cost of raw materials and returns”. Also, the retail price of
industrially-processed cowpea-Weanimix was re-estimated assuming that production
occurred when the prices of raw materials were the lowest during the year (i.e., using the
lowest estimated prices of maize, cowpeas, and groundnuts in Table 6.7). Then, the
derived retail prices of industrially-processed and self-prepared COWpea-Weanimix were
compared. The result is shoWn in Table 7.5.

The price of cowpea-Weanimix self-prepared by grain/flour-type product vendors
was estimated to range from ¢13,279 to ¢16,461 per kg during the year. On the other
hand, the price of industrially-processed cowpea-Weanimix was estimated to potentially
decline to ¢17,989 per kg, if produced by HB processors when the raw materials were the
cheapest (the estimated price for the LB did not change because the decrease in the cost
was not enough to offset the negative returns). This result means that even if industrial
processors could produce a large volume of cowpea-Weanimix during the season when
the prices of raw materials are the lowest and sell outputs throughout the year without
Changing the price, that price still could not fall below the highest price of cowpea-

Weanimix sold in the market.

 

'6 The mathematical procedure used is available from the author upon request.

209

Table 7.5 Estimated changes in the price of cowpea-Weanimix self-
prepared by grain/flour-type product vendors and the estimated
price of cowpea-Weanimix industrially-processed when the cost of
raw materials are the lowest

 

 

 

 

 

 

 

 

 

 

. Grain/flour-type

Processor Srgdld/cflegnilfi: product vendors Industrial

(type of budget) — Cheapest _ Most p rocess/(Lrs — “8
season (¢/kg) exp ensrve (¢ g)
season (¢/kg)

Retail price 13,279 16,461 17,989
(-2,096) (1,086) (-881)
Processor price 13,904
(-681)
Raw material 4,265 5,989 4,218
(-688) (1,036) (-681)
Other costs 1,316 1,316 6,691
Total cost 5,581 7,305 10,909
(-688) (1,036) (-681)
Return 7,698 9,155 2,995
(-1,408) (49) (O)

 

 

Note: the numbers in parentheses are the difference from the original values (in Table
7.2); HB: high profit representative budget.
Source: Calculated by the author.

7.3.4 Change in the Retail Margin

For all the analyses conducted so far, the retail margin has been fixed to 29%,
assuming that industrially-processed weaning foods were sold at small shops and the
VAT/NHIS was collected. The fourth sensitivity analysis was conducted with regard to
the effects of change in retail margins on the price-competitiveness of industrially-
processed cowpea-Weanimix.

As described in Section 6.2.5.], the following three different scenarios with
different ranges in retail margins were examined: (1) products sold at small shops that did
not collect the VAT/NHIS—with estimated retail margins of 3%-31%; (2) products sold
at small shOps that collected the VAT/NHIS—with retail margins of 19%-50%; and (3)

products sold at supermarkets that collected the VAT/NHIS—with retail margins of 38%-

210

57%.

A preliminary analysis found that the difference in the retail prices between
industrially—processed cowpea-Weanimix and industrially-processed Tom Brown or
soybean-Weanimix does not change much when retail margins change. With a retail
margin of 3%, the difference between cowpea-Weanimix and Tom Brown was
about 951,100 per kg (whether for LB or HB processors), while with a retail margin of
57%, the difference was about ¢1,700 per kg. Between cowpea- and soybean-Weanimix,
the difference was about ¢300 per kg with a retail margin of 3% and about ¢400 per kg
with a margin of 57%. This finding indicates that the level of retail margin has little
effects on the price-competitiveness among industrially-processed weaning foods. In
other words, the price-competitiveness of COWpea-Weanimix with Tom Brown or
soybean-Weanimix would not be greatly affected by the type of retailers where the
products are sold or whether or not the VAT/NHIS are collected”.

The analysis then focused on the price-competitiveness of industrially-processed
cowpea-Weanimix with cowpea-Weanimix self-prepared by grain/flour-type product
vendors. The relationship between the level of retail margins and estimated retail prices
were plotted in Figure 7.2 for each of cowpea-Weanimix produced by LB processors, HB
processors, and grain/flour-type product vendors.

The estimated retail price of cowpea-Weanimix processed by HB processors was
lower than the estimated price of self-prepared COWpea-Weanimix until the retail margin

reached 6%. In contrast, the estimated retail price of cowpea-Weanimix processed by LB

‘7 Holding the assumption that the same percentage mark-up would be charged across Tom Brown, cowpea-
Weanimix, and soybean-Weanimix. At least, two out of four supermarket respondents (i.e., current outlet of
industrially-processed weaning foods) indicated that they charged the same percentage mark-up for all
flour-type products.

211

Figure 7.2 Estimated retail prices of cowpea-Weanimix produced by LB
processors, HB processors, and grain/flour-type product vendors
for different levels of retail margins

 

 

 

 

 

 

 

 

 

 

 

_ 7
25,000 — —
, _ 1
3’ 20,000 1 1
.2 1
8 15.000 ~ 1
.9.
.8 l
3 10,000 « 1
'_§ 55 w/o 55 w/ sm w/ 1
a 5.000 - A 1
f \A I
r I I \ 1
1 O 7 T T I r i
0.00 0.10 0.20 0.30 0.40 0.50 0.60
1 retail margin
1 1 LB——MB----o1

 

 

 

Note: LB: low profit representative budget; HB: high profit representative budget; G:
grain/flour-type product vendors; ss w/o: small shops without collection of the
VAT/NHIS; 55 w/: small shops with collection of the VAT/NHIS; sm w/:
supermarkets with collection of the VAT/NHIS.

Source: Calculated by the author.

processors was always higher than the estimated price of the self-prepared product. The
difference in prices between cowpea-Weanimix processed by HB processors and the self-
prepared product was estimated to be about 933,700 per kg for a retail margin of 31%
(upper bound for small shops without the VAT/NHIS); about ¢6,500 per kg for a retail
margin of 50% (upper bound for small shops with the VAT/NHIS); and about ¢7,500 per
kg for a retail margin of 57% (upper bound for supermarkets with the VAT/N HIS).

The results indicate that the level of the retail margin greatly affects the price-
competitiveness of industrially-processed cowpea-Weanimix with self-prepared cowpea-
Weanimix. This implies that the outlet for industrially-processed COWpea-Weanimix

would need to be expanded to small shops, rather than continuing to be sold almost

212

exclusively at supermarkets, if the products are to be patronized by weaning mothers

from lower-income families.

7. 3.5 Combination at Different Scenarios
Finally, sensitivity analysis was conducted with regard to the combinations of four
different scenarios that were examined above (i.e., change in technical efficiency, volume

of production, price of raw materials, and retail margin).

7.3.5.1Industrially-processed cowpea-Weanimix vs. industrially-processed soybean-
Weanimix

As equation (17) indicates, the difference in retail prices between industrially-
processed cowpea-Weanimix and industrially-processed soybean-Weanimix becomes
smaller when: (1) technical efficiency improves (i.e., input-output ratio becomes larger);
(2) the difference in prices of cowpea and soybean grains becomes smaller; and (3) retail
margin becomes smaller”. For each of these variables, the minimum and maximum
values were selected from the data presented in the previous sections. Using these values,
the smallest and largest differences in retail prices of cowpea- and soybean-Weanimix

were estimated. The result is shown in Table 7.6.

18 - . . . . . .
. As equation (17) indicates, the volume of production does not affect the price-competitiveness of
industrially-processed cowpea-Weanimix with industrially-processed soybean-Weanimix.

213

Table 7.6 Estimated smallest and largest differences in retail prices of
industrially-processed cowpea- and soybean-Weanimix

 

 

 

 

 

 

Price- Price of raw materials Difference
_ * . .
competitiveness (1)135“; (¢fkg)* Retail m $565 (sf
of c-WM with .p ,1, margin*** 0 an
ratio Cowpeas Soybeans s-WM
s-WM
(it/kg)
Highest 0.85 5,155 4,548 0.13 120
Lowest 0.70 6,221 3,889 0.47 737

 

 

 

 

 

 

* For the reason to select 0.85 as the maximum value, see Section 7.3.1. The minimum
value of 0.70 was experimentally selected (as opposed to 0.77 [the mean input-
output ratio of the processor respondents]).

** From Table 6.7. These combinations assume that the price of cowpeas could become
the highest when the price of soybeans is the lowest, and vice-versa.

*** The minimum value of 0.13 is the mean retail margin of small shop respondents
without collection of the VAT/NHIS (see Table 4.12); the maximum value of 0.47 is
the mean retail margin of supermarket respondents with collection of the VAT/NHIS
(see Table 4.13).

Source: Calculated by the author.

Under the most unfavorable scenario for cowpea-Weanimix, the difference in
retail prices of industrially-processed COWpea- and soybean-Weanimix was estimated to
be only ¢737 per kg. This result indicates that cowpea-Weanimix could probably always
be price-competitive with soybean-Weanimix, unless customers prefer soybeans and are
willing to pay a high premium for soybean-Weanimix.

A key factor that affects customers’ willingness to pay a premium for COWpea- or
soybean-Weanimix would be the difference in the nutritional value between these two

19

products If soybean-Weanimix is nutritionally superior to cowpea-Weanimix,

customers might be willing to pay a premium for soybean~Weanimix.

 

'9 As long as the same ratio of ingredients are used (e.g., maize 75%, cowpeas/soybeans 15%, and
groundnuts 10%), the nutritional value of cowpea- and soybean-Weanimix are different, due to the
difference in the nutritional value of cowpeas and soybeans (as mentioned in Section 2.2.3, soybeans are
higher in protein and fat, but lower in carbohydrates than COWpeas). On the other hand, if cowpeas and
soybeans are used to produce Weanimixes that have the same protein content (e.g., the minimum protein
requirement for weaned children), the proportion of soybeans in soybean-Weanimix would be lower than
the proportion of COWpeas in cowpea-Weanimix, due to the higher protein content of soybeans than
COWpeas. In this case, soybean-Weanimix would be more price-competitive with cowpea-Weanimix than in
the results of the analysis reported in this chapter.

214

7. 3. 5. 2 Industriallvzgrocessed cowpea- Weanimix vs. industrially-processed Tom Brown
Using equation (18), a similar analysis was conducted to estimate the difference in
retail prices of industrially-processed cowpea-Weanimix and industrially-processed Tom

Brown. The result is shown in Table 7.7.

Table 7.7 Estimated smallest and largest differences in retail prices of
industrially-processed cowpea-Weanimix and Tom Brown

 

 

 

 

Price- . . Difference in
competitiveness :33”; Price Of raw materials (¢/kg) Retail prices of c-
of c-WM with P . margin WM and TB
ratio Maize Cowpeas Groundnuts
TB (Mg) 1
Highest 0.85 3,809 5,155 6,240 0.13 589
Lowest 0.70 2,541 6,221 8,486 0.47 2,415

 

 

 

 

 

 

 

 

Source: Calculated by the author.

Under the most favorable scenario, the difference in the retail prices between
industrially-processed cowpea-Weanimix and Tom Brown was estimated to be ¢589 per
kg. Under the most unfavorable scenario, the difference was estimated to be ¢2,415 per
kg. As mentioned in Section 6.2.3.2, the relative price of Weanimix and Tom Brown
varied among the processor respondents (the price of Weanimix was 1.2 to 1.9 times the
price of Tom Brown, except for one respondent who set the price of these products the
same). However, the result of the sensitivity analysis indicates that the relative price
between these two products could potentially be smaller than their relative prices set by
the majority of the processor respondents (if the price of Tom Brown is ¢17,488 per kg as
estimated for HB processors [Table 7.2], the difference of ¢2,415 means that cowpea-

Weanimix is 1.14 times more expensive than Tom Brown).

215

7.5.3.3 Industrially—processed vs. self-prevmzred cow’pea- Weanimix

Finally, using the combinations of most favorable and unfavorable scenarios for
cowpea-Weanimix 20, the lowest and highest retail prices of industrially-processed
cowpea-Weanimix were estimated. Then, the derived prices were compared with the

estimated lowest and highest prices of cowpea-Weanimix self-prepared by grain/flour-

type product vendors. The result is discussed below.

Table 7.8 Estimated lowest and highest retail prices of industrially-
processed cowpea-Weanimix

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Retail Bu d- lnput- Price of raw materials (¢/kg) Volume of Retail Retail

price et output Ground- production margin price

estimate g ratio Maize Cowpeas nuts (¢/kg)
Lowest LB Increased 14 773

. 4 . 3

estimate H8 0 85 2’54] 5’155 6’2 0 volume 0 13 12,213
Highest LB Current 24,895
1 estimate HB 0.70 3,809 6,22l 8,486 volume 0.47 24,052

 

 

 

 

Source: Calculated by the author.

As Table 7.8 shows, the lowest retail price of industrially-processed cowpea-
Weanimix was estimated to be around ¢12,000 per kg, while the highest price was
estimated to be around ¢25,000 per kg. On the other hand, as shown in Table 7.5, the
price of self-prepared cowpea-Weanimix was estimated to range from about ¢l3,000
to $016,000, depending on the season of the year. Therefore, the result indicates that, under

very favorable conditions, industrially-processed COWpea-Weanimix could become

cheaper than its self-prepared counterpart.

2° The combination of the most favorable scenarios consists of: (1) highest technical efficiency (i.e., highest
input-output ratio); (2) increased volume of production (for the HB, with the reduction of the processing
margin to one-half of the original value); (3) lowest prices of all ingredients (i.e., maize, COWpeas, and
groundnuts); and (4) lowest retail margin. Opposite is the combination of the most unfavorable scenarios.

216

7.4 Summary

In this chapter, enterprise budgets were constructed to estimate the retail prices of
Tom Brown, cowpea-Weanimix, and soybean-Weanimix, for two different categories of
processors and the corresponding outlet—(l) industrial processors whose products are
sold in small shops (potential outlet, as opposed to the current outlet of supermarkets);
and (2) grain/flour-type product vendors in the market.

Raw materials were estimated to be a major cost component of all three of the
products produced by industrial processors. For cost variables such as wage,
transportation, and printing and stationery, the estimated payment (per kg of output)
varied greatly across the respondents. Among 10 respondents, only one-half could
provide enough information for calculating the total cost of production and returns. Due
to the variation in the payment for each cost component, their returns also varied. For the
top two processors (in terms of the value of returns), the average retail price of their
products was estimated to be about ¢17,500 per kg of Tom Brown, about ¢18,900 per kg
of cowpea-Weanimix, and about ¢18,500 per kg of soybean-Weanimix, with an average
return to processors of about ¢3,000 per kg. On the other hand, for the bottom two
processors, the average retail price of their products was estimated to be about ¢18,300
per kg of Tom Brown, about ¢19,800 per kg of cowpea-Weanimix, and about ¢19,400 per
kg of soybean-Weanimix, with an average return to processors of about ¢-3,900 per kg.

Thus, the difference in retail prices of industrially-processed cowpea-Weanimix
and soybean—Weanimix was estimated to be only about ¢400 per kg. This is a minor
difference, which indicates that, under the conditions observed in Accra during February

and March 2007, cowpeas were competitive with soybeans as an ingredient in Weanimix,

217

if consumers valued the two products equally.

Also, the retail price of cowpea—Weanimix was estimated to be ¢l,400 per kg
higher than the retail price of Tom Brown. This finding indicates that the relative price
between industrially-processed cowpea-Weanimix and industrially-processed Tom Brown
could potentially be smaller than their relative prices set by the majority of the processor
respondents.

Estimated retail prices of weaning foods self-prepared by grain/flour-type product
vendors were about ¢14,300 per kg of Tom Brown, about ¢15,400 per kg of cowpea-
Weanimix, and about ¢15,100 per kg of soybean-Weanimix. Returns to vendors were
estimated to be about ¢9,100 per kg of output.

Thus, the difference in prices between industrially-processed and self-prepared
weaning foods was estimated to be from about ¢3,200 to ¢4,400 per kg. The only major
difference between industrially-processed and self-prepared weaning foods would likely
be that industrially-processed ones are sold in a sealed bag with a label containing an
authority certification logo, while self-prepared ones are sold in an open-air condition
without quality certification. It would be informative to conduct a consumer survey to
assess the willingness of Ghanaian mothers to pay a higher price for a better package and
quality certification, to find out whether the estimated higher price of industrially-
processed weaning foods would be accepted by customers.

Sensitivity analyses were conducted for different scenarios to examine the change
in the price-competitiveness of industrially-processed cowpea-Weanimix with: (1)
industrially-processed soybean-Weanimix; (2) industrially-processed Tom Brown; and (3)

COWpea-Weanimix self-prepared by grain/flour-type product vendors.

218

The results of the analysis showed that it is likely that cowpeas would always be
price-competitive with soybeans as an ingredient in industrially-processed Weanimix,
unless customers prefer soybeans over COWpeas.

The difference in the retail prices between industrially-processed cowpea-
Weanimix and Tom Brown was estimated to be about ¢600 per kg under the most
favorable scenarios for cowpea-Weanimix and ¢2,400 per kg under the most unfavorable
scenarios. Change in the price of raw materials was the main contributor to create these
differences.

The price-competitiveness of industrially-processed cowpea-Weanimix with
Weanimix self-prepared by grain/flour-type product vendors was affected by the change
in: (l) the technical efficiency of processing; (2) the volume of industrial production and
the processing margin; (3) the price of raw materials; and (4) the retail margin. Among
these examined scenarios, the level of processing and retail margins were found to be the
most important factors that determine the competitiveness of industrially-processed

COWpea-Weanimix, compared to self-prepared Weanimix.

219

CHAPTER 8

CONCLUSIONS

8.1 Summary

Cowpeas, a protein-rich and drought-tolerant indigenous African legume, are
popularly consumed at home and also as street foods in West Africa. However, industrial
processing of cowpeas is still negligible. For the goals of enhancing cowpea consumption,
utilization, and food security, many studies have been conducted by food scientists
associated with the Bean/Cowpea Collaborative Research Support Program to develop
nutritious and affordable cowpea-based processed products.

Cowpea-based processed products seem to have high potential because of the
popularity of cowpeas as an ingredient of various traditional dishes in the region, as well
as the growing urbanization and the increase in the opportunity cost of women’s time,
which will increase the demand for processed products. However, previous studies have
identified various constraints to creating and promoting such processed cowpea products,
including higher prices of cowpeas (compared to its substitutes), lower protein content
than soybeans (which are substitutes for cowpeas in the production of some of the
processed products), fluctuations in price and quality, lack of stable availability, and
possibly poor functionality of processed products.

This case study was conducted to examine the competitiveness of selected
cowpea-based processed products in Ghana, analyzing both price-related and non-price-
related factors. One of the two selected target products was ready-to-use dry cowpea meal

developed by the B/C CRSP for preparation of kosei (cowpea fritters, a popular street

220

food in Africa). If commercialized, the meal would save the time of street vendors of
kosei, who currently prepare it from cowpea grain using labor-intensive methods. The
other selected target product was Weanimix, a traditional roasted-maize-based weaning
food (called Tom Brown) fortified with cowpeas or soybeans (IO-15% of the output) and
groundnuts (10%), which was introduced in Ghana in 1987 by the Ministry of Health
Nutrition Division and the United Nations Children’s Fund.

