“1' w'v‘- —— ——_—-

 

19:3

DEMAND FOR ERG-ELERS {N THE
UNETEE‘J HATES

Them for Wu Dag?“ a? M. S.
MICHIGAN STATE UNEVERSETY

Glenn WV. Woodie};
1959

 

.- ”SI.
.5.

  
     

  

LIBRARY

Michigan State
University

DEMAND FOR BROIIERS IN THE mums) 5mm:
by

Glenn W. Woodley

AN ABSTRACT

Submitted to the College of Agriculture of Michigan
State University of Agriculture and Applied
Science in partial fulfillment of the
requirements for the degree of

MASTER OF SCIENCE

Department of Agricultural Economics

1959
Approved w W
_.__ r

 

 

The t
the relati
to detenm‘
States and
to beef at
has increa

The t
deland fo
the deman
relative
aIllllysis,
eImmune t
whether a
Studied.
u“ Peri
Sreseion
°btain m
in; the

In
the fin; a
1310:th
0r Consu
“are 19 4
1925 to

indlCate

Glenn W. Voodley
ABSTRACT

The two objectives of this study were: (1) to examine and explain
the relation of various factors affecting the demand for broilers and
to determine a model to estimate the demand for broilers in the United
States and (2) to examine the relative demand for broilers with respect
to beef and pork and to test the hypothesis that the demand for broilers
has increased relative to the demand for beef and pork.

The demand relationships examined in this study were: (1) the
demand for broilers, (2) the demand for beef relative to broilers, (3)
the demand for pork relative to broilers and (4) the demand for beef
relative to pork. Two approaches, graphic and multiple regression
analysis, were used in this study. The graphic analysis was used to
examdne the relationship between the price and quantity and to determine
whether any shifts occurred in the relationship during the period
studied. A further use of the graphic analysis was to determine the
time period and apprOpriate mathematical function to use in the re-
gression analysis. The purpose of the regression analysis was to
obtain more precise estimates of the relationships than was obtained
from.the graphic analysis.

In the graphic analysis, actual and relative time series data for
the retail price and per capita consumption of the commodities were
plotted with price or price ratio on the vertical axis and consumption
or consumption ratio on the horizontal axis. The time periods examined
were 1940 to 1958 for the actual and relative demand for broilers and
1925 to 1958 for the demand for beef relative to pork. This analysis

indicated changing relationships between price and consumption data

Glenn W. Woodley

prior to 1950 and more stability in the relationships after 1950. Since
the relationships after 1950 appeared to remain stable, single equation
least squares regression equations were fitted to the data for 1950 to
1958. Two models were fitted for the demand for broilers with retail
price and per capita consumption of broilers each considered as the
dependent variables; all relative demand models were fitted with the
price ratio as the dependent variable. Per capita disposable income
was included in each model as a demand shifter. Although each of the
models esthmated the annual observations with a high degree of accuracy,
reliability of the regression coefficients is questionable due to the
high intercorrelation existing between the independent variables in-
cluded in the models; this in turn limits the reliability of any
statements made concerning the various elasticities.

This study indicated that prior to 1950 the demand for broilers
increased relative to the demand for beef and pork while the demand for
beef increased relative to the demand for pork. Since 1950 the demand
for broilers and the demand for broilers relative to the demand for pork
appeared to remain stable while the demand for beef relative to the
demand for broilers and pork increased. Availability and improved
quality of broilers were considered as the most important factors
which caused the increased preference and demand for broilers prior to

1950.

 

DEMAND FOR BROILERS IN THE UNITED STATES

by

Glenn W. Woodley

A THESIS

Submitted to the College of Agriculture of Michigan
State University of Agriculture and Applied
Science in partial fulfillment of the
requirements for the degree of

MASTER OF SCIENCE

Department of Agricultural Economics

I

1959

ACKNOWLEDGEMENTS

The author wishes to express his appreciation to the many
peeple who made the completion of this thesis possible.

The author wishes to express a very special debt of
gratitude to Mr. Lester V. Manderscheid for his helpful
assistance during the preparation of this manuscript. The
author's association with Mr. Manderscheid was the most re-
warding experience of his graduate study.

The author is indebted to Dr. L. L. Boger and the Depart-
ment of Agricultural Economice for the Opportunity to study at
Michigan State University and the financial assistance while

at Michigan State.

Credit is due to Mrs. Helen Westman and other members of
the statistical pool for their c00peration with the statistical
computations. Many thanks are due to Miss Beverly Hamilton and
Mrs. Sandra Rogers for their c00peration in typing the first
draft of this thesis.

The author also wishes to express his appreciation to his
fellow graduate students for making his graduate study at
Michigan State most enjoyable; their association during his
graduate study was most appreciated.

(ZIDKPTER

II

JEII

'VI

TABLE OF CONTENTS

INTRODUCTION . . . . . . .

Objectives of the Study . . .

Growth of the Broiler Industry

CONCEPTS . . . . . . . . .

Demand Curves . . . . .
Relative Demand . . . .
Demand Elasticities . .

REVIEW OF PREVIOUS STUDIES
A

Aggregate Demand for Broilers . . . .

Consumer Attitudes and Preferences

Relative Demand . .
METHOD 0? ANALYSIS . . . .

Choice of Variables . .
Choice of Time Periods.
Choice of Functions . .
Data . .

RESULTS OF ANALYSIS . . .

Graphic Analysis. . . .
Regression Analysis . .

SOME CONCLUDING REMARKS .

Indications of the Analysis .
Some Reasons for the Changes in
Some Problems and Suggestions for Future Studies.

'EVIBLIOGRAPHY. . . . . . . . . . . .

AKPTENDICES . . . . . . .

Demand

Page

15
l6
l7
19
22
22
27
31
37
37
43
44
46
48

48
55

79
79
81
84
87

94

;£fiigpre

1.0

1.1

112

13

15

16

LIST OF FIGURES
Page

Commercial Broiler and Total Chicken Meat Production,
United States, 1934 to 1958 . . . . . . . . . . . . . 4

Gross Income from Commercial Broilers, United States,
1934 to 1958 O O 0 O I O O 0 O O O O O O O O O O C O 4

Average Retail and Farm.Prices of Broilers, United
States, 1940 to 1958 O O C O O O I O O O I C C O O O 9

Retail Price, Farm‘Value and Farm-Retail Spread for
Broilers, United States, 1950 to 1958 . . . . . . . . 9

Per Capita Consumption of Total Chicken Meat and Broilers,
Ready-to-Cook Basis, United States, 1909 to 1958 . . 12

Retail Price Ratios of Beef and Pork to Broilers,
United States, 1940 to 1958 . . . . . . . . . . . . . 14

Per Capita Consumption Ratios of Beef and Pork to
Broilers, United States, 1940 to 1958 . . . . . . . . 14

Demand for Broilers in the United States, 1940 to 1958 . 50

Demand for Beef Relative to Broilers in the United
StateS’ 1940 to 1958 I O O O O O O O O O O O O O O O 52

Demand for Pork Relative to Broilers in the United
States, 1940 to 1958 . . . . . . . . . . . . . . . . 54

Demand for Beef Relative to Pork in the United
States, 1925 to 1958 . . . . . . . . . . . . . . . . 56

Comparison of Actual and Calculated per Capita
consumption Of Brailers o o o o o o o o o o o o o o o 62

Comparison of Actual and Calculated Retail Prices of
BrOilerS O O O O O O I O I O O O O O O O O O O O O O

C,\
(h

Comparison of Actual and Calculated Relative Retail
Prices of Beef to Broilers . . . . . . . . . . . . . 69

Comparison of Actual and Calculated Relative Retail
Prices of Pork to Broilers . . . . . . . . . . . . . 72

Comparison of Actual and Calculated Relative Retail
Prices of Beef to Pork . . . . . . . . . . . . . . . 75

Chapter I
INTRODUCTION
As an introduction to this study, this chapter is to present the

objectives of the study and to describe some of the changes and trends

in the broiler industry.

objectives of the Study '

The first objective of this study is to examine and explain the
relation of various factors affecting the demand for broilers and to

determine a model to estimate the demand for broilers in the United

S cates.

The second objective of this study is to examine the relative
denand for broilers with respect to beef and pork and to test the

hypothesis that the demand for broilers has risen relative to that of
beef and pork.

Growth of the Broiler Industgz1

Technological advances in production, processing and packaging,

transportation and marketing have enabled the broiler industry to

Parts of the material presented in this section without footnotes
have been developed from previous articles and publications concerning
the broiler industry.. Among these are: Herman Bluestone, Delmarva-
l‘1~ar_:yland-De1aware Broiler Statistics and related data...1934-37, issued
COOperatively by the Agricultural Extension Services of the Unhr ersities
Of Maryland and Delaware and the Haryland-Delaware Crop Reporting Service,
Agricultural Marketing Service, United States Department of Agriculture,
August 1958; Edward Karpoff and John J. Scanlan, "Eggs and Poultry,"
fiarketing, The Yearbook of Agriculture-1954, United States Department

of Agriculture, pp. 467-75; A.B. Kennerly, "The Broiler Industry in
Transition," Texas Agricultural Progress, Volume 4, No. 4, July-
Angust 1958, pp. 7-9; H. P. Mortenson, Trends in the Poultry Mary...
Qfects on thLe Midwest, Bulletin 523, Agricultural Experiment Station,
University of Wisconsin, Madison, Uisconsin, June 1956; and Richard P.
Saunders, Contract:§roiler Growing in Main, Bulletin 571, Maine Agricult-
ural Experiment Station, University of Maine, Orono, Maine, Hay 1958.

expand in the last twenty-five years. A discussion of the changes

occurring in the broiler industry as a result of these factors is an

attempt to provide sufficient background information to describe the

setting of this study.

Prior to the 1930's poultry meat available to consumers was essen-

tially a by-product of egg production. Cockerels and adult birds that

had outlived their usefulness in laying flocks were the main sources of

Sunday chicken dinners. Therefore, the quantity of poultry meat avail-

able varied widely according to the seasons. When broilers were begin-

ning to become economically important during the early 1930's, three

important things happened in the poultry industry. (1) with the

greatest emphasis on comercial size flocks, light breed chickens grad-

ually replaced meat breeds for laying flock replacements. (2) As the

technique of sexing chicks became perfected, only female chicks were
801d with the male chicks being destroyed under the assmption that it

Was unprofitable to grow them out. (3) Feeding, breeding and manage-

ment practices were improved so that more eggs were produced per layer,

thus requiring fewer layers to provide the eggs necessary to supply

the market demand. These changes in production methods meant a de-

crease in the supply of poultry meat from laying flocks.2 These
developments in the poultry industry created a greater opportunity for

8pecialized production of poultry meat. As a result of these changes

Specialized or comercial broiler production was developed to meet this

Opportunity .

Diortenson, Op. cit., p. S.

3
Production. It was 1934 when broiler production had expanded to
the point where the United States Department of Agriculture began to
c: ollect separate statistics for the production of comercial broilers
and farm chickens; annual production estimates are shovm in Figure 1
for both broilers and total chicken meat.3 In 1934 farm chickens were
3 till the source of nearly all chicken meat produced in the United
states. At that time broilers made up only about 4 percent of total
chicken meat production. In the years following 1934 the broiler
industry rapidly expanded. From 1934 through 1951 broiler production
increased at an average annual rate of 21 percent. In 1951 broilers
became the most important source of chicken meat by accounting for 51
percent of the total chicken meat produced in the United States. Since
1951 the year to year percentage increases have not been as large as
those in the earlier years, and the average annual rate of growth was
only 12 percent. For the entire period 1934 to 1958 the average rate

Of growth was 19 percent.4 However, actual growth of production has

Comercial broiler production is meant to apply to specialized
broiler enterprises where broilers include all young chickens of the
heavy or cross-breeds raised only for meat production. More precisely,
the term broiler generally refers to a young chicken (usually under 16
Veeks of age) of either sex that is tender meated with soft, pliable
Smooth-textured skin and flexible breastbone cartilage. ‘

1 Farm chicken production is usually used to denote the production of
chickens on farms not including comercial broilers. Farm chicken meat
is largely a by-product of egg production in the form of old hens and
roosters.

Data for total chicken meat production shown in Figure 1 includes
broiler and farm chicken meat production. Farm chicken meat production
is the area between the two curves.

TAverages for the periods 1934 to 1951 and 1934 to 1958 include
the two years, 1944 and 1946, when production declined below the pro-
duction of the previous year.

million Dollars

Figure 1

Commercial Broiler and Total Chicken‘Heat Production
United States, 1934 to 1958

 

    
    

Total Chickens

“\a1

 

 

 

“
“c:
E
E
c-a Commercial Broilers
2:: “\94
em:
1.
C A a J_ J 1 L L a 4 a 1 1
1934 36 38 4O 42 44 46 48 50 52 S4 S6 58
Time
Figure 2
Gross Income from Commercial Broilers
United States, 1934 to 1958
?l()()() 1
£3<J<Jdl
(Stud
4:00-
2200

 

 

1934 36 38 4o 42 44 46 4s 50 Si 54 i3““§§

5

been larger in the 1950's than in the previous years. In fact actual

growth in the last three years has been the largest in the history of

the broiler industry. Production of broilers in 1958 exceeded 5 billion

pounds which was an all time high, an increase of 16 percent over 1957

and 179 percent over 1950 production, or represented 78 percent of the

total chicken meat produced.

During the period of rapid growth occurring in the broiler industry,
production of farm chickens increased with sharp increases occurring

in the early war years until reaching a maximm in 1943. Production

rose slightly in 1945, but since then production has continued in a

downward trend. Since 1952, farm chicken production has decreased

be low the 1934 level of production.

