THE IMPLICATIONS OF RECENT ECONOMIC AND POLICY.
CHANGES ON RETAIL AND FARM LEVEL DEMAND
FOR FOOD COMMOOIIIEs

Dissertation for the Degree Of Ph. D.
MICHIGAN STATE UNIVERSITY
SHIRLEY ANN PRYOR
1977

  

1’ Ll 1?!qu K3"?

.. 1.; ,
3‘, .intéan Stat:~ .

mecrsiqy (

 
 
 

  
 
 

ar‘fil"

 

This is to certify that the

thesis entitled

The Implications of Recent
Economic and Policy Changes
on Retail and Farm Level Demand

for Food Commodities
presented by

Shirley Ann Pryor

has been accepted towards fulfillment
of the requirements for

Agricultural Economics

 

Ph ' D ' degree in

Major professor

DateW

07 639

 

ABSTRACT

THE IMPLICATIONS OF RECENT ECONOMIC AND POLICY
CHANGES ON RETAIL AND FARM LEVEL DEMAND
FOR FOOD COMMODITIES

by

Shirley Ann Pryor

The major objectives of this work are to analyze some specific
United States farm commodities and isolate and estimate the relevant
variables affecting their consumption. The purpose is to evaluate the
substitutes and complements for the particular retail products as well
as the effect of income and specific marketing costs. The effect of
income dependent transfer payments and food stamps on demand is also
analyzed.

Econometric analysis is used to determine the relationship
among the variables. Unlike previous retail analysis covering many re-
tail commodities, no attempt is made to account for all foods using
the usual assumptions about elasticities. Instead, a system was de-
veloped for-each farm commodity. Retail and farm-level data are used;
the retail level data being considered especially relevant for analysis
involving consumer choices.

All data used is annual data published by the U. S. Department
of Agriculture and the Bureau of Labor Statistics.

Marketing margins are not estimated directly, rather farm-retail

price differentials emerge from the estimates of the price at both the

farm and retail levels.

Shirley Ann Pryor

In general farm-retail price differentials have declined or
stayed the same over the period in question, 1952—1972, except for bread,
flour packaged for retail sale and fluid milk. The percentage of the
retail price going to the farmer declined in all cases. The wage of
food manufacturing employees and the price of fuel has little impact
on these price differentials.

Food stamps and income dependent transfer payments have differ-
ential effects on the various commodities and thus on the various pro-
ducers, processors, retailers, and consumers of those commodities.
Income-dependent transfer payments have an effect on commodities in the
following order: fed-beef, non-fed beef, flour, broilers, eggs, cookies,
cheese, fluid milk, ham and butter; the largest impact being on beef.

Ham and butter receive a negative impact.

THE IMPLICATIONS OF RECENT ECONOMIC AND POLICY
CHANGES ON RETAIL AND FARM LEVEL DEMAND
FOR FOOD COMMODITIES

by

Shirley Ann Pryor

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

1977

Dedicated to my mother and father

ii

 

 

 

 

ACKNOWLEDGMENTS

I wish to thank Dr. Vernon Sorenson for his encouragement and
guidance throughout my academic program. I am also indebted to Dr. Glenn
Johnson, Dr. Lester Manderscheid, and Dr. James Johnson for their help
at various stages of my course and dissertation work.

Special thanks are due to the Department of Agricultural Eco-
nomics and the taxpayers of the United States for supporting me through-
out my stay at Michigan State University.

All the credit for any accomplishments and none of the blame for
any failures goes to my friends and most of all to my mother and father.
My father showed me the path. My mother widened it for me and pushed
me along it. "You don't have to help me with the dishes tonight, dear,

do your homework." I am still doing my homework.

iii

TABLE OF CONTENTS

LIST OF TABLES . C O O O O O O O O I O O O O O O 0 O O O O O O 0

LIST OF FIGURES . . . . . . . . . . . . . . .

Chapter

I.

II.

III.

INTRODUCTION AND OBJECTIVES . . . . . . . . . .

The Role of Demand . . . . . . . . . . . . . .
Objectives . . . . . . . . . . . . . . . . . . . . .
Method of Analysis . . . . . . . . . . . . . . . . .
Organization of Thesis . . . . . . . . . . .

FRAMEWORK FOR ANALYSIS . . . . . . . . . . . . . . . .

Demand Theory . . . . . . . . . . . . . . . . .
Framework I O O O O O O O O O O I O O O O O O O O
The Interrelationships among Income Distribution,

Income Redistribution Policies and Demand for Food .

Commodity Specific Structures . . . . . . . . . .
Estimation Procedure . . . . . . . . . . .

STATISTICAL ANALYSIS . . . . . . . . . . . . . . .

7 Outline for Presenting the Statistical Analysis . .

Wheat . . . . . . . . . . . . . . . . . . . . .
Modeling the Wheat Sector . . . . . . . . . .
Data. Correlation Matrix, Sources of Data and

Adequacy of Proxies . . . . . . . . . . . . .
Analysis of Regression Results . . . . . . . . . .

Milk. . . . . . . . . . . . . . . . . . . . . . . .
Modeling the Dairy Sector . . . . . . . . . . . .
Data: Correlation Matrix, Sources of Data and

Adequacy of Proxies . . . . . . . . . . . . . .
Analysis of Regression Results . . . . . . . . . .

Fed and Non-Fed Beef . . . . . . . . . . . . . . . .
Modeling the Beef Sectors . . . . . . . . . . . .
Data: Correlation Matrix, Sources of Data and

Adequacy of Proxies . . . . . . . . . . . . .
Analysis of Regression Results . . . . . . . . . .

Hogs . . . . . . . . . . . . . . . . . . . . . . . .
Modeling the Hog Sector . . . . . . . . . . . . .
Analysis of Regression Results . . . . . . . . .

iv

Page

vi

. viii

14
22
23

25

25
26
26

27
29
52
52

56
60
82
82

85
86
100
100
100

Chapter Page

ChiCken O O O O O O O O I O O O O O O O O O O O O 112
Modeling the Broiler Sector . . . . . . . . . . . 112
Analysis of Regression Results . . . . . . . . . 121

Eggs 0 O O O O O O O O O O O O O O O O O O O O 122
Modeling the Egg Sector . . . . . . . . . . . . 122
Analysis of Regression Results . . . . . . . . . 130
IV. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . 132
APPENDIX A
Description of Transfer Programs . . . . . . . . . . . 139
APPENDIX B

Variables used in Statistical Analysis . . . . . . . . 141

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . 145

LIST OF TABLES

Table Page
2.1 U. 8.: Transfer Programs (1950-1974) . . . . . . . . . . 16

2.2 U. S.: Benefit Outlays Under Public Income Transfer
Programs, Fiscal Year 1972 . . . . . . . . . . . . . . 17

2.3 U. 8.: Federal Food Stamp Program (1961-1976) . . . . . 18
3.1 The Wheat Model . . . . . . . . . . . . . . . . . . . . . 28

3.2 Simple Correlation Matrix for Variables Used in the

Wheat Equations . . . . . . . . . . . . . . . . . . . . 30
3.3 Regression Results for Wheat . . . . . . . . . . . . . . 32
3.4 Income and Food Stamps Elasticities for Wheat Products . 48

3.5 Wheat: Farm-Retail Spreads and Percent of Retail Price
Going to Farmer (1950-1970) . .A. . . . . . . . . . . . 51

3.6 U. 8. Consumption of Milk and Dairy Products (1950-1970) 57
3.7 The Dairy Model . . . . . . . . . . . . . . . . . . . . . 59

3.8 Simple Correlation Matrix for Variables Used in the
Dairy Equations . . . . . . . . . . . . . . . . . . . . 61

3.9 Regression Results for Dairy Equations . . . . . . . . . 62

3.10 Price, Income and Transfer Elasticities for Dairy
Praducts O O O O O O O O O O O O O O O O O O I O O 0 O 65

3.11 Dairy Products: Farm-Retail Spreads and Percent of
Retail Price Going to the Farmer (1950—1970) . . . . . 81

3.12 The Fed-Beef Model . . . . . . . . . . . . . . . . . . . 83
3.13 The Non-Fed Beef Model . . . . . . . . . . . . . . . . . 84

3.14 Simple Correlation Matrix for Variables Used in the
Livestock and Poultry Equations . . . . . . . . . . . . 87

3.15 Regression Results for Beef . . . . . . . . . . . . . . . 88

3.16 Income and Transfer Elasticities for Beef . . . . . . . . 101

vi

Table

3.17

3.18

3.19

3.20

3.21

3.22

3.23

3.24

4.1

Beef, Pork and Poultry: Farm Retail Spreads and

Percent of Retail Price Going to Farmer (1953-1970)

The Hog Model . . . . . . . . . . . . . . .

Regression Results for Hogs .

Income and Transfer Elasticities for Hogs and Poultry

The Broiler Model . . . . . . . . . . . . .
Regression Results for Broilers . . . . . .
The Egg Model . . . . . . . . . . ... . . .
Regression Results for Eggs . . . . . . . .

Comparison of Transfer Elasticities and Low
Elasticities . . . . . . . . . . .

vii

Income

Page

102
104
105
111
113
114
123

124

136

Figure

2.1

3.8
3.9

3.10

3.11

3.12

3.13

3.14

3.15
3.16

3.17

LIST OF FIGURES

Derived Demand .

U. 8.: Quantity of Wheat used for Food per Capita
{1952-1972) 0 o o o o o o o o o o o o o o o o 0

U. 8.: Farm Price of Wheat Deflated by the CPI (1952-1977)

U. 8.: Quantity of Wheat Flour used in Making Bread per
capita (1952-1972) 0 o o o o o o o o o o o o o o o 0

U. 8.: Quantity of Wheat Flour used in Making Flour
Packaged for Retail Sale per Capita (1952-1972) . .

' U. 8.: Quantity of Wheat Flour used in Making Cookies and

Crackers per Capita (1952-1972) . . . . . . . . .

U. 8.: Retail Price of Bread Deflated by the CPI
(1952-1972) 0 o o o o o o o o o o o o o o o o o 0

U. S.: Price of Flour Packaged for Retail Sale Deflated
by the CPI (19.52-1972) o o o o o o o o o o o o o o 0

U. 5.: Price of cookies deflated by the CPI (1952-1972).
U. 8.: Bread: Price and Consumption Indexes (1952-1972) .

Determination of U. S. Department of Agriculture's Farm-
Retail Spreads: An Example . . . . . . . . . . . . . .

U. S.: Farm-Retail Price Differential for Cookies Deflated
by the CPI (1952-1972) 0 o o o o o o o o o o o o o 0

U. 8.: Farm—Retail Price Differentials for Bread Deflated
by the CPI (1952-1972) 0 o o o o o o o o o o o o o o o 0

U. S.: Farm-Retail Price Differentials for Flour Packaged
for Retail Sale Deflated by the CPI (1952-1972) . . . . .

U. S.: Fluid Milk: Price and Consumption Indexes
(1952-1970) e o o o o o o o o o o o o o o o o o o o o 0

U. 8.: Cheese: Price and Consumption Indexes (1952-1970)
U. S.: Butter: Price and Consumption Indexes {1952-1970)
U. 8.: Quantity of Milk Produced per Capita (1952-1970).

viii

Page

35

36

37

38

39

40

41
42

43

49

53

54

55

66
67
68

69

Figure

3.18

3.19

3.20

3.21

3.22

3.23
3.24

3.25

3.26
3.27

3.28

3.29

3.30

3.31

3.32

3.33

3.34
3.35
3.36

3.37

U. 8.: Farm Price of Milk Deflated by the CPI
(1952-1970) 0 o o o o o o o o o o o o o o o o o o o 0

U. 8.: Quantity of Milk used for Making Fluid Milk per
capita (1952-1970) 0 o o o o o o o o o o o o o o 0

U. 8.: Quantity of Milk used for Making Cheese per
Capita (1952-1970). . . . . . . . . . . . . .

U. 8.: Quantity of Milk used for Making Butter per Capita
(1952"1970) o o o o o o o o o o o o o o o o o o o o o 0

U. 8.: Retail Price of Fluid Milk Deflated by the CPI
(1952-1970) 0 o o o o o o o o o 0

U. 5.: Price of Cheese Deflated by the CPI (1952—1970)
U. S.: Price of Butter Deflated by the CPI (1952-1970)

U. 8.: Fareretail Price Differential for Fluid Milk
Deflated by the CPI (1952—1970) . . . . . . .

U. 8.: Beef: Price and Consumption Indexes (l953ul971).

U. 8.: Price of Fed Beef Deflated by the CPI (1953-1971) .

U. 8.: Quantity of Fed Beef Produced per Capita
(1953-1971) 0 o o o o o o o o o o o o o o o 0

U. 8.: Quantity of Steak Consumed per Capita (1953-1971) .

U. S.: Price of Non-Fed Beef Deflated by the CPI
(1953-1971) 0 o o o o o o o o o o o o o o o 0

U. 8.: Quantity of Non-Fed Beef Produced per Capita
(1953-1971) 0 o o o o o o o o o o o o o o o o 0

U. 8.: Quantity of Hamburger and Ground Beef Consumed per
capita (1953-1971) 0 o o o o o o o o o o o o o 0

U. S.: Farm-Retail Price Differentials for Beef Deflated
by the CPI (1953-1971) 0 o o o o o o o o o o o o 0

U. S.: Hogs: Price and Consumption Indexes (1953-1971).
U. S.: Price of Hogs Deflated by the CPI (1953-1971)
U. S.: Price of Ham Deflated by the CPI (1953-1971)

U. 8.: Quantity of Ham and Bacon Consumed per Capita
(1953-1971) 0 o o o o o o o o o o o o o o o o o o o 0

ix

Page

70

71

72

73

74
75

76

80
91

92

93

94

95

96

97

. 103

106

107

. 108

. 109

Figure

3.38

3.39

3.40

3.41
3.42

3.43

3.44

3.45

3.46

3.47

3.48

Page

U. S.: Farm—Retail Price Differentials for Ham Deflated
by the CPI (1953-1971) c o o o o o o o o o o o o o o o o 110

U. S.: Broilers: Price and Consumption Indexes
(1953-1971) 0 o o o I o o o o o o o o o o o o o o o o o 116

U. S.:. Quantity of Broilers Produced per Capita
(1953-1971) . . . . . . . . . . . . . . . . . . . . . . 117

U. S.: Price of Broilers Deflated by the CPI {1953-1971) 118
U. S.: Price of Fryers Deflated by the CPI (1953-1971). . 119

U. S.: Farm—Retail Price Differentials for Broilers
Deflated by the CPI (1953-1971) . . . . . . . . . . . . 120

U. S.: Eggs: Price and Consumption Indexes (1953-1971) 125
U. 8.: Quantity of Eggs Produced per Capita (1953-1971) 126

U. S.: Farm Price of Eggs Deflated by the CPI
(1953-1971) 0 o o o o o o o o o o o o o o o o o o o o o 127

U. 8.: Retail Price of Eggs Deflated by the CPI
(1953‘1971) o o o o o o o o o o o o o o o o o o o o o o 128

U. S.: Farm-Retail Price Differentials Deflated by the
CPI (1953-1971) . . . . . . . . . . . . . . . . . . . . 129

CHAPTER I

INTRODUCTION AND OBJECTIVES

The Role of Demand

 

An economic system is composed of an infinite number and variety
of interdependent components. The problem in analyzing the system is to
define possible variables in the system and postulate on their interactions.
Various components of the economic system must be defined and isolated to
gain an understanding of the whole. This dissertation attempts to iso-
late, describe and analyze the internal domestic demand for a specific
set of U. S. farm commodities.

The demand for any particular product is hypothetically a func-
tion of individual tastes and preferences as well as the prices of com-
peting and complementary goods and income. On a national scale, variables
such as age, income and geographic distribution are important as are in-
flation, average income and total population. Public policies affecting
these variables are also important. Programs which have a direct effect
on the consumers and thus on demand are the income and in-kind transfer
(i.e. food stamps) programs.

There have been some large shifts in the use of some major U. S.
commodities between 1951 and 1971. The quantity of wheat used for food
per capita in the U. S. has declined by about 17 percent over this period.
There has been a shift in use from bread to cookies and flour packaged
for retail sale. The quantity of milk used has also decreased by about

22 percent. Use of fluid milk and butter declined considerably whereas

the use of cheese increased. Per capita consumption of fed-beef, non-fed
beef, hogs and broilers also increased, but egg usage decreased by 10

percent.

Objectives

 

Using traditional elementary demand theory and econometric analy-
sis, this work attempts to isolate the variables affecting demand for
some major food commodities. Particular emphasis is on food stamps and
income transfer programs. Variables considered most important due to
their possible implication on food demand are substitutes and complements
for various food commodities, income, income distribution, and, therefore,
policies to redistribute income. Only other food items are tested for
their complementarity or substitutability. No attempt was made to analyze
the interactions and trade offs between demand for food and demand for non-
food items.

The objectives of this study are to describe and analyze changes
taking place in the consumption of the major food commodities.1 Within
that framework are the following objectives: 1) to investigate retail
product consumption; 2) to analyze the effect of income distribution and
income redistribution programs on food consumption; and 3) describe and

analyze the role of marketing costs and their effect on retail products.

Method of Analysis

 

A system of equations is developed for each commodity. They are
assumed to be independent except where one commodity is a substitute or

complement for another.

 

1Milk, wheat, beef, pork, broilers and eggs.

Demand for farm commodities is a result of consumer perception
of various retail products, not farm products. The retail product is a
combination of farm commodities and marketing services. Demand for mar-
keting services and farm commodities is, therefore, expressed for the
retail product. Prices and quantities of major retail products are rep-
resented herein in a system of equations.

To analyze demand for farm commodities and retail commodities,
some concept of marketing margins and costs is essential. Marketing
margins are not estimated directly in this analysis but a farm-retail
price differential is obtained from price analysis at the farm and retail
levels. Factors proxying marketing costs are included in the farm and

retail-level equations.

Organization of Thesis

 

Chapter 2 contains the analysis which underlies the work to fol-
low. The first section presents the basic demand function and defines
relationships between farm and retail demand. The second section in-
cludes the theory and empirical evidence concerning relationships be—
tween income distribution, income redistribution policies and food demand.

The empirical results and analysis of the results form the major
bulk of Chapter 3. (Empirical results are presented commodity by com-
modity). The final pages of the chapter contain an analysis of the re-
sults across the commodities for farm-retail price differentials and
transfer payments. A summary is presented in Chapter 4 along with some

questions raised by the analysis.

CHAPTER II

FRAMEWORK FOR ANALYSIS

Demand Theory

 

The demand schedule represents the maximum commodity quantities
taken by consumers at each price level in a given time period. .For a
given demand schedule, everything else is assumed constant. For the in-
dividual consumer demand function, income, consumer tastes and prefer—
ences and consumer investments in consumer durables and disinvestment
in wives as factors in home production as well as prices of food substi-
tutes and complements are among the relevant variables assumed constant.
In the aggregate, the above factors, as well as income distribution and
demographic factors (such as total pOpulation size and age distribution)
are assumed constant. A shift in any of these factors will cause a
shift in the aggregate demand schedule. A shift in the demand schedule
will yield a different simultaneous solution of the demand and supply
functions, i.e., different values for market clearing prices and quanti-
ties.

Whether price and quantity are determined simultaneously is a
hypothesis which can be tested. If single demand functions are used for
each commodity, either own price or quantity must be assumed exogenous.
If quantity is exogenous, the supply function is assumed vertical and
predetermined in the time period at which we are looking at demand. If
prices are exogenous, the buyer is assumed a price taker. The quantity
bought has no effect on price. At the micro level for the individual

4

consumer, the function can best be expressed as quantity dependent, as

the individual has little effect on price. The more aggregated, the more
likely the amount bought will effect price. With agricultural products,
given the relatively fixed nature of supply, once crops are planted and
breeding herds established, price dependent demand equations are generally
used. If supply is not assumed fixed in the given time period, para-
meters can be estimated simultaneously.

The elementary traditional demand equation which is far from
ideal would include the commodity's own price and quantity as well as
prices of other possible substitutes and complements. The number of
parameters to be estimated must be less than the number of observations.
Given the limited number of years for which data are available, the nume
ber of variables must be severely limited.

Also, the number of years of data used must be limited because
the longer the time span, the more likely it is that some variables as-
sumed constant have actually shifted. (Tastes and preferences are two
examples).

One way to deal with the problem of degrees of freedom is to
divide all food into categories and use concepts suggested by economic
theory to find some of the parameters, thereby reducing the number which
must be directly estimated. There are costs to this procedure as well
as benefits.

1

This method is used by Brandow as well as George and King.

George and King's division of food into groups is less aggregated than

k

1G. E. Brandow, Interrelations Among Demands for Farm Products

and Implications of Market Supply, Pennsylvania Agricultural Experiment
Station Bulletin 680, (University Park, 1961); P. S. George and G. A.
King, Consumer Demand for Food Commodities in the United States with

.fggjgctiong for 1980. Giannini Foundation Monograph No. 26 (Davis, March
1).