Fieldwork was conducted in the Greater Accra Region of Ghana during February
and March 2007. Using structured questionnaires, interviews were conducted with 20
kosei vendors, 15 custom millers, 18 retailers (including grain/flour-type product vendors
in the market [outlet for everybody], small shops [outlet for everybody], and
supermarkets [outlet for wealthier families]), 30 weaning mothers, and 10 small- to
medium-scale local food-processing companies. Price data were collected from different
sources such as observations/personal communication in the market during the fieldwork,
secondary data obtained from the Ghanaian Ministry of Food and Agriculture, and an
online database. Using the qualitative data collected, descriptive analysis was carried out
to assess non-price related factors affecting the competitiveness of dry cowpea meal and
COWpea-Weanimix (i.e., Weanimix in which COWpeas are used rather than soybeans).
Quantitative data were used to prepare enterprise budgets and analyze the price-
competitiveness of the target products. Finally, sensitivity analysis was conducted to
examine the change in the price-competitiveness of these products under different
scenarios.

With regard to dry cowpea meal for preparation of kosei, descriptive analysis

found no new major non-price-related constraints to industrially producing dry cowpea

221

meal. The processor respondents, who were currently producing cowpea/soybean flour,
seemed to be capable of producing dry cowpea meal with their current equipment or with
a small additional investment. Most respondents used plate mills to grind grain—the
same technology used by custom millers.

Budgets for preparing kosei using dry cowpea meal were constructed using data
obtained from 4 industrial processor and 13 kosei vendor respondents. The analysis
indicated that, under the conditions observed in Accra during February and March 2007,
dry cowpea meal would not be price-competitive with cowpea grain for the majority of
kosei vendors. The difference in returns between wet-milled kosei (i.e., kosei prepared
from COWpea paste wet-milled by a custom miller; current method used by kosei vendors)
and dry-milled kosei (i.e., kosei prepared using dry cowpea meal) was estimated to range
from about ¢3,700 to ¢12,900 per kg of kosei (approximately US$0.41 to US$1.40; US$1
: ¢9,200 [Ghanaian cedisj during February and March 2007). This means that the meal
would only be attractive to vendors who, by switching from the wet-milling method to
the meal, could save an amount of labor that is equivalent of about 100% to 340% of the
minimum daily wage (¢19,000 [US$2.07]) for every 5 kg of kosei they prepare.

The sensitivity analysis showed that none of the four scenarios to improve the
price-competitiveness of meal—improvement in the technical efficiency, increase in the
volume of production (with or without reduction in the processing margin), bulk purchase
of meal by kosei vendors, and fluctuation in cowpea price—would greatly change the
results of the original analysis, if these scenarios occurred individually. If all of these
scenarios occurred simultaneously, the price of dry cowpea meal would decline sharply.

However, to adopt the meal, kosei vendors would still have to be able to save an amount

222

 

of labor that is equivalent of about 40% to 120% of the current minimum daily wage for
every 5 kg of kosei they prepare. The results suggest that, for the majority of kosei
vendors, dry cowpea meal could only be a price-competitive ingredient in kosei under a
combination of very favorable conditions.

With regard to the competitiveness of cowpea-Weanimix, a wide difference was
found among the prices of existing weaning foods available in Accra. Currently,
Weanimix is produced by small- to medium-scale local food processing companies and is
mainly sold at supermarkets, which are frequented by higher-income consumers. While
some of these companies also produced Tom Brown, grain/flour-type product vendors
also self-prepared Tom Brown (using custom millers) and sold it along with other
products.

Descriptive analysis of the data collected found two non-price-related constraints
to the commercialization of industrially-processed cowpea-Weanimix: (1) apparently low
awareness of Weanimix among consumers; and (2) preference for soybeans over cowpeas
among processor respondents (for reasons other than the lower price of soybeans).

Budgets were constructed, using data collected from five processor respondents
and seven grain/flour-type product vendors. The retail price of industrially-processed
COWpea-Weanimix (assumed to be sold at small shops) was estimated to be about ¢400
(approximately US$0.04) per kg higher than the price of industrially-processed soybean-
Weanimix, about ¢1,400 (US$0.15) per kg higher than the price of industrially-processed
Tom Brown, and about ¢3,500 to ¢4,400 (US$0.38 to US$0.48) per kg higher than the
price of a potential product—cowpea-Weanimix self-prepared by grain/flour-type product

vendors. These results indicated that, under the conditions observed in Accra during

223

 

February and March 2007: (1) cowpeas were price-competitive with soybeans as an
ingredient in Weanimix, if consumers saw these products as equally desirable; (2) the
relative price between industrially-processed cowpea-Weanimix and industrially-
processed Tom Brown could potentially be smaller than their relative prices currently set
by the majority of the processor respondents; and (3) whether industrially-processed
cowpea-Weanimix would be price-competitive with Weanimix self-prepared by
grain/flour-type vendors would depend on the willingness-to-pay among consumers for a
better package and quality certification—attributes that could be obtained only from
industrially-processed products.

Sensitivity analysis was conducted for five scenarios: change in (1) technical
efficiency of industrial processing; (2) volume of industrial production; (3) price of raw
materials; and (4) retail margins, as well as (5) combinations of different scenarios. The
results indicated that: (1) it is likely that cowpeas would always be price-competitive with
soybeans as an ingredient in Weanimix, unless customers prefer soybeans and are willing
to pay a high premium for soybean-Weanimix; (2) the difference in the retail prices
between industrially-processed cowpea-Weanimix and Tom Brown would range from
about ¢600 to ¢2,400 per kg depending on the conditions; and (3) the level of processing
and retail margins would significantly affect the price-competitiveness of industrially-

processed COWpea-Weanimix with self-prepared cowpea-Weanimix.

8.2 Policy Implications

8.2.1 Drv Cownea Meal for Preparation of K osei

The results of the study indicated that it would not be profitable for the majority

224

 

of kosei vendors to use dry COWpea meal processed by small- to medium-scale local
industrial companies, under the conditions observed in Accra during February and March
2007. Unless subsidized, processors could not sell dry cowpea meal at a price that is
attractive to kosei vendors, while maintaining the same level of returns that the
processors earned from similar products that they currently produced (i.e.,
cowpea/soybean flour). As reported in Section 4.2.3.3, the currently limited market for
dry cowpea flour seemed to mainly serve high-income consumers who prepare kosei at
home (high opportunity cost of their labor makes ready-to-use dry flour attractive), while
a smaller amount is sold to expatriate Ghanaians.

This finding indicates that, although the technology to industrially produce dry
cowpea meal is available in Accra, it is unlikely that at this time kosei vendors would
switch from the wet-milling method to the use of dry cowpea meal. As the study showed,
the higher the opportunity cost of vendor’s time becomes, the more attractive dry COWpea
meal would be to kosei vendors. Therefore, as Ghana’s economy grows, there will be a
point in the future when the value of time and labor saved by using dry meal would
exceed the cost of purchasing meal. Therefore, it would be best to wait to promote dry
cowpea meal to most kosei vendors until the economy reaches this point, unless the

policy makers are willing to support kosei vendors today by subsidizing the meal.

8.2.2 Cowpea- Weanimix
The study found that, although several types of industrially-processed Weanimix
were already available in Accra, the products were mainly sold in supermarkets which

targeted wealthier families. Also, the sensitivity analysis found that the level of the retail

225

 

 

margin would greatly affect the price-competitiveness of industrially-processed weaning
foods with those self-prepared by grain/flour-type product vendors. Therefore, it is
recommended that industrial processors of weaning foods expand their outlets from
supermarkets to small shops to make their products more available to weaning mothers
from lower-income families.

Also, collaboration between the government-run health clinics/hospitals and
weaning food processors could help to make Weanimix more available to weaning
mothers from lower-income families. Although a health clinic, which was visited for
conducting the weaning mothers’ interviews, sold Weanimix in small portions, the unit
price seemed to be more expensive than the average unit price of industrially-processed
Weanimix sold in supermarkets. This difference in unit prices appeared to be larger than
the cost of repackaging, which the health clinic would have to pay (assuming that the
health clinic purchased Weanimix and repackaged it to smaller portions at the clinic).
Therefore, health clinics/hospitals should seek ways to sell Weanimix to weaning mothers
for a lower price, such as bulk purchase of Weanimix from industrial processors.

The lack of availability of commercial Weanimix for lower-income families was
suspected to be found on the demand side, rather than the supply side. Interviews with
weaning mothers and industrial processors suggested that although consumer awareness
of Weanimix was increasing, the product was not yet widely known. Since it is unlikely
that small- to medium-scale local processors could afford to advertise on a large scale, the
government could initiate campaigns as a strategy to further increase consumer
awareness of Weanimix. Similarly, NGOs with a child survival mandate could assist in

promoting these nutritious weaning foods. Finally, if there exists a professional

226

 

organization that represents local processors of weaning foods, joint promotion programs
could be carried out in partnership with the government and such an organization.
Finally, among the weaning mother respondents, koko or banku, both of which
are made from fermented maize dough, was found to be a much more popular weaning
food than Tom Brown or Weanimix. If this is a general trend among Ghanaian weaning
mothers, the research on and the promotion of cowpea/soybean-fortified fermented maize

dough would need to be paid further attention.

8.3 Limitations

An important limitation of this study is uncertainty about the accuracy in the cost
calculation of industrially-processed products. In this regard, the following three issues
should be noted.

First, the shares of the target products in the total production of the industrial
processor respondents could be inaccurate, due to the assumptions used to derive these
values (i.e., approximated by the shares in weight, value of products, estimated
equipment use time, and number of packages). This problem would be most serious when
respondents processed products with very different characteristics than those of flour-
type products. For example, the model used for estimating the unit cost assumed that the
same amount of labor was needed to produce 1 kg of COWpea flour and 1 kg of honey,
which is most likely not true.

Cost variables that are subject to this type of inaccuracy problem include wage,
equipment, electricity, fuel, water, and printing and stationery. For instance, wages

accounted for 2% to 22% of the processors’ selling price in the representative budgets for

227

 

industrially-processed dry COWpea meal, and 9% to 35% in the processors’ selling price in
the representative budgets for industrially-processed weaning foods. If the actual wage
share of the target product in the total production was twice as large as the estimated
share, the unit wage should have been twice as large as the derived unit wage, which
could have changed the results reported in the derived budgets.

Second, since the budgets were constructed using monthly data collected in early
2007, seasonality in production was not taken into account. However, the volume of total
production of industrial processors’ facility may fluctuate from month-to—month (the
respondents indicated that their production of cowpea/soybean flour and Weanimix/quasi-

Weanimix was not constant throughout the year). Thus, if fluctuations in total production

Q}? P179}?
Z 9% Z 1’fo
8 g

occur overtime, the denominator of each share estimate (i.e.,

 

m m m
th/ (163/ A]!
g .1 8 ’ g
2Q jt / q ej Ajt
g

 

; see Section 5.1.1) fluctuates. Then, the estimated unit cost of the

gee
components for which respondents paid a fixed amount of money every month (e.g.,
wage and rent) or for which respondents paid once in a longer period than a month (e.g.,
printing and stationery) also fluctuates. For instance, if the volume of total production
doubles, the share estimate becomes one-half of the value used in the budgeting analysis,
and therefore, the unit cost of the affected components falls by 50%. Therefore, if the
volume of total production in the month for which the data were collected was unusually
high or low, the estimated unit cost of these components would be unusually low or high.
Third, information on the processors’ volume of inventory was not collected and

therefore not included in the analysis. The payment for cost components such as

228

transportation depends on the volume of sales, rather than the volume of production
during a month. However, due to the lack of information, the cost calculations assumed
that there was no storage from the previous month, and also that all the products
produced during the month were sold in the same month.

The small sample sizes are another limitation of this study. The number of
industrial processor respondents who could provide enough information to construct
budgets was limited to four for dry cowpea meal and five for cowpea-Weanimix. Also,
due to time and budget limitations, the number of kosei vendors interviewed was limited
to 13. Since the derived budgets had components whose values varied widely across the
respondents, a larger number of budgets (i.e., larger sample of respondents) would have
increased the reliability of the analysis.

With regard to the estimation of processor prices, to simplify the analysis, it was
assumed that processors would modify the processor prices so that the values of returns
remain unchanged. This assumption is the equivalent of assuming that the market for the
products analyzed is perfectly competitive (i.e., price levels are determined solely by
costs, including a “normal” return on capital). However, the processor price of
cowpea/soybean flour varied widely across the respondents. Also, as shown in Table 6.1,
there was a wide difference in prices among weaning foods produced by the respondents.
Moreover, as reported in Section 6.2.3.2, the share of each ingredient in weaning foods
varied across the respondents, implying that the taste of those weaning foods also differed
from product-to-product. Therefore, it would be more appropriate to describe the current
market for cowpea/soybean flour and weaning foods processed by local companies as a

market of monopolistic competition, in which similar (i.e., not identical but

229

differentiated) products are sold. Since each processor faces a downward-sloping demand
curve (rather than a flat demand curve in a perfectly competitive market), they own some
market power to change the selling price without losing all customers (Varian, 2003, p.
454). In such a market, processor prices might not increase as much as the increase in the
cost of production, if processors absorb a part of increasing costs by decreasing returns.
Similarly, processors might take advantage of an increase in the cost of production by
increasing the processor price more than the increase in the cost. However, information
on the pricing practices of processor respondents was not collected and therefore not
included in the analysis.

Finally, the use of the same representative prices of raw materials across the
processor respondents may have led to inaccurate estimates of unit costs of raw materials.
Since the respondents had different sources of raw materials, it is indeed expected that
they paid different unit prices for raw materials. This is a concern because the share of
raw materials in the total cost was estimated to be large. Also, the potential problem
related to the use of representative olonka-kg conversion rates was discussed in detail in

Section 4.2.5.3.

8.4 Future Research

8.4.1 Dry C owpea Meal» Prepamw

If dry cowpea meal for kosei preparation is produced by custom millers, the price
of their meal is expected to be lower than the price of meal processed by industrial
processors. However, this assumes that custom millers can produce the type of meal

required to make kosei. Thus, it would be useful to survey custom millers to find out

230

whether they could successfully mill cowpeas into the right particle size range for dry
cowpea meal (with or without a sieve) and whether they would be willing to do it, when
asked by customers.

Further research to determine the optimal soaking and whipping time necessary to
prepare a good quality kosei using dry COWpea meal is needed to accurately assess: (I)
the time that would be saved by the use of dry COWpea meal; and (2) the potential for
using meal to make fine adjustments of the quantity of kosei to prepare depending on the
sales of the day.

To better understand demand factors, a consumer survey needs to be conducted to
find out the potential incremental increase in cowpea consumption through: (1)
increasing home preparation of kosei using dry cowpea meal; and (2) the use of dry
cowpea flour/meal for daily cooking. With regard to home preparation of kosei, dry
cowpea meal has the advantage that customers could tailor kosei to meet their taste
preferences. However, it has the disadvantage that home preparation of a small amount of
kosei would require more labor and oil per unit of kosei than purchasing street-vended
kosei, which is prepared in a large amount. A preliminary analysis on the potential of
home preparation of kosei was conducted using the data collected from weaning mother
respondents. The results are reported in Appendix 7.

Finally, a higher proportion of agawu (fritter made from dry-milled cowpeas)
vendors interviewed had another source of income, compared to kosei vendor
respondents (although the sample size is very small). Since the processing procedure for
agawu is similar to the processing procedure to prepare kosei using dry meal, a survey of

agawu vendors could provide useful information to assess how much labor kosei vendors

231

could save by switching from the wet-milling method to the use of dry meal. Also, the
potential for creating commercial dry cowpea flour suitable for preparation of agawu
(rather than kosei), as well as its benefits and economic profitability, could be explored in

a future study.

8.4.2 Cowpea- Weanimix

This study did not collect any information regarding how much weaning mothers
in low-income families could afford to pay for Weanimix or weaning foods in general.
Thus, it is not clear whether weaning mothers with malnourished children can even afford
to purchase Tom Brown sold in the local market, which is among the cheapest
commercial weaning foods. These mothers’ willingness-to-pay for weaning foods should
be examined through a survey, if locally produced commercial weaning foods are to
target these weaning mothers.

Also, the difference in the nutritional value of Tom Brown, cowpea-Weanimix,
and soybean-Weanimix was not incorporated in the analysis of this study. The
willingness-to-pay for higher nutritional values among weaning mothers would be a key
factor that determines the price-competitiveness of different weaning food products.
Therefore, a study in this regard would be worth conducting.

The study found that grain/flour-type product vendors could sell Weanimix at a
lower price than Weanimix produced by industrial processors. No major constraints were
found for those vendors currently self-preparing and selling Tom Brown to produce
Weanimix. However, almost no vendors were observed selling Weanimix. Therefore, a

further research to assess the potential for grain/flour-type product vendors to produce

232

and sell Weanimix is needed. If such a Weanimix becomes available, weaning mothers
from low-income families currently using only Tom Brown, due to a budget constraint,
might be able to afford Weanimix.

The difference between industrially-processed and self-prepared weaning foods
was whether or not the products were sold in an air-tight package with a label containing
authority certification. A study could be conducted to find out if it is feasible and
economic to set up an independent testing/certification and packaging service that would
serve grain/flour-type product vendors. If such a service could be established, and if the
cost of service is lower than the difference between the estimated retail prices of
industrially-processed and self-prepared weaning foods, it might be possible to make
available to weaning mothers in low-income families a lower-price Weanimix with
greater quality assurance. Also, to accurately estimate the price-competitiveness of
industrially-processed Weanimix with Weanimix potentially self-prepared by grain/flour-
type product vendors, there is a need to conduct a consumer survey to assess their
willingness-to-pay for air-tight packages and quality certification.

The sensitivity analysis found that the level of processing margin would greatly
affect the price-competitiveness of industrially-processed weaning foods with self-
prepared counterparts. A study could be conducted to examine how low the processors
could set their processing margins, while gaining returns greater than the opportunity cost
of capital.

Finally, the study also found that because the cost of industrially producing
COWpea-Weanimix and soybean-Weanimix would not differ much, these products could

be sold for almost the same price. Thus, the competitiveness of cowpea-Weanimix would

233

depend on the consumers’ preference between cowpeas and soybeans. As reported, the
awareness and popularity of soybeans seem to differ across regions in Ghana. Therefore,
a consumer survey needs to be conducted among the target population of Weanimix to
assess which type of Weanimix would have the greatest potential among that population.
Also, there is a need to further investigate why the majority of the industrial processor
respondents used soybeans, rather than COWpeas, as an ingredient in their Weanimix (e.g.,
among the advantages/disadvantages of COWpeas/soybeans reported by the respondents,

which one was more influential than others?)

234

APPENDIX 1

Questionnaires

A.1.l Questionnaire for Street Vendors of Kosei

235

Competitiveness of Cowpea-Based Processed Products:
A Case Study in Ghana

- Questionnaire for Street Vendors of Kosei—

Respondent Number:

 

 

 

 

 

Enumerator:

Date: @907 Area in Accra

Time Interview Began: : Time Interview Ended:
INSTRUCTIONS:

1. Bring a kitchen scale to the interview site.
2. Visit vendors while they are doing business.
3. Buy kosei balls (at least 5 balls) (keep them to measure the weight after the

interview.)
4. Explain who you are and what you are doing.
5. Read the consent statement to the respondent. If she/he agrees, begin the

interview.

(Revised Feb. 20, 2007, final version)

236

1. EXPERIENCE AND PRODUCTS

1.]. How many years have you been selling kosei? ....................................

—-——-———

1.2. Do you sell kosei throughout the year, or are there seasons during which you don’t

 

sell kosei? ( 1 = Year around, 2 = Seasonal ) .......
If “Seasonal”,
1.2.1. In what months don’t you sell kosei, and why?
Months:
WM?

