TE :TEELTSES 0F TEE BEETLE!) E’GE
FOOD TE EGYPT

Thesis T's 9 {Ta Dawes GT T3335! A
TEin T6115: ST‘TE ' EKTT PRETTY

ngaa " Haas-an
296'?

{Hum

This is to certify that the

thesis entitled

AN ANALYSIS OF THE DEMAND FOR
FOOD IN EGYPT

presented by

Ragaa Abdel-Rassoul Hassan

has been accepted towards fulfillment
of the requirements for

Ph.D. Agricultural

degree in .
Econom1cs

:2; t, ,_ w m 027 (x

'!.A-§_.o~.‘-K— L'sL. I- :(..'.t- s;'§'."'
T K

I;g;<¢,dz»x'

L.
‘ C-

, " I “
La'ilx but i

 

 

Major professor

Due 15 August 1969

 

0-169

 

TJHn...ry ¢L

3553’." ‘.‘*‘MO§J\ 0"," '~ 994‘ 3

_ _.___.!fi

ABSTRACT
AN ANALYSIS OF THE DEMAND FOR FOOD IN EGYPT

By Ragaa A. Hassan

' The accumulating evidence of what seems to be a chronically
precarious food situation in Egypt has, in recent years,
increased the general concern about her food problem, and
heightened the interest in carrying out this analytical inves-
tigation. Much information is required, particularly for
planning purposes, about the changing structure of the Egyptian
food economy during the process of economic growth. Such
information would facilitate the more efficient use of planning
techniques in programming development projects in the country.

This study has focused on estimating parameters determining
the demand for food in Egypt. These estimates were used, along
with estimates for the annual rate of growth in GNP, and the
projected rates of population growth, to estimate the aggregate
demand for food products in l975 and I985. To accomplish this,
several types of data, cross-sectional and time-series, were
utilized, and more than one model of analysis was introduced.
However, it should be emphasized that, in a sense, the study
was one-sided as it focused on the demand side of the food
problem. Little attempt was made to analyze supply or to

project it.

Ragaa A. Hassan

In the cross-sectional analysis, which was based on the
published data of the household budget survey of 1958/59, the w
demand for food was assumed to depend on income, size of the
household, and location. Three alternative functional forms
were considered: (l) logarithmic, (2) semi-logarithmic, and
(3) log-inverse. Of these three functions, the logarithmic
relationship seemed to provide the best fit. The results
drawn from it were utilized in projecting demand.

Throughout the cross-sectional analysis, estimates for
income elasticity of demand for food groups and individual
food items were calculated for both urban and rural areas, and
a weighted average was introduced to provide an estimate for
the entire population.

The results of the cross-sectional analysis indicate
that two-thirds of the total expenditures in rural areas were
spent on food and beverages, compared to about one-half in
urban areas. Using Lorenz curves, total expenditures and food
expenditures were shown to be more equally distributed in
rural areas than in urban areas. Two explanations were offered
in this connection; First, that qualities and therefore prices
paid differ much more in urban areas, and second, that incomes,
occupations, and social status, and tastes vary greatly in

urban areas.

Ragaa A. Hassan

The income elasticity of demand for food in Egypt was not
as high as results from other countries' data suggested. The
.67 elasticity coefficient for Egypt is intermediate rather
than high in comparison with estimates for other countries.

The relations between quantity and expenditure
elasticity figures were as expected, expenditure elasticities
being higher than quantity elasticities.

High income elasticities for home-produced wheat, maize,
and millet induced a detailed discussion of grain consumption
patterns. This, in turn, resulted in the use of an indirect
method to project the demand for wheat and maize.

“ Although household size was expected to be an important
factor in the demand for food, its significance failed to
materialize when the effect of the household type, based on
its size, was tested by the analysis of variance procedure.
However, when the number of consumption units was included in
the regression analysis, this variable produced a coefficient
with a value that differed significantly from zero.

Two methods were used for estimating price parameters.

A combined model utilizing cross-sectional and time-series
data was applied to six products for which sufficient time-
series data were available. For other products, a newer

mathematical approach based on several assumptions about the

Ragaa A. Hassan

relationship between income elasticities and price elasticities
was utilized. The results of the combined model were not very
satisfactory in terms of statistical significance. However,
the mathematical model produced fairly reasonable results
suggesting that further exploration of the technique might
prove to be of a considerable value.

Projections of food demand were undertaken in the usual
manner. Two population projections were utilized based on
different assumptions about fertility rates. Rates of growth
in per capita income were estimated. Using these rates,
together with the income elasticity coefficients obtained in
this study, demand projections were made for 1975 and I985 for
commodity groups and individual food items.

Turning to the estimated percentage increase in total
demand for commodity groups, it was shown that the highest
projected increases were for fruit (about lOO percent in l975
and 260 percent in l985); meat, fish, and eggs (85 and 220
percent); milk and dairy products (80 and 200 percent); and
tea, coffee, and other stimulants (85 and 2l6 percent). The
lowest projected increases were for cereals and starches;
dried legumes; and oils and fats. The last three groups were
projected to increased by about 55 percent in I975 and about
l30 percent in l985. This, however, still implies a very

high increase in the demand for these products.

Ragaa A. Hassan

When these projections were made on per capita basis, an
increase in most items was observed. It should be emphasized,
however, that a part of this increase would be absorbed in a
shift to higher quality products rather than higher quantities
as per capita income rises.

-« The study has shown that rather large increases in the .—
demand for food products are to be expected. This demand has
to be fulfilled through increasing food production, increasing
food imports, or some combination of the two. Prospects for
increasing productivity seem best for certain crops such as
maize, vegetables, fruit, oil crops, sugar cane, and livestock
products. There is little doubt that increased productivity can
greatly aid in solving the food problem fim'one generation. Yet,
it is hard to see how it could do so for a population that will
almost double in about twenty-five years. The only feasible
path, given the present state of technology, seems to be a
drastic effort to decrease the rate of population growth.

The findings of this study raise questions about the
extent to which the nation would be able to provide its food
supplies, and also the question of trade policy with regard to
food products. It has been often argued that Egypt should

stop producing wheat, and concentrate her land and water

Ragaa A. Hassan

resources in products for which she has greater comparative
advantage such as cotton, rice, oil cr0ps, as well as sugar
and vegetables. This would imply greater dependence on trade,
but if imports were matched by exports of other cr0ps with
higher value, there should be no cause for concern. However,
for sounder decisions to be made in the area of trade policy,
more analytical studies are needed particularly for grains

for which shortages seem to be alarming.

AN ANALYSIS OF THE DEMAND FOR FOOD IN EGYPT

By

Ragaa Abdel-Rassoul Hassan

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

I969

 

 

’loflp
Q9(0C>-
2'70

a,

DEDICATION

To the memory of my father Abdel-Rassoul
Hassan, to whom I owe my education and the
delight in knowledge, and to my daughters:

Mona and Susan.

ACKNOWLEDGMENTS

In writing this thesis, I am deeply indebted to several
people who have given me both encouragement and help during
the course of my study.

Dr. Lester Manderscheid gave freely of his time and
experience throughout the study. His consulting and valuable
remarks contributed a great deal in its improvement.

Thanks are due to Dr. Robert Stevens for his encouragement
and help, not only during the course of this study but through-
out my stay at Michigan State. His comments and remarks on
the early drafts of this thesis were particularly appreciated.

Dr. James Shaffer's suggestions and comments has resulted
in many valuable additions, and I am deeply grateful to him.

I am deeply indebted to my major professor, Dr. Harold
Riley who supervised my study program and my thesis. His
patient guidance and assistance throughout my doctoral program,
and especially his constant encouragement and willingness to
help enabled me to pursue my study at Michigan State University.

Under the direction of Mrs. Arlene King, and later
Miss Laura Robinson, the many calculations needed for the study
have been programmed and performed efficiently. Special thanks
are due to Mrs. Laura Flanders, Mrs. Irene Chow, and Miss Sue

Tsai of the Computer Center.

Mrs. Vranna Salender typed parts of the early drafts.
The final copy was typed by Miss Patricia Riley. This has been
very much appreciated.

But the greatest debt is the most difficult to express.
Such is my debt to my wife, Afaf, who made my troubles her
own, and through her kindness and understanding helped easing
most of them.

Though the study has benefitted from the suggestions and
comments of several scholars, colleagues, and friends, the
responsibility of any errors or mistakes remains absolutely

my own.

Chapter I.
Chapter II.
Chapter III.
Chapter IV.

.u-o ——a-—I-—|
o o o I
w N-‘m

TABLE OF CONTENTS

PROBLEM. . . . . . . . . . . . .

Introduction. . . . . .

The Relation Between Economic
Development and Food Consumption.
The Objectives and Basic
Hypotheses. . . . .

THE BASIC FEATURES OF THE EGYPTIAN
ECONOMY. . . . . . . . . . .

2. I

WN-‘Z 0‘me N

O

wwwn NNNN N
0 O 0

w
J?

3.5

Introduction. . .

Land Area and Agricultural
Production. . . . .

The Demographic Picture . . .
The Trends of National Income .
Aggregate Food Consumption.
Summary . . . . . . . . . . .

CEPTUAL FRAMEWORK AND TECHNIQUES.

Source of Information . . . .
Analytical Procedures . .

The Cross-Sectional Models. .
3.3.I The Logarithmic Model. . .
3.3.2 The Semi-Iogarithmid Model
3.3.3 The Log-inverse Model.

The Combined Model for Cross-
sectional and Time-series Data.
Estimating Price Parameters from
Cross-sectional Information .

ANALYSIS OF THE CROSS-SECTIONAL DATA .

4.I

4.2

Some Considerations Concerning the
Utilization of Cross-sectional
Information in Analytical Income-
Consumption Studies . . . .
Definition of the Variables . .
4.2. I 'The Household. . . . .

4.2.2 Total Expenditures . .
4.2.3 Consumption Expenditures
4.2.4 Food Expenditures. .
4.2.5 The Household Size .
4.2.6 The Consumption Unit .

Page

61
64
64
65
66
68
69
69

Chapter V.

Chapter VI.

Chapter VII.

4.3 Results of the Analysis .

4.3.]
4.3.2
4.3.3

4.3.4

4.3.5
4.3.6

Introduction . .
Food Consumption Patterns.
Comparison of the Results
of the Three Models. . . .
Analysis of the Commodity
Groups . . .
Analysis of. Individual
Food Items . .

The Effect of Household
Size . .

4.4 Summary . . . . . : Z Z : :

ESTIMATING PRICE PARAMETERS. . . .

5. I Combined Analysis of Cross- sectional
and Time- -series Data. . .
5. 1. Definition of the Variables.
5.I.2 Results of the Analysis.

5.2 Estimating Price Parameters from

Cross-sectional

5.2.I
5.2.2

Information . . .
Assumptions and Techniques
Results of the Analysis. .

ROJECTION OF THE DEMAND FOR FOOD. . .
Introduction. . . . . . . . . . . .
P0pulation Projections. . . . . . .
Income Projections. . . . . . . .
Consumption Projections . . . . . .
Detailed Results for Food
Commodities . . .

6.5.I Grains . . . . . . . .
Meat and Poultry . . . . . .
Fish . . . . . . . . .
Milk and Eggs. . . . . . . .

A P
6. I
6.2
6. 3
6. 4
6.5

O\O\O\O\C\O\
O
O‘C‘U‘I .ITUJN

O
UWU‘IU1U1U'IU‘I

Potatoes and Onions. .
Citrus and Other Fruit .
Oils and Saturated Oils.

SUMMARY AND CONCLUDING REMARKS

BIBLIOGRAPHY . . . . . . . .
APPENDICES . . . . . . . . . . .

134
I38

I44
I44
I46
ISI
I56

I69
I7I
I74
I76
I77
I79
I79
I80

I8l
l88
I93

LIST OF TABLES

Table Page
1 Plan income targets by sectors . . . . . . 3
2 Net value added and the size of the working

force by sectors, in I959/6O and l964/65 . l4
Value of agriculture output, I960-I964 . . I6

4 Area, production, and value of total out-
put for major crops in I964. . . . . . . . l9
5 Urban and rural populations in successive
censuses from I927 to I960 . . . . . . . . 23
6 Projected size of total population . . . . 24
7 National income and national income

indices in current prices and in constant

I959/6O prices, I952/53 to I964/65 . . . . 27

8 Percentage shares of sectors in national

income, l959/60 to I964/65 . . . . . . . . 28
9 Per capita food consumption in Egypt,

l948/50 to l964/65 . .'. . . . . . . . . . 33
IO The size of the first household budget

survey and its percentage of the popula-

tion . . . . . . . . . . . . . . . . . . . 4I
II Periods covered and the month of study in

the I964/6S household budget survey. . . . 43

l2 Distribution of households according to
their total expenditures brackets,

1958/59 and I964/65. . . . . . . . . . . . 67
I3 Consumption units scale. . . . . . . . . . 70

I4 Average number of consumption units per
household for different household types. . 75

IS Percentage of items of total annual expen-
ditures in rural and urban areas, based
on the household budget survey,

I958/59. . . . . . . . . . . . . . . . . . 8I

Table

l6

l7

l8
I9
20

2l

22

23

24

25

26

27

28

Percentage of items of food expenditures
in urban and rural areas, based on the
household budget survey, l958/59 . . . . .

Values of R2 and S for the combined
analysis of urban and rural areas. . . . .

Values of'R-2 and S for urban areas . . .

Values of R2 and S for rural areas . . . .

Food consumption functions derived from
logarithmic and log-inverse models . . . .

Total expenditures elasticity of demand
for commodity groups and total food,
estimated from the logarithmic and the
log-inverse models for urban and rural
areas. . . . . . . . . . . . . . . . . . .

Total expenditures elasticity of demand
for commodity groups and total food-
weighted coefficients. . . . . . . . . . .

Estimates for B33, from the logarithmic
and the log-inverse models for commodity
groups C O O O O O O O O O O O O O O O O 0

Quantity elasticity coefficients from the
logarithmic and the log-inverse models in
urban and rural areas. . . . . . . . . . .

Expenditure elasticity coefficients from
the log and the log-inverse models in
urban and rural areas. . . . . . . . . . .

Total expenditures elasticity of demand
for the individual food items - weighted
coeffiCients . . . . . . . . . . . . . . .

Total expenditures elasticity coefficients
for wheat, maize and millet. . . . . . .

Home produced wheat, maize and millet as
percent of total expenditures for these
products, related to household income
levels . . . . . . . . . . . . . . . . . .

viii

Page

82

87
88
89

9I

94

99

IOI

IOZ

l03

lO4

l08

Table Page

29 Percentage distribution of expenditures
for cereals and starches in urban and
rural areas. . . . . . . . . . . . . . . . III

30 Total expenditures elasticity coefficients
for households, classified by size, from
the logarithmic and the log-inverse models
in urban areas . . . . . . . . . . . . . . II6

3] Total expenditures elasticity coefficients
for households, classified by size, from
the logarithmic and the log-inverse models
in rural areas . . . . . . . . . . . . . . ll7

32 Average annual household expenditures for
food in urban areas and the number of
reporting households . . . . . . . . . . . I20

33 Analysis of variance for food expenditures
in urban areas . . . . . . . . . . . . . . I2l

34 National disposable income, per capita
income, cost of living index and the
wholesale price index, l94S/46-I962/63 . . l28

35 Per capita consumption and prices for
selected items . . . . . . . . . . . . . . I29

36 Estimated relationships from the model
Zit= BO + 8}; log Pjt/Pt . . . . . . . . . I33

37 Derived price elasticity for the commodity
groups and their budget proportion . . . . l4l

38 Derived price elasticity for selected
commodities and their budget proportions . I42

39 Price elasticity coefficients from derived
mathematical model and from empirical
data for onions and tomatoes . . . . . . . I43
40 Projected population in Egypt, l970-l985 . I47
4I Projected population in Egypt by age

groups in I975, based on the constant
fertility rate assumption. . . . . . . . . 150

ix

Table Page

42 Population of Egypt according to location,
and percentage of rural population in
census years . . . . . . . . . . . . . . . I52

43 Projected population in I975 and I985 and
the percentage change from I964. . . . I62

44 Projected annual rate of growth in
national and per capita income and per-
centage increase in per capita income,

I964- I975 and I964- I985. . . . . . . . . . I62

45 Percentage change in demand for commodity
groups from I96 to I975 and I985. . . . . I63

46 Consumption expenditures projections for
commodity groups in I975 and I985 as
compared to their I964 level . . . . . . I64

47 Percentage change in demand for selected
commodities from I964 to I975 and I985 . . I65

48 Consumption projections of selected
commodities in I975 and I985 as compared
to their I964 level. . . . . . . . . . . . I66

49 Percentage distribution of annual total
expenditures on food products in I958/59,
I964/65, I975 projections and I985

projections. . . . . . . . . . . . . . . . I67

50 Projected per capita consumption of
selected commodities in I975 and I985
as compared to their I964 level. . . . . I68

5l Estimates of income elasticity of demand
in several countries . . . . . . . . . . . I83

Figure

LIST OF FIGURES

Page

Net food supply per capita I934/35-
l963/64. . . . . . . . . . . . . . . . . 34

Lorenz curves for total expenditures
and food expenditures in urban and
rural areas. . . . . . . . . . . . . . . 77

Lorenz curves for cereals and starches,
and meat, fish and eggs, in urban and
rural areas. 0 O O O I O I O O O O O O O 78

Lorenz curves for fruits, and milk and
dairy products in urban and rural
areas. I O O O O O O O O O O O O O O O O 79

A logarithmic and a log-inverse food
expenditure functions. . . . . . . . . . 92

xi

LIST OF APPENDICES

Appendix Page
A Some considerations in combining cross-
sectional and time-series data in a
single model . . . . . . . . . . . . . . . I94
B I Annual and per capita incomes and

investment, in current prices and in
constant I959/60 prices. . . . . . . . . . I98

B 2 Indices of national income, per capita
income and investment in current prices
and in constant l959/60 prices - I959/60
= 100. o o o o o o o o o o o o o o o o o o ‘99

B 3 National income by sectors, I959/60 to
I964/65 in constant l959/6O prices . . . . 200

B 4 National income indices by sectors,

I952/53 to I964/65 - I959/60 = IOO . . . . 20I

B 5 Average annual compound growth rates of
aggregate output and its components in
different periods. . . . . . . . . . . . . 202

B 6 Index numbers of agricultural output -
major commodity groups . . . . . . . . . . 203

B 7 Export market value of the U.S. agri-
cultural exports to the United Arab
Republic under specified government-
financed programs, I955-l963 . . . . . . . 204

B 8 Net food supply per capita (calories per

day) l948/49 to l963/64. . . . . . . . . . 205

B 9 Estimated coefficients from the double
logarithmic model for the commodity groups
in urban areas . . . . . . . . . . . . . . 207

B IO Estimated coefficients from the double
logarithmic model for the commodity groups
in rural areas . . . . . . . . . . . . . . 208

Appendix

B II

B l2

B l3

Page
Estimated coefficients from the double
logarithmic model for individual
commodities in urban areas . . . . . . . . 209
Estimated coefficients from the double
logarithmic model for individual
commodities in rural areas . . . . . . . . 2I4

Annual consumption per household of major

food products as related to income level
(expressed in total expenditures) in

urban areas - I958/59. . . . . . . . . . . 2I9

Annual consumption per household of major

food products as related to income level
(expressed in total expenditures) in

rural areas - 1958/59. . . . . . . . . . . 220

XIII

CHAPTER I

THE PROBLEM

l.l Introduction

 

The annual income per capita in Egypt is about $I40 which
places her among the poorer countries in the world. Where the
per capita income is low, a high proportion of it has to be
spent on agricultural products, since these are the first
necessities of life, especially in tropical countries. Also,
where the level of education is low, the productive capacity
of the population is tied closely to the natural environment.
A country with a low per capita income can spend very little
on education and health care and its people cannot look far
ahead, with the result that they tend to prefer activities and
investments with immediate benefits.

Egypt has shown a marked desire to develop economically.
This desire has been expressed by the government, but nonthe-
Iess is strongly felt by many different groups of the population.
This in turn, has resulted in the adoption of the first five
year plan that was put into operation in l960/I96I. The
principal objective of the plan, which contains a broad ten-
year plan composed of more specific five-year plans, was to

double the national income over a period of ten years. This

means an annual rate of growth of about 7.2 percent, which,
given the Egyptian population growth rates, implies an
annual growth rate in per capita income of about 4 to 4.5
percent.

The striking characteristic of the first five-year plan
was the major importance given to the industrial sector.

The sector was programed to contribute more than half the
planned increase in income (Table I). Together with agricul-
ture, the two sectors were expected to provide nearly 75
percent of the planned increase in income over that period.
The planned growth in services was relatively small. The net
result was planned to be a structural transformation as the
share of industry in national income could increase from

2l percent to 28 percent of total income. In the second
five-year plan, services were to play a larger role with a
relatively slow rate of expansion in both agriculture and
industry as compared to the first five-year plan.

By the end of the first five-year plan in I964/65, the
achieved growth rate was actually about 6 percent which
approaches the plan target. If this growth rate continued,
the national income would double over a period of l2 years.

I.2 The Relation Between Economic.
Development anleSOd Consumption
As the development process takes place, major changes occur

in the structure of the demand for food which affect the flow

Table I. Plan Income Targets by Sectors (In million LEl
at constant l959/6O prices)

 

 

 

 

 

 

National Flmplied AnnuaT

Income Target Income Compound Growth Rates

Sectors I959/6O I964/65
1959/60 1964/65 1969/70 to to

z, I964/65 I969/70
Agriculture 400 5l2 627 5.I 4.I
Industry 273 540 802 l4.6 8.2
Services2 609 743 Il35 4.4 8.9
Total I282 I795 2564 7.0 7.4

 

ILE stands for Eg ptian pound for which the official rate
of exchange is LE I = 2.30 at the present time.

2Includes trade and finance, transportation, housing,
education, health care, and other services.

Source: UAR National Planning Committee, General Frame of
the Five Year Plan for Economic and Social
Development, Julyij960-June I965, Cairo, I960.

of food products from farms to consumers' markets. As family
income rhes, there will usually be an internal transfer from
agricultural to non-agricultural employment. Labor and other
production factors will migrate to industry. These changes may
be slow due to psychological inhibitions, fewness of profitable
opportunities for industries, and the associated costs of
urbanization. They also may be slow in the economies where
gross domestic product is growing only slightly faster then
population. However slow, the important point remains that
these transfers must be compatible with the rise in per capita
consumption levels. It can be said that the most important of
these structural changes are the increasing income per capita
and the mass movement of labor and other factors of production
from rural to urban areas.I

The balancing problem is further aggravated in most less
developed countries by the fact that the sharp decline in the
death rates has not yet been followed by any significant change
in the gross reproduction rate. The result is a "population

explosion".2 Population growth alone can thus be estimated

 

lRobert Stevens, ”Elasticity of Food Consumption Associated

with Changes in Income in Developing Countries”, Foreign
Agricultural Economic Report, No. 23, Economic Research Division,
UO§YOD0K., 1965.

2United Nations, Dept. of Economic and Social Affairs,
”The Future Growth of World Population”, Population Studies
No. 28, N.Y., I958.

to require an increase in food production by as much as 3
percent annually to maintain present levels of food consump-
tion.

There is no doubt that the population increase can
hamper the process of economic growth. It has been estimated
that the developing countries as a group achieved a rate of
increase in GNP of 4.4 percent a year during the period
(l950-l960) as compared to a rate of 4.0 percent for the
European Free Trade Area countries, yet the population grew in
the former at an average annual rate of 2.0 percent, compared

1 The conclusion with

to a rate of .5 percent for the latter.

regard to the rate of growth of GNP per capita is quite obvious.
It has long been observed that "economic development and

population growth are related in mutual causation”.2 Egypt

provides an interesting case study for those wishing to test

theories about the interrelationship between the population

and economic growth.1 A high p0pulation growth rate of about

2.5 percent, together with rapid social change contributed to

the severe food shortages in certain commodities that the

country has recently experienced. This trend is expected to

continue as the interaction of economic deveIOpment and social

factors culminates in an increased demand for food, thus food

 

IH.J. Arnold, Aid for Development, the Bodley Head, London,
I966, p. 24. 7

2Report of the Ad Hoc Committee of Experts on Long Range
Program of Work in the Field of Population, Economic and
Social Council, United Nations, Oct. I964, p. 6.

production or food imports have to be increased greatly to meet
the growing demand which will be intensified by the suspected
high income elasticity of demand.

The accumulating evidence of what seems to be a chronically
precarious food situation in Egypt has in recent years increased
the general concern about its food problems, and heightened
the interest in carrying out this analytical investigation.
'While high population growth rates are a major factor in the
increasing demand for food, the high income elasticity of demand,
urbanization, and rapid social changes will also have a great
impact on the future food consumption patterns in the country.

In addition to its apparent importance, food has a
unique place in the economies of less developed countries, for
it represents the major product of the rural sector or the
so called ”subsistence sector”. It is possible with few
assumptions about the structure of the economy to determine
the outcome of the interacting prices and income elasticities
of demand in the two sectors on the consumption of food and/or
any other product, and see how the rural migration mechanism

would operate.I

 

ISee: Stephen Enke, Economics for Development, Prentice
Hall, Inc., I965, pp. (127-139).

l.3 The Objectives and Basic Hypotheses

Economic planning requires much information about the
changing structure of the Egyptian food economy during the
process of economic growth. Such information would facilitate
the more efficient and more economical use of planning techni-
ques in programming development projects in the country.

In this study, the objective was to estimate the income
and price elasticities of demand for major food products, and
then to use these estimates, along with estimates for the
annual rate of growth in GNP, and the projected rates of
population growth to estimate the future demand for food
products. Specifically the objectives of this study are:

(I) To investigate the determinants of the demand for

food and their magnitude.

(2) To estimate income and price elasticities of demand
for food in general, for food commodity groups, and
for basic food items.

(3) To utilize this information in projecting future
consumption levels for different commodity groups
and specific food items in I975 and I985, and to
relate these projections to the economic development

plans and the long-run potentials of agriculture.

The

study therefore focused on estimating parameters

determining the demand for food. To accomplish this, different

types of

and more

1.

The
(I)

(2)

(3)

(4)

data, time-series and cross-sectional, are utilized,
than one model of analysis is introduced.

underlying hypotheses for the study were as follows:
The income elasticities of demand for food in general
and for most individual food items, are relatively
high in Egypt as compared to developed nations.
Major differences in consumption patterns do exist
between urban and rural areas, with urbanites con-
suming more food in absolute terms, but spend
relatively less of their incomes on it.

The income expenditure elasticity differs signifi-
cantly from the income quantity elasticity, at least
for certain items such as meat, fish, and fruit,
where qualitative differences exist. This is mainly
because, other things equal, as income rises, a
family tends to consume better qualities and thus
expend more.

Increasing per capita income is the most significant

factor affecting the demand for food.I Further, the

 

tion.
for food.

IThe term ”the demand for foodll deserves a word of explana-

We recognize that there is no such things as a “demand“

Instead, there are demands for meat, wheat, eggs,

potatoes, and other products that provide food. However, in this
context, what is actually meant by the elasticity of demand for
food is not the quantity elasticity but rather the expenditure
elasticity based upon total food expenditures.

demand for animal proteins and fruits and vegetables
will rise faster as the income elasticity of demand
for these products is expected to be relatively high.
”(5) With the present trend in production and the rate of
population growth and unless some drastic measures
are taken to increase local food production,
increase food imports, or a combination of both; the
country will face severe shortages in many food
products by I975.

The study starts by introducing the basic features of the
Egyptian economy, which is believed to be very important as a
background within which the nature and scope of the food problem
can be evaluated. Chapter 3 discusses the concepts and techni-
ques used in the analysis, and introduces the reader to the
sources of information utilized in the study.

Chapters 4 and 5, concentrate on the results of the
statistical analysis. Each starts with an introductory section
that explores the ideas of using the types of data used, its
advantages and limitations. The two chapters are related to
each other and in fact complementary. The income elasticity
coefficients obtained and presented in chapter 4 are used in
formulating the combined model in chapter 5. The price
elasticities presented in section 5.I are used to estimate

money flexibility and therefore to estimate price parameters

ID

for commodities for which reliable time series data were
not available.

Chapter VI presents an overall projection of the demand
for food products in I975 and I985. This chapter is followed

by a summary and conclusions.

CHAPTER II

THE BASIC FEATURES OF THE EGYPTIAN ECONOMY

2.l Introduction

 

During the forties and early fifties, the Egyptian economy
was based on private enterprise system with a very limited
government role. Distribution of income and wealth were
extremely unequal. The new regime that took over in I952 had,
as its immediate purpose, the carrying out of land reforms

I on the ownership of

which placed a ceiling of 200 feddans
land. The ceiling was subsequently lowered to IOO feddans.
Land possessed by individuals in excess of that amount was
expropriated and gradually redistributed. The Suez war of

I956 carried with it the nationalization of British and French
properties, which involved the main part of the banking and
insurance system and other segments of large business and
industries. These prOperties, which were managed by specially
assigned governmental bodies, formed the foundation of what was
to be known later as the public sector.2

An economic planning was begun on a large scale in l958/59,

private enterprise was expected to play an important role in the

 

IA feddan, the area unit in Egypt, is equivalent to l.038
acres, or 4,20I square meters.

2See: Bent Hansen and Girgis Marzouk, Develgpment and -
Economic Policy in the U.A.R. (Egypt), North-HOlland Publishing
Company, Amsterdam, I965.

ll

l2

implementation of the plan by fulfilling the savings and invest-
ment targets laid down by the plan. Nevertheless, in I96l
large-scale nationalization followed which involved all big
industries and trade enterprises. The main economic justifi-
cation for such a decision was to bring investment and savings
under the direct control of the government. There were other
political motives behind the nationalizations, for the role

of private enterprises in the implementation of the plan had
never been well defined. Nor was there on the part of govern-
ment, any encouragement or incentives for private investors

to participate effectively in the plan.

These gradual and subsequent steps have in effect trans-
formed the country's productive system from a private enter-
prise market economy into an unique mixed system in which the
most significant economic decisions concerning production and

investment are taken by the state.
2.2 Land Area and Agricultural Production

The area of Egypt is aboutcne million square kilometers,
or nearly 238 million feddans, of which only about 5,967,000
feddans or about 2.5 percent, are cultivated. However, as the
average rate of crOpping per feddan is I.68 crops 3 year, this

gives the country an equivalent to IO,954,000 feddans of cropped

l3

area. It is important, in this respect, to distinguish between
the cultivated area and cropped area. The cultivated area is
the number of feddans actually planted, whereas the cropped
area, also expressed in feddans, is the cultivated area
multiplied by the number of crops grown annually under the
perennial irrigation where several crops can be planted in
rotation in the same area. The area that can be cultivated
depends entirely upon the elevation of the land in relation
to the nearest water supply.

The country's dependence on the agriculture sector can
be emphasized by pointing out that not only over one-quarter
of the national income and one-half of its employment is currently
derived from agriculture, but also large portions of industry,
commerce, finance, and transporation are centered around the

agricultural sector.1

Despite the notable expansion in the
industrial production and its concomitant impact on the economy,
agriculture has continued to play a principal role in the
country's economic Iifei'

The agricultural sector in Egypt has made four major contri-

butions to the process of economic deveIOpment:

 

IGamal Eleish, “The Applicability and Utilization of the
Input-Output Model in a Developing Economy: the Case of
Egypt Examined”, a paper presented at the International V
Conference on Input-Output Techniques, Geneva, September I96l.

l4

Table 2. Net Value Added and the Size of the Working Force
by Sectors in l959/6O and I964/65

 

 

 

 

 

 

Sector Net Value Added (LE mill.) Labor force (000)

l959/6O I964/65* l959/6O I964/65
Agriculture 399.9 5ll.7 4220 4660
IngT26EKIETEy 273.4 540.3 632 846

Transport and

communication 97.5 II7.5 2I9 226
Financial services l3.2 l8.6 2l 30
Others 593.5 725.6 I8ll 2098
Total I3775. l9l3.7 6903 7860

 

*plan targets

Source: National Planning Committee, General Frame of the First
Five-Year Plan, Cairo, I962.

 

IS

(I) It has stimulated the growth of basic agricultural
inputs industries such as chemicals and farm machinery. The
development of the fertilizers industry illustrates this point.
The local production of fertilizers which started in I947 at
a modest level of IS thousand tons increased in thirteen years
to reach 443 thousand tons in I960.‘l

(2) The implementation of many agricultural development
schemes and the rise in net cash income of the farm population,
provided a rising demand for many manufactured products.

(3) The expansion and diversification of agricultural
exports made available the foreign exchange needed to pay for
the import of machinery and capital goods upon which economic
growth depends.

(4) Finally, the agricultural sector has carried the
heavy burden of the provision of increased food supplies to
meet the growing and more diversified demand for food. This is
indeed a critical and major task. Failure to expand food
supplies to meet the rising demand is bound to jeopardize economic
grwoth since there is likely to be a notable rise in food prices
which could add to the inflationary pressure. Though the country
has had to supplement the locally produced food supplies by

Imports, particularly wheat, the local production has managed to

Ineet a good part of the food requirements.

 

IUAR Ministry of Agriculture, Annuaire Statistigue, I964.