Combined broiler and farm chicken production rose rapidly in the

Years following 1940. During the war years poultry made up most of the

Black in the meat supply, and this probably resulted in many people

cu ltivating a taste for chicken meat. Total production of all chicken

meat declined in the years following the war up through 1948 and has

tines tended to increase with slight declines occurring in 1952 and
1953.

Gross Income. From $34 million in 1934 gross income from broiler

Production as shown in Figure 2 has rapidly increased. Large increases

in gross income from broiler production occurred in the war years, and
even larger increases occurred in the late forties and early fifties.
In 1958, broiler production became a billion dollar business at the

farm level in the United States, the highest in the history of the

broiler industry. Only four times since 1934 has annual gross income

6

from broiler production decreased. Two of these were only very slight

declines of less than $6 million; the two much larger declines in gross

income occurred in 1946 and 1954. The 1946 decline in gross income was

accompanied by a decline in pounds produced with a rise in price per

pound while the 1954 decline was accompanied by a rise in pounds produced

with a decline in price per pound.

Vertical Integration. Dependent upon the quantity of broilers
produced were the processors and the feed dealers and manufacturers who

had developed businesses in the specialized broiler areas. As pro-

duction expanded in the specialized production areas, it encouraged

these businesses to expand their facilities to handle the increased

quantities of broilers being produced. Increased production required

additional investments for these businesses and, in order to keep their

Operating costs at low levels, they had to maintain a steady volume of

production throughout the year. This resulted in the integration of

various segments within the industry. Contract farming or vertical in-

t:egr:etion developed rapidly in the poultry industry. Broilers are

almost entirely produced on some type of integrated basis; in most areas

the feed dealers are the major contractors of broiler production but

80mg contracting is done by the processing plant.5 Integration in

the broiler industry has resulted in placing larger quantities of broilers
on the market with less seasonal variation in the supply and has trans-

E""‘l-‘I‘ed some of the uncertainties faced by growers to the contracting
a8‘tt'lozzies.

\

5United States Department of Agriculture, Contract Farming and
ation in Agriculture, Agriculture Information Bulletin

Vxfirtical Integr
“0- 198, July 1958, p. 17.

 

7

Technological Advancements. In spite of the downward trend in the

farm price of broilers since the late forties, expansion in broiler
production has been made possible by greater efficiency in production.
Increased feed efficiency of broilers and other technological advances
in production through research development have made much of this
possible.

In the earlier years of the broiler industry most of the processing
was done either on the farms or by small processors who had limited
equipmnt and facilities or the live birds were shipped from the pro-
duction areas to processing plants located in the central markets. In
recent years processing plants in the production areas have greatly
increased in size and have replaced obsolete equipment with modern
facilities which have resulted in lower unit cost and improved sanitary
conditions. One of the greatest developments in processing was the
adOPtion of the "assembly line" method of processing broilers.

Advancements in refrigeration and transportation have extended the
market areas of various broiler producing areas. Advancements made in
Infiltcn: transportation have caused a shift in the transportation of
broilers from rail to motor carriers. Time in transit for broilers
from: the processors to the market areas has been greatly reduced by
"19:01 carrier shipments due to the availability and flexibility of
their services, and hence better quality broilers now reach the retail
outlets.

Changes at the retail level have also occurred over the years.
YeQrs ago, poultry products were regarded as of minor importance in

retail sales and little attention was directed to the packaging and

8

merchandising of poultry products. In the years prior to about 1950,

most of the broilers sold at retail outlets were either New York

dressed or whole eviscerated. In recent years, most of the broilers

marketed at retail have been sold as ready-to-cook broilers either as

cut-up broilers or cut into specific parts; this added service was

done to meet the demands of consumers. Many retailers have recently

directed more attention to the merchandising aspects of displaying the

product and provide a continuous supply of broilers for their customers.

To this point the changes occurring in the production, processing

and marketing of broilers have been presented, but no attempt has been

made to describe the effects of these changes on the price level or per

capita consumption of broilers. The remaining portion of this chapter

will emphasize these changes.
Prices. Farm prices of broilers in the United States, as shown

in Figure 3, rose in 1940 above the 1939 level of 17.0 cents per pound,

which is still the lowest farm price in the history of the broiler

industry, and continued in an upward trend through 1948. By 1948 farm

Prices had risen to 36.0 cents per pound. Since 1949 the trend in farm

Prices of broilers has been downward. In only three years since 1949

h‘Ve farm prices showed an increase over the preceding year, and since

1953 farm prices have declined in every year except one. In 1958 the
a"etnmge price per pound received by producers declined to 18.5 cents.
“18 was the fourth lowest annual average price received by producers

.1t‘ce statistics have been recorded for counercial broilers and the

1%‘t price since 1941.

Cents Per Pound

Figure 3

Average Retail and Farm Prices of Broilers
United States, 1940 So 1958

  
 
 
  
 

  

 

 

     
 

   
   
 

 

 

 

60 ~
50 L (ready-to-cook)
40 a
W Farm Price
30 r (live broilers)
20 L
10 ~
0 L 1 1 l J L l l L 1'
1940 42 44 46 48 50 52 54 S6 58
Time
Figure 4
Retail Price, ParmflValue and Pareretail Spread for Broilers
United States, 1950 to 1958
60.
Retail Price
SOL (ready-to-cook)
40 .. arm Value
'8 13/?(per pound of broiler
3 ready-to-cook)
an 30+
H
o
a.
N “—f
m 20 L I:
g Pareretail Spread
c: (ready-to-cook basis)
lOr
0 l_ 4 L I J J _l 1

 

 

1950 51 52 53 54 55 56 57 58
Time

10

Retail prices of broilers, as shown in Figure 3, have tended to
move in the same direction as changes in farm prices since 1940 except
for two years. Retail prices of broilers increased in each year from
1940 through 1952 except for 1949 and reached their highest point of
61.6 cents per pound in 1952. Since 1953 the retail price of broilers

tended downward and declined to 47.4 cents per pound in 1958. Only one
year since 1952, 1955, showed a rise in the retail price above the
previous year.

The farm-retail price spread6 for broilers, unlike that for
my other food products, has not tended upward in recent years. This
is indicated by the data for the retail price and farm-value of broilers
shown in Figure 4. Since 1950 the farm-retail spread has remained
relatively stable and has fluctuated only in the range of 21.1 to 23.0
cents per pound. However, the farmers' share of the retail price has
fallen from 64 percent in 1950 to 54 percent in 1958 because of the
declining price of broilers.

Declining farm and retail prices and relatively stable farm-retail
3Preads indicate that the broiler and poultry industry have increased
ef-‘Eil.c:iency in production, processing and marketing enough to offset

°°3ts from increased services, higher wages for labor and price increases
of other inputs.

\

th 6Farm-retail price spread as used here means the difference between

due- retail price and farm value where farm value is the payment to pro-

DES-en for 1.37 pounds of live weight which is equivalent to 1 pound
ready-to-cook weight.

11

221' Capita Conmtion. Annual per capita consumption of chicken

meat in the United States, as shown in Figure 5, fell within the range
of 13 to 16 pounds per person in the period from 1909 through 1937.
Although a decline in per capita consumption of chicken meat of only 0.9
pound occurred in 1938, it was enough to drop annual per capita con-
sumption to the lowest point since 1909 which is the beginning date of
available statistics. Per capita consumption rose rapidly during the

war years with peaks in 1943 and 1945 of 23.0 and 21.6 pounds respectively.

By 1947 consumption had fallen to 18 pounds. This was followed by an

upward trend in per capita consumption in the following years. Since

1947 the only major decline in consumption occurred in 1955. Per capita

consumption of chicken meat rose to an all time high of 28.5 pounds in
1958. The rapid increase in per capita consumption of chicken meat in

recent years can be attributed to the increased per capita consumption

of broilers which is also shown in Figure 5. In the mid-thirties when

the broiler industry was in its infancy, per capita consumption of

brailers was less than one pound. By 1940 per capita consumption of

broilers had reached two pounds. Five years later broiler consumption

had increased to six pounds per person. Per capita consumption declined

in 1946 to five pounds but has increased rapidly in each of the years

Since then. In 1958 per capita consumption of broilers rose to 21.7

pollnds or 76 percent of total chicken meat consumption.

gelative 53”“. Only the trends in the actual broiler prices and

Per capita consumption have been mentioned above. Since broilers nust

e("“Pete at retail outlets it is perhaps apprOpriate to examine the

12

wmma as some .mmumum asses:

.mfimmm xooouOunmcnmz .muoafioum can unoz coxofiso anuOH mo coaueESmcou newemo pom

mm,

mafia
om mgr ow mm om mm on

ma

.m spawns

 

 

q

— fl 1 A1 d a A

    
    

muoaaoum

7 new: £33.30 .230“.

d

onH
.

 

OH

ma

ON

mu

spunoa

13

changes in retail broiler prices and consumption relative to the retail
prices and consumption of two products, beef and pork, assumed to be
their close competitors.
The trends of the ratios of the retail prices of beef and pork to
the retail price of broilers, as show in Figure 6, rose over the period

1940 to 1958. with the exception of the period 1948 to 1953, the ratios

moved in patterns similar to each other. In the period 1948 to 1953,

the ratios moved in diverging directions but came back together in 1953.
Through the entire period the ratio of the price of beef to the price
of broilers exceeded the ratio of the price of pork to the price of
broilers.

The ratios of the per capita consumption of beef and pork to the

per capita consumption of broilers, as shown in Figure 7, tend to change

over time in similar patterns. For both there is a definite downward

trend. In each of the years but two, 1947 and 1953, changes occurring
in the ratios have been in the same direction. The ratio of the per
capita consumption of pork to broilers exceed or equaled the ratio of
the: per capita consumption of beef to broilers from 1940 through 1952,
but since 1953 the pork ratio has been less than the beef ratio.

To sumarize this section, new technology in the broiler industry
in recent years has made possible the marketing of larger quantities of
bro11ers at lower prices. Because of this new technology and steadily
flee-1I‘l.ning retail prices relative to competing meats, broilers now

cmrise a larger prOportion of the diet of the American consumer.

14
Figure 6

Retail Price Ratios of Beef and Pork to’Broilers
United States, 1940 to 1958

  
 

 

 

1s75r
1.50‘
__1.s_e_e_£_ ,
1.25r Broilers -~\¢;; ,/
/
O \ /
‘9' L \/
3.000 I y-A‘J
3
H .. / \~../
2. '75 Fork
., Broilers
”.1
u .50“
0
cs
.25’
0‘ L L 1 L _1 I J. a A L
1940 42 44 46 48 50 52 54 . 56 58
Time
Figure 7
Per Capita Consumption Ratios of Beef and Pork to Broilers

35

30

25+

20

15

10

Per Capita Consumption Ratios

United States, 1940 to 1958

 
   
   
 

Fork
Broilers

lr\fflbv-

Beef
Broilers

 

Chapter II
CONCEPTS OF DEMAND

Construction of economic models involves knowledge of the relevant
economic theoy before statistical estimation because statistical
methods require knowing something about the model underlying the problem
under consideration. Therefore the purpose of this chapter is to
present economic theories and concepts of demand which are believed to
be relevant to this study; discussion will be directed to demand curves,
relative demand and demand elasticities.1

Demand, as used in economics, is defined as a schedule of the
various quantities of a good which consumers are willing to buy from
the market at various alternative prices, other things being equal.

Leftwich2 lists the most important factors affecting the quantity
of goods which consumers will take from.the market as: (1) the price
of the good, (2) consumers' tastes and preferences, (3) the number of
consumers under consideration, (4) consumers' incomes, (5) the prices
of related goods, and (6) the range of goods available to consumers.
when the latter five factors are held constant, as required by the

stated definition of demand, we assume an inverse relationship to

 

1Generally accepted economic theories and concepts which are pre-
sented in this chapter without footnotes are develOped from material
presented in various textbooks in economics. See: Joe S. Bain, Price
Theoyy, Henry Holt and Co., New‘York, 1952; Kenneth E. Boulding,
Economic Analysis, 3rd Edition, Harper &‘Brothers, New York, 1955;
J. R. Hicks, Value and Ca ital, 2nd Edition, Oxford University Press,
London, 1946; Richard R. Leftwich, The Price §ystem.and.Resource
Allocation, Rinehart & Co., Inc., New'York, 1955; George J. Stigler,
The Theogy of Price, Revised Edition, Macmillan Co., New York, 1952;
Sidney‘weintraub, Price Theogy, Pinman Publishing Corporation, New‘York,
1949.

zLeftwicb, 02. cit., p. 27.

l6

exist between the price of the good and the quantity taken. Hence, the
higher the price of the good the less consumers will take, or the lower
the price of the good the greater the quantity consumers will take.
Thus, the stated definition is essentially the same as the "law of
demand" as formlated by liarshall.3 It states that the greater the
amount to be sold, the smaller nust be the price at which it is offered
in order that it may find purchasers; or in other words, the amount
demended increases with a fall in price and diminishes with a rise in
price. Harshall also stated that the ultimate regulator of all demand
is consumers' demand.4
Qemand Curves

In the presentation and discussion of the definition of demand the
inverse relationship existing between the price and quantity taken of
the product was introduced. This relationship can be sumarized into
a demand schedule which shows the amount of the product that will be
taken at various prices. When this schedule is presented graphically
with price on the vertical axis and quantity on the horizontal axis and
the various points are connected, the resultant is referred to as a demand
curve. Since there is an inverse relationship between price and quantity,
the demand curve slapes down and to the right. In order that the quantity
demanded at each price level has a meaning, the time period involved
mast be specified. Thus, a demand curve shows the quantities of a pro-
duct per unit of time which consumers will take from the market at

various alternative prices, other things being equal.