 

 

1.2.2. What do you do during the period you don’t sell kosei?

 

 

 

1.3. In the past year, have you sold any food products at the same time you sell kosei?
(0=No, l==Yes) ......

If“Yes”,
1.3.1. What products? Enter all that apply

1 = Koko, 2 = Hausa koko, 3 = Other (specify):

 

1.4. How many days per week do you sell kosei? ............................................

If the respondent doesn’t sell kosei everyday,
1.4.1. Which day of the week is your day off? ......

 

237

2._ COST OF PRODUCTION

2.1. What ingredients do you use to make kosei?
Check the box of the cost components in the table below.
If the respondent doesn’t mention any item listed, ask if she/he uses that item.
Ask about the varieties or traits of cowpeas, reasons to choose that variety or
traits, and types of seasoning.

2.2. From whom do you usually buy these ingredients? Enter the code.

2.3. How often do you buy each of these ingredients? Enter the answer.

2.1. 2.2. 2.3.
From whom
Codes:
1 = vendor in the market

2 = retail store Frequency of purchase
Cost components 3 = wholesaler

4 = farmer
5 = other ( ify)
Cowpea grains*
Onion
Pepper
Ginger

Salt

Seasoning“

 

Water
Oil
Fuel

[:1
D
E]
13
[3 Garlic
[:1
Cl
[:1
El

 

 

 

El
E1
1:]
I]
V

Ilt

arieties or traits (e. g., color, size)

 

Why that variety or traits
** Types of seasoning

 

2.4. Do you use the same ingredients throughout the year, or do you change the

 

ingredients for different seasons? ( l = Same, 2 = Change ) ........ _
If “Change”,
2.4.1. What ingredients change and when?
Chamge: Months:
Change: Months:

 

238

2.5. How much of each ingredient do you buy each time, and how much does each cost?
Enter the answers in the table below.

2.6. If you were to buy the ingredients you used yesterday, how much would you have
paid?
Enter the answers in the table below.

2.5 2.6
an and teach rchase Cost of

g, kg ingredients and

Payment (Cedis) fuel used
yesterday (Cedis)

Cost components Amount Unit .or
litter

/ unit

Cowpea grains

Onion

Pepper

Ginger

Garlic

Salt

Seasoning

Water

Oil

Fuel

 

El
E]
Cl
1]
C]
D
E]
D
Cl
Cl
Cl
[:1
CI

2.7. Do you make the same amount of kosei evegyday, or does the amount change in
different days of the week? ( 1 = Same, 2 = Change ) ................................ _

If “Change”,
2.7.1. By how much does the amount of kosei you make change, and why?

By how much:
Why:

2.8. Depending on the availability of cowpeas, do you change the amount of kosei you
make at different times of the year? ( 0 = No, 1 = Yes ) ............................ _

If “Yes”,
2.8.1. By how much does the amount of kosei you make change?

thow much:

239

3. PROCESSING PROCEDURE

3.1. Do you prepare kosei by yourself? ( 0 = No, l = Yes ) ...............................

 

If “Yes”,
3.1.1. Do you usually prepare kosei alone or does somebody assist you?
( 1 = Alone, 2 = With (specify): )...__
If “N0”,

3.1.2. Who prepares the kosei you sell?. ..

 

3.2. How do you [or the answer to Q. 3. 1.2] prepare kosei? Please explain step by step.
Ask (a) the time of the day each step is undertaken, and (b) the time needed for
each step (including the time for moving from one place to another. e.g.,
from home to custom miller).

 

 

 

 

 

 

 

 

(1) ( 1:1 AM 1:] PM)( [:1 min E1 hour)
(2) ( [:1 AM Ci PM)( 13 min [:1 hour)
(3) ( 1:1 AM 1:1 PM) ( 1:] min [:1 hour)
(4) ( 1:1 AM [3 PM)( [:1 min :1 hour)
(5) ( 1:1 AM [:1 PM) ( 1:] min 13 hour)
(6) ( [I AM 13 PM) ( E1 min [:1 hour)
(7) L C1 AM 1:1 PM) ( C1 min 13 hour)
(8) ( 1:] AM 1:1 PM) ( 1:1 min 13 hour)
(9) L 13 AM 1:1 PM) ( CI min Cl hour)

 

3.3. Do you fry kosei balls here or home? ( 1 = Here, 2 = Home ) .......................

240

3.4. In the past month, did you ever run out of kosei balls when there were still
customers who wanted to buy kosei? ( 0 = No, 1 = Yes) ...........................

If “N0”,
3.4.1. How do you make sure that you don’t run out of kosei balls?

 

If “Yes”,
3.4.2.1. How often did this happen? ..................

 

3.4.2.2. How did you feel when you ran out of kosei balls. Which of the
following examples best indicates to your feeling? ...................... __
= I should have prepared or brought more kosei balls! I could have
made more money! Tomorrow I will prepare or bring more balls.
2 = Good! It’s sold out today. I made enough money. Tomorrow, I will
prepare or bring the same amount of kosei balls.
3 = Other (specify):

 

 

3.5. In the past month, did you ever have remaining paste or kosei balls when you
wanted to end business for the day? ( 0 = No, 1 = Yes ) ............................

 

If “N0”,
3.5.1. How do you make sure that you don’t have remaining paste or kosei
balls?
If “Yes”,

3.5.2.1. How often did this happen? ..................

 

3.5.2.2. How many leftover balls did you usually have? Or if you had leftover

paste, how many balls could you have made from that amount of

paste? .................................................................. balls

3.5.2.3. What did you do with the leftover paste or balls?...___ __ __ __
Enter all that apply

I = Threw away, 2 = Gave to family, 3 = Gave to friends
4 = Other (specify):

 

3.5.2.4. Did you have any leftover paste or kosei balls yesterday?
(0=No, 1=Yes) ......

If “Yes”,
3.5.2.4.1. How many lefiover balls did you have yesterday? Or if

you had leftover paste, how many balls could you have made
from that amount of paste? ............................ balls

24]

4. SALES

 

4.1. At what price did you sell kosei yesterday? ............ Cedis/( )balls

Reminder: Buy kosei balls (at least 5 balls) from the respondent.

4.2. Do you usually keep the siz_e of kosei balls the same during the Qty or do you change

 

the size at different times of the day? ( l = Same, 2 = Change ) ........
If “Change”,
4.2.1. By how much do you change the size during the day, and why?
B45 how much:
Why:

 

4.3. Depending on the availability of cowpeas, do you change the siz_e of kosei balls at
different times of the year? ( 0 = No, l = Yes ) ......................................

If “Yes”,
4.3.1. When do you change the size during the year, and by how much?
When:
By how much:

4.4. How much was your total sales from kosei land the gnswer to 0.1.3.11 yesterday?
Cedis

Cedis

4.6. Were yesterday’s sales from kosei normal for this time of the year?
' (0=No, 1=Yes)......

If “No”,
4.6.1. How were they different?

 

 

 

 

242

4.7. Depending on the availability of cowpeas, do your daily sales from kosei change at
different times of the year? ( 0 = No, l = Yes ) ......................................

If “Yes”,
4.7.1. When do they change, and by how much?

When:
By how much:

4.8. Do you have any other sources of income other than selling kosei [and the answer to
Q.1.3.1]? ( 0 = No, I= Yes) ...........................................................

If “Yes”,
4.8.1 What are these income sources? .......

 

4.8.2. In the past year, was your income from selling kosei land the answer

to Q.1.3.I|: ....................................................................

I = the most part of your total income
2 = not the most part but more than half of your total income

3 = half of your total income
4 = not little part but less than half of your total income

5 = the little part of your total income

 

243

5. USE OF COWPEA FLOUR AND MEAL
5.1. Have you ever used commercial dry cowpea flour to make kosei? ..................

I = Currently using, 2 = Once uSed but stopped, 3 = Never used

5.1.1.Why?

 

 

 

Say, “There is a new type of cowpea flour specially developed for kosei preparation. You
can soak the flour in water for between 30 and 60 minutes to make paste, and the
kosei made from that paste tastes as good as the kosei made from cowpea grains.”

5.2. Would you be interested in using this new product to make kosei, if you could buy
this product? ( 0 = No, 1 = Yes ) .......................................................

If “Yes”,
5.2.1.1. How much would you be willing to pay for 1 olonka or margarine tin

of this product?... Cedis/ D olonka [:1 margarine tin

5.2.1.2. From whom would you prefer to buy this product?
Enter all that apply. ...................................... _ _

 

I = Vendor in the market, 2 = Small retail store,
3 = Big retail store such as supermarkets

4 = Wholesaler, 5 = Flour processor,

6 = Other (specify):

 

If “N0”,
5.2.2.Why not?

 

 

 

Say, “Thank you for answering my questions.”

[END OF INTERVIEW]

6. WEIGHT OF KOSEI BALLS

After the interview, measure the weight of kosei balls purchased.
kg/f ) balls

244

A.l.2 Questionnaire for Processors of Cowpea or Soybean Flour, Weaning Foods
and/or Garil

 

' As mentioned in Chapter I, COWpea-fortified gari (grated, fermented, and roasted cassava) was originally
among the suggested target products for this case study.

245

 

 

 

€228 Ea .88 .m .32 Easy:

. a
3232:. 2: Smog .82? 0505 .2 .Eot:o%8 2t 8 €088me E0809 2: 93m .mZO—PUDM—Hmz

 

"co—Em— 32335 25... ”:anm Barr—3:— 05:.

 

25:32 353395 SEQ “3.5

2215.2 Bow—8&3”

 

"..Sabfisam

$.30 .5ha £59m wfiawog mic—m anon—«mam .8 «2:30
we £88939 .8“ 3353530 ..

«52.0 5 beam «30 < "£2695 5338.5 comamiaoakeu .3 awe—SZEQQEQU

246

 

3:0:

..........
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

 

 

Ego:

 

0..
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

II.

830;

I

......................................................... AWU>HH

If

......
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

0 Add;
..................... .0103 3:25: .85 ow x83 ..om m5“: “HM” Mom .311
............................................ .xAH—finoflm mwgw F: V—HOB WDDNAo—QEU a “VD .5
............................................. o . . . .

. . .V.fi
................................ 38>» 333: L85 ow x993 ..om mac: NM” ”Mm Mme.—
....................................... . 326$ £5 E x83 3&2an 253:... BEE
......................................... o . . . .

 

................... $53 333m: 35 on a 2:0: mam“ ”HM“ WWW”
............................................ .35 c
....................................... bamHm—mowm mEu Em v10>> 50% .523 HOJHO when—:9:— N_. .—

 

.IIIlm base 302 .Zl
wbzmofl So» E x83 33%: so» ow :83 .3 £30:

ӣ3 ES Sci
”23 =25 so; .
2538 353 8m 98 5:3 Ea: can Bow 2: =2 3.on ._ m _

.amo>s .=

.02 u o V $883 mm 326$ So» 55>» So,» of E mwotom 205 2.4. an.“

 

 

 

 

 

 

 

 

”>5:
as?
f

   

ooooooooooooooooooooooooooo

”c055
nowamso : moon Bo: tam ES .555 ._._.N._

fax»: =
.02 n o v $8» 2: mo £583 2: co wfiwcomoc owcmsu 83mg

2: moon .mmg

 

 

“.UOB m: “ousrhl a“: u.:§l ............
E 3230 350$ So» meow H3. ..om 9:5.— zcmE 305 .m._

oooooooooooooooooo
IIIIIIIIIIIIIIIIII
0000000000000000
000000

mmmocmmsn 29m biog.“ So» Be So» 3:3 5 .:

maps. any .52? 22355.2. 4.5—H.501—

247

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.25m
to.
Z
S
o
m
m
o
F . . w
v
N . .
j 7
fifhl N
i fill! _
41%.. 13.3 am a @2181me m m m m m m m w m. m. m
N [1 m. n e M w m m a n u. u
g AC m. M. m. m m 1 m 252. 23.5
@ J m u
[E {liar .. m till
3569qu
n; um 04 m;
Watooqfl 850 n v .5me 2: E 833/ .1. m £22233 0 N .558“. .1.
3.8 2: 33m $5629: :28 35 so» ow 50:3 Sci 5 _
3::— ofi p.925 A8 3:“ A3 A5 2:55 »_:o 23.. 3830.5 :08 mo £560qu 2: 2a :23» o _
.»_o>=8nmo.. .taw .8 . cc usages? ..Sc
53.335 3 52—38 m_ 335.:— ofi .= 2:5: 3:2: 2: ~28: Ag .8 .36 A5 823 .332. 035 o c .— . a
.88“: 285 05 2: =8

5 2 .8323 2: SEN
a 3:3 .23» Esq 2t 5 .m._

835 $2603 so» Be 3260

248

 

3 .858: =
: 9.: 8:388 .8: 80:3 8 8:: .=

.O—A—Nu an: :—

9::u
8 o a 29:: 6% m :8 w:_::o3
8338:: “Mom—HM“: “9.8:: .8838 new: “Eu. “Ma—”m 5.10:.— 2: 29::
J m 08 a: o E. . . .

.-.mhfiwouh. “.3: 8:3: 1:“ :5: :88 u:_::e3 8::

3:2:8w5 :: 8 80 3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

Q
:
K 2
lll‘l'lll o
l w
ll 5
o
l I m
4 l f.}{ v
mmmmmmmmmm [11:s;||:|;[1IIIIIIJIIJIllnlllvllxtll4Iltlla|||¢|||alllllrw N
8818] ~
1 [ [{II M: 828.. 9qu
km 8: :83: :2 A: .
is: E among... :23 as E 3.5%.:
It." 0 I/ f 1 88/1
3 w _
[88,.\“ it

 

 

 

:33 :o» E: 38:08 08:: mo :28 .8 :88 3o: .3 0:: E d:
48:3: .5828 9:: o
:88 8: b8 88:: :8 88:83. :5 mi
83 8 w: :2: 8:8 3:: : E 8038: :

8:888 9:: :8

:: :8 5:: .8:
.8888 3:: :2: w:

3:88.. 388.:— .3
8:8 3:: :m:

.838: :_ biog; m
:. 83w 888:8. 0

sec 2: no :23 MO
3::m 888:8.— o.: .=

_: Z .9338: a: 88:
::.:e:. :8. 338:: :0: E: 38:08 08:: mo :23 .8 :88 302 .w:

 

249

 

 

a a CO WGMCQUB
80:00:. 00: So: :0: Wfiaflflfia 0:: E .N...
>838:
8.: 8.00 .80: 80
0:000 : . 00.808 :0» 3038.. 0:: ..0 =:
.80: 80820088300 8.8.0:. U

0:00: 0:: =
. . .838: 8:::0 80::

:0 0: ::0 8.:0 .53 2: :2: 00 £038.... 2: =: 8.. a — O 0.0:

.88. :2:.:0.

 

. . 88:03
............................. $8808 0:3 So: :0: .8: hhwwhwhowfi .0 E :. .. ...
............ :3 0:8 30
8000 0:880:38 8:00.05 608008 :0» 38.00.. 2: .8 =: 8.: 00.00 .80: 80: m :
.80: :00 . .

 

 

.8: N00820: m

 

 

:00\.ME§03:200N0W N

 

 

 

80....=3::Ae.w\:0&30© .
7

 

 

\ . . 2:3 :.:0E .0:—
:.::\ :28:— am:. 2:3 :82: 8:. fix: . .
0x; 0.. 8.0.0. 8:: 0.x... 0: 0:50. 8:: 00. 2008 8::

838.5
1 i3 “whiff
d0... 32... ét 5.5
- . 0 _ . _

 

 

 

 

 

 

 

 

 

 

 

 

 

. a. :0 00:: 2:: 2: .=
0:: 0:: 0:0 0 08 m w:.. 00 2: 80:: 0.0: 00:58:: 808:5 .8: 00w:::0 80:8.— . .
0. : .. : . .08 20:: 0080.02. .8 38.0. =: .8: 0.8N

2.8% B08200. 88:: 5000\sz003 “00:00:00. .80: 5:80:00 0.: 00.8 :::3 :0 :80:— um:. 0:: 8:3. .0. ..

82.300 8% :00 :0»

 

£2808: :80: ::0::%0m\:0:300: 08:3

.80: 88:80 MOZ 8:300 walk—ml 008.00.... 80.82.00: 2: .2

0. 08:0.— 8::08: :80: ::0::8m\:0:300

2:0: 88:80
3 .80: 82.980 ED:

.80: :8300 0038.... EOZ 80: 80.82.00.— 0:: .2

8:0: :8300 0. 08,—0.2 8:88: :80: 8088038300

3 .80: 8:300 008.002. 80:82.00: 2: .2

"0087.

250

 

 

 

I“
_U

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

D
302 D
302 m 0:...008 w:..:0m D
302 m 8308.3 D
302 m SEQ D
302 m 30.5 D
302 m .0800: D
302 m ...8 0088:: D
302 M :8 80.8.3: 08:... D
307. m .5820 D

...mmwwwwfiu .0800. 00.... .0=0m 00.08 ...wwmmz 2%. 808.83.
5: ed Wm .mm Wm Nd ..m

 

 

 

 

 

 

 

 

.0038: 2: .08: 08088300 0...: 8.80:0. .0: .00.. .0: 80:0 :9. 0.0 02202: :0 2.08: 30.. ...N

0.8... 30.00.00. 88:: .800\m§§0

.0038: 2: .08... 0.0.00 808880 0.... 3.0 ..08: 30.. .oN

.0038: 2: .08... 08088800 0.... 8.. :0... .0... 82.3 80.”. .

...038: 2: .08... 08088100 0.... ..0 .00.:8 2.. m. 2.3 .

.80.. 2: .08: A... 880.. :0: 02.3 808880 0... 0:3 20 30: ...m.m 5.00033 .—
3.307... 08.. 2: 0.02.0 ...307... ..— 0:0:: .0 30: .. bx. :0» En. .

30.2. 0.2.3 2: :. .038: 2: .08.... 0808880 0.... 8n. :0: .0..0 .00.. .83 a. .
.030. 2: :. :2. 2: 0.02.0

.: 30.00.00. .80: 000:0..02009300 00300.: 0: 0m: :0» 0.0 80898.0 .055 .

m.N
.mm

Wm
NN

..N

 

8:28:00.L~

251

Guam 3Q: 8 3.3.56»

 

 

 

 

 

 

 

 

 

 

 

oEwcu

 

336 m

 

ES

C]

 

.2502.—

€o€=<v a“:

 

25020 D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oEwco

 

 

386 D

 

 

pOuOE

5:80 B

 

25029 D A

 

f

#L__LL

 

 

E5} E5

 

.:=oE<

 

=5: ..om 55035.5

.532.
-35:

33¢; 335:. 3 com:

ESE—Eva

 

 

ZN

 

Ed

 

3 V 2

 

 

£55.59 :5. :2: w: >52: >5: is .w: 52: .35.. :5. a E 9532:“ of 82» 253.5%“: 2: .= .322. 2:3 2: :5 Eu
$2605 :9» :8 $2608 805 we :03 («o :02: Bo: £52. 28 .8.— m 2.532: 55:. was Leeann 33;; £5 S: no» .3 A fin

£952? 2: ..Sum ““96: m 2.222: 55:. can .83..“ 2: oc $3888: :02: 30E .ofim
48: 2: 28:0 $530 33% go 565 oEo2o as E @2958 m 2.5an 55.: new .53..“ So» 8< dd

.322. 033 a... 5 mafia: .2.—5.3 2: Swan.—
358353 08am 2t 95: $805 :0» oc $2605 655 .wN

252

 

 

2::an mam—mom

 

5328*?