Table 3.
(in L.E. million)

l6

Value of Agriculture Output I960-I964

 

 

 

 

Groups of products I960 l96l I962 I963 I964
field crops 380 326 397 405 453
vegetables 39 42 5l 60 67
fruits 26 24 29 32 33
animal products I05 IIO II6 l3l I63

Total 550 502 593 628 72l
Source: UAR Ministry of Agriculture, The Monthly Bulletin of

Agricultural Economics and Statjstics, Vol.7F7,

July I966, pp. (55-56).

 

I7

A variety of crops are grown in Egypt. With the exception
of cotton, rice, and onions, they are grown mainly for domestic
consumption. The index of agricultural output shows an
average rate of increase in total production of about 6 percent
during the last decade.I It is desirable, for the purpose of
this study, to separate food and non-food products.” Over the
period from I937 to I960, there were times when food output
grew ahead of the population. However, over the whole period
food production rose by 46 percent while the population
increased by 64 percent, a divergence that resulted in a lO
percent fall in per capita food supplied from domestic sources.2

The major field crops grown in Egypt are cotton, wheat,
corn, rice, maize, clover, onions and sugar-cane. Of less
importance are beans, nuts, sesame, and flax. The area put
to field crops amounts to about 9,750,000 feddans annually, or
about 94 percent of the crOpped area. To this group of farm
products about 63 percent of the value of agricultural output
is attributed.3 Cotton is the principal cash crop and the main

source of foreign exchange from the agricultural sector. It

makes up about I6 percent of the field crops cropped area and

 

IDonald C. Mead, Growth and Structural Change in the
Egyptian Economy, Richard Irwin, Inc., I967, P. 56.

2Donald c. Mead, lbid., p. 59.
3The figures refer to I964 data.

l8

accounts for about 35 percent of the group's output value.
Egypt is the main supplier of long-stable cotton to the world;
it produces about 40 percent of the world's long varieties

and nearly 50 percent of the extra long-stable cotton. About
30 percent of the total annual cotton production which amounts
to around 450,000 tons is locally manufactured.

Wheat normally occupies about l3.5 percent of the field
crops area and makes up to l2 percent of its value. Rice is
the second largest agricultural export commodity and is one of
the crops that will be expanded in acreage after the completion
of the Aswan High Dam. Onions form the third largest agri-
cultural export as about 40 percent of the crOp, or about
260,000 tons are exported annually. Major legume crops include
beans, lentils, and peas. The principal oil crops grown in
Egypt include groundnuts and sesame.

Vegetables and fruits together occupy nearly 7.I percent
of the total cropped area and normally make up about I4 percent
of the value of agricultural output.

Livestock production represents an important segment of
the agricultural production as it accounts for about 23 percent
of the value of agricultural output of L.E. I68 millions.

Table 4 shows the area, production and value for total

output for major agricultural products.

I9

 

 

 

 

 

 

 

 

 

Table 4. Area, Production, and Value of Total Output for
Major Crops in I964
value of Percent of
Area Production total output cropped value
area of ag.
output
fEddans metric tons LTE.
(000) (000) (000)
A. Field Crops
cotton I6Il 504 l57.083 l4.7 2l.8
wheat l295 l,499 53,959 ll.8 7.5
corn 1660 1,934 57,172 15.2 8.0
rice 962 2,036 40,098 8.8 5.6
beans 408 340 2I,58l 3.7 3.0
millet 494 740 I8,932 4.5 2.6
sugar-cane I34 5,l50 I2,389 I.2 I.7
onions* I34 647 6,923 I.2 .9
others 348I 84,486 3I.8 ll.7
Total
field
crops lO,l79 452,62l 92.9 62.8
B. Vegetables 608 4,278 67,303 5.6 3
C. Fruits I67 l,l44 33,223 I.5 6
D. LiEestock
8 aTry
Products l67,763 23.3
Total 10,954 720,910 100 100

 

*Production and value of onions

Source:

U.A.R. Ministry of Agriculture, The Monthly Bulletin
Agricultural Economics and StatistTCs, Vol. 17, July

include interplanted crop.

of

I966.

20

Egypt's annual per capita income places her among the
underdeveloped regions of the world. Despite this it is hard
to say that the Egyptian agriculture is primitive and backward
as is often the case for underdeveloped countries. According
to the figures of the Food and Agriculture Organization of
the United Nations for output per acre in I965/66, Egypt
ranked fifth for cotton (nearly 30% above the United States),
fourteenth for wheat, twelfth for maize, third for rice
(ahead of Japan), first for chick-peas, lentils and millet.
Japanese yields per acre are roughtly at the same level as
in Egypt for all major cr0ps grown in both countries.I This
is not to deny the possible scope for improvement and
increases in agricultural output particularly through the use
of better seeds with higher yields, but the point is that with
the present level of technology and organization, the agri-
cultural sector in Egypt stands close to intensive cultivation,
and that the scope of simple improvement by applying better
techniques would therefore seem limited.

This might explain the fact that the problem of rural

poverty in Egypt is much more difficult to eliminate than is

 

1Food and Agriculture 0r anization of the United Nations,
Production Yearbook, I966, Vo . 20, Rome I967.

“1:;—

2I

the case for many less devel0ped countries. Poverty seems t0'
beya result of ”unfortunate factor proportions and not merely
of an inefficient agriculture.Ill
After I970, the scheduled completion date for the Aswan

High Dam, the cultivable land is to be increased by almost one-
third. Yet, the high rate of the population growth is bound

to outstrip any possible increase in the land area. Unless
some economical method was to be found for desalting sea water
for agricultural purposes,2 or the rate of population growth

to decline rapidly, the land-man ratio is expected to fall

still further.'

2.3 The Demographic Picture

 

The Egyptian demographic picture represents an interesting
case study of a rapidly growing population with almost constant
age structure., Egypt was cited in the development literature

as an example of the overpopulated areas in which the famous

 

IDonald C. Mead, Growth and Structural Change in the
Egyptian Economy, op. cit., p. 74.

ZAt the present writing, there is a project under discussion
by a special committee in the U.S. government to study the utili-
zation of the nuclear power in desalting sea water. The U.A.R.
government has expressed its interest in the project, but it is
doubtful if any funds could be made available for the project.
See: Dwight D. Eisenhower, “A Proposal for Our Time”, Reader's

D’gest, June I968.

22

Arthur Lewis' type theories for economic development with
”unlimited” supply of labor is applicable.I Though these
theories have been challenged on various grounds, the fact
is that the population has continued to grow more rapidly
than the cultivable area, and that despite the substantial
expansion in employment opportunities outside agriculture,
the land-man ratio for both total and agricultural popula-
tions has continued to fall.

The population of Egypt, according to the I960 census,
had reached about 26 millions, and is estimated to be slightly
less than 30 millions in I969. Of the total p0pulation, about
62 percent are rural. The annual compound rate of population
growth based upon the period I947 - I960 is estimated to be
about 2.4%.2 This high rate of population growth creates a
very different problem for a country already believed to be
overpopulated, as it makes it harder for the expansion of the
so-called “modern sector” to match the increasing supply of
labor. Despite the continuous efforts to apply a policy

of population control, and the measures that have been taken

ISee: W. Arthur Lewis, ”Economic DeveIOpment with
Unlimited Supplies of Labor”, in: A. Agrawala and S. Singh
(eds.), The Economics of Underdevelopment, Oxford Univ. Press,
I963, pp, (40045497.

 

2The rate, r was estimated by applying Pn=Po (I+r)n

23

in the field of family planning this high rate of population
growth is expected to continue, the reason being the sub-
stantial decrease in the crude death rates which is estimated
-to be I6 per thousand in recent years as compared to a rate

of 3l.9 per thousand in I947/48.I

Table 5. Urban and Rural Populations in Successive
Censuses from I927 to I960

 

 

 

Year Urban Rural Total
000 000 000
I927 3,770 lO,707 I4,l77
I937 4,436 Il,484 I5,920
I947 6,262 I2,704 I8,966
I960 9,864 l6,l20 25,984

 

Source: U.A.R. Department of Statistics and Census,
ngulation Census, I960, Cairo, I963, Vol. 2

The committee of experts which was formed in I962 by

the government to study the population trends in the country

 

IH. Abdel-Aty, "Life-Table Functions for Egypt”, The
Mil-Bank Memorial Fund anrterly, Vol. 39, No. 2, April I96l,
PP. (350-7).

24

made five different assumptions with regard to the future trends
of the fertility rates for the period from I960 to I995.I

Five projections were made based upon the five assumptions
about the fertility rates. The projections are shown in
Table 6, with the first projection considered the maximum or

the upper limit and the last being the lower limit.

Table 6. Projected Size of the Total Population (in millions)

 

 

 

 

 

 

Assump- Estimated Population Compound Annual Rate of Increase
tion I970 I985 I960;I970 I970%l985 l960%l985
,I 34.5 52.5 2.8 2.8 2.8
2 34.0 48.3 2.7 2.4 2.5
3 34.5 48.4 2.8 2.3 2.5
4 33.4 43.6 2.5 I.8 2.l
5 3l.7 38.8 2.0 l.4 l.6

 

Source: U.A.R. Central Statistical Committee, Population
Trends in the U.A.R., Cairo I962, p. 53.

The fertility rates have major implications on the projected

p0pulation size as well as on the age distribution. However,

IThese assumptions are discussed in Ch. VI, Sec. 6.2.

25

we tend to believe that fertility rates are not likely to decline
substantially within a period of one or two decades. Therefore,
the Egyptian demographic picture in the coming twenty-five

years is expected to be dominated by some decline in mortality,
with fertility rates constant or falling very little. This

would imply a further rise in the rate of population growth.

It is thus reasonable and safe to assume that the growth rate
will be in the range of 2.5 to 2.8 per year at least within

the next fifteen years.

2.4 The Trends of National Income

Over the period from I952/63 to I969/60, national income
measured in current prices increased from L.E. 806 million
to L. E. I285 million,with a total increase of 59.4 percent
and an annual compound rate of annual increase of 6.9 percent.
However, over that period the wholesale price index for all
commodities increased by about l8 percent. In constant prices
of I959/60, national income increased from 949 million to L.E.
I285 million with 35.4 percent increase over the seven year
period, and an average compound rate of annual increase of
about 4.5 perCent.

Over the First Five-Year Plan period, from l960/6I to
I964/65, national income increased from its original level

of L.E. l285 million in I959/6O to L.E. I905 million in

26

I964/65 in current prices or L.E. I762 million in constant
prices of I959/60. The increase being about 48.2 percent in
current prices and 37.l percent in constant prices, with an
average compound rate of annual increase of 8.2 percent and
6.5 percent, reSpectively. If this trend is sustainable in
the future, it would be considered a remarkable achievement.
The original objective of the first Five-Year Plan was
to increase national income in constant prices over the five-

I The realized

year period by L.E. 5IO million or 40 percent.
increase over the plan period amounted to L.E. 477 million
or 37.l percent comprising 94 percent of the planned increase
in national income. I

As can be shown in Tables 7 and 8, the increase in
national income in l96l/62 was slight and as low as 2.4
percent in current prices and 3.3 percent in constant prices
of l959/6O over the previous year. This was mainly due to
the bad cotton crop in that year, and a resulting decline in
agricultural income by about L.E. 30 million below I960/6l
and l959/6O levels. Fluctuations in the cotton cr0p was

responsible also for the relatively high increase in national

income in l959/6O and for its decline in constant prices in

l955/56.

 

IU.A.R. National Planning Committee, 93. cit.

27

Table 7. National Income and National Income Indices in
Current Prices and in Constant I959/6O Prices,
I952/53 to I964/65

 

 

National Income NationaT’Ificome Indices

 

 

 

L.E. million l959/6O = IOO

Year current l959/6O current l959/60

prIces prIces prIces prIces
l952/53 806 949 62.7 73.8
I953/54 847 l026 65.9 79.8
I954/55 920 I096 7I.6 85.3
l955/56 965 IO37 75.I 80.7
I956/57 I067 l057 83.0 82.3
l957/58 II26 ll29. 87.6 87.9
l958/59 Il57 II6O 77.8 90.3
I959/6O I285 l285 IO0.0 IO0.0
I960/6l I389 I364 IO8.5 lO6.l
1961/62 1422 1412 110.7 109.9
I962/63 l537 l533 ll9.6 II9.3
I963/64 I7I9 I648 l33.8 I28.2
I964/65 I905 I762 I48.2 l37.l
Source: Calculated from Table Bl in the appendix.

28

Table 8. Percentage Shares of Sectors in National

Income, I959/6O to I964/65

 

 

/ I960] l96l7 T9627 I963/ I964]

 

 

Sectors I968 l96l I962 I963 I964 I965
Agriculture 3l.5 29.5 26.4 27.8 27.5 26.7
Manufacture and

Electricity 20.7 2l.9 23.I 22.7 23.5 23.5
Construction 3.7 3.2 5.2 5.4 5.8 5.3
Transport 7.2 7.5 8.3 8.3 8.7 8.6
Housing 5.7 5.4 5.4 5.I 4.8 4.6
Trade and Finance, 10.0 10.6 10.8 10.1 9.0 9.2
Other Services 2l.2 2l.9 20.8 20.6 20.6 22.l

Total IOO I00 IOO IOO IOO IOO

 

Source: Calculated from table 83 in the appendix.

29

All sectors of the economy contributed to the increase
in national income, but the proportion of that contribution
as well as the rate of growth varied from one sector to
another. Thus, over the period from l952/53 to l959/60, income
from agriculture increased by L.E. l53 million or about
6I percent. Income from manufacturing and electricity increased
by L.E. l39 million or IO9 percent. Income derived from
construction, transport, trade and finance together increased
by L.E. ll8 million or 78 percent. Income from other services
increased by L.E. 69 million or some 25 percent.

Over the same period, I952/53 to l959/60, the share of
agriculture in the national income remained around 32 percent.
Manufacture and electricity increased from l5.8 percent to
20.7 percent. Trade and finance showed a slight increase,
while construction and transportation remained around 3.5
percent and 6.5 percent, respectively. The share of other
services declined from about 34 percent to 27 percent.‘

During the first Five-Year Plan income from agriculture

increased by L.E. 65 million or about l6 percent. Income from

1See tables B3, B4, and BS in the appendix.

30

manufacturing and electricity increased by L.E. I48 million
or 55.6 percent. Income from construction increased by

L.E. 46 million or 98 percent. Transportation increased its
contribution to national income by L.E. 59 million or about
63 percent. Trade and finance increased by L.E. 33 million
or 25.6 percent. Income from other services increased by
about 42 percent. Therefore, of the overall increase in
national income, agriculture contributed l3.4 percent, manu-
facture and electricity 3I.O percent, construction and
transport contributed together 27.5 percent, while other
services contributed 3I.7 percent.

The low rate of growth in the agricultural sector as
compared to that of manufacture and electricity reflects the
emphasis placed on the later sectors in the allocation of
investments of the five year plan. However, the rate of
growth in agriculture is expected to increase over the second
five-year plan period due to the completion of the Aswan High
Dam, scheduled for I970, and other associated programs.

Because other sectors of the economy expanded at a higher
rate than agriculture during the first five-year plan, the
share of agriculture in the national income declined from about
3I.5 percent in I959/6O to 26.7 percent in I964/65, while the
share of manufacture and electricity increased from 20.7 per-
cent to 23.5 percent. This trend is expected to continue over
the years to come, because of the greater emphasis the planners

PUt on industrialization and urbanization.

3l

This expansion was accompanied by a growing deficit in the
balance of payment. Economic theory indicates that such a
deficit means either a higher level of consumption with the
production level and the rate of investment constant, or an
increase in capital formation with the consumption level
constant, or some combination of the two. Needless to say,
Egypt is not the only deveIOping nation with a balance of
payment problem. Though aggregate consumption has increased,
the apparent growing deficit in the balance of payments can
‘be mainly attributed to the planning efforts in recent years
which increased investments share of national income from

I3.3 percent in I959/6O to 2l.6 in I963/64.
2.5 Aggregate Food Consumption

Food consumption plays an important role in the economies
of less developed nations for several reaons. In addition of
being the major product of the so-called l'subsistence sector”,
the ratio of food expenditures to total income is a well
known criterion to measure the well-being of an individual
or of a nation. As Engel believed that "the poorer is a

family, the greater is the proportion of the total outgo

32

,which must be used for food".I

This famous relation depicts
the simple fact that the higher the proportion of income spent
on food, the lower will be the expenditures on other items
which, together with food, measures the individual's or the
nation's prosperity.

Food deficiency is a well known phenomenon in less
developed areas. However, in this respect one should dis-
tinguish between two different aspects of food deficiency.
Quantitative deficiency as measured in terms of calorie
intake is usually referred to as undernourishment, while
qualitative deficiency, or the nutritional side as measured
by the lack or adequacy of particular essential nutrients
is known as malnutrition or malnourishment.2

In dealing with the quantitative aspects of the food
problem in Egypt, we have to consider two sources of bias in
the aggregate figures: (I) that information on actual consump-
tion levels or the balance sheets is based upon several de-

rivations and cavmwsions that are subject to certain errors,

 

IAs quoted in C. C. Zimmerman, “Ernest Engle's Law of
Expenditures for Food”, anrterly Journal of Economics,

1932-3, p. 80.

2George R. Allen, llThe World's Food Shortage: Nutritional
Requirements and the Demand for Food”, Iowa State University
Center for Agricultural and Economic Adjustment, Food--One
Tool in International Economic Development, Ames, Iowa,7l965,

pp. 32-59.

 

33

and (2) that the national averages do not reveal the dispari-

ties in income distribution and food consumption.

Table 9. Per CaBita Food Consumption in Egypt, l948/50
to I96 /65

 

 

Total Animal

Years Calories Fat: Protein Protein

 

number/day grams/day grams/day grams/day

 

I948/50 2360 38.5 69.4 l2.5
I95l/53 2340 26.0 67.8 l0.7
I954/56 2470 37.0 7I.7 l2.2
l957/59 2530 39.5 73.5 ll.9
I960/62 2670 42.0 79.0 l2.2
I963/64 2890 47.8 85.5 I2.6
I964/65 2930 49.I 84.I l2.6
Source: Collected from: Food and Agriculture Organization

Food consumption in Egypt,

several volumes

of the United Nations, Production Yearbook,

 

in terms of calories, has shown

some improvement over the last three decades as it is shown

in figure I.

The composition of calories sources has remained

almost constant, with cereals, potatoes and other starchy

roots constituting about 70 percent of the total calorie intake.

 

34

mtocuo mmK

muozboca
>L_mp

6cm x__e saw.

mmmm w
cm_m .umme mmw

mm_nmuomo>

a m6_atu haw
m_mocou NM“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

x_pcoaa< "mULJOm
o
t
.4 2, 7. I. O 69 Ru 7. ,0 a, h. 25 9. I. no a; agnae
666 6655555555554639
//////////////////a
25 7. l 0 q, ‘8 7. ,o :5 h. 25 2 1. O a; no.8n4r
66665555555555hh33e
mow mmwwmmmwwwwwwmwm
00m
- ooo._
\
I \\ 1. .
\ \\\\\\ ooo N
\\ x x
- I oom.N
____ _ _ __L_
_ _ a _ t .s ooo.m
m mo_co m
“>66 Ema m0_Lo_muV p\ . _ u
ao\mmm_-mm\amm_ .ma_amo tom >_aa:m cook “62 ._ 6t:m_u

35

Calorie figures reveal the fact that the food situation in
Egypt cannot be characterized as a case of undernourishment,
as it is well above the calorie requirements in Egypt as
estimated by the Food and Agriculture Organization. It would
be rather a case of malnutrition. There is no clear-cut
definition of the extent of malnutrition in a certain country,
but it is a well-established criterion that if the ratio of
,calories derived from cereals, sugar, and starches to total
calorie intake exceeds two-thirds, that is a clear evidence
of serious malnutrition.l
The pressure of population growth and rising incomes upon
food supplies is increasing and will continue to increase.
The present increases in p0pulation practically swallow up
all increases in production and leave the general standards of
living still very low. From I937 to I947 the population grew
by about 20 percent, while food production seems to have risen
by slightly less than that, with the result that food con-
sumption per capita in I948 was somewhat around its prewar

level. Therefore, with the decline in the rate of increase

of food production combined with the p0pulation growth, Egypt

 

1Food and Agriculture Organization of the United Nations,
Second World Food Survey, Rome I952, p. I3.

36

shifted from being a net food exporter in I947 to becoming a
net food importer. These imports which accounted for L.E. 27
million in I950, and L.E. 35 million in I95l, have made it
possible to maintain the level of food consumption. The
failure of agricultural production to keep pace with the rapid
growth of population was a major factor behind the launching
of the planning schemes in l959/60.

From I952 to I960, the situation changed drastically.
Food production increased by more than 30 percent,1 while
the population grew by about 22 percent. Net food imports
decreased. Wheat imports decreased from over one million tons
in I952 to only IO thousand in I954, and none was imported in
I955. Maize imports declined substantially and rice exports
rose. Since I960, food imports began to rise again mainly
under U.S. Public Law 480, and there was a massive increase
of net food imports reaching about L.E. 62 million in I963.
The share of net imports to total domestic food supply rose
from 2-3 percent in I960 to lO-l2 percent in I963. Egypt's
agricultural imports from the United States under P.L. 480
increased from I2 million dollars worth of agricultural products
to $782 million in I966.2 These were mainly food products of

which wheat constituted a major part.

 

ISee table B6 in the appendix.

2See table 87 in the appendix.

37

Of all food items, cereals stand as a major source of
nourishment to the Egyptian population. Wheat, maize, and
rice provide almost half of the calorie intake. Over the
past two decades these cr0ps have continued to increase in
production. Cereal crops as a group increased by about 40
percent over the last two decades. Fruits and vegetables
almost doubled and livestock production increased by some 85
percent. Yet the population pressure allowed only a slight
improvement in the levels of consumption, as measured by

calories.

2.6 Summary

The population of Egypt in I969 is estimated to be slightly
less than 30,000,000. It is increasing at an annual rate of
about 2.5 percent, a fact of major concern to the government,
since the per capita income is extremely low and the amount of
arable land is limited. More than 60 percent of this population
is rural.

The most notable feature of the geography of Egypt is the
Nile river, in whose valley almost all of the pe0ple live. The
completion of the Aswan High Dam, scheduled in I970, will not
only increase the arable land by about one-third, but also
will put the river and all water resources under full control

for the first time in its long history.

38

The government of the UAR exercises substantial powers
over the national economy. The second five-year plan is under
way with full government control. Though efforts were directed
toward industrialization in recent years, agriculture is still
the basis of the national economy. The fertile Nile valley
yields grains, cotton, fruits, and sugar cane, at levels.
nearing the highest in the world.

The serious challenge is one of fostering rapid progress
in economic development, including production of more food to
help feed the growing pOpulation and to meet the rising demand
for higher quality diet. To achieve such a goal would require
enormous efforts if the p0pulation continues to increase at
the present rate and practically swallow up any increase in
production and leave the general level of living still very

low.

CHAPTER III

CONCEPTUAL FRAMEWORK AND TECHNIQUES

3.l Source of Information

This study depended heavily upon the information collected
in two household budget surveys. The first was conducted
during the period from October I958 to November I959 by the
Egyptian Central Statistical Committee with technical
assistance from the International Labor Office of the United
Nations. The second survey was conducted over the period from
April I964 to January I966 which was subdivided into four
sub-periods during which information was collected.

These surveys had, as a major purpose, the development of
a representative picture of family living conditions by
recording the amounts spent and the quantities consumed by a
sample of household units and to evaluate the economic status
of the households with regard to their total outlay and the
changes in their assets and liabilities over time.

Because of the difficulties of sample selection and data
collection, household budget surveys are rarely designed to
be completely representative of the entire population of a

country. In addition, such inquiries often omit certain

39

40

elements such as family income, tax payments, or cash holdings,
giving more emphasis to the consumption expenditures of the

households. In order to approximate a near ”cetris paribus”

 

situation, households were stratified into ”homogeneous
groups“. This stratification is based on the location of the
household, rural or urban,and household size.

Households were grouped by total expenditure levels. The
average expenditure by household groupings are the basic data
used for fitting the consumption function. In the calculation
of the regression coefficients these group averages were
weighted by the number of the households in each group. The
procedure used will be elaborated in section (3) of this
chapter.

The Central Statistical Committee of Egypt stated that
the household surveys were conducted in order to permit the
achievement of, among other things, the following two principal
objectives: (I) To obtain data useful in the construction of
a general cost-of-living index, and two sub-indices, one for
urban and another for rural communities, and (2) to measure
income elasticity of expenditures for various expenditure

categories.I

 

IInternational Labor Office, Reggrt to the Government of
The United Arab Republic on Family Budget Survey and on the
New Consumer PriceTTndices in theTSouthernTRegion of the U)A.R.
(Egypt), Geneva I960, p. 9.

4l

The reliability of information derived from such household
surveys depends upon, among other factors, the representativeness
of the sample. There were in fact two samples taken, one for
urban households and another for rural households. Table lO
shows the size of these two samples as compared to the total
population.

Total household consumption expenditures were measured
during both a month and a year for different categories of
goods. It was measured in both value and quantity for current
consumption commodities, and in value only for durable and
semi-durable goods, services, transfer payments and other

outlays.

Table ID. The Size of the First Household Budget Survey
and Its Percentage of the Population

 

 

 

 

Location Total number Total number Total %

of House- of households of individuals population total

holds In the sample in the sample (Jan.'60) population

Urban 3,I45 I7,468 9,864,000 O.l8

Rural 3,037 l6,487 l6,l20,000 0.l8
Total 6,l82 33,955 25,984,000 0.l3

 

Source: UAR Central Statistical Committee, Household Budget
Survey in the Egyptian Region, I958-59, Cairo,
April l96l. (in Arabic).

42

The major methodological difference between the second
survey of I964-I966 and that of I958-I959, is that it consists
of four different data collecting rounds in which a stratified
representative sample was drawn and each sample was studied in
a specific month as shown in table II.

The total number of the households studied in this second
survey was I3,8l7, out of which 9,3l7 were living in urban
areas and 4,500 were living in rural areas.

The procedures were used for collecting information were
as follows:

(I) For some items, mainly food,information was

obtained about the month preceding to the study
month.

(2) For other items, such as clothes, data were

collected for the preceding three month
period.

(3) For durable goods, data were collected for the

preceding year.

The complete results of this second survey are not yet
available. Some general information, together with the pre-
liminary results of its first round, were published in Arabic
in I967. These preliminary results were the principal source

of information used in this study.

43

Table II. Periods Covered and the Month of Study in the
I964-66 Household Budget Survey

 

~—

 

 

Survey Period Covered Method of
rounds from to Study
First April I, I964 March 3I, I965 April I965
Second July I, I964 June 30, I965 July I965
Third October I, I964 September 30, I965 Oct. I965
Fourth January I, I965 December 3l, I965 Jan. I966

 

Total expenditures have been used in many studies as a
fairly reliable estimate of the household income. The latter
is usually recorded with less accuracy than the former which
is equal to the sum of all expenditures on various items, food
and non-food. Total consumption expenditures consists of the
sum of actual consumption expenditures of the household plus
the value, at the local retail prices of the home produced
goods consumed by the household.

It should be pointed out, in this regard, that the share
of private gross disposable income in gross national income in
Egypt has fallen since I960, while public net disposable
income has risen. The household's control on the volume of

present consumption deferred for the sake of higher future

44

consumption, in other words, on saving, is no longer signi-
ficant. Although households are still free and are, in fact,
encouraged by the government to save, their savings form
only a small proportion of the total. Therefore, since
savings are not likely to change greatly as a proportion of
income, the eleasticity of demand based upon total expendi-
tures is not expected to differ much from that based upon
income.

As it will be explained in section (4) of this chapter,
time-series data will be used to supplement the cross-sectional
information. It is assumed that prices and indeed other
market variables are held constant in a single cross-section
sample, thus the income parameters or, more precisely, the
relationship between food consumption and total expenditures
can be estimated efficiently using such data. Although time
series data alone can be used to estimate both price parameters
and income parameters, there are methodological reasons for
pooling the two types of data that will be discussed later.

Per capita consumption figures were derived through
several conputations. The production, exports, imports,
losses, seeds and population data were used for calculating
per capita consumption for most products. With respect to

meat, data are collected for all meat from the slaughter

45

houses together with imports of frozen meats. Local production
data for milk and dairy products are for all sources of milk
together with exports and imports data. Consumption figures
are calculated for fresh milk, butter, ghee, and cheese.

The source of the population projections is the findings
of the research carried out by the committee of experts which
was formed by the Egyptian government in I962 to study the

population trends in the country.I

3.2 Analytical Procedures

Increasingly, it is being realized that aggregate con-
sumption functions derived from time-series data are not
sufficient for forecasting consumption or for demand analysis.
In addition to the difficulties involved in obtaining a homo-
geneous sequence of data that covers a reasonable length of
time, its enlargement in size by taking more frequent obser-
vations will raise the question of the presence of serial
correlation and the observations can no longer be considered
independent. In such a situation three alternatives can be

considered: (I) The use of cross-sectional data or as it is

 

IUAR Central Statistical Committee, Population Trends in
the U.A.R., Cairo, I962.

46

called the household surveys, (2) Pooling the two types of
data, and (3) The utilization of mathematical relationships,
based upon certain assumptions, to estimate the required
parameters. The present study, moved in these three
directions.

Three models of analysis will hence be introduced:

(I) A cross-sectional model of analysis, (2) A combined
model for cross-sectional and time series data, any (3) A
- model for estimating price parameters from cross-sectional
information.

In establishing this procedural framework, and in the
analysis there were four assumptions:

I. It is assumed that all other variables, not explicitly
included in the analysis were constant. This applies mainly
to institutional and political situations.

2. No substantial change in the price relationships
among different products within the country.

3. That there are differences in the consumption patterns
as between rural and urban areas to the extent that might
justify our dealing with the two sectors separately in various
occasions.

4. That a reasonable amount of homogeneity does exist in

the data. This assumption applies to both types of data used

47

in this study. In time-series data it is assumed that periods
of time are homogeneous except for the differences in the
explicit variables, and for random effect. This seems likely
in view of the relative shortness of the time period that the
data covers. The assumption for cross-sectional information
implies the homogenity of different households, except for

the differences accounted for in the measured variables and

a random error. This assumption cannot technically be proven

but it is necessary.

3.3 The Cross-Sectional Model

Calculations were performed on the original data to
provide information about the average annual expenditures for
food per household.I The households are classified according
to their total annual expenditures and their size within rural
and urban location categories.

The task was to analyze the relationship between consump-
tion (in terms of physical consumption or consumption
expenditures), annual income (as indicated by annual total
expenditures), location, and family size. The number of
consumption units rather than the actual number of household
members (also available), is used as a more meaningful indicator
of the household size, in order to take account of the age

variation.

 

ISee Tables B l3 and B l4.

48

The economic model we assume to identify the relationship

between these variables can be written symbolically as follows:
Df = f (I, S, L)

or that the demand for food Df, is a function of annual income

I, household size, S, and location L.

It is apparent that these variables are not the only ones
that influence the demand for food. The price factor is the
most important variable that we have omitted from the demand
function.I This has been done purposely because of the nature
of our cross-sectional sample which assumes a constant price
at the time of data collecting. With the exception of price,
the other major variables influencing the demand for food have
been included.

The choice of a function or a mathematical form to express
this relation is a matter of great importance. The calculated
Income elasticities will depend on the type of function that
has been fitted. The criteria used in the selection of the
proper function were: (I) statistical accuracy, (2) its
economic relevance within the framework of the economic theory,

and (3) simplicity of the computation.

 

IDifferences between urban and rural price relatives may
show up in the location variable.

49

Practical considerations limit the choice of the function
to four types: linear, semi-logarithmic, the logarithmic and
log-inverse functions.

A linear function is apparently the simplest and was
frequently used in the literature, but it is now generally
regarded as unsuited for the analysis of income-food consump-
tion relationships in particular. Its basic assumption is that
the coefficient of elasticity tends toward unity as income
increases indefinitely, which is not the logical relationship

I For this reason the linear

in the case of food consumption.
function was not utilized in this study.
The logarithmic function which takes the form:
log Y = a + 8 log X + H
has been extensively used in the literature. Its utilization
has been popular for its known convenient advantage that the
regression coefficient, B, is equal to the elasticity

coefficient. It permits an easier introduction of the effects

of the household size2 and is also considered satisfactory

 

IAs an example of the use of a linear function in a
consumption income relationships, see: R. G. Allen and A. L.
Bowley, Family Expenditure, London, I935.

 

2See: S. J. Prais and H. S. Houthakker,The Analysis of
Family Budgets, Cambridge (England), I955.

50

when “food consumption is expressed in terms of expenditures
rather than in terms of quantities"I and when the income range
Is sufficiently narrow. Like the linear function, the log-
arithmic function does not allow for a ”saturation” level of
food consumption as income grows toward infinity. This
llsaturation level” or the hypothetical highest possible level
of consumption is very high and approaches infinity.
The semi-logarithmic function that takes the form:
Y = 0+ 8 log X + u
is also used very often, and is preferred when food consumption
is expressed in terms of quantitites.2 It has been recommended
to fit the consumption of food ”necessities” though it does
not allow for a saturation level.3
The log inverse function with the general form:
log Y =<2_..§ + H
which has a saturation level seems to provide a better fit when
”food consumption is expressed in terms of quantities and when

1+

the data available covers a broad income range.