 

3A1fred Marshall, Principles of Economics, 8th Edition, Macmillan 6:
Co., London, 1920, p. 100.

m0: 90 92-

17

Demand curves as discussed in this section may apply to either the
individual consumer or market demand curves for products for both are
defined in essentially the same way. The market demand curve for a
product is an aggregation of the demand curves of all individual con-
sumers for the product.

A distinction should be made between a change in demand and move-
ment along a single demand curve for a change in the quantity taken does
not necessarily mean that there has been a change in demand. ,Movement
along a given demand curve refers to a change in the quantity taken
resulting from.a change in the price of the product with other things
influencing demand remaining unchanged. A change in demand refers to a
shift in the entire demand curve. It implies that consumers are now
willing to take a different quantity of the product at each price level.
An increase in demand refers to the case where the quantity taken
increases at all prices; similarly, a decrease in demand is where the
quantity taken decreases for each price. Circumstances causing shifts
in the demand curve to occur are changes in the "other things being
equal" phrase of the definition of demand, such as changes occurring
in the incomes or the tastes and preferences of the consumers.
Relative_yemand

Since relative demand is being studied in this thesis, a section
on the theory of relative demand appears to the relevant of this

chapter.

18

Articles by Clawson5 and Staehle6 were attempts to revive the use
of relative prices to the position in economic analysis which they lost
to the influence of income during the "keynesian revolution." Although
the article written by Clawson preceded the one by Staehle, only the
latter will be discussed in detail in this chapter. The reason is
that the article by Staehle pertains more to the theoretical development
of the analysis while Clawson directs his presentation more to the
features and flmplications of graphic methods which relate the ratio of
prices of two commodities to the ratio of their quantities. A discussion
of the article by Clawson will be included in Chapter III.

Staehle states that:

"In those old-fashioned branches of economic reasoning
relative prices used to play a preponderant role" and "he

(Staehle) believes that the moment has come to put in a

‘modest plea in favor of increased attention to the influence

of relative prices."7 "If the neglect of the effect of changes

in relative prices were rigorously limited to theoretical

work, with conclusions clearly labeled as inapplicable to

the real world, there would be little harm in this type of

economics. The results would simply fall into that already

so heavily filled box marked 'Partial, i.e., incomplete
analysis: Never to serve economic policy'.”

 

5Marion Clawson, "Demand Interrelations for Selected Agricultural

Products," Quarterly Journal of Economics, Vol. 57, February 1943, pp.
265-302.

6Hans Staehle, “Relative Prices and Postwar Markets for Animal Food
Products," Quarterly Journal of Economics, Vol. 59, February 1945, pp.
237-279.

7;§id., p. 237.

81bid., p. 239.

19
Staehle developed a theoretical foundation for the individual
consumer demand for a commodity. Re assumed the demand of an individual
for a given commodity (x) to be dependent upon his real income (r) and
the price (p) of the commodity. He assumed the particular relationship
to be
x =- f(r) + (pm).
If this function is considered to be comon to all consumers and there

are n of them, then the consumption per person would be
i = Sf‘r) + géng .
n n

If the influence of income were linear,
f(r) 3 a + br,

consumption per person would become

1? - £65) + age)...

That is to say that if the individual function connecting consumption
and income, f(r), were linear, average consumption equals the consumption
that corresponds to the average income. In that case the average income
is the only parameter of the frequency distribution of incomes which
needs to be taken into account. The simple additive relation assumed
between the influence of income and price is quite arbitrary. It means
that a price change will induce individual consumers to change their
levels of consumption quite independent of their incomes.9
Demand Elasticities

The theory presented above assumed that the factors affecting

demand were constant except for the price of the commodity; however, in

 

9;bid., p. 247.

20
real world the other factors are constantly changing. The purpose of
this section is to present some of the elasticity concepts which are
used to distinguish and measure the functional relationships existing
in the real world between prices, consumer incomes and quantities of
products demanded.

Price Elasticity. Price elasticity of demand, commonly referred
to as elasticity of demand, measures the relationship between prices and
the quantities taken. It is defined as the percentage change in quantity
taken associated with a one percent change in price.

Price elasticity of demand can be measured between two points (arc
elasticity) or at a single point (point elasticity). If the two points
between which the arc elasticity is being measured are moved closer to-
gether, they will eventually merge into a single point; therefore,
point elasticity is the same as are elasticity when the distance between
the two points. is infinitesimally small.

For a very small change in price, the mathematical expression for
the price elasticity of demand is:

41G
q

Ea_________-
den... 4?

P

l?

.43..
q
where p is the price and q is the quantity taken.

Income Elasticity. The income elasticity of demand is defined as
the percentage change in the quantity taken associated with a one per-
cent change in income or as the ratio of the relative change in quantity
taken to the relative change in income. iHathematically this can be

expressed as:

21

 

E 43L.
Ida.___£_=d_1_c I
an; A: q
I

where q is the quantity taken and I is the income.

If the income elasticy of demand is positive, then an incease in
income results in an increase in purchases; this is the case for most
goods. If the income elasticity is negative, an increase in income
results in a decline in the quantity taken; goods of this nature are
usually referred to as inferior goods.

Cross Elasticity. Cross elasticity of demand measures the extent
to which commodities are related to each other. It is defined as the
percentage change in the quantity taken of one commodity associated with
a one percent change in the price of a related commodity. Cross elasti-
city of demand for a commodity with respect to a second commodity can

be expressed mathematically as:

(JQI
XEd a q]_ =- 4 ‘11 P2
A P2 4 P2 ‘11

P2
where q1 is the quantity of the first commodity and p2 is the price of-

the second commodity.
If the cross elasticity between the two commodities is positive,
the commodities are considered as substitutes for each other. If

negative, the two commodities are considered to be complements.

Chapter III
REVIEW OF PREVIOUS STUDIES

The purpose of this chapter is to discuss some of the previous
demand studies for poultry and competitive products. Individual studies
to be discussed in this chapter vary in nature from the more advanced
mathematical and statistical approaches to those largely descriptive
and based upon interviews with consumers at the retail level. Some
attention will be directed to the techniques used and variables included
in the analyses as well as to brief descriptions of the results obtained.
Aggregate Demand for _B_t;gi_l_e_r_§_

Christensen and Mighell1

suggested that the changes in national
demand at the consumer level for chickens and eggs between the two
5~year periods, 1925 to 1929 and 1945 to 1949, were affected by changes
in (1) total pOpulation, (2) real incomes of consumers and their dis-
tributions, (3) food preferences and (4) supplies of substitute foods.
Changes in the quantity and valLe of chicken-meat consumption have been
affected by changes in character of the product. The character of
chicken meat available to consumers has changed with the large expansion
in the commercial broiler business. This would have the effect of
expanding consumption of chicken meat even though there had been no

change in price relationships between products.2 They indicate that

unpublished studies of the Bureau of Agricultural Economics suggest

 

1Raymond P. Christensen and Ronald L.‘Mighell, Competitive Position
of Chicken and Egg Production in the United States, Technical Bulletin
No. 1018, Bureau of Agricultural Economics, United States Department of
Agriculture, August 1950, p. 35.

 

21bid., pp. 37-38.

23
that the elasticity of demand for chickens is higher than for eggs,
perhaps as high as -l.0.3

Riely,4 Gieseman,5 and Parsons6 used data obtained from the
Michigan State University Consumer Panel to study consumer demand for
various meat products in Lansing,‘Michigan.

Riley found that in any given week approximately one-third of the
panel families bought some kind of poultry meat and that poultry ex-
penditures made up about 12 percent of the total meat bill over a two-
year period. Fryers were the most flmportant chicken product purchased?
He used a seven variable regression model to estimate the quantity of
poultry purchased. The independent variables included in the model for

the period JUIy 19514December 1952 were:

X1 - price of beef

X2 ‘ price of pork

x3 - price of sausage

X4 - price of poultry

X5 ' price of fish

X6 - average weekly income per family

The multiple correlation coefficient obtained was 0.58. Prices of poultry

 

31bide , pp. 430
4Harold D. Riley, Some'Heasurements of Consumerégemand for Heats,
Unpublished Ph.D. Thesis, Department of Agricultural Economics,‘uichigan
State College, 1954.

5Raymond U. Gieseman, Changes inQEeat:gurchases During_l952 and
1953 Related to Family Characteristics, Unpublished‘M.S. Thesis, Depart-
ment of Agricultural Economics, Michigan State University, 1956.

Querril He Parsons, Newspaper Advertising;of:Meattgroducts and:;£g
Relation to Consumer Purchases ingggpsingl‘uichiggg, During 1956, Un-
published M.S. Thesis, Department of Agricultural Economics, Michigan
State University, 1958.

7Riley, Op. cit., p. 185.

24
and pork were both significant explanatory variables with the price of
poultry being the most important. The price elasticity of demand
measured at the mean was -l.43, and the cross elasticity of demand with
reSpect to pork prices was 1.02.8

Parson's analysis indicated that when newspaper advertising Space
during 1956 was allocated among the broad meat groups, poultry ranked
third in the classification behind beef and pork. However, when advertis-
ing Space was allocated according to specific meat items, broilers were
given more space than any of the other meat items studied.9 He used a
regression equation which expressed the quantity of broilers purchased

as a function of the following variables:

X1 price of beef

X2 = price of pork
X3 = price of broilers

x4 3 temperature

X17= advertising space featuring broilers as a Special
The adjusted multiple correlation coefficient obtained was 0.89. Adverti-
sing space for broilers proved to be a highly significant variable in the
equation. The regression coefficient for advertising Space indicated that
at the mean a ten percent increase in the amount of advertising space fea-
turing broilers as a Special was associated with an increase of 1.49 percent
in the quantity purchased. The amount of advertising Space was more closely
related to the quantity of broilers purchased than for any other meat
studied. The coefficient of simple correlation between advertising space
and the quantity of broilers purchased was 0.79. Regression coefficients

for the prices of broilers and pork were significantly different from

zero at the five percent level.10

 

§;g;g., p. 189.
9Parsons, Op. cit., pp. 19-22

1°Ib1d., pp. 71-72.

25

Baker and Farris11 investigated the short run consumer demand
relationships for broilers. Data used in this study were obtained from
thirteen retail food stores in the Lafayette-West Lafayette, Indiana
area covering an eleven week period, April 1, 1956 through June 16, 1956.
Multiple correlation techniques were used to analyze the data. The
combination of independent variables chosen as the best predictor of
changes in broiler sales included the prices of broilers, short loins
and chuck roast. The coefficient of multiple determination obtained
from.these variables was 0.933. Using these variables the price elasti-
city of demand for broilers in the Lafayette-West Lafayette area for the
thme period studied was estimated to be -1.2. It was pointed out that
this coefficient also reflected broiler advertising influence; advertis-
ing and price influences of broilers could not be separated because of
the relatively high intercorrelation between the two variables. The two
cuts of meat most closely associated with broiler sales were short loins
and chuck roast; both were found to be competitive with broilers. The
cross elasticities of demand for broilers with respect to the price of
short loins and the price of chuck roast were 1.6 and 1.3 respectively.
It appeared in general that broiler sales were more influenced by the
price and advertising of beef cuts than pork cuts.

Nordin, Judge and Wahby12 set up a detailed model to estimate the

retail demand conditions for beef, pork and poultry products. As

 

“George 1.. Baker, Jr. and Paul L. Farris, An Empirical Estimate of
Broiler Price Elasticity and Cross glisticities at Retail, Paper pre-
sented at the Joint Annual Meeting of The American Farm Economic Association
and The Canadian Agricultural Economics Society, Winnipeg, Canada, August
20-22, 1958.

12J. A. Nordin, George G. Judge and Omar Uahby, Application of
Econometric Procedures to theppemands for Agricultural Products, Research
Bulletin 410,.Agricultural Experiment Station, Iowa State College, Ames,
Iowa, July 1954.

26
originally constructed, the model was overidentified. Consequently,

some of the variables were drOpped out to make the model uniquetly de-

termined. The variables included in their uniquely determined model

were:
Y1 per capita quantity of pork sold at retail
Y2 retail price of pork
Y3 retail price of beef
Y4 retail price of poultry products

retail price of dairy products

retail price of oleomargarine

per capita quantity of beef sold at retail

per capita quantity of poultry products sold at retail
time

retail price of other foods

per capita disposable income

weighted average of 24 for the 5 preceding years.

r4
\1
IIIIIIIIIIIIIIIIIII

Retail prices and disposable income were deflated by the consumer price
index. The data were converted to logarithms and covered the period
1921 to 1941. The three demand equations included in their uniquely

13

determined model were:

Pork
Y1 = -O.91Y2 + 0.60Y3 + 0.87Y4 - 1.23Y5 - 0.91Y6 - 0.0321 + 0.1623

+ 0.7624 + 0.29Z5 + 2.70

Beef
Y7 =- 0.533!2 - 0.77313 + 0.67Y4 - 0.221:S - 1.09Y6 - 0.0221 + 0.2923

+ 0.6524 - 0.1225 + 3.06

Poultry products
Y8 3 0.12Y2 + 0.28Y3 - 0.68Y4 + 0.22Y5 + 0.31Y6 + 0.00221

+ .O.3623 + 0.5324 + 0.2825 + 0.42
Since the variables are expressed in logarithms, the elasticities
can be read directly from.the equations. Hence, it can be seen that a

one percent rise in the price of poultry products can be expected to

 

13Ibid., p. 1016.