 

SEQ

 

mmoi

 

DECIDE]

summed

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ocmwco

33% D

58 BEENZ D

BEE

05020 D

€wa :5 ESE< :32.
.52. ..2— 552695 -33: $39— 8355 3 new: EoEngcm
EN EN 0N wN

 

«maxi 33.33% Eeuxfiwafiatab

 

253

 

 

 

£50 :23» .23 05 E Easing ow?» .53—Ea 39:5» 33 :02: 302 .N.N.m.m

 

 

 

£qu ....... $28:— Ea. :88me owmz, 13028 28% so» Ev 50:8 >51 ANNA
.355: =
$qu ........... 2:5:— uom Easing own? 3028 95% so» ow 5:8 303 ._.~.m

£99?» £5.58 $83.. .=

........ A owsmco H N 6me bficoE BxE H _ v
@103 mo oEEO> 2: 8 @6808 uwcmco $me ow S 58:03 So» 2 owma 35:08 onc a man so» 0Q .m.m
62:22.: .3 .89 a ma gin an? own? ..—

 

 

............................................................................ oAEQum
2: ..8 ~39 3 ma.— oéoam .= :3 693: 58. ES .535:— ramocc “song—mob 2: .5 bags :2: 318 on. =a ..Sam

 

”95033 550 u 3 Jab H a
.cocmoEzfifioQoEF H w ,bocosflm av wEEcm H n A2022, 5m mam wfiwflocc 203928889 n o
scum H m .883 H v Acosmtommcmb Sm mam mafia—0x3 .25 H m 56585 H N hows? H #

5..me 8 may: no» ow £80 mo 22 :33 £8:on $50 .50» mo =m
9:0:on H8 02m :5 4wa 5:38 ,mwoow chaoa 55G :aBQOmthz/oo wEosvoa 28 Eco Ho: SE63 2053 So» 8830 0% ._.m

flbfiofi So» wcsfiuqo *0 $00 2: Sons 30525 V7,“ 8 ox: 2:03 _ 302: Sam

 

ZO—Péma—O >P—Q—U<h hO hmOU .m

254

 

£50 . . ...2808 A55 583 5a Ana so» 05 owm5>m :0 5:8 302 ._.m.m.m

 

 

 

{cw—=30: .—

l ................. A 0320 u m “08% u _ V 28»
2: 5 $8: 880mb“. Hm owcmno 8:085 2: 805 5 0558 A55 583 5m 5.58 .5 8:08m 08mm 05 A3 39A OD .m.m.m
£30 ............................... “...—:58 En— 808A8 533 5m 55% 39A Ev 5:8 30E .m.m.m

ll ................................ A 50m H m . Afifioamv 550 H m $85800 53>? «EEO u A v

@033 A5 5% cc 80:3 89E 4.0m
£32225 5 .59 a ma nozw $3 533 .:

 

 

mmwou ............. 2:58 A55 58 5m Ana so» ow «NF—98 .5 :58 >52 ._.N.v.m
...owufiu: 8

ll

................. A QWGNH—U H N .Dawm H — v WHNQA
2: .5 38: “8:0me 3 owswco 858m 2: $05 5 8808 >55 63 5m A258 :5 8.58m 083 2: A5 so» 00 dim

 

 

mmUOU ............................. Ava—«EOE ww—w— “Coexfiwfl —®a MOW UGDQm 50% U—U £038 30: ...—Vm
£53325 .5 ~59 « 3 :03» 33 .5355 was... ..8 ma Eta—:8 .2..— .=

 

 

 

£80 ..... 2:58 596 555520 88 Ana 5% ow away: .5 5:8 303 ._.N.m.m
...»?20: =
l ................. A chwfl—U H. N nOEQm H — V «VHNOAA D£H

5 88: 805b€ Ha owcwno 858w 05 $05 5 858 555 A5583 5% A258 5 858m 088 05 A3 5% OD .N.m.m

 

 

£30 ..................... “...—:58 um.“— 80882 558020 88 55% :0» 25 5:8 303 4.0m
£32225 .5 ~59 « 5 52a 53 ”$3.58.“. .2

255

 

£qu ...EEoE .005 cocmoEDEEoQBB 8g .92. 8% ow cushy»: .8 :03: 30$ ._.N.o.m
c.9329. =

l ................. A OWENSU H N «DEQW H — v NHMD~A DSHmO moemu
“Sohowmmfi Hm owcwfio “CH—GEN Ufiu mDOU ho “Sufi—OE %.~®>v COZNUMCDEEOUBOH how >0COE ““0 “EDGE“ osww 06: %6Q 50% OD N.0.M

 

 

£on ............................. 2.28:. «ma. 8338588828 88 95% so» Ev 5:8 303 Add
8388.8 .8 «89 a as :02» 33 :o_89_==EEeuo_8 .=

 

 

 

 

 

 

£on .2282: bog bucocfim was maria 88 >3 :0» on «war—98 :8 :02: 302 ._.N.w.m
$329. :
ll ................. A owcwnu H m 68% H g v $8» 2: mo 38: 88056
E owcwno 8:08.. 2: 808 8 .588 >85 5283mm was wcsfia H8 .386 .8 888m 25% 05 >3 so» 0D .N.w.m
£on .......................... “...—:8:— «ma. bucoufim van 9558 8a 28% so» Ev none 303 ._.w.m
£32228 .8 .89 a ma EZw 83 3283.“? 6.8 «.553 .2
£50 ...EEoE >88 coswtommcw: 8m awn so» ow awake: .8 5:8 302 ._.N.h.m
vow—35.. =
l ................. A owcmnu n m 68% u _ v $8» of .8

$8: ERNEHV Hm owcmco 850:8 2t meow 8 .58.: >55 coumtommcab 8m 388 no 88:8 25% 2: >8 so» 0Q .N.\..m

 

 

£on ............................. 2:8:— umfl cozmtoamcwb 8m 98% so» Ev :02: 303 4.5m
£22228 .8 ~89 a ma :o>_w 33 3.939» .8.— mau “Eu—:25 838.838.: .:

 

 

hmo> D £HCOE D \mWfivU ooooooooooooooooooooooooooooooooooooooooooooooooooooooo NVHCQH ”SH mm gone 30: team
£32228 .8 .89 a ma 53w an? 80.. .=

256

.80 .8588. 80.2.0 .wEE... 80.2.0 8.. 9.0.— 808830. 05 .= «.008 .m=0_00._mm .082

 

 

 

860000... B .00.?D 5808 U 8500 Amv
080.08.ng .00.. B 5:08 B 2000 ANV
0880008 U .00.. B 5:08 B B500 AC

...805 8.. x00 :0» 00 £008 >20: 0:0 ..A05 0.0 .055 53>: .—

ul .................................................................................................................... Am0>nfl oOZHOV~uU>O£N
500208 .0: 80 80.0. .0800 .0000. 8800.5 .80: 8000080808500 8.0.50.0 0. 2.500% 0.0 .05 0.000 .050 .80 0.05 080. .m..m

 

 

 

 

 

0:00U ......5808 80.6 A a 8.. >00 :0» 00 0u0..0>0 ..0 £008 302 ...N. ..m
500.55.. =
................. A 080.5 M m .080m H _ V 0.00% 05 8 .085 80.0.55
.0 080:0 8:080 05 .000 .0 .5808 .85 A H 8.. .0808 .8 8:080 080. 05 >00 00% 0Q N. ..m
.800 ................... 05.88 .00_ A _ .0. 8.0% :0» 05 558 30: .... ..m

0.85.0.8 .8 00:0. 0800 05 ....0 .838 0.03 850.08 .8 380 .050 .—

 

£HCOE\mmUUU .................................................. A...FfiH—hhvfi‘c \AHU>U \ANQ 30% OHV 3 xNH £038 303 .o—.m
$850.28 .8 .80 0 00 ..02w .0... a .=

257

ANNA—v :0: 0.0» 000 A505 ..m: .000.— 8.800.: .305 ...... 0. 0w .0080—00 00.5 .00: 0000080800300 .=

983.00 0000 8 00800.80 0000.00 0.508.020 .8 808.... 00.
08.006008 0.8 80.808_ .008 00. 000000 ....00 0080—00 .2000 .0000 $0000..» 00.00.00 .80: 0000080800300 00. «.880.

 

 

 

 

 

 

000.5 .00 00008...— 0.0000...—
m.m_.m m.m_.m ..m..m

 

 

 

 

 

 

0000380 080.0 00. 0.03 0003 .m.m . .m

.0520 2.. 82.5.. 2.3 .02..
00808 03000. :0» 000 30000.0 00003 .0 0 ...m_.m

I 600?. ..—

........ AWU>H~ nOZHOV
0800080000 $080.00 .0 00000080000 _0808...0>0w-:0: ._0808...0>0m 00. 80.0 088w 0080080 .80 00.0000. .0>0 00.. 0>0I .m _ .m

258

 

:33
”:02: Bo: hm
”5:3
Q33 28 £02: Bo: 3 .owcwno : $06 555 Add
...owufiu: .:
III ......................... A OWGN£U H N 698% H — v
$3» of («o 88: Buyout—U Hm owcmso EsoEm of meow 8 .3608 Dog Soc amonwxomRoEsoo Mo 252% 2:8 2: 825a so» 0D .mé

 

 

as”; _ A as; Ego TTT EEE I as: :5;

$850 so»
:8 So: :02: Bo: @5me mo EsoEm 85 80¢ 98 55a 3355 39A 2:: :08 8: so» ow mEEw amona>OmRum>>oo :02: 25$ 66

.............................................................................. h—QGN wan—u =5 hogs-m—
A ”35025 .850 H o 45% m .mExSm H v 6on H m .8me H N £5me H # v
mason 8:on 8 8: no» ow £338 («0 bomb; £033 .mé

......... A mu> H _ noZ H o V 2:5 9 30% x8 wan 925 2080530 85 36% 39:50 Sci Soc 3858 0260.5 so» 0Q .mé
...:o: .3958 2:55 wfikfitoefi u—

0000000000000000000000000000000000000000000000000000000000

mhsoc amoQ§OmRoa>>oo wEosvoE tSm 526$ So» Eu #3» 3:3 5 ._.v

:55: :wosmxahumkg mo =26:on So» Sons $5525 080m in 8 BE 2303 _ 302: sham

 

mDOJm Z<mm<>0m2mm>900 ho ZO—HUDGOy—n— .v

259

 

 

smoocubtzu 2: 8a 655 .~.N.o.v
.amo>: .=

. Am»; n ~ .02 n o V 2:0: 28938 3%? 8858 $805 9 8: so» 8:63qu 2: E moocuhotav is 205 8< .Ndé
...—5: anon—38 % «2.38 52. 8933.:— Eoenonmoh 2: .=

 

 

32 a as a

a:

 

dso; D 56 D

av

 

dson D 5.: D

as

 

dsoc U SE B

S

 

go: U SE B

6V

 

dsoz D as D

5

 

Ago; D 58 D

a;

 

A50; D 58 D

6

 

6:0: D EE D

5

 

A50; D 5.: D

.333: 2: :33. 83$ .36 some 5m c250: b—moab 2::

 

3

05 03:58 28 new 3 new 5298 333 $20: cmunmxoflmomzéo 835.:— E; «E: :23 30:8 no» 353005 2: 2m 23>? .cé

260

 

mfiucoe .................................................. mafiamxme Hm HDOG goflw%0m\dugoo UuOum 30% GNU WCO— >90: ..éV
...ououms .=

. A 2£mmoo mm »_v_oSc mm :om H N ,2on H _ V
Auo—£33 ma »_v_oSU mm Soc of :m :8 so» ow ho Soc amongomhoesoo 8on so» oo .26

 

§ V ex. V $38325
Ac». V £96365 Ac». V may—«Snowmen .cxo V moo—88M Ac». V £28223
no“ :8 So» ow :oEoooS 855 .So: amongomhonaoo no 23853 So» 5on so» xmm oo 8:— EsoB _ 6%

 

 

“EB
SSE 3o: »m
”:23?
of?» SS SSE 3o: »n .0985 x mooo 555 .wd
{$52.03 =

 

 

................ A UWCNF—U H N oOEdm H fl v
$3» of V0 3E: Eoootzo ow owcmno corn 2: mooo go ,So» 2: Sonwsofio 8% 2:3 2: S» Soc amongomhogoo :% so» oQ .wé

 

 

 

mommxoao » VEBU 3V
mommy—own» V230 3V
WUWNv—mefi V\mH~UUU ANV ...muHmOU —MtUHNE wCMWNMUNQ OF: mUOmv £038 303 .N..N.V

 

StomamV .550 H v coma; H m :80 H m “wan o:o_»5o»_om H _

CV m 3V ” 3V: . é}— .c—.—O oomV mason :monm»om\woo>>oo :8 oo 8: so» AmVowmv—oao 2: V0 3838 2: E 355 45¢

261

$535.3 2: one was Rmcosmozo »E watoBmcm how so» Vigil: »am size—thy
4523:: 9.5:. fl sot—B «no 2: ANNdV aw: ..o ASA—V :m: .556 o. ow $33.5 «no—Somme.— ofi 3 _Sa Jo» 2.3. «o: .=

 

 

 

 

mSoc a wESooS o“ mofimbmcoo SSS ,mowficgowflo ,mowficgom 05 08 x55 :0» oo 355 .mfiv

 

 

 

mSoc 8938 wEosooS @on so» So 33 4.2%
arom0>3 M—

....................................................................... A m®> H — noz H o V mange—7w monk/CU VOUDfiOHQ h0>0 30% Q>NI N~.V
.Soc gongom »_:o moo—6o...— »zuotao Eon—Same.— 2: =

 

 

 

EB .2 3
............. A 265— ::oQ H m €82:th n v dwcwco :65 H M 453835 N N 458805 N fl V
noggoosa S howcmso Ho: 63886 .8385 Soc amongoflmogoo Mo 5535:. So» Bo £23» 09:: .man 05 E g ..v

 

262

 

”>55
ESE 3o: »m
”555
3:3 USN ESE 3o; »9 .owcwso : moon 5:3 ._.v.m
.355: =

........................ A UWCNH—Q H N «OENW ||.I — V Nhflu~h
2: mo 8E: 25:56 E owcmso EsoEm 05 mooo S .EcoE »._o>o _hooxwfitooi R3333 .Vo VSoEm 2:3 2: 02605 so» on Jim

 

 

 

 

 

ESE V A “VVSV
953 V A HVVEV TTTTT 2:33 I use
Ecgux V » HESV
ESEJ V » HSSV

39:0 ANA— .e.—.O 03V 8:23..ch

AVE—«Bo so» SS 8355 05 («o :SE 3o:
.mEoVonE Vo 8:95 85 an can .coooxmeSoP: V3333; @26ng so» 0E: some 8: so» oo EoEEwE come mo :SE 262 .m.m

.............................................................................. ~h—Qfl—N flan—fl -—N boas-m
2 ”E525 55o u 0 .8m m .2315 u v 5on u m .5sz u N £532 u _ V
onthEooi .3383: 8:2: 8 8: so» oo 80958 .Vo but? £255 .N.m
£3933 2.23:3 coo.— wE—EoB 63923 2: ..—

 

.......................................................... hcotcx.%2.~23m§ .BMFQQNNWH— wcmofifinfla taam %H___owm .2.—OVA UJV udvk HN£>> CH ..—m

:.§o€®=.§§.: V3339: Vo :ocSooS So» Sons mcocmoso oEom xmm ow 8:— EsoB _ 3oz: .»am

 

”OCR UZ—Z<m>> GmFUmAm—m HEP mO ZO—FUDDOmm .m

263

 

lAUv mlllIAnv mlllAav ...................

3:033 550 H v ,8me H m aEmu H N 93 use—3830; H V

C}— .c—AO oomV wTQQMEngi 333333 :8 8 um: 30% 35 $3338 2: 30 33858 of 3 355 Add

 

 

 

 

 

 

 

 

 

 

dso; D EE D V S:
dzo; D £8 E V AoV
dso: D 55 D V AwV
dzo; D EE D V CV
95: D EE D V 3V
duo: D 58 D V AmV
dso: D EE D V QV
duo: D EE D V AmV
dsoc D 55 D V ANV
A50; D EE D V CV

£355“ 2: 9,53 32>? .35 some 5V B38: 33295 2::

 

2: 83:33 3cm 3% E 3% £398 385 mwéfimEzsmi 32333 335.5 :9» 2:: :23 30:8 so» 3338th 2: 2m :23 .m.m

264

 

 

E3 .33
II ............ T505. rcoa n m 6838:: n v 6985 :65 H m £88500 H m £8885 n _ V
mofiBosc
5 .owcmno Ho: “885% .0882: g @8550 mwooV wcmcmoa («o :28:on So» 2v 6.50» 99:: “man 05 5 .o_.m

 

$535.— 95 E .38 uni—.33 E3 335.:— 3 com: a...“ 23959 .=

 

mat—MOE ................................................... NESEMXNE um EQQ\M=.~=B§S 32.939; DHOHm 30% GNU wCO_ >90: ...—@m
{9.3%. ..—

..... A Bflmmom mm x2023 mm =om H m “2on H V V
Auo—£80m 8 3x056 mm 825E 2: :3 50% ow go _hcoxwfizcmi 3:832 8on so» 0Q .ah

 

ax. V 6x. V ”EtoommV 850 AX. V 30505 SE £08925 12636:—
..ch V 8:50 5.8: van 2858: A..\.. V 32320:? AX. V 30:90”—
ro :3 :0» oc ~590qu V95? .Qeowazamfi B2933; 2:st 25»: 3:53 .3 808930 So» 50% so» xmm 9 95 2:03 _ .w.m

 

 

:53
”some Bo: >m
”5:3
QEB 98 £03: Bo: 3 .owcmno : 80¢ :23? 4.5m
{$5.63 :

 

l

................ AOWCNSU H N «DENW H # v
mag 2: V0 8:5 :58wa Va owcwco 8E 2: 30¢ S 58» 2: 505.1385 coca 2:3 2: 8V Esfimficzgmx 32833 :3 so» 0D Sm

 

moumxommd VEBU 3V
8322: V280 3V
woumV—UNAMA V\mmfiou AQV ......................................................o.m\.HmOU ~wmu~®HNE wCMwmxomQ UF—H WUOU £038 30E .N.©.m

 

 

265

 

$8958 wEmz ¢on so» Ev £3 .N.m_.m
.gmo>: .=

 

so: 23 :2
.223:

III .................................... A mm; H V .oZ H C V AhmoooV woman”; So» E 0568?: on 3 98858 com: 85 so» 96: .m_.m

 

.5598 2: E 68.. wfiuaot »=a 333.:— 3 com: H02 a...“ 23,933 .=

 

 

fiacosms 2: 2“ EB._.N.~_.m
...3»..=

....... A mo> H _ 62 n C V mmcmonm»om mar; $338 mmoooa ow om: so» 85385 of E moocobtfio »:m 22: 8< .N.N_.m

 

 

mmUOOw wcmfiwva uSQA OOSUOHQ 0H mCdDDQkOm UCM mwvflgoo «.303 Om: 30% mCOmdUu 0:”— Oum “$33 _.N—.m

 

52:5...— 28 E »:..ammooo: 35 552“.— 23 :_ £58 win-33 335.:— 3 tom: =P0m a...“ 23213:; 6.3 23953 .=

 

 

EB .2 2
I ............ A265. :50 H m 629502“— H v .owcmno :65 H m 45,3280 H N #53855 H _ V
mega—«03¢

newcmco Ho: .ommooooo 69885 232—«How wEEmEoo mooom wESBB mo :26:on So» En £23» 3...: “man 2: 5 .A _.m

 

 

$5.05.— ofi E toe mam—=33 »=w 335.3 3 com: a...“ 252—.33 .=

266

.32285 2: 1:0 ES ...mooumoov »E motes?“ “on :o» 452;: »am antic—=0
.2.—3.895 9.2: m_ 533 «no a... ANNA—V so: .5 Am—dV :3. .353 3 ow £32? 253598.. 2: .: _Ea Jo» «not .o: .=

 

 

 

 

Scflooowfi
on mm @0938 £15 woom mean?» a 356th 9 Bfiabmcoo “Boom ,mowficgoaflo “muwmfigom 05 8a x55 so» oo :33 .m V .m

 

wmcmoowom wEm: @on so» Bo 33 NE .m
.amo>: .2

 

so: 33 .33
.22.. =

................................ A we» H _ 62 H AV V mmooow mica?» So» E Eozoohwfi on ma £5338 wow: ~96 so» 02% .36

 

£56.“.— 2: 5 zoom wE—SPS »=a 333.:— o. .52. HOZ 9:“ mason—«Now .=

267

 

”>55
hose Bo: mm
”:25,
Q33 can £02: Bo: 3 ,owcaso : 30v 5:3 466

{$5.61.