 

 

 

that of Brandow because they made further assumptions about the utility
of goods. However, use of these concepts requires an accounting of total
income. There must be information about non-food use as well as disag-
gregated food use of income.

It is difficult to explain differences existing between retail
and farm level demands for food. Usually, the product bought at the re-
tail level is quite different from that sold at the farm level. For ex-
ample, milk at the farm level becomes ice cream, cheese, yogurt, whipped
cream, etc., at the retail level. Though some eggs undergo no essential
transformation in form they make a transformation in place and time.
Thus, there is a difference in the product whether it be form, time or
place.

The difference in product price at the retail and farm level re-
flects supply and demand for services which change the product in time,
form and place. Waldorfl attempts to measure demand for these marketing
services. He believes a total consumer demand exists for marketing ser—
vices, including those services which will be bought and those provided
at home. Those services which will be bought, i.e. demand for marketing
services, are, according to Waldorf, consumer demand for these services
minus the household supply of these services.

(Finding a proxy for a quantity and price of marketing services
is difficult. Waldorf substitutes Bureau of Census data for "value added
by manufacturers" adjusted for price changes. The quantity of marketing
services consumed is the difference between value of gross output de-

flated by an index of wholesale food prices and costs of materials and

¥

1William Waldorf, "The Demand and Supply for Marketing Services,"
iggrnal of Farm Economics, XV, 42-60.

supplies deflated by an index of materials and supplies. The price of
food manufacturing services is value added in current prices divided by
value added in constant prices. 4

When estimating farm level demand, the effect of marketing ser-
vices on farm demand must be considered. Farm level demand without the
demand for marketing services cannot be theoretically justified; it
derives from the retail demand and must be analyzed as such.

Demand for the retail product includes demands for marketing
services and for farm products that are inputs into the retail product.
Neither George and King nor Brandow attempt to handle the question of
marketing services within the supply and demand framework. Instead,
they link up farm and retail prices by hypothesizing about marketing
firm pricing policies. Brandow, for example, assumes a constant markup
plus a percentage markup.1 Thus, only retail price and commodity quan—
tities will affect farm prices. George and King investigate various
pricing policies and combinations of pricing policies and, as with
Brandow, derive farm level demand from the estimated retail demand func-
tions. 1

To obtain derived demand Friedman2 assumes inputs into the re-
tail good are in fixed proportions (Figure 2.1). The demand for one of
the joint inputs is derived from the difference between the supply curve
for the other and the demand for the total product. Thus, the derived

demand for farm goods (inputs into a retail product) is the vertical

 

lBrandow, Interrelationships Among Demands, p. 44.

 

2Milton Friedman, Price Theory (Chicago: University of Chicago
Press, 1962).

 

Price for Retail Commodity
Price for Marketing Services
Price for Farm Commodity

Retail Commodity Supply

    
 
  
  

Supply of Marketing
Services

P b--------

Supply of Farm
Commodity
Retail Demand

 

 

Quantity of Retail Commodities

O 1 2 3 Quantity of Marketing Services
0 1 2 3 Quantity of Farm Commodity

P

PR 3 -- - - -

Supply of Marketing Services

Retail Commodity

 

Derived Demand for Farm Commodity

 

Q

Figure 2.1

Derived demand

difference between the demand for retail products and supply of market-
ing services.

The maximum amount paid for a retail good is shown by the demand
curve for the retail good. The minimum price for marketing services is
given by the supply curve for those services. Since the retail product
is produced with fixed proportions of inputs, according to the assumptions
of the analysis, the demand for farm products is a residual, or the dif-
ference between the two curves.

Assumptions of the static model include: 1) a fixed relation-
ship between farm foods and marketing services as inputs into the retail
good, and 2) demands for marketing services and farm goods are not simul—
taneous (except at the original equilibrium point) since each is derived
from the other's supply curve.1

Richard Foote2 presents the following model of derived demand
consisting of relationships for consumer demand, marketing group supply
and demand, and producer's supply.

Q 8 quantity consumed

c
Qp = quantity produced
Pr = retail price

P = price paid to producer
D = disposable income

Z 8 other factors affecting consumer demand

 

lIbid., p. 151.

2Richard Foote, Analytical Tools for Studyingjpemand and Price
Structure, U. S. Department of Agriculture Agricultural Handbook No. 146
(Washington, D. C., August 1958).

10

Z a other factors affecting marketing groups demand and
supply
Z i other factors affecting producer supply

1. F(Qc, P D, 21) - 0 consumer demand

I,

2. g(QC, Pr’ P , 22) 0 marketing group's supply

W

3- h(Qp, Pr’ PW, 22)

4. 1(Qp, Pw’ 23)

0 marketing group's demand

0 producer's supply

Assuming certain equilibrium conditions, the number of equations
to be estimated is limited. If Qc = Qp the marketing supply and demand
equations become one.

5. j(Qc, Pr’ P , Z = 0 marketing group's behavior

w 2)
The system then consists of functions for marketing group's be-
havior, producer supply, and consumer demand (1, 4, 5).
If Q supplied by the marketing group equals the quantity demand-
ed by the consumer group, retail price can be eliminated from 5 and l to
give the following derived demand for the producer's product.

6. k(Qc, P = 0

w, 22, D, Z

1)
Notice that 22 or factors affecting the marketing group's behavior must
be included in the estimation.

Foote's original four equation framework is useful. Z2 represents
factors affecting both the marketing group's supply and demand. A case
can be made for a differentiation of these factors affecting supply and
demand. Also, the assumption that Qc - Qp will not hold except when the
farm product becomes only one retail product. When there are several
retail products the system of equations becomes bigger. As with previous

methods, assumptions were made to eliminate variables and thus produce

a derived demand. This is a "partially reduced form equation derived

11

from a structural equation by the elimination of a price through the use

 

of equilibrium condition."1

 

Framework
Because of the interest in livestock and food grains only,
schema by Brandow and George and King of accounting for all income
not appropriate for this analysis. Rather than consider each food
a substitute or complement for another one, each commodity will be
tinized along with its most likely substitutes.
The general schema for commodity estimation is the same as

of Foote except exogenous factors affecting retail supply, Z , and

2

demand, Z differ. In Foote's schema they are the same.

 

3

Retail demand: QR = F(PR, 21)
Retail supply: QR = F(PR, PF’ 22)
Farm demand: QF = F(PF, PR, 23)
Farm supply: QF = F(PF, 24)
QR = quantity of retail good
PR = price of retail good
PF - price of farm commodity
QF 8 quantity of farm commodity
Zl = exogenous variables affecting retail demand
22 = exogenous variables affecting retail supply
Z3 = exogenous variables affecting fanm demand
Z4 - exogenous variables affecting farm supply

1

Ibid., p. 102.

the

is

item

scru-

that

farm

12

Because of the adjustment period for some agricultural commodities,
supply and demand at the farm level may not be simultaneous. The quan-
tity of the commodity supplied may be determined by the price of the

previous year. This hypothesis is testable. Q is then an exogenous

F
variable in the farm demand equations.
The schema changes to:
1. Farm supply: QF = F(PF t-l’ Z4)
2. Farm demand: PF = F(QF, PR, 23)
3. Retail supply: QR = F(PR, PF’ 22)
4. Retail demand: QR = F(PR’ 21)

Functions 2, 3, and 4 can be estimated simultaneously or PR can
be excluded from 2 and the farm demand equation estimated independently
of retail supply and demand. If the latter approach is taken, farm de-
mand must be estimated as a derived demand and the effects of Z1 and 22
included in the estimation.

The marketing margin or difference between price at retail and
price at the farm emerges from this relationship and supply and demand
for marketing services are accounted for, although not specifically test—
ed for. It is not assumed that QF and QR are the same as in many in-
stances, they are not. QF may become many different products at the re-
tail level. Thus, the system of equations needed to explain farm level
demand becomes larger as more retail products are produced from the raw
product. The system for each commodity expands to several equations of
type 3 and 4, two for each retail product as well as 1 and 2 (farm supply
and demand).

The new system becomes

1. Retail demand: R = l,...,n QR = F(PR E

 

l3

2. Retail supply: R = l,...,n QR = F(PR, PF’ Z2)

3. Farm demand: R = l,..,n PF = F(QF, PR, Z3)

4. Farm supply: QF = F(P , E )
Ft-l 4

QF’ PF’ QR’ PR are self-explanatory equation variables. Z1 vari-
ables are the substitutes and complements that retail customers would be
choosing among. These would also include more general variables such as
average income, population, income transfer programs, and the number of
bonus food stamps or total transfers made to low income consumers.

22 variables in the retail supply equation represent inputs into
retail product production. This includes changes in the farm product at
the wholesale, processing and retail level. Important inputs are: labor,
transportation and fuel costs, substitutable and complementary farm com—
modities. Commodities labeled substitutes at the farm level are those
which the processor or wholesaler considers substitutes. In processing,
for example, soybean protein may be considered a partial substitute for
beef, especially in longer runs; at the retail level, pork and chicken
would be substitutes.

Z3 variables in the farm demand equation resemble those in the
retail supply equation except they must include substitutes and comple-
ments for all retail production processes and the variables applicable
to all, such as labor. The price or the quantity of substitutes and
complements can be used. When quantities are used, it is assumed that
they are predetermined in the given time period. If prices are used, it
is assumed that the buyers are price takers and the prices of the various

goods are not determined simultaneously.

l4

Z4 variables are the inputs into the production process and
other exogenous variables affecting farm supply such as policy prices.

If it is hypothesized that supply is predetermined, then Z4 variables do

not need to be included in the estimation.

The Interrelationships Among Income Distribution
Income Redistribution Policy and
Demand for Food

 

Income distribution has an affect on the demand for food; hence
the implication is that income redistribution programs also affect food
demand. The connections among these three variables must be established,

at least partially and approximately.

Transfer Payments and Income Redistribution

The purpose of government transfer payments and taxation is to
redistribute income to designated target groups. There are three aspects
to the redistribution. One is taxation; another is deficit financing and
the third is transfer. No attempt is made to take into account all as-
pects of the former two processes.

Much taxation and subsequent transferral of income, goods and
services occurs among groups of people whose particular characteristics
would be difficult to isolate. Payments are made to groups, e.g., vet-
erans and the aged, based on their particular characteristics.

Low income groups also obtain special consideration. Programs
catering to them specifically are called income-tested programs.

It is important to distinguish between programs which transfer
income directly and in-kind programs which provide services or a par—
ticular good instead of cash. Most relevant to this analysis are income-

tested, direct income transfers and food stamp programs.

15

Tables 2.1 and 2.2 give some perspective on the particular pro—
grams as related to other transfer programs and to social welfare payments
in the United States. Table 2.1 lists the total social welfare programs
and more specifically the public aid programs. Public aid programs in-
creased from 10.6 percent of all social welfare programs in 1950 to 11.2
percent in 1970.

The three income transfer payments used in this analysis are also
listed. Aid to families with dependent children (AFDC) has had the
largest increase in spending. A brief description of these programs ap-
pears in Appendix A.

Table 2.2 compares income dependent programs with other transfer
programs sponsored by various government levels. Table 2.3 outlines the
historical data along with the most current data on participation num-
bers and total value of food stamps issued.

In 1970, the program was liberalized greatly increasing the num-
ber of participants and government expenditures. If April 1976 is in-
dicative of the rest of 1976, federal contributions to the program for
the year would amount to $5,466 million-~twice the 1974 amount and ten
times the 1970 amount.

Looking briefly at income distribution in the United States re-
veals two major trends.1 First, the percent of aggregate income re-
ceived by each fifth of the families has not changed significantly since

1947. Second, the difference between the lowest fifth and highest fifth

 

1Sylvia Lane, "Effectiveness of Public Income Redistributive
Policies and Programs on the Redistribution of Income to Low-Income Groups"
(paper presented at the meeting of the American Agricultural Economics
Association, Columbus, Ohio, August 12, 1975), p. 1.

16

 

 

 

TABLE 2.1
U. S.: TRANSFER PROGRAMS (1950-1974)8
1950 1960 1970 1974 1974
1950
TOTAL SOCIAL WELFARE PROGRAMS
Federal, state and local.
Includes social insurance,
public aid, health and medical
programs, veterans programs,
education, housing and other
social welfare.
(million dollars) 23,508 52,293 145,962 242,386 10.3
TOTAL PUBLIC AID PROGRAMS
AFDC, 0AA, GA,
Aid to permanently and
totally disabled,
Aid to blind, b
Medical Asst. for aged
(million dollars) 2,496 4,101 16,488 33,628 13.4
TOTAL PUBLIC AID AS PERCENT OF
TOTAL SOCIAL WELFARE PROGRAMS 10.6% 7.8% 11.2% 13.8%
SPECIFIC PUBLIC AID PROGRAMS -
AFDC million dollars 556 1,055 4,857 7,991 14.3
thousand recipients 2,234 3,073 9,659 11,006 4.9
0AA million dollars 1,469 1,922 1,866
thousand recipients 2,786 2,305 2,082
GA million dollars 353 422 731 1,131 3.2
thousand recipients 413 431 547 585 1.4

 

aIncludes administrative costs.

b

Assistance; GA=General Assistance

AFDC=Aid to Families with Dependent Children; OAA=Old Age

Source:

Abstract of the United States, 1975 (Washington, D. C.:

Printing Office, 1975), p. 284.

U. S. Department of Commerce, Bureau of the Census, Statistical

 

Government

17

TABLE 2.2

U. S.: BENEFIT OUTLAYS UNDER PUBLIC INCOME
TRANSFER PROGRAMS, FISCAL YEAR 1972
(in billions of dollars)

 

Benefit outlays, fiscal year 1972

 

 

 

 

 

 

 

 

State and
Program Total Federal Local
INCOME—TESTED PROGRAMS:1
Aid to families with dependent Children 6.7 3.7 3.0
Old age assistance 2.5 1.7 .8
Aid to the blind .l .06 .04
Aid to the permanently and totally
disabled 1.5 8 .7
General assistance .7 .7
Veterans' pensions 2.5 2.5
National school lunch program (free
or reduced-price lunches) .5 .5
Food stamps 2.0 2.0
Food distribution (to individuals and
families) .3 .3
Public housing .8 .8
Medicaid 7.0 3.9 3.1
Total, income-tested programs 24.6 16.3 8.3
OTHER INCOME TRANSFER PROGRAMS:
Old age and survivors insurance 34.5 34.5
Disability insurance 4.0 4.0
Railroad retirement 2.1 2.1
Civil service retirement 3.4 3.4
Other federal employee retirement 4.0 4.0
State and local retirement 3.3 3.3
Unemployment insurance 6.4 6.4
Workmen's compensation 3.0 .2 2.8
Veterans' medical care 2.2 2.2
Veterans' compensation 3.6 3.6
Medicare 8.5 8.5
Total, other programs 75.0 68.9 6.1
Total, all programs 99 6 85.2 14.4

 

1These programs base benefits on current needs of recipients.

Source: U. S. Department of Commerce, Bureau of the Census, Statistical
Abstract of the United States, 1975 (Washington, D. C.:
Government Printing Office, 1975) p. 281.

 

l8

 

.Nom .a Amama .muammo mcauaaum

 

 

unmaauo>ou ".0 .Q .aouwcaSmmzv mmma muaumaumum amuouaouauww .musuaauauw< mo .uaoa .m .D "mouaom
.ooo.oomm ammo mmmam
saw o.ao seem «mam osam comma asmuumaouav
omma
mom w.oo Noom mmmm oamm oomma .umo mmma
NON n.5n maNN meow hmes mamN .umm snaa
Qua m.qm amam mmma «mmm mmmm mnma
mma m.qm nmna aama momm Nmam umma
moa a.om mNma amaa mamm amma amma
“Na q.om omm can omoa mama omma
om m.nm mum «mm moo mmqa moma
mm m.mm mma mum «we muoa mema
mm o.mm ooa ama omm mmm Roma
mm ~.~m mo moa «ma «mm coma
om o.mm mm mm mm oaa mama
mm o.mm mm mm mm mm aoma
mm «.mm ma am om Ne moma
mm e.nm ma mm mm m mama
w Aa.osv Ame Ame a o asma
Amumaaomv oaam> ammumu Aw .aaav Amumaaov :oaaaaav Amumaaov aoaaaaav Aooo.av
mommaoauumm amuou mo unouumm amuoa mammaoauumm ou umoo osam> aawus.H amuoa mucmmauauumm
mom meadow mason voamma maamum mo moam>

 

aonmauaomav z<muomm mz<em noes magmamm ".m.a

m.N mam<H

19

of the families has widened considerably. One measure of the difference
is obtained by dividing all families into five income groups and examin-
ing the top income in each income group. In constant dollars the dif-
ference between the highest income in the fourth income group and in the
first income group was $7,384 in 1947 and $13,945 in 1974.1
According to a study by Reynolds and Smolensky,2 the trend with-
out transfer payments would have been towards more inequality. Transfer

payments considerably slowed the rate of increase in income inequality

in 1961 and 1970.

Income Distribution and Demand for Food

Income level is a major determinant of the value of food bought.
Income distribution should affect the amount of money spent on food as
well as the basket of commodities bought.

Theoretically, money received from a transfer payment should not
be spent any differently than money from any other income source. Cash
transfers, unlike food stamps, for example, do not lock spending into any
particular good. Thus, with knowledge of income elasticities for income
groups and knowledge of income distribution one should be able to project

food consumption with and without transfer payments.

Food Stamps and Demand for Food
Food stamps have a more direct effect on food demand. Theoreti-

cally, they should lock some of the money spent on them to food buying.

 

1U. S. Department of Commerce, Statistical Abstracts, 1975, p. 392.

2M. Reynolds and E. Smolensky, "Post—Fisc Distribution of Income:
1950, 1961 and 1970" (paper presented at the meetings of the Western
Economic Association, June 27, 1975).

20

Stamp issuance depends on income. The participant pays a sum for a cer-
tain amount of stamps. The difference between the total amount of food
stamps and their cost is the amount of "bonus" stamps. The increase in
demand for food can at the maximum be equivalent to the bonus food stamps
or at the minimum zero.

If the family chooses to spend the same amount of cash as previ-
ously spent on food on food stamps, all bonus stamps will represent a
demand expansion for food. However, if they choose to spend less cash
on food and purchase the number of food stamps that will buy them the
same amount of food as previously, there is no demand expansion at all.
The average is somewhere in between and depends on income and household
size.1 The lower the income, the larger the family size and the lower
the nutritional level of the family, the more effective the food stamps
are in increasing food demand.

The U. S. Department of Agriculture (USDA) estimates that with
a cash supplement of $1.00, demand for food is expanded by $.20 to $.30
and with $1.00 of bonus food stamps, demand is expanded by $.50. Indi-
vidual food items were also singled out, using data from the Household
Food and Consumption Surveys (HFCS) of 1955 and 1965 to estimate total
increase in demand for these various foods due to food stamps. Food

stamps have a particularly large impact on red meat.

 

1U. S. Department of Agriculture, Economic Research Service,
Bonus Food Stamps and Cash Income Supplements: Their Effectiveness in
Expanding:Demand for Food, Marketing Research Report No. 1034 (Washington,
D. C.: Government Printing Office 1974), p. iv.

21

Mittelhammer and West1 set up a theoretical framework for com-
paring the impact of food stamps on the recipients with that of a cash
transfer on demand for food. Theoretically, demand for food would de—
pend on the indifference curves but generally expenditure on food is
expected to be more with food stamps than with a cash subsidy. By
studying the number of people spending more for food than their food
stamp allotment, they could estimate the numbers that would tend to
spend the same amount on food with food stamps as with a cash transfer.
0r, conversely, they could estimate the numbers of households where the
food stamp programs increase demand for food above that which would be
present with a direct cash subsidy. Thirty-eight percent of the families
in the 1965 Household Survey used less food than would have been allocated
to them under the food stamp program. This represents an estimate of
those families for which the food stamp program would increase the food
bought above that for a cash subsidy of value equal to the bonus amount

in the food stamp program.

Analysis of Transfer Payments
Little work has been done on inclusion of policy variables in food
demand equations. One problem is the dearth of relevant data linking
retail demand to food stamps. The ideal data for food stamps would be
a detailed study of consumer purchases of relevant food categories, in-
come, various socio-economic criteria and whether or not food stamps

were used.

 

1Ron Mittelhammer and Donald A. West, "Impact of the Food Stamp
Program on the Demand for Food, Some Theoretical Considerations," (paper
presented at the 1974 American Agricultural Economics Association meet-
ings in College Station, Texas, August 1974).