27

result in a 0.68 percent fall in the quantity of poultry products pur-
chased at retail and that a one percent rise in income can be expected
to result in a 0.53 percent rise in the quantity of poultry products
purchased at retail.

Demand studies for products competitive with poultry, which are

similar to those presented above, have been published by Shepherd,14

Wallace and Judge15 and Working.16
Consumer Attitudes and Preferences

Consumers have preferences which influence their choices. It is
difficult to isolate these preferences, but since consumers make a
selection it is possible to observed some of the factors which may have
influence their choice. Discussion to this point has been directed more
to the effects of retail prices on conmmmptien. It now seems apprOpriate
to present some of the studies which concern individual consumer pre-
ferences and other factors affecting individual purchases.

Harold Smith17 interviewed 1,156 housewives in Baltimore, Maryland;

Trenton, New Jersey; and New‘York, New'York to examine consumer purchases

 

14Geoffrey Shepherd, Chagges in theigemand for‘Meat and Dairy Pro-
ducts in the United States_§ince 1910, Research Bulletin 368, Agricultural
Experiment Station, Iowa State College, Ames, Iowa, November 1949.

15Thomas D. Wallace and George G. Judge, Egonometricfigpalysis of
thg%§eef and Pork Sectors of the Econom , Technical Bulletin T-75, Agri-
cultural Experiment Station, Oklahoma State University, Stillwater,
Oklahoma, August 1958.

16Elmer J. Working, Qemand for Meat, Institute of Meat Packing,
University of Chicago, Chicago, Illinois, 1954.

17Harold D. Smith, Consumer Preferences and Buying Habits for Chicken,
Bulletin x-s, September 1951 and Consumer Demand for Chicken as Related
to Demand for Total Meat, Miscellaneous Publication No. 112, October 1951,
Department of Agricultural Economics and Marketing, Maryland Agricultural
Experiment Station, University of Maryland, College Park, Maryland.

28

and consumer demand for chicken as related to the demand for total meat.
Special attention was given to the effects of factors such as consumer
income, religion, race and size of family. He found some variations in
consumer demand for meat and chicken between the three cities. High
income peOple were found to consume more meat and chicken than low
income peOple. In Baltimore and Trenton per capita consumption of
chicken was highest for families making $6,000 or more a year, and for
New‘York consumption was highest in families with incomes between $4,000
and $6,000. Families of less than three members consumed more meat and
chicken per capita than larger families. Although large families consumed
less meat and chicken per capita, chicken represented a larger part of
their total meat consumption.18 The kind of meat served most frequently
in all three cities was beef, pork, chicken and lamb in that order. More
than 70 percent of those surveyed served beef more frequently than other
‘meats. Chicken was generally purchased on Friday and Saturday and
was served on Sunday. Low income families tended to serve chicken on
Sunday more often than high income families.19

Rask0pf20 surveyed 3,611 families in Tennessee in 1955 to evaluate
the factors believed to affect per capita consumption of broilers. The
family characteristics studied in relation to their effects on per capita

consumption of broilers were religion, race, occupation, residence, size

 

18Smith, Consumer Demand for Chicken as Relatedito Demand forgggtal
Meat ’ pp a 34-35 a

19Smith, Consumer Preference and Buying Habits for Chicken, pp. 21-23.

20B. D. RaskOpf, Consumer Ratinggoffi Broilers in Tennessee, Bulletin
No. 255, November 1956 and Factors Affecting Per Capita Consumption of
Broilers in Tennessee, Bulletin No. 288, August 1958, Agricultural Experi-
ment Station, University of Tennessee, Knoxville, Tennessee.

29
of family, education of family head, family income, members of family
purchasing broilers, preferred method of cooking broilers and per capita
consumption of other meats. RaskOpf found that small families consumed
‘more broilers annually per capita than large families and that families
with higher incomes generally consumed more broilers than did lower
income families. He also found that each increase of one year in the
education of the head of the family was related to about one-third of a

pound increase in annual per capita consmmption of broilers.21

22 suggested that apparently an "educational

Branson and Mountney
rebound" exists against chicken. In their survey of consumers in Houston,
Texas they found that the more educated the housewife the greater the
likelihood that her family will perfer some meat other than chicken. In
the medium income families they found that only 6 percent of the families
where the homemaker had a grade school education served chicken less
than once a week, compared to 16 percent where the homemaker was college
educated.

Chicken was found to be the first choice meat of only 17 percent
of the families surveyed in Houston. However, when asked for their
second preference, chicken ranked far ahead of any other meat. Chicken
was definitely a second choice meat among both high and middle income

families; only among low income families was chicken the preferredmeat.23

 

21RaskOpf, Factors Affecting Per Capita Consumption offigroilers in
Tennessee, pp. 3-4.

22Robert E. Branson and George J.‘Mountney, Consumer Attitudes and
Preferences RegardingyChicken, Bulletin 895, Texas Agricultural Experi-
ment Station, College Station, Texas,‘Harch 1958, p. 9.

zalbids’ pp. 3-40

30

Most families that had increased chicken purchases during the year
indicated that they did so because the family increased in size or
because chicken was more economical than other meats.

25 indicated that of the Portland and South Portland,‘uaine

Lebrun
consumers interviewed 78 percent purchased their poultry meat on Friday
and Saturday and that nearly 60 percent of the poultry meat was consumed
on Sunday. The factors which Lebrun found to be related to purchases of
poultry meat included family income, occupation of the wage earner, size
of the family and religion.

R. C. Smith26 examined the relationship between consumer purchases
of fryers and such factors as prices of different kinds of meats, quality
factors, merchandising practices, income, size of family, and nationality
of consumers. He found that approximately 63 percent of the changes in
weekly volume of fryer sales in two large retail chain firms in the
Philadelphia, Pennsylvania - Wilmington, Delaware area were associated
with changes in prices of fryers, stewers, beef, pork loin, veal and
lamb. Fryer sales were found to form a pattern Opposite that of pork
loin. A positive relationship existed between the price of fryers and
the price of beef. Price and volume of fryer purchases varied inversely;

changes in retail price were found to be associated with approximately

41 percent of the changes in the volume purchased.

 

241b1d., p. 7.

25Edmond J. Lebrun, Consumerggurchase and Use of Poultry Heat,
Bulletin 524, Maine Agricultural Experiment Station, University of Maine,
Orono, Maine, January 1954, pp. 10-12.

26R. C. Smith, Factors Affecting Consumer gurchases of Paying
Chickens, Bulletin No. 298 (Technical), Agricultural Experiment Station,
University of Delaware, Newark, Delaware, July 1953, pp. 30-31.

Relative Demand

A somewhat different approach to the measurement of demand is
through the use of the techniques of relative demand. In addition to
the previously mentioned articles by Clawson and Staehle, other studies
using this type of analysis have been made by Szatrowski,27 Shepherd,

28 and Waugh.29

Purcell and Manderscheid
The method used by Clawson was completely graphic. In his analysis
he presented illustrative charts to portray the‘relationships between
commodities within the meat, fruit and grain groups. The commodities
included in the meat group were beef, pork, veal, lamb and chicken. He
constructed the charts mentioned above by plotting the price ratios of
the two commodities on the vertical axis and the consumption ratios on
the horizontal axis. After he plotted the data for the period studied,
he then determined the relationship between the variables by inspection
and drew a line through the observations to represent the average
relationship. He mentioned that the advantage of this type of diagram
was that it showed the relation between four variables rather than two

as most price-quantity diagrane do. However, he added that about all

that can be said from this type of diagram is that a close relation exists

 

27Zenon Szatrowski, "Time Series Correlated with Beef-Pork Con-
Sumption Ratio," Econometrics, Vol. 13, No. 1, January 1945, pp. 60-78.

28Geoffrey Shepherd, J. C. Purcell and L. V. Manderscheid, Economic
Analysig of Trends in Beef Cattle and Hog Prices, Research Bulletin 405,
Agricultural Experiment Station, Iowa State College, Ames, Iowa, January
1954.

29Frederick‘v. Waugh, Graphic Analysis in Economic Research, Agri-
cultural Handbook No. 84, Agricultural‘uarketing Service, United States
Department of Agriculture, June 1955, pp. 48-49; "A Partial Indifference

Surface for Beef and Pork." 102W, Vol. 38, No. 1,

February 1956, pp. 102-112; ra hic Analysis in Agricultural Economics,
Agriculture Handbook No. 128, Agricultural Mhrketing Service, United

States Department of Agriculture, July 1957, pp. 46-47.

Lo)
[*3

or does not exist between the four variables or that this relation has
shifted during the time period studied. Since the extent of a shift is
not apparent until a few years after it occurs, this imposes a limitation
on using this type of diagram.for forecasting purposes.

In one of the charts Clawson plotted the cattle/hog price ratio
against the beef/pork consumption ratio, then he drew in two lines to
represent the relationship between the two ratios. Years 1899 to 1914
and 1934 to 1939 were found to have identical relationships and were
represented by one line. The other line drawn was for the period 1919
to 1933. Years 1915 to 1918 were a transition period and did not fit
well with either the earlier or later periods. .mmong other charts pre-
sented with the observations for the chicken/cattle price ratio and
chicken/beef consumption ratio, but Clawson did not draw a line of
average relation for the observations due to extreme clustering of the
dots. He suggested that the relation was slight and uncertain and that
the clustering of years probably reduces the reliability of the results
considerably.3o Clawson claimed that none of the charts presented or
made and not presented provided any evidence that the general price level
or real incomes of consumers affected the relation in any way. Years
of widely divergent price and income conditions not infrequently fell
close together. The price ratio apparently eliminated most of the effect
of changing prices and changing real income.31
4 Clawson stated that if the quantity of beef consumed rose in com-

parison with the quantity of pork consumed, it would not be unreasonable

 

3oClawson, 02. cit., pp. 271-83.

31;bid., p. 278.

33
to expect the price of beef to fall in comparison with the price of pork.

I

The closeness of this relation and the extent of the price fall would
be significant to producers of each commodity.32
The elasticity of the lines of average relation depend upon the
shape and position of the line. Straight lines inclined downward to
the right always have decreasing elasticity as X increases, whereas
curved lines convex when viewed from the point of origin generally
have constant or increasing elasticity, although they may be inelastic
throughout.33

To prove his hypothesis that the principle factor responsible for
the deviations in consumption of meat was the relative price of meat,
Staehle presented the relationships by graphic analysis. His first step
was plotting time series data for annual consumption of meat per person
and annual income per person on a diagram with a derived function re-
lating annual meat consumption to levels of income per person.34 He
named the deviations between the two levels of consumption the "in-
fluence of price." He then plotted the influence of price and relative
retail price of meat which he obtained by establishing the money price
for each year and deflating it by the index-number for the cost of food.
He then estimated the relation between the two variables and drew a

line through the observations to represent the relationship.

 

32 Ibid. p. 255.
33 Ibid., p. 278.

34The function relating annual meat consumption and levels of in-
come per person was derived from family budget data for three levels of
income. The function, referred to in Chapter II as f(r), was fitted
from this data by graphic interpolation and was found not to deviate
substantially from a straight line.

 

34

In order to verify his hypothesis he stated that it was necessary
to study the relation between the relative price and the influence of
price and if the relation was close and exhibited a negative slope, his
theory could be accepted as verified.35 This he obtained. Staehle then
proceeded to examine various commodities with the methods described
above and with the relative price and quantity ratios as used by Clawson.

Szatrowski, whose article was published about the same time as those
by Clawson and Staehle, used multiple regression analysis to statistically
examine the relationship between the beef-pork consumption ratio and

other time series data for the period 1901 to 1941. In the analysis the

beef-pork consumption ratio was as a function of the following variables:

X1 ‘ time trend

X2 - beef-pork price ratio

X3 - consumption ratio of preceding year

X4 = cattle-hog papulation ratio

X5 = corn yield of previous year in bushels per acre
X5 = income, dollars per capita.

The data were analyzed for two separate periods, 1901 to 1920 and 1921
to 1941, and then for the whole period. Results of the study indicated
that the relationship between the beef-pork consumption ratio, the beef-
pork price ratio and the beef—pork consumption ratio of the preceding
year were essentially the same for the two separate periods.

Shepherd 55_gl, used techniques similar to those described in the
above studies to examine the relative demand of beef to pork. In their
analysis beef and pork retail price ratios were plotted against beef
and pork consumption ratios for the period 1925 to 1952. The plotted
observations showed a tendency to cluster around three different lines

of relationship, with negative slaped like demand curves. Their analysis

 

351bid., p. 257.

35

indicated that a shift occurred in the relationship during the years

1932 to 1934 and again in 1947. The line drawn to represent the relation-
ship during the war years was nearly horizontal; they attributed this

to the price ceilings in effect during the war which kept the price
relationships unchanged regardless of changes in relative consumption.
Their analysis revealed that the relative demand for beef rose over the
period studied. Further investigation indicated that the demand for

beef had increased over the period while the demand for pork had de-
creased.

Waugh used a still different approach in his analysis. He con-
structed and attempted to justify what he called "a partial indifference
surface for beef and pork." In his analysis he plotted the annual per
capita consumption of pork on the vertical axis and per capita con-
sumption of beef on the horizontal axis of what appears to be a cross
section of a three-demensional surface. Through each dot representing
the combination of beef and pork consumed in the year indicated, he con-
structed a short heavy line representing the negative adjusted retail
price ratio of beef to pork for that year.36 He then drew in light
contour lines to represent a smooth three-dimensional surface. The
slope of the contours were drawn at about the same slope as the neigh-
boring observed price ratio lines. These lines conformed to a character-

istic feature of indifference curves for when they were extended beyond

 

36Price ratios were adjusted for the effects of changes in consumer
incomes. 'Waugh in his gpurnal of Farm.Economics article showed contours
for both adjusted and unadjusted price ratios and found the two to be
essentially the same. He attributed this to the fact that the elasticity
of the price ratio with respect to income was small and not significant.
In this case he found both price ratios to be a good approximation to a
partial indifference surface, but stated that in the general case price
ratios need to be adjusted for income effects.