......................... A DWCN:U H N nogm H _ V
23% 2: go 385 “:80“va 3 $550 330:3 2: meow 5 fiEoE >55 33% .3333; go 539% 08mm of 02605 no» 0Q .vd

 

 

 

 

ESE _ A sea
3:33— : 5E3 TTTTT Ecsfiwx _ A ”3:3
E53 _ A gas

59:0 ANA— .e.—.O 83 8:33qu

@5990 so» 58 825E 05 mo
:02: >6: ,mucomnewfi mo EsoEa 35 80¢ can 33% @2333 026th so» 2:: :23 3: so» ov 333%.: Sumo 90 50:8 302 40:0

I I III. II I II .............................................................................. xh—H—A—a Han—H ——N hub—um
A ”3625 .25 u o ,3 m afiEm u v 5%; u m 692 u N ,éowmz u _ V
2.29% @2933, 23% 2:32 .6233 8:2: 2 om: so» on $3958 («0 but? 523 No
.3333 323:3 tau 688—8 2: ..—

 

...................................................................

mwtsm nmtfimfi @6353 93m 5:63 So» Eu :3» :23 5 ._.©

2.33% .3283; go 20:26th SPA Sosa 28525 08% x8 2 8E 2:03 _ 302: Sam

 

E<U GHPUmAm—m ”al.—L hO ZO—FUDA—Omm .w

268

 

atoommv 550 H v 59$ H m :80 H N ,wan wcoREPA—om H _

 

 

 

 

 

 

 

 

 

 

IIIGV ”IIIISV mllulAmv ........ Cl— .c—AO 83 Myriam @2333 :3 8 8: so» 55 Amvowwxowa 05 mo 3888 05 mm :33 4.0.0
d3; D EE D 8:
fine: D £8 D g
duo: D 58 D 93
go; D 58 D 6
C30: D 58 D A8
A50: D EE D Amv
C30: D EE D Go
duo: D 58 B CD
duo: D EE D Amv
C30: D EE D A:

:95: 2: :33. 8:3 .9on some how 330: 3:8me

 

2:: 05 0:2:th 98 ¢on 3 ¢on 52qu 885 $3M 32,333 335.3 59m 2:: ...—23 30:8 :0» $530th 2: 8a 355 .m6

269

 

 

 

 

 

 

EB .32

............ T505. {50 u m “BEBE—m n v .owawno C65 H m 6332er H N .3882: H V V
mgflzosc ho newsman. Ho: “838% .8385 g wEEchoo cam Vo cocosuoa Box Ev .25?» 99:: ~22.— 2: E .26

 

 

£53.“.— ofi E 6.59:— _._aw has 335.:— 3 new: a...“ 33958 .=

 

mSHEOE ............................................................ mezemxme Ha TKBMVmNUmemQ UVOHm 30% GNU wav— 30: ...—00
0.98%.. .=

....A 07:38 .8 bon—V ma 2% H m «8on H V V $5682 ma 3x015 mm 6:on 2: :3 so» ow Ho HEM 33939; 2on :0» OD .06

 

g V fix. V ”3:8qu 350
Ag V 82533 3322ch AR. V mVafisfimom 6x. V whozfiom 6X. V 3282055
“8 :8 so» ow comtoaoa 355 ._gm @29me 22st 253 Each: Vo £05930 So» Sosa so» xmm 8 BE 2:03 V .w.©

 

“>55
”52:: Bo: >m
”55$
Q33 95 £38 >20: 3 “owcmco z meow 5:3 .26

...owEEU: :
ll ................ A Uwcmfiu H N aDENW H H V

Eng 2: V0 38: 380%? Va owcmzo 8E of meow .5 d3» 05 Soswsefi 8E 25% 2: 8V Tam 32.833 :3. so» OD .50

 

 

 

wmmfioa A $68 3
3&3an :2on 3V
amxomafl V\mmwvflv va u...-u.-n...

.........................................AVHWOU flmmhoHaE mamwNV—QNQ 0:“ moov SCH—E Bo: .N.©.©

270

 

 

$3853 wEm: no; so» 20 .33 4.2.0
v.88»: .=

II ................................................. A m0> H — ooz H o V Numhww HDOxA Cm Hcomfivohwcm cm MN WNUQBOU “Vow: hv>o 30% U>NE .m—.©

 

00:35.. 2: E tau .35 333.3 3 ....8: POZ 9.8 83933 .=

 

 

 

mmooceotso 05 2a was? ._.m.m_.0
aa¢m0>3 “—

....... A mo> H V .oZ H o V mmomonwxom 83803 83958 838m 9 8: so» 3538a 35 E 338806 >3 82: PE .N.NV.0

 

 

pram So» 323E 8 823038 03 83958 Son 8: so» 8:8on of 8m 30>? .V.N_.0

 

A335...— oao E 3:833: 35 .352». 2: 5 flow 335.:— 3 08: Fix—FPS 230:8 can 83433 =

 

 

...»;3._.VV.0

............. A 255— tcoQ H m .3833; H v .owcmco “EEO H m .08380D H N .08335 H V V
moVwBooc Ho dwcmso Voc .8333 .8385 masonawom wEEScS tow .Vo .5350th Box 20 .mp3» 3:: 8mm 2: E .V _.0

£56m.— 2: E .2.—5.:— muaw ES 335.... 3 08: «.3 8:30“ 8 .=

271

$9232.: 2: .25 .23 ...25526 >8 maroamca .8 so» 235.: mam .om_Eo.:O
...—5395 9.2: mm :03? 2:. 2: A52: am: .5 93.3 :3. .553 3 cw £33..“ 2.22.2.8.— 2: .= .23 do» 25: 22. u—

 

 

 

 

wEoGEwE cw mm $0958 53» Cum a “.36:on 9 3:65:00 SEE“ .muwflcggflc .muwficazum 2: 2m 25: so» ow :25» 6:0

 

 

$283an mag 85 so» EU 33 4.3.0
...mo>s .=

............................................. A m®> H w .02 H O v btww h50% am ucvmfiohwcm Cw mm m26®£§m0m Cum: uo>® 30% 2rd: ..V—C

 

£535.. 2: E 2% an.“ 02:52. 8 NW3: POZ a...“ 2.33%.; .=

272

A.l.3 Questionnaire for Custom Millers

273

Competitiveness of Cowpea-Based Processed Products:
A Case Study in Ghana

- Questionnaire for Custom Millers —

 

 

 

Enumerator: Respondent Number:
Date: /2007 Area in Accra
Time Interview Began: : Time Interview Ended:

 

 

INSTRUCTIONS: Read the consent statement to the respondent. If she/he agrees,
begin the interview.

(Revised Mar. 3, 2007, final version)

274

1.1. In what year did you start your milling business? .............................

 

1.2. What do you mill? Check all that apply.

1.3. How often do you mill each? Ask only about cowpeas, soyabeans and maize.
1.4. How much do you charge for milling one olonka of each?

1.5. How much would you charge if your customer bring one olonka of [(a) to (6)]?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.2 1.3 1.4 & 1.5
Grains/Dough to mill Frequency Fee (Cedis)/olonka
Cl Cowpeas (dry)
[I Cowpeas (wet)
D Soyabeans (dry)
I] Soyabeans (wet)
[I Maize (dry)
D Fermented maize (wet)
I] Fermented maize dough (dry)
Cl Millet (dry)
E1 Millet (wet)
D Cassava (dry) [kokonte]
Cl Cassava (wet)
j C! Groundnuts
E1 Spices
C! Other: _ El dry [3 wet
‘ C] Other: _ E] dry El wet
C] Other: _ El dry [I] wet
(a) [Fermented maize] +-Cowpea (wet)
(b) Fermented [maize + cowpea] dough (dry)
(0) Roasted maizgdw)
(d) RoastedLnaize + cowp_eas + grgundnuts] (dry)
(e) Roasted [maize + soyabeans +figroundnuts] (dry)

 

 

If the fees are NOT the same between dl_'y and wet,
1.6. Why do you charge different fees for dry and wet product?

 

 

 

If the fees are NOT the same between different grains,
1.7. Why do you charge different fees for different grains?

 

 

 

 

 

275

If the respondent does only wet milling of cowpeas,
1.8.1 Would you be willing to do dry milling of cowpeas if your customers request?
(O=No, 1=Yes) .......

_-

 

If “Yes”,
1.8.1.1. How much would you charge for one olonka of dry cowpeas?
....... Cedis
If “N0”,

1.8.1.2. Why not?

 

 

If the respondent does only dgy milling of cowpeas,
1.8.2 Would you be willing to do wet milling of cowpeas if your customers request?
(O=No, 1=Yes) .......

 

If “Yes”,
1.8.2.1. How much would you charge for one olonka of wet cowpeas?
....... Cedis
If “N0”,

1.8.2.2. Why not?

 

 

1.9. If your customers bring a large quantity of grains or dough, do you usually offer a

 

 

discount? ( 0 = No, l = Yes ) ....... __
If “Yes”,
1.9. 1. Could you give some example of such discount?
What to mill:
How much to mill: olonka
How much (charge): C edis

 

1.10. 1 would like to ask you about the customers who bring you cowpeas for milling.
What prOportions are:

[:1 Kosei vendors( %), El Individuals( °/o), Cl Restaurants( %)
[:1 Otheris‘rgcifj): ( ”/oL ( %)

1.1 1. How much cowpeas do you typically mill per day?
(Dry milling) .............................. olonka
(Wet milling) .............................. olonka

1.1 1.1. Does the amount of cowpeas you mill change depending on the seasons of
the year? ( O = No, l = Yes ) ...... _
If “Yes”,
1.1 1.1.1. When does it change, by how much, and why?
When:
By how much:
Why:

 

 

 

276

1.12. Which grades do you normally mill cowpeas for customers?
( 1 = very fine, 2 = fine, 3 = coarse, 4 = very coarse)
(Dry milling) ......................................................................... __
(Wet milling) ....................................................................... __

1.13. Is it easy to mill cowpeas to different flour grades? ( O = No, 1 = Yes ) ........... __

If “N0”,
1.13.1. Why?

 

 

 

1.14. What do you have to do when you switch from milling cowpeas to milling other
things, and vise-versa?

(from fly cowpeas to others)

 

 

(from v_ve_t cowpeas to others)

 

 

(from fly others to cowpeas)

 

 

 

(from w_et others to cowpeas)

 

1.15. Is the machine you use to mill cowpeas a plate mill, hammer mill or other? ......

1 = Plate mill, 2 = Hammer mill, 3 = Other (specify):

 

1.16. What is the brand of the milling machine? ......................

 

1.17. Is your mill powered by an electric motor or diesel engine?
( 1 = electric motor, 2 = diesel engine) ................................................

1.18. What are the horsepower and capacity per hour of the milling machine?

........... L Lhorsegower; ( ) [:1 kg D ton of (product: )/ hour

1.19. In what year did you buy the milling machine? ................................

 

1.20. From whom did you buy the milling machine? ......
1.21. How much did it cost? ...................... Cedis
Say, “Thank you for answering my questions.”

[END OF INTERVIEW]

277

A.l.4 Questionnaire for Retailers

278

Competitiveness of COWpea-Based Processed Products:
A Case Study in Ghana

- Questionnaire for Retailers —

 

 

 

Enumerator: Respondent Number:
Date: /2007 Area in Accra
Type of Retailer .................................................................................

1 = Grain/flour vendor in a market, 2 = Small shop in a market,
3 = Small shop located other than in the market, 4 = Supermarket
5 = Other (specify)

 

Time Interview Began: : Time Interview Ended:

 

INSTRUCTIONS: Read the consent statement to a retailer. If she/he agrees, begin
the interview.

(Revised Mar. 3, 2007, final version)

279

l. COWPEA/SOYABEAN FLOUR
1.1. In the past year did you sell cowpea or soyabean flour? ( 0 = No, l = Yes )......___

If “No”, go to “2. WEANING FOODS” (p. 4).
If “Yes”, continue.

1.2. What are the brand and company names of those products?
For small local firms without a brand or producer name, ask for the
firms’ location. Enter the answers in the table below.

1.3. Do you have each of these products on the shelf throughout the year?
Check the answer in the table below.

For the brand(s) “No”,

1.3.]. In what months do you have shortage and why?
1 Put the brand number(s) for which the same answers were given.

 

 

 

 

 

( ) Months:
Why:
( ) Months:
Why:
( ) Months:
W by:

 

1.4. In the past month, which one did you sell the most? Which one did you sell the
second most? Rank the products.

 

1 .2 l .3 1 .4
Year round Sales
availability rank

UNO DYes
DNo DYes
DNo DYes
DNo DYes

 

Brand Producer/Location

 

 

 

 

 

MAWN—

D No [:1 Yes

 

 

 

 

 

 

1.5. From whom do you buy each of these products?

1 = Producer, 2 = Wholesaler, 3 == Grain/flour vendor in a market,
4 = Other (specify):

1 Brand number

Q) 12) (31 (4) ( 5)

 

280

 

1.6. How much do you w for each of these products when you purchase them?
If the respondent gives the price in a unit other than kg, ask how many kg that
unit contains. Enter the answers in the table below.

1.7. Are there seasonal changes in the prices you p_ay to buy these products?
( O = No, l = Yes ) Circle the code in the table below.

If “Yes”,

1.7.1. In what months does the price change, and by how much?
1 Put the brand number(s) for which the same answers were given.

( ) Months:
By how much:

L ) Months:
By how much:

( ) Months:
By how much:

1.8. At what price do you £11 each of these products?

1.9. Are there seasonal changes in the prices at which you §e_l_! these products?
( O = No, 1 = Yes ) Circle the code in the table below.

If “Yes”,
1.9.]. In what months does the price change, and by how much?
1 Put the brand number(s) for which the same answers were given.

( ) Months:
By how much:

 

 

 

 

 

 

 

 

L 9 Months:
By how much:
( i Months:
gyjow much:
1.6 1.7 1.8 1.9
Purchase , Selling ,
(32"?) W ”n“ 152?“ i212”. Price Unit if. 322?".
p' Qedisy g LCedis) g
0 . 1 0 . 1
O . 1 O . 1

 

 

 

 

MAWN~

01
0.1
01

 

 

 

 

 

 

 

 

 

 

281

 

1.10. I’d like to ask you about your customers who buy each of these products. What
proportion of [brand I] do you sell to: Restaurants? Caterers? Individuals?
Other customers? Put the percentage in the table below.

Repeat the same questions for each brand.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.10
Customellc/o
v:
5’ '3
Brand (see p.1) 5 if _g "‘_ 3:, 3,,
a 3 '5 a: a: a
Q) ....
in: *5 “D 5 E. '5
a: U E. o o o
1
2
3
4
5
*Other 1 (specify): ; 2: ; 3:

 

 

If the respondent sells both cowpea and soyabean flour,

1.1 1. Do you sell COWpea flour more or less than soyabean flour? Please select from
the following choices. “COWpea flour sells: .....................................

1 = Much more than soyabean flour,
2 = A little more than soyabean flour
3 = The same as soyabean flour,

4 = A little less than soyabean flour
5 = Much less than soyabean flour

282

2. WEANING FOODS

2.1. In the past year did you sell Tom Brown, Weanimix, any other maize-based flour

type products, or maize dough? ( 0 = No, 1 = Yes ) ................................

If “No”, go to “3. GARI”. (p. 7)

If “Yes”, continue.

2.2. What are the brand and company names of those products?

For small local firms without a brand or producer name, ask for the firms’
location. Enter the answers in the table below.

2.3. Do you have each of these products on the shelf throughout the year?
Check the answer in the table below.

For the brand(s) “NO”,

2.3.]. In what months do you have shortage and why?
1 Put the brand number(s) for which the same answers were given.

 

 

 

 

 

( ) Months:
Why:
( ) Months:
Why:
( ) Months:
Why:

 

2.4. In the past month, which one did you sell the most? Which one did you sell the
second most? Rank the products.

 

2.2

2.3

2.4

 

Brand

Producer/Location

Year round
availability

Sales
rank

 

[I] No C] Yes

 

E] No [3 Yes

 

D No D Yes

 

E] No 1:] Yes

 

M-RWN—e

 

 

 

 

II] No 1:] Yes

 

 

2.5. From whom do you buy each of these products?

1 = Producer, 2 = Wholesaler, 3 = Grain/flour vendor in a market

4 = Other (specify):

 

1 Brand number

(1)

42)

(3)

283

(4)

(5)

 

 

2.6. How much do you w for each of these products when you purchase them?
If the respondent gives the price in a unit other than kg, ask how many kg that
unit contains. Enter the answers in the table below.

2.7. Are there seasonal changes in the prices you p_ay to buy these products?
( O = No, 1 = Yes ) Circle the code in the table below.

If “Yes”,

2.7.]. In what months does the price change, and by how much?
1 Put the brand number(s) for which the same answers were given.

L ) Months:
By how much:

( ) Months:
By how much:

( ) Months:
By how much:

2.8. At what price do you s_efl each of these products?

2.9. Are there seasonal changes in the prices at which you ggfl these products?
( 0 = No, l = Yes ) Circle the code in the table below.

If “Yes”,

2.9.]. In what months does the price change, and by how much?
1 Put the brand number(s) for which the same answers were given.

( ) Months:
By how much:

 

 

 

 

 

 

 

 

L ) Months:
flhow much:
( LMonths:
BLhow much:
2.6 2.7 2.8 2.9
Purchase , Selling ,
ZZZ“? ...... U... 42?. 321?“. ...... Unit “if. 5213.3".
P- (CedisL g (Cedis) g
0 . l 0 . 1
O . 1 O . 1

 

 

 

 

MAWN—t

 

 

 

 

 

 

 

 

 

 

284

 

Ask only to grain/flour vendors in a marke ,

2.10. I’d like to ask you about your customers who buy each of these products.
What proportion of [brand 1] do you sell to Individuals? Retailers?
Other customers? Put the percentage in the table below.