 

I .

L. M. Goreux, ”Income and Food ConsumptIon”, Monthl
Bulletin of Agricultural Economics and Statistics, FAD, Vol.
IX, No. IO, October I960, pp. (l-I3).

2L. M. Goreux, Ibid.
3S. J. Prais and H. S. Houthakker, op. cit.

AL. M. Goreux, op. cit.

5I

The analysis was confined to the last three functions,

namely, the logarithmic, and semi-logarithmic, and the log

inverse.

The Logarithmic Model

A logarithmic function is used to express the economic

model as a statistical model. The function will take the form:

log-ELI- : Bo, + I3Ii 16,314"-E + 82; log x1r+ 831 x2 + u.
r

r I
xIr
Where:
Xri = the average household expenditures of the r-th
total expenditures group, on the i-th commodity,
measured in Egyptian Pounds.
Xlr = the average number of consumption units per house-
hold in the r-th total expenditures group.
Yrt = the average households total expenditures in the
r-th groups, measured in Egyptian Pounds, at time t.
X2" = a variable that takes a value of zero for any
observation from the rural households, and a value
of one for any observation from the urban households.
Boi’ Bli’ B23, B3i = parameters to be estimated
”i a disturbance term, assumed to satisfy the ordinary

least squares assumptions.

X2r’ the dummy variable, is introduced in order to take some

account of locational and indeed social diffenances in consump-

tion between rural and urban households. It permits the mean

consumption to differ between them while the responses in

income or household size is equal for each. It is expected

that, in general, the coefficient B3; will be negative, partly

52

because urban households are expected to make less demand on
physical energy, and partly because social and educational
differences between rural and urban areas are likely to make,
other things equal, for a higher non-food expenditures by the
households of urban areas, with the consequent need for
economy in food expenditures.

The coefficient Bli represents the total expenditures
elasticity estimate. The logarithmic relationship has the
property of constant elasticity.l It will, ideally, represent
the way a typical household would behave under different
income situations, holding other variables constant. It is
used as a criterion to classify items consumed as necessities
or luxuries. Assumptions can be made as whether to consider
this coefficient as fairly representative of the income
elasticity of demand, or to reduce it by a certain factor
related to the marginal propensity to consume, to give a better
estimate of the income elasticity of demand. However, it can

be argued that total expenditures, which is used here as an

 

ITo prove this, assume a linear log relation for both
quantity and income log q = a + b log I, applying the principles
of differentiation we have I/q. dq/dl = b/I, or dq/dl. I/q = b.

53

indicator of income is a better explanatory factor because it

is more closely related to the permanent economic status than

income which is more likely to include transitory and unexpected

elements.1
The coefficient BZi represents the extent to which house-

holds consume more or less per consumption unit simply a

because, other things equal, they contain a large or small

number of consumption units. A positive value of B2i indicates

that large households consume more per consumption unit than (7

 

do small ones for a given income per consumption unit. We
may expect this factor not to be very important because the
commodities are broadly defined, and because our data were not
sufficiently detailed to permit the inclusion of such factors
as the "size“ of the product in our analysis. This coefficient
also may tell us something about the households' economies of
scale. As the size of the household increases, its food
outlay is expected to increase but at a lower rate than that
of the increase in size.

Having defined our model, a simple mathematical trans-
formation can be performed on it in order to facilitate

computation:

 

ILawrence R. Klein, An Introduction to Econometrics,
Prentice-Hall, Inc., I962, p. 58.

54

log Xrl = B . + Bll log Yrt + (1+ BZI -' Bll) log Xlr + BBIXZF 4' UI

OI

or

log Xri Boi + Bli log Yrt + B'Zi log er + B3i X2r +1.1i

where: B'Zi = l + 82: + B];

A final point should be considered. Since the data we
are using are averages for groups of households with certain
annual total expenditures rather than individual observations
for each household, it is desirable to weigh each observation
by the number of households on which it is based in order to
take into account the sampling error affecting the estimated
averages.

Let Nr indicate the number of households in the r-th
group, and let Zi and Zj any two variables in our original

equations, then,

ZN Z. - X N Z. Z N

z. />3N
I’I" II’ Y‘ I' II" I” r JI" 1"

r
Will be the weighted sum of products of Zi and Zj adjusted
for means. This process should be performed before applying

our model as well as any of the subsequent models.
The Semi-Logarithmic Model

The semi-logarithmic model as used in our analysis will

take the following form:

 

55

Xri Y
——=B.+log rt+BZ-Iogx +B3iX2r+u-
Ir OI YT: I lr I
where:
Xri = the average household's annual expenditures of the

r-th total expenditure group, on the i-th commodity.

xlr = the average number of consumption units per house-

hold in the r-th total expenditure group.

-<

the average household's annual expenditures on all

rt
commodities in the r-th group, at time t.
er = a variable that takes a value of zero for any observa-
tion from the rural households and a value of one for
any observation from the urban households.
BOi' Bli’ 82,, 833 = parameters to be estimated.
u; = a disturbance term, assumed to satisfy the ordinary
least squares assumptions.

X2r is a dummy variable introduced for the same purpose as
in the logarithmic model, that is to take account of the
locational and social differences in the consumption patterns
as between rural and urban areas.

The coefficient 8]: represents the sl0pe of the relation-
ship curve that corresponds to each value of the ratio XrI/XIr'
In this model we end up with a set of elasticities each corres-
ponding to a total expenditure group and an average number of

consumption units. The coefficient of income elasticity of

demand, or more precisely the total expenditure elasticity for

56

the i-th commodity will be equal to B13. Xlr/Xri-
Again, for the purpose of facilitating computations, the

model can be expressed in the following convenient form:

X'ri = BOI + BII I09 Yrt + 3'2i 109 Xlr + B3ix2r + U1
Where:

X'ri = Xri/Xlr

3'21= 321- BIi

The Log-Inverse Model

The following function implies that the consumption of a
particular item or group of items, expressed in log terms, is
a function of the total annual expenditures expressed in log-
inverse form, and the average number of consumption units
expressed in log form. Therefore, except for the total annual
expenditures,the function is similar to the double logarithmic
function explained earlier in this chapter.

Symbolically, the function can be written as follows:

Xlr + BZI log Xlr + 33] er + U-

Log-iil = BOI Bli I

Xlr rt

OI':

log Xri = BOi Bli xIr/Yrt + 8'23 log Xlr + BBi X2r + “i

as we define 8'2; l + 823

57

xri = the average household's annual expenditures of the
r-th total expenditures groups, on the i-th commodity.

xlr = the average number of consumption units per house-
hold in the r-th total expenditures groups.

the average household's annual expenditures on all

.<

fl

('1'
ll

commodities in the r-th group, at time t.

a variable that takes a value of zero for any

X
N
‘1

II

observation from the rural households and a value of
one for any observation from the urban households.

BOiI Bli» B23, BBi = parameters to be estimated.

u; = a disturbance term, assumed to satisfy the ordinary

least squares assumptions.

The total expenditures elasticity coefficient for this
function will be equal to Bli XI/Y. However, this coefficient
is calculated on the assumption that the logarithms used in the
function are Neperian logarithms. If decimal logarithms were
used a correction factor would be necessary and the elasticity

coefficient will be equal to 2.3026 B.X]/Y.

3.4 The Combined Model for Cross-Sectional
and Time-Series Data

In a single cross-section sample, it is assumed that

prices and other market variables are held constant, thus the

58

net relationship of demand and income can be estimated using
such data. The idea of pooling the cross-sectional and time-
series data, as will be elaborated in chapter 5, implies the
use of an independent estimate of demand parameters associated
with income in estimating the demand parameters associated with
prices.

Our model will be, with few modifications, similar to that
applied by Stone1 in his study to estimate the demand for a
number of food products in the United Kingdom.

The demand function for items of food products is assumed
to take the following logarithmic form:

log Xrt = 6+ 8 log Yrt + ylog Prt/Pt +<Slog Pst/Pt + urt

Where:
Xrt = consumption expenditures on the r-th food item per
consumption unit in period t.
Y = total expenditures per consumption unit in period t.

Prt = price index of the r-th food item in period t.

Pst = price index of the related s-th product in period t.
Pt = price index of all consumer goods in period t.
a,B,y,6= parameters to be estimated.

urt = the error term, assumed to satisfy the ordinary

least squares assumptions

 

IRichard Stone (and others), Consumers' Expenditures and
Behavior in the United Kingdom, l920él938, Cambridge University
Press, I954, Vol. I.

 

 

    
  

iv'u‘. " ’r‘Hu—LA T f. . '. 'Lrn‘r'firn“..‘ ‘ .' ‘ :v-ol
vaVfitW:Sfiwflfiwugfihwwwwuw-_

its

0— ' ‘ 7‘ Hui—.7
L: m jrwfilu-ln I . mat
. c I “ ,4 o - .

“"— ,v n' ..__5"l‘. ii?»
1 ‘A.~1oJ)2-'.-> ..g

P A, c- .

59

The conSUmption unit scale was constructed in the way
explained in the following chapter. As for the price index
of the related products Pst’ this variable will be omitted
for some equations, and will be expanded to include more than
one related product for other equations, depending upon the
nature of the product in question.

Total expenditures, as mentioned earlier, will be used as
an indicator of total income, a practice that has the effect
of scaling down the computed elasticity by a constant factor.

The pooling technique will be as follows:

(a) An estimate of B, the expenditure elasticity of demand
is to be obtained from the cross-sectional analysis
as described earlier. This estimate will be sub-
stituted directly into our new equation by forming
a synthetic variable from aggregative time series
data, call it 2t,

A
zit = log xit - B log Yt

Where:
’\ . . . .
B = an estImate of B, derIved from cross-sectIonal analySIs.
Xit = the per capita consumption expenditures on product i,

O

in time period t.

Yt = per capita income in time period t.

The multiple regression equation can be estimated for that
variable and the relative prices for the product in question

and also for the related product or products.

6O

Zit =°‘I + Bij

(2) If a serial correlation, which is defined as the corre-

Iog Pjt/Pt + ”i

Iation between a series of observations and the time series
lagged by one or more units, if found to be significant

and positive, the systematic disturbance can be partially
substracted by converting the variables entering the
equation into the first differences of the actual

observations.

3.5 A Model for Estimating Price Parameters
from Cross-Sectional Infbrmation

The idea behind Introducing this model is to estimate price
elasticities for commodities or groups of commodities for which
time series data are not available, by drawing conclusions about
these parameters from the budget pr0portion spent on the
commodity and the Engel elasticity.

The model is based upon certain mathematical relationships
and assumptions about the ”representative consumer”, and the
want structure independence of certain groups of commodities.

If these assumptions are accepted as reasonable it would be
possible "to go very far in this direction”.I

The model is applied for a number of commodities, and will

be further elaborated in chapter V.

 

IRagnar Frish, "A Complete Scheme for Computing All Direct
and Cross Demand Elasticities in a Model with Many Sectors”,
Econometrica, Vol. 27, April I959, p. I77.

CHAPTER IV
ANALYSIS OF THE CROSS-SECTIONAL DATA

4.I Some Considerations Concerning the
Utilization of Cross-Sectional Data
in Analytical‘Income-Consumption Studies

The use of cross-sectional data has made a great contri-
bution to econometrics that dates back to the days of Le Play
and Engel. Engel used Belgian cross-sectional data to give
empirical support to his famous assertion that the lower the
family's money income, the greater the percentage of that
income spent for food, and that the percentage of family income
spent for food is, therefore, the best measure for determining
the level of living. The term ”Engel curve" is still widely
used to show the variations in food expenditures associated
with the variations in the family income.

In recent years, cross-sectional data have been commonly
used in econometrics to form a basis for part of the macro-
economic model structure. This tendency has been influenced
to a great extent by the discontent about the nature of aggre-
gate data. Though the practice has its associated complica-
tions, household-to-household variation in expenditure, size,
and other variables; estimates of income elasticity of demand
and estimates of other parameters, can be calculated and used

in constructing an oVeraIl macroeconomic model.

6l

62

The popularity of using household expenditure surveys for
a wide variety of purposes stems from two basic reasons. The
first is the frequent occurence of collinearityI in time-series
data, a problem that seems to be less severe in cross-section
dataz. The second is the desire to construct more complicated
behavioral models3.

An argument can also be made that under the specific
conditions of less developed countries, the estimation of
the consumption levels can probably be obtained more efficiently
and accurately through the household expenditure surveys. Under
such conditions we usually lack consistent time-series samples
over a reasonable length of time. Though the size of the
sample can be enlarged by taking more frequent observations,
say by months or weeks instead of years, the problem of serial
correlation, or the interdependence between successive sample
values, sets a limit to the enlarging process, and large sample
theory can no longer be applicable. Cross-sectional surveys
provide large samples of hundreds or thousands of observations,
to which large sample theory is applicable, and useful estimates

of the parameters can be obtained.

ICollinearity is the tendency of many economic series to
move together in the same trend over time.

2Edwin KUh, ”Validity of Cross-Sectionally Estimated
Behavior Equations in Time-Series Applications“, Econometrica,

v0] 0 279 I959.
3Edwin Kuh, Ibid.

 

63

Although prices paid, and qualities consumed by different
households differ, and it is reasonable to expect that these
differences are systematically related to income levels, it
is assumed that over the given period of data collection,
the whole sample faces the same ”market situation“. Prices
paid, interest and wage rates, and other variables can be
said to be held constant over that period of time.I

Several problems that represent limitations in the use
of aggregate time-series data can be easily handled through
the use of such ”disaggregated information”. The study of
the frequency distribution of estimated random disturbances,
the study of the distributional effects, and the inclusion
of demographic, attitudinal, and other variables in the
behavior equations are examples of such pointsz. The data
can be handled so as to yield the form of the functional
relationships between dependent and explanatory variables.

These advantages are best realized when data on indivi-

dual households are available, but they can be captured to a

considerable extent even if, as in our samples, households

 

ILawrence Klein, An Introduction to Econometrics, op. cit.
p. 550

2Lawrence Klein, "Estimating Patterns of Savings from
Sample Survey Data”, Econometrica, Vol. I9, l95l.

64

are grouped by income levels or total expenditures brackets.
Such groups of households are rendered nearly homogeneous by
their cross classification on several characteristics simul-
taneously. In calculating estimates for the parameters,
these averages should be weighted by the number of house-
holds in each bracket to take into account the sampling
error affecting the estimated averages.

Yet, these surveys have their own set of problems that
include “the possibility of bias due to educating the reporting
families, loss of randomness due to 'fall out', and some
difficulty in keeping track of changing economic and social

characteristics of the sample”].

4.2 Definition of the Variables

The Household

The household unit is defined as comprising of persons
sharing the same roof and making their expenses, at least for
major items of the budget, out of a common fund. According
to such definition, family budget surveys should be referred

to as “household budget surveys". In addition to this general

 

IMarguerite C. Burk, I'Some Analysis of Income-Food
Relationships”, Journal of the American Statistical Association,

Vol. 53, I958.

65

definition, the Egyptian surveys retained as an additional
criterion, that any person staying with the household for
less than fifteen days during the preceding month of data
collection should be treated as a visitor and not as a member

of the household.

Total Expenditures

Total household expenditures were measured during both a
month and a year. The expenditure schedules used in the budget
make provisions for recording separately quantities and values
of home-produced products or products taken from the household
stocks that were consumed during the period in question. The
imputed values of services paid for in kind were shown
separately in the category of miscellaneous items and non-
consumption outgo. No general measure was taken, however, to
distinguish purchased consumption from consumption in kind so
as to arrive at an estimate of the degree of monetisation for
different categories of consumption items and various groups
of consumers. As it will be shown later in this chapter, home-
produced consumption of wheat, maize, and millets was distin-
guished from purchased consumption.

Expenditure categories include: (I) current consumption
commodities, for which information on value and quantity Was

obtained; (2) durable and semi-durable goods, which include

66

clothes, linen, furniture and household equipment, and other
durables such as automobiles, appliances, etc.; (3) services
that include medical expenses, education, sports, and cultural
pursuits, transportation and other expenses; (4) other outlays
which include rent and other related expenses, and other
expenditures on goods and services; and (5) transfer payment
such as gifts, donations, etc.

Although data has been obtained with regard to personal
savings and investment, and hence estimates for families'
personal incomes were computed, for a number of reasons this
information has never been published.

Data were published for rural and urban areas, and the
households were classified into thirteen groups based upon
their total annual expenditures. The number of households

within each group is shown in Table l2.

Consumption Expenditures

Consumption expenditures include all the items considered
in total expenditures except for the transfer payments. Data
were given during a year and a month for the consumption
expenditures classified into twenty-two broad categories.
These are: (I) cereals and starches, (2) dried legumes,
(3) meat, fish, and eggs; (4) fats and oils; (5) milk and

dairy products; (6) vegetables; (7) fruits; (8) sugar and

67

Table l2. Distribution of Households According to their
Total Expenditures Brackets,l958/59 and I964/65

 

 

Expenditures Number of households

 

 

 

 

E3;;%9;;es (in I958/59 survey I964/65 survey
pounds) urban rural total urban rural total
less than 25 6 '42 48 IO I8 28
25 65 279 344 IIO l23 233
50 I65 460 625 2II 2l8 429
75 25l 442 693 339 352 69l
I00 602 748 I350 lll2 997 2l09
l50 5l2 443 955 I329 854 2l83
200 356 246 602 lI65 6l5 I780
250 269 I34 403 96l 420 l38l
300 377 l27 504 l4l9 469 I888
400 307 89 396 I356 290 I646
600 I3I I8 I49 588 85 673
800 45 4 49 277 23 300
IOOO and more 59 5 64 440 36 476
Total 3l45 3037 6l82 93l7 4500 l38l7

 

Source: For I958/59, U.A.R. Central Statistical Committee,
Household Budget Survey in the Egyptian Region,
Cairo, April I96l. (in Arabic)

For I964/65, U.A.R. Administration of Public
Mobilization, The Preliminary Results of the
Household Budget SUrvey, I965, Cairo, May I967.
(In Arabic).

68

sugaries; (9) other foodstuffs; (I) beverages and refreshments;
(II) stimulants; (I2) fuel and lighting materials; (I3)
cosmetics; (l4) domestic cleaning materials; (l5) clothes

and linen; (I6) furniture and household equipment; (l7) other
durable goods; (l8) educational, cultural and sports pursuits;
(I9) medical expenses; (20) transportation; (2l) other

services; (22) miscellaneous.

Food Expenditures

 

Though food expenditures were not specified as a single
item, we consider food expenditures to include the first eleven
expenditures categories.

Further data for specified food items were made available
both on a yearly and monthly basis, and in terms of quantity
and value. Of these we will utilize information on the
following items: (I) wheat (home-produced and purchased);

(2) maize (home-produced and purchased); (3) millet (home-
produced and purchased); (4) flour; (5) bread; (6) rice;

(7) macaroni; (8) beans; (9) lentils; (IO) meat; (ll) poultry;
(I2) fresh fish; (I3) eggs; (l4) oils; (l5) saturated oils;
(I6) milk; (I7) white cheese; (I8) whey cheese; (l9) butter;
(20)ghee; (2l) potatoes; (22) onions; (23) tomatoes; (24)
citrus fruit; (25) dates; (26) sugar; (27) tea; and (28)

coffee.

69

The Household Size

Households, in both rural and urban areas, were classified
generally to three groups, small, medium, and large. Small
households are those with one to three members, medium with
four to six, and large with seven members and more. Data
were given for different expenditures brakcets for each group

of households simultaneously.

The Consumption Unit

Consumption expenditures on specific commodities are
usually experessed on a per capita or per household basis. It
is an established argument that both are rather misleading
concepts due to the differences in the age-sex composition

from one household to another.1

One way to handle this problem
is by constructing an age-sex equivalent scale in which each
age-sex type is expressed as a proportion of some standard
type.2 In the household budget surveys in Egypt, the adult
male is assumed to be equivalent to one consumption unit. The

classification is based on calorie requirements by age and

sex groupings. The scale is shown in Table I3.

 

'H. Wold and L. Jureen, Demand Analysis, John Wiley and
Sons, New York, N.Y., I953, p. 233.

2David William Price, Age-Sex Eguivalent Scales for United
States Food Expenditures - Their Computation and AppTication,
Ph.D. Thesis, MiEhigan State University, I963.

70

Table I3. Consumption Units Scale

 

 

 

Age Sex Consumption
units

Less than I3 years male and female .6

l3 years to less

than 50 male l.O
female .7

50 years and older male .8
female .5

 

Source: Same as Table I2

Using a scale of this type would improve the data, and

give a more meaningful comparison of food expenditure between

two individual households, or groups of households with

different age-sex composition.

Therefore,

in our analysis,

the household size will be expressed not in terms of number

of individuals, but rather in numbers of consumption units.

With such age-sex variation accounted for, other different

characteristics can be analyzed.

7l

4.3 Results of the Analysis

4.3.l Introduction

The present section will deal with the results drawn
from the analysis of the cross-sectional survey. It will
start with a depiction of the food consumption patterns as
revealed by the household budget survey. Then the results
of the statistical analysis in which the relationships
between the consumption expenditures for different food items,
i's and total expenditure brackets, r's are discussed.

Symbolically, xr; represents the average household's
expenditures of the r-th total expenditures bracket on the
i-th item. Variables xr] and xrlz represent the average
household's annual expenditures on l2 commodity groups as
follows:

xr] : cereals and starches.

sz : dried legumes.

Xr3 : meat, fish, and eggs.

xru : oils and fats.

er : milk and dairy products.

Xr6 : vegetables;

Xr7 : fruits.

xrg : sugar and sugaries.

72

xr9 : beverages.
erO : tea, coffee, and other stimulants.
xrll : other food stuff.

xr12 total food.

Variables from xr13 to xr46 stand for average household's
expenditures on particular food items measured in Egyptian
Pounds, while variables x'rl3 to x'r46 represent the quantities
consumed of these items measured in kilograms. These food items
are: .

xr 13 and x'rl3 : wheat (home produced).

xr 14 and x'rl4 : wheat (purchased).

xr 15 and x'rlS : wheat (total).

xr l6 and x'r16 : corn (home produced).

x 17 and x'r17 : corn (purchased)
x l8 and x'rl8 : total corn.

xr 19 and X'r19 : millet (home produced)

xr 20 and x'r20 : millet (purchased).
xr 2] and x'r2] : total millet.
xr 22 and x'r22 : flour.

x 23 and x'r23 : bread.

xr 24 and X'r24 : rice.

XI. 25 and x'r25 : macaroni.
xr 26 and x'r26 : beans.

xr27 and x'r27 : lentils.

xr 28
r 29
xr 30
Xr 3I
"r 32
xr. 33
xr 34
xr 35
Xr 36
xr,’37

xr 38'

Xr 39
xr 4O

xr 4l
xr 42
xr 43
xr 44
r 45
xr 46

and
and
and
and
and
and
and
and
and
and
and
and
and
and
and
and
and
and

and

Variables

x'r28
X'r29
x'r30
x'r3l
x'r32
x'r33
xlr34
x'r35
XIr36
x'r37
x'r38
XIr39
XIr40

-x'r4l

x'r42
x'r43
X'r44
x'r45
x'r46

73

meat (all kinds)
poultry.

fresh fish.

eggs. (in numbers)
oils.

margarine and saturated oils.
milk.

white cheese.

whey cheese.
butter.

ghee.

potatoes.

onions.

tomatoes.

citrus.

dates.

sugar.

tea.

coffee.

from x]; and x13] stand for average household's

annual categories, where r = I, 2, ..., l3. The value of this

variable for any certain bracket differs between rural and

urban areas and between different household types, as the

actual average total annual expenditures is used in quanti-

fying its value.

74

 

XII total annual expenditures less than L.E. 25.
x23 total annual expenditures from L.E. 25 to
less than L.E. 50.
X3i total annual expenditures from L.E. 50 to
less than L.E. 75.
X41 total annual expenditures from L.E 75 to
less than L.E. 100. F
i
x5i total annual expenditures from L.E IOO to E
less than L.E. ISO. 1
|
x6; total annual expenditures from L.E. ISO to
less than L.E. 200. I
x7i total annual expenditures from L.E. 200 to
less than L.E. 250.
X8i total annual expenditures from L.E. 250 to
less than L.E. 300.
x9; total annual ex enditures from L.E. 300 to
less than L.E. OO.
xlOi total annual expenditures from L.E. 400 to
less than L.E. 600.
x”i total annual expenditures from L.E. 600 to
less than L.E. 800.
Xl2i total annual expenditures from L.E. 800 to
less than L.E. IOOO.
x133 : total annual expenditures of L.E. IOOO and more.

The impact of the household size is expressed in terms of
consumption units Xlr: that is, the average number of consump-
tion units per household in the r-th total expenditures
brackets, where r = I, 2, ..., I3. The values that this
variable take are shown in Table I4 for different types of

households.

75

 

 

 

 

 

Table I4. Average Number of Consumption Units per Household
for Different Household Types.*

Annual

total Urban Areas Rural Areas

expendi-

tures

groups large medium small total large medium small total
I - - .83 .83 - 2.00 .78 .8l
2 4.00 3.00 .95 I.2I 5.00 3.I2 I.27 1.57
3 4.80 3.42 I.4I 2.09 5.09 3.25 I.82 2.49
4 5.I4 3.35 I.72 2.78 5.2l 3.40 I.89 3.09
5 5.27 3.5l I.85 3.43 5.48 3.58 2.IO 3.76
6 5.52 3.6I I.82 3.78 5.96 3.84 2.03 4.64
7 5.55 3.74 2.l3 4.I5 6.39 3.93 I.89 5.46
8 5.84 3.84 l.92 4.46 6.99 3.94 2.33 6.07
9 6.Il 3.84 l.92 4.90 8.l4 3.96 2.50 7.I6
IO 6.67 3.8l 2.07 5.28 9.07 4.08 2.25 8.09
II 7.04 3.86 2.00 5.37 9.94 3.00 9.55
I2 7.27 3.67 2.40 5.29 I4.00 - l4.00
I3 7.50 3.95 2.00 5.98 l3.25 4.00 ll.4O

 

*Small households are those with one to three members,

medium with four to six, and large with seven and more.

Source:

Calculated from:

UAR Central Statistical Committee,
The Household Budget Survey, op. cit.

 

 

76

As indicated earlier in chapter 3, x2r is a dummy variable
that takes a value of zero for any observation from the rural
areas, and a value of one from any observation from the urban

areas.
4.3.2 Food Consumption Patterns

The household budget data reveals certain information
about the equality or inequality in the distribution of total

annual expenditures, which is considered as a meaningful

 

indicator of income, in the distribution of food expenditures,
as well as expenditures on specific individual items or groups
of commodities.

This distribution can best be shown by a collection of
Lorenz curves. A Lorenz curve is a graphic device showing
the cumulative percentage relationships between two variables.
The diagonal line represents complete equality; that is IO
percent of the spending units spend IO percent of the total
expenditures. In figures 2, 3, and 4, cumulative percentages
of consumption units is shown against cumulative percentages
of expenditures.

It is obvious that total expenditures and total food
expenditures are more equally distributed in rural areas than
in urban areas. Two explanations can be offered in this

connection. First, qualities as well as prices paid, differ

Percentage of annual expenditures (urban)

77

Figure 2. .Lorenz Curves for Total Annual Expenditures
and Food Expenditures in Urban and Rural Areas

Percentage of consumption units (rural)

 

   

 

 

 

I00 90 80 7o 60 so _30 30 ' 20 10 o
90 _ Rural Areas ', I IO
80 * '1 20 2
3
L
70 I- ‘ 30 In
0
L
3
60 a « 40 .2
' ‘D
C
so » 3
f 50 §
1+0 ‘I‘ -I 60 E
. . c
c
m.
30 ~ . 70 u.
o
0
20 r « 80 31
H
c
. O 3
'0 Urban Areas 3 9 E1
7 _L 1 l l L 1 l J . .J
0 lo 20 30 4O 50 6O 7O ‘80 90 . -IOO

Percentage of consumption units (urban)

Source: Calculated from UAR Central Statistical
Committee, 22, cit.

Percentage of total expenditures (urban)

78

Figure 3. Lorez Curves for Cereals and Starches, and
Meat, Fish, and Eggs, in Urban and Rural Areas

Percentage of consumption units (rural

       
 
    

 

 

 

 

 

I00 90 80 7O 60 SO 40 30 20 IO 0
T r y y 1 w l I I
90 7 Rural Areas '0
80 _ . 20
70- '30
60*r 4 4O
50 - - 50
40I- 4 60
BOIP ‘ 7O
20 Urban Areas q 80
'01 d 90
0 ,. .0 20 30 40 30 60 70 750 90 100

Percentage of consumption units (urban)

Source: See Figure 2.

Percentage of total expenditures (rural)

 

Percentage of annual expenditures (urban)

79

Figure 4. Lorenz Curves for Fruits, and Milk and Dairy
‘1 Products In Urban and Rural Areas

Percentage of consumption units (rural)
100 29 89'. 20 60 SO" 40 30 20 lo 0

 

      
 

 

 
   

 

   

 

v 3 r 1 Ar r 1
900- I I '0 A
. 7.:
30L Rural Areas 20 a
. m
7o.a a 30 8
60°IP -II ‘50 g
' 3'
5°lr ~-so .3
3
“0-”- 1> 60 g
o
“5
30,1— _I. 70 0
m
t . - 3
20 Urban Areas 4 '80 é
L
'0‘ -II 90 :
J j J l 3 4 1 .!

 

o 16 20‘ 50 do so so 70 60' 90‘ . 100

Percentage of consumption units (urban),

Source: See Figure 2.

8O

much more in urban areas; and second, incomes, occupations,
social status, and tastes vary greatly in urban areas.

Fruits and meat are the most unequally distributed in
both rural and urban areas. Fifty-two percent of the consump-
tion units share only 32 percent of the expenditures on meat,
fish, and eggs; and 27 percent of the expenditures on fruits a
in urban areas. The corresponding shares of expenditures on
these items for 49 percent of consumption units in rural areas
are 38 percent and 34 percent. This inequality in the distri- ,7
bution of expenditures on these ”expensive” items of the food
budget, can be understood in the light of price and quality
differentials, as they are not considered among the "basic"
items, and therefore are not subject to price controls. The
lower income groups do not necessarily consume relatively
less quantities, but rather lower qualities.

Expenditures on food and beverages account for about 50
percent of the total annual expenditures in urban areas, and
more than 66 percent in rural areas, with an overall average
of about 58 percent, a very high percentage as compared to
22 percent in the United States, 23 percent in Canada, 47

percent in Yugoslavia, and 52 percent in Ceylon.I

 

IUnited Nations, Statistical Yearbook, New York, l96l,
pp. (297-9).

Table I5. Percentage of Items of Total Annual Expenditures
in Rural and Urban Areas, Based on the Household

 

 

 

 

Budget Survey, l953/59
Budget Items Urban Rural Weighted
. ,fi average

percent percent percent

Food and beverages 50.l 64.4 58.2

Clothing and linen 8.2 7.8 8.0

Fuel and lighting

material 3.5 3.5 3.5

Cosmetic and domestic

cleaning materials 9.7 l.6 5.8

Education 3.6 l.0 2.0

Medical expenses 3.0 2.0 2.5

Transportation 3.9 1.8 2.9

Other expenses (inc.

housing) l8.0 l5.9 l7.l

Total 100 IOO lOO

 

Source: Calculated from:

UAR Central Statistical Committee,

The Household Budget Survey, op. cit., pp. (355-6)

82

More'than one-third of the expenditures on food items goes
to cereals and starches. In rural areas, the share of cereals
and starches is as high as 42 percent. Meat, fish, and eggs,
the sources of animal protein, account for about one-fifth.
However, the distribution of total expenditures, as shown in
Table I6, does not reveal the whole picture as the prices of
most cereals and other ”basic'I items are usually controlled

and/or subsidized, while other items, such as meat, fish,

 

fruit are relatively more expensive, and out of reach for a

considerable segment of the consumers.

 

 

 

 

 

Table I6. Percenta e of Items of Food Expenditures in Urban
and Rura? Areas, Based on the Household Budget
Survey, 1958/ /59
Food items Urban Rural Weighted
average
.percent percent percent
Cereals and starches 27.4 4l.8 34.5
Dried legumes 2.6 4.2 3.4
Meat, fish, and eggs 2l.2 l7.l l9.2
Oils and fats 3.8 2.4 3.1
Milk and dairy products l2.8 l2.6 l2.7
Vegetables 7.4 5.I 6.3
Fruit 6.4 3.0 4.7
Sugar and sugaries 6.4 6.5 6.4
Other food stuff 5.8 2.0 3.9
Beverages 6.2 5.3 5.8
Total IOO IOO lOO
Source: Same as Table l5.

83

4.3.3. Comparisons of the Results
of the Three Models

The three single-equation models used in the cross-

sectional analysis were:

(I) A double logarithmic model: 5%
log Xri/Xlr = BOi + Bli log Yr/Xlr + BZi log Xlr
+ BBI log er + Ui

 

(2) A semi-logarithmic model: i
Xri/Xlr = Bo: + BII log Yr/X]r + BZI log Xlr +
BGI log X2: + Ul

(3) A log-inverse model:
‘09 Xri/Xlr = 30: - B]: XIV/Yr + 823 log XIV + B3l
log X2... 4' Ui

These three models are further simplified to take the
following more convenient forms:
(I) log xri = BOi + Bli log Yr + B'ZI log Xlr + B3i log X2r + Ui
where: B'Zi = 1 + BZi - Bli

 

IFor definition of the variables and mathematical symbols,
the reader is referred to Chapter 3.