22

Most data collected cannot be used for this type of analysis.
Because of the interest in nutrition of food stamp users, much of the
data collected outline differences in food eaten by food stamp users and
non-food stamp users rather than food bought.1

Transfer payments as well as food stamps affect the consumer's
expenditure pattern. In this study food stamps and transfer payments
are analyzed for their effect on consumption. The two different programs
are summed and a variable representing the two is used to estimate the

effect of these variables on food demand.

Commodity Specific Structures

Retail and farm demand for each commodity is estimated using a
different set of equations and variables. The equations vary according
to hypothesized simultaneity among the supply and demand curves and the
number of retail products considered in the analysis of a particular
farm product. Retail products considered in the analysis are determined
partly by the importance of the retail product that the farm commodity
is used in, and by availability of data.

Variables in the equations at the retail level depend on the sub-
stitutes and complements for the retail product. All retail equations
include a food stamp or transfer variable as well as an income variable.

To account for the effect of aggregate population change and inflation,

 

1Dae Sung Lee and Gus T. Ridgel, Income and Food Expenditure
Patterns for 3 Selected Group of Families in Frankfort-Franklin County,
Kentucky State University, Cooperative State Research Service (Frankfort,
May 1974); J. Patrick Madden and Marion D. Yoder, Prpgram Evaluation:
Food Stamps and Commodity_Distribution in Rural Areas of Central
Pennsylvania, Pennsylvania State University Agricultural Experiment
Station Bulletin 780 (University Park, June 1972).

 

23

all quantities are per capita and all prices are deflated by the Consumer
Price Index (CPI).

Variables in retail supply equations include inputs into the
processing of the farm product and its retailing. In all retail supply
equations, wages of food manufacturing employees and the fuel price index
were used to proxy these inputs into retail products. The wage used here
is only that of food manufacturing employees. It excludes wages in other
areas such as transportation, where the wages might be higher. The trends
are similar. The simple correlation between the wage of food manufactur-
ing employees and wages of employees in the private non-agricultural sec-
tor is .73. There is no proxy for labor—saving capital and new technology.
In cases where a possible substitute or complement exists for the farm
input into the retail product, it was included as a variable.

Equations representing demand for a farm commodity include vari-
ables from all retail supply equations for that particular farm commodity.
Thus, aside from wage and fuel, other variables included are substitutes
or complements used to produce any of the retail products from that come

modity.

Estimation Procedure

 

The two-stage least squares estimation is used in an attempt to
capture the simultaneity in allocating the quantity of farm commodities
among the various retail products. This procedure requires that all but
one endogenous variable be estimated in the first stage of the process
using ordinary least squares (OLS). All exogenous variables in the whole
system of equations are the independent variables in this first stage of
estimation. In the second stage, the estimated values of the endogenous

variables in the first stage are used to estimate the one remaining

24

endogenous variable.

Each equation in the system is estimated separately. Thus, more
than one equation may be used to predict a particular variable. For ex-
ample, both retail supply and demand equations can be used to predict
prices and quantities at the retail level by using included and excluded
exogenous variables to project all endogenous variables. Alternatively,
the supply and demand equations can be solved simultaneously and used

for prediction.

CHAPTER III
STATISTICAL ANALYSIS
Outline for Presenting the Statistical Analysis

The results are presented commodity by commodity. A model is
presented for each specific farm commodity along with a discussion of
the model. Data problems for the specific commodities and a simple cor-
relation matrix for each set of data are presented.

A measure of the multi-collinearity between variables is provided
by a simple correlation between the variables. Because three separate
sets of data were used, three simple correlation matrices are presented.
There is a listing and analysis of regression results.

The analysis emphasizes: 1) price and quantity relationships;

2) production of a variety of retail products from the farm products;

3) price differentials between farm and retail products and 4) the ef—
fect of income and transfer payments on the various retail commodities.
Excluded exogenous variables for each equation, i.e., the exogenous vari-
ables for the other equations in the system are listed along with the
results for each equation. The results include the coefficients for each
variable, the standard error for each coefficient, and the R2 and Durbin-
Watson statistic for each equation.

The first stage of the two-stage process is used for estimating
some variables. Equations used for that purpose are listed along with
the other regression results. Because they are not structural equations

their estimated parameters are not included in the analysis of the

25

26

regression results.

Wheat

Modeling the Wheat Sector

To model and estimate the wheat sector, it is necessary to find
data for not only a relevant category of retail products, i.e. those
products which are the major retail products from the farm commodity,
but also a matching retail price for a sufficient time span. Bread, flour
and cookies meet those criteria best.

The major use of wheat for domestic food consumption is bread.
According to the Census of Manufacturers,1 71 percent of the wheat flour
in 1972 went into making commercial breads and cakes. In 1954, 74 per-
cent of it was used for those products. This represents 40 percent of
the total wheat used for food in the United States in 1954 and 1972.

Other major uses of wheat are flour packaged for retail sale,
cookies, crackers, cereals and pastas. Retail flour, and cookies and
crackers changed from two percent and seven percent of the quantity of
wheat used for food in the United States in 1951 to four percent and 10
percent in 1972. There is some specialization in use of various types
of wheat for retail products. A more accurate picture of the retail
wheat market might be obtained if wheat were disaggregated according to
these specializations for estimating purposes.

Wheat is not only used for domestic food consumption, but also
for feed, seed and alcohol. Larger proportions of the United States crop

are exported every year. In 1950, 20 percent of the United States crop

 

1This categorization excludes small-scale bakery production
where goods are sold directly to the consumer.

27

was exported. The percentages have continued to increase from 40 percent
in 1960 to 52 percent in 1970 and 56 percent in 1975.
Data: Correlation Matrix, Sources of Data
and Adequacy of Proxies

There are several problems with the available retail data. Many
retail prices from the Bureau of Labor Statistics, for example, have no
equivalent retail quantities which can be used for analysis. Thus, many
variables must be proxied. How each particular commodity is proxied is
explained in the section dealing with the data problems of that commodity.

Retail data for wheat products are no exception. First, the data
on quantities of the relevant retail products are taken from the Census
of Manufacturers. These data are collected only every 5 years. A linear
trend is used between points to estimate the quantities of wheat used in
the various retail products for the years in between.

Second, the category used in this analysis to represent bread
quantities includes not only bread but also "cakes and related products."
Also the cooky variable is represented by data for cookies and crackers.
Third, the amount of cookies, bread and wheat flour used are proxied by
the amount of wheat that goes into these products as inputs rather than
the amount of the retail products produced. Fourth, the retail product
prices used are very product specific. The price of bread, for example,
is the price of white bread and the price of cookies is sandwich cookies.
A more thorough analysis would require finding better matches for the
quantities and prices used.

The variables affecting the quantity of wheat used for other
than domestic consumption (QWHOTHER), i.e., export and feed demand, are

proxied here by lagged quantities of wheat produced and exported. In

28

TABLE 3.1

THE WHEAT MODEL

 

System of equations for wheat

 

Retail demand:
Retail supply:
Retail demand:
Retail supply:
Retail demand:
Retail supply:

Farm demand:

Farm supply:

QBREAD = f(PBWHR, 2)

QBREAD = Herman, PWHF, Z)

QFLOUR = f(PFLR, 2)

QFLOUR = f(PFLR, PWHF, Z)

QCOOKIE = f(PCKR, Z)

QCOOKIE = (PCKR, PWHF, 2)

QWFOOD = f(PWF, PBWHR, PCKR, PFLR, 2)
QWOTHER = f(PWF, Z)

QFOOD + QWOTHER = Ofififfififi

QWHPROD a f(PWFt_1’ Z)

Endogenous variables

 

QBREAD
PBWHR
PWHF
QFLOUR
PFLR
QCOOKIE
PCKR
QWFOOD
QWOTHER
QWHPROD

quantity
price of
price of
quantity
price of
quantity
price of
quantity
quantity
quantity

Exogenous variables 2

 

PBWHR
PSUGAR
FUEL
STAMPS

QWHXLAG
WAGEFME
Y

PFLR

price of
price of
price of

of bread

bread

wheat at farm

of retail flour

retail flour

of cookies

cookies

of wheat used for food

of wheat used for other than domestic food
of wheat produced in U. S.

bread
sugar
fuel, power and light bought by manufacturing firms

federal contribution to food stamps
QWHPRODLAG quantity of wheat produced in the U. S. in time t-l

quantity

of wheat exported from the U. S. in time t-l

wage of food manufacturing employees

disposable personal income

price of

retail flour

29

further analysis, they might be proxied by foreign production of wheat
and the number of cattle on United States farms. (These are more specific
indicators of export demand and feed demands). The number of cattle on
farms indicates the amount of food needed and the overseas production
of wheat indicates the amount of competition for United States suppliers.
Table 3.2 lists the simple correlation matrix for variables in
the wheat equations. However, many variables are highly correlated with
time (YEAR). Thus, in many equations, the coefficients could be biased
because of the correlation of an included variable with a variable which
has been excluded from the system, but is also highly correlated with
time. Also, transfer payments (TRANS) and disposable personal income
(Y), which are included in most retail equations, are highly correlated

(.80). This particular problem is discussed later in the chapter.

Analysis of Regression Results

Quantity and Price

 

As evident from Figure 3.1, the amount of wheat used for food per
capita in the United States has dropped significantly. This represents
the decline in wheat used for making bread (see Figure 3.3), and the more
recent decline in wheat used for cookies and retail flour between 1967
and 1972 (Figures 3.4 and 3.5). From 1951 to 1972, (the time period used
for analysis), the total amount of wheat used for food increased but the
amount declined. Over the same time period, there was a perceptible de-
cline in price of cookies and retail flour, and an increase followed by

a decrease in the price of bread (see Figures 3.6, 3.7 and 3.8).

30

TABLE 3.2

SIMPLE CORRELATION MATRIX FOR VARIABLES USED IN THE WHEAT EQUATIONSa

 

"I 'U "U '13 "U "U IO 0 O 0 O O U3 "3 5: "C
C: on O p m C W C '11 S 1: a: "i > F1
1" O :6 :9 O O O 'U :5 m 73
5 > 3 7S C O 50 U)
:9 an H :9 c o n m a
m C
CD I"
3'
o
FUEL 1
PBWHR .35 1
PCKR .89 .32 1
PFLR .77 .44 .95 1
PSUGAR .71 .53 .78 .80 1
ewe .81 .11 .92 .84 .68 1
QBREAD .65 .31 .90 .96 .70 .81 1
QFLOUR -.66 .17 -.71 -.53 -.27 -.73 -.56‘ .85 1
qwroon .65 -.13 .86 .80 .56 .89 .86 -.91 -.76 1
QWHPRODLAG -.42 -.38 -.35 -.33 -.39 -.20 -.22 .01 .13 -.08 1
qwxrac -.41 .27 -.60 -.49 -.25 -.69 —.54 .86 .84 -.77 -.04 1
STAMPS -.46 -.66 -.70 -.84 -.72 -.54 -.84 .19 .22 -.52 .97 .25 1
TRANS -.61 -.65 -.82 -.92 -.79 -.65 -.9o .30 .32 -.61 .32 .30 .97 1
WAGEFME -.80 —.16 -.93 -.89 -.72 -.93 -.87 .76 .65 -.90 .28 .60 .62 .73 1
YEAR -.78 —.11 -.96 -.91 -.70 -.94 -.92 .83 .77 -.96 .21 .73 .66 .75 .95 1
r -.83 -.25 -.97 -.91 —.78 -.93 -.89 .72 .72 -.89 .25 .66 .70 .80 .95 .97

 

aAll quantities are per capita and all variables expressed in dollars have been deflated by the
CPI (l967=100). TRANS. STAMPS and Y are deflated per capita. Further explanation and data sources for
all variables can be found in Appendix B.

31

Before 1967 the decline in bread consumption could be explained
by a response to price changes.1 After 1967 this was no longer possible
as both prices and consumption declined (Figure 3.9). Statistical analy-
sis verifies that price is not the most explanatory variable although it
has an effect. The coefficient on the price of bread in Equation 1.5,
Table 3.3, is very significant but the price elasticity is small. My
results indicate a price elasticity of -.33. The analysis of King and
George for comparison produced a price elasticity of -.15.~

Consumption of both flour purchased for retail consumption and
cookies is unresponsive to price according to the statistical analysis
(Equations 1.1 and 1.3). Price coefficients for both cookies and flour
are insignificant. A search for possible substitutes and complements
for retail products made from wheat was unsuccessful. The only sta-
tistically significant substitute for wheat flour is bread and for bread
the only substitute is wheat (Equations 1.1 and 1.5).

The idea that there is a substitute for an item which takes such
a small portion of the budget is difficult to accept. Therefore, further
research is in order. What is labor in the house worth? At what in-
come level, for example, would the consumer make bread instead of buy it?

Equations 1.6 and 1.7 indicate that marketing costs may have an
effect on consumption. The price of fuel is significant in both bread
and total wheat consumption because bread is the major component of wheat
used for food in the United States. Data show that fuel costs decreased

relative to other goods between 1951 and 1971 and thus would have a

 

1U. S. Department of Agriculture, Economic Research Service, Food

Consumption, Prices and Expenditures, Agricultural Report No. 138
(Washington, D. C., Government Printing Office, 1968).

32

 

U<Ax30

 

 

 

 

wmz<em AmN.v Amm.~av Am~.v
oaaaommmzo amen Na. + au<2 om.a I m<u=mm ms.+
mm.a mm. m mmaxoou mo e.a
mama Ama.nV Aoo.v aaaaam aamumm
magma mam mo.na I axon mm. + ma.mm I mmaeoooc
oaaxzm
u<aaommmso
amps “as.v
mu. ma. azamo<3 mmzaam o~.a I
m<o=mm mmasoou mom m.a
mama Aea.mv Ame.v sesame aamumm
mazes m Am.na + meow No. + mm.o~ n mmHaoooo
o<axsm
mmzaam Ao~.v Amw.msV
oaaaommmzo amps so. + .mzmmmaz -.me I
mm.a em. m macaw mo N.a
m<o=mm Aoo.a~v mo.av magnum aamumm
magma mam am.om I mama mm. + am.omI u maoamo
o<axzo Ama.mv Aam.eav
ueaaommmzo mmzaam mm.a I m q~.mm +
mm. NA. amen uaoau now a.a
azma6<3 Aaa.~v amm.av causes aamumm
m<oamm magma oo.~ + mama mm. + ma.~m I n maoamo
3a m mzoaowoxo muasmom mauaa coaumomm
N umenauxa

 

MH<mm3 mom mHADmmm onmmmmwmm

m.m mam<s

33

 

umamzo

Aoo.v Aa.av Aea.v Ama.v

 

 

 

 

mammam amaa mo. + mammomz mm. + mmoa No. I mama «o. +
mo.a am. umaaomamsc names you
w Aeq.v Aam.mv menace Emma m.a
mmoama magma mo. I aooamo Na.~ + mo.mI I ama
umamzo moo.v moo.v aaa.v Aao.v
mammam aaaa ao.Io<aa3a mo. I ozaaumz ma. I mmua so. +
aw.a mm. w<aoommm30 . umoss you
a Aao.v Amo.v sesame aama a.a
mausma mama mooo. + mmzma mo. I mo.e I aooamm
omamzo
mammam aaN.av Aqm.mov
mq.a no. umaaomamzo amaa mm.~ I azamumz oo.m I
W vamun— mo
mammma mom.smV amo.mv aaaaam aamumm o.a
mama aza am.mo + mmzma mo.ma + mm.omm I ammmmo
Am~.av Ao~.av
mammam co.m I m mm.~ +
o<aaomm30 vmoun mow
Aoc.v Amm.av .
ea.a am. amaa . . I . I sesame aamaom m a
azamumz mama on m + mmzma ma m am was ammmmo
3a m msoammoxo. muasmom oauam coaumomm
N wowsauxm

 

Ae.aaouv m.m mamma

34

.mmanmaum> msoaowoxo men oumuamoa moanmaum> onu um>o wumm .muouuo mumvcmum msu omoauom muamaoawmoou
onu Home: mamosuomumm .Aooaumomav Hmu mnu mp mommammm soon m>mn mmanmaum> ouaum aam was wuaamo mom was
moanmamm> muauamsm aa< .m xawcmma< ca mcsom on emu mumv aam mo muuaom was coaumauomom muoaaaou 4m

 

 

 

 

 

 

aqo.o
m<wzma No.+
Aaoo.v Ama.v Aoa.v
m.a omamZo mooo. + omaoomamzo no. I mammam oo.I comma
soaumsoo . mo ooaum aa.a
mm.a mm. scum Aam.av Amo.v Amm.av Amo.mv
mao .w mq.a + amaa oa. + mw<3 mo.~ + we. a mmzma
Amoo.v Amm.v Am~.v Aw~.¢v
ma omamzo moo. I omaoomammo om. I mammam oof w 3.? «.6368
ooauoomo mo moaum oa.a
oo.a mm. Boom Aoa.v Amo.v Aaw.mv Aaa.mav
mao mmobma No. I amam mm. + mu<3 ca.q I mm.mq u mmoa
$3
m<osma so. +
Amoo.v Amc.v Aom.v
m.a oaamso moo. I omaoomamzo oo. I mammao oo.aI so
5...... .. “3......
m~.a am. aoum Amu.cv Ama. Amn.mv Aqm.mmv
mac m mo.q + amph «N. + mzmmw<3 aa.~a I mn.mm I mama
an m msoaowoxo nuanmmm mauaa ooaumsum
N ouooaoxm

 

Ao.uo66o m.m mamma

35

Bu. per Capita

.15

.05

.95

.85

    
 

.75
Estimated from Equation 1.7

. Real Values
.65 "

 

.56

 

52 54 56 58 6O 62 64 66 68 7O 72

Year

Figure 3. 1

U. S.: Quantity of wheat used for food per capita (1952—1972)

36

Dollars per Bushel

2.80

2.40

2.00

1.60

1.20

 

 

 

52 54 56 58 60 62 64 66 68 70 72

Year

Figure 3.2

U. 3.: Farm price of wheat deflated by the CPI (1952-1972)

37

Lb. per Capita

5.70

5.50

5.30

5.10

52

U. S.:

 

 
 
    

Estimated from Equation 1.5

‘(////’Reae Values

I
II... ..I.
..

  

 

54 56 58 6O 62 64 66 68 7O 72

Year
Figure 3.3

Quantity of wheat flour used in making bread per capita (1952-1972)

38

Lb’ per Capita Estimated from Equation 1.1

6.20 Real Values

5.80

4.60

4.20

3.80

3.40

 

 

 

52 54 56 56 6O 62 64 66 68 7O 72 Year

Figure 3.4

U. S.: Quantity of wheat flour used in making flour
packaged for retail sale per capita (1952—1972)

12.

12.

12.

11.

11.

11.

11.

11.

10.

10.

39

Lb. per Capita

.uu. ‘(Estimated from Equation 1.3

° ;;/Real Values

 
    

 

 

52 54 56 58 . 6O 62 64 66 68 7O 72
Year

Figure 3.5

U. S.: Quantity of wheat flour used in making cookies
and crackers per capita (1952-1972)

23

22

21

20

19

18

40

Cents per Lb,

    

 

 

Real Values
. "H. 7stimated from
. Equation 1.11
52 54 56 58 6O 62 64 66 68 70 72

Year

Figure 3.6

U. S.: Retail price of bread deflated
by the CPI {1952—1972)

66

62

58

54

50

46

42

41

Cents per 5 Lb.

  
    

Real Values

00 '00.... /
0
Estimated from

"0..” Equation 1 . 9

 

 

52 54 56 58 6O 62 64 66 68 70 72

Year
Figure 3.7

U. S.: Price of flour packaged for retail sale
deflated by the CPI (1952-1972)

42

Cents per Lb.

62 '

58

54

50

46

 

42

 

52 54 56 58 60 62 64 66 68 70 72
Year

Figure 3.8

U. S.: Price of cookies deflated by the CPI (1952-1972)

43

 

Z of 1967

110

Nominal Price
100

\£:.O.
~s.\
“5’...
90
“tr. Deflated Price
per Capita Consumption

 

 

67 68 69 7O 71 72
Year

Figure 3.9

U. S.: Bread: priceand consumption indexes (1967-1972)

44

positive influence on consumption. However, it would be interesting to
see the effects of this for the period after 1971 when fuel costs rose
rapidly.

Income and Transfer Payments: Questions
on Analysis

 

 

Multicollinearity. Income and transfer payments are included in

 

the analysis although there is a high correlation between them. The cor-
relation between food stamps and income is .70 and between transfer pay-
ments and income is .80. Also,all these variables are highly correlated
with time (YEAR), stamps (STAMPS), .66; transfer payments (TRANS), .77; and
income (Y) .97. These high correlations present some problems in de-
ciphering the relationship between the exogenous variables. Time is ex-
cluded from most equations because it provides little insight into the
structural relationships involved in the system. Yet, it is likely cor-
related with some relevant factors excluded from the analysis either be-
cause of inability to proxy them such as tastes and preferences or tech-
nology, or because of lack of knowledge about their existence.