Repeat the same questions for each brand.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.10
Customer o/o)

Brand (see p.4) ,5 g ”... it: :0

E E 2: s 2:

1: 8 5 5 5

E a: o o o
1
2
3
4
5

*Other 1 (specify): :2: :3:

 

 

 

285

3. GARI

3.1. In the past year did you sell gari? ( O = No, 1 = Yes ) ................................. __

If “No”, Say “Thank you for answering my questions”, and end the interview.

If “Yes”, continue.

3.2. What are the brand and company names of those products?

For small local firms without a brand or producer name, ask for the firms’
location. Enter the answers in the table below.

3.3. Do you have each of these gari brands on the shelf throughout the year?
Check the answer in the table below.

For the brands “No”,

3.3.1. In what months do you have shortage and why?

1 Put the brand number(s) for which the same answers were given.

 

 

 

 

 

( ) Months:
Why:
( ) Months:
Why:
L ) Months:
Why:

 

3.4. In the past month, which one did you sell the most? Which one did you sell the
second most? Rank the products.

 

3.2

3.3

3.4

 

Brand

Producer/Location

Year round
availability

Sales
rank

 

I] No [I] Yes

 

D No [3 Yes

 

E] No D Yes

 

E] No D Yes

 

Lh-bWNH

 

 

 

 

C] No D Yes

 

 

3.5. From whom do you buy each of these gari brands?

I = Producer, 2 = Wholesaler, 3 = Grain/flour vendor in a market,

4 = Other (specify):

 

1 Brand number

(IL

12)

(31

286

(4)

(5)

 

3.6. How much do you pay for each of these gari brands when you purchase them?
If the respondent gives the price in a unit other than kg, ask how many kg that
unit contains. Enter the answers in the table below.

3.7. Are there seasonal changes in the prices you my to buy these products?
( O = No, 1 = Yes ) Circle the code in the table below.

If “Yes”,

3.7.1. In what months does the price change, and by how much?
1 Put the brand number(s) for which the same answers were given.

( ) Months:
By how much:

L ) Months:

By how much:

L ) Months:
By how much:

3.8. At what price do you s_efl each of these gari brands?

3.9. Are there seasonal changes in the prices at which you fill these products?
( 0 = No, 1 = Yes ) Circle the code in the table below.

If “Yes”,

3.9.1. In what months does the price change, and by how much?
1 Put the brand number(s) for which the same answers were given.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

( ) Months:
By how much:
( ) Months:
By how much:
( 1 Months:
By_how much:
3.6 3.7 3.8 3.9
Purchase , Selling ,
(32mg) Pm" unit Iii?“ 2:15;): Pm" ”n" Itiiigit 221?:
P- LCedifl g (Cedis) g
1 O . l O . 1
2 O . 1 O . 1
3 O . 1 0 . 1
4 O . 1 0 . l
5 0 . 1 0 . l

 

 

 

 

 

 

 

 

 

 

287

 

Ask only when the respondent sells either cOWpea- or soyabean-fortified gari,
3.10. I’d like to ask you about your customers who buy each of these gari brands.

What proportion of [brand 1] do you sell to Restaurants? Caterers?
Housewives? Other customers? Put the percentage in the table below.

Repeat the same questions for each brand.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.10
Customers (%)
Brand (see p.6) E e g 4;, 3:. *m
‘5 2 E I... s. i...
i it =5 .53 E E
32 u .E o o o
1
2
3
4
5
*Other 1 (specify): ; 2: :3:

 

 

 

Say, “Thank you for answering my questions.”

[END OF INTERVIEW]

288

A.l.S Questionnaire for Weaning Mothers

 

 

289

Competitiveness of Cowpea-Based Processed Products:
A Case Study in Ghana

- Questionnaire for Weaning Mothers —

 

 

 

Enumerator: Respondent Number:
Date: /2007 Area in Accra
Time Interview Began: : Time Interview Ended:

 

 

INSTRUCTIONS: Read the consent statement to the respondent. If she agrees, begin
the interview.

(Revised Feb. 20, 2007, final version)

290

 

l. WEANING FOODS

1.1. Do you give your baby weaning foods, that is, special foods easy for babies to eat?

If “Yes”, continue.
If “No”, go to “2. KOSEI”.

1.2. How many children do you have, how old are they, and to which child do you give
weaning foods?

Write down the age specifying month(s) or year(s).

Check the box if the child is weaned.

Age
Weaning [:1 E1 _ C] D D

 

1.3. How often do you give weaning foods to your child? .................................
1 = 3 times or more/day, 2 = Twice/day, 3 = Once/day, 4 = Less than once/day

1.4. What are the characteristics of the weaning foods that you prefer? Please rate each of
the following criteria by using the number 1, 2 and 3: the 1 means “Very important”,
2 means “Somewhat important”, and 3 means “Not important”.

1 = Very important, 2 = Somewhat important, 3 = Not important

(1) Easy for you to prepare ........................................................
(2) Cheap ...........................................................................
(3) Nutritional value is high .......................................................

(6) Easy for your child to digest ..................................................

(5) Your child likes the taste ......................................................
(7) Your mother or grandmother prepared in the same way ..................

1.4.1. Are there any other characteristics of weaning foods that you think are
important?

 

 

 

1.5. Do you prepare your own weaning foods, or do you buy ready-to-use commercial
weaning foods, or do you use both types? ..............................................

1 = Always self prepared, 2 = Always commercial, 3 = Both

If “Always self prepared” or “Both”, go to the next page.
If “Always commercial”, go to p. 3.

291

Questions for “Always self prepared” and “Both”:

1.5.1.1. Ask only if “Always self prepared”,
Why don’t you use commercial weaning foods? .................. ____ _ __
Enter all that apply.

1 = Your child doesn’t like the taste of commercial weaning foods
2 = Commercial weaning foods are expensive
3 = Other (specify):

 

 

1.5.1.2. To make weaning foods. do you use: Check all that apply.

[:1 Maize D Sorghum C] Millet El Rice DCassava E] Yam
C1 Cowpeas Cl Soyabeans [:1 Groundnuts [:1 Bambara beans

Cl Milk [:1 Egg [:1 Fish El Meat

C] Other (specify):

 

1.5.1.3. What are the relative amounts of those ingredients in the weaning foods that
you prepare? If you prepare different types of weaning foods, please begin

with the one you prepare the most often.
Enter the answers in the table below.

[Codesz L = Lar e amount; M = Medium amount; S = Small amount

 

 

 

 

 

m
E a a 5 ..
5 > 0 E
E ‘50 E a Q 3% g 3 a: "’
a 15 :=: "3. g E E >. I? .n = g i 8
E m E a: u > u (S 0 cl: 2 as E: 2
Formula]
Formula2
Formula3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.5.1.4. How do you prepare [Formula 1]? Please begin with the materials you

LX-
11)

 

 

 

 

292

 

 

If the respondent prepares dough or Tom Brown from raw materials,

1.5.1 .4.1. How often do you prepare dough/Tom Brown?

 

 

1.5.1.4.3. How long does it take to prepare dough/Tom Brown each
time? ..........

 

1.5.1.4.4. How long can you keep dough/Tom Brown before it goes
bad? ................................................................. days

1.5.1.5. From whom did you learn how to prepare weaning foods?... __ _
[Enter all that apply.

1 = Mother, 2 = Health worker, 3 = Other (specify):

 

Questions for “Always commercial” and “Both”:

1.5.2.1. Ask only if “Always commercial”,
Why don’t you make weaning foods by yourself? ........... _ ‘ ___
Enter all that apply.

 

1 = It takes time to prepare by yourself
2 = Your child prefers the taste of commercial weaning foods

3 = Commercial weaning foods are more nutritious
4 = Other (specify):

 

(Continue to next page)

293

1.5.2.2. What brands of weaning foods do you buy? If you buy different brands,

please begin with the one you buy the most often.
Enter the answer in the table below.
If the answer includes products of local small firm without a brand

name, ask the respondent to describe the products.

1.5.2.3. Where do you buy each of these weaning foods, and how much does each
cost? Enter the code and answer.

1 = Market, 2 = Small store, 3 = Supermarket,
4 = Other (specify)

 

1.5.2.2 1.5.2.3
Brand/Descri tion of weaning foods Where Price (Cedis)

 

1.5.2.4. Do you add any other ingredients to any of the commercial weaning foods
you buy before you give them to your child? ( 0 = No, l = Yes ) ..........

If “Yes”,
1.5.2.4.1. What do you add to each commercial weaning foods, and why?

(1) Add:
Why:

(2) Add:
WM:

(3) Add:
WILL

If “YES”,
1.5.2.5. 1.1. Where did you buy? ......

 

1.5.2.512. How often do you buy? ......

 

If“No”,
1.5.2.5.2. Why not?

 

 

294

 

1.5.2.6. Have you ever bought weanimix? ( 0 = No, 1 = Yes ) ....................

If “Yes”,
1.5.2.6.1.1. Where did you buy?

 

1.5.2.612. How often do you buy?

 

If “N0”,
1.5.2.6.2. Why not?

 

 

Questions for “Both”:

1.5.3.1. How often do you use commercial weaning foods compared to weaning
foods that you prepare by yourself? .............................................

1 = Most of the time, 2 = More often than self-prepared

3 = As often as self-prepared, 4 = Less than self-prepared, 5 = Rarely

1.5.3.2. Does the answer change depending on the seasons of the year?
(0 = No, l= Yes)...

 

 

If “Yes”,
1.5.3.2.1. When does it change, by how much, and why?
When:
By how much:
Why:

 

295

 

2. KOSEI
Say, “Now I would like to ask you some questions about foods that you prepare.”

2.1. How often do you prepare a meal with cowpeas as an ingredient?. ..

 

2.1.1. Does the answer change depending on the seasons of the year?
(0 =No, 1= Yes)...

 

 

If “Yes”,
2.1.1.1. When does it change, by how much and why?
When:
by how much:
Why:

 

2.2. What dishes do you prepare using cowpeas? Please begin with the one you prepare
the most often.

L1)
(2)
(3)
(4)
(5)

 

2.3. Do you eat kosei? ( O = No, l = Yes) ...................................................

If “Yes”,
2.3.1. How often do you eat kosei? ............................

 

2.3.2. Does the answer change depending on the seasons of the year?
(0 = No, 1= Yes )......

 

 

If “Yes”,
2.3.2.1. When does it change, by how much and why?
When:
Bihow much:
Why:

 

2.4. Do you know how to prepare kosei? ( O = No, 1 = Yes ) .............................

If “No”, go to 2.5.

If “Yes”,
2.4.1. Do you prepare kosei at home? ( 0 = No, 1 = Yes ) ...................

296

If “Yes”, go to 2.4.2.
If “N0”,
2.4.1.1. Why not? Enter all that apply ..............

 

1 = It’s tedious to make, 2 = It’s cheap to buy
3 = Other (specify):

 

 

 

2.4.2. If a commercial cowpea powder was available that you could use to
easily prepare good quality of kosei, would you be interested in buying it
and make kosei at home? ( O = No, 1 = Yes ) ...........................

[f “N0”,
2.4.2.1. Why not?

 

 

2.5. Have you ever bought cowpea flour? ( 0 = No, 1 = Yes ) ............................

If “Yes”,
2.5.1 .1. How often. do you buy cowpea flour? ............

 

2.5.1.2. Where do you usually buy cowpea flour? ..............................

1 = Market, 2 = Small store, 3 = Supermarket,
4 = Other (specify)

 

2.5.1.3. What do/did you use it for?

 

 

 

 

If “No” or “Once bought but not any more”,
2.5.2. Why not?

 

 

 

297

3. CAR]

3.1. Do you eat gari? ( 0 = No, 1: Yes) ...................................................

If “Yes”,
3.1.1. How often do you eat gari? ........................

 

3.1.1.1. Does the answer change depending on the seasons of the year?
(0 = No, 1= Yes ).......

 

 

If “Yes”,
3.1.1.1.1. When does it change, by how much and why?
When:
By how much:
Why:

 

3.2. What dishes do you prepare using gari, when you prepare foods for your family?
Please begin with the one you prepare the most ofttap.

(D
(2)
(Q
(4)
L5)

3.3. Do you make gari from cassava yourself? ( 0 = No, 1 = Yes ) ..........................

3.4. Where do you usually buy gari? ..........................................................

1 = Market, 2 = Small store, 3 = Supermarket,
4 = Other (specify)

 

 

3.5. How much do you usually pay for 1 olonka ofgari?.................. Cedis

3.6. Let’s assume that you found two different types of gari on the shelf in the store. One
is normal gari that you always buy. The other one has a label which reads “New

Gari — Supplemented by Cowpea —— More Nutritious”. Do you think you would
want to try this new product if the price is the same? ( 0 = No, l = Yes ) ....... __

298

If “NO”,
3.6.]. Why not?

 

 

 

If “Yes”,
3.6.2. If thflrice of the new product was higher than the price of
gari you usually buy, how much more would you be willing to
pay for the new product? ...... Up to Cedis of increase

3.7. Let’s assume that you found two different types of gari on the shelf in the store. One
is normal gari that you always buy. The other one has a label which reads “New
Gari — Supplemented by Soyabeans — More Nutritious”. Do you think you would
want to try this new product if the price is the same? ( 0 = No, 1 = Yes) ....... __
If “No”,
3.7.1. Why not?

 

 

 

If “Yes”,
3.7.2. If the price offithe new product was higher than the price of
gari you usually buy, how much more would you be willing to
pay for the new product? ...... Up to Cedis of increase

Say, “Thank you for answering to my questions.”

[END OF INTERVIEW]

4. NUTRITIONAL STATUS OF THE CHILD

Observe the respondent’s child. Does she/he show an apparent symptom of
malnutrition? ( 0 = No, 1 = Yes) ....................................................

299

 

 

APPENDIX 2

 

Volume-Weight Conversion Rates

Olonka-kg conversion rates obtained from the samples of different commodities are
presented in Table A2]. The samples were purchased and weighed as described in
Chapter 3. The ratios of “heaped up” olonka to kg (figures in the right column in Table
A.2.1) were used throughout this study when the unit conversions were necessary

between olonka and kg.

Table A.2.1 Olonka-kg conversion rates

 

 

 

. kg/olonka kg/olonka
Commodity (flat to cup) (heaped up
Cowpea* 2.2291 2.5094
Gari 1.7174 1.9782
Groundnut 1.9803 2.2426
Maize 2.3043 2.5344
Millet 2.3550 2.6504
Onion 1.6037 3.1123
Pepper 0.6193 0.8642
Sorghum 2.3666 2.6878
Soybean 2.2377 2.5039
Tom Brown 1.4406 1.7080

 

 

 

 

* type called “Niger”

Source: (all except Tom Brown) samples purchased at Nima Market, Accra, on March 23,
2007; (Tom Brown) G. A. Annor, personal communication, November 6, 2007;
samples were weighed at a laboratory of the Department of Nutrition and Food
Science at UGL.

To measure the volume of fermented maize dough, vendors commonly used a measuring
cup called the “rubber cup,” which was different in size from the olonka cup. The rubber
cup-kg conversion rate obtained from the sample of fermented maize dough is presented
in Table A.2.2. This rate was used to derive the price per kg of fermented maize dough
sold by the grain/flour-type product vendor respondents (see Section 6.2.1).

Table A.2.2 Rubber cup-kg conversion rate

 

kg/rubber cup kg/rubber cup
(flat to cupL (heaped up)
Fermented maize doughi 3.2998 6.8033

Source: Sample purchased at Nima Market, Accra, and weighed at a laboratory of the
Department of Nutrition and Food Science at UGL, on March 23, 2007.

 

 

 

 

 

300

For the prices of raw materials, the same representative prices calculated for each
commodity were used across the industrial processor respondents, whether the
respondents purchased raw materials from markets or middlemen or grew them on their
own farms. Also, when the kosei vendor respondents could only recall the monetary
value of onion, pepper, and/or salt that they used, the weight of these ingredients was
estimated using their representative price per kg. The values and sources of these

APPENDIX 3

Representative Prices of Raw Materials

representative prices are summarized in Table A3. I.

Table A.3.1 Representative price per kg of raw materials

 

 

Raw
material

Rep.
price/kg
(¢)

Source

 

Cowpeas

5,474

See Section 4.2.5.

 

Groundnuts

7,650

See Section 6.2.6.

 

Maize

3,000

See Section 6.2.6.

 

Millet

4,450

Median of ¢4,800 per kg at Makola market (commodity classified as “millet
(sanio, grain)”) and ¢4,100 per kg at Tema (commodity classified as “millet
(sounna, grain)”) in February 2007, obtained from http://www.tradenet.biz/
(retrieved October 1, 2007). These two locations and commodity
classifications were selected because the price data obtained were close to the
wholesale price data (urban average) obtained from the MoFA up to
December 2006.

 

Onion

7,013

Median of: (I) ¢20,000 per olonka observed by the author at Nima market on
March 23, 2007, converted to price per kg using the olonka-kg conversion
rate of 3.1 123 (see Appendix 2); (2) ¢25,000 per olonka observed by the
author at Madina market on March 27, 2007, converted to price per kg using
the olonka-kg conversion rate; and (3) ¢7,6OO per kg at Makola market
(yellow variety) and ¢3,750 per kg at Agbogbloshie market (violet variety) in
March 2007, obtained from http://wwwtradenet.biz/ (retrieved July 13,
2007).

 

Pepper

32,545

Mean of: (1) eight prices per olonka each reported by eight kosei/agawu
vendor respondents, respectively; (2) ¢30,000 per olonka observed by the
author at Nima market on March 23, 2007; and (3) ¢10,000 for 0.5158 kg
observed by the author at Madina market on March 27, 2007. (Prices per
olonka were converted to prices per kg using the olonka-kg conversion rate of
0.8642Jsee Appendix 2].) (standard deviation = 7,066.)

 

Salt

5,126

Mean of 19 prices per kg each reported by 18 kosei/agawu vendor
respondents, respectively (one of the respondents reported 2 different prices
for different p_urchase units). (standard deviation = 1,007.)

 

 

Soybeans

 

4,069

 

See Section 6.2.6.

 

 

301

 

APPENDIX 4

Notes on the Calculation of Representative Budgets
for Dry Cowpea Meal and Tom Brown/Weanimix

The data collected from the industrial grain flour/weaning food processor respondents
needed modifications to calculate budgets for producing dry cowpea meal, Tom Brown,
cowpea-Weanimix, and soybean-Weanimix, because no respondent was actually
producing all of these products. In other words, the data had to be adjusted to reflect the
difference in the cost of production between our target products and the products for
which the data were collected. Also, some values needed to be assumed or estimated
because of missing or apparently erroneous information. The main methods used for
these modifications, assumptions, and estimations were noted below (except those ones
already explained in the text of this thesis). Further details are available from the author

upon request.

Raw material

(I) The input-output ratio (or [1 — waste rate]) was assumed not to change between
soybean flour, cowpea flour, and cowpea meal production (for the analysis of dry
COWpea meal), as well as between maize, cowpea, soybean, and groundnut processing

(for the analysis of Tom Brown/Weanimix).

(2) There was a respondent whose cowpea flour contained onion as an ingredient. To

construct a budget for producing dry COWpea meal with COWpeas as the only
ingredient, the cost of onion was subtracted from the raw material costs and an

estimated cost of COWpea grain that would be needed to obtain the same end product

weight was added.