84

(2) X'ri = 303 + Bli '09 Yr + B'zi ‘09 X'Ir + 53: '09 X2r + Ui
where: X'ri = Xri/Xlr

and B'Zi = BZI - Bli

(3) '09 Xri = BOi ' Bli XIr/Yrt + B'Zi ‘09 Xlr,+ B3i er + Ui
Where: B'Zi = I + BZI

The purpose of this transformation was mainly to facilitate
the computation process. Thus, the coefficients can be estimated
as desired as long as there is a mathematical relationship
between BZi and B'2i.]

In presenting the results, the following format was used:

The standard error of the estimated parameters were given in
parenthesis below the estimates. The coefficient of determin-

2

ation , adjusted for the number of degrees of freedom, was

 

1Though the estimate b2: can easily be calculated there is
a problem in calculating the standard error of the coefficient,

 

5b2i =I52bli + 52b2i - 2 COV- b1: ' b'zi

However, (kaowingz that the2 upper limit of the quantity Cov. b]°,
b'z; is (s bli - $2 b2i according to Cauchy- -Schwarz inequality,
which in most cases was a very small value, an assumption was
made that this covariance is equal to zero. This particular
coefficient was not used in the combined model, nor does the
calculation of its exact standard error affect the results in

any way.

2R2 = l - (l-Rz) N-l/N-P-I, gives the exact split of the
variance of the dependent variable into explained and unexplained
variance.

w“ .‘ ' "‘- ~r -.".W
I

85

denoted by R2. The standard error of estimate (or the standard
error of the residuals, i.e. the square root of the error sum
of squares divided by the number of the degrees of freedom),
was denoted by S - S is an estimate of how well the regression
line fits the data or the "goodness of fit."

The close examination of the results of the three models,
particularly the value of R2 and S, reveals that the logarithmic
function provided the best fit in terms of R2 and S, or the
analysis of variance for overall regression. The log-inverse
model provided a fit almost close to the logarithmic, while
the semi-logarithmic function provided for the least satis-
factory results in terms of the goodness of fit. It is for
that reason that the emphasis in the analysis was put on the
logarithmic and the log-inverse fittings.

The estimated relationships for total food expenditures
from the three models were as follows:

The double logarithmic function:

log Xri/Xlr = .3845 + .6428 log Yr/Xlr + .0259 log Xlr

(.0169) (.0093) (.0141)
+ .0150 log er
(.0040)
R'2 = .994 S = .0124

86

The semi-logarithmic function:

xri/xlr = -63.062l + 60.9745 log Yr/Xlr + 3.4556 log Xlr

(4.2787) (2.2537) (3.9620)
'oLl‘283 log er
(1.0152)
.2
R = .9651 s = 3.1282

The log-inverse function:

log Xri/Xlr = l.9824 + l6.l3l4 Xlr/Yr + .0698 log Xlr
(.0576) (.8576) (.9296)
- .OOIS log er

(.0011)
fi2 = .9873 s = .0437

The total expenditures elasticity coefficient for total
food was .64 from the logarithmic model, .84 from the semi-log
model, and .70 from the log-inverse function.1 The elasticity
coefficient based upon the logarithmic function is constant
over the range of the function. The log-inverse elasticity
depends on the total expenditures level, and the above value

was calculated at its mean. The semi-logarithmic elasticity

 

IThough these values are considered here, there is a
Iweighting problem involved and was discussed in the.next

section.

Table I7.

Values of R2
of Urban and Rural Areas

87

and S for the Combined Analysis

 

 

Semi}

Log-inverse

 

 

 

Commodity Logarithmic logarithmic ‘

Groups R'l' S RT 5 R" S
Cereals 8 starches .9972 .0343 .7573 .9644 ..98l9 .044]
Dried legumes .8943 .0609 .7702 .1796 .9645 .0650
Meat, fish, eggs .9967 .003I .934l .309] .9749 .0717
Oils 8 fats .9947 .0355 .9094 .l5l9 .9413 .0949
Milk and dairy

products .9890 .0638 .9224 .7405 .983] .058l
Vegetables .9907 .O49I .9225 .3366 .9559 .0809
Fruit .9934 .0555. .9133~ .5217 .9749 .0865
Sugar 8 sugaries .9929 .9425 .9245 .2882 .9785 .0546
Beverages .9922 .0466_ .9l54 .3I66 .9635 .9756
Tea, coffee, 8

other stimulants .9760 .lO66 .87I0 .8235 .9527 .I2l2
Other food stuff .9494 .lO98 .8769 .3465 .83l4 .l9OO
Total food .9994 .Ol24 .965l 3.I282 .9873 .0437

 

88

 

 

 

 

 

Table 18. Values of'R—2 and S for Urban Areas

Commodity Logarithmic Iongithmic Logzinverse
9"°”PS R" s R' s R",- s
Cereals 8 starches .9992 .0142 .8281 .2959 .9980 .Ol3O
Dried legumes .9877 .0626 .6585 .1068 .9826 .0394
Meat, fish eggs .9979 .0333 .9755 .Oll9 .9907 .0458
Oils and fats .9947 .0423 .8745 .l824 .9729 .9548
Milk and dairy

products .9938 .0596 .9825 .4591 .9937 .0366
Vegetables .9974 .0325 .9726 .2424 .9877 .0431
Fruit .9940 .0655 .9834 .2892 .9948 .0395
Sugar and sugaries .9973 .0457 .9521 .262l .9877 .0402
Beverages .9940 .0503 .9447 .3121 .9814 .0519
Tea, coffee 8

other stimulants .9918 .0721 .9789 .5339 .9581 .1065
Other food stuff .9916 .0406 .8366 .2779 .9570 .0583
Total food .9996 .0117 .9738 3.2745 .9875 .0422

Table I9.

89

.2
Values of R and S for Rural Areas

 

 

 

 

 

Commodity Logarithmic 109:??1hmic Log:inverse
grows R" s R" 5 RT 5
Cereals 8 starches .9979 .0170 .9759 .3662 .9975 .Ol83
Dried legumes .9984 .0283 .9801 .0761 .9965 .0232
Meat, fish, eggs .9971 .0228 .9440 .7008 .9893 .044]
Oils and fats .9828 .0476 .64l4 .0694 .9805 .0503
Milk and dairy

products .9843 .0555 .9527 .3089 .9971 .0238
Vegetables .9934 .0275 .9ll6 .1456 .9860 .0402
Fruit .9965 .0293 .9653 .1324 .9930 .0415
Sugar and sugaries .9931 .0317 .8689 .3199 .9746 .0610
Beverages .9880 .0424 .8384 .2761 .9680 .0709
Tea, coffee 8 1

other stimulants .9884 .0994 .8l8l .9637 .9424 .l429
Other Food stuff .9124 .lOIl .0380 .1072 .9049 .1053
Total food .9993 .0106 .9655 2.2065 .9934 .0322

__

90

/’

depends upon the level of food expenditures, and its value

I In fact, the

presented above was calculated as its mean.‘
log-inverse elasticity coefficient differs from l.8575 for a
total expenditures level of L.E. 20 per consumption unit to
.1061 for a total annual expenditures level of L.E. 350 per
consumption unit.2

Two consumption functions for food were derived from
estimated relationships between total annual expenditures
and food expenditures from the logarithmic and the log-inverse
fitting; they are shown in table 20 and in figure 5.

The preceding discussion has focused merely on intro-
ducing the results of the three models. However, a major
point should be emphasized at this stage, that is the sample
proportional representation of the urban and rural population.
As the discussion of the results was pursued further, a
"weighted” set of elasticity coefficients based upon the

separate estimates for urban and rural areas was introduced,

as elaborated in the following section.

 

B
IFor the semi-log model e = .4343 I ' Xlr
2For the log-inverse model e = 2.3026 (BI/Yr) - er

91

Table 20. Food Consumption Functions Derived From Logarithmic

and Log-inverse Models

 

 

Total Annual
Expenditures
per Cons. Unit
Yr/xlr

Food Consumption Functions

 

A log function derived

from: loghégi/Xlr -

.384 +

log Yr/

xIr + .0295 log x]r

A log-inverse function
derived from: log Xr;/
Xlr = 1.9824 + l .1314
XIr/Yr '1' .0698 log Xlr

 

 

12:84:29.8:- . 52.381.329.18: .5.

L.E. 4[.E. L.E.
20 17.8 .64 12.8 1.86
50 32.0 .64 39.0 .74
100 50.0 .64 56.4 .37
ISO 64.7 .64 64.0 .24
200 77.9 .64 68.1 .18
250 90.1 .64 70.7 .14
300 101.4 .64 72.4 .12
250 111.9 .64 73.4 .11

 

*Elasticity coefficient for the log-inverse function was
measured at the mean of the total annual expenditures (income).

92

Figure 5. A Logarithmic and a Log Inverse
Food Expenditure Functions

 

     

 

 

 

2.5 i 1 ,
2.0)..
A log inverse
function
'°5 “ double
logarithmic
function
I, I ,L
1.0 1.5 2.0 2.5

I09 Yr/Xlr

Source: Table 21

93

4.3.4 Analysis of the Commodity Groups

Commodity groups analyzed in this study include: (I)
cerealsgand starches, (2) dried legumes, (3) meat, fish, and
eggs, (4) oils and fats, (5) milk and dairy products, (6)
vegetables, (7) fruit, (8) sugar and sugaries, (9) beverages,
(IO) tea, coffee, and other stimulants, and (11) other food
stuff. Relationships were estimated for each of these groups
for urban and rural areas separately and for the whole sample,
with a dummy variable included for the locational difference.
Estimates were also obtained for different household sizes
within urban and rural areas.

The estimated relationships from the double logarithmic
function are shown in the Appendix]. The elasticity coefficients
were derived from these relationships and are shown in table 21.

Further consideration of the figures presented in these
tables suggested that the estimated elasticity coefficient for
the sample as a whole,when the two samples are combined in
one model,cannot be considered as a reliable estimate for

further analysis. The reason is that it is based on the total

number of observations in the household budget sample. The

 

ISee Tables B9 and BIO in the Appendix.

94

Table 21. Total Expenditures Elasticity of Demand for Commodity
Groups and Total Food, Estimated from the Logarithmic
and the Log-Inverse Models for Urban and Rural Areas

 

 

 

 

 

Commodity groups logarithmic log-inverse*
urban rural urban rural
Cereals 8 starches .1035 .4081 .0757 .5083
Dried legumes .1752 .5790 .1104 .7281
Meat, fish, 8 eggs 1.0199 .9745 .7428 1.2000
Oils and fats .4387 .2912 .2974 .3530
Milk 8 dairy products .9409 .8194 .6792 1.0574
Vegetables .6845 .7292 .4880 .6420
Fruit 1.2395 1.2900 .9075 1.6032
Sugar and sugaries '.6323 .7206 .4498 .8558
Beverages .7120 .7005 .5001 .8219
Tea, coffee 8 other .7318 1.2866 .5133 1.5105
Other food stuff .3545 .5455 .2461 .4914
Total food .6331 .6887 .4533 .8446

*Estimates for elasticity coefficients are estimated at the
mean of total annual expenditures for each region.

95

Table 22. Total Expenditures Elasticity of Demand for
Commodity Groups and Total Food - Weighted
Coefficients*

 

 

 

 

Commodity groups logarithmic log-inverse
function function**
Cereals 8 starches .2923 .3429
Dried legumes .4255 .9934
Meat, fish 8 eggs .9917 1.0263
Oils and fats .3472 .3318
Milk and dairy products .8656 .9137
Vegetables .5882 .5835
Fruit 1.2708 1.3388
Sugar and sugaries .6870 .7012
Beverages .7049 .6996
Tea, coffee, 8 others .9758 1.1316
Other food stuff .4729 .3982
Total food .6676 .6959

 

*Based upon the separate estimates of elasticity coefficients
in the two regions with the appropriate weight given to each.

**The original estimates are calculated at the mean of total
annual expenditures.

96

sample, though randomly selected within the urban and rural
areas, is not a pr0portionate sample as only 48.5 percent of
the individuals in the sample are rural while the correspond-
ing rural percentage of the total population at the same

time was about 62 percent. Another path through the
complexity of the aggregation problem, though not necessarily
an unerring one, is to calculate a set of ”weighted” coeffi-
cients based upon the separate urban and rural estimated
coefficients.

Examination of these "weighted'| elasticities makes it
possible to classify the studied commodity groups into three
categories, according to the magnitude of their elasticity
coefficients:

(1) Commodity groups with a relatively low income
elasticityl, which include cereals and starches, dried
legumes or pulses, and oils and fats.

(2) Commodity groups with intermediate income elasticity
which include sugar and sugaries, vegetables, and beverages.

(3) Commodity groups with a relatively higher income
elasticity; this includes meat, fish and eggs, milk and dairy
Products, fruit, tea, coffee and other stimulants.

—‘

IUnless otherwise cited, the income elasticity and the
total expenditures elasticity were used interchangeably.

 

97

_2
The coefficient of determination, R , was generally high.

It has a value more than .95 for all commodity groups except
for the logarithmic fitting of the dried legumes which was

.89. 82

was exceptionally high for meat, fish, and eggs,
oils and fats, and fruits. For total food, the logarithmic
fitting provided a value for R2 as high as .9994.

.The elasticity coefficients were the lowest for cereals
and starches and were lower in urban than in rural areas. The
low elasticity coefficient for the group in general was expected
due to the inclusion of many items in the group and the wide
range of substitutability involved. The lower coefficient for
the urban areas can be attributed to the relatively low budget
proportion spent on the group in urban areas as will be shown
when the individual commodities are analyzed. For similar
reasons, dried legumes elasticities vary significantly between
urban and rural areas. It rises from around .15 in urban
areas to about .60 in rural areas.

Tea, coffee, and other stimulants which include cigarettes,
tobacco, and liquor provide a significant variation between
rural and urban areas. While the elasticity coefficient for
this group was about .7 for urban areas, it reaches a high 1

level of about 1.3 in rural areas. With the lack of other

recreational facilities in rural areas, as income rises

98

consumers tend to spend more time and money in the café,

the only social institution available. The variation is

also observed between the estimates of the logarithmic

function and the log-inverse function. The basic differences
between the two functions as to the constant coefficient in

the log function, and the existence of a saturation level in
the log-inverse function, are responsible for this discrepency.

When urban and rural areas are combined in one model and
a dummy variable that takes a value of one for any observation
from urban areas and a value of zero for any observation from
rural areas, was included, the coefficient B3; associated
with that variable was generally negative. Except for oils
and fats, vegetables, fruits, and beverages, this coefficient
had a negative value. This result confirms the early
expectations. Households in urban areas tend to make less
demand on physical energy products as income rises. Further,
environmental and social factors would encourage households,
other things equal, to spend more on non-food items as income
rises.

Table 23 shows the values obtained for B33, the dummy
variable coefficient. It should be mentioned here that
despite the problem of the proportional representativeness of
the sample, the values of that coefficient are not affected

and are val id.

99

Table 23. Estimates for 83;, from the Logarithmic and
the Log—inverse Modeksfor Commodity Groups

 

 

Commodity Groups Logarithmic Log-inverse
Model Model

 

Cereals and starches - .0917 - .0055
Dried legumes - .1419 - .0079
Meat, fish, and eggs - .0302 - .0030
Oils and fats + .2003 + .0132
Milk and dairy products - .0667 - .0052
Vegetables + .0918 + .0064
Fruit + .0632 + .0041
Sugar and sugaries - .0306 - .0081
Beverages + .0183 + .0009
Tea, coffee, 8 others - .0588 - .0033
Other food stuff + .4720 + .0303

 

Total food - .0150 - .0015

 

100

4.3.5 Analysis of Individual Food Items

Individual food items analyzed in this study include
wheat, maize, millet, flour, bread, rice, macaroni, beans,.
lentils, meat, poultry, fresh fish, eggs, oils, saturated
1 oils, milk, white cheese, whey cheese, butter, ghee, potatoes,
onions, tomatoes, cirtrus, dates, sugar, tea, and coffee.

Elasticity coefficients was computed for both quantities
consumed and total expenditures. The results derived from
both the double logarithmic and the log-inverse models are
presented in Tables 24, 25, 27.

Another set of ”weighted” elasticity coefficients was
also calculated, with the proper weight given to the separate
urban and rural estimates. These coefficients are presented
in Table 26.

It would be noted that, in general, the expenditure
elasticity is higher than the quantity elasticity; in certain
cases they were much higher. This is a result of the fact that
households with higher income tend to consume superior quality
items at higher prices in the same general line of product
than do, other things equal, low income consumers. It is for
that reason that the disagreement between the two elasticities
was most apparent in commodities with a high level of quality

differential, such as rice, macaroni, meat, poultry, fresh

101

Table 24. Quantity Elasticity Coefficients from the Logarithmic

and the Log-Inverse Models in Urban and Rural Areas

 

 

Logarithmic Model Log-inverse Mode1*

Commodities

 

 

Urban RuraT’ ‘UrEan Rural

areas areas areas areas
Flour -.6661 -2.3752 .5776 -2.7720
Bread .1961 -1.2429 .1872 1.3837
Rice .3308 .7311 .4594 .9141
‘Macaroni .8457 2.3588 .9020 2.9104
Beans .1527 .5275 .1620 .6669
Lentils .1181 .2722 .1242 .3074
Meat .8160 .7464 .8321 .9011
Poultry 1.6439 1.8360 .7279 2.2331
Fresh fish .3934 .8128 .4268 1.0098
Eggs 1.1415 1.2966 .1872 1.5432
0115 .2018 .1633 .2041 .2162
Saturated oils .5361 .7643 .5486 .8904
Milk .8357 1.7240 .8563 2.1182
White cheese 1.1327 1.6163 .1610 1.9338
Whey cheese -.6336 - .7274 .6930 - .8487
Butter .8124 - .0644 .8531 .1478
Ghee .9573 .9479 .9907 1.1268
Potatoes .4852 .8988 .4841 1.1126
Onions .1506 .0298 .1370 .0366
Tomatoes .3743 .2183 .3908 .2768
Citrus 1.0175 1.4752 .0617 1.7901
Dates .3024 .4950 .3224 .6131
Sugar .2412 .4079 .2510 .4543
Tea .0248 .2864 .0399 .3030
Coffee 1.6002 2.1692 .6796 2.5144

 

*Calculated at the

mean of total annual expenditures.

 

Table 25.

102

Expenditure Elasticity Coefficients from the Log

and the Log-inverse Models in Urban and Rural Areas

 

 

Commodities

Logarithmic Model

Log-inverse Model

 

 

Urban RuraT’ ’Urban TRural
areas areas areas areas
Flour - .5201 -2.3883 - .4319 -2.7917
Bread .1961 -1.2429 .1892 -1.3837
Rice .4042 1.1297 .4242 1.4014
Macaroni .9101 2.6782 .9634 3.2832
Beans .0205 .6150 .1636 .7612
Lentils .0716 .4196 .0809 .5043
Meat .9358 .7574 .9583 .9216
Poultry .7301 1.9251 .8150 2.3288
Fresh fish .7636 1.1527 .8069 1.4247
E9 5 .2351 1.3298 .2821 1.5917
0115 .2888 .3532 .2844 .4247
Saturated oils .2876 .7245 .6051 .8425
Milk .9897 1.7463 .0394 2.1509
White cheese .1884 1.9283 .2165 2.3495
Whey cheese .5368 - .5381 .5997 - .6277
Butter .9218 .0591 .9839 .0012
Ghee .0141 .9882 .0476 1.1750
Potatoes .5768 1.0234 .5765 1.2537
Onions .2348 .1273 .2175 .1408
Tomatoes .5452 .3633 .5619 .4537
Citrus .2126 1.5420 .2688 1.8727
Dates .4492 .7327 .4673 .9018
Sugar .3963 .5012 .4065 .5649
Tea .0973 .3286 .0753 .3606
Coffee .5889 2.1692 .6589 2.6740

 

 

J. 1.."1WN"AHUEPAM.U§1 FIVWIUSTII Ali ~

[11

103

Table 26. Total Expenditures Elasticity of Demand for the
Individual Food Items, Weighted Coefficients*

 

 

 

 

Commodities Double Log Function Log-inverse Function**

 

 

Quantity Value Quantity Value
Flour -1.7257 -1.6784 -1.9381 -1.8950
Bread - .6961 - .6961 - .7867 - .7867
Rice .5790 .8540 .7033 1.0300
Macaroni 1.7838 2.0063 2.1472 2.4017
Beans .3850 ..3891 .4750 .5341
Lentils .2136 .2874 .2377 .3434
Meat .7728 .8252 .8749 .9355
Poultry 1.7630 1.8510 2.0411 2.1335
Fresh fish .6534 1.0048 .7883 1.1899
Eggs 1.2377 1.2938 1.4079 1.4740
Oils .1779 .3287 .2116 .3714
Saturated oils .6776 .5585 .7607 .7523
Milk 1.3864 1.4588 1.6386 1.7285
White cheese 1.4325 1.6471 1.6401 1.9189
Whey cheese - .6917 - .5367 - .7895 - .6171
Butter .2680 .3888 .2325 .3746
Ghee .9515 .9980 1.0751 1.1266
Potatoes .7416 .8537 .8738 .9964
Onions .0757 .1681 .0747 .1699
Tomatoes .2772 .4324 .3201 .4948
Citrus 1.3013 1.4182 1.5132 1.6432
Dates .4218 .6250 .5026 .7367
Sugar .3445 .4613 .3770 .5047
Tea .1870 .2407 .2030 .2522
Coffee 1.9530 1.9487 2.1972 2.2883

 

*Based upon separate estimates of elasticity coefficients in
rural and urban areas.

**The original estimates are calculated at the mean of total
annual expenditures.

104

.mmmcm .mczc ocm cone: Low wmumE_umo mumcmamm ozu cod: nmmmm44

.moczu_ocmaxm .mzccm _Mu0u ozu 80 some mcu um omum_:o_mo«

 

.mHOu
oommcocaa
omoaooLaumEo;

“um___z

_mu0u
oommzoeza
nmozooLQImEo;
”oN_mz

.muOh

ommmsocaa

boozpoLaImEo;
"ummzz

 

 

k_mm.- make. 6km6.- om_:.- mom_. oomm.-
mass..- :mmm..- mkom._- Nem_._- :m__._- N_m_._-
ooo~._ k6mo.~ mmm_.- smum.. 6:86.. m::~.-
Nmmo.- 4.66. m6-.- “mo..- 6-o. m_m~.-
o_:m.- mom_._- mamm.- 6__m.- :NNm.- m-6.-
mumm._ omno.~ Nmmm. ~N_N._ .o:m.. ckmm.
“Nam. mom~._ Name. mmko. Ammo._ mmmo.
m_mm.- mmm6.- N_~m.- mm_:.- 646m.- .m~k~.-
m_mm.~ :mN_.m mmm_._ m_mm._ mmmm.~ ~m~o..

44666;m_63 .mtam 666.: 666;m_6z .mtam 666.:

4.0602 omcm>c_um04 _mpoz o_E;u_me04

 

>u_o_umm_u >u_ucm:o

 

po___z ocm oN_mz .ummgz LOm mucm_o_mmmoo >u_o_umm_m moc:u_cchXm

_mu0H .mm o_nmh

.mmmLm —mL3L Ucm EGO—L: 1.0% mQUMEwumm GHMhNQMm 0r: COQD Ummmmyczn

.mmL:u_ocmdxo .mzccm .muop ozu mo cmoE mcu um vmum_:o_mu«

 

105

 

 

:oom.- .moo.- ~:mm.- moum.- memo. nmmm.- , .muou

JOOm._I mm:m._u 50mm..- m:o~._- mmo~._- m_ou._1 oommcuLaa

:m~_._ .mmm._ Noo~.- mNQm. _~mm._ m_NN.- 66656666-6so;
.66___z

mam..- m_oo. nom~.- .zmo.- :NNO. m_m~.- .6000

“Noo._ .mm~._- mmmm.- zk~m.- mwmm.- mzkm.- nmmmzutsa

nmm:._ nmoo.N :mmm. _m_~._ moow.. comm. omoapoLaumEo:
“mN_mz

mam“. m~o~u_ memo. m_mo. :_~m. omuo. .mHOu

m_~w. ~mom.- om_m.1 Nmm:.u m_mm.u mmm~.1 oommcocna

mmmfi.m mm~_.m Nmmo._ wmmm._ mom:.u m:m_._ omoaooLaneo;
”umocz

44666;m_62 .mtsx 6662: 666;m_63 _mtam cant:
*_mooz mmcm>c_umou .0602 o_E;u_me04

 

>u_o_umm_m 0L:u_UCMQXm

 

 

om:c_ucoo .NN m_nmh

106

fish, eggs, milk, cheese, ghee, citrus, dates, and coffee.

A curious observation is that all of these products, perhaps
with the exception of meat, are not subject to price control.
Meat prices are officially fixed. However, these fixed
prices are neither enforced by the government nor observed

in the market.

The studied commodities can be classified into four
general categories:

(1) Commodities that can be labeled ”inferior”, with
negative income elasticity. These include flour, brea, and
whey cheese.

(2) Commodities with low positive income elasticity,
below .5, which include beans, lentils, oils, butter, onions,
tomatoes, sugar, and tea.

(3) Commodities with moderate income elasticity between
.5 and .7, which include saturated oils and dates.

(4) Commodities with a relatively high income elasticity,
that include rice, macaroni, meat, poultry, fresh fish, eggs,
potatoes, milk, white cheese, ghee, citrus, and coffee.

This classificationI gives a good idea of the general

picture, yet it conceals many differences between urban and

 

1In this classification the coefficients of whole grain
wheat, maize, and millet were not included, as they were further
discussed in detail by the end of this section.

107

rural areas. Butter, for instance, has a high elasticity in
urban areas, around .9, where refrigeration facilities are
available, while in rural areas the elasticity coefficient
is at a very low level, other major differences can be noted
for rice and potatoes, both have a low elasticity coefficients
in urban areas and a very high coefficients in rural areas.
The analysis of wheat, maize, and millet, where data
were available on home-produced and purchased consumption,
produced some puzzling results. The inputed home-produced
consumption expenditures as a percentage of total expenditures
for these products are presented in Table 28 for the three
grain items. For rural areas, the trend of this percentage in
relation to income is generally upward. It is a reasonable
generalization to assume that in rural areas where the major
activity is agriculture, the higher the farmer's income, the
higher the “volume” of the home-produced products. Households
with high incomes in the rural areas are likely to be land-
owners while low-income households are likely to be part or
full time laborers whose products are in terms of wages rather
than physical products. The relatively high percentage of
”home-produced" items in urban areas was surprising and may be
attributed to the existence of many big land owners who prefer
to dwell in cities. However, as Stevens noted, I'experience in

the United States supports an increase in the per capita value

108

 

.u_o .mm .mmxmmm_ .co_mmm cm_uo>om mcu c_ >m>cnm
ummoam o_o;mm:oz .oouu_EEou .mo_um_umum .mcucou .m.<.: Soc» boum_:o_mu "mULJOm

 

.mmcsu_ocmaxo .muOu .maccm ommco>m oumo_oc_ >__m:uom mmcam_m 05h4

 

 

 

00— mm Nm o~.mom_ wN mm N: m_.m:m_

00. mm _m Om.mom Om .m we o~._mm

mm mm mm .o.uoN om :m m. mm.:mm

we mm mm mm.muz 4m .0 m: ::.Nm:

on OK No mo.mmm w. mm m. mm.::m

mm on :0 mm._m~ m: nu mm o~.:m~

mm om mm mm._~u w. om _~ m~.:N~

A: m: o: mo.~u_ m. mN NN No.m~_

mm m: mm oo.m~_ :_ m_ m -mm.m~_

mm mm mm om.mm m. n m .m.mm

@— NN m_ :_.mo 0— m m om.~m

m m o 0:.mm m o. m um.o:

N _ : um.o~ o o 0 N0.@.
ucoocoa ucoocoa ucoocoa .m.4 ucoocma ucoocoa ucmocoa .m.u
um___z oN_mz umocz *_o>mu 0500c. um___z mN_mz mums: *_m>ou 0500c.

prosmmaoz n_ommm:or
mmoc< .mL3m mmmc< can:

 

 

m_m>mu mEOoc.
o_o;om:oz Ou ovum—mm .muunboL¢ omozu L08 mmL:u_UCQQXm
.mu0h mo ucouLoa mm um___z pcm .mN_mz .ummzz ooozooLanEOI .mN m_bmp

109

of home-produced food even at high income levels. It was not

until after World War II that the farm value of home-produced

foods per capita decreased”.I
Examination of the elasticity coefficients shows that

purchased wheat, maize, and millet all have a negative elas-

ticity coefficients in urban and rural areas. Both the home-

produced and purchased elasticity is higher in rural areas

than in urban areas. In the two regions, the purchased

products elasticity is higher for wheat, and is at its lowest

level for millet where it becomes negative. R2 was high,

about .90 for quantities of home-produced wheat and maize in

both areas. The coefficient of determination was lower when

the imputed values of these products was introduced for no

obvious reason except the possibility of underestimating or

overestimating these products when estimating its value by the

households. This coefficient is lower for purchased wheat,

or maize, and much lower (about .7) for millet. The apparent

reason for this low coefficient for millet is its unique geo-

graphical consumption patterns. It is mainly consumed in

upper Egypt and is almost unknown in other parts of the country.

The coefficient of determination would have improved signifi-

cantly had this geographical information been available and

included in the mOdel.

 

IRobert Stevens, Elasticity of Food Consumption Associated
with Changes in Income in Developing Countries , 22. cit., p. 3

110

It was not surprising, however, to have a negative value
for B33, the dummy variable coefficient, for all fittings of
wheat, maize or millet. This indicates the obvious fact that,
other things equal, consumption of these products is higher

per consumption units in rural areas than in urban areas.

Further Considerations of the Results
of Wheat and Maize

The most puzzling results in these data are the high
positive income elasticity for home-produced wheat and maize,
and the negative income elasticity for purchases of these items.
In examining these results, it should first be pointed out that
for urban areas the percentage of expenditures for whole grain
wheat and maizeI represents only 13 percent of cereals and
starches expenditures, and less than one percent of total
expenditures. The corresponding percentages for rural areas
2

reaCh about 66 percent and 27 percent, respectively.

There are at least three possible explanations concerned

with data, price relationships, and quality.

 

IThese quantities of whole grain wheat and maize are
usually custom-milled at small mills.

28ee Table 16 (Ch. IV), and Table 29.

111

Table 29. Percentage Distribution of Expenditures for
Cereals and Starches in Urban and Rural Areas

 

 

 

 

 

 

 

Items 3:882 836:; 33692§§d
percent percent percent

Wheat:

home-produced 1.7 16.6 10.9

purchased 5,4 16.7 12.4

total 7.1 33.3 23.3
Maize: _

home-produced 1.6 15.9 10.5

purchased 4.3 16.4 11.8

total 5.9 32.3 22.3
Millet:

home-produced .5 5.2 3.4

purchased 1.8 7.5 .5.3

total 2.3 12.7 8.7
Rice 9.8 11.2 10.7
Flour 19.8 6.7 11.7
Bread 49.1 2.1 20.0
Macaroni 4.1 .7 1.3
Others 2.0 1.0 123
Total cereals and
starches 100 100 100

 

Source: Calculated from:

U.A.R. Central Statistical Committee,
The Household Budget Survey,gp. cit., pp. 210-241.

112

With regard to the first, a likely explanation is the
possibility of data discrepancy that resulted in nfisclassi-
fying a greater percentage of wheat or maize as home-produced
than was actually the case. This could be due to a common
Egyptian practice in which a landlord often gets his land rent
or a part of it in kind. He may classify this portion of his
wheat or maize consumption as "home produced”. Urban dwellers
who have rural connections and receive quantities of wheat or
maize directly from the farms for their domestic consumption
may also refer to these quantities as "home-produced”, though
they were provided to them as a compensation for their share
in a common land-ownership, or for other services they rendered.

Another factor that may influence data misclassification
relates to the legal requirement for farmers to deliver a
percentage of their produced wheat or maize to the government-
owned Agricultural and Cooperative Bank. This requirement may
encourage clandestine purchases to be reported by buyers as
home produced. It is, however, unclear to what extent this
would increase the reported home-produced wheat and maize.

The previous line of argument suggests that reported data
on home-produced wheat and maize may be unduely high, and
offers a partial explanation of the unusual elasticities for
home-produced and purchased. Two additional points may throw

some light on that complex relationship: (1) The price

113

relationship. Wheat, for instance, is sold to the government
at a fixed price which, in most cases, is lower than the
market price. A partial delivery of the product is compulsory
in repayment for advanced loans. Large farmers (or high-
income households) can afford to repay their loans and keep
large quantities of their product for their own consumption.
Small farmers, (or low-income households), may find that
selling their product to the government and purchasing the
government-subsidized flour mix provides more income. (2)
Quality differentials. Domestic varieties of wheat are
usually regarded as having better baking qualities, and are
particularly preferred for their high gloten contents.I
Imported wheat or wheat flour are usually regarded as being
of a lower or of uncertain quality.