This analysis accepts the fact of somewhat high correlation be-
tween the variables. The implications are that they are fine for pre-
diction purposes but of little use in isolating the affect of either
variable on the endogenous variables. The results are checked against
theory and against the results of Purcell and Raunikar's study conducted
in Georgia.1

Income distribution. Both income and transfer payments or food

stamps have been included in an attempt to capture the food consumption

 

1J. C. Purcell and Robert Raunikar, "Quantity-Income Elasticities

for Foods by Level of Income," Journal of Farm Economics, XLIX (1967),
1410-1414.

45

of low and higher income groups. The transfer payments (TRANS) or food
stamps (STAMPS) variable proxies income to low income groups, and average
income (Y) captures that to higher income groups. Purcell and Raunikar
use cross-sectional data to measure income elasticities for various in-
come groups. Theoretically, there should be consistency between transfer
and food stamps elasticities and Purcell and Raunikar's low income elas-
ticities, and between my income elasticities and Purcell and Raunikar's
income elasticities for higher incomes.

The problem of the excluded variable is now apparent. High cor-
relation may exist between transfer payments, a variable included in the
statistical analysis, and personal taxes, a variable not included in the
statistical analysis. If this is true, the estimated coefficient for the
transfer payment variable represents not only the increase in income to
low income people but also the decrease in income of the higher income
people or the taxpayers. This biases the coefficient.

There is some correlation between transfer payments and taxes,
but the transfer payments here are only income dependent transfers and,
therefore, are only a portion of the total transfers. As noted earlier,
in 1972, for example, these income-tested programs represented only 24.6
percent of the social welfare transfers. Other types of transfers, such
as transfers through the farm programs, are not included under social
welfare income-tested transfers. A more exhaustive analysis is needed
to analyze the possible relationships between the income-tested transfers
presented here and taxes.

Doublecounting; Doublecounting becomes a problem when both in-
come and transfer payments are considered. Income is defined as per-

sonal disposable income (income after taxes and transfer payments, but

46

before payments in kind). Therefore, transfers are actually included
twice, in the average income and as a separate variable. In this analysis
average income is used as a proxy for the income to those not receiving
transfer payments.

Income and transfer elasticities. The expectation for income

 

elasticities aggregated over all income groups is that they will be posi-
tive with a few exceptions for low preference products. In the King and
George's analysis, for example, only lard has a negative income elasticity.
For disaggregated income groups, we might expect large differences in
elasticities among the various income groups.

At very low incomes, elasticities would tend to be positive, in-
dicating that a high percentage of each additional dollar is spent on
food. There would be little substitutability between food items and ad-
ditional dollars would be spent on inexpensive generally low preference
foods.

As incomes increase and there is more substitutability among
foods, the elasticities for some foods might be negative. The higher
the income, the more substitutions there are for any product, because
the absolute price is less of a constraint, and response to income change
is likely to decrease. This decrease is less likely to occur if the
good is highly preferred. Beef is a good example.

Apalysis of Income and Transfer Payments
in the Wheat Equation

For all commodities, both the transfer payment variable (TRANS)
and the food stamp variable (STAMP) are tested. In the wheat equations

the food stamp variable produces better statistical results.

47

Results of the analysis of food stamps and income are presented
in Table 3.4. For flour, the income elasticity is rather high. A de-
crease in elasticity is expected as income rises. An increase in desire
for home-baked products is a possible explanation for increased elasticity
among high income groups. Again, a more thorough analysis of demand for
marketing services is needed.

The results for cookies are somewhat unbelievable. There is no
ready explanation for the decreasing elasticity from low income to high.
The elasticities for bread are quite consistent with other analyses and
suggest a constant income elasticity for bread in all income categories.

Qpestions on Analysis of Farm-Retail Price Differentials:
Farm-Retail Price Differentials versus Marketing Margins

 

 

The farm-retail price differentials here are not equivalent to
the "marketing-margins" or "farm-retail spreads" used by the U. S. Depart-
ment of Agriculture. Figure 3.10 below illustrates the U. S. Department
of Agriculture's concept. First, an equivalency is established between
the retail product and the average amount of farm product that goes into
the retail product.1 Because one unit of a particular farm commodity may
not be used for only one retail product, other uses of the farm product
must also be considered. For example, a by-product of butter is dry
skimmed milk powder. Therefore, the contribution of skimmed milk powder
must be subtracted from the gross farm value to obtain a net farm value
for butter. The net farm value for the particular retail product is

then subtracted from the retail price per pound of that product to arrive

 

1U. S. Department of Agriculture, Economic Research Service,
Conversion Factors and Weights and Measures for Agricultural Commodities
and their Products, Statistical Bulletin No. 362 (Washington, D. C.:
Government Printing Office, 1965).

48

TABLE 3. 4

INCOME AND FOOD STAMP ELASTICITIES

FOR WHEAT PRODUCTS

 

 

 

 

 

Packaged

Flour Cookies Bread
Income elasticity 1.72 .37 .013
t-test for income coefficient 2.22 2.18 .40
Food stamps elasticity -.036 -.013 .008
t-test for food stamp coefficient .90 3.61 3.95
High income elasticitya -3.35d b -.04
Low income elasticitya -.30d b .19
Aggregate income elasticityc .08 b .00

 

aSource: J. C. Purcell and Robert Raunikar, "Quantity-Income
Elasticities for Foods by Level of Income," Journal of Farm Economics,

XLIX (1967), 1410-1414.

 

bNo Analysis made of demand for cookies.

cSource: P. S. George and G. A. King, Consumer Demand for Food
Commodities in The United States with Projections for 1980, Giannini

 

 

Foundation Monograph No. 26, (Davis:

March, 1971), p. 51.

dIncludes cornmeal as well as white flour.

49

 

BEEF, CHOICE GRADE

Retail Price, Form Value, and Form-Retail Spread, 1970

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Retail price, per lb. Net lorm value _ Form-retail spread
98.6: ‘ use 37.1:
I
1
Choice heel '°"" 9'°"""
cattle lam enuivalent 0”” '9'” '0'". Bypraduct Net lam value
- ' ll) 05- 'i“ pet ll). of heel allowance per lb. of heel
PH“ P" ' X weight per 3 'l - = 1
live weight lb. sold at °' "'0' at retau
retail) 66.3¢ 43¢ “‘5‘
2908‘ 228 a.
A,
_ 1
Gross lam value Percentage byproduct Byproduet allowance
X allowance = a
663‘ 7"5% 4.86

 

 

 

 

 

 

 

l

Wholesale value of products per l00 lb. live weight*

 

it 1040155315 VALUE or
C‘RCASS 529.3‘.‘ BYP'OdUCls ...... 7.2% .......
avpnooucrs $2.26

/’
TOTAL ”L60. Carcass ------- 92.8% %

 

 

 

 

 

 

 

 

 

 

Source: U. S. Department of Agriculture, ERS, Farm Retail Spreads

 

for Food Products, Miscellaneous Publication No. 741 (Washington,
D. C.: Government Printing Office, 1971) p. 66.

 

Figure 3.10

Determination of U. S. Department of Agriculture's
farm-retail spreads: an example

50

at the marketing margin or farm retail spread. If other ingredients are
used to prepare the retail product, the costs of those ingredients must
be subtracted from the retail price to arrive at the marketing costs.

In this analysis the marketing margins are not estimated directly
but emerge from the estimations of the price at both the retail and farm
level. By estimating the by-product allowance and delineating other in-
gredients into the retail product, the marketing margin can be determined.

The farm-retail price differential, not the spread or the margin,
is estimated here. The differential is merely the difference between the
farm and retail prices of the farm product per retail quantity, after
the farm price has been converted by determining the amount of the farm
commodity that is an input into the retail product. The farm-retail
price differentials measure the amount of the retail price that goes to

marketers.

Farm-Retail Price Differentials for Wheat

The farm-retail price differential has decreased for flour and
cookies. This means an absolute decrease in deflated dollars to inputs
other than wheat. The farm-retail spread, as calculated by the U. S.
Department of Agriculture, has also decreased. For bread, the retail
price differential and the farm-retail spread have increased. At the
same time, the percent of the retail product going to the farmer for all
three wheat products has decreased (Table 3.5). Thus, whether the farm-
retail price differential and retail price increase or decrease, the per-
cent going to the farmer tends to decrease relative to the suppliers of

other inputs into the retail process.

51

TABLE 3.5

WHEAT: FARM-RETAIL SPREADS AND PERCENT OF RETAIL
PRICE GOING TO FARMER (1950-1970)a

 

 

White bread Cookies Flour
Percent of retail price going
to farmer
1950 39 4.2 39
1960 35 3.4 35
1970 35 3.2 35
Farm-retail spreadb
1950 15.6 c 41.4
1960 19.0 51.8 39.6
1970 17.8 39.6 32.8

 

aSource: U. S. Department of Agriculture, ERS, Farm-Retail
Spreads for Food Products, Miscellaneous publication No. 741 (Washington,
D. C.: Government Printing Office, January 1972).

 

bFarm-retail spread as defined by U. S. Department of Agriculture
and deflated by CPI (19678100).

cData not available.

52

Farm Level Demand

At the farm level, the price of bread, cookies and fuel are es-
pecially important in determining how much wheat goes into food (Equation
1.7). The price of bread, as expected, is the most important factor.
It has a strong negative impact, but the price of cookies has a positive
impact. This is difficult to justify theoretically.

Labor costs do not have an impact on the quantity of wheat used
for food. Fuel prices have a very small impact (Equation 1.7).

There is no simultaneity in determining the price or quantity

of wheat that goes into food. Farmers respond to the price of the pre-

vious year.
Milk

Modeling the Dairy Sector

The dairy industry in the United States is heavily regulated.
Milk is graded for manufacturing and fluid use and prices are regulated
by both federal and state governments. The grade of milk is determined
by the sanitary conditions under which it is produced. There has been
a large shift in production from manufactured grade to higher standard
fluid grade. Presently, only about one-fourth of the milk sold by farmers
is sold as manufacturing milk. Regulation of prices by federal and state
governments has increased from the late 1940s when 50 percent of the

fluid grade milk was regulated to presently when 96 percent is regulated.1

 

1U. S. Department of Agriculture, Economic Research Service, Milk
Pricing, by Alden Manchester, Agricultural Economics Report No. 315
(Washington, D. C.: Government Printing Office, 1975), pp. 9-10.

60

56

52

48

44

40

53

Cents per Lb.

 

 

 

52 54 56 58 6O 62 64 66 68 70 72
Year
Figure 3.11

U. S.: Farm-retail price differential for cookies
deflated by the CPI (1952-1972)

20

10

18

17

16

54

Cents per Lb.

+—Real Values
"% Estimated from

   
  

Equation 1.8 and 1.11

 

 

52 54 56 58 60 62 64 66 68 70 72
Year
Figure 3.12

U. S.: Farm-retail price differentials for bread
deflated by the CPI (1952-1972)

46

44

42

40

38

36

34

32

30

U.

55

Cents per 5 Lb.

 

   

Real Values
Estimated from

 

52

S.:

54 56 58 60 62 64 66 68 70 72
Year

Figure 3.13

Farm-retail price differentials for flour
packaged for retail sale deflated by the CPI (1952-1972)

56

As with wheat, the major retail products are used in the estima—
tion. Table 3.6 illustrates major uses of milk. About 50 percent is
used for fluid milk production and 40 percent for manufactured products
such as butter and cheese. These products are considered the major re—
tail products to be used in analysis. As evident from Item 2 in Table
3.6 virtually all the milk consumed in the United States was produced in
the United States that year; little is stored, exported or imported.

It is difficult to determine what price to use in the estimation
procedure which proxies the price set by government policy. First, it
is supposedly exogenously determined. But this is not completely true
because the price is set to represent what would be a longer-run market
price, i.e., an endogenously determined variable.

Second, milk prices are regionally determined. Therefore an
analysis of the milk market by state or region may present a more precise
picture of the dairy sector. In this model a lagged price on the farm
and lagged quantity of milk produced are used to proxy the government
policy price. Farm price is tested for simultaneous determination with
quantity because policy price and farm price are not always the same.
The policy price only determines the lower limit.

Supply was tested for simultaneous determination with demand by
including the most significant variables from previous independent estima-
tions of the farm supply equation. Those variables were corn, price of
non-fed beef and private wage.

Data: Correlation Matrix, Sources of Data
and Adequacy of Proxies
The quantities of cheese, butter and fluid milk used were milk

equivalents of the quantities. The estimates of the products can be

57

TABLE 3.6

U. S. CONSUMPTION OF MILK AND DAIRY
PRODUCTS (1950-1970)

 

 

 

 

 

 

 

 

1950 1960 1970
1. Milk8 used for domestic consumption
(millions of pounds) 117,358 123,102 117,493
(per capita pounds) 663 691 582
2. Percent of milkb used for domestic
consumption that was produced in
the U. S. that year. 99.35 100.005 99.54
3. Percent of milkb used for:
manufacturing 47 48 51
fluid 36 43 44
farm use 15 7 3
residual 2 2 2
100 100 100
4. Percent of milkb used in various
manufactured products as per-
centages of total consumption
in U. S.:
butter 23 23 20
cheese 10 ll 17
canned milk 5 4 2
dry whole milk .8 .5 .4
ice cream
other manufactured products 3.2 2.5 .6
milk used for manufacturing 47 48 51

 

aSource: U. S. Department of Agriculture, Agricultural Statistics,

 

(Washington, D. C.: Government Printing Office, 1972).

bComputed from above source.

58

improved by using actual product quantities.

The retail price used for cheese is the price of American cheese.
However, there has been a shift in demand for more specialty cheeses and
the price of American cheese may not represent this trend.

The price of milk at the farm used in the model is an aggregate
price (the price of milk sold to plants and dealers). A disaggregated
price series is also available for fluid milk and milk used for manufac-
turing. The manufacturing grade can be disaggregated further to milk
used for butter, cheese, and other manufactured products. Because 50
percent of the milk is used for fluid, this blend price may reflect milk
price fluctuations for the fluid milk more than the price of milk for
other uses.

Fluid milk costs more than manufactured milk; thus, the aggregate
price of milk sold to plants and dealers used in this model is higher
than that used for manufacturing. In 1973, for example, the price in

dollars per 100 pound of various grades of milk was as follows:

Eligible for fluid milk $ 7.42
Of manufacturing grade ' 6.20
Milk for butter 6.10
Milk for all cheese 6.35
Milk for evaporated milk 5.64
All milk wholesale1 7.14

Further disaggregation is needed to determine marketing margins.
The differential between the price of all milk for wholesale and the price
of manufactured grade is large enough so that if the price of all milk is
used as a proxy for manufactured milk, the wholesale price of the milk

used in making butter is higher than the retail price of the butter-~an

 

lThis blend price is used in the estimation.

59

TABLE 3.7
THE DAIRY MODEL

 

System of equations for milk

 

Retail demand:
Retail supply:
Retail demand:
Retail supply:
Retail demand:
Retail supply:
Farm demand:

Farm supply:

Endoggnous variables

 

QFLUID
PFLUID
PMILKF
QCHEESE
PCHEESE
QBUTTER
PBUTTER
POLICYPRICE

Exogenous variables 2

 

TRANS

Y

PBURGER
PMARG
PBREAD
WAGEFME
PFUEL
QMILKFMLAG
QMILKFARM

(QFLUID,
(QFLUID,

(QCHEESE,
(QCHEESE,
(QBUTTER,
(QBUTTER,

(PMILKF,
(QMILKF,

quantity
price of
price of
quantity
price of
quantity
price of

PFLUID, Z) = 0
PFLUID, PMILKF, Z) = o

PCHEESE, 2) = o
PCHEESE, PMILKF, 2) = 0
PBUTTER, 2) = o
PBUTTER, PMILKF, Z) = o

PBUTTER, PFLUID, PCHEESE, QMILKF, 2) = o

 

POLICYPRICE, 2) = o

of fluid milk at retail level
fluid milk at retail level
milk received by farmers

of cheese at retail level
cheese at retail level

of butter at retail level
butter at retail level

policy price for farm level milk

transfer

payments

disposable personal income

price of
price of
price of

hamburger meat
margarine
bread

wage of food manufacturing employees

price of
quantity
quantity

fuel
of milk produced in time t-l
of milk produced in time t

60

unlikely conclusion.

Simple correlations for all variables used in the dairy equations
are presented in Table 3.8. As with the wheat equations, there is high
correlation between transfer payments (TRANS) and average income (Y). In
addition, the index price of fuel is highly correlated with many of the

relevant milk variables. This causes difficulty in analysis.

Analysis of Regression Results

Quantity and Price

 

There have been some major changes in dairy product usage in the

United States. Milk use both for fluid milk and manufactured products
has decreased from 663 pound per capita in 1950 to 582 pound in 1970.

Figures 3.14, 3.15 and 3.16 illustrate some of the per capita changes in

consumption trends. Use of fluid milk and cream per capita as well as
butter has decreased considerably while that of cheese has increased
(Table 3.6). The total amount of milk consumed has declined by 135 mil-
lion pound or by .11 percent. This represents a per capita consumption

decline of 12 percent (see Figure 3.17).

Deflated prices of butter, cheese and fluid milk are illustrated
in Figure 3.22, 3.23 and 3.24. All prices declined steadily except that

of cheese which begins to increase after 1963. Figure 3.14, 3.15 and

3.16 compare prices both deflated and undeflated with quantities of the
various retail products consumed. As illustrated in Table 3.8, there is
a positive correlation between deflated prices and quantities consumed.
With cheese the correlation is .42; with butter .48; and with fluid milk
.73. This is, of course, the opposite of what is expected. A negative
correlation is obtained only when undeflated prices are used (except for

cheese). Again there is a positive correlation of .82 with cheese, but

mmHBmHum> HHm now mwousom mumv cam GOHuwammem umsuusm

.m waawam< :H venom on cmo

.muHamo “on mam vmumemo mum w mam mz<ma

.Aooauecaav

Hmo mnu ma wmumHmww sown m>ms mumHHoo CH commoumxm meanum> HHm mam muHamo you mum mmHuHucmsv HH<m

 

H

61

YEAR

«m. Hm.
H «a.
H

WAGE

mm.
mm.
mm.

TRANS

00.!

mm.|

ow.l

mm.1
H

QMILKFARMLAG

mm.n

mo.l

00.:

om.l

mm.
H

QMILKFARM

qw.l Hm. om.l om.| HH.I q¢.I qw.l
mw.l ow. mw.l om.l HH.¢ wH.I om.l
mm.l Om. mm.l cm.n «H.I m~.I wo.l
Hm.| mm. mm.l ow.l 50. Ho. mm.l
mm. mm.l mm. cm. Ho.l mm. mm.
mm. No.1 mm. «m. mo.l mm. mm.
H mm.1 mm. mm. mo.l mo. mu.
H 00.: mo.l o¢.I HH.| mo.|
H mm. MH.I OH.I «0.
H OH. Hm. cm.
H me. mm.

H mm.
H

c m

E R R A
D co pa "U “r nu
T. E T nu K O
m m m R m u a
P; no nu mm M“ M
n4 n4 ms 6. n. or p.

«q.l
Na.
wq.l
00.
«N.
mm.
mo.

PCHEESE

mw.l
om.l
«5.1
nm.|
on.

mm.
mm.
um.

PBUTTER

mo.l

OH.I
NH.I
mm.l
we.

Hm.l
wN.I

PBREAD

aw.l
mn.l
mm.l
ow.l
mm.
em.
Ho.
50.:
mm.
om.
«0.1
HH.
ow.
mm.l
No.
mm.
H

FUEL

m<mw

»

m2mmo<3
mzama
o<qzm<mxaHzo
zm<mMaHzc
anqmo
mmmmmoo
mmayamo
amaze
omamz<mm
zm<mmaHzm
oomozm
mmmmmom
mmaaomm

Q<mmmm
Hmsm

 

mmone<aom me<n may 2H new:
mmam<Hm<> mom mee<z monH<ammmoo mqmsz

w.m mgm<H

62

 

 

 

 

 

 

Aam.mv AmH.wv Aea.Hv a :
wmsmzlzm n<mmmm mm.m + mzama ma.m . amaze oq.H + u up a
. . oaqzmxaHzo unnlln. .||||. .1111. you .
on N on swam eqmame a N
mzmmo<3 “NH.mHV ANm.V assume
.w Hm.om u mmaaamm om.a u hm.qo~ u mmaypmo
mz<ma
HHN.V Aoo.v
o<azmxasz amps NC. I mammoaz as. + mmmmgu «o
-.H ma. Aaaaam m.~
omDNMMWM Amm.v Aso.Hv Hfimumm
nammmm zm<mzm cm H u mmmmmom as m + no «a u mmmmmoo
Am~.v Aso.mV AH~.V
mzama Ho. n.w -.a + om<zm o~.H +
H.~
mo.~ Hm. sausages Amo.v Aqq.v Aao.v mmmmgu mo ~.~
aoum mmomsmm Ho. - nammmm as. u gene so. u muaum
mac
Amo.v Amo.v An~.H~V
mzmmo<3 no. . o<azmmaHzo co. u aa.no u mmmmmom
uaqzemgHzo Asm.av Aqm.av
amam um=m2<mm on. ..W mm.m + mmmmgu you
o~.~ em. mzmmo<3 uamamc H.~
amaze Amm.v ahfl.v Hfimumm
nammmm mz<ma ca.m + mmmmmom me. + ne.qn u mmmmmoo
39 NM wwmwmwmmm mud—mom / oHuHH condo—um

 

mmZOHH<DOm MMH<D mom mHHDmmm onmmmmumm

m .m mafia.