Equipment

(1) The purchase prices of the equipment were inflation-adjusted using the CPI obtained
from the IMF.

(2) Costs for owned equipment were calculated as follows:

Plate mill: to estimate the monthly share of the target products in the purchase cost, a
lifetime of 30 years was arbitrarily assumed, and the purchase price (inflation-
adjusted) was divided by 360 (= 30 X 12). When a respondent mentioned that she/he
paid maintenance costs but did not know how much, it was assumed that she/he re-
sharpened the plates once a month paying ¢30,000 (the charge of ¢30,000 is
information obtained from custom millers). When a respondent mentioned that there
was no maintenance cost, we thought it was unrealistic. Therefore, the same
assumption was made about the re-sharpening of the plates, and the cost was added.

302

 

 

Sealer: to estimate the monthly share of the target products in the purchase cost, a
lifetime of three years was arbitrarily assumed, and the purchase price was divided
by 36 (= 3 x 12). However, a wide range was observed in the lifetime of sealers
across the respondents. If a sealer was already used for more than three years when
the interview was conducted, it was assumed that the lifetime would end this year
(2007); if a respondent replaced sealers more frequently than every three years (e. g.,
buy a new sealer every 6 months), the given answer was used for calculation of the
monthly cost.

Gas oven, roaster, roasting pan, dryer, sieving machine, and dehuller: to estimate the
monthly share of the target products in the purchase cost, a lifetime of 15 years was
arbitrarily assumed, and the purchase price was divided by 180 (= 15 X 12).

Transportation (including fuel for vehicles)

(1) Vehicle purchase and maintenance costs were included in the transportation costs.

The purchase prices of the vehicles were inflation-adjusted using the CPI obtained
from the IMF. To estimate the monthly share of the purchase cost, a lifetime of 15
years was arbitrarily assumed, and the purchase price was divided by 180 (= 15 X 12).

Miscellaneous

(1) One of the respondents mentioned that a certification from the Ghana Standards

Board (GSB) cost ¢8,000,000 per year for all the products and a registration from the
Food and Drugs Board (FDB) cost ¢1,000,000 every three years per product. These
payments were added to the miscellaneous costs of all the respondents (the cost of
GSB certification was multiplied by the share of the target products in the value of all

products produced), whether or not the respondents mentioned that they paid these
fees.

Share calculation

(1) Some respondents did not produce but bought, packaged, and resold an oil product.

Due to lack of information, this product was not included in the share calculation of
the budgeting analysis. (However, if the payment for different cost components
reported by the respondents in fact included the payment for the resources used to sell
this oil product, the calculated share of the target products in the payment for such
cost components would be overestimated.)

303

APPENDIX 5

Data and Notes on the Calculation of Representative Budgets
for Kosei Preparation

The data collected from 13 kosei vendor respondents on the costs and sales of kosei
business are presented below. Many numerical values that were supposed to be collected
from kosei vendor respondents were missing or found to be inconsistent with other pieces
of information obtained from the same respondents. As a result, the unit cost calculation
involved a number of estimations and assumptions. Also, since all the respondents were
not preparing kosei in the same way, some modifications were made to several figures in
the data to adjust for uncommon payments made by some of the respondents. The notes

following Table A.5.1 explains major estimations, assumptions, and modifications that
were made. Further details are available from the author upon request.

304

n

A

Room 5.22 2... basic... .88

a

< E 50322..— bosom

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. mos—g 358mm voxoa
. :3 2: mafia .053 0.: .3 38.3.3 :5 35.2.2.3. 2: .3 v8.2.8 8: 032.9 .. 8qu
8... 3.... .....m E 83.: 83: 53.. o as; 8.6 2......” 83 .S c 8.... .....m. ......8 .3
mm... ..3 3m .2 83: 48.8. 84...... 0 8n... 8...» 82m 83 .8 o 83... an... 28.2 ......
m8. :3 8 .... 3.... 88.8 ”3.: o 83 83 83 8.. S. o 8.... 8.... mg... m.
..8. _E 8. 3 2...: 3.8 .23.. .. 83 83 $3. 88.. em ... $3 23 83. N.
..8. _ m3 _ 3 n... 83.. E3: 838 8 Sum 83 .3.: 83 8.. o 83 8...... e83 ..
.8. :3 8m ...... Ed. .88. R3... .. 8.3 83 88.8 8%. 83 ,o 2...: 83. 48...... o.
8... 8.. 8m 3 83. 88.8. 2.3.. o 83 83 34“.... 83 ...8. ... 23 8mm 83“.... 8
NS. _ N3 _ 8.... ....m 83:. 838 33.: o 2...... .33 5.3 8.. 8... .. 8% 88... 38.8. m
88. N3 _8 3 :3 83... 38.8 o 83 .3.. _ 3... .3. 4:. .. o8.~ .8... a... a
.8. $41.3... ...... ...... .83... 88.8. 83: o 8% 8...... 2%.: com 5 8... o .8... «2.: 82...... e
88. :3 8m ...: 88.8. 838 88...... o 83 8mm. 8.88 o . 8... o 8...... 88.8 28.8 m
.8. 3.. . .... 3 83 893 84.3 o 83 8.? 88.8 8m. 8.. c 8.... 88m 8...... ..
. . 3 ”2...... 88.8. 32.2 o 83 8.. m 8.. mm 8.. N 83 .. 8mm 8cm 3...... m
..8 18L... 3 ...: 3.....- 838 .3.; o 83 E, 83.. 83. 3.. .. 8.. mm 8.. 8 Re E m
8... Mm ...Wm M... 83.- 82.3 83%... o 83 , 8.3 8...... 83 8 a mmm mmm $3. .
.8. o
... 3 no
m l W W. M W. w. 3 MW. d O m .m
... “m m a ...... m m. m m. mm .n. m m m... ...u. ....w. m. .... ....
a m. ... .. a m m .u. a w. W o w l m 1 m a u a... w
W. m m m. 1.x. m. u: m. m. m. .l: w m. m N
”In .... ..m 0.. n .... W .l. W I.
m m M. a s s
a... S... D. i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82:85 .83. no .3.» Se. 2.8 can “moo _.m.< 2an

 

305

Other ingredients

(1) Respondents 2, 3, 7, 8, 9, and 12: Only these respondents used ingredients other than
cowpeas, onion, pepper, and salt. Even if used, the share of other ingredients in the
total cost was small for most respondents. Furthermore, other ingredients reported by
the respondents (ginger, garlic, green leaves, and spices) seemed to have light weight
relative to their volume. Therefore, the contribution of all such ingredients to the cost
of preparation and weight of kosei were ignored (the payment for these ingredients
was subtracted from the cost). By doing so, the representative budget derived using
the data collected from the 13 respondents became a budget for preparing kosei using
the same ingredients (cowpeas, onion, pepper, and salt)2.

_Sa_1t_

(1) Respondent 9: The frequency of purchase and the quantity at each purchase were
missing information. To assume the cost of salt the respondent used the day before
the interview was conducted, the mean quantity of salt used for 1 kg of cowpea grains
was calculated using the data obtained from 11 other kosei respondents (0.030 kg).
Then, using the price of salt the respondent could recall (¢2,500 for 0.5 kg) and the
amount of cowpeas the respondent used (6.02 kg), the cost of salt needed for the
amount of kosei prepared by the respondent was calculated as follows: 0.030 X 6.02 X
(2,500 / 0.5) = 893.

(2) Respondent 12: The respondent answered that she paid ¢20,000 for l olonka of salt
and used it also for family cooking. Since the weight of l olonka of salt was missing,
the mean price of salt per kg was calculated from the information obtained from the
17 respondents (including agawu vendors) who could recall the cost of salt (¢5,126).
Then, the same calculation done for Respondent 9 was applied: 0.030 X 2.08 X 5,126
= 316.

Water

(l) Respondent 13: The respondent did not know how much she paid for the water that
she used for kosei preparation (she paid a water bill to the Ghana Water Company
every month). To assume the cost of water, the mean value of the ratio of water cost
to COWpea cost was calculated using the data collected from the other kosei
respondents, excluding one outlier. The value obtained was 0.081. Then, this number
was multiplied by the cost of cowpeas that the respondent used the day before the
interview was conducted (¢4,975): 0.081 X 4,975 = 403.

 

2 As shown in Table A.S.l, no obvious difference was found between the daily revenue as well as return of

those respondents who used additional ingredients and the daily revenue and return of those who used only
cowpeas, onion, pepper, and salt—a wide range in both revenues and returns was observed in both groups.
If it is appropriate to use the size of revenue and/or return as a measurement of success in business, adding
extra ingredients to kosei does not seem to be a key for success, at least among the 13 respondents.

306

Fuel

(1) Respondent 2: The respondent was the only vendor interviewed who used LP-gas as
fuel. However, the respondent could not recall its cost. Therefore, the mean ratio of
the cost of charcoal to 1 kg of kosei prepared was calculated (¢895) using the
information obtained from 10 other respondents (excluding 1 outlier and 1 respondent
using firewood), and multiplied by the estimated quantity of kosei prepared by this
respondent (16.46 kg) to assume the fuel cost paid by this respondent: 895 X 16.46 =
14,736.

(2) Respondent 7: The respondent purchased charcoal for the preparation of kosei, hausa
koko, and pinkaso. There was no information available to identify the share of kosei
relative to hausa koko and pinkaso. She reported ¢10,000 as the cost of fuel she used
for the business during the day before the interview was conducted. It was assumed
that: (1) the ratio of charcoal used for kosei to the charcoal used for pinkaso was
proportional to the ratio of oil used for kosei to the oil used for pinkaso (3 / (3+13):
see the notes for oil); and (2) hausa koko needed the mean amount of charcoal that
was used for kosei and pinkaso, that is, the ratio of charcoal use for kosei, hausa koko,
and pinkaso was assumed to be 3 : (3+13)/2 : 13. Therefore, the assumed cost of
charcoal used by the respondent for preparation of kosei was calculated as: 10,000 X
3 / {3 + (3+13)/2 +13}= 1,250.

Custom dehulling and milling

 

(1) Respondents 1, 2, 3, 4, 5, 6, 7, and 8: The cost of custom dehulling and/or milling was
missing. The value was estimated from the quantity of cowpea grain and/or dehulled
cowpeas that the reSpondents used as well as the representative custom milling charge
for that amount of cowpeas—which was calculated based on the data collected from
custom millers and other kosei respondents who could recall the custom milling
charge (see Table 4.9).

(2) Respondent 11: The respondent’s answer included ¢6,000 as transportation cost. No
other respondent mentioned paying transportation cost to go to custom milling. The
reason might be that the other respondents had a custom miller within walking
distance from their house, or it might be that they did not think of transportation cost
when asked about the cost of custom milling, even though they were actually paying
for transportation. There is no information available to know which was the case.
Therefore, the transportation cost paid by this respondent was ignored (936,000 was
subtracted from the calculated cost of custom milling [¢11,291]). By doing so, the
representative budget derived using the data collected from the 13 respondents was
not reflective of any transportation cost.

Hired assistants

(1) Respondent 12: The respondent was the only vendor who hired assistants among the
13 kosei respondents. The payment for assistants on the day before the interview was

307

conducted (¢20,000 per day for 2 assistants [i.e., ¢10,000 per assistant]) was removed
from the cost variables. By doing so, the estimated returns of each respondent
included the return to all the labor needed to prepare kosei.

Revenue

(1) Respondents 4 and 13: The respondents reported that they had leftover balls the day
before the interview was conducted. The values of leftover balls reported by the
respondents were added to their revenue so that the figures reflect the number of balls
they prepared times the price per ball.

The values in Table A.5.1 were divided by the estimated weight of kosei prepared by
each respondent to obtain the unit cost, revenue, and return. The results are shown in

Table A52.

308

.2... 03.... ...... 2.m.< 03m... .2520”. o. .88.. 8.... o... 8.526.... ..m.< 03¢... .858
Amddé actoom E S : .5233 08. .83 ..o w: _ 2305 o. .532. .8...
89:8 Eu .0 Emma? .«QEQ 629 8:.st ”.38.. BEES»... Emacs... ......u. “.6 528% E w 55 8 39:82. 22.80:. o... 3.2.2.8

8.8.3.8 ”vow—on 82...; ..oEzmmm 888.. A2283“... 32.58 N H ..on ..o ems... 2.. “.2330 83g 8 “.2503 80:50 ”BszswE 83oz

 

 

 

 

 

 

.m. 8... m2... 8...: :3 :8 N8 3:... .... 8 8. :8 $3 .....5
8. :8. an..- «8.2 8%... .8 8... 8m.“ ..m 8 as 8.... 8m... .... + .m.
... .8. SN 38 28.. .3 R 83 8.. 8:. .... H32...... ... .93.. .33 ......
N... 8... 88. 32.2 83.... 88 m... .8... ..8 Q 5. 38.. R3 3. H22...... .... .38 ....c 8.8...
.3. 88. 2.....- 8...... 83 08 ...... .- E..- 2..- N 08. o8- an. .93 . ...
8. .8. 38.2 «.83 88.8 88.. 83 88.... . z... 8. 88.. 83 :3 3.3 + ...
... 8... 88.8 83 8% N8 8... 83 88 3.. 33m 8. 8.... .3. 8...? .8538 .38 ....m
m... 38. m3...“ .....8. $3. 3.. 3.. 8.... 8.. 8 v8 8.... .8... ... ”328 .5533 ..ea :8...
_ .... _ mm... 83 8mm. 82 88.. 8.... 83 .mm 8 8.. ...: 83 ...
E, 8... .8... 8...: 83. E. m; 8?. 8... 8 .3.... 88 83 m.
8. 8... .8... ... ..m. s . N... Na. 8.. a3 :8 : m8 8... 83 ..
N... .8. 3.... 8...... :3. ..8 ...... 8:3 ...... 8. 88 a... 8.3 ...
a. 8... :3 38.: 8...... 88 0:... .85 SM 8. 8.. 8... 8...“ a
_i_ 28. .86. 83.8 8% 8m ...... _ ...... o. 8 o... 83 83 M.
....er 8... 83 83. m8... 8. 8.. 83 8 8 ...: 8m. 83 h
m... 8... 8...... 8...... .8... :8 88 _ .8... 88 8. 8.. :N.. 2.... m.
INHII 8... 83.8 80.... .8... ...... 8.. 83 .. ...: 8m 8... 2...... m
mm“ .8. :8 83. m8... 8.... E... 88.... .... 2. SN 5... .... ..m ..
8 s 8... 83 .....m. 8m... .8 on... 83 ..8 8 8m 88 .8. m
.... 8... E- 8...: 8...... Rm . 8.. 98.8 8. Q 83 R3 «8... N
...V. .8. 5.....- 2...... 3&8 38.. a8... _ R3. :3 S 33. SN 83 .
P mu
m .w. .... m. m, M... .... .. m s .... o .. m. w ......
.0 e m m u m w m m m w w m. m m m. m
.7 H w «u. m. WI w l m 1 m u u w m... m.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82:85 .80.. .0 303.3 0333823.. ...... 3 8:. 2.8 can 800 N.m.< 038...

309

APPENDIX 6

Mean Retail Margin Calculation

The mean retail margin for each of the three supermarket and seven small shop
respondents was calculated as follows:

(1) Supermarket Respondent 1: since the respondent answered that the margin for all
flour products was 20%, this value was used as it was;

(2) Supermarket Respondent 2: the purchase prices reported by the respondent were
VAT/NHIS excluded (i.e., the respondent reported processor prices). The formula Pc

= 1.15 (1 + m) Pp (see Section 4.2.4) was used to obtain the value of m for each
product, and then the mean value was calculated;

(3) Supermarket respondent 3: the respondent could not provide the purchase prices of
the products. Therefore, the processor prices of those products obtained through the
processor interviews were used. The formula Pc = 1.15 (1 + m) Pp was used to obtain
the value of m for each product, and then the mean value was calculated;

(4) Small shop respondents: no respondent mentioned the VAT/NHIS during the
interviews. Although it is uncertain, their margin might not be subject to the
VAT/NHIS because these shops are small in scale, large in number, and more
importantly, their business records might not be kept well enough for VAT/NHIS
calculation; all of these factors could undermine the efforts of the government to
check and collect the VAT/NHIS from them. On the other hand, the suppliers of
goods to these small shops most likely collect and pay the VAT/NHIS, because the
products sold at the small sh0ps observed during the fieldwork were well-packaged
and brand-named products, including imported goods; the government could easily
collect the VAT/NHIS from such processors. For this reason, it was assumed that the
purchase prices reported by the respondents included the VAT/NHIS (i.e., 1.15Pp).
As for the selling prices, regardless of what has just been discussed, it was still
assumed that the respondents collected the VAT/NHIS on their margin. This is
because it was expected that after the VAT/NHIS was introduced and the purchase
price for the respondents increased, the respondents kept the rate of margin to the
purchase price (i.e., they receive the margin of m{(1+0.15)Pp}, of which 0.15mPp is
the additional revenue for them unless it is collected by the government), rather than

kept the absolute value of margin (i.e., mPp). The formula Pc = 1.15 (1 + m) Pp was
used to obtain the value of m for each product;

As mentioned in Section 3.2.3.2, since few small shops were observed selling
COWpea/soybean flour or Weanimix/Tom Brown, it was decided to: ( 1) ask

310

respondents about Cerelac3 , a p0pular weaning food produced by the company
Nestlé4, and also about other industrially processed grain flour products, where such
products were sold by the respondents; and (2) show respondents a sample bag of
cowpea flour purchased from one of the processor respondents and ask them at what
price they would sell the product, proposing three different purchase prices (¢7,000,
¢10,000, ¢12,000). Then, the mean values of m were calculated using the formula Pc
= 1.15 (l + m) Pp for Nestlé’s products and for a sample bag of cowpea flour (i.e.,
mean of the 3 self-assumed selling prices), separately from other products;

Finally, the mean values of m for each respondent were calculated using: (1) the
values of m for each of the targeted products and other industrially processed flour
products sold by the respondent; (2) the mean value of m for Nestlé’s products sold
by the respondent; and/or (3) the mean value of m for a sample bag of cowpea flour.

 

3 Cerelac had a variety of flavors, and it was too time consuming to ask about all different flavors that the
respondents sold. Therefore, only two flavors (maize [seemingly cheapest among Cerelac] and “3 fruits”
[seemingly most expensive]) were selected. Cerevita (instant maize porridge) was also included when it
was sold by the respondent.

4 There was a concern that the retail margin for products produced by a large-scale company might not
provide a good estimate for the potential retail margin for our targeted products. In fact, one respondent
mentioned that the selling prices were determined by the processor. Another respondent mentioned that the
selling prices were determined by the wholesaler. However, for each respondent, the calculated retail
margins for Cerelac were not substantially different from the self-assumed retail margins for cowpea flour.

311

APPENDIX 7

Weaning Mothers as Consumers of Kosei

As mentioned in Section 4.2.1.9, weaning mothers were asked a set of questions about
their consumption and home preparation of kosei. The results are shown below.5

There was a wide range in the frequency of eating kosei among the respondents (Table A.
7.1). One third of the respondents knew how to prepare kosei. Only two respondents had
ever bought cowpea flour: one of them bought it only once, while the other bought it
regularly once a week. Both of them bought it from a hospital and used it for fortification
of weaning foods (i.e., they did not buy it to make kosei). Four respondents prepared
cowpea flour by themselves.