The conclusions drawn from the previous discussion and
from reviewing elasticity estimates for cereals and starchy
products can be summarized as follows:

(1) In urban areas the percentage of home-produced grains
of total expenditures for cereals and starches is less than 3

percent, and represents a very negligible portion of total

 

IFawzy Refai, ”Importance of Testing Grain Varieties and
New Lines for Quality”, a Lecture presented at the U.A.R.
Agricultdral Training Center for Afro-Asian Countries, ATC 22,

1965.

114

income. As income rises, urbanites tend to consume less whole
grain wheat and maize, less flour; but more bread,‘ more rice,
and more macaroni. However, for that small proportion of
whole-grain products consumed in urban areas, the percentage
of ”home-produced” products increases as income rises. This
is due to the fact that the higher the income, the higher the
probability of having land-owners who dwell cities included

in the sample.

(2) In rural areas, the percentage of home-produced
whole-grain consumption to total expenditures for cereals
and starches is more than 37 percent. In these areas, as income
rises, households tend to consume more of home-produced grains
and purchase less flour and less bread.

This extended consideration of whole grains consumption
helps explain why flour has a negative income elasticity. This
is believed to be a result of aggregating all types of flour
in a single item, and due to the fact that the home-produced
whole grain wheat is considered a "superior” source of carbo-
hydrates in rural areas. Several types of flour are available

in the market. The most important is the government-subsized

IThe income elasticity for bread in urban areas is positive
and equals about .2. _

115

flour mix which is a mixture of wheat, maize, and other products.
Certainly if the data were broken down to its components,

several elasticities would have been obtained. This belief

is supported by the fact that income elasticity for cereals

and starches as a group is positive, and amounts to about .3.

It would be reasonable to assume a positive income elasticity
for wheat flour, and negative income elasticities for other

mixed types.l
4.3.5 The Effect of Household Size

Tables 30 and 31 show that in both urban and rural areas
for certain ”basic“ commodities, the larger the household size
the larger the elasticity coefficient. This applies to cereals
and starches, dried legumes, meat, fish, and eggs, milk and
dairy products, vegetables and fruit. This applies also to
oils and fats, and sugar and sugaries in urban areas. The '
opposite is true for beverages, tea, coffee, tobacco and other.
stimulants for the obvious reason that the adults' proportion
in small households is greater.

The household size is introduced in the original models
to allow for variation in expenditures per consumption unit

due to economies of size.“ The coefficient BZi indicates the

Table 30.

116

Households, Classified by Size, from the

Logarithmic and the Log-inverse Models in

Urban Areas

Total Expenditures Elasticity Coefficients for

 

 

Commodity Group

Logarithmic Model

Log-inverse Model

 

 

 

Small Medium Large Small Medium Large
Cereals 8
starches -.O478 .1729 .1849 -.0236 .1375 .1490
Dried legumes .1838 .2175 .3547 .1517 .1710 .2599
Meat, fish 8 eggs .9109 .9586 .1450 .8191 .7542 .9018
Oils 8 fats .3939 .4509 .4939 .3453 .3627 .3459
Milk and dairy
products .9599 .8847 .1066 .8541 .6989 .8244
Vegetables .7022 .7030 .8048 .6136 .5546 .6119
Fruit .2707 .2315 .3722 .1149 .9710 .0000'
Sugar 8 sugaries .6080 .6403 .7461 .5424 .5030 .6017
Beverages .8698 .7647 .8332 .7469 .6016 .6631
Tea, coffee 8
other stimu-
lants .2190 .8557 .6971 .0508 .6688 .5810
Other food
stuff .6090 .3783 .4087 .5128 .2917 .3487
Total food .6907 .6342 .6948 .6016 .4988 .5467

 

Table 31.

117

Households, Classified by Size, from the

Logarithmic and the Log-inverse Models in

Rural Areas

Total Expenditures Elasticity Coefficients for

 

 

Commodity Groups

Logarithmic Model

Log-inverse Model

 

 

 

Small Medium Large Small Medium Large

Cereals 8

starches .4249 .4536 .4467 .5055 .4796 .3910
Dried legumes .7231 .5744 .7416 .8673 .6316 .6703
Meat, fish 8 eggs 1.0021 .0571 .0150 .1810 .0806 .9068
Oils 8 fats .6051 .5350 .5382 .7321 .5739 .4617
Milk and dairy

products .9566 .0572 .9624 .1996 .0979 .8837
Vegetables .5944 .8701 .7586 .6984 .8649 .6878
Fruit 1.4115 .0361 .1547 .6685 .0749 .0108
Sugar 8 sugaries .9682 .8045 .7000 .1509 .7959 .6195
Beverages .9778 .8437 .7736 .0674 .8405 .6842
Tea, coffe 8

other stimu-

lants 1.7123 .6148 .9817 .0471 .6513 .8716
Other food 9

stuff .7294 .3830 .3528 .6905 .4482 .3255
Total food .7952 .7608 .7180 .9343 .7771 .6374

 

118

extent to which households consume or Spend more or less per
consumption unit because other things equal, they contain a
larger or smaller number of consumption units.

In urban areas, a positive value for 821 was obtained for
dried legumes; meat, fish and eggs; milk and dairy products;
vegetables; fruit; sugar and sugaries; beverages; stimulants
and total food in general. In rural areas, fewer positive
values of B2; were obtained, namely for dried legumes; meat,
fish and eggs; milk and dairy products; fruit; beverages;
stimulants; and total food. A positive value for that coeffi-
cient for a given product indicates that, for a given income
per consumption unit, large households spend more per consump-
tion unit than do small ones. Although this coefficient was
also obtained for households classified by size in rural and
urban areas, its importance appears to be negligible due to
two factors: first, the homogeneity of the classified observa-
tions, and second, because the commodity groups themselves are
broadly defined.

In order to test for the significance of the household
size effect, the analysis of variance procedure was applied to
the urban sample. The rural sample is not included in the
analysis because of the obvious correlation between the house-
hold size and its income. The larger the family size, the

greater the number of wage-earners, and consequently higher

119

income. The two-way classification of the sample was according
to the income level, as indicated by the total annual expendi-
tures, and the household type as defined earlier in Chapter 111.
Since the number of reporting households vary greatly from one
cell to another as shown in Table 32, the calculation of the
analysis of variance was modified to allow for this fact.I

The results of the analysis of variance are given in
Table 33. The analysis indicates a significant income effect
at both .05 and .01 levels of significance, while the household
type effect was not statistically significant.

It is possible to pursue this kind of analysis still
further and test the hypothesis of additivity of income level
and household type, or test the hypothesis of interaction
between the two factors. However, it is felt that this problem,

though interesting, is outside the major focus of the study.

4.4 Summary

This Chapter presented the analysis of the cross-sectional
data drawn from the first household budget survey.

Three models of analysis were utilized, of which the
logarithmic and the log-inverse functions provided better

fits than did the semi-logarithmic model.

 

IThe analysis of variance model applied here is explained
in detail in: Vernon G. Leppitt, Determinants of Consumer
Demand for Home Furnishings and Equipment,HarvardUniversity

Press, 1959, pp. 19-35.

120

Table 32. Average Annual Household Expenditures for Food in
Urban Areas and the Number of Reporting Households*

 

 

 

Average .
Total Annual Fam'1Y TYPe

ExpendItures Large Medium Small
40 25 L.E. 34 L. . 29
(I) (7) (57)

(5) (47) (113)

88 69 65 61
(21) (119) (111)

125 92 88 30
(131) (304) (167)

173 121 112 99
(135) (282) (95)

224 I46 134 122
(130) (I71) (55)

274 172 156 136
(108) (136) (25)

345 200 180 156
(209) (130) (38)

487 261 219 191
(175) (104) (28)

684 332 272 241
(69) (50) (12)

871 400 31] 277
(22) (18) (5)

1547 577 410 472
(36) (19) (4)

 

Source:

*Figures in parentheses indicate number of reporting households.

UAR Central Statistical Committee, Household Budget
Survey in the Egyptian Region, 92,

121

 

mm om:.mmm.-

.6668

 

 

moo.m:m NN m_~._o~._ co_uomtmuc_

6~.N _~.¢ ___.mm~.~ __ m-.om~.m~ Anmumamumv

0500c.

moo.~o_._ Aomumsmomcsv

mN_m U_ocom:ox

::.m m~.~ oom.o-._ N _o~.mm:.~ Avmumsaumv

mN_m o_ocom:o:

omo.mmm.m~ Avmumzmnmcav

0800c.

_N.N NA.m ~A_.0mo.N m. _m~.Nmm.6~ 6~_m 6_osmmso;

new 6800:.

Amo.v o_umc Eooomcm co_um_cm>

u u motmnum com: 80 mocmzdm 80 Sam mo ooLJOm
ovum—nopmu mmmcmoo

 

 

mmmc< can: :_ mocsu_ocoQXu boom Low mocm_cm> mo

m_m>_mc< .mm o_nmh

122

Elasticity coefficients based upon expenditures were
estimated for commodity groups. The, elasticity coefficients
based on both quantity and value were estimated for selected
individual items. All of these coefficients were estimated for
urban and rural areas, and then, a weighted average was intro-
duced to provide a general view for the entire population.

The examination of the estimated relationships indicates:

(I) That the food consumption patterns differ significantly
as between urban and rural areas.

(2) That there are differences between quantity and
expenditures elasticity, and that in general, expenditure
elasticities are higher than the quantity elasticities.

(3) That although the household size seems to be an
important factor in the demand for food, its significance
failed to materialize when analyzed after adjusting for income
effects. However, when the number of consumption units was
included in the regression analysis, the variable produced a
coefficient with a value that differs significantly from zero.

The estimated income elasticity coefficients, or more
precisely, total expenditures elasticity coefficients have been
taken as a fairly reliable indicator of income elasticity of
demand. Some investigators in similar studies chose to reduce

the total expenditures elasticity by 10 percent so that they

123

would approximate the income elasticity of demand more closely.
This practice is based on several assumptions regarding the
relationship between total income and total expenditures.
However, it is felt that this practice of arbitrarily scaling
down all elasticities by a constant percent would not be a

significant improvement where the marginal propensity to save

is relatively small.

CHAPTER V

ESTIMATING PRICE PARAMETERS

This chapter is devided into two distinct sections. The
first deals with estimating price elasticities for some food
items based on analysis of time-series data for prices and
consumption, adjusted for income effect, for certain items
for which reasonable data were available. The second, attempts
to utilize the mathematical relationships between income and
price parameters to estimate price elasticities for most

commodity groups and individual items dealt with in this study.

5.1 Combined Analysis of Cross-sectional
and Time-series Data

5.1.1 Definition of the Variables

The objective of the analysis in this section is to intro-
duce the price factor in the food consumption analysis. This
is done by combining the income elasticity coefficients estimated
in the previous chapter together with the ”price” as a new

element in the analysis.I

ISee Appendix A.
124

125

Recent price data in Egypt are suspect due to the existence
of governmental price controls over a wide range of products,
and the use of these prices in published sources as the actual,
going prices. However, over the periods used in the analysis
we find no reason to seriously question the price data, at
least through 1963.

The variables introduced in the analysis include:

(1) Per capita consumption in terms of quantity.

(2) Per capita income.

(3) Income elasticity of demand.

(4) Prices or price indices for different products.

(5) Certain deflators: the cost-of-living index, and

the wholesale price index.
Though data about per capita consumption are available over a
wide range of commodities and groups of commodities, the price
data represented a limiting factor for extending the analysis
to include all the commodities that were included in the cross-
sectional analysis. Data are either unavailable, or subject
to various limitations and not quite comparable. After a
thorough investigation of that matter, it was necessary to
limit the analysis to few products. However, the elasticity
coefficients obtained are utilized in the following section to
determine the money flexibility and thus to compute the price

elasticities for all the other items.

126

Six products were included in the analysis. These are:
wheat, millet, onions, oils, tomatoes, and total food. The
data of annual per capita consumption are the official
estimates published by the Egyptian Ministry of Agriculture.
They are based on total production, net trade, storage, losses,
and use for other purposes. The analysis covers different
periods of time for each commodity. From 1945-46 to 1962-63
for wheat, from 1946-47 to 1958-59 for maize and millet,
from 1946-47 to 1962-63 for onions, from 1949-50 to 1962-63
for tomatoes, and from 1948-49 to 1962-63 for oils and total
food.

Prices used are wholesale prices except for oils and total
food where the wholesale price indices were used; there was no
specific reason for using the latter except its availability.
These prices are deflated by the wholesale price index for all
commodities.

Per capita incomes were calculated through the use of the
figures on private disposable income and population data.
These per capita income figures were deflated by the cost of
living index.

The unfortunate limitation of price data was an obstacle
in extending the analysis to deal with rural and urban areas,

separately. The nature of the price and per capita income

127

data as overall averages for the whole country necessitated
use of an overall income elasticity coefficients. The co-
efficients used are weighted averages of the urban and rural
coefficients.

Because the combined model is a logarithmic model, it
was logical and justifiable to use the coefficients derived
from the logarithmic model. However, there were not much
divergence between the logarithmic model coefficients and
those of the log-inverse model. The goodness of fit for both
of them was similar and at a satisfactory level.

The use of a single equation model, in other words the
inclusion of only one endogeneous variable in the model,
is justifiable. Domestic consumption is considered the major
single outlet for the items dealt with in the analysis. The
share of exports is very small and storage outlet cannot be
considered of any major importance. In addition, under price
control policies, the world market price for food can hardly
be related to local prices or the expenditure decisions in
Egypt. Had this not been the case, price would have been
regarded as a function of local production and world prices,
then consumption would have been made a function of local
prices and income. It is believed, therefore, that a single
equation model with domestic consumption as the only endogeneous

variable can depict the nature of demand for these products.

Table 34.

128

National Disposable Income, Per Capita Income,
Index, and the Wholesale Price

Cost of Living

Index,

1945/4

to 1962/63

 

 

 

NatTOnal Cost-of- Wholesale Price

Disposable Living Index All
Year Income Per Capita Index Commodities

(current Income 1953-54 (1939 = 100)

price) = 100

L.E.million L.E.
45/46 470 25.5 99 321
46/47 450 23.9 97 304
47/48 499 26.2 94 330
48/49 613 21.4 95 311
49/50 700 35.2 94 344
50/51 770 37.9 99 383
51/52 816 39.1 108 372
52/53 754 35.1 107 355
53/54 801 36.4 100 345
54/55 846 37.5 96 351
55/56 860 37.3 96 389
56/57 924 39.1 98 422
57/58 943 39.0 102 417
58/59 1003 40.5 102 417
59/60 1107 43.7 103 418
60/61 1194 45.9 103 425
61/62 1280 47.9 103 421
62/63 1359 49.6 100 419
Source: Donald Mead, Growth and Structural Change in the

 

Egyptian EconOm ,
Homewood, 111., 1967), Table VI-F-l, pp. 400-401.

(RicharafD.

Irwin,

Inc.

129

.m0E:.o> .mc0>0m
.0oo.um.umum 0L.m:cc< .momc0o ocm mo.um.u0um mo uc0EuLma0o «<2 "co.umELOmc. 0o.cm

 

 

.m0E:.o> .0L0>0m .mo_um.umum ocm m0.EOcoum .mLou.:o_Lm<
mo c_u0..:m >.;ucoz 05h .0c:u.:o.cm< mo >Lum.c.z m<a "co_umELo.c. >u.uc0:o "00L30m

 

.o.9\mtmumm_a .m00_ta ..mumt .mmoumEOH to. ..oo. u :m\mmm.v mmu_oc_ 00.66 0—mm0.o;3 ooo.
.muOu ocm m._o Lo. ..m.x\me0umm.a m0o.co 0.0m0.o;3 .0N.mE ocm mco.co .u0...E .um053 co. "m
.mEmLmo..x :. .co_uoE:mcoo .moccm mu_omo L0o no

 

 

 

om.~ .J..m om. :.m:: No. m.o :.~ :.om w... m.o. mm.~ m.... mm\Nmm.
om.~ ...mo Nu. ~.:o: :o. m.: m.. m.n~ m... m.m ~m.~ m.m.. ~o\.mm.
om.~ 5:.mm NN. m.mo: No. ~.m m.. m.- o~.. m.m ~m.~ ~.om .o\oom.
om.~ .m.~m m.. :.:o: .0. m.: 3.. o.m~ mo.. o.m ~m.~ o..o. oo\mmm.
mm.~ .m.mw 4.. o.mwm no. w.: m.. o.:~ mm.. 5.x m.~ :.m. .m.~ m.mo. mm\mmm.
mm.~ .m.wm m.. m..mm mo. o.m m.. N..~ oo.. n.m m.N m.m. .m.~ o.mm mm\mmm.
om.~ mu.~m m.. ..Nmm mo. m.: m.. m.o~ m... m.: :.~ ..NN .m.~ o.nm mm\mmm.
~..~ mN.mw 0.. 0.0m: oo. m.m N.~ m.o~ ow. “.0 m.~ :.m. .m.~ m.mn mm\mmm.
...N .o.mm .o. o.o.: .m m.m n.. m.m. om. m.m m.. N.o~ mn.~ m.mw mm\:mm.
u... mw.:m mm m.mo: .m m.m m.. m.m. om.. m.: 5.. o.- mo.m n..m :m\mmm.
.N.. ow._o oo. N.o~m oo. m.m m._ m.- .m.. m.m n._ :.o. mm.~ m.mm mm\~mm.
om.. o~.~m mm m.oum no. o.m m.~ m.m. N... m.u 0.. :.NN m..~ ~.~m ~m\.mm.
.m.. Nm.om .o. m..~m m.. m.m o.~ m.~. ow. m.o n.— N.- m..~ m.~m .m\omm.
on.. mm.mm mm ~.mmm mm m.m m.. N.@. mm. ~.m m.. m.m~ m..~ o.mm om\m:m.
.m.. mm.:n om m.mmm um m.m m~._ 8.: m.. o.mN m..~ m.mn m:\m:m.
.m.. ~o.mo on. o.m m.. o.m~ m..~ m.:m m:\N:m.
.m.. oo.uo om.— m.m m._ m.m. .mdm o.mm N:\o:m.
mm.~ m.mo o:\m:m.
6 o n. o a o n. o 6 o n. o n. o

0~.mz 6000 .0008 m..o m00umEOF mco.co 00...: 000:3 mLm0>

 

 

m0o.Lm ocm co.uoE:mcou mu_omo L00 .mm 0.30»

130

5.1.2 Results of the Analysis

It was apparent from the previous discussion that the
price data, and to some extent consumption data, represent a
restriction on the extension of the analysis to the desired
dimensions. It is true enough that the results of any study can
be no better than the data used in the analysis. Under such
circumstances, we attempted to extend the price analysis as far
as possible, bearing in mind all the difficulties involved.

The combined model was applied to nine products. The
results were less satisfactory in terms of statistical signi-
ficance, than hoped for.

Several explanations can be cited for the failure of the
time-series analysis to provide significant results:

(1) The reliability of the data. In most cases, data for
prices represent the official fixed prices rather than the
actual prices. In many instances, these were underestimated
indicators of real prices.

(2) The problem of aggregation, as the data represented
averages for different types and qualities with a wide range
of substitutability, price data were assumed to be a weighted
average that reflected the change in price level over the
studied period. Yet, the quantities consumed are simple

averages that do not show quality differences. The aggregation

131

problem is of a particular importance when logs are used, as
the sum of the logs is not equal to the log of the sum.

(3) The locational differences that have been overlooked
in the time-series analysis. While the cross-sectional analysis
Showed basic differences that exist in the consumption patterns
of the two sectors, in terms of income levels, income elas-
ticities, and budget proportions spent on different items. A
weighted elasticity Coefficient was used to adjust location
effect, while price data were not weighted.

Table 36 shows the estimated relationships between per
capita consumption adjusted for income and deflated prices for wheat,
maize, millet, onions, tomatoes, oil, and total food.

The results show a price elasticity coefficients, which
are equal to the estimate 331’ of -.6 for wheat, -.01 for
maize, -.4 for millet, -.25 for onions, -.6 for tomatoes,

-.8 for oils, and -.1 for total food. The values of R, R2 and

‘R2 for all the commodities included in the analysis were extremely
low with the exception of tomatoes which has a value of R of

about .7.

When testing the hypOthesis 811 = 0 against the alternative
hypothesis 311 f O, the nul hypothesis was rejected for all
the products, which indicate that the estimated values for B];

differ significantly from zero.

132

Yet the analysis of the overall regression produced signi-
ficant values of f ratio at a level of significance .05 for
only wheat, millet, tomatoes and oils. The overall regression
for the other products was not statistically significant at
that significance level .05.

It is an obvious conclusion that the results of this
analysis was not statistically satisfactory, and should be
looked at with caution. However, when discussiong the commodity
projection, an assumption was made with regard to the stability
of price relationship, and hence the results of the present
analysis were not in fact utilized in the projections.

When testing for serial correlation, the results showed
values of d' (Durbin Watson statistic) higher than dU (Durbin-
Watson upper limit), which implies acceptance of the hypothesis
of random disturbance.

The inclusion of the prices of substitutes, in the case
of wheat, maize, and millet did not significantly improve the

results.

133

Table 36. Estimated Relationships from the Model

Zit = BC + 811 log Pjt/Pt

 

_____===L_;==================================================
Estimated Coefficients, Z; =

 

 

Commodities Eeriodd
BO 8] R S overe

Wheat .8100 -.6353 .4734 .0645 1945/46
.0577 .2956 I962/63

Maize 1.9161 -.Ol37 .0106 .0471 1945/46
.0924 .3327 I962/63

.Millet .8533 -.4230 .3919 .0615 1946/47
,6859 .2994 1958/59

Onions .0675 -.2527 .2787 .1254 1946/47
' (.5635) (.2248) I962/63
Tomatoes -1.4529 -.5658 .6830 .0613 1949/50
(.4054) (.1747) I962/63

Oils .0556 -.8340 .3946 .0633 1948/49
(.0211) (.5386) 1962/63

Total food 1.5389 -.1016 .0272 .0495 1948/49
(.0134) (1.0366) I962/63

 

134

5.2 Estimating Price Parameters from
' Cross-sectional Information

5.2.1 Assumptions and Techniques

It is usually easier to obtain cross-sectional data on budget
proportions and estimates for income elasticity than to obtain
estimates for price elasticities. In the present study,
reliable price data over a reasonable length of time were
either unavailable or unreliable for several items. This short-
coming compelled the use of mathematical relationships among
elasticities to avoid the data obstacle.

The model used here,1 is based upon certain basic mathe-
matical relationships, these are:

(l) The Homogenity Condition, which implies that the
demand function is homogeneous of degree zero in money income
and absolute prices. It follows that the sum of price elasticity,
income elasticity, and cross elasticity is equal to zero, or in

mathematical notations:

e-- = - Ei for all i

 

IUnless otherwise cited, this portion depends heavily upon
the work of Lester Manderscheid, "Price and Income Elasticities,
Relative Growth rates: Some Implications for Understanding
Agriculture's Terms of Trade”, a Paper Presented at the
Agricultural Development Council Seminar, Stanford University,
February 1966.

135

where:
eij = price elasticity for good i with respect to the
price of good j, all other prices constant. Zeu-
indicates that sum of i-th own price elasticity and
its cross price elasticity with respect to changes
of other goods taken one at a time.
E: = income elasticity for good i.

(2) The_Symetry Condition, which indicates that the cross
elasticities of two commodities, each with respect to the change
in the price of the other, are approximately equal with the
budget proportions spent on them are equal and the income
elasticities equal, or:

Ei+eij/Wj=Ej+eJ°i/Wi

where w; = budget proportion spent on good i. From this
relation we can derive:

eij=wj/wiej1-W_I(E1-E,l)forallland-1'

(3) Igtgl Expenditure Condition, which specifies that
the summation of the product of income elasticity and the
budget pr0portion, for all commodities is equal to one, or:

Z:vq E; = 1
this relation includes saving as a commodity or excludes it

from income.

136

(4) Cournot Aggregation, which implies that the summation
of price elasticities of all goods with respect to changes in
the price of one commodity j, is equal to j's proportion of
expenditures with sign reversed, or:

{w; eij = - w; for all j

In addition, to make the model workable we will accept,
as a reasonable assumption, Wold and Jureen's argument that
the fundamental relations involving income elasticity will not
strictly hold unless sector incomes vary proportionately. In
addition, the relation is based on the "typical consumer” with
a stable preference field and non-decreasing satisfaction
associated with increased quantities of each good consumed and
continuous derivatives everywhere. This assumes that ”the
behavior of the market can be described by the behavior of a
'representative individual' in much the same way as Alfred
Marshall spoke of a representative firm."l

If Frish's notion of “want structure independence” which
means that the quantity of good i influences only the marginal
utility of good i and no other marginal utility is accepted,
all direct and cross elasticities can be calculated from

knowledge of income elasticity and the money flexibility.

 

IRagnar Frish, “A Complete Scheme for Computing All
Direct and Cross Demand Elasticities in a Model with Many
Sectors“, Econometrica, Vol. 27, April 1959, pp. 177-196.

137

Though the want independence is a strong assumption, it might
be more acceptable if products are aggregated in some broad
fashion.

Money flexibility can be defined as:

W' = 53m . a
363 m
where:
m = marginal utility of money

a = money income.
The knowledge of money flexibility is a basic assumption
in Frisohfis formulas for computing direct and cross price

elasticities:

 

 

Kk

(l) E;k=-E; Wk(1 +W' )
(2) e.. = -E. w. - (1 'WI 51)

11 I I W'

where:

e;k = cross price elasticity for good i with respect to k .
e;; = i's own price elasticity . i
E; = i's income elasticity.

W1
WI

budget proportion spent on i .

money flexibility ,

Because W' is not known, it can be estimated from knowledge

0F the income elasticities and one or more price elasticities

138

by inserting these estimates in the previous two equations, or,

WI = !k Ek E-k
eik+ Ei wk

= E; (l - W; E. )

W' I

e.. + W. E.
11 1 I
W' can be estimated from several sets of elasticities,

then an average of these estimates can be used.

6.2 Results of the Analysis

The analysis started with knowledge of few price elasti-
city coefficients, namely, for wheat, maize, millet, oils,
onions, tomatoes, and total food.

The first step in the analysis was to estimate money
flexibility, or the relative change in the marginal utility
of money associated with change in money income. As indicated
before, money flexibility can be estimated indirectly through
the knowledge of price elasticity, income elasticity, and the
budget proportion of the commodity in question.

It was unfortunate that the few price elasticity coeffi-
cients obtained from the combined analysis of section 1, were
1:Or the population as a whole, while it was apparent that major
differences do exist between the consumption patterns and
behavior, and consequently the spending decisions in urban

and rural areas.

139

In estimating money flexibility, several Options were
open:

(1) To calculate money flexibility through the empirical
price elasticity coefficients and using a weighted income
elasticity coefficient for the entire population.

(2) To calculate money flexibility in rural and urban
areas separately using the estimates of income elasticity for
each, and assuming that there was no difference between the
two areas with regard to price elasticity.

(3) To use a set of values for money flexibility suggested
by Frisch as a useful first approximation until further data
and estimates become available.

Frisch's suggested set is:I

-10 for an extremely poor and apathetic part of the
population;

-4 for the slightly better off but still poor part
of the population who want to become better off;

-2 for the median part of the population;

-.7 for the better off part of the population;

-.1 for the rich with ambitions toward conspicuous

consumption.

ISee, Lester Manderscheid, 92. cit., p. 7.

140

The calculations of money flexibility for the entire popu-
lation based upon weighted income elasticity coefficients and
weighted budget proportions produced an average of -.85 with a
wide discrepency between an estimate of -.49 based on price
elasticity calculated from oil, and -l.32 derived from the
tomatoes elasticity. When urban and rural areas were dealt
with simultaneously, two different sets of estimates resulted
with a wide margin in between. However, for rural areas money
flexibility was higher, in absolute terms, and around -2.5,
while it was a little lower in urban areas.

In view of the inaccurate nature of price coefficients and
of the assumption of homogeneous price elasticity in both areas,
it was decided to use an approximate value of -2 for urban areas,
and -4 for rural areas, the two values suggested by Frisch and
not far from the calculated averages of money flexibility.

With these values of money flexibility and previous
estimates of income elasticity coefficients (based upon total
expenditures) available, price elasticity coefficients were
calculated for all commodity groups and individual items.

Table 37 shows these derived estimates for commodity groups
in urban and rural areas. Similar calculations for individual
commodities are shown in Table 38.

It should be kept in mind, however, that although these

calculated coefficients are useful as approximations, they

141

Table 37. Derived Price Elasticity for the Commodity Groups
and Their Budget Pr0portions

 

 

 

 

'Urban Areas RUral Areas
Commodity money flexib. = :2 money flexib. = -4
groups budget derived budget derived—
proportion elasticity proportion elasticity
percent percent
Cereals and
starches 13.70 - 1.40 27.78 -10.28
Dried legumes 1.31 - .30 2.80 - 1.53
Meat, fish 8 eggs 10.65 - 4.81 11.38 - 8.64
Oils and fats 1.89 - .85 1.60 - .50
Milk and dairy 6.09 - 3.70 8.36 - 5.65
Vegetables 3.71 - 2.01 3.43 - 2.22
Fruit 3.18 - 4.09 2.04 - 2.10
Sugar and sugaries 3.24 - 1.72 4.29 - 2.72
Beverages 3.08 - 1.42 3.49 - 2.19
Tea, coffee and
others 5.60 - 2.96 6.74 - 6.21
Other food stuff 2.89 - 1.02 1.26 - .73

 

Source: For budget proportions data, UAR Central Statistical
Committee, The Household Budget Survey, op. cit.,
pp. 344-6.

142

Table 38. Derived Price Elasticity for Selected Commodities

and Their Budget Proportions

 

 

 

 

 

W

Commodities money flexib. = -2 money flexib. = -4

—budget HerTVeET' budget derived

propoation elasticity pr0p0%tion elasticity
Rice 1.34 - .63 3.11 - 2.80
Macaroni .55 - .73 .19 - .83
Beans .45 - .05 1.02 - .68
Lentils .50 - .17 .92 - .45
Meat 6.57 - 3.74 7.75 - 4.49
Poultry 1.53 - 1.23 1.78 - 2.25
Fresh fish 1.58 - 1.13 1.08 - 1.17
Eggs .68 - .94 .52 - .80
Milk 1.93 - .97 3.85 - 3.87
White cheese .79 - .97 .09 - .58
Butter .67 - .80 1.14 - .08
Ghee 2.30 - 1.66 2.04 - 1.77
Potatoes .56 - .52 .63 - .74
Onions .34 - .19 .60 - .ll
Tomatoes .83 - .60 .82 - .36
Citrus .75 - .96 .43 - .79
Dates .34 - .34 .62 - .60
Sugar 2.11 - .87 3.36 - 1.60
Tea 1.68 - .20 2.90 - .96
Coffee .47 - .94 .32 - .87
Source: For budget proportions data, UAR Central Statistical

Committee, The Household Budget Survey, op. cit., pp. 355-6.

143

are based upon certain unverified assumptions. These assumptions
and conditions may fit the commodity groups rather than the
individual items. The “want independence” notion may be better
applicable to them. For individual items, the want independence
notion may seem to be a rather heroic assumption as it is hard

to accept the idea flat quantity consumed of meat, for instance,
does not have any influence on the marginal utility of poultry

or fish. Despite these shortcomings, these calculated coeffi-
cients are accepted as an approximation until more information

is made available.

It is interesting in this context to compare the estimated
price elasticity coefficients for onions and tomatoes with the
derived coefficients. As can be shown from Table 39, there is
not much divergence between the two sets, which might support
the usefulness of the derived coefficients.

Table 39. Price Elasticity Coefficients from Derived

Mathematical Model and from Empirical Data
for Onions and Tomatoes

 

4_

e;; derived from mathematical model Calculated

 

Commodity

 

eii
Urban Rural Weighted

Onions -.1878 -.1056 -.1368 -.2527
Tomatoes -.5997 -.3633 -.4531 -.5653

 

CHAPTER VI

PROJECTION OF THE DEMAND FOR FOOD
6.1 Introduction

The elasticities and coefficients presented in the
previous chapters represent an intermediate step toward
increasing our ability to project the future demand for food
products.

Two projections were made. A medium range projection
for 1975, and a long range projection for 1985. 1975 was
selected for several reasons, by that year the majority of
the effects of the Aswan High Dam and other development
projects will have been realized. The 1985 projection was
designed as a long-term forecast.

The first step in demand prOjection was to project popu-
lation, the second, to project income, the third, to proceed
to commodity projections.I To estimate the increase in
potential per capita demand, it is necessary to select the in-
come elasticity coefficients referring to the base year used

‘

. ISee: L. M. Goreux, ”Economic Growth and Commodity Projec-‘
tIons”, Monthly Bulletin of Agricultural Economics and Statistics,

Vol. 10,3uly/Augustil96TT’p.1-T7.

144

 

145

projection. From the value of the per capita demand estimated
for the target year, the level of the total demand can easily
be derived by multiplying the per capita demand by the size

of the population projected for the target year.