63

 

 

 

 

 

 

 

Ama.v Anm.mv Am~.v
mzama so. +.m na.m u amaze a“. +
2 so; a: 83 ham. 3
-.N No. aofiumsum mmumsmm so. + n<mmmm ea. . swam mN. +
aoum -illlltu IIIIII IIII. mo onum
mac Amo.v x~o.v Ama.muv
mzmmoas mo. + u<g=mmgHzo no. . aN.ms u oomozm
u<qzmxaHzo .
Amps mz<maamm.mw . saws
. . mzmmuaz .nnnn. waaam pom .
ma mm amaze Ams.av haw.v uamame A N
ompmz<mm . u . . u Hamumm
nammmm mpoq SH oomozm so + OH Nam angmo
mz<me
Axo.v Aq~.v
zm<mMasz gmnm mm. + mzmmoaz mm. . umuusw
o~.H om. .Inn. .IIIIII. m o.~
om<zm Ako.~v Aoc.av hflaaam
umsmz<mm . . . Hfimuom
n<mmmm zm<mzm oh m u mmaasmm cm a + mN ama u awesome
Amm.av “Hm.ov “H¢.V
mz<me a~.m n.w -.m_+ amaze m¢.~ +
q.~
w~.m om. aofiumsum Aao.v Anm.v Ana.v “muss; mo m.~
aoum «momamm ma. nammmm mm.~ u game an. + mowum
mac
Amo.v Amo.v Am~.oqv
mzmmo<3 0H. u o<qzquHzo mH. u mq.HmH u mmaasmm
n msoamwoxm muHsmom oHUHH coHumavm
3 ~m amusauxm

 

Au.ucoov m.m mqm<a

.mmHanum> mooamwoxm onu oumoncH mmHanum> onu uo>o wumm .muouuw uumvnwum on» momoHodo muamHonmmou «no
nouns mHmonucoumm .AOOOH n momHv Hmo onu up wouMHmmv soon o>m£ moHQMHum> mUHNQ HHm paw wuHamu you mum
moHAMHum> huHuamsu HH< .m vaaman< :H vdsom on emu mama HHm mo muusom mam aoHumHuommv ouonaou <m

 

 

 

64

 

 

Aqm.oV Haw.va HNN.HV
mz<ma oH.H +.m mo.NoH . amaze mm.q +
. . Amq.v Aso.ev Hmw.v HHHa No .
mN N am mHo anemone mq.H u panama No.a n Hmam Nm.H + AuHuamno HH N
HNN.V HmN.v HNN.HNHV
mammo<3 NH. + o<HzmxHHzc N0. + NN.NNN u zmMHHzo
xNH.v Aeo.v
mz<ma Hmam HN u manage; Ho +
N xHHa you
HHo.v ANm.v
sq.H No. omazm . n . \ census OH.N
umamz<mm zm<HMHHzo moo ummmmom N6 + aumm
alumna AHN.V 14¢.v
mmeaamm He. + oomozm mm. + NN.NN- u zm<mzm
mz<ma
Hom.v Hmo.v
ulemeHHzo Hana 0N.H +.mzmmo<z NH. . xHHa
ms.H Nm. N uHaHN a.
om<2m . . No AHaaam N
n<mmmm zm<mzm an oomozm Nm + on moN u anHm
msoamwoxm
3n um vovsHoxm muHamoM oHuHa aoHumsvu

 

an.ucoov a.m NHH<H

65

TABLE 3.10

PRICE, INCOME AND TRANSFER ELASTICITIES
FOR DAIRY PRODUCTS

 

 

 

 

 

 

Fluid milk Cheese Butter

Price elasticity .12 .32 —.82
t value on price coefficient .82 4.00 1.95
King and George's price

elasticity3 -.34 —.46 -.62
Income elasticity -.13 .14 -.87
t value on income coefficient 2.19 3.47 3.10
Transfer elasticity -.79 1.28 -.87
t value on transfer coefficient 3.52 6.44 .38
Low income elasticityC .35 b b
High income elasticity .35 b b
Aggregate income elasticitya .203 .24 .318

 

aSource: P. S. George and G. A. King, Consumer Demand for Food
Commodities in the U. S. with Projection for 1980, Giannini Foundation
Monograph No. 26 (Davis: March, 1971).

bData not available.

 

cSource: J. C. Purcell and Robert Raunikar, "Quantity Income
Elasticities for Foods by Level of Income," Journal of Farm Economics,
XLIX (1967), 1410-1414.

 

124

116

108

100

92

84

66

 

      
 

Z of 1952

Undeflated

I Price

I
4’
I
/
1'
[/7
\‘J
$’0‘.‘g40IIssues-lost".....‘o.
v "0... Quantity per Capita
%’ ” Deflated

Price

 

 

52 54 56 58 60 62 64 66 68 70
Year
Figure 3.14

U. S.: Fluid milk: price and consumption
indexes (1952-1970)

67

 

 

 

 

Z of 1952
Undeflated
140 ’ Price
I
I Quantity
130 Consumed
1' per Capita
1'
120 1'
I!
110 I'
l ,. Deflated
’ g...
100 ’4 run-fl L Price
x , f
.\’ -‘-./l u. 9..
90 .’. “\jouunuobull fi. "in..."".o|“
r‘.. .....
0v
80
52 54 56 58 6O 62 64 66 68 70
Year
Figure 3.15
U. S.: Cheeseprice and consumption indexes (1952-1970)

124

116

108

100

92

84

76

68

68

 

 

Z of 1952
Undeflated
“.Price
\
‘1‘ .~
0‘.‘
\
¥—-‘

Quantity
Consumed
per Capita

 

“‘~ Deflated Price

 

 

52 54 56 58 60 62 64 66 68 70
Year

Figure 3.16

U. S.: Butter: Price and Consumption Indexes (1952-1970)

69

Lb. per Capita

760

720

680

   

Estimated from

64
0 l Equation 2 . 11

600

560

520

 

 

52 54 56 58 6O 62 64 66 68 70

Year

Figure 3.17

U. S.: Quantity of milk produced per capita (1952-1970)

62

58

54

50

46

42

U.

70

Dollars per 1,000 Lb.

 

  
  
 

Real Values

Equation 2.10

 

S.:

52 54 56 58 60 62 64 66 68 70
Year
Figure 3.18

Farm price of milk deflated by the CPI (1952-1970)

71

Lb. per Capita

         
   

305 .
0....”

295

Real Values
285

Estimated from

275 Equation 2.7
265
255

 

 

52 54 56 58 6O 62 64 66 68 70

Year
Figure 3.19

U. S.: Quantity of milk used for making fluid
milk per capita (1952-1970)

72

Lb. per Capita

     

100

96

92 .. 'Real Values

88 ‘FF‘F‘Estimated from

Equation 2.1

84

80

76

 

 

52 54 56 58 6O 62 64 66 68 70

Year

Figure 3.20

U. S.: Quantity of milk used for making cheese
per capita (1952-1970)

73

Lb. per Capita

180

170

  
   
 

160
Real Values

150

140
stimated from

Equation 2.4
130

120

 

 

52 54 56 58 60 62 64 66 68 70

Year

Figure 3.21

U. S.: Quantity of milk used for making butter
per capita (1952—1970)

58

56

54

52

50

48

46

74

Cents per Half Gallon

 

Estimated from
Equation 2.8

 

 

52 54 56 58 60 62 64 66 68 70

Year

Figure 3.22

U. 8.: Retail price of fluid milk deflated
by the CPI (1952-1970)

47

45

43

41

39

37

35

75

Cents per Half Lb.

  
 

0
' Estimated from
Equation 2.2

 

 

52 54 56 58 60 62 64 66 68 '70

Year
Figure 3.23

U. S.: Price of cheese deflated by the CPI (1952—1970)

76

Cents per Lb.

100

96

92

    

88 Real Values

Estimated from

84 ‘ {Equation 2. 5

80

76

 

 

52 54 56 58 60 62 64 66 68 70

Year

Figure 3.24

U. S.: Price of butter deflated by the CPI (1952-1970)

77

it is -.85 and -.90 with butter and fluid milk, respectively.

Statistical analysis should ferret out the effect of the other
variables as well as own price on consumption. Equations 2.1, 2.4 and
2.7 in Table 3.9 are the retail demand equations for fluid milk, cheese
and butter. In Table 3.10 my own computed income and price elasticities
are compared with those of others.

Own price coefficients for cheese and butter are significant. The
coefficient for milk is not. Price has a large impact on the amount of
butter consumed and a much smaller impact on quantities of fluid milk and
cheese consumed.

Price elasticity signs for fluid milk and cheese are positive,
indicating an increase in consumption with an increase in price, a result
not justified by the theory. The coefficient for fluid milk is insignifi—
cant and thus can be excused. The insignificance of the sign argues for
the lack of responsiveness of milk to its own price. This is a justifiable
conclusion. However, the coefficient for cheese is highly significant with
a t value of 4 (see Table 3.9). The price of cheese must be highly cor-
related with another possible explanatory but excluded variable to sug-
gest that variable is actually the cause of this unusual result. A
change in tastes and preferences is unlikely the reason for the positive
price elasticity because the price of cheese is not highly correlated
with time (YEAR); in fact, the correlation is very low, .06.

The results suggest two possibilities: 1) consumers will increase
their consumption of cheese if the price increases; and 2) the data used
could be the source of the problem. The price of cheese more than likely
is relatively accurate. The prices of cheese, butter, fluid milk and

milk at the farm are highly correlated (Table 3.8) indicating there are

78

no real absurdities in those particular data sets. But the price of
cheese does represent only one variety of cheese, American cheese, whereas
the quantity used represents all cheese varieties.

If American cheese substitutes well for other varieties, its
price is probably a good proxy because the price of all cheese varieties
should rise and fall together. But if they are not substitutes, the
price of American cheese is not a good proxy and could upset the analysis.
Further work must be done in order to clarify whether or not the vari-
eties of cheese are highly substitutable for each other.

No statistically significant substitutes or complements were
found for any milk products. Even margarine as a substitute for butter
has a t value of only .74. This suggests it is not the price of margarine
which makes it such a good substitute but a change in tastes and prefer—

ences as indicated by the high t value for income (Y), 3.30.

Income and Transfer Payments

From Table 3.10 we can see that both fluid milk and butter have
negative income and transfer payment elasticities. The elasticities for
cheese are what might be expected. The elasticity is high at low income
and decreases rather abruptly at higher incomes. This suggests cheese
is a luxury food at low income levels, but relatively unresponsive to
income changes at higher incomes.

The elasticities are negative and constant for butter. The large
negative income elasticity is likely a result of changing tastes and
preferences, since margarine and butter are by common sense highly sub—
stitutable for each other, yet this is not indicated in the analysis.

The elasticity for mdlk is negative and increasing for higher in-

comes, i.e., there is less of a negative impact on consumption with

79

income change. This result, too, could be explained away by suggesting
a change in tastes and preferences but there is nothing to indicate that

in the analysis.

Farm-Retail Price Differentials

 

Farm-retail price differentials can be estimated only for fluid
milk. The reasons were explained thoroughly in the section on data but
the main thrust is that the farm level price estimated is an aggregate
and of all retail products from milk; only fluid milk has a per unit re-
tail price higher than the per unit farm level price. But even the re-
sults for milk are unsatisfactory due to the wide divergence of estimate
farm-retail price differentials from the actual price (Figure 3.25).

Table 3.11 shows that the share of the retail price going to the
farmer has declined. At the same time the farm-retail spreads have
fluctuated for the various retail products. From 1950—1960 the percentage
of the retail price for fluid milk going to the farmer decreased while
the farm-retail spread increased. The percentage going to the farmer in-
creased from 1960-1970, but the farm—retail spread decreased.

Analysis suggests the price of fuel has affected the quantity of
fluid milk supplied at the retail level (Equation 2.9, Table 3.9). The
high correlation between retail price of fluid milk (PMCROC) and price
of fuel (FUEL), .80, suggests multicollinearity and inability to separate
the effect of either variable. It is likely that the price of milk rather
than fuel (because fuel should theoretically have a negative sign and it
has a positive one) is responsible for the high correlation. Wage has
some impact on the amount of fluid milk supplied and thus on margins.

In addition, price of fuel also affects the price of milk at the farm

(PMFARM).

36

32

28

24

20

16

 

80

Cents per Half Gallon

Real Values

  
 

.,.°’ Estimated from
‘. “.0" Equation
.....o‘ 2.8 and 2.10

 

52 54 56 58 6O 62 64 66 68 70

Year
Figure 3.25

U. S.: Farm-retail price differential for fluid
milk deflated by the CPI (1952-1970)

81

TABLE 3.11

DAIRY PRODUCTS: FARM-RETAIL SPREADS AND PERCENT OF
RETAIL PRICE GOING TO FARMER (1950-1970)a

 

 

 

Fluid milk Cheese Butter
Farm-retail spreadsb
1950 21.9c 21.0 22.9
1960 28.3 21.3 23.0
1970 24.5 24.1 "21.0
Percent of retail price
going to farmer
1950 55 47 77
1960 47 44 77
1970 50 44 71

 

aSource: U. S. Department of Agriculture, ERS, Farm-Retail Spread
for Food Products, Miscellaneous Publication No. 741 (Washington, D. C.:
Government Printing Office, 1971).

 

bDeflated by CPI (1967=100).

cPer half gallon.

82

With cheese, however, the amount as well as the percentage going
to the marketers have increased. At the same time the cheese consumption
has increased greatly and cheese prices have risen despite the correspond-
ing decline in milk prices at the farm.

As with fluid milk, the farm-retail spread for butter increased
in 1960 and decreased in 1970. At the same time, percent of the retail
price going to the farmer decreased (Table 3.11). The postulated shift
in tastes and preferences away from butter, especially after 1960, and
the subsequent shift in demand for butter caused its decline in retail
price. Thus, the percentage going to the farmer has also declined. The

farmer seems to have absorbed the impact more than the intermediaries.

Farm Level Demand

 

Cheese has the highest coefficient in the farm level demand equa-
tion, Equation 2.10 and may explain changes in milk price at the farm
more than any other retail product made from milk. The price of fuel
(FUEL) and the retail price of milk (PNGROC) are highly correlated.

Fuel is perhaps picking up some of the effect of fluid milk at the farm
level explaining why fluid milk seems to have a smaller impact on the
farm level demand for milk.

The amount of milk produced does not effect its price nor are
the two simultaneously determined (Equation 2.10). Neither fuel nor

labor costs have any impact on the price of milk.

Fed and Non—Fed Beef

 

Modeling the Beef Sectors
Fed and non-fed beef were treated as separate commodities and,

therefore analyzed separately. Both fed and non-fed were tested for

83

TABLE 3.12
THE FED BEEF MODEL

 

System of equations for fed beef

 

Retail demand: (QSTEAK, PSTEAK, 2) = 0
Retail supply: (QSTEAK, PSTEAK, PFEDBF, 2) = 0
Retail demand: (QHAMBURG, PHAMBURG, Z) = 0
Retail supply: (QHAMBURG, PHAMBURG, PFEDBF, Z) = o
. Retail demand: (QCHUCK, PCHUCK, Z) = 0
Retail supply: (QCHUCK, PCHUCK, PFEDBF, 2) = 0
Farm demand: (QFEDBFPR, PFEDBF, PCHUCK, PHAMBURG, PSTEAK, i) = 0
Farm supply: (QFEDBFPR, PFEDBF, Z) = o

Endogenous variables

 

QFEDBFPR quantity of steers and heifers slaughtered
PFEDBF price of fed-beef received by farmers
PCHUCK price of chuck meat

PHAMBURG price of hamburger meat

QSTEAK quantity of steaks, roasts and chops
BURGPLUS quantity of hamburger and ground beef

Exogenous variables 2

 

WAGEFME wage per hour of food manufacturing employees

PHOGS price of hogs received by farmers

PBROIL price of broilers received by farmers

QFEDBFPR quantity of steers and heifers slaughtered

QNONFED quantity of beef slaughtered other than steers
and heifers

QHOGSPR quantity of hogs slaughtered

QBROILERPR quantity of broilers produced

Y disposable personal income

TRANS income dependent transfer programs

PFRYING price of frying chickens

PHAM price of ham

84

TABLE 3.13

THE NON-FED BEEF MODEL

 

System of equations for non-fed beef

 

Retail demand:
Retail supply:
Retail demand:
Retail supply:
Farm demand:

Farm supply:

Endogenous variables

 

PNONFED
PHAMBURG
PROUND
QNONFED

BURGPLUS
PCHUCK

Exogenous variables, 2

 

QNONFED

PmmL
QFEDBEEF
QHOGSPR
QBROILPR
WAGEFME
PBROIL
PHAM
PFRYING
Y

TRANS

(QHAMBURG, PHAMBURG, Z) = 0

(QHAMBURG, PHAMBURG, PNONFED, 2) = o
(QCHUCK, PCHUCK, 2) = )

(QCHUCK, PCHUCK, PNONFED, 2) = o

(QNONFED, PNONFED, PHAMBURG, PCHUCK, Z) = 0
(QNONFED, PNONFED, 2) = 0

price received by farmers for other than steers
and heifers

price of hamburger

price of round steak

quantity of beef other than steers and heifers
slaughtered

quantity of hamburger and ground beef sold

price of chuck

quantity of beef other than steers and heifers
slaughtered

price of fuel

quantity of steers and heifers slaughtered

quantity of hogs slaughtered

quantity of broilers produced

wage of food manufacturing employees

price of broilers received by farmers

price of ham

price of fryers

disposable personal income

income dependent transfer payments

85

lagged response to price. If farm supply is determined by expected price,

let us say the price of the previous year, the model changes so that the

farm demand equation becomes:

P fed beef = F (P chuck’ P roast’ Q fed beef’ Z)
For this model the quantities of other possible substitutes produced are

used instead of prices.

As with other products it is difficult to obtain retail data for
the various cuts; consequently, only one retail product was used as the
retail equivalent for the farm product.

Data: Correlation Matrix, Sources of Data and
Adequacy of Proxies

The data used to represent fed-beef includes the number of steers
and heifers slaughtered. Non-fed is a residual or the difference between
total slaughter and steer and heifer slaughter.

There is little information on retail level beef products. The
Bureau of Labor has collected a series of prices of retail beef cuts.
The U. S. Department of Agriculture publishes a series of uncanned meat
products. From these series the quantity of steaks, roasts and chops
produced is used to represent one of the retail products from fed-beef.
A combination of hamburger and ground beef is used to represent one of
the major retail products of non-fed beef. Two problems are encountered
when these data are used.

First, the category steak does not include only fed-beef. There
is a growing trend towards steaks from non-fed beef. Fast food chains
are the major procurers of these steaks and an industry has grown around
the processing of meat for these chains. It would have been more appro-

priate to account for the various cuts of fed—beef but the data on

86

quantities are not available. Much of the cutting is done at the butchers
and there is little way of measuring what happens to the carcasses at that
retail level.

Second, in the data series on retail cuts no information on the
amount of the various types of meat going into each category of cut is
available. For example, hamburger meat is usually a combination of pork
and beef, but there are no data on the amount of actual beef in hamburgers.

The correlation matrix for livestock and poultry is presented in
Table 3.14. A difficulty in the analysis arises because the quantity of
fed-beef produced (QFEDBFPR) is so highly correlated with time (YEAR) and

thus with income (Y), .98.

Analysis of Regression Results

Quantity and Price

 

Over the time period studied, 1953 to 1971, the quantity of fed-
beef produced increased by 83 percent while the quantity of non-fed beef

decreased by 25 percent (Figure 3.26). Data from the Livestock and

 

Meat Statistics1 show that the amount of hamburger and ground beef has
increased by 390 percent and the amount of steaks, chaps and roasts has
increased by 726 percent over the same time period.

Over the same period, the price of fed-beef increased by 9 per-
cent and that of non-fed increased by 25 percent. The price of hamburger
and round declined over the same period--hamburger by one percent and

round by two percent.

 

1U. S. Department of Agriculture, Economic Research Service,
Livestock and Meat Statistics, Statistical Bulletin No. 522 (Washington,
D. C.: Government Printing Office, 1973).