When asked why they did not buy cowpea flour, the majority of the respondents
answered that they had never seen commercial cowpea flour. Two respondents mentioned
that they did not know what to use it for. Another respondent mentioned that she
generally did not like commercial flour products due to safety concerns.

Table A.7.l Frequency of eating kosei, knowledge on how to prepare
kosei, and experience in purchasing cowpea flour among weaning
mother respondents

 

 

 

 

Consumption and home preparation of kosei / $23112?
Experience in use of cowpea flour (n 2 30)
0 5
HO 7
1 1-20 6
How many times per month eat 21-30 5
kosei Rarely 2
Often I
Depends 3
Don’t know I
. No 20
Know how to prepare kosei Yes I 0
No“ 28
fl
Ever bought cowpea our Yes 2

 

 

 

 

 

* Includes four respondents who prepared cowpea flour by themselves.
Source: Field survey in Accra, February and March 2007.

 

5 It is not clear whether weaning mothers tend to have any particular behavior towards the consumption of
kosei that is not found among general consumers. For example, one of the respondents mentioned that she
used to eat kosei but stopped when she got pregnant. Therefore, it should be emphasized that the data were
collected from a very specific sample population (i.e., weaning mothers visiting a health clinic/hospital for
weighing their babies) and the sample size is very small. The results are not presented here to derive any
general characteristics among Ghanaians about their kosei consumption but rather to stimulate discussion
and help to design further research on this issue.

312

As Table A.7.2 shows, among the 10 respondents who knew how to prepare kosei, 5
actually prepared it at home, while the other 5 did not.

Of the five respondents who prepared kosei at home, three answered that they would try
commercial cowpea meal/flour for making kosei if such meal/flour was available, while
the other two answered that they would not try. One of these two respondents did not buy
commercial flour but prepared it by herself, while the other was not sure whether kosei

would taste the same when using such meal/flour.

Of the five respondents who did not prepare kosei at home, two did not eat much kosei
each time, and therefore it was more convenient for them to just buy the amount they
wanted to eat. Another two were not sure whether they could prepare the right amount
they wanted to eat. The remaining one thought that the preparation was tedious. All of
these five respondents were interested in trying commercial cowpea meal/flour.

Table A.7.2 Home preparation of kosei and interests in dry cowpea
flour/meal among weaning mother respondents

 

 

 

 

 

 

 

 

Weaning mother who Would try Would NOT
knows how to prepare cowpea try cowpea Total
kosei (n = ID) meal/flour meal/flour
Currently prepare kosei 3 2 5
at home
Currently do NOT
. 5 O 5
prepare kose1 at home
Total 8 2

 

 

 

 

Source: Field survey in Accra, February and March 2007.

313

APPENDIX 8

Cowpea-Fortified Gari

Gari (grated, fermented and roasted cassava) is popularly used in Ghanaians’ daily diet. It
is relatively inexpensive but low in protein. Studies conducted by food scientists
associated with the B/C CRSP have found that cowpeas could replace cassava up to 20%
by weight while still guaranteeing an acceptable end product quality (G. A. Annor,
personal communication, March 22, 2007). As reported in Chapter 1, cowpea-fortified
gari was initially among the target products of this case study. However, due to lack of
time during the fieldwork in Accra, only one gari processor was interviewed. The
respondent produced cowpea-fortified gari, soybean-fortified gari, and yellow gari only
when she was requested by customers who brought or paid for additional ingredients. A
budgeting analysis of cowpea-fortified gari was conducted using the data collected from

this respondent and applying the same methods used to construct the budgets for

industrially produced dry cowpea meal and Weanimix. The purpose of this analysis was
not to derive general conclusions but to stimulate discussion and provide useful

information to design further research on the profitability of cowpea-fortified gari. The
results are reported below.

Table A.8.1 Budgets for producing 1 kg of cowpea-fortified gari (¢)

 

 

 

 

 

 

 

 

L Regular (no 10% 20% J
fortification) fortification fortification
Broccssor price 5,055 5,849 6,642 1
Raw materials 978 1,569 2,160]
Equipment 235 438 640
Wage 1,795
Electricity 0
Fuel (excl. fuel for vehicles) 233
Water 65 Assumed to Assumed to
Rent 24
Transportation (incl fuel for vehicles) 204 be the same be the same
, , , ' as regular as regular
Printing & stationery 0
Telecommunication 37
Packaging material 73
Miscellaneous 0
metal cost 3,695 4,489 I 5,2821
[ Return 1,360 Same as regular —l

 

 

 

Notes: Raw materials: (cassava) the respondent purchased cassava from a farmer;
(cowpeas) the waste rate for processing cowpeas was assumed based on the data
collected from cowpea/soybean flour processor respondents; Equipment: (regular)
costs of custom-grating and ~pressing cassava + own roasting pan; (cowpea-fortified)
costs of custom-grating and -pressing cassava + custom-dehulling and -milling
cowpeas (the charge was assumed based on the data collected from custom miller
and kosei vendor respondents) + own roasting pan; Transportation: the cost for
buying inputs (i.e., transportation cost for selling the output is not included).

Source: Field survey in Accra, February and March 2007.

314

Figure A.8.] Budgets for producing 1 kg of cowpea-fortified gari

 

 

 

           

7,000

6,000 -~——

5 000 **-~
on i I Return
5 4,000 fliA 7- Equipment
'° 3,000 «e ....... ERaw meat
---------- X Other costs

 

 

 

\

iii: fr§ \\

Regular 10% fortification 20% fortification

 

 

I r

 

 

 

 

Source: Table A.8. I.

As shown in Table A.8.l and Figure A.8.1, the cost of raw materials was expected to
increase from about 911,000 to about ¢1,600 with a 10% replacement of cassava with
cowpeas and from about ¢1,000 to about ¢2,200 with a 20% replacement. This sharp
increase in the raw material cost is due to the difference in prices between cassava and
COWpeaS. Price data collected from the MoF A indicates that the monthly wholesale price
per kg of cowpeas was on average 4.0 times the price per kg of cassava during 2002-2006
at urban markets of the Greater Accra Region (with a standard deviation of 0.57 times).
In fact, the same data show that cowpeas (fortifier) have been more expensive than gari
(end product)—for the same period (i.e., 2002-2006) and location (i.e., urban markets of
the Greater Accra Region), cowpeas were on average 1.6 times more expensive than gari,
with a standard deviation of 0.39 times.

In addition to the increase in the raw material cost, the equipment cost also increased
because of the charge for custom-dehulling and -milling to process cowpeas. As a result,
the total cost was estimated to increase from about ¢3,700 per kg of gari to ¢4,500 (10%
replacement) and to ¢5,300 (20% replacement).

The processor price of 20%-cowpea-fortified gari was estimated to be about ¢6,6OO per
kg. In fact, during the interview, the respondent mentioned that if she had to replace 20%
of cassava with cowpeas, she would have to increase the selling price of gari from
¢10,000 per olonka (estimated to be ¢5,055 per kg)6 to ¢12,000 per olonka (estimated to
be ¢6,066 per kg). However, she also mentioned that at this price the customers would
complain, and therefore she did not produce COWpea-fortified gari for normal customers.

 

6 This respondent sold her gari for a much higher price than the processor prices of gari reported by retailer
respondents. The mean processor price of gari reported by three retailer respondents was ¢3,707 per kg
(with a standard deviation of ¢292).

315

 

Table A.8.] and Figure A.8.] also show that, if it is assumed that the respondent would
have to sell cowpea-fortified gari for the same price as regular gan', the return would
decline to less than one-half of the original value for 10% replacement, while it would
turn negative for 20% replacement.

With regard to the advantages and disadvantages of fortifying gari, the respondent
mentioned that fortification with either COWpeas or soybeans would improve the taste and
nutritional value of gari, while it makes processing more tedious.

These results suggest that cowpea-fortified gari would face a challenge to be price-
competitive with normal gari, because of a large difference in prices between the original
ingredient (i.e., cassava) and the fortifier (i.e., cowpeas), unless customers have high
willingness to pay for cowpea-fortified gari7.

 

7 It should be noted that the price of gari fluctuates seasonally. Therefore, the results of budgeting analysis
might look different if it is conducted in a different season of the year.

316

 

REFERENCES

Afoakwa, E. 0., Sefa-Dedeh, S., & Sakyi-Dawson, E. (2004). Effects of COWpea
fortification, dehydration method and storage time on some quality characteristics
of maize-based traditional weaning foods. African Journal of Food, Agriculture,
Nutrition and Development, 4, l.

Annan, N. T., & Plahar, W. A. (1995). Development and quality evaluation of a soy-
fortified Ghanaian weaning food. Food and Nutrition Bulletin, 16, 3. Tokyo:
United Nations University Press.

Bean/Cowpea Collaborative Research Support Program (B/C CRSP). (2006a). Regional
project research and training workplans (October IL2006 to September 30, 2007).
Technical Committee Meeting, Memphis, Tennessee, May 10-12, 2006.

Bean/Cowpea Collaborative Research Support Program (B/C CRSP). (2006b). Research
and training highlights 2006.

Bentley, M. E., Dickin, K. L., Mebrahtu, S., Kayode, B., Oni, G. A., Verzosa, C. C., et al.
( l 99]). DeveIOpment of nutritionally adequate and culturally appropriate weaning

food in Kwara state, Nigeria: An interdisciplinary approach. Social Science &
Medicine, 33, IO, llO3-llll.

 

Bernsten, R. H. ( I993). The role of cowpeas in the Ghanaian economy: A rapid appraisal
subsector analysis. Michigan State University, Department of Agricultural
Economics.

Bernsten, R. H. (2005, Summer). IV. Survey sampling AEC 874: Michigan State
University.

 

Briggs, P. (2004). Ghana: The Bradt travel gpide (3rd ed) Guilford: The Globe Pequot
Press.

Crawford, E. W. (2002, Summer). International farm management analysis methods. AEC
890: Michigan State University.

Davis, D. W., Oelke, E. A., Oplinger, E. S., Doll, J. D., Hanson, C. V., & Putnam, D. H.
(1991). Cowpea. In University of Wisconsin-Extension, Cooperative Extension.
( I 992). Alternative field 0mg manual. Retrieved August 2, 2006, from
http://cornagponomgwiscedu/AltemativeCrgps/Cowpea.htm.

Debertin, D. L. (1986). Agricultural fimoduction economics. New York: Macmillian
Publishing Company.

317

 

 

 

 

 

Dovlo, F. E., Williams, C. E., & Zoaka, L. (1976). Cowpeas: Home preparation and use
in West Africa. Ottawa: lntemational Development Center.

Egounlety, M., Aworh, O. C., Akingbala, J. 0., Houben, J. H., & Nago, M. C. (2002).
Nutritional and sensory evaluation of tempe-fortified maize-based weaning foods.
International Journal of Food Sciences and NutritionLS3, 15-27.

El-Habashy, M., Phillips, R. D., Eitenmiller, R. R., Mostafa, M. M., Rahma, E., & El-
Bedaiwy, A. (1995). Computer optimization of weaning food formulas. Journal of
Food Quality, 18, 295-312.

El-Habashy, M., Phillips, R. D., Eitenmiller, R. R., Mostafa, M. M., Rahma, E., & El-
Bedaiwy, A. (1997). Protein quality assessment of computer optimized weaning
formulas. Journal of Food Quality, 2Q 267-277.

Faye, M. D. (2005). Investigation of key aspects for the successful marketing of cowpeas
in Senegal. Ph.D. Dissertation, Department of Agricultural Economics, University
of the Free State.

Food and Agriculture Organization of the United Nations (FAO). (1997). Agriculture,
food and nutrition for Africa: A resource book for teachers of agriculture. Rome:
Author.

Fulton, J. (2004). Trip report: Senegal May 11-26, 2004. Bean/COWpea Collaborative
Research Support Program (B/C CRSP).

Harsh, S. B., Connor, L. J., & Schwab, G. D. (1981). Managing the farm business.
Englewood Cliffs, NJ: Prentice-Hall.

Hayes, R. E., Mwale, J. M., Tembo, P., & Wadsworth, J. I. (1995). Computer-optimized
weaning food blends. Food and Nutrition Bulletin, 16, 3. Tokyo: United Nations
University Press.

Henshaw, F. 0., McWatters, K. H., Oguntunde, A. O., & Phillips, R. D. (1996). Pasting
properties of cowpea flour: Effects of soaking and decortication method. Journal
of Agricultural and Food Chemistg, 44, 1864-1870.

International Institute of Tropical Agriculture (IITA). (2006). Cowpea. Retrieved August
2, 2006, from hppM/www.iita.or_g/crm_3_/COWpea.htm.

Kethireddipalli, P., Hung, Y. C., McWatters, K. H., & Phillips, R. D. (2002). Effect of
milling method (wet and dry) on the functional properties of cowpea (Vigna

unguiculata) pastes and end product (akara) quality. Journal of Food ScienceL67,
1, 48-52.

318

 

Lambot, C. (2002). Industrial potential of COWpea. In C. A. Fatokun, S. A. Tarawali, B. B.
Singh, P. M. Kormawa, & M. Tamo (Eds), Challenges and opportunities for
enhancing sustainable cowpea production (pp. 367-375). Proceedings of the
World Cowpea Conference III held at the lntemational Institute of Tropical
Agriculture (IITA), Ibadan, Nigeria, 4—8 September 2000. IITA, Ibadan, Nigeria.

Langyintuo, A. S., Lowenberg-DeBoer, J ., & Amdt, C. (2005). Potential impacts of the
proposed West African monetary zone on cowpea trade. Agricultural Economics,
3;, supplement 411-421.

Langyintuo, A. S., Lowenberg—DeBoer, J ., Faye, M., Lambert, D., Ibro, G, Moussa, B., et

al. (2003). Cowpea supply and demand in West and Central Africa. Field Crops
Research, 82, 215-231.

Langyintuo, A. S., Ntoukam, G., Murdock, L., Lowenberg-DeBoer, J., & Miller, D. J.

(2004). Consumer preferences for cowpea in Cameroon and Ghana. Agricultural
Economics, 30, 203-213.

 

Lartey, A., Manu, A., Brown, K. H., Peerson, J. M., & Dewey, K. G. (1999). A
randomized, community-based trial of the effects of improved, centrally
processed complementary foods on growth and micronutrient status of Ghanaian

infants from 6 to 12 mo of age. The American Journal of Clinical Nutrition, 70,
391-404.

Leung, W. W. (1968). Food composition table for use in Africa. Bethesda, Maryland: US
Department of Health, Education, and Welfare, and Rome: Food and Agriculture
Organization of the United Nations (FAO).

Masters, W. A., & Sanogo, D. (2002). Welfare gains from quality certification of infant
foods: Results from a market experiment in Mali. American Journal of
Agricultural Economics, 84, 4, 974-989.

McWatters, K. H. (1986). Use of peanut and cowpea flours in selected fried and baked
foods. In R. L. Ory (Ed.), Plant proteins: Applications, biological effects, and
chemistry. Washington, DC: ACS Symposium, Series 312, American Chemical
Society.

McWatters, K. H., & Brantley, B. B. (1982). Characteristics of akara prepared from
COWpea paste and meal. Food Technology 36, 66-68.

McWatters, K. H., Chinnan, M. 8., Phillips, R. D., Beuchat, L. R., Reid, L. B., & Mensa-

Wilmot, Y. M. (2002). Functional, nutritional, mycological, and akara-making
properties of stored cowpea meal. Journal of Food Science. 67, 6, 2229-2234.

319

 

 

McWatters, K. H., Resurreccion, A. V. A., & Fletcher, S. M. (1990). Response of
American consumers to akara, a traditional West African food made from cowpea
paste. International Journal of Food Science and Technology 25, 551-557.

Mensa-Wilmot, Y., Phillips, R. D., & Hargrove, J. L. (2001a). Protein quality evaluation
of cowpea—based extrusion cooked cereal/legume weaning mixtures. Nutrition
Research,21, 849-857.

Mensa-Wilmot, Y., Phillips, R. D., Lee, J., & Eitenmiller, R. R. (2003). Formulation and
evaluation of cereal/legume-based weaning food supplements. Plant Foods for
Human Nutrition, 58, 1-14.

Mensa-Wilmot, Y., Phillips, R. D., & Sefa-Dedeh, S. (2001b). Acceptability of extrusion
cooked cereal/legume weaning food supplements to Ghanaian mothers.
International Journal of Food Sciences and Nutrition, 52, 83-90.

Mishili, F. J. (2005). Cowpea markets and consumer preferences in Ghana. MS. Thesis,
Purdue University.

Nimoh, F. (2004). Consumer preference and market opportunities for processed cowpea-
based products in the coastal regions of Ghana. M.Phi1. Thesis, Department of
Agricultural Economics and Agribusiness, University of Ghana-Legon.

Nti, C. A., & Plahar, W. A. (1995). Chemical and biological characteristics of West
African weaning food supplemented with cowpea (Vigna unguiculata). Plant
Foods for Human Nutrition, fl, 45-54.

Onofiok, N. O., & Nnanyelugo, D. O. (1998). Weaning foods in West Africa: Nutritional
problems and possible solutions. Food and Nutrition Bulletin, 19, 1. Tokyo:
United Nations University Press.

Phillips, R. D., McWatters, K. H., Chinnan, M. S., Hung, Y., Beuchat, L. R., Sefa-Dedeh,
et al. (2003). Utilization of cowpeas for human food. Field Crmps Research, 82,
193—213.

Plahar, W. A., Nti, C. A., & Arman, N. T. (1997). Effect of soy-fortification method on the
fermentation characteristics and nutritional quality of fermented maize meal. Plant
Foods for Human Nutrition, 51, 365-380.

Plahar, W. A., Onuma Okezie, B., & Gyato, C. K. (2003). Development of a high protein

weaning food by extrusion cooking using peanuts, maize and soybeans. Plant
Foods for Human Nutrition, 58, 1-12.

320

 

Reardon, T., & Timmer, C. P. (2007). Transformation of markets for agricultural output in
developing countries since 1950: How has thinking changed? In R. E. Evenson &
P. Pingali (Eds), Handbook of agricultural economics, 3: Agricultural
development: Farmers, farm production and farm markets. Amsterdam: Elsevier
Press.

Sefa—Dedeh, S. (2005). Final report for USAID associate cooperative agreement No. 641-
A—00-03-0003. Ghana Private-Public Partnership Food Industry Development
Program.

 

Singh, A., Hung, Y., Corredig, M., Phillips, R. D., Chinnan, M. S., & McWatters, K. H.
(2005). Effect of milling method on selected physical and functional properties of
cowpea (Vigna unguiculata) paste. International Journal of Food Science and
Technology 4p 525-536.

 

Smith, I. F. (1995). The case for indigenous West African food culture. BREDA Series No.
i United Nations Educational, Scientific and Cultural Organization (UNESCO),
Bureau Regional pour I’Education en Afrique (BREDA).

United Nations Statistics Division (UNSD). (2005). Per capita GDP in US dollars: All
countries for all years — sorted alphabeticalli Retrieved September 6, 2006, from
http://unstats.un.org/unsd/snaama/dnllist.asp.

United Nations Statistics Division (UNSD). (2006). Millennium Development Goals
Indicators. Retrieved October 4, 2006, from
http://mdgs.un.org/punsd/mdfig/Defaultasp}:

Varian, H. R. (2003). Intermediate microeconomics: A modern approach (6th ed.). New
York: W. W. Norton & Company.

World Bank. (2006). Africa development indicators 2006. Washington, DC: Author.

321

 

     

ll11711111111111:11;l

3 1293 02956 8