In projecting consumption, several underlying assumptions

should be made clear:

(1) that income is the independent variable while
consumption is a dependent variable determined by
income level, household size, and location. The
elasticity coefficients estimated earlier in this
study can be used to forecast demand assuming that
such coefficients would remain constant over time,
and that consumers would behave in the same way, if
any of the independent variables change.

(2) that no substantial change in relative prices within
the country or in the world market would take place
during the projection period.

(3) policitical stability.

(4) no major inflation that involve drastic changes
in the purchasing power of money. It would be unrealistic
to assume no inflationary pressure which, together with
the balance-of—payment difficulties are most typical
among "distrubances" that affect underdeveloped countries.I

ISee: Albert 0. Hirschman, The Strategy of Economic Develop-
EEHLE, Yale University Press, 1964, p. 1956-166.

 

146

6.2 Population Projections

 

Population is the single largest factor which decides the
trend of most economic activities. During the process of
economic development, changes in population are a cause, an
effect, and an indicator of economic change.l F

As indicated in Chapter II, the Committee of Experts which

was formed by the Egyptian Government in 1962 to study the pop- 1

 

ulation trends of the country, made five different projections
each based on different assumption with regard to the fertility
rate:2
I. Fertility rates will remain constant from 1960 to
1985 at the actual 1960 level of 190.0 per thousand.
II. Fertility rate will decline by one percent of the 1960
rate each year up to 1985, with a total decline of 25
percent over the period.

111. Fertility rates will remain at the 1960 level for 15
years and decline thereafter by 5 percent each year
for 10 years.

IV. Fertility rates will decline by 2 percent of the 1960

rates each year during the projection period, with a

total decline of 50 percent.

 

IJ. J. Spengler, "Population: Cause, Effect, Indicator”,
.ESQnonfic;Deve109ment and Cultural Change, April 1961.

2Fertility rates are calcultaed in terms of births per
thOusand females at the age group 15-49 in a given period.

147

V. Fertility rates will decline by 5 percent of the 1960

rate for 10 years and remain constant thereafter.

Mortality rates up to age 45 are assumed to decline exponen-
tially over the projection period toward certain ultimate values
obtained from a study of the minimum relevant mortality rates
in various countries. For ages 45 and over, mortality rates
were assumed to remain constant at their 1960 level.

Projections based on the first assumption were considered
as an upprt limit, while those based on the IaSt assumption
were considered as a lower limit.

The projected population was calculated by following the
base population at each age group and the expected births
according to the relevant fertility rate. The application of
the appropriate survivorship probability to each age group,
including the new borns, at a certain date would provide an
estimate of the population at that date.

Table 40. Projected Po ulation in Egypt, 1970-1985
(in millions

 

Fertility rate assumptions*

 

 

Years

I II IFF 1V V’
1970 34.5 34.0 34.5 33.4 31.7
1975 39.7 38.5 39.7 37.1 33.8
1985 52.2 48.3 48.4 43.6 38.8

 

*Assumptions 1 to V as stated in the present section.

Source: UAR Central Statistical Committee, Population Trends
in the U.A.R., 92, cit., p. 53.

7*

 

148

These projections imply a compound annual rate of popula-
tion growth between 2.8 (for the upper limit) and 1.6 (for the
lower limit), over the projection period. Without being
pessimistic, it is unlikely that the population growth rate
would drop to a rate as low as 1.6 within twenty-five years.
The most probable rate would lie between 2.5 and 2.8. Though
the population growth rate reached high levels in the past,
higher than 2.8, it is believed, based upon examination of
consecutive pOpulation data that it was a temporary phenomenon
attributed to other factors.I

For the purpose of this study, the projections under the
first and the second assumptions were utilized. It is believed
that these assumptions are consistent with the conditions in
most less developed nations where the fertility rates remain
stable while expectation of life at birth increases. For
practical purposes ”the influence of likely changes in the age
structure on average nutritional requirements can be disregarded
in most low income countries, particularly in the case of popu-

lations with stable fertility rates“.2

‘See: Bent Hansen and Girgis Marzouk, Development and
.Economic Policy in the U.A.R., _2, cit-. PP- 27‘45-

2L. M. Goreux, op. cit., p. 6.

 

149

The impact of population growth on the demand for food is
very great and may outrank the effects of income growth.
However, it is believed thatin a rapidly developing economy
with a relatively high income elasticity of demand for food
"any decline in the rate of population gorwh may be accompanied
in a few decades by an accelerated increase in the demand for
food, or at any rate will not be accompanied by a corresponding
lower demand for food”.1

The population projections were made for the population
as a whole, and were based on the demographic factors without
including other geographic or economic factors. Projection
by location was not made by the Committee. But it has been
shown from the previous chapters that major distributional
and characteristic differences do exist between urban and rural
areas to the extent that they should be dealt with separately.
Table 42 shows the declining percentage of the rural population
to the total due to migration. During the 1950's rural popu-
lation grew at about 1.9 percent, while growth in urban areas
was much greater, about 4.3 percent. This trend is expected
to continue and has, in fact, been pushed in that direction by
the implementation of the two five-year plans. More than one-

third of the planned investment was allocated to projects

 

1R. P. Sinha, Food in India, Oxford University Press, 1961,
pp. 63-64.

 

150

 

 

 

 

 

 

 

Table 41. Projected Population in Egypt by Age Groups in
1975, Based on the Constant Fertility Rate
Assumption (in thousands)
Age 1960 census 1975 projection
Groups
male female total male female total

0 - 4 '2209 1909 3118 3627 3069 6696

5-- 9 1979 1834 3813 2953 2468 5421
10 - 14 1658 1534 3192 2484 2094 4579
15 - 19 1118 1044 2162 1997 1715 3712
20 - 24 925 878 1805 1925 1800 3725
25 - 29 864 1058 1921 1609 1505 3114
30 - 34 810 847 1657 1079 1020 2099
35 - 39 851 883 1734 884 852 1736
40 - 44 663 617 2180 819 1018 1827
45 - 49 569 575 1144 747 800 1547
50 - 54 496 506 1002 759 824 1563
55 - 59 324 317 641 563 563 1126
60 - 64 322 555 677 884 569 957
65 and over 420 486 906 703 907 1610
Total 13207 12847 26054 20592 19149 39741
Source: UAR Central Statistical Committee, Population Trends

 

in the U.A.R., 93. cit.

151

associated with the Aswan High Dam. Cairo and Alexandria, along
with their suburbs were slated to receive more than one-half

of the remaining planned investment. Another sizable portion

of the investments was allocated to the extractive industries
along the Red Sea, far from the settled rural population of

the country. With the exception of a few textile and food
processing industries, the rural areas and even small cities
were to receive little attention as potential sites for
industry.I This allocation was affected by the lack of trans-
portation facilities and other infrastructure in rural areas.

It would be, therefore, a reasonable conclusion to assume that
the percentage of rural population will drop to about 57
percent in 1975, and 54 percent in 1985. Accordingly, of the
upper limit projected population of 39.7 millions in 1975,

about 22.5 millions are rural, and of the 52.5 millions projected

for 1985, about 28.3 millions are rural.

6.3 Income Projections

The information on Egypt's national and per capita income

are very scanty. However, the available data were reexamined

 

1Janet L. Abu-Lughod, ”Urbanization in Egypt: Present
State and Future Prospects", Economic Development and Cultural

Change, Vol. 13, No. 3, April T965.

 

152

 

 

 

 

Table 42. Population of Egypt According to Location, and
Percentage of Rural Population in Census Years
(in thousands)
Years Population Percentage
Urban rural total Of rural
ypopulation
1917 2,640 10,030 12,670 79.1
1927 3,770 10,407 14,178 73.4
1937 4,436 11,485 15,921 72.1
1947 6,262 12,704 18,967 67.0
1957* 8,223 14,773 22,996 64.2
1960 9,864 16,120 25,984 62.0
1975** 45.687 57.0
1985** 52.533 54.0

*Estimates.

**Projections for total population, based on assumption 1.

Source:

UAR Central Statistical Committee, Population Trends
in the U.A.R., .2: cit.

 

153

and evaluated by Hansen.] He estimated that from 1952/53 to
1960/61, the national income of Egypt rose from LE 972
million to LE 1341 million in 1960/61, or at approximately 4.3
percent compound annual rate. Later, he and Marzouk estimated
that from 1952/53 to I962/63 the real gross national income
rose at about 5.2 percent compound annual rate, and at about
6 percent during the last six years of the same period.2

The principal objective of the general Ten-Year Plan was
to double the national income over a period of ten years.3
This means an average annual rate of growth of 7.2 percent.
For the first Five-Year Plan a 40 percent increase was planned,
corresponding to a growth rate of 7 percent annuall. This
implies an annual growth rate in income per capita of about
4.0 to 4.5 percent.

Recent estimates show that national income has increased
to LE 1905 million in 1964/65 in current prices (or LE 1762
million in constant 1959/60 prices).l'1 That makes for an annual

compound rate of growth around 7 percent.

 

lBent Hansen, ”The National Income of UAR (Egypt)", Memo
No. 355, Institute of National Planning, Cairo, 1963.

2Bent Hansen and Girgis Marzouk, 92,cit., p. 320.
3Recently, and in view of the circumstances of the Middle
East conflict, the plan period was extended to 1972 instead
of 1970.

“See Table 7, Ch.11.

 

154

It can be concluded therefore that the national income
has increased at a rate near 6.5 percent during the last decade,
and is expected to grow at a level near that rate or slightly
higher.

The agricultural sector has contributed around 13 percent
of the overall increase in the national income during the plan
period, while industry contributed more than 31 percent. This
reveals the emphasis put on the industry in the investment
plans. Over one-third of the total investment in the first
Five-Year Plan was directed to industry and electricity, the
share of agriculture and irrigation in the investment program
was no more than one-fourth and a good part of it was for the

I It followed that the

construction of the Aswan High Dam.
agricultural sector grew at a rate around 3 to 3.5 percent
during the last decade. This indicates that other sectors of
the economy should have grown at a rate from 7 to 8 percent.
All the previous figures on national income refer to
gross national product. It is realized that the private

disposable gross income may be more relevant when projecting

demand is concerned, however, the difference between the two

 

INational Bank of Egypt, "Egypt's Economic Growth, 1952-
1963", Economic Bulletin, Vol. 17, No. 1, 1964, p. 50.

155

terms is the total public net income, which grew nearly
proportionately with the GNP,‘I this means that the previous
rates can still be valid for disposable income.

It is not quite true that the rural income coincides
with the agricultural income, a part of the income drawn from
the agricultural activities goes to non-rural residents and vice
versa. Of the estimated labor force of 7.8 millions in 1960,
4.4 millions or about 56 percent were involved in agriculture
or related activities. It is thus a reasonable approximation
that during the plan years the annual rate of increase in
income in rural areas was about 3.5 to 4 percent, while in
urban areas was about 7 to 8 percent.

In arriving at consumption projections it was assumed
that these rates would be sustained up to 1975, namely about
4 percent for rural areas and 8 percent for urban areas. For
1985 projections a rate of 4.5 percent in rural areas and
8.5 percent in urban areas was used. Considering the trend of
the agricultural sector's share in national income over the
past, these rates would imply a growth rate of national income
of 6.5 percent over the period 1964-1975 and of about 7 percent
over the period 1964-1985.

 

IThis is true for a long series of data with the exception
of few years in the early sixties, as a result of the major
structural changes in the economy. See Chapter 11, sec. 2.4.

 

156

6.4 Consumption Projections

In projecting consumption, three basic variables are
involved: (1) rate of growth in per capita income, (2) the
rate of population growth, and (3) income elasticity of demand.
The relationships between these variables and food consumption

were expressed mathematically by OhkawaI as follows:

d = p + gn + pgn
where
d = percentage change in national demand for food
in a given period
p = percentage change in population
g = percentage change in per capita income

n = income elasticity of demand

Ohkawa dropped in the last term of the formulation Pgn
reasoning that it is of small importance.

Using this formula to estimate the percentage change in
demand does not require knowledge of the actual per capita
income or the actual size of the population either in the base
year or in the target year. If the compound rate of growth for

each is known,the percentage chan be derived by applying

 

_ IAs cited by Robert Stevens, Elasticity of Food'ConSumption
Associated with Changes in Income in DeveIOping Countries,
*23. cit., p. 4. ‘ “

157

the compound rate formula

From Yn = Yo (1+r)n
where Yn is the size of the population or income in the target
year, Yo the size of the population or income in the base year,
n the number of years between the two dates, and r the compound
rate of growth,it follows that

Yn -0Yo = (1 - r)" - 1

In projecting the percentage change in national demand for

food products, the study faces a difficult choice:

(I) to estimate the percentage change in demand for the
country as a whole using weighted elasticity co-
efficients for rural and urban areas. In so doing
one might question the accuracy of the weighting
procedure under the apparent dynamic condition. In
declining percentage of rural population has already
been discussed and assumed to continue though at a
slower rate.

(2) to estimate the percentage change in demand for urban
and rural areas separately using the appropriate rates
of change in population and per capita income, and thus
arrive at the volume of total demand in the target
year. This procedure would have been preferred had

information been available on the actual consumption

 

158

level in both areas in the base year. The figures are
available on national level, and any attempt to break
them down would at best be some kind of judgment.
None of the two paths is absolutely unerring. However,
the problem involved in estimating the national consumption
is the expected decline in the percentage of rural population
from about 60 percent in the base year to an estimated percen-
tage of about 54 in 1985. The difference, over a period of 21
years would not be of a substantial size that makes for a signi-
ficant change in the weighted elasticity coefficients, and
therefore can be disregarded. This can further be justified
that though rural laborers migrate to cities and urban centers,
it is doubted that their consumption behavior and decisions would
drastically change. It is most probable that they keep their
rural consumption patterns under slum living condition at least
for an adjustment period.I
The percentage change in population was calculated based
on population projections under the first two assumptions.
These percentages are shown in Table 43.
From the projected compound rates of change in national

income, the implied rate of change in per capita income can

 

IArthur Lewis, ”Unemployment in Developing Country", Lecture
to Midwest Research Conference, USA, Oct. 1964.

 

159

be calculated based on population projections. Therefore,
the percentage change in per capita income can be estimated
throughout the projection period.

The estimates of total expenditure elasticity of demand
calculated in Chapter IV using the logarithmic model were
utilized. Elasticities used in projecting the demand for
commodity groups (in money terms based on 1964 prices); are
expenditure elasticities, while those used in projecting the
demand for selected individual products are quantity
elasticities.

In projecting the demand for wheat and maize, the study
faces a difficult choice. These two products are demanded by
the consumers in several forms: whole grain, flour, and bread.
The last two forms are basically mixtures of several products
as discussed earlier in Chapter IV.

Three possible alternatives can be explored in order to
project the aggregate demand for the two products:

(1) Project the demand for each form of the two products
separately, then add them up after being converted to uniform
units. This would be valid if the interrelationships between
these different forms remain constant over the projection
period. These projections, however, require knowledge of the
level of consumption of each form separately in the base year
and its elasticity coefficient. These data are unfortunately

not available.

 

160

(2) Convert all forms of consumption of each product
to the grain form, and utilize an “adjusted" income elasticity
for each product to project its aggregate demand. However, we
lack a solid base to adjust the whole grain elasticity for the
two products due to the complex relations that exist in the
consumption patterns of the two products.

(3) Project the demand for wheat and mazie indirectly.
This is done by projecting the increase in demand for cereals
and starches as a group. Subtract from it the projected
increase in demand in rice and millet for which separate pro-
jections could be made. The rest, as was shown earlier in
Table 29, is largely wheat and maize in different forms. This

procedure was chosen.1

 

IThe calculations were performed as follows:

(1) Projections were made for the percentage increased in
total expenditures for cereals and starches using
the projections of population and income growth
together with the estimate for elasticity of demand
for the group.

(2) Separate quantity projections were made for rice and
millet, which were then multiplied by the base year
prices of the two products. These total values were
then subtracted from the projected total expenditures
for cereals and starches.

(3) Assuming that the pr0portion of wheat to maize in the
diet will change slightly in favor of wheat in the
target years. It was assumed that wheat's share will
increase from 51 percent in 1964 to 60 percent in
1975 and to 65 percent in 1985.

161

For commodity groups the highest projected increase in
demand was for fruit (about 100 percent increase is projected
for 1975 and about 260 percent for 1985). High rates of
increase were also projected for meat, fish and eggs; milk
and dairy products; tea, coffee and other stimulants. The
lowest projected rates were for cereals and starches; dried
legumes; and oils and fats. The three groups have a projected
percentage of increase of around 55 percent for 1975 and
around 130 percent for I985.

The percentage increase in demand for selected commodities
are projected for 1975 and 1985 under the two popu1ation
hypotheses.I 1

High increases were projected for poultry, eggs, milk,
citrus, meat, and to some extent potatoes.2 The lowest
projection was for millet, mainly because of its negative
income elasticity of demand. Wheat was projected to increase
by about 100 percent by 1975 and 220 percent by 1985. Rice
was projected to increase at a slower rate. Other products
show reasonable increase between 50-60 percent for 1975 and

IOO-125 percent for 1985.

 

INamely hypothesis 1 and hypothesis II as were defined in
section 6.2.

2The high rise in the demand for potatoes is mainly due
to the high income elasticity of demand for that product in
rural areas.

162

Table 43. Projected Population in 1975 and 1985 and the
Percentage Change from 1964 (Base year)

 

 

 

Year Assumption* Population Efggeggggnggfinge
million percent
1964 28.6
1975 | 39.7 38.8
11 38.5 34.6
1985 | 52.5 83.6
11 48.3 68.9

 

*See Section 6.2

Source: Calculated based on:

Committee, op. cit.

UAR Central Statistical

Table 44. Projected Annual Rate of Growth in National and
Per Capita Income, and Percentage Increase in

Per Capita Income

1964-1975 and 1964-1985

 

 

 

(Base Year: 1964)
Population
Projection 1975- 1985-
Assumption
Compound annual rate
of growth in national
income 6.5 7.0
Implied annual rate | 3.7 4.2
of growth in per
capita income 11 3.8 4.5
Percentage increase I 48.8 136.4
in per capita income 11 50 3 15] 3

 

163

Table 45. Percentage Change in Demand for Commodity Groups

 

 

 

 

 

from 196 to 1975 and 1985
Target
Year: 1975 1985
Population I II I ;;
assumptIon:
Cereals and starches S3 49 123 113
Dried legumes 6O 56 142 133
Meat, fish, and eggs 87 84 219 219
Oils and fats 52 52 131 121
Milk and dairy products 81 78 202 200
Vegetables 67 64 164 158
Fruit 101 98 257 261
Sugar and sugaries 72 69 177 173
Beverages 73 70 180 175
Tea, coffee, and other
stimulants 86 84 217 216
Other food stuff 62 58 148 140
Total food 71 68 175 170

 

 

 

164

.0 >0>c3m u0mozm 0.0;0mooz 0gp mo mu.:m0m >c0c_E..0cm 0;

 

.3385 .39 :16... dom—

.eo_66~___eoz

.mmesm_c mo\:om_ toe

 

 

 

 

 

 

 

o..oom mo co.u0Lum.c.Eo< .m.<.: "Eccw 0000.30.00 ”00L30m
~o~.m:o.~ omm.moo.~ :om..mo._ mNm.moo._ ~mm.mom om.:m 666. .666.
omm.__~ oom.m_~ oo:.mm_ .mm.~:_ ooN.oo m_.m 00:66 6660 Lasso
Nmm.ko~ m:..mo~ mmm.ko_ o~¢.mo_ ooo._m m~.m memeuo
0,000.00 008
omm.o_~ _NN.:_~ .mm.om_ mmo.~m_ 608.6“ N..N mmmmtm>6m
mo~.mo_ www.mm_ mmm.:m Nmm.mo :_m.m: A... mm_emmsm a tmmam
oom.mm_ mom._m_ me_.m. mm~.:. Nem.om _m._ e_ste
_::.om_ o_o._o~ oko.:~_ :m_.k~_ o:_.ok ok.~ mm_emummm>
:o_.:NN .mm.mk~ mmo.~o_ okm.mo_ mom._m 6N.m .6686 >t_me a x__z
:oo.om omm.mm ~m_.mm mok.oo o_m.mm mm._ meme 6:6 m__o
Noo.k~o Noo.k~o mmo._om 6mm.mom :mm.om_ No.6 6666 w ;m_c .6662
m:~.:. o__.kk __~.me oom.om mom._m m... 662566. 66_to
.:~.ko: Nom.o~e Nmm.:m~ omm.~mN om_._m_ ok.o 6656.666
0:0 m.00c0u
--mmcmmoo;u .m..wi .Mmu
,0m0c0>0 monotm
__ . _ __ _. .muOF o0u:m_03 >u.ooeeoo
co.umE:mmn 00.000 L06
mom. m~m_ mom—\aom. mcmm>
.0>04 :wm. c.0sh Ou 00L00EOQ m0 mmm. 0:0
mum. c. monotu >u.ooeeoo Low mco.uo0aoLm m0cou.oc0axM co_u Eamcou .0: 0.00p

165

Table 47. Percenta e Change in Demand for Selected Commodities
from 196 to 1975 and 1985

 

 

Target 1975 1985,

Year:

 

Population I II I II
assumption:

 

Wheat 111 105. 229 217
Maize 42 41 85 78
Millet 23 18 40 20
Rice 67 64 153 156
Beans 58 54 136 127
Lentils 49 45 103 101
Meat 76 73 189 186
Poultry 125 123 324 335
Fresh fish 71 67 173 168
Eggs 99 97 252 256
Oils 47 43 108 96
Saturated oils 72 69 176 171
Milk 106 104 273 279
Potatoes 75 72 185 181
Onions 42 38 94 8O
Citrus 102 100 261 266
Dates 59 56 141 133

Sugar 56 52 131 121

 

166

Table 48. Consumption Projections of Selected Commodities in

1975 and 1985 as Compared to Their 1964 level

(in thousands of metric tons)

 

 

 

 

Efiifiiiiion 1964 I975

assumption 1 1| 1 11
Wheat 1880 3968 3862 6201 5958
Maize 1806 2644 2574 3343 3212
Millet 614 755 724 859 737
Rice 889 1485 1458 2338 2275
Beans I62 256 249 382 368
Lentils 22 33 32 47 44
Meat 227 399 393 656 649
Poultry 69 155 154 292 300
Fresh fish 139 238 232 379 372
Eggs 36 72 71 127 128
Oils 167 259 252 366 345
Saturated oils 30 51 51 83 81
Milk 1278 2633 2607 4767 4844
Potatoes 241 422 414 687 677
Onions 366 520 505 710 659
Citrus 274 553 548 989 1003
Dates 239 380 373 576 557
sugar 402 627 611 929 888

 

Source: For 1964 consumption figures, UAR Ministry of Agri-

culture, “Food Balance Sheet, 1964”, MonthlygBulletin
of Agricultural Economics and Statistics, July T965.

 

167

Table 49. Percentage Distribution of Annual Total Expenditures
on FOOd Products in I958/59, 1964/65, 1975 Projections
and 1985 Projections

 

 

1975 7T985
Commodity l958/ 1964/ pgpjections projections

groups 592 65 I II I II

 

21.8 19. 19.4 17.7 17.3

Cereals 8 starches 34.5 5
Dried legumes 3.4 3.6 3.4 3.4 3.2 3.2
2 5

 

Meat, fish 8 eggs 19. 22.4 24. 24.6 26.1 26.6
Oils 8 fats 3.1 4.4 4.1 4.0 3.7 3.6
Siéfiuitia‘ry 12.7 10.4 11.0 11.1 11.5 11.6
Vegetables 6.3 8.7 8.5 8.5 8.5 8.3
Fruit 4.7 4.2 4.9 5.0 5.5 5.7
Sugar 8 sugaries 6.4 5.7 5.7 5.7 5.8 5.8
Beverages 3.9 8.7 8.8 8.8 8.9 8.9
Other food stuff1 6.8 10.1 9.6 9.5 9.1 9.0
Total 100 100 100 100 100 100

 

IExcludes tea, coffee, liquor and other stimulants.

2Weighted average for rural and urban areas.

ures: Table 17 Ch. IV.

Source: l9
19 75, and 1985: Calculated from Table 49.

9 f’
5

\OLQ

168

 

 

 

 

 

Their 1964 level (in kilograms)
Year and 1985
population 1964
assumption I II I 11
Wheat 66.7 99.9 100.3 118.7 123.3
Maize 64.0 66.6 66.8 64.0 66.5
Millet 21.8 19.0 18.8 16.4 15.2
Rice 31.5 37.4 37.9 44.5 47.1
Beans 5.7 6.4 6.5 7.3 7.6
Lentils 1.1 9 .8 .9 .9
Meat 8.0 10.0 10.2 12.5 13.4
Poultry 2.4 3.9 4.0 5.5 6.2
Fresh fish 4.9 6.0 6.0 7.2 7.7
Eggs 1.3 1.8 1.8 2.4 2.6
Oils 5.7 6.5 6.5 6.9 7.1
Saturated oils 1.1 1.3 1.3 1.6 1.7
Milk 45.4 66.3 67.7 90.8 100.3
Potatoes 8.5 10.6 10.7 13.1 14.5
Onions 13.0 13.1 13.1 1.35 13.6
Citrus 9.4 13.9 13.2 18.8 20.8
Dates 9.4 13.9 14.2 18.8 20.8
Dates 8.5 9.6 9.7 11.0 11.5
Sugar 14.2 15.8 15.9 17.7 18.4
Source: Calculated from Table 48.

 

169

These projections assumed no major changes in relative
prices not only for domestically produced products but also
for imports and exports as well. These prices, are subject
to world supply and demand and also other exogenous factors,
and their stability depends to a great extent on the stability
and direction of movement of these factors. Aside from these
economic factors, the Egyptian market is influenced by
political considerations. Under such conditions, the country
may decide or be obligated to Shift from one course of action
to another. These frequent moves might imply a revision of the

1 The main

whole cost structure of the production sector.
point is that changes in relative prices or other politically
inspired changes are likely to occur and their frequency is

higher than in other countries with a relatively stable market

economy.

6.5 Detailed Results for Food Commodities

The projections made earlier in this study were based on
the assumption of stable relative prices. How valid was this

assumption and how valid are the projections themselves?

 

IAn example is the drastic change in cost structure as
a result of the U.S. Government decision to stop its sales of
surplus food products to Egypt under P.L. 480 at mid 1967.'

 

In.

170

The Egyptian agricultural policy is incorporated in, and
is, an integral part of the country's overall plan to double
the gross national product within the present decade. The
basic aim of this policy is to increase the degree of self-
sufficiency. In view of the rapidly growing demand for food,
price and production policies of food products and the
direction of trade policies need to be re-evaluated.

Without introducing any exogeneous factors, there are
certainly factors that would affect the projections from within.

This increase in demand has to be fulfilled by a corresponding

increase in local production, in imports, or some combination

of the two. Unless this is done, excess demand is more likely

to put pressure on prices. Whether the government policies

would allow the price mechanism to function for all products

is yet to be seen.

The present section examines the implications of such

structural changes of demand, and discusses the future prospects

of the Egyptian food economy.I

. ISource of information on production levels: U.A.R.
Ministry of Agriculture, The Monthly Bulletin of Agricultural

Egonomics and Statistics, Several issues.

 

171

6.5.1 Grains

In this group, wheat and rice were projected to have the
highest increase in demand. Wheat would increase by some 90
percent of its 1964 level in 1975, and by about 200 percent
by 1985. This implies that wheat consumption levels would
increase from its 1964 level of 1.9 million tons to some 3.6
and 5.6 million t)ns in 1975 and 1985, respectively. These
figures refer to the demand at the consumers' level. When
converted to the farm level, using the same rates of extractions,
loss, and other uses, as in 1964, we end up with about 4.1
million tons of wheat in 1975 and 6.8 million tons for 1985.

Average wheat production for 1960-1964 was about 1.5
million tons. If local production continues to grow at the
same rate of growth as it was during the last two decades,
large imports are required in order to fulfill the projected
increase in demand. It has been argued that wheat productivity
in Egypt can be increased by as much as 140 percent if a
prOper "package approach" of techniques, varieties, improved

drainage, was applied.1 However, a more likely path is an

 

ISee: Bushra Abdel-Sayed, ”The Potential Use of
Fertilizers for Intensification and Development of Agriculture
in the United Arab Republic", Ph. D. Thesis, Michigan State
University, 1969.

 

172

increase in the administered prices by as much as 40-50
percent, which would lower the volume of demand by about

25 to 30 percent. If the local wheat production follow the
present trend, it will reach a level between 1.8 and 2.0
million tons by 1975, and this would leave about one to one
and half million tons to be imported by that year. Wheat
competes with cotton in the crop rotation, and farmers are
required, by law, to plant one-third of their total crop
acreage to wheat each year. If over the next two or three
decades, world cotton prices declined drastically, wheat
may be grown instead. If that happens, agricultural and

, national incomes in Egypt will be affected and the whole
demand structure would change.

Demand for maize shows an increase of about 60 and 100
percent in 1975 and 1985, respectively. Taken by itself,
local production is capable of fulfilling that demand
particularly with the recent trend of increasing productivity.
The average yield per feddan of maize has increased from .92
ton in 1954 to 1.17 tons in 1964, or some 25 percent within
ten years. Yet, there is a great potential of increasing
Iaroductivity by introducing hybrids and proper fertilization.
It is feasible to double the average yield per feddan in two
decades period. If this can be realized it would be of

[Maramount importance for: (l) substituting wheat, to certain

 

173

extent, particularly in rural areas, and (2) increasing live-
stock production.

It Should be emphasized here that the projections of both
wheat and maize were made indirectly as described earlier in
Chapter VI. The hatio of wheat to maize in the national
diet in the target years was arbitrarly estimated.

Millet demand was projected to show the lowest increase.
The projected demand in 1975 and 1985 can be easily fulfilled.

The demand for rice was projected to increase by about
65 percent in 1975 and by 160 percent in 1985, reaching a
volume of about 1.5 and 2.3 million tons, respectively. Local
production, with a level of more than two million tons in
1964, is expected to increase after the completion of the
Aswan High Dam to a level as high as 3 million tons.I However,
rice is a major export commodity, and It is the country's
policy to give priority to the foreign demand. Rice exports
fluctuates around an average of half million ton a year and

have a good market potential.

 

IThe Aswan High Dam Reservoir of water would make it
possible to expand rice acreage to around 1.2 million feddans.
See: M. El-Kateb, ”Agriculture Production after the High Dam",
in U.A.R. Ministry of Agriculture, The Monthly Bulletin of
A ricultural Economics and Statistics,TVol. 16, Dec. 1965,
pp. 7-T18.

 

174

6.5.2 Meat and Poultry

Meat was projected to increase over its 1964 level by
about 75 percent in 1975 and by about 187 percent in 1985,
reaching about 400 and 650 tons, respectively. Of the
227,000 tons of different types of meat consumed in 1964,
30,000 tons were imported to supplement local production.

In recent years, the country has experienced severe shortages
in meat to the extent that the government, after a long,
unsuccessful experience in controlling meat prices, finally
admitted that the price mechanism should function to clear
the market. Certain days only were allowed for selling meat
in an effort to surpress demand in a country where day-to-day
shopping is the common practice.

1964 figures Show a total population of livestock of
about 1.6 million head of cattle, 1.5 million head of buffaloes,
1.7 million head of sheep, and about one million goats. Camels
total about 200,000 and pigs around 19,000. Large numbers of
cattle, sheep, and in some years camels are imported mainly
for slaughtering purposes from Libya and Sudan.

Forage production in Egypt is a limiting factor in
expanding the livestock population. Bersim (Egyptian clover)
is the only forage crop grown in the country. However,

possibfle imporvement lies in replacing native cattle which

 

175

has no distinct breed by adapted highly productive breeds. If
livestock production is to rise, better and well balanced
rations must be made available. The problem is that livestock
has been always regarded as a by-product, or at least given a
second priority after crop production, and this in turn, has
resulted in meat shortages in the past few years.

Poultry consumption was projected to Show the highest
increase of about 125 percent from its 1964 level in 1975,
and more than 325 percent in 1985. This increase is mainly
attributed to the high income elasticity of demand and implies
a projected consumption level of about 150,000 tons for 1975
and about 300,000 tons for 1985. Though large amounts of
slaughtered chicken imports were received during the late
1960's from the U.S. under P.L. 480, it is believed that these
imports were a by-product of the whole P.L. 480 package. The
local production in 1964 made up more than 98 percent of national
consumption of poultry. However, it is doubtful that local
production can be more than doubled by 1975 and tripled by
1985. More likely, prices would increase by as much as
40-80 percent by 1975, which would reduce demand by as much
as 50 to 100 percent from its projected level of 1975, in
'view of the high price elasticity. If this happens, the
demand would be at a more manageable level. For 1985, it

is hard to predict a certain path. In such a period, long

 

176

enough to give more attention to poultry production and to
substantially increase it, it is more likely that production
would increase partly in response to higher prices (if left
uncontrolled), or that some combination of local production
and imports may fulfill consumption requirements.
At any rate, meat and poultry situation is expected to
be critical in the future for three reasons:
(1) the high income elasticity of demand for these
products;
(2) the present shortages, particularly in meat;
(3) the low level of animal protein in the national
diet, and the need to improve its pr0portion if the

diet is to be balanced.