36

the length of the price ratio lines they did not cross. Waugh believed
these contours to be a partial indifference surface for beef and pork
alone, indicating the amount of satisfaction obtained from the two com-
modities. This analysis implies that for each level of money income
there is a definite indifference surface for beef and pork.

Although the nature of the studies presented in this chapter vary
widely, these studies have been used in formulating the analysis for
this study especially in the techniques used and in the selection of the

variables for the statistical models.

Chapter IV
METHOD 0F.ANALYSIS

The purpose of this chapter is to describe and justify the methods
of analysis used in this study. Discussion will center on the choice
of variables, time periods and statistical functions included in the
analysis.

The method of analysis used in this study is similar to that used
by Clawson, Staehle and Shepherd et al. This analysis is most closely
related to the analysis by Shepherd g£_gl.; the main difference being
the introduction of a third commodity, broilers, in the analysis. The
methods used in these analyses were preyiously discussed in Chapter III
of this thesis. As a general summary of the graphic techniques used in
these analyses, retail price ratios and per capita consumption ratios
of the commodities studied were plotted for time series data with price
ratios on the vertical axis and per capita consumption ratios on the
horizontal axis. Through the plotted observations lines of average
relationship were drawn to show the relationship between the price and
consumption ratios and the shifts occurring in the relationship during
the period. Similar diagrams have been used in this study to examine
the relationships, and these have been supplemented by a statistical
analysis of the relationships in recent years.

‘Choice of Variables

If an analysis is designed primarily for use in forecasting and it
is assumed that no change in structure has taken place over the period
included in the analysis or for which the forecast is made, a single

equation using as the dependent variable the item for which a forecast

 

38

is desired will be apprepriate. If the analysis is designed mainly
to estimate the elasticity of demand or other structural coefficients
and a single-equation approach is applicable, the independent variables
should be those that can be considered as predetermined as by weather,
by economic forces in an earlier period or by broad economic forces.1

In economic theory demand is usually spoken of in terms of the
quantity demanded as a function of prices, but for purposes of prediction
most demand studies are now concerned with predicting prices and hence
consider the quantities consumed to be predetermined. The author of this
study is also more interested in predicting price, and hence price is
considered to be the dependent variable in the economic model to be
deve10ped.

As a basis for determining the variables to be included in
the statistical analysis, an economic model for the demand for broilers
can be deve10ped as:

P =3 f (X1, X2, X3, X4, X5, X6, X7. X8, X9, X0)

where
P - the price of broilers
X1 = consumption of broilers
x2 3 consumption of farm chickens
X3 = consumption of beef
X4 = consumption of pork
X5 = consumption of turkey
X6 3 consumption of lamb and mutton
X7 = consumption of fish
X8 = disposable income
x9 = papulation

changes in consumption habits or tastes and preferences
of consumers

 

lRichard J. Foote and Karl A. Fox, Analytical Tools for Measuring.
Demand, Agriculture Handbook No. 64, Agriculture marketing Service, United
States Department of Agriculture, January 1954, p. 8.

39

A flow diagram.of the demand and supply structure for chickens and
broilers constructed by Fox2 supports this economic model. The diagram
indicates that the retail price of chickens is directly affected by the
consumption of farm chickens and broilers, supply of red meats and turkeys
and disposable consumer income.

The retail price of broilers was selected as the price to be included
in the model. This would be expected since this is an attempt to
measure the consumer demand for broilers. This selection is justified

3 for he suggests that equations designed to measure the demand

by Foote
for consumer goods should be based preferably on prices at the retail
level or, if they are not available, then at the wholesale rather than
the local market level.

POpulation can be eliminated as a variable within the model by ex-
pressing consumption of the commodities included on a per capita basis.
By doing this trends in consumption due to the effects of population
changes are eliminated. Inclusion of data in this form appears to re-
present the consumption trends of an "average consumer” within the
population; however, data in this form can be multiplied bythe population
and total market data for the commodity are obtained.

Inclusion of all the other variables mentioned previously would

result in an equation which would be extremely difficult to handle. If

the more important and relevant variables were included first, then it

 

2'Karl A. Fox, The Analysis of Qemand for Farm Products, Technical
Bulletin No. 1081, United States Deparmment of Agriculture September
1953, p. 49.

3Richard J. Foote, Analytical Tools for Studyingjggmsnd and Price
Structures, Agriculture Handbook No. 146, Agricultural‘Marketing Service,
United States Department of Agriculture, August 1958, p. 23.

40
would be expected that each additional variable added in a statistical
model would add little to the coefficient of multiple determination.
Because of the short time period chosen for the statistical analysis,
the number of independent variables included should be held to a minimum
so that there will be sufficient degrees of freedom to test whether the
regression coefficients for the included variables differ significantly
from zero. Through the use of economic theory and knowledge of the
specific commodities the more important and revelant variables can be
selected for possible inclusion in the statistical analysis. Likewise
the less important can be eliminated.

Since this study is an attempt to examine the demand for broilers,
no questions should arise from the inclusion of the per capita consumption
of broilers as a variable in the statistical model. Beef and pork are
believed to be the closest substitutes for broilers and have been included
in the analysis. Previous studies support this belief.4

Fish and lamb are perhaps minor competitors of broilers at the retail
outlets. Since per capita consumption of these commodities was very small
compared to beef, pork and broilers, fish and lamb were deleted in the
statistical analysis. Turkey consumption was also small relative to the
‘meats included in the statistical model. Since consumption of turkeys
has been highly seasonal in the past, the author believes that the annual
estimates do not represent the real influence of turkey consumption on
the demand for broilers; hence, it has been deleted from the analysis.

Unfortunately, the author knows of no statistical studies from.time series

 

48ee the discussion in Chapter III concerning the studies by Riley,
Parsons and Baker and Farris.

41

data that justify the deletion of these variables or reveal the influence
of these variables on the demand for broilers. However, BransonS found
that among the Houston, Texas families interviewed only 2 percent of the
families listed fish as their favorite. Turkey and lamb were each listed
as the preference of 1 percent of the families interviewed. It is harder
to justify deletion of the consumption of farm chickens or other chicken
meats excluding broilers from the analysis. However, since the thirties
broiler consumption has accounted for an increasing percentage of total
chicken meat consumption. In 1950 broilers accounted for about one-

half of total chicken meat consumption and have since continued to
increase steadily. The downward trend in consumption of farm chickens
leads the author to believe that consumption of farm chickens now has
little effect on the demand for broilers, and for this reason it has

been omitted from the statistical analysis.

Disposable income is generally considered to be important in approxi-
‘mating demand functions. Fox6 states that the disposable income of
domestic consumers has proved to be the best over-all indicator of their
demand for agricultural products. Per capita disposable personal income
was chosen as a variable in this study to represent the general level
of consumer demand.

Consumption habits or tastes and preferences of the consumers affect

their demand for broilers. No really adequate quantitative variable is

 

5Robert E. Branson, The Consumer Market for Beef, Bulletin 856,
Texas Agricultural Experiment Station, College Station, Texas, April
1957, p. 7.

6Karl Fox, "Factors Affecting Farm Income, Farm Prices and Food
Consumption," Agricultural_§conomic Research, Volume 3, No. 3, July
1951, p. 67.

42

available to measure the relationship; however, this relationship is
often indicated in demand studies by the introduction of a time variable.
Schultz7 indicated that the time variable is "a catch-all for the dis-
turbing factors." The inclusion of time in the analysis assumes a

linear change in these variables, and it is the change that is measured
not the level of preference. Time was included in the original equations
for this study as a measure of the influence of the variables not in-
cluded in the analysis. A high intercorrelation (r = .98) existed
between income and time. Time was then deleted from the equations, and
the equations were refitted with income as the demand shifter.

The above discussion has mentioned only the absolute values of the
variables included in the analysis. It should be mentioned in this
section that price and consumption variables were included as ratios in
parts of the analysis. The general form of the equations using relative
variables included the price ratio of two commodities as the dependent
variable with the pre-determined variables including the consumption
ratio of the same two commodities, actual consumption of the third com-
modity and income. The general form of these equations can be expressed
mathematically as:

Pa
Pb

 

Q3
3 f a , Y)
( Qb Qc

where P is price, Q is per capita consumption for the commodities a,

b and c and y is per capita disposable income.

 

7Henry Schultz, The Theory and.ypasurement of Demand, University
of Chicago Press, Chicago, Illinois, 1938, p. 184.

43

Choice of Time Periods

The selection of the time period for the statistical analysis was
made on the basis of the relationships shown in the statistical demand
curves for the commodities; statistical demand curves indicate whether
a change occurred in the price-quantity relationships over the range
of the data.

The problem of identification of statistical demand curves was
pointed out by Working.8 He stated that:

"If we construct a statistical demand curve from data

of quantities sold and corresponding prices, our original

data consist, in effect, of observations of points at which

the demand and supply curves have met. Altho we may wish to

reduce our data to static conditions, we must remember that

they originate in the market itself. The market is dynamic

and our data extended over a period of time; consequently

our data are of changing conditions and must be considered

as the result of shifting demand and supply schedules."9
Working suggested that if the shifts occurring in the supply curve were
greater than the shifts in the demand curve, then the points representing
the intersections of the two curves trace out what appears to approximate
a statistical demand curve. He further suggested that a curve could be
fitted to these points, and from it we can detenmine the elasticity of
demand.

Both actual and relative broiler data used in this analysis were
first graphed in price-quantity diagrams for the period of available data.

_Hhen this was done, the time sereis data were examined for possible

statistical demand curves. In the earlier years of the data wide

 

8E. J. UOrking, "What do Statistical'Demand Curves' Show?", Quarterly
Journal of Economics, Volume XLI, No. 2, February 1927, pp. 212-235.

9;b1d., p. 218.

fhactuation existed between the annual observations for both actual
broiler data and for beef and pork relative to broilers, indicating
large shifts in both the demand and supply curves. About 1950 the points
representing the annual observation began to show less variation between
the years and to form a pattern resembling a statistical demand curve.
For this reason only the data for 1950 to 1958 were included in the
statistical analysis. .
Choice of FUnctions
For all models included in this analysis, single equation least
squares regressions were used to estimate the structural relationships.
A choice existed between natural numbers and logarithms as the
form for including the variables in the equations. In general, logarithmic
equations should be used when the relationships between the variables
are believed to be multiplicative or the relations are believed to be
more stable in percentage than in absolute terms or both. If the re-
lationship between the variables is believed to be additive, arithmetic
variables should be used.10 A function which is linear in logarithms
has constant elasticity while a function which is linear in natural
numbers has a constantly changing elasticity. Although logarithmic
equations have the mechanical advantage of yielding curves having a
constant elasticity, this is not a valid criterion for deciding on
their use; in fact this may be a rather unrealistic assumption, eSPecially

at the extremes.11 Although the advantages and disadvantages of each

function can be

 

10Foote and Fox, 02. cit., pp. 9-10.

11Foote, Op. cit., p. 37.

45

pointed out, it is difficult to choose by a priori reasoning the
appropriate'mathematical function to use. Perhaps a valid a priori
grounds exist for fitting the relative demand curves in logarithms for
it seems unrealistic to assume a constant marginal rate of substitution
over any range of reasonable length. However, all models were fitted
using both natural numbers and logarithms.

The regression coefficients obtained were first checked to see if
they were consistent with expectations. Expectations were based upon
economic theory and results of previous studies. The t test was then
used to test the null hypothesis that the regression coefficients did
not differ significantly from zero against the alternative hypothesis
that they did differ significantly from zero. Since the test of signific-
ance was designed to test whether the regression coefficient was zero,
the calculated t value was the ratio of the regression coefficient to
its standard error. The calculated values were checked against the t
values obtained from the "Student's" distribution. If economic theory
lead to the expectation of either a positive or negative sign for the
regression coefficient, a one-tailed test was used to test the null
hypothesis. If no expectations were made regarding the sign of the
coefficient, then a two-tailed test was used. Tests of significance
were made at the .01, .05 and .10 levels.

Previously it was mentioned that functions linear in logs have a
constant elasticity; hence, the elasticities are the regression co-
efficients and can be read directly from the regression equations. Since
the elasticities change along functions which are linear in natural

numbers, the points along the functions where the measures are taken

46
should be specified; the elasticity are usually specified at the mean
values. The regression coefficients of functions including relative
data have been referred to as the "elasticity of substitution," but there
appears to be some questions concerning the restrictive conditions and
the validity of these estimates.12 Therefore the "elasticity of sub-
stitution" will hereafter be used in quotes.

Originally, it was planned that the unexplained residuals in the
equations would be tested for serial correlation by the Durbinéwatson
test, but the available tables did not extend down below fifteen obser-
vations; below fifteen observations it is almost impossible to get
results other than inconclusive from the Durbin-Watson test. As a
result serial correlation among the residuals was examined by inspection
only.