87

.0 xHuconm< cH 06:00 on can meanum> HHw now moousom muev 0cm coHumcmHaxo uwnuusm

.kumH000 vcm muHauo

 

awn— OHQ V Hue—m mghm .mgu. .AOOHINOOHV HAD 050 N05 vOuwwav C000 N>M£ mthHOO CH vommOMQXU mwdnwwhmei HAG Ucm fiuwawu hum Ohm. wwwuwucmav HH<Q
H N0. 00. on. N0. 0N. 00.: N0.: «0.: 00. H0.: 00.: 00. 0H.: H0. 00.: 00.: 0a. 0N.: 00.: H~.: 0H. 00.: 00.: 0~.: H0.: H0. ~0.: N0. m<0>
H 00. N0. 00. mm. 0H.: n~.: 00.: 00. 00.: ms.: 00. Hm.: me. m~.: 00.: 0H. 0H.: «0.: 0~.: 00. n0.: 0H.: H0.: H0.: N0. n0.: Ho. r
H 05. mm. Nu. 00.: n~.: mm.: 00. 55.: nu.: N0. 0a.: Nn. H0.: H0.: HN. NH.: «0.: 0H.: 0H. 00.: H0.: BN.: 00.: 00. 00.: N0. mzmmo<3
H 00. N0. 0~.: -.: 00.: H0. 00.: 00.: 0h. 0H.: 00. NH.: 00.: NH. 00.: 00.: H0.: 0N. Hn.: 00.: 0a.: H0.: 0N. NN.: 0o. mz<¢H
H 00. NH.: 0H.: 0H.: «0. 00.: 00.: mm. -.: mm. 00.: 00.: HH. 00.: 00.: 00.: 0H. 00.: 50.: 0~.: 00.: NH. 00.: m0. mmz<em
H HH.: 0H.: 00.: mm. 00.: 00.: 00. 00.: 00. 00.: n~.: NH. 00.: 00.: 00.: 0H. 00.: ~0.: 00.: 00.: 0h. 0N.: 00. x<mhmo
H 0m. 00. 00.: 00.: 00. 00.: 00.: N0.: 00. 00. 0N.: mm. 00. 0H.: 00.: 00. an. Hm.: 50. 00.: 0H. 0H.: 0002020
H 00.: 00.: H0. 50. 00.: 00.: 00.: «0. 00. «0.: 00. 00. ~0.: 0N.: N0. 00. 0~.: H0. No. 00. 0~.: 0<Hmmzo
H 00.: 00. NN. 50.: NH. -.: 00.: NH.: HH.: 00. 0N. 00. 00.: Ha. 0H. 0H. 00. NH.: 00. «0.: 000:0
H 00.: 00.: 0o. 0H.: Ne. 00.: 00.: 0H. 0n.: H0.: 0H.: N0. 00.: 00.: 00.: 00.: 00. 00.: 00. 00000000
H . 00. ~0.: 00. 00.: N0. 00. 00.: HN.: 00. HH. 0H.: 00. 0s. 0N. 00. NH.: ON. 00.: 0<Hmmmoomo
H Ho.: N0.: 00.: 00. 00. 00.: 00. N0. 0H. 0H.: 00. 05. 00. 00. 0H.: 00. 00.: 0000000
H 00.: 00. H0.: 00.: «0. 0H.: 00.: 0~.: mm. «0.: 00.: 0H.: «0.: 00. 0H.: 00. meHommo
H 00. 00.: 00. 00. 00. 00.: mm. 00. 0H. 00. 00. H0.: Hn.: 0H. 00.: 020000
H NH.: 0a.: 00. 00.: 00.: «H. 00. 00.: Hn.: mm. 00.: 0H. 00.: 00. amazozm
H N0. 0H.: 00. mm. 00. 00. N0. 00. Ho. 00. 00.: No. 0H.: U<Hmoozm
. H 0H.: 05. N0. 00. 0H.: 00. H0. 00. 00. H0.: 00. 00.: 000:0
H 0H.: H¢.: 00. 00. 0H.: 0~.: 00. 50.: 00.: NN.: 0H. 0000:400
H N0. Hm. 00.: «0. 00. eu. N0. «0.: 00. 00.: z<:m
H ~N. 00.: 00. N0. n0. 00. 0~.: 00. 00.: ozHrmmm
H 0m. 00. mm. 00. 0m. 00. 00. 00.: 0<Hm000
H 0H.: nm.: 00. NN.: co. 0H.: 00. mama
H 00. 0~. ~0. 00.: 00. 05.: 000000
H «H. N0. H0.: 00. n0.: 000000
H 00. 0N. 0N. 00.: goanm
. H 00.: on. 50.: HHoxmm
H H0.: n0. u<mz<=
H 00.: Hunk
H mDHmoxam
G
V. V. nu T S Q Q Q Q Q Q Q Q P P P D. p. D. m” D. P D. D. P P m B

 

monH<Dom >xHH=om oz< MUOBmm>HH 02H zH ammo wmmHn<Hx<> mom meH<z ZOHH<HmmmOU mHAZHm

0H.m MHQ<H

88

 

HNo.Nv HmN.eV Hom.0
mzame No.e +.w oH.HH : oszmmm mm. :

 

 

 

N.o game 6N. + odemm mo. + HHommm Hm. + No uuHem
aoHumsdw
scum AmN.v H00.v H00.0HV
mHo ooze NH. : mzmmodz oo. + No.mH u mmomme
Hmo.0 Hmo.v
E a. - E a. +
Hume:oum
. . mz<ma HHo.V Hoo.v mom .
so N No N mzmmoaz so. + omomzdme He. + osnamo N m
quwmmm Bush
same HNH.V ANN.V
oaammme Moomoe oe. : mmommm HH. : mH.oH u maemommo
HoN.momv
mz<me No.mooN + enuun use
oN.H 0o. oanamo H.o
m2emo<2 Hoo.oosv Hso.0ov HHnuum
oaHmmme .m eo.NeN + moomoe No.0 + mH.oooo: u e<mamo
30 m msoamwoxm muHsmmm mHuHH :OHumsvm
N ouoaHuxm

 

demon mom maHomme onmmmmomm

0H.m mHm<B

89

 

Hoo.oH0 Hoe.Hv
.w No.0H + oaHomzo oo.H :

 

 

 

HoN.NV Hmoo.v
om.H so. mzams eN H +_mmHHoemo Noo + sameness; .m
m.m HoN.o HNo.Ho mo muHue o
doHunaom moomo NH. : mammommo on m :
seam HNH.0 Hom.Nv HHH.oNo
mHo Hmom on. : ommzozo om.e : eo.oHH : omomz<mm
HHoo.0 Hoo.0
eNHHomoo moo. + Nomoomo oH. + Home
O Idod
oo.N so. Hoo.o HoH.o Hoo.v o e sou m.m
oaHmomzo mammommo No N : Hmom nH : ommzozo HN + oceans
mz<ma HoH.V HeN.v seam
N ozoome so. : omomzame Ho. + oo.oo u omezozm
HmN.HNv Hom.Hmv
mzame NN.oNN +.m Ne.ooH + uumuaaans
NH.N No. Hoe.Hv HNa.H0 HNo.ov new e.m
oszmmm Ho. +_z<mm me. + eoomom mm. + oceans
mzmmo<3 HmN.ov HHmuum
oaHemezo oeomzmmm Ne.N : o.eoo: u monomom
30 0598098 3 H500 o u .H. aOHumsvm
Nm ouoaHuxm H H H

 

Ho.usouv nH.m mHm<s

9O

maoamwoxm ecu oumoHvsH mmHAMHum> may um>o mumm
mmmonucmumm
uuHucmsv HH<

.moHHMHHm>
.muouuo unaccoum man mmoHoco muamHonmmoo mnu Haws:
.A00Hns00Hv Hmo mau up wmumHmmv coon o>ms mmHHmHum> moHun HHm 0cm muHemo nun mum moHanum>
.m vacmma< aH vasow on use mumv HHm mo mousom was :oHuaHuommv mumHaaoU <m

 

 

 

Hmo.0 HNo.v
moose no. + HHommN No. +
. . . moon
No.N oo. oonmmzm Hoo.v HHo.v Hoo.v omm:doe Ho o.m
mz<ms Hope Ho + m2mmo<3 Ho + ozoomm mo :
» ::::: IIIIIII. huHucmso
ozHHmmm Hoo.0 ANH.0
ammo omomz<me HH. : ommzozm oH. + mo.N n ommzozo
aNN.oHv Heo.v
.m 0m.H : omHmmzo oH.m :
HoN.HV aNoo.V
mzdms mN.N + NHHHomoo ooo. +
O m 0 “:30“ N O"
NN N N AoH.o HHN.HV mo ueHum
m.m moomo oo. T mmmmommo no.0 :
:OHumsvm
anew HoH.v Hom.Hv
mHo Hmomm oe. : omezozo Na. + oN.ooN u ozoomm
so a maoeumoxm muHamum uHuHH soHunaom
N onoaHuxm

 

Ho.usouo mH.m mHoae

160

140

120

100

80

6O

40

U. S.:

Z of 1953

91

 

 

Quantity of
[“xFed-Beef

'6’ Produced
1' per Capita
.Undeflated
’ Price of
I Hamburger
I

I h- - -~- ’
.,.°’ ’3'? Deflated
° 3; Price of

Hamburger

Quantity of
.°° ,.°" Non-f ed Beef

Produced

per Capita

 

53 55 57 59

Beef:

Figure 3.26

69 71
Year

Price and consumption indexes (1953-1971)

28

26

24

22

20

92

Dollars per 100 Lb.

 

Real Values

  

   

Estimated from
Equation 3a3

 

U.

52 54 56 58 60 62 64 66 68 70

Year
Figure 3.27

S.: Price of fed-beef deflated by the CPI (1953-1971)

93

Lb. per Capita

eal Values
Estimated from

120 .-
F—Equation 3. 2

  

110

90

7O

50

 

 

52 54 56 58 60 62 64 66 68 70

Year
Figure 3.28

U. S.: Quantity of fed-beef produced per
capita (1953-1971)

39

33

27

21

15

94

Lb. per Capita

U.

 

Estimated from
Equation 3.1
Real Values

 

 

S.:

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.29

Quantity of steak consumed per capita (1953-1971)

95

Dollars per 100 Lb.

    

20.00
18.00 Real Values
16.00 Estimated from
Equation 3.5

14.00
12 . 00 0.. .0

.0
10.00

 

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.30

U. S.: Price of non-fed beef deflated by
the CPI (1953-1971)

44

4O

36

32

28

24

20

 

96

Lb. per Capita

    

:gEstimated from
' % Equation 3.8

.0
0,: ‘

   
 

Real Values

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.31

U. S.: Quantity of non-fed beef produced
per capita (1953-1971)

97

Lb. per Capita

Estimated from
Equation 3.4

{Real Values

 

 

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.32

U. S.: Quantity of hamburger and ground beef consumed
per capita (1953-1971)

98

According to the statistical analysis at the retail level, the
most substitutable product for steak is chicken. At no level does fed-
beef have a significant price coefficient, indicating no significant
response to price change.

Chicken, pork and fed-beef all appear as substitutes for non-fed
beef (Table 3.15). Consumption of non-fed beef is also unresponsive to
price change.

Results of the analysis might be more significant using aggregated
beef rather than fed and non-fed beef categories. The prices of non-fed
and fed-beef are highly correlated, .86, and the quantities are relatively
correlated, -.38, both indicating substitutability. At the same time,
Equation 3.5 indicates high substitutability of fed-beef for non-fed
beef although the opposite does not appear to be true (Equation 3.1 and

3.2).

Income and Transfer Payments

 

The most interesting results are the effects of income (Y) and
transfer payments (TRANS) on consumption of fed-beef and non-fed beef
(Equation 3.1 and 3.4, Table 3.15). The income and transfer elastici-
ties have been computed and compared with other elasticities for beef
from other sources (Table 3.16). Although the data for steak and hamburg
are highly questionable, it is interesting that the retail level equa—
tions should have such good statistical results and that it is transfer
payments and income that are so significant.

The income elasticity for fed-beef is inaccurate but the t value
for the income coefficient is small, verifying the inaccuracy of the in-
come elasticity. Transfer payments had a strong impact on beef quanti—

ties consumed, both fed and non-fed. These elasticities are consistent

99

with expectations. Transfer payments, for example, have a bigger elas—
ticity for fed-beef than non-fed, and have a higher elasticity than in-

come .

Farm-Retail Price Differentials

Farm-retail differentials for fed-and non-fed beef when adjusted
for inflation remain quite constant over the period 1953 to 1971 (Figure
3.33). Table 3.17 verifies that this is also true for farm-retail
spreads, as defined by the U. S. Department of Agriculture for choice
beef. Over this same period the percentage going to the farmer has de-

clined from 66 percent of the retail price to 63 percent.

Farm Level Demand

 

The quantity and price of fed-beef are simultaneously determined
according to the farm demand regression result (Equation 3.2). With non-
fed beef the quantity is determined exogenously to the price in time t.
For fed-beef both chuck and hamburger appear to be major retail products
but the high multicollinearity between the prices of retail products
made from beef makes it impossible to tell which one is most important
in determining the quantity of fed-beef produced.

Retail prices are a major determinant of the quantity of fed-beef
sold. Wages have an important effect on fed-beef production; fuel prices
do not. Hogs and broilers are the only other products which appear to
be substitutes and complements to fed-beef. Broilers are the most sig—
nificant but appear as a complement--an unlikely result.

Fed—beef is a substitute for non-fed beef (Equation 3.5). No
other commodity substitutes according to the regression results. The

substitution goes one way. The amount of fed-beef produced affects the

100

price of non-fed but the reverse is not true. Neither fuel nor labor

:costs have a definite impact on non-fed beef (Equation 3.5).

' fleas
Modeling the Hog Sector

As with beef, the hog model is tested for simultaneity in price
.and quantity and lagged response to price changes. Ham and bacon com-
‘bined are used as the retail product. As with beef there are limitations
to the data available for retail products. Thus, ham and bacon are com-
‘bined to form some measure of retail pork use. The price of ham (from
the Bureau of Labor Statistics) is a proxy for the prices of the various

types of ham and bacon.

Analysis of Regression Results

Figure 3.34 illustrates changes in the hog market between 1953
and 1971. The retail and farm price has declined considerably, whereas
the quantity of hogs produced, although it has fluctuated, has on the

average remained about the same. However, the use of hogs for ham and

bacon has increased although there are no data to indicate which product

the shift is from.

No possible substitutes are significant in the regression results.
As with beef, the main variable affecting hog quantities is transfer pay-
ments (TRANS). The elasticity is quite high (.42) and significant. The
results are consistent with those expected because the elasticity for
high income or the average income (Y) is lower than that for the lower
income group or the transfer payments.

The farm level demand analysis indicates the quantity of hogs
produced by farmers is not determined simultaneously with the price of
the same year. Fed-beef is possibly a substitute for hogs (Equation

4.2). Broilers and non-fed beef appear to be possible complements.

101

TABLE 3.16

INCOME AND TRANSFER ELASTICITIES FOR BEEF

 

 

 

 

Fed Non-fed Total
beef beef beef
Income elasticity 511619 2.22 c
t value on income coefficient .30 3.86
Transfer elasticity 6.83 2.48 c
t value on transfer coefficient 7.54 8.66
Low income elasticitya c c .37
High income elasticity c c .37
Aggregate income elasticityb c c .28

 

8Source: J. C. Purcell and Robert Raunikar, "Quantity-Income
Elasticities by Level of Income," igurnal of Farm Economics, XLIX (1967)

bSource: P. S. George and G. A. King, Consumer Demand for Food
Commodities in the United States with Projections for 1980, Giannini
Foundation Monograph No. 26 (Davis: March, 1971).

 

 

cNot available.

102

TABLE 3.17

BEEF, PORK AND POULTRY:
FARM RETAIL SPREADS AND PERCENT OF RETAIL PRICE
GOING TO FARMER (1953-1970)a

 

 

Eggs
Choice Frying grade A
beef Pork chickens large
b
Farm-retail spreads
1953 28.8 25.7 25.7 26.0
1960 31.6 30.6 19.9 22.2
1970 31.9 33.0 19.0 19.6
Percent of regail price
going to farmer
1953 66 67 65 70
1960 65 51 57 65
1970 62 51 46 63

 

aSource: U. S. Department of Agriculture, ERS, Farm-Retail Spread

 

for Food Products, Miscellaneous Publication No. 741 (Washington, D. C.:
Government Printing Office, 1971).

bDeflated by CPI (1967=100).

 

7O

6O

50

4O

30

20

103

Cents per Lb.

    

Estimated _
"'t‘from
Equations
3.3 and 3.7
Real

 
   

.‘u u o r.
.0 O

 
  

/Real
,. fl, 'stimated from
HAMBURGER ” Equations 3. 5
and 3.6

    
 

O
.'.....

 

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.33

U. 8.: Farm retail price differentials for beef
deflated by the CPI (1953-1971)

104

TABLE 3.18

THE HOG MODEL

 

System of equations for hogs

Retail demand:
Retail supply:
Farm demand:
Farm supply:

Endogenous variables

 

PHAM
PHOGS
QHAM
QHOGS

Exogenous variables 2

 

PFRYING
FUEL
PHAMBURG
PHOGSLAG
QBROIL
QFED
QNONFED
TRANS
WAGEFME
Y

(QHAM, PHAM, 2) a o

(QHAM, PHAM, PHOGS, 2) a o
(QHOGS, PHOGS, PHAM, 2) - o
(QHOGS, PHOGS, 2) = 0

price of ham

price of hogs received by farmers
quantity of ham

quantity of hogs produced

price of chicken at the retail level
indexed price of fuel

price of hamburger meat sold at retail
price of hogs received by farmers in t-l
quantity of broilers produced

quantity of fed beef produced

quantity of non-fed beef produced
transfer payments

wage of food manufacturing employees
personal disposable income

.moHanum>

moocowoxm mnu oumoHvaH moHanum> onu uo>o mama .muouuo vumvcwum may mmoHoao muaoHuHHmoou ecu nova:

mononucmumm .H00Hum00Hv Hmu onu >0 woumHHov coon o>mn mmHanuw> ooHua HHm 0cm ouHmmo use was moHanuw>
muHuemsv HH< .m vacmma< aH venom on can mumw HHm mo mousom 0am aOHumHuowov ouoHoaoo <0

 

Am<.v A00.0mV
E a. J. as +

 

105

 

 

Hoe.mv Hoe.v HoN.0
mz<ma mN.o : Home as. : mzmmo<3 NH. : an
oH.N so. .:::. .::::::: a o.e
N.e Hom.ev HHN.NV No nuHum
soHunaou ommzozo 0o.H + mammommo No.m :
anew Hooo.v Heo.0 Hon.oNv
mHo NNHHoemo Hoo. +.mmmoomo 0e. : om.HoH a z<me
Hom.o HNH.0 HHo.Hv
Hoop mm. : mzemodz HN. + omezozo No.N +
mwoc How
. . Hoo.NV Hmoo.o
0N H 0N . : . osnamo N.e
ofimoooa E N N E moo + Ens
mzame HHm.0 Hem.0
» memoomo 0o. +_z<mm oH. + oH.eHH : moomm
Hom.NHo
oeomzame mo.N :
HH
HNe.HHo Hoo.eoHo ommm ”MW
o0 H 0N Hmom ozHoNeN oN o + mz<ea mN Hoe + osnano H e
mzmmomz Hoo.moNv HeN.mHv HHnumm
oaHmoomm .w oe.oN : z<me oo.eH : No.0moN : o<mz<m
30 m msoaowoxo mu 3mm 0 u so umsv
N vocsHoxm H m H Ha H m

 

amooz mom mHHDmmm ZOHmmmmUmm

0H.m NHm<H

106

Z of 1953

Quantity of Ham
and Bacon Consumed
.a‘per Capita
120
Hogs Produced
per Capita

   

100

\‘- ‘ A. \\. //\\
8° \\ I, \ ’ -‘ - \‘ —/“\

\\Price of Ham
\’ \ Deflated b the
60 ‘ I \ CPI y
‘Price of Hogs
Deflated by
the CPI

 

4O

 

 

54 56 58 60 62 64 66 68 70

Year

Figure 3.34
U. S.: Hogs: price and consumption indexes (1953-1971)

107

Dollars per 100 Lb.

Estimated from
Equation 4 2

 

 

11
a
e
R
cocoons-onus.
-
oH.. ..
00
o
a
once-no.
C.-
3.
no...
so...
30
000
000
I...
coo
0
000000
J.
o
o
of
o
o
If
0
00
ounces-0
0
0......-
.C.‘
no.
so
use.
out
one.
so...
so
AU L4 04 nu 00 (U
0‘ 0‘ 0L 0‘ 11 11

71

69

67

65

63

61

59

57

55

53

Year

Figure 3.35

Price of hogs deflated by the CPI (1953-1971)

S.:

U.