6.5.3 Fish

Consumption of fresh fish is expected to increase by as
much as 70 percent in 1975 and around 170 percent in 1985.
Yet, this increase is not an alarming one for a country with
a long coast of more than 1,500 miles on the Mediterranean to
the north and the Red Sea to the east, more than 26,000 miles
of the river Nile and its connected network of canals, and

the huge lakes of Manzala, Borolus, Mariut, Idko, and Qarun.

177

Navigation is possible all year round. Fish production

which totalled around 155,000 tons in 1963,I can be tripled

if sufficient investment is directed to provide modern

fishing equipment. A great potential lies in this area and
its expansion would release some pressure on meat and poultry,

and help in balancing present diets.

6.5.4 Milk and Eggs

Milk and eggs are associated with livestock production.
Demand for milk is expected to double its 1964 level by 1975
and to triple it by 1985. Unless production and marketing
facilities improve significantly, milk is likely to be in short
supply. Prices have risen in the past few years and such a
trend is likely to continue. Milk yields at the present time
are very low, between 1,000 and 3,000 lbs. per year.

Demand for eggs was projected to increase almost 100
percent by 1975 and around 250 percent by 1985. Egg prices
have increased and would continue to increase unless poultry
production is given prompt attention. The present average

egg production is small, between IOO-120 eggs per year per hen.

 

'Ibid., Dec. 1965, p. 95.

178

With the exception of two native breeds the majority of
chickens in the country can scarcely be recognized as a
particular breed.

For both milk and eggs, it is difficult to predict future
trends. However, few points seem to be in order:

(1) Introduction of improved livestock and poultry
breeds are bound to raise productivity to a great extent.
Several livestock breeds, mainly the Holstein-Freisian and
Herford have been introduced as specialized milk and meat
cattle, respectively. Also imported standard chicken breeds
are now being raised in massive numbers on experimental farms
for breeding and crossing purposes.

(2) Expansion in maize production would increase live-
stock production substantially, as the availability of year-
round forage crops is a major limitation on expanding live-
stock production.

(3) Pricing policies of the past have had a discouraging
effect on production. Indeed, under the unrealistic price
controls, farmers lacked the incentive to pay much attention
to milk and eggs and other products with unfavorable cost-price
relationships.

(4) Particularly for milk and eggs, refrigeration and'

other improved marketing facilities are of a vital importance.

179

6.5.5 Potatoes and Onions

These two “products have two common characteristics, (1)
both are major export products, and, (2) their projected
increase in demand seem to be reasonable. Domestic production
can certainly fulfill the projected increase in the demand for
both of them if the price and cost relationships for each of
them remain constant.

The choice, therefore, would be between the export market
and the domestic market, and would depend to a great extent

on the relative prices between the two markets.

6.5.6 Citrus and Other Fruits

A substantial increase in the demand for citrus and
most likely for other fruits seems to be a natural course, in
the light of its high income elasticity.

A rapid expansion of about 60 percent in fruit acreage,
particularly citrus, was made during the past decade. This
expansion made it possible for sizable quantities of citrus
to be exported annually. This acreage increase will be
reflected in a continued expansion in production in the coming

decade as new orchards reach their full production stage.

180

Citrus production in 1964 was about 475,000 tons, and
is expected to reach more than 600,000 tons by 1975. However,
it is again the choice between exports or domestic consumption
depending on the foreign markets' capacity and prices in

relation to local markets, that would determine the future

path for fruit consumption.

6.5.7 Oils and Saturated Oils

Oilseed production in Egypt centers mainly around cotton-
seed and to some extent peanuts and flax. In the last few
years, local production, which is a by-product of cotton
industry, fell short of consumption requirements. Increased
quantities of either processed oils or cottonseeds were
imported. In view of the projected increase in the demand for
oils by about 50 percent in 1975 and about 100 percent in 1985,
imports must continue at an increasing rate. These projections
may underestimate the actual demand as oils and saturated
oils may substitute for butter and ghee if their prices rise

significantly, which is likely to happen.

 

CHAPTER VII

SUMMARY AND CONCLUDING REMARKS

This study has focused on estimating parameters deter-
mining the demand for food in Egypt. These estimates were
used, along with estimates for the annual rate of growth in
GNP and the projected rates of pOpulation growth, to estimate
the aggregate demand for food products in 1975 and 1985. To
accomplish this, several types of data, cross-sectional and
time-series, were utilized, and more than one model of analysis
was introduced. However, it should be emphasized that, in a
sense, the study was one-sided as it focused on the demand
side of the food problem. Little attempt was made to analyze
supply or to project it.

In the cross-sectional analysis, which was based on the
published data of the household budget survey of I958/59, the
demand for food was assumed to depend on income, size of the
household, and location. Three alternative functional forms
were considered: (1) logarithmic, (2) semi-logarithmic, and
(3) log-inverse. Of these three functions, the logarithmic
relationship seemed to provide the best fit. The results

drawn from it were utilized in projecting demand.

181

182

Throughout the cross-sectional analysis, estimates for
income elasticity of demand for food groups and individual
food items were calculated for both urban and rural areas,
and a weighted average was introduced to provide an estimate
for the entire population.

The results of the cross-sectional analysis has indica-

ted that two-thirds of the total expenditures in rural areas
were spent on food and beverages, compared to about one-half

in urban areas. Using Lorenz curves, total expenditures and
food expenditures were shown to be more equally distributed

in rural areas than in urban areas. Two explanations were
offered in this connection: First, that qualities and therefore
prices paid differ much more in urban areas, and second, that
incomes, occupations, social status, and tastes vary greatly

in urban areas.

The income elasticity of demand for food in Egypt was not
as high as results from other countries' data suggested. The
.67 elasticity coefficient for Egypt is intermediate rather
than high in comparison with estimates for other countries
(Table 51).

The relations between quantity and expenditure elasticity
figures were as expected, expenditure elasticities being higher

than quantity elasticities.

183

Table 51. Estimates of the Income Elasticity of Demand
for Food in Several CountriesI

 

 

.‘—_

 

Country Date of estimate Income Elasticity
United States2 1955 .25
Japan2 1955-56 .52
1ta1y3 1952-53 .62
Panama2 1952-53 .63
Egypt[-1 1960 .67
Greece2 1957-58 .70
YugosIavia2 1955 .72
Brazil3 1953 .79
Libya3 1950 .80
Puerto Rico2 1952 .80
Ghana2 1954-55 .81
Ceylon3 1953 .81
India, Punjab3 1950 .81

 

'All these estimates were derived from budget analysis
studies.

2Robert Stevens, Elasticitygof Food Consumption Associated
With Changes in Income inDDeveloping Countries, 22. cit.,
p. 16.

3H. S. Houthakker, ”An International Comparison of Households'
Expenditure Patterns", Econometrica, Vol. 25, Oct. 1957,
pp. 532-5].

uWeighted average for rural and urban areas, see Ch. IV.

 

184

High income elasticities for home-produced wheat, maize,
and millet induced a detailed discussion of grain consumption
patterns. This, in turn, resulted in the use of an indirect
method to project the demand for wheat and maize.

Although household Size was expected to be an important
factor in the demand for food, its significance failed to
materialize when the effect of the household type, based on
its size, was tested by the analysis of variance procedure.
However, when the number of consumption units was included in
the regression analysis, this variable produced a coefficient
with a value that differed Significantly from zero.

Two methods were used for estimating price parameters.

A combined model utilizing cross-sectional and time-series data
was applied to six products for which sufficient time-series
data were available. For other products, a newer mathematical
approach based on several assumptions about the relationship
between income elasticities and price elasticities was utilized.
The results of the combined model were not very satisfactory in
terms of statistical significance. However, the mathematical
model produced fairly reasonable results suggesting that further
exploration of the technique might prove to be of a considerable

value.

185

Projections of food demand were undertaken in the usual
manner. Two population projections were utilized based on

1 Rates of growth

different assumptions about fertility rates.
in per capita income were estimated.2 Using these rates,
together with the income elasticity coefficients obtained in
this study, demand projections were made for 1975 and 1985
for commodity groups and individual food items.

Turning to the estimated percentage increase in total
demand for commodity groups, it was shown that the highest
projected increases were for fruit (about 100 percent in 1975
and 260 percent in 1985); meat, fish, and eggs (85 and 200
percent); milk and dairy products (80 and 200 percent); and
tea, coffee, and other stimulants (85 and 216 percent). The
lowest projected increases were for cereals and starches; dried
legumes; and oils and fats. The last three groups were projected
to increase by about 55 percent in 1975 and about 130 percent
in 1985. This, however, still implies a very high increase in

the demand for these products.

 

IThe two projections imply a compound annual rate of
growth of about 2.8 and 2.5, respectively.

2These rates depend on the chosen population projection.
They range from 3.7 to 4.5 per year.

186

When these projections were made on per capita basis, an
increase in most items was observed. It should be emphasized,
however, that a part of this increase would be absorbed in a
shift to higher quality products rather than higher quantities
as per capita income rises.

The study has shown that rather large increases in the
demand for food products are to be expected. This demand has
to be fulfilled through increasing food production, increasing
food imports, or some combination of the two. Prospects for
increasing productivity seem beSt for certain crops such as
maize, vegetables, fruit, oil crops, sugar cane, and livestock
products. ,There is little doubt that increased productivity
can greatly aid in solving the food problem for one generation.
Yet, it is hard to see how it could do so for a population
that will almost double in about twenty-five years. The only
feasible path, given the present state of technology, seems to
be a drastic effort to decrease the rate of population growth.

The findings of this study raise questions about the
extent to which the nation would be able to provide its food
supplies, and also the question of trade policy with regard to
food products. It has been often argued that Egypt should
stop producing wheat, and concentrate her land and water

resources in products for which she has greater comparative

187

advantage such as cotton, rice, oil crops, as well as sugar

I This would imply greater dependence on

and vegetables.
trade, but if imports were matched by exports of other crops
with higher value, there should be no cause for concern.
However, for sounder decisions to be made in the area of trade
policy, more analytical studies are needed particularly for
grains for which shortages seem to be alarming.

Finally, it Should be mentioned that in several stages
of this study, insufficient or unreliable data set a limit on
pursuing the analysis in the desired directions. However, it
is hOped that in the future, when sufficient data become
available, it may be possible to re-evaluate the results of
this study, and examine the impact of time trend, economic,

social, and institutional factors on food consumption patterns.

The present writer, at least, would h0pe to do so.

 

ISee: W. F. Owen, ”Land and Water Use in the Egyptian
High Dam Era", Land Economics, Vol. XL, August 1964, pp. 227-93.

BIBLIOGRAPHY

Abu-Lughod, Jane L., ”Urbanization in Egypt: Present State
and Future Prospects”, Economic Development and Cultural
Change, Vol. 13, April 1965.

Allen, R. D., and A. L. Bowley, Family Expenditure, London, 1935.

Amin, Galal, Food Supply and Economic Development, University
of London, 1966.

Barlow, Robin, Intersectoral Income Redistribution in Egypt,
Ph. D. Thesis, University ofMichigan, 1961.

Black, G., ”Variations in Prices Paid for Food by Income Levels”,
Journal of Farm Economics, Vol. 34, 1952.

Brady, 0. S., and H. A. Barber, ”The Pattern of Food Expendi-
tures”, Review of Economics and Statistics, Vol. 30, 1948.

Burk, Marguerite, "Some Analysis of Income-Food Relationships”,
Journal of the American Statistical Association, Vol. 53,
1958.

Clarkson, Geoffrey P. E., The Theory of Consumer Demand: A
Critical Appraisal, Prentice-HSTI, Inc., 1962.

Duesenberry, James 5., "Income-Consumption Relations and their

 

Implications”, in: Incomey Employment and Public Policy,
Eszgys in Honor of A. H. Hansen,TW.TNorton 8 Company,
19 . ‘

Elbogdadi, A. Latif, The Five-Year Plan for the Economic and
Social DeveIOpment of the U.A.R., 1960-1965, Cairo, 1960.

Eleish, Gamal, ”The Applicability and Utilization of the Input-
Output Model in a Developing Economy: The Case of Egypt
Examined”, a Paper Presented at the International Con-
ference on Input-Output Techniques, Geneva, Sept., 1961.

El-Kammash, Magdi M., Economic Development and Planning in
Egypt, Frederick A. Praeger,New York, N.Y., 1966.

188

I89

Enke, Stephen, Economics for Development, Prentice-Hall, Inc.,

1965.

Farrell, M. J., ”The New Theories of the Consumption Function”,
The Economic Journal, Vol. 69, Dec. 1959.

Friedman, Milton, The Theory of the Consumption Function,
National Bureau of Etonomic Research, PrincetonTPress,

I957.

Frish, Ragnar, "A Complete Scheme for Computing All Direct and
Cross Demand Elasticities in a Model with Many Sectors“,
Econometrica, Vol. 27, April 1959.

Goreux, L. M., “Economic Growth and Commodity Projections”,
Monthly Bulletin of Agricultural Economics and Statistics,
FAO, Vol. 10, July/August, 1961.

., llIncome and Food Consumption”, Monthly Bulletin
76f Agricultural Economics and Statistics, FAO, Vol. 9,
Oct. 1960.

 

Haavelmo, T., and M. Grishick, "Statistical Analysis of Demand
for Food: Examples of Simultaneous Equations of Structural
Equations”, Econometrica, Vol. 15, 1947.

Hansen, Bent, and Girgis Marzouk, Development and Economic
Policy of the U.A.R., North-Holland Publishing Company,
Amsterdam, 1965.

 

Hildreth, Clifford, ”Combining Cross-Section Data and Time-
Series Data“, Cowles Commission Discussion Paper No. 347,
May 1950. (unpublished).

., llPreliminary Considerations Regarding Time-Series
and/or Cross-Section Studies”, Cowles Commission Discussion
Paper No. 333, July 1949. (unpublished).

 

Hoch,|rving, ”Estimates of Production Function Parameters
Combining Time-Series and Cross-Section Data”, Econometrica,
Vol. 30, 1962.

Houthakker, H. S., An International Comparison of Household
Expenditure Patterns”, Econometrica, Vol. 25, Oct. 1957.

., “The Econometrics of Family Budgets”, Journal of
Royal Statistical Society, Series A., Vol. 65, T952.

 

190

International Labor Office, ”Report to the Government of the
United Arab Republic on Family Budget Survey and on the
New Consumer Price Indices in the Southern Region of the
U.A.R. (Egypt), Geneva, 1960.

Iowa State University Center for Agriculture and Economic Adjust-
ment, Food: One Tool in International Economic Development,
Iowa State University Press, Ames, Iowa, 1962.

Issawi, Charles, Egypt at Mid-Century: An Economic Survey,
Oxford University Press, 1957.

Johnston, J., Econometric Methods, McGraw-Hill Book Company,
Inc., 1960.

Kardouche, George, The U.A.R. in DeveIOpment: A Study in
Expansionary Finance,Fredrick A. Preager, Inc., T966.

Klein, Lawrence, Introduction to Econometrics, Prentice-Hall
Inc., 1962. '

., ”Estimating Patterns of Savings from Sample
Survey Data”, Econometrica, Vol. 19, 1951.

 

Kuh, Edwin, "Validity of Cross Sectionally Estimated Behavior
Equations in Time-Series Applications”, Econometrica.
Vol. 27, 1959.

Kutzents, G., ”Measurement of Market Demand with Particular
Reference to Consumer Demand for Food”, Journal of Farm
Economics, Vol. 35, 1953.

Lippitt, Vernon G., Determinants of Consumer Demand, Harvard
University Press, Cambridge, Massachusetts, T959.

Liviatan, Nissan, ”Errors in Variables and Engel Curve Analysis”,
Econometrica, Vol. 29, July 1961.

Lubell, Harold, ”Effects of Income on Consumers' Expenditures“,
The American Economic Review, Vol. 37, 1947.

Manderscheid, Lester, "Price and Income Elasticities, Relative
Growth Rates: Some Implications for Understanding Agri-
culture's Terms of Trade”, A working paper presented at
the Agricultural Development Council Seminar on Industrial-
ization and Agricultural Development, Stanford University,
Feb. 25, 1966.

191

., ”Some Observations on Interpreting Measured Demand
Elasticities", Journal of Farm Economics, Vol. 46, Feb. 1964.

 

Marschak, J., ”Review of Schultz, Theory and Measurement of
Demand", Economic Journal, Vol. 49, 1939.

Mead, Donald, Growth and Structural Changp in the Egyptian
Economy, Richard C. Irwin, Inc., 1967.

National Bank of Egypt, "Egypt's Economic Growty, 1952-1963”,
Economic Bulletin, Vol. 17, 1964.

 

O'Brien, Patrick, The Revolution in Egypt's Economic System,
Oxford University Press, 19661

Owen, W. F., ”Land and Water Use in the Egyptian High Dam Era“,
Land Economics, Vol. XL, August 1964.

Prais, S. J., and H. S. Houthakker, The Analysis of Family Budgets,
Cmabridge, England, 1955.

Prais, S. J., "Non-Linear Estimates of the Engel Curve”, Review
of Economic Studies, Vol. 20, 1953.

Saab, Gariel, The Egyptian Agrarian Reform, 1952-1962, Oxford
University Press.'_T967.

Sabri, All, The Years of Social Transofmration, An Evaluation
of the First Five-Year PTan, Dar-ETmaaref,TCairo, 1966.
(Arabic).

 

Sinha, R. P., Food in India, Oxford University Press, 1961.

 

Spengler, J. J., "Population: Cause, Effect, Indicator”,
Economic Development and Cultural Changp, Vol. 9, April 1961.

Sparks, Willard R., Estimates for the Demand for Food from
Consumer Panel Data, Ph. D. Thesis, Michigan State
University, 1961.

Stevens, Robert, Elasticity of Food Consumption Associated with
Changes in Income in Developing Countries, FAER No. 23,
Economic Research Service, U.S. Department of Agriculture, 1965.

192

Stone, Richard (and others), The Measurement of Consumers'
Egpenditure and Behavior in the United Kingdomb (1920-1938),
Gam5riage University Press, 1954.

Tendulkar, Suresh D., 'Econometric Study of Monthly Consumption
Expenditures in Rural Uttar Pradesh”, American Journal of
Agricultural Economics, Vol. 51, No. 1, 1969.

U.A.R. Administration of Public Mobilization, The Preliminary
Results of the Household Budget Survey of T965, CaTro,
May 1967. (Arabic). '

U.A.R. Central Statistical Committee, Household Budget Survey
in the Egyptian Regionngairo, April 1961. (Arabic).

U.A.R. Central Statistical Committee, Popplation Trends in the
U.A.R., Cairo, 1962.

U.A.R. Ministry of Agriculture, The Monthly Bulletin of Agricul-
tureal Economics and Statistics, several issues.

U.A.R. National Planning Committee, General Frame of the Fiye-
Year Plan for Economic and Social—Development-July 1960-
JDne T965, Cairo, 1960.

United Nations, Dept. of Economic and Social Affairs, The Future
Growth of8World Population, Population Studies No. 28,
N.Y., 195 .

United Nations, Dept. of Economic and Social Affairs, The United
Nations Development Decade, United Nations, N.Y., T962.

United States Department of Agriculture, Public Law 480 and
Other Economic Assistance to United Arab RepubTic (Egypt),
ERS No. 83, 1964.

, The Agricultural Economy of the United Arab Republic

IEgyptI, FAER No. 21, 1964.

Wold, H., and L. Jureen, Demand Analysis, John Wiley and Sons,
New York, 1953.

Woytinsky, W. S., and E. S. Woytinsky, World Population and
Production, The Twentieth Century Find: New York,—T953.

 

 

APPENDICES

Appendix A

SOME CONSIDERATIONS IN COMBINING CROSS-
SECTIONAL AND TIME-SERIES DATA
IN A SINGLE MODEL

-In a demand function the two major sets of parameters,
namely, the price parameters and the income parameters, can
be obtained jointly from a set of time series data. Though
it can be argued that there is a tendency to estimate short-
run parameters from time-series samples, and long-run parameters
from cross-sections, this need not be true in all cases.
Nevertheless, there are serious reasons that suggest the pooling
of the two types of samples. These reasons include:I
(l) multicollinearity.
(2) the effects of income distribution.
(3) the problem of identification.
(4) the least square bias.
Furthermore, it is apparent that the pooling technique
will enable us to get in close contact with the decision-making
unit, the household, at the microeconomic level through the

cross-section sample, and also to look over the aggregates at

the macro-economic level through a set of time-series data.

 

IFor an explanation of these mathematical terms, see:
Lawrence R. Klein, pp. cit., pp. 62-69.

194

195

A simple formulation of the standard technique in pooling
the two types of data can be presented as follows: An estimate
of income elasticity is to be obtained from cross-sectional
data, where other variables such as location, size of families,
are considered. This estimate is to be substituted directly
or after certain adjustments in the time-series model, such

as the following single-equation model:

log Qt = BO + Bl log Pt + B2 log Yt
where Q is the quantity consumed, P the price, and Y the income.
log 0 can be adjusted to the income estimates by taking:

log 0% = log Qt - Bi log Yt
and thus estimating the relation:

log Q; = BC + Bl log Pt

The practice of combining the two sets of information has
been applied in a number of studies in the last few years.
HildrethI emphasized the idea of pooling time-series and cross-
sections in his two unpublished papers. Sparks used one of the
several Statistical models suggested by Hildreth to combine the
two types of data in his study to estimate the demand for food

from consumer panel data. His combined analysis estimates a

 

IClifford Hildreth, ”Preliminary Considerations Regarding
Time Series and/or Cross Section Studies”, Cowles Commission
Discussion Paper, No. 333, July 1949. (unpubTiShedT. anH;
”Combining Cross-Section Data and Time-Series Data”, Cowles
Commission Discussion Peper, No. 347, May 1950, (unpuEIished).

 

196

parameter for each family that is constant over time but varies
among families. He viewed these parameters as measuring ”the
effect of unobservable characteristics that were particular

to a family over time”. When these family constants were
estimated, the coefficient of determination was thirty-five
percent greater than the coefficient obtained from the
relationship that did not estimate these constants.l

Stone2 used the pooling technique in estimating the
demand for a large number of food products in the United
Kingdom. He estimated the income elasticity from the Engel
curves and substituted it directly in a Single equation model.

In estimating income elasticity from cross-sectional information,
he took the family size (expressed in equivalent adult scale)
and social class as Separate variables.

Another example of the utilization of the pooling techni-
que can be found in Hoch's estimation of production function
parameters,3 which is based on Hildreth's unpublished papaers.

He suggested a pooling procedure by combining a cross-sectional
sample, and data of a set of firms over a period of years, and
using the analysis of covariance which is a generalized regre-

ssion procedure utilizing aspects of the analysis of variance.

 

IWillard R. Sparks, ”Estimates of the Demand for Food from
Consumer Panel Data”, Ph.D. Thesis, Michigan State University, 1961.

2Richard Stone (and others), 9p. cit.

3Irving Hoch, ”Estimates of Production Function Parameters
Combining Time-series, and Cross-Section Data", Econometrica,
Vol. 30, January 1962, pp. 34-53.

197

When combined data are used, it usually allows a wide
choice of models available to the investigator. But other
problems arise that are not generally encountered when either
type of data is used separately. Among diese problems is the
selection of variables in making cross-sectional estimates of
income elasticity. In addition to the family's income and I
size, region, occupation, education, price expectations, and T

0
other variables may be important. The inclusion of such

 

variables usually adds a burden of additional analysis. A
second problem is the implied assumption of constant elasticity
between households and over time which is hard to substantiate
but, nevertheless, should be made. Other problems include the
assumptions of homoscedasticity and of serial correlation.
Further, though the equations of a logarithmic model are
expressed as linear in logarithms, there is a problem of
aggregation involved, as the log of a sum is not equal to the
sum of the logs. Most values in the time series part of the
pooling are sums. If we assume that prices are the same in a
certain time period for all consumers, which is hard to sub-
stantiate, income cannot be the same, and certain assumptions
have to be made with regard to income distribution in the time
series data. Further investigations may encounter other aggre-
gation problems and may eventually lead to improved aggregation

theories and procedures.

198

.emiam .ee ..moma

 

”chemo. aomflimmma .mpppmaaozH Heoapaaaaem

 

 

 

 

 

.moflpmfip0pm 0cm coap0uflafipoz oHHbsm go moc0w¢ H0LQC00 "Eopm o0p0azoa0o "0ohsom
0.8mm o.mom m.mm m.ao moea mama mo\aoma
o.emm a.mem m.>m m.Ho mama maea no\moma
m.mmm o.mmm :.mm o.mm mmma emma mo\mooa
a.mam m.amm m.mm o.mm mHaH mmaa mo\aomfi
m.amm m.mmm H.mm m.mm :mma mmma Hm\omma
H.38H :.H>H H.om H.om mmma mmma om\mmma
wnama :.Hma :.m: m.m: omHH anH mm\mmma
m.mmH :.mma m.m: m.m: mmaa mmaa mm\>mma
o.maa o.HmH m.aa 0.6: emoa oaofl am\omma
m.aoa H.mea m.aa a.Ha emoa mom om\mmma
H.3pfl m.m:a o.m: H.o: mmoa omm mm\:mma
:.oma :.NMH m.mz H.mm mmoa New :m\mmma
o.mmH o.mHH e.m: H.8m mam mom mm\mmoa
COfiHHHE .m .q . . . . .m .q . . . . QOHHHHE .m .q
mmofipo m0oHpo wmofiho m0oHaQ m0oflpo moofipo
pampmcoo 0:09030 pampmcoo pc09030 pompmcoo 0:00030
2H CH QH CH ca CH 000%
pc0Epm0>cH meoocw mpfiomo e00 0Eoocfi H0C0Hpmz

 

.m0ofimo om\mmma pcmpmcoo
ma oc0 m0oapd pC0LpSo ca .pc0Epmm>cH 0:0 mmEoocH mpfiamo 900 new H05QC<II.H m mqm<e

.H 0 6H669 seam ampaaaaaeo

“000:0m

 

199

 

 

 

 

o.~mH o.mHm m.maa :.mma H.5ma m.m:a mm\ooma
m.wom m.eam o.mHH :.mma m.wma w.mma :m\moma
:.:>H m.:>H m.oaa o.HHH m.mHH w.mHH mm\mmma
m.m:a m.m:a m.:oa z.moa m.moa p.0HH mw\amma
m.m:H m.oza m.:oa m.moa H.mOH m.moa Hm\ommH
o.ooH o.ooa o.ooH o.ooH o.ooa o.ooa om\mmma
H.6oa o.moa a.mm m.mm m.om o.ee mm\omma
e.om m.om m.mm o.mm m.em o.em mm\emma
m.em H.@@ H.@@ m.mm m.mm o.mm em\omma
m.woa :.ooa >.mm m.mm n.om H.m> mm\mmma
m.aoa m.mm m.mm m.om m.mm m.a~ mm\:mma
o.mm m.~e o.mm m.me m.me o.mo am\mmma
m.Hm m.mo H.em m.ma w.ma e.mo mm\mmma

w0oapo m0ofipo m0oflpa m0ofieo m0oape m0ofiee

pcopmcoo 0:00030 ucmpmcoo 0:00050 pompmcoo 0:00050
Ga Ga ca Ga 0H CH 000%

pc0Epm0>cH 0Eoocfi 00H00o 00m 0Eoocfi H0cofipmz

 

.OOH u oo\amafiiiaaaaee ooxmmma peeppeoa ca 626 mmaaee
0:00030 CH pc0Epm0>cH 0:0 0Eoocfl 0pfiomo L00 .0EoocH H020Hpmc go m0oflocHII.m m mqm<e

200

 

 

 

 

.mm .6 ..moma ”oeamov somaimmma .meoamaaoeH
H00Humfip0pm .moflpmau0pm 0c0 coap0ufiawnoz ofiansm mo zoC0w< H0hpc0o ”0ohdom
moea mama mmma mHaH aoma mmma H6668
WIWMI WWI I©1Hm1 1....IWN11 Mlmlml INHIN1 mm0H>chm .HmQPO
moa mad ama mma maa ama mpemeao
0:0 00009
Hm me we we a» me wcamsom
mma aaa ama BAH moa ma 666666609
mm mm aw aw as ea eoapasepmcoo
2H: mom mam 0mm mom 00m moaoaepopam
000 00300005002
05: mm: om: mem mo: mos meaaaaaapma
. . . . . . . . . eoaaaae .0 .o . . . . . . . . . . . .
mo\aoma ao\moma mo\moma mo\aoma Hoxooma oo\mmma 6666660

 

.mmeaee ooxmmaa passages ea mo\aoma pp oo\mmma .mpppaam so agoaea Hmeoa66211.m m mamae

201

.m m 0H009 Eoem 0000H30H0o "0ohsom

 

 

H.5MH m.mmH m.mHH m.QOH H.ooH o.OOH H0668
mnmmm mummm 0.0HH H.00H mummm o.ooH 666H>eam paepo
o.mmH a.aHH a.aHH 0.8HH a.mHH o.OOH aoeaeHe
000 00009
o.HHH m.oOH m.oOH H.aOH a.HOH o.OOH weraom
a.moH m.amH m.omH m.mmH 8.60H o.ooH peoememse
a.aaH m.aom 0.05H m.emH o.mm o.OOH COHpeaeameoo
o.mmH m.maH m.omH m.mmH o.mHH o.OOH aeamaocH
000 00300003202
o.oHH m.HHH m.moH H.mm m.mm 0.00H meapHaaHewa
mo\eomH ao\momH moxmomH mo\HomH Ho\ooaH ooxmmmH 6660660

 

.Aooa u ow\mmmav molawma on mm\mmma .mLOpo0m an 000H0CH 0E00QH H0C0Hu0211.: m mqm<e

202

.0:.0 .hammH ..eeH .8030H o0aeaHm ".HHH .ooosmeomv
0:0:00m :0000000 0:0 :0 00:0:0 H0050oz00m 0:0 :03000 .0002 0H0:om "000000

 

 

0.: m.: :.: :.m m.: 0.m m.0 0.0 :.H m0oHIm:mH
m.m m.0 m.m m.w m.m m.HH m.:H :.m N.H m01m0ma
00 o0immmH
0.: 0.: m.H H.m 0.: H.0 0.0 m.0 0.m o0ImmmH
00 :mImmmH
m.m1 :.H 0.: m.m1 w.m w.m m.m1 :.m o.o :mmaiamma
0.0 o.m H.0 :.m m.: m.mH m.HH :.0 o.o HmmHim:mH
0 0 0 0 0 0 0 0 0
0500:H 0030000 m0ofi> 00:0:00 w:H 0000 :000 I MWWMH 00: do
H0:0000z H0:0000z 1000 0:0 Imsmm 16:009 10:00o:oo 0 0:0 m IHWW< 0000000
00000 00000 00:00 00008800 t r z00m:0:H

 

.0000000 0:0000000 :H 00:0:00800
00H 0:0 030000 000000000 00 00000 :0300w.0:500500 H05::0 0w000>¢11.m m mqm<e

 

203

 

 

.000 .0 ..mmmldmm “0002 ”000000
0.000 0.000 0.000 0.000 0.:00 0.::0 0.000 0.000 0.0:0 0000
0.000 0.000 0.000 :.000 0.0:0 0.0:0 0.0:0 0.0m0 0.000 0000
0.:00 0.000 0.m00 0.000 0.000 0.00 0.00 0.000 0.000 0000
0.000 0.000 0.m00 0.000 0.000 0.0:0 0.000 0.000 :.:00 0000
0.000 0.000 0.000 0.000 0.000 0.:m0 0.:00 :.000 0.0:0 0000
0.000 0.000 0.000 0.::0 0.000 0.000 0.000 :.:00 0.0:0 0000
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0000
0.000 0.000 0.00 0.:00 0.000 :.000 0.000 0.000 0.000 0000
0.000 0.000 0.000 0.000 0.000 :.000 0.000 0.000 0.000 0000
0.000 0.000 0.00 0.m00 0.000 0.000 0.000 0.000 0.000 :000
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0000
0.00 0.000 0.00 m.00 0.00 :.000 0.000 0.00 0.0:0 0000
:.00 0.000 0.00 0.00 0.00 0.000 0.000 0.00 0.000 0000
0.00 0.000 0.00 0.00 0.00 :.000 0.00 0.00 0.000 0000
0.:0 0.000 0.:0 0.00 0.00 :.000 0.000 m.00 0.000 0:00
0.00 0.000 0.00 0.00 0.00 0.000 0.000 0.:0 0.:00 0:00
0.:0 0.:0 0.0: 0.00 0.00 0.000 0.000 0.00 0.000 0000:0000
0000000000 0
0000000 00000 000000 a .
mmmme 008000 000000 000000000> 00000 000 r00000 0000000 000000 00000

 

.0000um0000 000000 000008800 00008 "000000 000000000000 00 0000800 x0000ll.0 0 00009

20“

 

‘1.