M

All data included in the analysis of this study were obtained from
series published by the United States Department of Agriculture. Since
no comprehensive publication of poultry statistics was available at the
time when the data were sought, data were obtained from.numerous
publications including Thg_goultry_and;§gg Situation; Chickens and Eggs -
Egrmggroduction,ggisposition, Cash Receipts, Gross ggcome by States;
‘Liyestock anngeatfigtatistics; Consumption of Food in the United States,
.1909-52; and, Agricultural Outlook Charts, 1959. Data used in the

analysis are included in Appendix.A; more precise references as to the

 

12See: Irving‘Horrisett, "Some Recent Uses of Elasticity of Substi-
tution-eA Survey", Econometrica,‘Volume 21, No. 1, January 1953, pp. 41-
62 and Kenneth W. Meinken, Anthony S. Rojko and Gordon A. King, "Measurement
of Substitution in Demand from.Time Series Data--A Synthesis of Three
Approaches," ggggnal of Farm Economics, Volume 38, No. 3 August 1956,
pp. 711-735.

47

source of the data are also given there. Egg and Poultgy Statistics
Through 1957, Statistical Bulletin No. 249, was published during the
analysis phase of this study; it contains most of the statistics used
in the analysis.

The retail price series for broilers included in the analysis is
questionable. This series was based on prices collected by the Bureau
of Labor Statistics which included prices for the first three days of

the week.13

Since these prices are based on the first three days of the
week, studies previously mentioned in Chapter III indicate that these
prices are based on only a small portion of the volume of broilers sold
at retail outlets since the largest sales of broilers occur during the
last days of the week. Therefore these prices do not measure the
effects on sales as a result of broilers being featured as weekend

specials in the retail outlets and probably do not represent a true

estimate of the retail prices of broilers.

 

13 For a more complete discussion see: Leo R. Gray and Uilliam L.
Mitchell, garketingipreads for Eggs and Frying Chickens in the U. :54
and Selected Cities, ANS-296, Agricutural'Marketing Service, United States
Department of Agriculture, January 1959 and "Weekend Prices of Food"
Retail Food Prices by Citig§_(mimeograph), Bureau of Labor Statistics,
United States Department of Labor, September 1956 and October 1956.

Chapter V
RESULTS OF ANALYSIS

The purpose of this chapter is to present and explain the results
of the analysis obtained from the methods and procedures discussed in
the previous chapters of this thesis. This chapter is organized into
two main sections; the first section contains a discussion of the results
of the graphic analysis while the second contains the results of the
regression analysis. Each of these sections is divided into subsections
for the results obtained for the demand for broilers, demand for beef
relative to broilers, demand for pork relative to broilers and demand
for beef relative to pork.

Graphic Analysis

The purpose of the graphic analysis is to examine the relationship
between the price and quantity data, to determine whether any shifts
occurred in the relationship during the period and to determine the
time period and appropriate mathematical function to use in the re-
gression analysis.

In Chapter I it was hypothesized that the demand for broilers has
been increasing relative to the demand for beef and pork. If the demand
for broilers has been increasing relative to the demand for beef and
pork, the increase in the relative demand for broilers could have re-
sulted from: (1) the demand for broilers increasing while the demand
for beef and pork was decreasing, (2) the demand for broilers increasing
while the demand for beef and pork was constant, (3) the demand for
broilers, beef and pork increasing with the demand for broilers increasing
faster, (4) the demand for broilers being constant while the demand for

beef and pork was decreasing and (5) the demand for broilers, beef and

49

pork decreasing with the demand for beef and pork decreasing faster.
Similar reasons could be stated for the case of the relative demand

of beef to pork, but they would be duplicates of the above and will not
be stated again. The following sections have been deve10ped to examine
the average relationships between price and quantity data and to pro-
vide evidence for later conclusions concerning the nature of and the
reasons for changes in the demand for broilers relative to the demand
for pork and beef.

ngand forggroilers. The consumer demand for broilers at retail

 

appears to have fluctuated ‘more than the supply of broilers in the
1940's, but both were increasing. This is indicated by the data shown
in Figure 81 for the retail price and per capita consumption of broilers.
A positive relationship existed between the retail price and per capita
consumption of broilers during this period. If a line were drawn
through the observations to represent the relationship between the
retail price and per capita consumption of broilers during this period.
If a line were drawn through the observations to represent the relation-
ship between the retail price and per capita consumption, it would more
nearly approximate a supply curve than a demand curve for the line

would slepe upward and to the right indicating that demand had been
changing greater than the supply during the period. Further inspection
of the diagram reveals that a change occurred about 1950 in the relation-
ship between the retail price and per capita consumption. A line fitted

by inSpection to the observations for the period 1950 to 1958 slapes

 

The years when price celings and rationing were in effect due to
World War II have been indicated in Figure 8 by circles around the
observations for those years.

 

 

. :l..tfsulf1.1 .m‘l'i’ Fir

50

NN

 

33 0u oqu .333 woods: on”. a.“ 8530.5 now panama .w shaman
apnoea Ca mumaaowm mo cowumESmcoo muaamo pom
ON ma 0H «H NH OH w o a N o-
a _ q fl 1 + 1 q d q N
3 a
. m
T as U
1 .3
Na 9
U
- -3 m
See a
«a 0
min 0 I.
m
Ifie "W
I
8
1
3
came I.
u
Jon no
“a. m
n
ma
. a
1 1
mm .a
mg m
u
p
loo

 

 

 

51

down and to the right, which indicates that supply was changing greater
than the demand; and appears to more nearly represent a demand curve
for broilers for the period. The downward slope indicates that the
relationship between the retail price and per capita consumption was
negative. Annual observations during the period tended to move con-
sistently down a demand curve. There are no indications of any shifts
occurring during this period in the relationship between the retail
price and per capita consumption of broilers.

Demand for4§eef Relative to groilers. The demand for beef relative

 

to the demand for broilers, as shown in Figure 92, has been declining.
This is indicated by changes in the relationship between the ratio of
the retail price of beef to broilers and the ratio of the per capita
consumption of beef to broilers. Inspection of Figure 9 reveals that
large year to year changes occurred in the ratio of the per consumption
of beef to broilers in the 1940's while relatively small changes occurred
in the ratio of the retail price of beef to broilers except in 1947
when price ceilings were removed. The ratio of the per capita consumption
of beef to broilers was decreasing which indicates that consumption of
broilers was increasing relative to beef. Since consumers were increasing
their consumption of broilers relative to beef at fairly constant relative
prices, this indicates that the demand for broilers was increasing relative
to the demand for beef. About 1950 the large shifts between annual obser-
vations eeaSed to appear, and the relationship between the price and
per capita consumption ratios appeared to remain stable during the period
1950 to 19 8. A line fitted by inspection to

2The years when price ceilings and rationing were in effect due

to World War II have been indicated in Figure 9 by circles around the
observations for those years.

wmoa cu oema .mouwum mamas: ea... a.“ uueaaoum ou 0233mm «mom new panama
mueawoum\moem caumm ceaumaamcoo muaamo new

.m seam:

 

52

 

as.

«N . on ma ea an Na oa m
d _ a a 1 J1 J 4
Q
3 o
e .3
30 e
S
0
3

me

 

a
a

l

 

\O
o
O

I

Q
0

O
r-C

N
H

<2'
r-l

o.H

w.a

szexroaa/;aaq noises 33114 IIQQGH

53

the observations for the period 1950 to 1958 appears to more nearly
represent a demand curve for the relative data; a curved line appears
to fit the relationship better than a straight line. The downward
slepe of the curve represents an inverse relationship between the ratio
of the retail prices of beef to broilers and the ratio per capita con-
sumptioncxfbeef to broilers. No indications are present in the diagram
to suggest that any shifts occurred in the relationship between the
ratios during the period 1950 to 1958.

Demand for Pork Relative tgigroilers. The retail price ratios of
pork to broilers and the per capita consumption ratios of pork to broilers,
which are shown in Figure 10,3 indicate that the demand for pork relative
to the demand for broilers has been declining. Annual observations
for the 1940's indicate that changes occurred in the relationship be-
tween the retail price ratio of pork to broilers and the per capita
consumption ratio of pork to broilers; this is indicated by the large
changes that occurred in the per capita consumption ratio while changes
in the retail price ratio were much smaller. The relationships appears
to have remained relatively stable in the 1950's. A line which was
fitted by inspection to’represent the average relationship between the
price and per capita consumption ratios for the period 1950 to 1958
slopes down and to the right and appears to more nearly represent a
demand curve for the relative data; a curved line appears to be a better

fit for the relationship than a straight line. The downward slope of

 

3The years when price ceilings and rationing were in effect due to
World War II have been indicated in Figure 10 by circles around the
observations for those years.

54

o:

.mmma cu oama amouaum woman: one aw muoaaoum cu o>wuoaom xuom How panama .oH shaman

muoaaoum\xuom seamed segueasmcou muwmoo Ham

on ma

ON

ma

0H m o

 

 

.J

H

41

he.

* l

we

m w

 

 

¢.o

o.o

w.o

o.a

N.H

¢.H

szaIIOJQ/nlod uorasu 93:1; 119393

V.._.

 

 

55

the curve indicates the existence of an inverse relationship between
the ratio of the retail price of pork to broilers and the ratio of the
per capita consumption of pork to broilers.

Demand for geef:§g}ative to Pork. The time series data for the
retail price ratio of beef to pork and the per capita consumption ratio
of beef to pork shown in Figure 114 indicate that the demand for beef
has been increasing relative to the demand for pork.5 Since continuous
series of beef and pork retail prices are available for a longer period
of time than were broiler prices, this diagram has been extended back
to 1925 to show the changes in relative demand over a longer period of
time. In the period from 1925 to 1958, with the war years deleted, two
shifts and possibly three have occurred in the relationship between the
ratio of the retail price of beef to pork and the ratio of the per capita
consumption of beef to perk. Shifts in the relationship occurred be-
tween the years 1932 to 1934 and 1947 to 1950. It is questionable whether
there has been a shift in the relationship since 1950 or whether the
scattering of the observations indicate a single relationship with wide
deviations. It is conceivable that one may consider two relationships
for the period since 1950 with the separation occurring in 1954 to 1955;
this is the situation shown in Figure 11 where the two lines of average
relationship have been indicated by broken lines to indicate that they
. are questionable. However, the statistical analysis for the period 1950 to
1958 was based on the assumption that the relationship during the

4The years when price ceilings and rationing were in effect due to
World War II have been indicated in Figure 11 by circles around the
observations for those years.

5A similar diagram.was published in the study by Shepherd et a1.

Op. cit., p. 733; this diagram was discussed in Chapter III of this
thesis. Since the study by Shepherd g£_al, was published, the price and

consumption data have been revised so there are some differences in data
included in the two diagrams.

56

o.H

 

wmma ou mNmH .woumum woman: «am :a xuom 0» o>wumfimm moon wow panama .HH musmwm

xu0A\moom owuom ceauaasmcou ouwmwo mom
m.H ¢.H m.H N.H H.H o.H m.o w.o 5.0
.1

J A! 1

a 4 d 1 1

 

 

0.0

w.o

N.H

¢.H

o.~

axed/gaag organ 90114 119333

57

period was unchanged. Since the curves indicating the lines of average
relationship have been shifting up and to the right, this gives some
indication that the demand for beef relative to pork has been increasing
over the period. The four lines representing the relationships between
the retail price ratio and per capita consumption ratio slepe down and
to the right. The downward slopes indicate the existence of an inverse
relationship between the price ratios and per capita consumption ratios
of beef to pork.

gymmary of the Graphic Results. Actual and relative time series
data used in the graphic analysis indicate that prior to 1950 the re-
lationships existing between retail price and per capita consumption of
broilers were not clear. Only since 1950 has the relationship between
retail price and per capita consumption for actual and relative data
for broilers resembled anything like a demand curve and can be approximated
by a single equation regression model. This analysis tends to indicate
that the relationship between the price and quantity data in natural
numbers is linear for the actual data and curvilinear for the relative
data.
Eggression Agalysis

To obtain more precise estimates of some of the demand shifters,
income and competing products will be included in the regression analysis.
The estimates obtained from the regression analysis also provide a basis
for making probability statements about the results providing the assump-
tions mentioned below are approximately satisfied.

To facilitate the discussion in this section a few comments on
notation are in order. The dependent variables included in the regression

analysis are:

58

Y1 ' per capita consumption of broilers in pounds
Y2 a retail price of broilers in cents per pounds

relative retail price of beef to broilers

r4
to
II

Y4 = relative retail price of pork to broilers

Y5 = relative retail price of beef to perk

The independent variables included in the analysis are:

X1 3 retail price of broilers in cents per pound

X = retail price of beef in cents per pound

X3 = retail price of pork in cents per pound

X4 - relative per capita consumption of beef to broilers

X5 = relative per capita consumption of pork to broilers

x6 = relative per capita consumption of beef to perk

X7 = per capita consumption of broilers in pounds

X = per capita consumption of beef in pounds

X9 = per capita consumption of pork in pounds

X0 = per capita disposable personal income in dollars
and the u's are the unexplained residuals which are assumed to be
normally and independently distributed with a uniform variance. Obser-
vations on each variable are in terms of annual data for the United States.

It was mentioned in Chapter IV that all models were fitted in both

natural numbers and logarithms. Both of these were examined for close-
ness of fit, and one was selected on the basis of both the best fit to
the data and the graphic analysis in the preceding section. Natural
numbers were selected as the best fit for the demand models using actual
data while logarithms were selected as the best fit for the relative
demand models. Only the equation selected as the best fit for each

model is included in this chapter; the other equations are included in

59
this chapter; the other equations are included in Appendix B. For each
of the models presented in this chapter, the standard errors of the
regression coefficients are included in parentheses below the regression
coefficients. The adjusted coefficient of multiple determinatimn(§2)
and the standard error of estimate (S?) are included with each model.