85

79

73

67

61

56

50

108

Cents per Lb.

Real

 

Estimated from
Equations 4.3

 

 

53 55 57 59 61 63 65 67 69 71
Year
Figure 3.36

U. S.: Price of ham deflated by the CPI (1953-1971)

109

 

 

Lbs. per Capita
25
Real
24 .HEstimated from
Equation 4.1
23
22 '
2 l a... ’0'.
...o : .0. ‘0‘... a...
2 0 3.. ‘, 3 °‘ ' ' 0 O“ o.
o. .. .. :0. 00...
19 ‘
53 55 57 59 61 63 65 67 69 71
Year
Figure 3.37
U. S.:

Quantity of ham and bacon consumed
per capita (1953—1971)

48

44

40

36

32

28

24

 

110

Cents per Lb.

a; s ‘vfiu. Real Values

"""h Estimated from

Equations
4.2 and 4.3

 

53 55 57 59 61 63 65 67 69 71

Year

Figure 3.38

U. S.: Farm—retail price differentials for ham

deflated by the CPI (1953-1971)

111

TABLE 3.20

INCOME AND TRANSFER ELASTICITIES FOR
HOGS AND POULTRY

 

 

 

 

Ham and

Bacon Broilers Eggs
Income elasticity -.09 1.23 .76
t value for income coefficient .27 7.94 2.43
Transfer elasticity .42 -.002 0
t value for transfer coefficient 2.2 .43 .11
Low income elasticitya .31c .13d .16
High Income elasticity -.26c .13d -.22
Aggregate income elasticityb .13c .17 .05

 

8Source: J. C. Purcell and Robert Raunikar, "Quantity and Income
Elasticities by Level of Income," Journal of Farm Economics, XLIX (1967),
1410-1414.

 

bSource: P. S. George and G. A. King, Consumer Demand for Food
Commodities in the U. S. with Projections for 1980, Giannini Foundation
Monograph No. 26, (Davis: March, 1971).

 

 

cElasticity for pork.

dElasticity for poultry.

112

None have t values larger than two, although they do approach it. Again,
wages and price of fuel do not appear to have a significant effect on
the price of hogs although the t values are close enough to two to sug-
gest that they might have an effect.

The percent of the retail price going to farmers has decreased
over the period of analysis. At the same time the farm-retail spreads
have increased according to the U. S. Department of Agriculture for the
aggregate, pork (Table 3.20).

Figure 3.38 illustrates the declining farm-retail price differen-
tial for ham only. The differences between the spreads and the differen-
tials can be explained in two ways. First, the definitions between the
farmrretail price differential and price spread differ. Second, the
United States Department of Agriculture figures represent an average
price for all cuts of hogs including ham and pigtails. In other words,
we might be looking at Spreads between the various cuts of pork indicat-
ing a declining spread for more expensive cuts and an increasing spread

for less expensive cuts.

Chicken

Modeling the Broiler Sector
Broiler production was used as a proxy for retail consumption
because there are no data on consumption of various types of chicken.
Also, chicken consumption per capita and broiler production are highly
correlated (.98). Table 3.14 shows that broiler production per capita
(QBROILPR) is highly correlated with time (YEAR) which could result in

misinterpretation of results.

113

TABLE 3.21

THE BROILER MODEL

 

System of oquations for broilers

 

Retail demand:
Retail supply:
Farm demand:

Farm supply:

Endogenous variables

 

PBROILERS
PCHICKEN
QBROILERS
QCHICKEN

Exogenous variables 2

 

FUEL
PFEDBF

PHAM
PHAMBURG
PHOGS
PNONFEDBEEF
QBRPRLAG
STAMPS
WAGEFME

Y

(QCHICKEN, PCHICKEN, 2)

(QCHICKEN, PCHICKEN, PBROILERS, 2)
(QBROILERS, PBROILERS, PCHICKEN, 2)
(QBROILERS, PBROILERS, Z)

price
price

of
of

broilers at farm level
chicken at retail level

quantity of broilers produced
quantity of chicken consumed

price
price

price
price
price
price

of
of

of
of
of
of

quantity
quantity
wage of food manufacturing employees
disposable personal income

fuel

fed beef (steers and heifers) received
by farmers

retail ham

hamburger meat at retail

hogs received by farmers

non-fed beef (cows) received by farmers

of broilers produced in year t-l

of bonus food stamps

114

 

$0.3 $5.3 $0.3

odeNmoo oo.oH : mmz<am Ne.m +.m_oo.NH

 

 

 

Ho.m 0o. HoH.o HoH.V HNo.v neuHHous m.m
N.m Hmoe mm. : moome so. : omezoza oN. : Ho uoHsN
GOHHmsvo .Il::
aouw AON.V H00.v Am0.500
mHo mmomee mo. + mzmmoa: mH. : ee.oN u HHomom
Heo.v Heo.o HmH.o
Home Ho. : moomm Ho. : ommzozm HH. +
mumHHouH
. . HeH.o AHo.V new .
mm H no mmomme HH. : mammoes Noo. + osnaao N n
odemmmo seem
mmz<am HHN.V Hoa.v
N oszemm oH. : HHoemm NH. + No.0 : meHHomoo
HNooHv HNNHV HNHHV nemHHoss
mmmmmmmm Ho +.mmmmmm oH :.w.mm H + new
oN.N oo. Hope unmade H.m
szmoaz Hooo.0 Hooo.0 HHneme
oaaummmo z<om Noo. : oszmem No. : «No. u eeHHoNoo
msoamwoxm
o
3o Ne oooaHuxm nuHamum mHuHH soHuna m

 

mmmmaHomm Mom mHHDmMM onmmmmwmm

NN.m NHQ<H

115

.moHHMHum>

msoamwoxm may mumoHvaH moHHMHum> mnu uo>o mumm .muouum wumvdmum mnu mmoHoeo muaoHonmooo mnu nova:

mooonuaoumm .H00Hun0mHv Hmo may 00 vmumHmmc coon m>mn moHHmHum> moHun HHm 0cm muHamo use mum mmHanum>
huHucmsv HH< .m vaamea< aH venom on cmo mama HHm mo mousom 0am coHumHuommv muoHaaoo <0

 

 

Hoo.Nv Hoo.nv HNo.HHV

odHNNNoo om.mN : mez<em No.e +.w NN.oN +
as o o

«.0 Hush 00. : moomm 00. : numzozm H0.H :
.III: ::::I: :IIIIII. mo moHum

cOHuwsvo
scum Hwo.0 HoH.v HmN.NeV
0H0 magnum 00.H.+ Mtho<3 mu. : 00.00H I quthm
3D msoaowoxu mu zoom m u H :OHumsvm
Ne ouoaHuxN H H H

 

Ho.ucoov ~N.0 mHm<e

116

 

  

 

 

Z of 1953
II“ Quantity of
1' Broilers
260 ’.-I Produced
4’ per Capita
1’
‘g’
220 /__"”
z ,I
140 I
I
Price of
100 Frying Chickens
Deflated by the
. CPI
6O ..Q'I..‘CIIIIIIIIIIIIIIIC'I. ‘\ Price Of BrOilerS
-~---------.....,_./Def1ated
~20

 

 

53 55 57 59 61 63 65 67 69 71

Year

Figure 3.39

U. S.: Broilers: Price and consumption
indexes (1953-1971)

37

33

29

25

21

17

13

Lb.

 

per Capita

 

117

Equation 5.1

 

53

U.

55

S.:

57 59 61 63 65 67

Figure 3.40

Quantity of broilers produced per
capita (1953—1971)

69

71

Year

34

30

26

22

18

16

10

118

Cents per Lb.

 

. "-I.. Estimated from
°’. Equation 5.3

 

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.41

U. S.: Price of broilers deflated by the
CPI (1953-1971)

74

66

68

50

42

34

119

Cents per Lb.

 

Real
gEstimated from
Equation 5.4

 

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.42

U. S.: Price of fryers deflated by the
CPI (1953-1971)

30

26

22

18

14

120

Cents per Lb.

 

    

Real

0

7 Estimated from

Equations 5.3 and
5.4

 

U.

53

S.:

55 57 59 61 63 65 67 69

Figure 3.43

Farm-retail price differentials for broilers
deflated by the CPI (1953—1971)

71

Year

121

Analysis of Regression Results

Since 1953 there has been a constant decrease in farm and retail
prices of chicken (Figure 3.39). Concurrently, production and consumption
of chicken has rapidly expanded. Chicken production has increased by 190
percent while the 1971 prices of broilers and retail fryers are 33 per-
cent and 45 percent, respectively, of their 1953 price when adjusted for
inflation. Prices at both the retail and farm level are highly correlated,
.99.

From the regression analysis, we see that the amount of chicken
consumed definitely responds to price change (Table 3.22). The price co-
efficient is significant with a t value of 3.33 with our own price elas-
ticity of -.36. According to the results, hamburger is a substitute for
chicken.

However, since all the retail prices for beef are so highly cor-
related, there is no certainty that hamburger is actually a substitute
but rather some or all retail beef cuts may act as substitutes.

The coefficient for food stamps (STAMPS) is negative and insig—
nificant, but for income it is very high and significant (Table 3.20).
One problem.with elasticity interpretation is that personal disposable
income (Y) is highly correlated with time (YEAR), .97, and there is no
certainty that the large response with respect to increased income is
not a result of the huge technological changes in the industry over the
last 20 years.

Figure 3.43 and Table 3.17 illustrate the decrease in the farm-
retail price differentials and spreads for broilers. According to the
U. S. Department of Agriculture statistics the farm-retail spread de-

clined by 26 percent from 1953 to 1971. At the same time the percentage'

122

of the retail price going to the farmer also declined. Those data are
indicative of the huge decline in retail price. Again, the decline is
explained by technological advances in the marketing sector.

In the analysis of farm level demand no variable has a t value
of two or more although the R2 (.95) and Durbin-Watson statistics (1.85)
are within reasonable range (Equation 5.2). Thus, according to the farm
level demand regression results, there are no definite substitutes or
complements for broilers. Wage rate, price of fuel, and own price are
not significant in determining broiler quantity demand at the farm level.

Since two-stage least squares was used it is likely that in the
first stage results, the overwhelming importance of technological change
was manifested through the income variable (Y). This variable is highly
correlated with time and thus highly correlated with the technological
change in the industry within the 20 year time span of the analysis.
This explains the high R2.

As is evident from the correlation matrix in Table 3.14, income
(Y) explains .94 percent of the variation in quantity of broilers pro-
duced. This quantity (QBROILERS) would have to be run against time to
see if the system used here has provided a better understanding of the
structure of the system than simple use of time. Indeed it has because
when this particular analysis was done, although the R2 is high, .98,
the Durbin—Watson is low 1.16, indicating poor structural analysis when

only time is used.

Eggs

Modeling the Egg Sector
The quantity of eggs produced and eggs consumed is so highly cor-‘

related (.99) that egg production was used as a proxy for consumption.

123

TABLE 3.23

THE EGG MODEL

 

System of equations for eggs

 

Retail demand:l
Retail supply:
Farm demand:

Farm Supply:

Endogenous variables

 

PEGGS
PEGGSF
QEGGS

Exogenous variables 2

 

FUEL
PFRYING
PHAM
PHAMBURGER
QBROILPR
QEGGSPR
QEGGSPRLAG
QFEDBRPR
QNONFED
TRANSFER
WAGEFME

Y

YEAR

(QEGGS, PEGGS, 2)
(QEGGS, PEGGSF, PEGGS, Z)
(QEGGS, PEGGSF, PEGGS, i)
(QEGGS, PEGGSF, 2)

price
price
quant

price
price
price
price
quant
quant
quant
quant
quant

of
of
ity

of
of
of
of
ity
ity
ity
ity
ity

eggs at retail
eggs received by farmers
of eggs

fuel indexed

retail chicken

ham

hamburger

of broilers produced

of eggs produced

of eggs produced in year t-l

of steers and heifers slaughtered
of cows slaughtered

dollar value of income dependent transfer payment
wage of food manufacturing employees
disposable personal income

trend

124

.mmHHMHum> moocmwoxo may ouonvaH mmHanum> mnu uo>o mumm
nova: mononucwumm

.muouum vumstum msu omOHoam musoHonmmoo mnu
.A00Hun0mHv Hmo ocu 0H wouMHmmv soon o>ms mmHHMHum> onua HHm was suHamo Hog mum moHnm

:Hum> kuHuomsv HH< .m wasoaa< aH venom on :00 mumv HHm mo muusom 0am :OHumHuommv mumHaaoo <Hw

 

 

 

 

Anoo.v
E aooo : ammo
. . ANN.0 HNo.v new .
No N No Nanoommo HN. + omezozo No.N + osnauo m o
o<Hmamoomo ::::::::. .::::::: sumo
mz<ma HHN.mo HNn.o
N ammoomo oH.HH : emoomm mm. + sm.oH : Nmoome
HNo.mv
m<mw NN. +
HNoo.v AoN.NHv Ame.v
mz<ma Hoo. : odHeeomo No.0 : Hmoe no. + HHmuue
. . m 006 .
om N as H.o HoN.o HHo.mmo Hom.o N“ nMHnm N o
soHenaom mammodz oo +.w 00 N : oeomz<me oe +
acne Hom.v Hem.v HeH.NNHNo
mHo same «H. + oszemm 0N. + oN.NHe: : emoome
Hmo.o Hoooo.v
,mmmm NH :.mmmmm Noooo + nmwu
oo.H No. Hom.v AeHo.V Hmoo.o pom H.o
Hoop .w mo.+ omomzamm eoo. +_z<mm Noo. + osnauo
mzemoos HNo.V HNo.V HHnunN
omHmomooNo ozHNNNN eooo. : emoome Ho. : no.0aN : Nemoomo
msoamwoxo
30 NM covsHoxm muHsmmm mHuHH aoHumsum

 

mmoom Mom mHHDmmm onmmmmumm

<N.0 MHm<H

125

Z of 1953

 

  

 

 

120
100 --‘:.........
.~..."""'UO.||Io-Icanon"...."tloussooni"‘ Quantity Of
80 Eggs Produced
per Capita
60
Retail Price
of Eggs Deflated
40 Farm Price of
Eggs Deflated
by the CPI
53 55 57 59 61 63 65 67 69 71
Year

Figure 3.44

U. S.: Eggs: price and consumption indexes (1953-1971)

33

32

31

30

29

28

126

Dozen per Capita

f
O
O
O

.0

Estimated from
° Equation 6.1
Real

 

 

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.45

U. S.: Quantity of eggs produced per capita (1953-1971)

58

52

46

4O

34

28

127

Cents per Dozen

   

Real
Estimated from

Equation 6-3

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.46

U. S.: Farm price of eggs deflated by the CPI (1953-1971)

88

8O

72

64

56

48

40

128

Cents per Dozen

. Estimated from
Equation 6.2
Real

 

 

 

53 55 57 59 61 63 65 67 69 71

Year
Figure 3.47

U. S.: Retail price of eggs deflated by the CPI
(1953-1971)

129

Cents per Dozen

  
   

 

   

 

38
34
30 '°
26 '-.
22
i-Estimated from
18 : Equations 6.2
7 a f and 6.3
'7.....~' Real
14
53 55 57 59 61 63 65 67 69 71
Figure 3.48
U. S.:

Year

Farm—retail price differentials for eggs
deflated by the CPI (1953-1971)

 

130

Analysis of Regression Results

Per capita egg consumption has declined considerably since 1953.
Per capita production, which is highly correlated (.99) with per capita
egg consumption, has declined by 13 percent since 1953. From 1951 to
1969 the amount of eggs used for breaking has increased by 52 percent.
This represents an increase in percentage of total egg production from
6.6 to 8.5.1 Various parts or the whole part of the liquid egg is ex-
tracted for use in the manufacture of baked goods, mayonnaise, noodles,
baby food and many other products. On the whole, per capita consumption
of processed eggs has increased but processed eggs still comprise only
about 10 percent of total per capita egg consumption.

Statistical analysis shows there is a significant and negative
relationship between price of eggs at the farm and the quantity of eggs
produced (Equation 6.3, Table 3.24). The quantity of eggs produced
according to this model is exogenous to the present price.

Non—fed beef according to the farm level equation (6.3) is a com-
plement for eggs and has a high t value (2.37). Fed-beef also appears
as a complement but the t value is smaller (.94). In the retail level
(Equation 6.1), the coefficient price of hamburg has a very low t value
and, therefore, contradicts the results above. It may be that beef is
acting as a proxy for an excluded variable or that the multicollinearity
between the retail price of eggs (PEGGSR) and the quantity of non-fed
beef produced (QNONFED) is interfering with the accuracy of the coeffi-

cients.

 

1U. S. Department of Agriculture, Economic Research Service,
The EgglProducts Industry, Marketing Research Report No. 917 (Washington,
D. C.: Government Printing Office, 1971) p. 4.

131

Time (YEAR) was included as a variable in the retail demand
equations to test for the possibility that changes in tastes and prefer-
ences have caused the decline in per capita egg consumption. The t
value for the coefficient is four, indicating the possibility of a change
in tastes and preferences away from eggs over this period. Recommenda-
tions by doctors and subsequent advertising on the relationship between
eggs, cholesterol and heart disease may have affected the preference for
eggs.

U. S. Department of Agriculture data for farm-retail spreads ap-
pear in Table 3.17; Figure 3.48 illustrates the decline in farm-retail
price differentials between 1953 and 1971. Recent technological advances
in egg production and marketing have led to the sharp decline in price.
But the swing away from egg consumption despite the retail drop in price
has caused an even larger drop in price. The amount going to the marketers
has decreased but at the same time the percentage as well as amount going

to the farmer has declined.

CHAPTER IV

SUMMARY AND CONCLUSIONS

The major objectives of this work are to analyze some specific
United States farm commodities and isolate and estimate the relevant
variables affecting their consumption. The purpose is to evaluate the
substitutes and complements for the particular retail products as well
as the effect of income and specific marketing costs. The effect of
transfer payments and food stamps on demand is also analyzed. Many
variables cannot be studied because they are either assumed to be insig-
nificant or their effects cannot be measured by this analysis. Some
variables which could be isolated in further analysis were mentioned
previously.

Econometric analysis of the demand function is used to determine
the relationships among the variables. Unlike previous retail analysis
covering many retail commodities, no attempt is made to account for all
foods using the usual assumptions about elasticities. Instead, a system
was develOped for each farm commodity. Retail and farmrlevel data are
used; the retail level data being considered especially relevant for
analysis involving consumer choices.

The major results for each commodity are given in Chapter 3.

But there are also some general observations about the farm-retail price
differentials and the transfer payments and food stamps that can be made.

In this analysis, marketing margins are not estimated directly;
rather farm-retail price differentials emerge from the estimations of

the price at both the retail and farm level. Thus, to determine the

132

133

marketing margins (as defined by the U. S. Department of Agriculture) a by-
productnallowance must be estimated. This is omitted in our analysis. Also
there is no allowance for the other ingredients in the retail commodity.

Instead, conversion factors1 specified by the U. S. Department
of Agriculture are used to determine the amount of the farm commodity
used in the relevant amount of the retail product. The farm price for
the retail equivalent amount of the product is subtracted from the re-
tail price to ascertain the marketing margin.

Farm-retail price differential estimates are more successful
in some cases than others. Because two separate estimates of two difu
ferent prices are used to estimate the farm-retail price differential,
errors in the separate equations may counteract each other or build on
each other so that the real and estimated differentials tend to diverge.
Estimates for bread and cookies were generally more accurate. Fluid
milk, however, is a good example of the actual diverging from the esti-
mate. Cheese and butter are not estimated because the aggregated price
of milk at the farm used in this analysis is higher than the retail
price of the equivalent quantities of butter and cheese. Other commodi-
ties tend to follow the appropriate trends, but do not provide accurate
year to year estimates.

A major problem in analyzing the marketing costs is finding ap-
propriate data to proxy these costs. Fuel prices and wages of food manu-
facturing employees are used here for that purpose. Fuel costs play a

role in the amount of wheat and fluid milk consumed and the quantity of

 

sq

1U. S. Department of Agriculture, Economic Research Service,

Conversion Factors and Weights and Measures for Agricultural Commodities
and their Products, Statistical Bulletin No. 362 (Washington, D. C.:
Government Printing Office, 1965).

134

milk demanded at the farm level. Labor costs at the retail level affect
the supply of fluid milk. A more direct analysis of marketing margins
is needed to better understand the various components of the margins.

This analysis indicates only fluid milk has experienced any in-
crease in the percent of the retail price going to the farmer. The per-
centage going to the farmer has declined whether or not the retail price
and/or the farm-retail spread has increased or decreased. Between 1960
and 1970 the deflated farm-retail price spread increased for cheese,
beef and pork. Only for these products, then, has the actual amount go-
ing to marketing costs increased.