 

.qmma .mcsm .mm .02 mmm .Apazmmy OHHnsqmm pma< cmufica on mocmpmflmm¢

oHEocoom mmapo cum ow: 3mg oaansm .00H>nmm copmmmmm OHEocoom n<QmD

“mopsom

.mUOHan w30H>me Eopm umpmdnnm poz*

 

 

 

 

 

5mmgomm mom.:oa :om.oa mmm.mmm mpo.aa oow.mm Ham.om www.man Hmpoe
m::.m:a mmo.:m *qu mmm”:mfi mm mm:.mfi mamfim 5mm.moa mmma
mom.ama mam H mmm.oma HOH.N Hod.mfl mmm.ma :mmnmma mmmfl
mmmaHOH :moaza mam“fi mam.om .. mmo.ma u: mm:.>m Hmma
Ham.mm Hmm.o mmm.m mq:amm :Hm opfi.m mma mm>.o> coma
omm.mm oo~.m .. ommnmm mm mmm.: .. Hom.om mmma
mmm.ma mpm.oa I- mmo.a mm: omH.H .. I- mmmfl
Hmmnmfi omo.ma I- mom mzm mm: . -u .. Nmma
mmm.mm mmH.:H :mo.m :mfi.mm mam.~ HmN.HH .. pmo.pfi mmma
30>.Nm mpm.aa I- Hmm.ma mmm moo.ma .. I- mmmfi
. . . . . . . . . . . . mamafloc unmmsosp . . . . . . . . . . . . . .
mmfiamm mocmghso
mpmoaxm Hmpoe pmppwm mCOHpmcom mocmmpmem cwfimmom
amps» hmcuo Emmmopm om mcsw
uH50fipm< HH¢ zpfipsomm HHH mapfia HH mfipfle H mapfie mcfincm
Hmpoe manpsz mew

ow:

.A .m

 

Qmm< Umpflsb ms» on mpgoaxm amasuHSOprm

.mmmanmmma .mempmoma Umocmqfimnpcmscnm>om Umflwaomgm mmcc: oaanzamm
.m .2 map mo msfim> pmxgme pnoaxmuu.u m mqm<e

205

 

 

comm mwmm :mmm mfimm mmmm mzmm mmmm qum q<eoe
mm Fm om mu mp am mm mm mpmm new mHHo
mma OHH moa moa Hma mza mma mma mposuopa
mpflmc ucm xafiz
mm :H ma ma NH NH ma ma zmfim
: z z z m m m : mwmm
Fm om u: as am mm m: o: paw:
mHH mmH moa moa HHH Hm mp am pfisnm
H: mm 5m mm mm mm mm pm mmanmpmmm>
mad moa NOH NOH NHH NOH NNH :NH mass can
mommm .mmHSm
:5H :5H mod 35H Nod mma 53H mmfi mmfipmwsm
cam pawsm
0H om NH NH om ma am am uoog
anopmpm hmspo
dam mmOpwpom
wowa Gama Bmma mama woma mooa mmmfl :oma mammpmo
mm\mmma mm\:mma :mxmmma mm\mmma mm\HmmH Hm\omma omxmzma mzxmzma mmopsom

 

.:m\mmma on mz\m:ma Ammc pmm mmfipoamov mpaamo 9mg magnum voom pm2||.m m mqm<e

206

 

 

 

.mmESHo> Hapm>mm «goonmmmw

 

 

 

soaposuopm «macapmz cmpHCD map.mo coaummficmwpo HandeSoHaw¢ 6cm ooom "meadow

ommm ommm ommm ommm mmzm ommm 02mm mmmm 4¢Boe
55H mma moa mma moa mHH wma moa mpmm 6cm mHHo
mod moa Boa MOH moa OOH moa moa mposcopa
zpfimc cam xafiz
ma ma ma om ma ma mm om swam
m m m m m z m m wwwm
:m om mm b: m: mm mm mm ummz
00H mm paupm

wma HNH mma HHH mma mma

we a: mmanmpmwm>
moa mHH on mHH mm moa oma :m mus: vcm
mummm .mmasm
”NH wma Nma mqa NNH moa mod NNH mofipmmsm
new pawsm
om :m mm mm mm 0H mm mm @009
mcohwpm mmcpo
cam mmOpmuom
mmom Hmom Hmma mwufi mmma mmwa omma mmwa mammpmo
:©\mmma mm\mmma mm\amma Hm\ommH oo\mmma mm\mmma mm\wmma mm\mmma mmopsom

 

 

UmSQHpCOOII.w m mqm<B

207

TABLE B 9.-—Estimated coefficients from the double
logarithmic model for the commodity groups in urban areas.

 

 

—2
Dependent B B B R S
variable 0 l 2

X1 ' .7878 .1035 -.1035 .9992 .0142
(.0236) .0143) (.0245)

X2 -.5798 .1752 .1113 .9877 .0626
(.1042) .0631) (.1095)

X3 —.l767 .0199 .0798 .9979 .0333
(0555) .0336) (.0583)

X“ -.7719 .4387 .0483 .9947 .0423
(0704) .0426) (.0735)

x5 -1.4676 .9409 .2044 .9938 .0596
(.0993) .0601) (.1049)

X6 -.9985 .6845 .0795 .9974 .0325
(.0542) .0328) (.0566)

X7 -2.3997 .2395 .2292 .9940 .0655
(.1091) .0660) (.1144)

X8 —.8970 .6323 .0464 .9973 .0457
(.0762) .0461) (.0800)

X9 —1.1226 .7120 .0778 .9940 .0503
(.0838) .0507) (.0877)

X10 -1.4648 .7318 .3787 .9918 .0721
(.1200) .0702) (.1261)

Xll .0200 .3545 -.2043 .9916 .0406
(0677) .0401) (.0714)

Xl2 .3754 .6331 .0229 .9996 .0117
(.0195) .0118) (.0200)

 

TABLE B lO.--Estimated coefficients from the double

logarithmic model for the commodity groups in rural areas.

 

 

Dependent -2
variable B0 B1 B2 R S

X1 .3722 .4081 -.0061 .9979 .0170
(.0433) ( 0229) (.0360)

x2 —.9584 .5790 .0304 .9984 .0283
(.0725) (.0382) (.0592)

x3 -.9574 .9745 .0194 .9971 .0228
(.0584) (.0308) (.0480)

X“ —.6570 .2912 -.0205 .9828 .0476

~ (.1217) (.0642) (.0995)

x5 —1.0159 .8194 .1086 .9843 .0555
(.1419) (.0749) (.1157)

x6 —.5208 .5292 -.1172 .9934 .0275
(.0704) (.0372) (.0574)

x7 -2.3890 1.2900 .1175 .9965 .0293
(.0750) (.0396) (.616)

x8 —.8658 .7206 —.0397 .9931 .0317
(.0811) (.0428) (.0663)

x9 -1.0035 .7005 .0058 .9880 .0434
(.1109) (.1343) (.2078)

x10 —2.4433 1.2866 .4309 .9884 .0994
(.2543) (.1343) (.2078)

Xll .0552 .5455 —.0808 .9124 .1011
(.2587) (.1365) (.2114)

x12 .3155 .6887 .0130 .9993 .0106
(.0271) (.0143) (.0244)

 

209

TABLE B ll.--Estimated coefficients from the double loga-

 

 

rithmic model for individual commodities in urban areas.
Dependent —2
variable B0 B1 B2 R S
Xl3 -2.2l46 1.1345 .3543 .9267 .2430
(.4048) (.2450) .4253)
X13 -2.3834 1.0752 .3138 .5756 .4610
(.7677) (.4646) .8066)
X14 .8955 -.2776 .5914 .8183 .2973
(.4951) (.2996) .5195)
Xi“ -.6322 -.2725 .4483 .8444 .2798
. (.4660) (.2821) .4898)
X15 .4192 .0780 .3885 .8955 .2292
(.3818) (.2311) .4012)
X15 -1.1055 .0838 .4239 .9239 .2137
(.3560) (.2155) .3741)
X16 —1.0474 .5806 .3602 .9524 .1674
(.2788) (.1687) .2931)
16 -l.2485 .5270 .3352 .7432 .2342
(.3900) (.2360) .4098)
Xl7 1.6871 -.5748 .2412 .8656 .2106
(.3508) (.2123) .3687)
Xi7 .3656 —.6228 .1803 .8925 .1759
(.2931) (.1774) .3080)
X18 1.1930 -.2318 .2667 .9071 .1839
(.3063) (.1854) .3209)
18 —.l640 -.2519 .1958 .9229 .1561
(.2599) (.1573) .3731)
X19 .2006 —.2219 .2179 .8600 .2218
(.3694) (.2236) .3873)
X19 —.O4O3A -.2443 .5026 .4719 .2101
(.3500) (.2118) .3674)

TABLE B ll.--Continued

210

 

 

Dependent —2
variable B0 B1 B2 R S

X2O 3.7380 -1.2109 -.5087 .7349 .2061
(.3433) (.2078) .3605)

x50 2.0969 -1.1512 .5163 .7021 .2125
(.3539) (.2142) .3715)

x21 3.3379 -.9427 .4733 .8202 .1371
(.2284) (.1382) .2402)

X51 1.7029 -.8900 -.4687 .7907 .1444
(.2405) (.1455) .2527)

x22 2.2686 —.6661 .2786 .8693 .1519
(.3651) (.1535) .2362)

XéZ .5541 -.5201 .3053 .8872 .1375
(.3306) (.1390) .2139)

x23 1.4243 .1961 .0699 .9913 .0268
(.0643) (.0270) .0416)

X53 .4243 .1961 .0699 .9913 .0268
(.0643) (.0270) .0416)

x2“ .6162 .3308 .1137 .9827 .0468
(.1125) (.0473) .0728)

Xéu -.0811 .4042_ .0911 .9797 .0513
(.1232) (.0518) .0797)

X25 1.4020 .8457 .3443 .9747 .0800
(.1923) (.0808) .1244)

X55 2.9501 .9101 .4433 .9775 .0808
(.1943) (.0817) .1257)

x26 1.0841 -.1526 .0382 .9379 .0809
(.1706) (.0717) .1104)

X56 -.5303 -.0205 .0140 .9410 .0734
(.1765) (.0742) .1142)

X27 1.0403 5.1181 .1013 .9837 .0535
(.0891) (.0539) .0766)

TABLE B ll.—-Continued

211

 

Dependent

 

variable B0 B1 B2 R S

x57 -.2690 -.0716 -.0511 .9735 .0422
(.1086) (.0457) (.0703)

x28 .2085 .8160 -.0656 .9982 .0165
(.0396) (.0166) (.0256)

X58 1.0276 .9358 -.0296 .9968 .0243
(.0585) (.0246) (.0378)

X29 -2.8737 1.6439 .2201 .9754 .1030
(.2475) (.1041) (.1602)

X59 -3.7352 1.7301 .2552 .9772 .1035
(.1359) (.0571) (.0879)

Xéo -1.7505 .7636 .1956 .9800 .0634
(.1525) (.0641) (.0987)

X31 .5359 1.1415 .1118 .9879 .0562
(.1352) (.0568) (.0874)

Xél -2.9861 1.2351 .1873 .9878 .0604
(.1452) (.0610) (.0939)

X32 .3944 .2081 .0490 .9869 .0616
(.1025) (.0621) (.0881)

Xé2 -.7355 .2888 .0004 .9773 .0480
(.1155) (.0485) (.0747)

X33 -.9022 .5361 .1038 .9835 .0464
(.1116) (.0469) (.0722)

Xé3 1.6296 .5876 .0756 .9638 .0737
(.1771) (.0745) (.1146)

X3“ -.5075 .8357 .1973 .9907 .0445
(.1070) (.0450) (.0692)

X34 —2.8620 .9897 .6218 .9692 .1704
(.2838) (.9897) (.6322)

x35 —2.3143 1.1327 .3074 .9773 .0833
(.2003) (.0842) (.1296)

x§5 -3.1358 1.1884 .3431 .9789 .0831
(.1998) (.0840) (.1293)

TABLE B 11.--Continued

 

212

 

Dependent B

 

variable 0 B1 B2 R S

x36 2.3692 —.6336 -.3965 .8153 .1121
(.2694) (.1133) .1743)

x§6 .8636 -.5368 .3597 .7801 .1184
(.2847) (.1197) 1842)

x37 —1.1710 .8104 .1444 .9276 .1037
(.2492) (.1048) .1612)

X§7 -1.9578 .9268 .0679 .9643 .0800
(.1922) (.0808) .1244)

X38 -1.5050 .9573 .1254 .9846 .0583
(.1402) (.0589) .0904)

x§8 -1.9735 1.0141 .1434 .9852 .0592
(.1424) (.0598) .0921)

x39 .0247 .4852 .1202 .9867 .0436
(.1048) (.0441) .0678)

ng -1.7126 .5768 .1272 .9870 .0447
(.1075) (.0452) .0695)

X40 1.1082 .1506 .1549 .9662 .0479
(.1151) (.0484) .0745)
(.1227) (.0516) .0794)

X41 .7886 .3734 .0461 .9945 .0233
(.0560) (.0234) .0362)

X51 -1.1826 .5452 .0376 .9874 .0547
(.0474) (.0199) .0307)

x42 —1.0111 1.0176 .1427 .9874 .0547
(.1314) (.0553) .0850)

x52 -2.8143 1.2162 .1348 .9857 .0636
(.1528) (.0642) .0989)

X43 .2808 .3024 .0405 .9822 .0412
(.0991) (.0417) .0641)

x53 -1.5177 .4492 .0275 .9875 .0388
(.0933) (.0392) .0604)

TABLE B ll.--Continued

213

 

Dependent -2
variable B0 B1 B2 R S
X4“ 8065 .2412 -.Ol49 .9747 .0493
(.1184) (.0498) (.0766)
X44 4304 .3963 -.0763 .9927 .0266
(.0638) (.0268) (.0413)
X45 .1779 .0248 -.l226 .9706 .0370
(.0889) (.0374) (.0575)
X55 .0303 .0973 -.0894 .9834 .0352
(.0847) (.0356) (.0549)
X46 -3.8690 1.6002 .2289 .9698 .1129
(.2713) (.1141) (.1756)
X46 -3.8486 1.5889 .1841 .9747 .1012
(.2433) (.1023) (.1574) '

 

TABLE B l2.-—Estimated coefficients from the double loga-
rithmic model for individual commodities in rural areas.

 

Dependent

—2

 

variable B0 B1 B2 R S

X13 -3.7228 2.4705 .7705 .9735 .1382
(.3537) (.1866) (.2893)

X13 -5.4697 2.5538 .8489 .9668 .1612
(.4125) (.2177) (.3377)

X1“ 2.4274 -.6318 .1169 .8939 .1422
(.3637) (.1919) (.2974)

Xi” .5972 -.5076 .2152 .9228 .1275
(.3263) (.1722) (.2668)

x15 .2954 .9714 .0811 .9967 .1275
(.0651) (.0344) (.0529)

X15 -1.4665 1.0387 .1881 .9969 .0269
(.0689) (.0363) (.0566)

X16 -3.0482 1.6008 1.1856 .9872 .0984
(.2519) (.1329) (.2059)

Xi6 -4.7490 1.6401 1.2545 .9833 .1158
(.2963) (.1564) (.2425)

Xl7 3.5509 -.9823 —.1754 .9743 .0576
(.1474) (.0778) (.1204)

X17 1.8802 -.9274 —.1314 .9760 .0567
(.1451) (.0766) (.1187)

x18 1.9301 .0224 .0287 .9716 .0629
(.1608) (.0849) (.1315)

X18 .3853 .0226 .0845 .9714 .0663
(.1606) (.0895) (.1386)

x19 -1.8419 1.5921 .8278 .9245 .1687
(.4316) (.2277) (.3535)

x19 -3.4361 1.6440 .2915 .9276 .1639
(.4193) (.2213) (.3434)

x20 3.8400 —1.2087 -.3280 .7771 .1812
(.4637) (.2447) (.3807)

TABLE B 12.--ggn§inueg

215

 

Dependent

 

variable B0 B1 B2 R S
Xéo 2.0489 -1.1174 -.2780 .7797 .1793
(.4588) (.2421) (.3753)
>51 2.0489 .0608 -.2621 .8285 .1183
(.3028) (.1598) (.2476)
Xél .2981 .1303 -.2216 .8394 .1208
(.3091) (.1631) (.2527)
x22 3.2859 —2.3752 .8796 .9598 .1534
(.6711) (.2373) (.2757)
XéZ 1.7871 -2.3883 .9357 .9574 .1605
(.7018) (.2482) (.2884)
x23 3.0114 —1.2429 -.3882 .6315 .2437
(1.0658) (.3769) (.4379)
x53 2.0114 —1.2429 -.3882 .6315 .2437
(1.0658) (.3769) (.3479)
X214 -.6701 .7311 .5449 .9776 .0674
(.2946) (.1042) (.1210)
X24 -2.5937 1.1297 .4633 .9498 .0744
(.3256) (1.1297) (.3337)
x25 -5.0560 2.3588 1.7077 .9456 .1238
(.5415) (.1915) (.2225)
Xés —7.3296 2.6782 .9926 .9532 .1309
(.5724) (.2024) (.2352) -
x26 -.3368 .5275 .2126 .8993 .0811.
(.3548) (.1255) (.1458) ‘
x§6 —1.7410 .6150 .1821 .9759 .0380
(.1660) (.6150) (.1825)
x27 .6104 .2722 —.2112 .9838 .0287
(.1257) (.0445) (.0517)
x57 -1.0794 .4196 -.0290 .9942 .0149
(.0654) (.0231) (.0269)

TABLE B l2.--Cont1nued

216

 

 

Dependent —2

variable B0 Bl B2 R 3

x28 -.0925 .7464 -.0315 .9940 .0168
(.0734) (.0259) (.0301)

X28 -.6684 .7574 -.0386 .9954 .0262
(.0671) (.0354) (.0419)

X29 -3.5039 1.8360 .5593 .9802 .0603
(.02636) 9.0932) (.1083)

x29 -4.2374 1.9251 .5137 .9883 .0473
(.2068) (.0731) (.0850)

x30 -1.3979 .8128 .3782 .9681 .0520
(.2274) (.0804) (.0934)

Xéo —3.0084 1.1527 .4271 .0656 .0604
(.2642 (.0934) (.1086)

x31 -.4255 1.2966 -.0605 .9938 .0224
(.0982) (.0347) (.0403)

X31 —2.8609 1.3298 -.0173 .9951 .0208
(«0912) (-0323) (.0375)

x32 .2103 .1633 -.0551 .9822 .0287
(.1255) (.0444) (.0516)

XéZ -.9500 .3532 0.9418 .9874 .0217
(.0950) (.0336) (~0390)

x33 -1.7284 .7643 -.0298 .8364 .0956
(.4182) (.1479) (.1718)

x§3 -2.3409 .7245 -.0123 .8739 .0810
(.3544) (.1253) (.1456)

x3“ -3.6501 1.7240 -1.2917 .9643 .0969
(.4239) (.1499) (.1742)

X34 -5.1805 1.7463 1.3478 .9587 .1068
(.4670) (.1652) (.1919)

x35 -3.7268 1.6163 .3077 .8414 .1566
(.6848 (.2422) (.2814)

TABLE B 12.--Continued

217

 

 

Dependent —2
variable B0 B1 B2 R S

x§5 -4.5824 1.9283 .0154 .9361 .1042
(.4557) (.1612) (.1873)

x36 2.0800 -.7274 -.0070 .9810 .0422
(.1846) (.0653) (.0758)

X36 .5264 -.5381 -.0970 .9687 .0460
(.2011) (.0711) (.0826)

x37 .3965 -.0644 -.1008 .6170 .1464
(.6403) (.2265) (.2631)

Xé7 -.2864 .0591 -.1109 .5993 .1427
(.6239) (.2206) (.2564)

x38 -1.2497 .9479 -.0023 .9776 .0364
(.1592) (.0563) (.0063)

x§8 -1.6580 .9882 -.0176 .9765 .0378
(.1655) (.0585) (.0680)

x39 4.9102 .8988 .2430 .9644 .0519
(.2270) (.0803) (.0933)

x§9 -2.6580 1.0234 .2148 .9766 .0431
(.1882) (.0665) (.0773)

Xuo 1.2772 .0298 -.1387 .9921 .0162
(.0709) (.0251) (.0291)

X50 -.4917 .1263 -.1997 .9836 .0211
(.0923) (.0326) (.0379)

Xul .8070 .2183 .0015 .9810 .0282
(.1233) (.0436) (.0506)

x51 -1.1046 —.3633 .0050 .9791 .0295
(.1290) (.0456) (.0530)

qu —1.9499 1.4752 .1641 .9896 .0341
(.1492) (.0528) (.0613)

Xfi2 -3.5721 1.5420 .1715 .9894 .0355
(.1552) (.0549) (.0638)

x43 -.0064 .4950 .0411 .9531 .0467
(.2043) (.0722) (.0839)

TABLE B 12.--Continued

_1

218

 

Dependent —2
variable B0 B1 B2 R 3
x53 -2.0658 .7327 .1458 .9740 .0396
(.1731) (.0612) .0711)
X44 .8917 .4079 .2130 .9152 .0487
(.2129) (.0753) .0875)
x5“ -.3242 .5012 .3030 .9088 .0527
(.2305) (.0815) .0947) .
XLl5 -.2385 .2864 .1313 .8788 .0639
(.2796) (.0989) .1149)
x55 -.3083 .3286 .1028 .9130 .0551
(.2412) (.0853) .0991)
x46 —4.0971 2.0423 .0206 .9459 .1003
(.4385) (.1551) .1809)
x56 —4.6442 2.1692 .1250 .9437 .1094
(.4786) (.1693) .1967)

 

AOHOmaav

 

 

2:19

 

 

.AmsmuonHx ch mm\wmmHnnmmmLm coop: ca

  
  

Hood Hvao< .opHmp .conmmlcmHAmmmm on» :« mowpsm powvzm UHocomsox .moupHEEoo HmoHumHuMum Hmpucmo m<a ”sopu UmumHsono "monsom
mm m: HMH sOm ham OON Omm mHm mOO Hmm OOH mo O mOHocmmsoc co 966532
mm.m mm.m sm.m mm.m om.: o:.z mH.: ms.m mz.m ms.m mo.m Hm.H mm. pHoc
omson\muH:: :oHuQESm
1coo go ambasc mm¢pm><
mH.~:mH om.H~m mm.amm 33.nm: mm.:zm om.:>m mm.:mm mo.m~H mm.mmH Hm.mm om.mo >m.o: ww.wH A.m.qv mmLSpHocoaxm
, Hmsccm Hmpou mwwnm><
quemw who pzonm :oHumELomcH
m.m m.o o.m m.m o.m H.H o.H o. a. m. H. H. .. mouuoo
m.m m.O m.m O.m ~.: O.a 3.: O.m m.m O.m O.m O.H m. «we
m.O:H m.moH m.HoH s.mm m.m~ m.m~ m.om o.ms m.:: m.om m.mm m.mH o.» pmwjm
m.mm m.mm :.mm H.2m o.Hm :.sm m.mm :.mH ~.mH m.HH s.m o.o H.m mmumo
O.mmm 0.0~H 0.0mH 0.00H m.ma H.mO 0.00 m.mm O.mm m.:H m.O m.O H.m waspHO
A.Omm m.OOH m.m~H m.OOH :.HmH H.OOH 0.00 o.oa m.:O m.Oa m.Om H.Om m.OH mmOpmsoe
m.moH m.zm m.ms O.~w m.mw o.H: w.mm m.m: H.o: :.mm o.om o.mH H.~ mCOHco
m.maH O.HHH m.moH m.mw m.ms m.m: :.mm o.o: m.mm m.mm m.OH H.m o.H mmOmeom
m.:m m.mm m.mm m.=m m.~H m.oH m.HH ~.m o.o m.m :.m o.H .. mono
m.sH o.mH m.HH m.m m.o H.m m.a m.m m.H :.H m.H m.H m. Louusm
m.:H m.m o.m w.HH m.mH m.m m.mH m.mH :.:H m.oH H.mH m.mH m.m mmmoco mocz
m.~m m.om m.~m m.mH :.mH :.w m.w :.m s.m m.m m.H m. .. ommono muHc3
m.m>m o.omm H.22m H.mmH m.HaH o.mm 0.:0 m.z~ m.Hm m.Hm m.om m.HH m. xHHz
m.mH H.~H m.mH m.MH o.o o.» o.~ m.m 0.: m.m H.m m.H m. ”HHo omumusumm
~.:m o.om 0.x: o.mm H.5m 0.36 m.Hm m.sm s.mm m.~H m.mH m.~ m.m mHHo
O.OOHH O.HmO 0.0m~ O.HOO O.sHm m.OOm A.OOm m.OHH O.OO 0.0m O.mO 3.5m m.HH Amaonsaz cHO mmwm
m.oO m.qm 5.3m o.m: m.m: nwsm m.cm m.om m.mm c.1H H.m H.o o.m cmHm cmmpm
m.om m.zs w.mm o.sm o.mm ~.wH m.mH m.~ 0.: m.H :.H o.H m. appHsom
m.:mm o.moH o.:mH H.m0H ~.z~ m.Hw o.om s.mm m.mm m.om :.wH ~.HH m.m umoz
o.~m m.mH o.mH m.Hm m.Hm s.mH m.oH m.mH m.mH N.MH :.HH m.w m.m mHHpcmq
o.am z.mH m.am m.mm o.om m.nm w.mm H.Hm m.mH 9.0H m.mH :.m o.m mcmom
o.Hs O.H© m.Ho o.o: 0.5m H.Om m.:m m.~H H.mH m.o 5.: H.m .. «cosmomz
m.mom H.mom H.m~H ~.mmH m.HmH m.mmH m.mHH o.mm m.mm m.Hm m.:m m.Hm m.n oOHm
m.smm m.mmm m.msm m.©mm s.~Hm m.~OH m.owH m.smH H.o:H o.nHH m.mm m.mm m.>m ommbm
m.mm o.moH H.m0H m.mmH m.m:m m.msm ~.mmm ~.oow :.mOH H.0HH :.mm m.mm m.mm L30Hm
m.HH m.O O.m m.a m.am O.mH O.mH a.mm O.sm H.m: H.mm O.:: A.Oa Ommmaoaaa
m.: m.O O.m m.O a.m N.HH m.m O.m O.a H.s m.m m.m .. Omoacoaaumeoc "uoHHH:
~.ma :.Om s.m: m.Om O.sa A.OO O.:m m.Om o.om O.mO O.:: m.am m.m 66mmzoasa
O.Hm :.mm m.:m H.m: F.3m o.mm o.mH m.mH m.@ m.: m.m H.: .. noosuopanmso: "mNHmz
:.mMH H.HH H.m~ w.mw m.mo m.mm m.mm o.om m.m: :.m: H.o: m.:m o.m nmmmcopza
m.om m.mm o.~H m.ma o.Hm m.mm =.©H m.zH H.m :.H m.H o. .. voodoopaumEo: "amen:
Lo>o QOOH com 000 00: com omm oom omH OOH mu om mm mmfipaooesoo
can» can» cmcu can» cm:» can» can» cmcu can» can» can» . .
mmMH wmoH mmoH mmmH mmmH muoH mme mme mmmH wmmH mmmH mmmH ”mug A mmwwnmasmmwx
66 OOO 6» OOO 6» OOO 6» OOm 6» omm 6» OOm 6» OOH as OOH 6» ms 6» om 6» mm 4 HO m

Hmscc< kuoe

HmohzuHccodxo Have» cH ommmmuaxmv Hm>mH mEoocH on pwumHop mm moospopd ooow poOmE mo vHonomson you coHpassmcoo Hascc<nl.MH m mqm<9

.MH m mHDwB 00m "OOLSOW

 

m : mH ow OmH :mH Ozm ma: OOO m2: OOO mfim m: mcHocmmson Mo 609232
o:.HH oo.:H mm.m mo.m OH.O wo.w w:.m 20.: O~.m mo.m m:.m mm.H Hm. mnaocmmson
. , mom made: coHuQESm
Icoo mo amass: ommho><
O0.000H O0.000 HO.AO~ O0.0AO O0.0mm NO.Hsm OO.Hmm OO.maH OO.mmH O0.00 OH.OO O0.0m OO.ON H.O.OO mapspHucmaxm
szccm kuou mwmuo><
mHQEmm on» psoom coHumELomcH

 

2230

 

 
 

O.OH ~.m 0.0 O.m O.H O.H O. O. m. m. H. H. .. OOOOOO
O.Hm O.OH O.O O.O O.O 0.0 O.: m.: O.m O.m H.O O.H O. «we
O.OON O.H:m O.OOH O.OOH O.OOH O.OO O.Os O.OO O.OO O.OO O.Om O.Hm O.HH amwam
O.OOH O.OOH 0.0HH H.OO 0.00 O.OO O.OO O.Hm O.OO O.OO O.OH m.O O.m mapmo
0.00m 0.00H O.HOH O.Oa m.mm 0.0m O.OO 0.0m 0.0H 0.0 0.0 O.m O.H OOLOHO
O.OOO O.OOH O.OOH H.HmH H.OOH 0.00 O.OO O.OO O.HO 0.0m O.Om O.Hm O.OH OOOOOEOB
O.OOH O.OOH O.OOH O.OO O.Os O.OO O.OO O.OO O.OO 0.0m H.Om O.HO O.OH mcoHcO
m.mHm 0.00H m.OsH 0.0HH 0.00 0.00 0.0m 0.00 O.Hm m.Om m.mH m.O m.: mmouauom
O.OO O.OH 0.0m O.OH 0.0H O.OH O.O O.O O.m O.: H.O O.H O. 66cc
0.0m 0.0 O.m m.OH o.o a.» H.O m.m m.m 0.0 m.m O.m m.H pmuusm
O.OO O.OO 0.00 0.00 O.OO 0.0a 0.00 0.0m 0.0m 0.0m 0.00 0.0H 0.0 666626 Own:
O.OH .. m.O m.O O.m m.m O.H O.H a. O. m. H. H. mmmmno mOan
O.OOO O.OOOH O.OOO O.HOO :.OOm 0.00N 0.000 O.OOH O.OHH O.OO O.Om O.m m.m xHH:
O.OH 0.0 O.m O.m O.m O.H O.H O.H O.H O. O. O. m. mHHo oaamaspam
O.H: O.OO O.OO O.Hm H.Om O.OO O.Om O.OH O.OH O.O O.A 0.0 O.m mHHO
O.OOHN O.OOO O.HHO 0.0ma O.HOO H.Omm O.OOH O.HOH O.OO H.OA O.OO O.Om H.OH Hmpmneac cHO ammm
O.OO O.OO O.OO H.H: 0.0m O.Om O.OO O.OH O.OH O.O 0.0 H.O O.H cmHO nmmaa
O.OOH O.OO O.OO O.OO O.Hm O.Om H.HO O.OH O.O 5.: s.m H.H O. OapHsom
O.OOm O.HOH O.OOH O.OHH O.OO O.OO O.OO O.Oa O.Om O.Om 0.0H O.OH 0.0 One:
O.OO 0.0s O.OO 3.5m 0.0m 0.0m O.OO O.OH 0.0H O.OH O.HH 3.5 O.= OHHOcOO
O.HOH O.OOH O.OOH 0.00 O.HO O.OO O.OO 0.0m m.~m O.OH O.OH O.O O.m mamom
0.0s 0.0m 5.3: O.Hm 0.0H O.OH A.» O.m m.m O.H O. m. H. Hcoumom:
O.OO: O.OOO O.AOO O.Hmm O.OOO O.Omm H.OOO O.OOH O.OHH 0.00 O.H: O.Om 0.0H OOHO
O.O O.OO O.OH O.Om O.OH 0.0 O.O O.O O.a H.O O.OH :.HH O.O ommam
O.m O.HMH 0.0H 0.00 0.00 H.OO 0.00 0.00 O.H» 0.00 H.Om 0.0m 0.0m uonm
.. .. 0.0s 0.00 0.0s 0.0m :.OO 0.00 O.OOH 0.0mH O.ONH O.HO. 0.00 OOOOOOOOQ

O.OOO O.OOO 0.0HO :.AOH O.OOH O.OOH 0.0HH O.OO 0.0a 0.00 O.Om O.s O.m Omoaooaaumsos "umHHHz
0.0mm O.OOO O.Omm O.OOO O.Hmm O.OHO 0.0am O.OON O.OHO H.OHN O.ONH O.OmH H.HO Ommacoaaa

O.OOO O.OOO O.HAO O.OOO A.AOO O.HHO 0.0am O.Oam 0.00H 0.0s 0.00 0.0H O. omoanoaanmson "mNHm:
O.Oam O.OOH H.HOO O.OOO :.OOO 0.0mm O.OO 5.23m 0.00H 0.00H H.OHH O.OO O.OO Ommmnoaaa

:.mmOm O.MOAH O.OOOH O.OHO O.OHO :.AOO O.Omm O.OHO O.OOH O.Om O.Om O.O O.H cooaooaaamsoc "Oman:
66>o OOOH OOO OOO OOO OOO OOO OOO OOH OOH ON OO Om amHuHOossoo
DEM can» Ems» cm:» can» :62» can» can» Cmnu can» can» can» can» A.m.dv mQSOLG

OOOH mme mmom mmoH meH wme mmma mmmd mmmd mmwd mama mmmd mmvd mmhzudccwaxm

0» com 0» coo 0» co: 0» oom 0» 0mm ow oom 0» omH 0» OOH 0» mu 0» om 0» mm Hmscc< Haves

.HmemponHx :HV mm\mmmH .mmopm Hausa CH
HampsuHucmaxm Hmuou cH ummmmpqxov Hm>oH oEoocH o» coumHop mm muosvopa ooou gonna go oHonmmsoz hog coaua55mcoo Hmscc<nu.=H m mqm<9