Demand for Broilers. Although it was indicated in the previous

 

chapter that the author was more interested in predicting prices than
quantities consumed, separate models using actual data have been
develoPed using both price and quantity of broilers consumed as the
dependent variable.

Consumption of Broilers. The per capita consumption of
broilers was considered to be influenced by the retail price of broilers,
retail price of beef, retail price of pork and income. The variables
were fitted in the additive form as:

(1) Y1 = -8.29598 - 0.3361211;1 +-o.091282x2 +0.067933x3 + 0.018851x0
(.06267) (.02465) (.04186) (.00253)

'2

R a .985 SM- .524

y
ISignificant at the .01 level

Significant at the .05 level
asignificant at the .10 level

Signs of the regression coefficients were consistent with expecta-
tions; therefore one-tailed tests were used for the test of significance.
The coefficients of each of the variables included were significantly
different from zero. These variables were associated with 98.5 percent
of the variance in the per capita consumption of broilers.

The model shows an inverse relationship between the per capita con-

sumption of broilers and the retail price of broilers and a positive

relationship between consumption and the retail price of beef, retail

60

price of pork and income. The simple correlation between per capita
consumption and retail price of broilers was -O.92, and, the correlation
between per capita consumption and income was 0.96; correlation between
per capita consumption and the other independent variables, retail price
of beef (r2Y B -.24) and retail price of pork (r3Y = .26), was much
lower. The intercorrelation between the retail price of broilers and
income was -0.85; an intercorrelation of this magnitude would usually
be expected to cause large standard errors of the regression coefficients
,and would provide sufficient evidence to question the validity of the
regression coefficients as well as the elasticities calculated from
these coefficients. Little can be said about the effects, if any, that
the high intercorrelations will have on the estimates obtained from
the model. However, in this case reasonably small standard errors were
obtained for the regression coefficients in despite of thehigh inter-
correlation.

The price elasticity of demand for broilers measured at the mean
values was -l.29. It should be noted again that a demand curve which
is linear in arithmetic form has a constantly changing elasticity of
demand which is highest at the upper end and becomes lower with movement
down the curve. At the upper end of the estimated demand curve for
broilers the price elasticity of demand calculated from.the 1950 data
was -2.26, and at the lower extreme the elasticity of demand from the
1958 data was calculated to be -0.73. The high intercorrelation be-
tween the retail price of broilers and income provides sufficient
evidence to question the validity of these estimates of price elasticities

of demand. This model indicates that the demand for broilers

61

was highly elastic in the early fifties and has since decreased until
now appears to be inelastic. This indicates that one must use elastic-
cities with extreme caution and especially the estimates of elasticities
calculated at the mean values. This especially applies to commodities,
such as broilers, where there has been a secular trend in the price of
the commodity over the period studied which would cause continuous
movement in one direction along a demand curve.

The cross elasticity of demand for broilers at the mean values
was 0.47 with respect to the price of beef and 0.28 with respect to
the price of pork. This indicates that substitution exists between
broilers and the two commodities and that broilers are in competition
with each commodity. The model indicates that a ten cent increase in
the retail price of beef would increase per capita consumption of broilers
by 0.91 pounds while a ten cent increase in the retail price of pork
would increase per capita consumption of broilers by 0.68 pounds.

The regression coefficient for income indicates that each $100
increase in per capita income increases per capita consumption of
broilers by 1.88 pounds; the income elasticity of demand for broilers
at the mean values as calculated from this coefficient was 2.11. However,
this coefficient probably includes more than the influence of income.
Income and time are highly correlated so one might expect this coeffi-
cient to include the influence of changing tastes as well as the in-
fluence of income. The validity of this coefficient is also questionable
as a result of the high intercorrelation existing between income and the

retail price of broilers.

Per Capita Consumption of Broilers

Actual as a Percent of Calculated

110 ‘

105

62

Figure 12

Comparison of Actual and Calculated
Per Capita Consumption of Broilers

 

Actual

- - - Calculated /

 

 

 

100 \/ V

95

 

Eltimltes of the annual per capita consumption of broilers obtained
from the equation are shown in Figure 12 along with the actual per
capita consumption. The estimates indicate that a relatively close fit
was obtained for the data. In all cases the direction of the year to
year movement was correctly indicated. The model appears to give
consistently close estimates for each year with the unexplained residuals
between -0.54 and +0.47 pounds, or in a range of eight percent about
the actual values. The largest percentage deviations occurred in the
earlier years; however, this can be accounted for by the lower level of
consumption in the earlier years. No pattern could be detected in
the residuals, therefore, the assumption of randomness and independence
was not unreasonable.

2gigg_2fi_fi;gilgzg. The price of broilers was considered to
be a function of the per capita consumption of broilers, per capita
consumption of beef, per capita consumption of pork and income. The

model was fitted in the additive form as:

Y2 = 82 + b27X7 + b28x8 + b29X9 + bZOXO + u2

where u2 is the unexplained residual of Y2. The results from the fitted

equation were:

Y2 = 46.22536 - 2.463451x7 - 0.382872x8 - 0.25425x9 + 0.055671x0
(.73387) (.17689) (.28103) (.02616)

I12 = .873 39 = 2.055
1Significant at the 0.5 level
2Significant at the .10 level
The signs of the regression coefficients for the per capita con-

sumption of broilers and income were consistent with expectations, and

the coefficients were significantly different from zero using a one-

64

tailed test. Since no expectations were formulated for the coefficients
for the per capita consumption of beef and pork, two-tailed tests were
used for the test of significance; the coefficient for per capita con-
sumption of beef was significantly different from zero. The variables
included in the model were associated with 87.3 percent of the variance
occurring in the retail price of broilers in the period 1950 to 1958.

This model also indicate an inverse relationship between the price
and per capita consumption of broilers; each one pound increase in the
consumption of broilers results in a decrease in the price of broilers
of 2.46 cents. The elasticity of demand for broilers calculated at
the mean values was -l.56.

The model indicates that each $100 increase in per capita income
increases the retail price of broilers by 5.6 cents per pound; however,
this coefficient also reflects the changes in the tastes of the consumers
over time which could not be separated from income.

The signs of the coefficients for beef and pork consumption were
the same as the sign for broiler consumption; this would indicate that
beef and pork were complementary with respect to broilers which is
contrary to the results of the previous model when per capita consumption
of broilers was considered to be a function of prices. The regression
coefficient for the per capita consumption of beef indicates that a
one pound increase in beef consumption results in a decrease of 0.38
cents in the price of broilers.

Each of the independent variables included in the model was highly
correlated with the retail price of broilers; consumption of broilers

was the highest (r7Y =-.92) and was followed by income (r0Y = -.85),

65

consumption of beef (r8Y = -.79) and consumption of pork (r .68).

9Y—
Per capita disposable income was highly correlated with consumption of
broilers (ro7 - .96) and consumption of beef (r08 = .84) while con-
sumption of beef was also highly correlated with consumption of broilers
(r87 = .75) and consumption of pork (r89 = -.77). These high inter-
correlations are perhaps the cause of the high standard errors of the
regression coefficients; the standard errors in this model are much

larger than those obtained in theprevious model. Since all of the
independent variables are highly intercorrelated, it is doubtful that

the values for the regression coefficients and the elasticities calculated
from them have any real meaning.

Estimates of the retail price of broilers obtained from the model
were fairly close to the actual prices as can be seen in Figure 13. In
all cases except two, 1951 to 1952 and 1956 to 1957, the direction of
the year to year movement was correctly predicted. The unexplained
residuals were between -2.21 and + 1.85 cents per pound; overestimates
of the retail price were less than 4.5 percent while underestimates
were less than 3.5 percent of the actual values. No pattern
could be detected in the residuals; therefore, the assumption

of randomness and independence was not unreasonable.

Comparison of the results from the two models presented above, where
retail price and per capita consumption of broilers were each considered
to be the dependent variable, reveals that both equations give relatively
good fits to the data but give quite different coefficients and elasticities.

The results of the two models indicate that the model using price as a

Retail Price of Broilers

Actual As a Percent of Calculated tr

66
Figure 13

Comparison of Actual and Calculated
Retail Prices of Broilers

59 ’

57 F

    
 

Actual
- - - Calculated

 

51 t

49 t

47 b \

 

1950 51 52 53 54 55 56 57 58
Time

110 P

105 r

f’\ _ /

95

90 r

i _ L I L I L A

1950 51 52 53 54 55 56 57 58

 

Time

67

dependent variable involves higher intercorrelations among the indepen-
dent variables. For this reason the model using quantity as the de-
pendent variable probably is more meaningful for estimating elasticities.
This is true even though one might consider the quantity available for
consumption to be predetermined while price varies so as to clear the
market.6 The high intercorrelations do not appear to affect the pre-
dictive power of the models within the period studied but may affect
forecasts for future periods if the highly intercorrelated independent
variables become less intercorrelated.

Demand for Beef Relative to Broilers. The relative retail price of
beef to broilers was considered to be influenced by the relative per
capita consumption of beef to broilers, per capita consumption of pork

and income. The variables were fitted in the multiplicative form as:

_ b34 b39 b30

where u3 is the unexplained residual of Y3. The results obtained from
the model fitted in this form for the period 1950 to 1958 were:

1
(3) log Y3 = 5.03131 - 0.91909 log X4 - 0.17680 log X9 - 1.219291 log X

(.10419) (.22131) (.27486) 0

13.2 = .920 s;= .015

lSignificant at the .01 level

 

6This is a relative and not an absolute argument. Since stocks of

broilers are normally small, the quantities produced in any year will
reflect prices in the previous years to the extent that fixed costs were
incurred. However, it must be recognized that broiler production, un-
like pork and beef production, can be varied to a considerable degree
within a year.

68

The sign of the regression coefficient for the relative per capita
consumption of beef to broilers was as expected, and the coefficient
was significantly different from zero with a one-tailed test. No
expectations were formulated concerning the coefficient for income;
it was significantly different from zero using a two-tailed test. The
per capita consumption of pork was not significantly different from zero.
These variables were associated with 92 percent of the variance occurring
in the relative retail price of beef to broilers during the period 1950
to 1958.

The regression coefficient for the relative per capita consumption
of beef to broilers indicates an ”elasticity of substitution” between
the two commodities of -l.09 which is constant over the period. .The
income coefficient indicates that a one percent increase in income
results in a 1.22 percent decrease in the relative retail price of beef
to broilers; however, this coefficient also reflects changes in the
tastes of consumers which could not be separatedofrom.income.

A simple correlation of -0.84 existed between the dependent variable,
relative retail price of beef to broilers, and the relative per capita
consumption of beef to broilers; simple correlations between the depen-
dent variable and income (rOY =-' .39) and per capita consumption of pork
(r9Y =';ll) were considerably lower. Income was highly correlated with
the relative per capita consumption of beef to broilers (r40 = -.80)
and moderaterately correlated with per capita consumption of pork
(r90 = -.63). These intercorrelations are not as high as those in the
previous models but are perhaps high enough to reduce the reliability

of the regression coefficients.

Retail Price of Beef/Broiler

Actual as a percent of Calculated

09

Figure 14

Comparison of Actual and Calculated Relative

1.3r

1.2.

1.1

 

10

95

 

Retail Prices of Beef to Broilers

 

- - - Calculated

7O

Estimates obtained from the model for the relative retail price of
beef to broilers as shown in Figure 14, indicate that the function is
a relatively glad fit for the data. Year to year changes in the estimates
were consistent with changes occurring in the actual price ratio;
however, this is not very surprising in this case for changes between
the annual observations were rather pronounced. The model underestimated
the price ratio in the first year, overestimated the ratio for the next
four years and then underestimated the ratio in the last four years;
however, the number of observations is insufficient to support a state-
'ment that serial correlation exists among the residuals. The unexplained
residuals were between -0.04 and +0.06 or in a range of less than nine
percent of the actual values of the observations.

Demand for Pork Relative to Broilers. The relative retail price

 

of pork to broilers was considered to be influenced by the relative per
capita consumption of pork to broilers, per capita consumption of beef

and income. These variables were fitted in the multiplicative form as:

Y4 = aaxsbASXBbABXObQOu4
where u4 is the uneXplained residual of Y4. The results obtained from
this model for the period 1950 to 1958 were:
(4) log 14 3 7.87131 - 0.938111 log X5 - 0.10982 log X8 - 2.183571 log X0
(.12280) (.13194) (.47762)
112 = .945 s;;= .014

ISignificant at the .01 level.
The regression coefficient for the relative per capita consumption
of pork to broilers was negative as expected and was significantly dif-

ferent from zero using a one-tailed test, and the regression coefficient

71

for income was significantly different from zero using a two-tailed

test. The coefficient for the per capita consumption of beef was not
significantly different from zero. These three variables were associated
with 94.5 percent of the variance occurring in the relative retail price
of pork to broilers during the period 1950 to 1958.

The regression coefficient for the relative per capita consumption
of pork to broilers indicates the "elasticity of substitution” between
the two commodities to be -l.07. The coefficient for income indicates
that a one percent increase in income results in a 2.18 percent decrease
in the relative retail price of beef to pork; it should be mentioned
again that this coefficient also represents changes in the tastes of
consumers over time which could not be separated from income.