From 1950 to 1960 there was also an increase in spread for milk
and bread. Consumption of beef and cheese increased, too. (Chicken is
not included here because the technological changes have greatly lowered
the costs of marketing broilers). Consumption of high quality pork cuts
has increased also although pork slaughter per capita has not. Thus, a
possible correlation exists between increases in consumption per capita
and increases in farm—retail spreads.

As shown earlier, chosen proxies for marketing costs have not
proved significant in determining quantities consumed of the various
products. The correlation between quantities consumed and farm-retail
spreads might give some hint as to why some spreads have increased. If
they cannot be explained by costs, they can perhaps be explained by other
costs not included in this analysis or profits gained by marketing firms
from increased demand of these products. Structural analysis of the in-
dustries involved would help answer this question.

The general decline in marketing margins remains unexplained.

One assumes they would remain rather steady, but technological changes

135

may be responsible for their continuing decline. This hypothesis re—
quires further investigation.

Income-dependent transfer payments in 1971 are 3.37 times larger
than their 1953 value. After adjustment for inflation and population
increase, the deflated per capita figure is 1.75 times the 1953 value.
Similarly, disposable personal income adjusted for inflation and popula-
tion change increased by only 50 percent.

In order to believe the statistical results from the analysis
of income and transfer payments we expect some kind of coincidence be-
tween the transfer elasticities derived in this study and low income
elasticities derived from other sources. Table 4.1 compares the transfer
elasticities with the low income elasticities presented in Raunikar and
Purcell's study. Where comparisons are available, except for fluid milk
and chicken, the signs are coincidental. The coefficient for transfer
payments for chicken was not significant and, therefore, does not con-
tradict the low income elasticity found by Purcell and Raunikar. However,
the statistical results of the milk analysis contradict the results from
other sources.

This study contributes to the information about food stamps and
transfer payments combined over time. It does not single out food stamps
and analyze the effect of this particular policy alone. This has been
done by the U. S. Department of Agriculture in a study entitled, EEEEE
Food Stamps and Cash Income Supplements: Their Effectiveness is Expand-

ing Demand for Food.1

 

 

1U. S. Department of Agriculture, Economic Research Service,

Bonus Food Stamps and Cash Income Supplements. Their Effectiveness in
Expanding Demand for Food, Marketing Research Report No. 1034 (Washington,
D. C:: Government Printing Office, 1974).

136

TABLE 4.1

COMPARISON OF TRANSFER ELASTICITIES AND
LOW INCOME ELASTICITIES

 

 

 

 

 

 

 

 

 

 

 

 

t value of transfer LOW

Transfer coefficient in income a

elasticity retail equation elasticity
Flour -.036 .90 -.30
Cookies -.013 3.61 b
Bread '.008 3.95 .19
Cheese 1,28 6.44 b
Butter -.87 .38 b
Fluid Milk -.79 3.52 .35
Steak 6.83 7.54 .37
Hamburger 2.48 8.66
Ham and bacon .42 2.52 .31
Chicken - . 002 .43c . 13
Eggs .43 1.16 .16

 

3Source: J. C. Purcell and Robert Raunikar, "Quantity Income
Elasticities for Foods by Level of Income," Journal of Farm Economics
XLIX (1967), 1410-1414.

 

bNot available.

cElasticity for poultry.

137

Only with better micro-level cross-sectional data specifying con~
sumption with and without food stamps can a more complete analySis be
undertaken. This particular analysis could be improved if the results
found here were compared with those of a study, similar to that done by
the U. S. Department of Agriculture study mentioned above, analyzing food
stamps and transfer payments together rather than food stamps alone.

Results from this study indicate transfer payments and the food
stamp program have had a significant impact on food consumption in the
United States. These impacts have been felt by the farmers, wholesalers
and retailers of farm commodities as well as the whole economy. This im—
plies that any policy analysis of these programs must consider carefully
not only the effects on the recipients but also the effects on others
in the economy.

Studying the effects on the various factors in the economy, the
transfer and food stamp programs as an aggregate seem to shift diets
away from grains and toward more expensive sources of protein and calcium.
The effect on the producers is to shift demand to beef producers and to
producers of feed for beef and away from producers of wheat.

The effect on the marketers is somewhat obscure. Data imply that
where there is a strong increase in demand there is an increase in market-
ing margins. But this may mean profits rather than marketing services.

A more direct analysis should yield some information on the effect of
transfer payments and food stamps on demand for these services.

The analysis sheds some light on the adequacy of the food stamp
program. The two stated objectives of the food stamp programs are to
1) increase the nutritional standards of the low income population; and

2) increase demand for food. This study highlights the fact that there

138

are distribution impacts regarding both of these objectives. Various
income groups use their new income differently and, as a result of
these programs, demand is increased from some farm products and de-

creased for others.

APPENDICES

APPENDIX A

DESCRIPTION OF TRANSFER PROGRAMS

 

APPENDIX A

DESCRIPTION OF TRANSFER PROGRAMS

Aid to Families with Dependent Children (AFDC). Assistance is
provided to needy families with dependent children where the father is
absent, incapacitated, or in 25 states, unemployed. Benefit amounts
vary by family size, and are decreased as the family's other income
rises. The program is administered and partially funded by State and/or
local governments. Non-Federal costs are matched with Federal funds,
the Federal percentage ranging from 50 percent of total costs in New
York to 83 percent in Mississippi. AFDC programs operate in 54 juris-
dictions (all 50 states, the District of Columbia, Guam, Puerto Rico,

and the Virgin Islands).

Food Stamp_. The Department of Agriculture, through local wel-

 

fare agencies, provides food coupons for needy families and individuals
in the 2,027 counties electing to operate the program. A county cannot
operate both the food stamp and food distribution programs except in
emergency situations. Eligible persons may purchase the coupons; the
purchase price increases as other income increases. Coupon allotments
vary with family size. The coupons are redeemed at face value by grocery
stores for food items purchased by recipients. State and local govern-

ments bear most of the administrative costs.

General Assistance (GA). Welfare payments made in cash or in

kind by states and localities to low-income persons ineligible for

139

 

140

assistance under one of the four Federal categories (families with de-
pendent children, the aged, the blind, and the disabled) are termed
general assistance. These programs are authorized, administered, and
financed solely by state and local governments. Rules of eligibility
and benefit levels vary widely from place to place, with some locali—

ties having virtually no general assistance at all.

Old Age Assistance (0AA). Assistance is provided to needy in—
dividuals who are age 65 or over. Benefits decrease for increases in
other income. (For administration, financing and jurisdictions covered,

see "aid to families with dependent children").

Source: U. S. Congress, Joint Economic Committee, Public Income Transfer
Programs: The Incidence of Multiple Benefits and the Issues Raised by
their Receipt, by James Storey, Joint Economic Committee Print, Paper No. 1
(Washington, D. C.: Government Printing Office, 1972).

 

 

APPENDIX B

VARIABLES USED IN STATISTICAL ANALYSIS

 

 

 

 

Nu0HI00mH 0H0 .0H\mueoo :oonLo 0cH>HH pom ooHHm .330.H UZHHMmm
NNoH:omoH mam .aH m\o Hm>mH HHnumn an saon No mUHNa NHNN
mumaumm 00 vw>HmomH
mom NNoH ma .eH ooH\w AnemNHus osm mummenv Hume:omN Ho mUHHe moomma
H:u was» aH mumaumm 00 00>Hmowu
mom NNoH ma .aH ooH\o ananHue one mammunv Nume:omN «6 uUHNe oaammee
NNoHuomoH mHo .eH\o mmeoou eUHSosnn sou 66Hua HHnume meom
Nan:000H 0H0 .0H\muooo Mezzo mo ooHum HHmumH Monmom
~N0HI000H 0H0 .0H N\H\0 mmmono amoHums< mo moHuu HHS?H mmmmmum
NH¢HI000H 0H0 .0H\0 HHmuoH um 0mmu0 ouHca mo moHHm mmzmm
NNoH:omoH mHm .0H\m ensues No uUHNe HHmuus mmegoom
NNmH m< .0H\muaoo mumaumm 00 00>Hmomu muwHHoun How moHum HH000m
Nan:0mmH 0H0 .0H\mucwo Ho>oH HHmuoH 0moun mana mo moHum 0<000m
.soHuomamcH Hmnmvmw
nova: vommoooua 0am wmumamua an Huwnuo
Huom .muusn Huoa "mmmHmson was 0000a
Hmumz .Hmstwu moHaoHu “mmmHmoon .0000m
Hom mNoH mzH .aH ooo.H sauna .unHamus an: momUHHm .smHamue sounm o<mz<m
eNoH nasHN NsHeaennu
eN .snz we: ooHnNooH :asna as Samson usNHH one amass .Hmao Hope
000 NNmH m< 00Hum0mH xmwsH ooHum umasmcou H00
COHuummmcH Hmumvmw Haws: vmmmmooua
Hom oNoH m2H .aH ooo.H one onenamee Nude osaosw one unmeananm monoNoN
00mm sumo, :oHumoHHasm
mousom muHaD COHuaHuome mam: mHanHm>

 

 

mHme<z< H<UHHmHH<Hm zH 0mm: mmHm<Hm<>

m xHszmm<

141

142

uanma amulou:00mmu

 

 

Noe .Hoa NNoH m< .eH ooo.ooo.H .soHuuaooee emHHosa HnHeemeaou HHono
muoswoum vmumHmu
Nu .N0 .00 0am oxmo .0m0u0 mosvoud ou mmHuumavcH
.00 .000H :0 .uBo ooo.H vmuomHmm sH 00m: uson ummna mo muHuemav 0<mmmo
N.H NNmH m< .an\0 mumEumm >0 00>Hmomu umocz mo moHum 030
H:u mosh cH
~.H NNmH m< .=0\0 mumsumm an 00>Hooou amass mo moHua 0<Hzmmzm
NNoH:omoH mHo .eH m\e unmam No sUHHa HHnumn Naoome
NNOH:000H 0H0 .0H\mucmo umma ooHoso .0 .0 mo ooHHn H.230.H 020000
H 0H=0 so moHHmaam doom oHummaow
Noo NNoH ma ooo.ooo.H No use NsHenm soHunHaaoe anHH>Ho .m .o Noe
museumm mp
000 NumH m< .nH 00H\0 00>Hmoou Amsoov moon noHIGoG you moHue ommzozm
HIu you» aH museumm 00
000 man 0< .0H 00H\0 00>Hmoou AmSOov moon wwHIcos you ooHum 0<Hmmmzm
. monoum muooouw
NNoH:omoH mHo .Hnm N\H\o ea oHom HHHa oHaHH Ho uane HHnumn ooeoze
museumm mp
000 Nan m< .0H 00H\0 00>Hmowu mHmmoHona HHHE HHm mo moHun zm<m20
NumH:00mH 0H0 .nH.mu:mo Ho>oH HHmumu um oaHumwuma mo moHun um<zm
H:u aH
on NNoH ma .mH ooH\o nenaenH as om>Hnume noes new neHsa oaHmoomN
0N0 «~0H 0< .nH 00H\0 mumsumm 00 00>Hoomu mmo: How moHua 000:0
~NOH:000H 0H0 .0H\muamo Hm>oH H230.H um umma nowuonamn How ooHun omomz<mm
«NmHI00mH 0H0 .0H\mucoo am: voxoam oHons now ooHum HHmuoH z<mm
00mm mama GOHuooHHnam
momwaom : muHap aoHuaHuomma mama mHanum>

 

A0.uaoov m NHazmmm<

 

coauumamcw Hmumvmm Mucus

 

 

.na wmmmmuoum cam wmumamum muse mmmamaon
Hon mNmH ng coo.H new mummou .maoau .mxmmum mo xufiuamac M<memo
Auwunwamaw umwfim: vam ammum
.na macfia umucwsmHm Hmuouv umumunwsmam
mum Numfi m< ooo.ooo.H AmsouV mmmn ammucoc mo hufiuamsu ammzozo
.na HIu umm% ca wmumunmzmHm
mum Nan m< ooo.ooo.H Amsoov mama cmmuaoa mo zufiuamsu o¢qmmmzo
.na Hun ummz ca
qu.oqq Numfi m< ooo.ooo.a aha“ as“ so wmoauoua xafia mo mufiuamaw o<qzquHzo
.na
Hmm «Rafi m< ooo.ooo.a umumunmsmfim mmo; mo mufiuamaw ammoomo
.na cowuusvoum
qu.oqq Numa m< ooo.ooo.a xafia nfiaaw ca gem: saga mo zufiuamac aHsgmo
uaoam wwumamua can
14 NB .mo .mo wmvcwan waaxma How mmfiuumscaw vmuumamm
“m .wm .qm¢a so “so ooo.H an umasmaou “scam “muss mo aufiuamsu msoqmo
.AH wmuounwsmam
mum «Rafi m4 ooo.ooo.a AmumMfim: cam mummumv mama cum mo hufiuamsu mmmmawmc
mhmxumhu flaw.
an .no .mo mmfixoou wafixma How mwfiuumsvca wmuumamm
.wm.qmmH so .u30 ooo.H up wussmaou “scam “muss mo mufluamsu mmHmoooo
.AH uwcuo van mmmmno mwMuuou
aqq.oqq Numa m< ooo.ooo.a .amUHHma« wafixma a“ umma xHHa mo zufiuamau mmmmmoo
.nH umuuan
H¢¢.oqq «Rafi m< ooo.ooo.H K$5233 wafixma aH cums xafia mo zufiuamau mmaaamo
mwmm mama coaumoaansm
mmuusom . mufiap dowuaauummn mama mapmwum>

 

Av.uaoov m NHszmm<

144

 

.muumuumn< Hmowumfiumum
“mumuauummscm: mo mamamo

 

“<m ”coaumsuwm coaumuuommcmuH wcm mdfiumxumz
"Eu ”mafiumHuMum gonad mo ammusm

 

 

“mam ”muwumwumum Hmuzuasowuw<‘ um<m

"was “muwumwumum ummz wam MUOumm>fiA umzq

 

 

 

omma
wcwaafiwmn manmfium> wcwuu m<mw
mom Nmma m< .Hom.afim waouca Hmaomumm manmmommww M
N mnmalomma m9: anon\m mummoamam wcfiuauummacma voom mo mmma mzmmw<3
on .mo .00 muamuwfimmm Hmumamw .440 .oom< .maamum
mo .Nnoa <m ooo.ooo.aw coon "mamuwoua Hummamuu wmumwu maoocw mz<ma
AmQEMum voom maconv
No ~m¢a m< ooo.ooo.aw mmswum woom ou coausnfiuucou Hmumwmm mmz<ym
.39
NH «Rafi m< ooo.ooo.H H-“ ummz afl vmuuoaxm amass mo mufiuamau u<gxzo
.sn Hlu ummz CH
NH NN¢H m< ooo.ooo.H .m.: mag :H umoswopa “mugs mo hufiuamav o¢goommmzo
.33
NH NNmH m4 ooo.ooo.H .m.: a“ woo“ you cum: ummns mo xufiuamsu noomo
mwmm mama :owumoaansm
mmouaom mafia: coaumauummo wamc mfinmfium>

 

Aw.uaouv m xHazmmm<

1'}

BIBLIOGRAPHY

BIBLIOGRAPHY

Arthur, H. B. and B. W. Marion. Dynamic Factors in Vertical Commodity
Systems, Research Bulletin 1065. Wooster: Ohio Agricultural
Research and Development Center, 1973.

Brandow, G. E. Interrelations Among Demands for Farm Products and Implica—
tions of Market Supply. Pennsylvania State Agricultural Experiment
Station Bulletin 680. University Park: 1961.

Foote, Richard. Analytical Tools for Studying Demand and Price Structure.
U. S. Department of Agriculture, Agricultural Handbook No. 146.
Washington, D. C.: August 1956.

Friedman, Milton. Price Theory. Chicago: University of Chicago Press,
1962.

 

George,Pu S. and King, G. A. Consumer Demand for Food Commodities in the
United States with Projections for 1980. Giannini Foundation
Monograph No. 26. Davis: March 1971.

Hiemstra, Stephen and Egbert, Alvin. "Shifting Direct Government Payments
from Agriculture to Poor People: Impacts on Food Consumption and
Farm Income." Agricultural Economics Research 21 (1969), 61-69.

Lane, Sylvia. "Effectiveness of Public Income Redistributive Policies
and Programs on the Redistribution of Income to Low-Income Groups."
Paper presented at the meeting of the American Agricultural
Economics Association, Columbus, Ohio, August 12, 1975.

Lee, Dae Sung and Ridgel, Gus T. Income and Food Expenditure Patterns
for a Selected Group of Families in Frankfort-Franklin County.
Kentucky State University. Cooperative State Research Service.
Frankfort: May, 1974.

 

Madden, Patrick J. and Yoder, Marion D. Program Evaluation: Food Stamps
and Commodity Distribution in Rural Areas of Central Pennsylvania.
Pennsylvania State University Agricultural Experiment Station
Bulletin 780. University Park: June 1972.

Mittelhammer, Ron and West, Donald A. "Impact of the Food Stamp Program
on the Demand for Food; Some Theoretical Considerations." Paper
presented at the 1974 American Agricultural Economics Association
Meetings in College Station, Texas. August, 1974.

Okner, Benjamin A. and Pechman, Joseph. "Who Paid the Taxes in 1966?"
American Economic Review, LXIV (May, 1974), 168n174.

Purcell, J. C. and Robert Raunikar. "Quantity-Income Elasticities for

Foods by Level of Income." ,Journal of Farm Economics, XLIX
(1967), 1410-1414.

145

 

146

Reynolds, M. and Smolensky, E. "Post—Fisc Distribution of Income: 1950,
1960 and 1970." Paper presented at the meetings of the Western
Economic Association, June 27, 1975.

Smith, James D. and Franklin, Stephen D. "The Concentration of Personal
Wealth." American Economic Review, LXIV (May, 1974) 162-167.

U. S. Congress. Joint Economic Committee. Public Income Transfer Pro-
grams: The Incidence of Multiple Benefits and the Issues Raised
by Their Receipt, by James R. Storey. Joint Committee Print,
Paper No. 1. Washington, D. C.: Government Printing Office,
1972. '

 

 

 

U. S. Department of Agriculture. AAgricultural Statistics. Washington,
D. C.: Government Printing Office, 1950-1975.

 

.‘ Economic Research Service. Bonus Food Stamps and Cash Income
Supplements: Their Effectiveness in EXpanding Demand for Food.
Marketing Research Report No. 1034. Washington, D. C.: Govern-
ment Printing Office, 1974.

 

 

. Economic Research Service. Conversion Factors and Weights
and Measures for Agricultural Commodities and their Products.
Statistical Bulletin No. 362. Washington, D. C.: Government
Printing Office, 1965.

 

. Economic Research Service. Economic Effects of the U. S. Food
Stamp Program Calendar Year 1972 and Fiscal Year 1974. Agricul-
tural Economic Report No. 331. Washington, D. C.: Government
Printing Office, 1976.

 

 

. Economic Research Service. The Egg Products Industry. Mar—
keting Research Report No. 917. Washington, D. C.: Government
Printing Office, 1971.

 

. Economic Research Service. Farm-Retail Spreads for Food
Products. Miscellaneous Publication No. 741. Washington, D. C.:
Government Printing Office, 1971.

 

. Economic Research Service. Food Consumption, Prices and Ex—
penditures. Agricultural Report No. 138. Washington, D. C.:
Government Printing Office, 1968.

 

 

. Economic Research Service. Livestock and Meat Statistics.
Statistical Bulletin No. 522. Washington, D. C.: Government
Printing Office, 1973.

 

. Economic Research Service. Marketing and Transportation
Situation Washington, D. C.: Government Printing Office, May,
1974.

 

 

147

. Economic Research Service. Milk Pricing, by Alden Manchester.
Agricultural Economics Report No. 315. Washington, D. C.:
Government Printing Office, 1975.

 

. Economic Research Service. Vertical and Horizontal Integra-
tion in the Market Egg Industry, 1955-69, by George Rogers.
Economic Research Service 477. Washington, D. C.: Government
Printing Office, 1971.

 

 

United States Department of Commerce. Bureau of the Census. Census of
Manufacturers. Washington, D. C.: Government Printing Office,
every five years.

 

. Bureau of the Census. Statistical Abstracts of the United
States. Washington, D. C.: Government Printing Office, 1975.

 

United States Department of Labor. Bureau of Labor Statistics. Esti-
mated Retail Food Prices. Washington, D. C.: Government Print-
ing Office, monthly.

 

Waldorf, William. "The Demand and Supply for Marketing Services."
Journal of Farm Economics, XV, 42-60.

 

West, Donald A. and Price, David W. "The Effects of Income, Assets, Food
Programs and Household Size on Food Consumption." American
Journal of Agricultural Economics, LXIII (November, 1976), 725—730.