Each of the independent variables was substantially correlated with
the relative price of pork to broilers; relative per capita consumption
of pork to broilers (r6Y = -.89) was negatively correlated and per capita
consumption of beef (r8Y = .65) and income‘rOY = .75) were positively
correlated. Income was highly correlated with relative per capita con-
sumption of pork to broilers (r50 = -.96) and per capita consumption of
beef (r80 = .83) while the per capita consumption of beef was correlated
with the relative consumption of pork to broilers (r58 = -.82). In spite
of the high intercorrelations the standard errors of the regression co-
efficients do not appear as high as one might expect; however, the high
intercorrelations raise questions about the validity of the regression
coefficients.

Estimates of the retail price ratio of pork to broilers, as shown

in Figure 15, indicate that the predicted year to year changes between

 

Retail Price of Pork/Broilers

Actual as a percent of Calculated

1.4 -
1.3 *
1.2 '
1.1 .
1.0 r
0.9 -

 

Figure 15

Comparison of Actual and Calculated Relative
Retail Prices of Pork to Broilers

Actual
- - - Calculated /

 

 

72

 

110 -

105 t

100

_L L L J— #L_ I l L I
1950 51 52 53 54 55 56 57 58
Time

 

 

 

/\\//\/ x

L L L I l l

1950 51 52 53 54 55 56 57 58

73

the first three years were in the wrong direction. This should not be
considered as a serious fault of the model for the changes occurring in
the actual observations were small relative to changes in the ratio in
other years and the estimates were relatively close to the actual values.*
In all other cases the direction of the year to year movement was cor-
rectly predicted. The unexplained residuals were between -0.036 to +0.03?
or in a range of seven percent about their actual value. No pattern
could be detected in the residuals; therefore, the assumption of ran-

‘domness and independence was not unreasonable.

 

Demand for Beef Relative to Pork. The relative retail price of

 

beef to pork was considered to be influenced by the relative per capita
consumption of beef to pork, per capita consumption of broilers and

income. The variables were fitted in the multiplicative form as:
b56 b57 bSO
where u5 is the unexplained residual of Ys. The results obtained from

the model for the period 1950 to 1958 were:
1

(5) log Y5 = -0.22350 - 0.79777 log x6 + 0.27601 log X7 + 0.01559 log X0
(.13081) (.27055) (.9330)
£2 = ,855 s4 = .021

y
lSignificant at the .01 level

The sign of the regression coefficient for the relative per capita
consumption of beef to pork was negative which was consistent with ex-
pectations, and the coefficient was significantly different from zero
with a one-tailed test. The other two variables were not significantly
different from zero. The independent variables included in the model
were associated with 85.5 percent of the variance occurring in the

relative retail price of beef to pork.

The regression coefficient for the relative per capita consumption
of beef to pork indicates the "elasticity of substitution" between the
two commodities to be -1.25.

Each of the independent variables included in the regression
equation was negatively correlated with the relative retail price of
beef to pork; relative per capita consumption of beef to pork (r6Y = -.86)
being the most highly correlated was followed by income (rOY = -.46) and
per capita consumption of broilers (r7Y = -.41). Income was highly
correlated with per capita consumption of broilers (r70 = .98) and
relative per capita consumption of beef to pork (r60 = .80) while relative
per capita consumption of beef to pork and per capita consumption of
broilers were highly correlated (r67 = .78). The high intercorrelations
perhaps point to the source causing the large standard error of the
regression coefficient for income and cast doubt on the validity of the
other coefficients.

Estimates of the relative retail price of beef to pork as shown in
Figure 16, indicate that the model is perhaps the least effective estimating
model included in the analysis; this does not mean to imply that the
model is a poor estimator, but dust it is not as effective as the other
equations. The direction of the predicted year to year changes was con-
sistent with the actual changes except in one case, 1952; the difference
in the direction of the change may be related to the large overestimate
for 1951. The residuals were between -0.068 to + 0.065 or in a range
of ten percent about the actual values. The residuals do not appear to
be inconsistent with the assumption of randomness, but they do cast

doubt on the validity of the assumption that the relationship between

 

Retail Price of Beef/Pork

Actual as a percent of Calculated

1.3e

H
N
l

H
a
p—I
T

H
D
r

75

Figure 16

Comparison of Actual and Calculated Relative
Retail Prices of Beef to Pork

Actual
‘- Calculated

 

 

 

m
r
l
T
l
r

110-

lose

10

95

9G

 

1950 51 52 53 54 55 56 57 58
Time

1950 5l 52 53 54 55 56 57 58
Time

 

76

the variables remained the same over the period 1950 to 1958; the
residuals tend to support the relationships between the retail price
ratios and the per capita consumption rations that were indicated by
dashed lines in Figure 11.

Summary of Rggression Results. Although relatively good fits were
obtained from each model, the high intercorrelations between the in-
dependent variables caused questions to arise concerning the validity
of the regression coefficients. This restricts any meaningful statements

that could be made concerning the elasticities obtained from the co-

 

efficients; these high intercorrelations give reasons to emphasize the
graphic as well as the regression results.
In the following chapter these results will be further deve10ped

and conclusions drawn as to their usefulness in predicting the future.

Chapter VI
SOME CONCLUDING REMARKS
The hypothesis stated in Chapter I was that the demand for broilers
has been increasing relative to the demand for beef and pork. If this
hypothesis can be accepted as being true, the increase in the relative
demand for broilers could have resulted from changes in.the demand for
either or both of the commodities. Five possibilities were stated in
Chapter V as reasons for the increase in the relative demand for broilers;

they were: (1) the demand for broilers increased while the demand for

 

beef and pork was decreasing, (2) the demand for broilers increased
while the demand for beef and pork was constant, (3) the demand for
broilers, beef and pork increased with the demand for broilers increasing
faster, (4) the demand for broilers was constant while the demand for
beef and pork was decreasing and (5) the demand for broilers, beef and
pork decreased with the demand for beef and pork decreasing faster.

Results of the analysis were presented in the previous chapter;
therefore, this chapter is more of an attempt to restate the more im-
portant results and to state the conclusions that can be drawn from the
evidence indicated by the analysis.
Indications of the Analysis

Indications of shifts in demand that were observed in Chapter V
and are presented in this section will be discussed with reference to
two periods, the period prior to 1950 and the period 1950 to 1958.

Results of the analysis indicate that the demand for broilers rel-
ative to beefami pork did increase during the period 1940 to 1950. This

is indicated by the relative data shown in Figures 9 and 10 where the

78

relationship between the retail price ratio and consumption ratio changed
over the period. These diagrams were expressed in terms of beef and pork
relative to broilers; since the changes occurring in the relationship
between the retail price ratio and consumption ratio during the period
were shifts to the left, this indicated that the demand for broilers
increased relative to beef and pork. Further comparison of these two
diagrams reveals that shifts in the relationships between the retail
price and consumption ratios of beef to broilers and pork to broilers
appear to have disappeared at about the same time. Since shifts in the
average relationship between the retail price and per capita consumption
of broilers (Figure 8) also ceased to appear at about tne same time,

this may be an indication that the demand for beef and pork was relatively
stable during the period, at least when compared with the demand for
broilers.

The analysis indicated that the demand for beef relative to pork
increased in the period 1925 to 1958. This was indicated by the shifts
which occurred in the relationship between the retail price and con-
sumption ratios. Shifts in the relationship were not as continuous as
the shifts occurring in the relative demand of beef and pork to broilers.
This was indicated in Figure 11 where four lines of average relationship
were drawn to show the relationship between the price and consumption
ratios; since the lines of relationship more nearly resembled demand
curves and tended to move up and to the right over time, this is an
indication that the demand for beef relative to pork increased over the
period studied. Since urban consumers ate more beef than farm peOple

and the income elasticity of demand for beef was higher than the income

79

elasticity of demand for pork, Shepherd‘g§__a__l_.1 attributed the increase
in demand for beef relative to pork.mainly to increased urbanization
and consumer incomes.

Since the demand for beef relative to the demand for pork increased
during the period studied, this indicated that the demand for pork rel-
ative to broilers declined more during the period than the demand for
beef relative to broilers.

Since the demand for broilers increased, the last two possibilities

 

stated as reasons for the increase in the relative demand for broilers
can be rejected. This analysis does not contain sufficient evidence
to determine which of the remaining three possible reasons is responsible
for the changes in the relative demand, but the increase in the demand
for beef relative to pork does indicate that the demand for either
beef or pork was changing. This analysis has examined only the relative
demand for beef and pork and was not concerned with the absolute demand
for either commodity; however, diagrams plotted from the actual data for
!
beef and pork, which was the same data used to construct the relative
data, indicate that the demand for beef was increasing over the period
while the demand for pork was decreasing. From this it can be concluded
that the increase in the relative demand for broilers resulted from an
increase in the demand for broilers and an increase in the demand for

beef which was less than the increase in the demand for broilers (Reason

3) and a decrease in the demand for pork (Reason 1).

 

lShepherd et al., op. cit., pp. 739-742.

80

The demand for broilers appeared to remain stable over the period
1950 to 1958. Both the graphic and regression analyses indicated that
the relationship between the retail price and per capita consumption
of broilers remained stable during the period. The graphic analysis
indicated that the annual observations tended to move consistently
down a single line of average relationship, and the close fits for the
data obtained from the regression models further indicated that no
changes occurred in the relationship during the period.

The graphic analysis indicated that the demand for beef relative

 

to broilers appeared to remain relatively stable during the period 1950
to 1958, but the estimates obtained from the regression model give reason
to question whether the relationship between the retail price and con-
sumption ratios remained stable over the period. Residuals derived from
the estimates indicated that the model underestimated the price ratio
in the first year, overestimated the ratio in the next four years and
then underestimated the ratio in the last four years. Since the relative
quantity in the last four years was taken at a higher relative price than
was predicted by the model, this could be an indication that the demand
for beef increased relative to the demand for broilers during the latter
years of the period. The graphic analysis did indicate that there was
a tendency for the price ratio to increase in the latter years of the
period, but this appeared to be an upward movement along the line of
relationship rather than a change in the relationship.

The demand for pork relative to the demand for broilers appeared
to remain stable since 1950. The graphic and regression analyses gave

no indications that suggested any changes in the relationship between the

retail price and consumption ratios of pork to broilers during the period.

81

The demand for beef relative to the demand for pork increased
during the period 1950 to 1958. This was first noted in the graphic
analysis in Chapter V where there appeared to be a shift in the relation-
ship between the retail price and consumption ratios of beef to pork.
The possibility of a shift occurring in the relationship was weakly
supported by the residuals of the estimates obtained from the regression
analysis.

The analysis indicated that broilers now compete with beef and pork
on an aggregate level. This was shown in the actual demand equation
for broilers where quantity consumed was considered to be a function
of the various prices; in this model an increase in the prices of beef
and pork resulted in an increase in the quantity of broilers consumed.
This was also indicated by the retail price and consumption ratios of
beef and pork to broilers where a change in the price ratio resulted in
a change in the opposite direction in the consumption ratio.

The analysis indicated that prior to 1950 the demand for broilers
increased relative to the demand for beef and pork while the demand for
beef increased relative to the demand for pork. Since 1950 the demand
for broilers and the demand for broilers relative to the demand for pork
appeared to remain stable while the demand for beef relative to the
demand for broilers and pork increased.

Some Reasons for the Changes in Demand

Shifts occurring in the actual and relative demand for broilers
have resulted from changes in tastes and incomes of the consumers.

Consumers have acquired an increased taste or preference for
broilers over time for per capita consumption of broilers increased from

2.0 pounds in 1940 to 21.7 pounds in 1958. Unless consumers had increased

 

82

their preference for broilers, it is extremely doubtful that per capita
consumption of broilers would have increased by this amount as a result
of changes in the price of broilers. Availability and improved quality
of broilers at the retail outlets have probably been the two most
important factors leading to the increased preference for broilers.
Increased production of broilers and the related vertical integration
occurring in the industry have made available a larger and more continuous
supply of broilers. In the early years of the broiler industry the

housewife was never sure when she would find broilers or poultry meat

 

available in the grocery store, This may have eliminated broilers or
poultry meat from the diets of families where the housewife planned the
family menu in advance of her grocery shOpping. The housewife can now
be sure that a supply of broilers is available at her local grocery
store, and she can plan in advance to include broilers in the family
menu.

Improved and standarized quality of broilers available at the
retail outlets has permitted consumers to place a higher degree of
reliability on the broilers which they view in the retail outlet and
may have encouraged greater consumption of broilers by consumers. Im-
provedquality of broilers at retail has resulted from younger and more
tender meated birds, which are raised in confinment, and more careful
handling of the birds during processing combined with faster trans-
portation between the processing plants and the retail outlets.

The form of the product has changed over time as well as the
quality of broilers. In the earlier years broilers were marketed as

either New York dressed or ready-to-cook, but over the years processing

83

shifted to ready—to-cook broilers; this shift is indicated by the retail
price series for broilers published by the Bureau of Labor Statistics
and the United States Department of Agriculture. Retail prices for
broilers are now reported only on a ready-to-cook basis whereas they
were reported on New York dressed basis until 1953.

Retail outlets recognized the potential market for broilers and
Spent more effort to make the product attractive and eye-appealing to

potential consumers. Broilers are now displayed in self-service displays

 

along with competing meats where consumers can inspect the quality of
broilers and other meats and can choose the meat which they desire.

Nest of the changes discussed above mainly occurred prior to or
shortly after 1950 and were perhaps the main reasons for the increase
in the preference and demand for broilers with the price of broilers,
which was increasing, being relatively unimportant. However, these
changes along with other changes in the broiler industry which were
discussed in Chapter I have not been as pronounced since 1950, and the
consumer preference and demand for broilers now appears to be more
influenced by the downward trend in the price of broilers. This is
supported by the results of Chapter V for the analysis indicated changing
relationships between price and consumption prior to 1950 and more
stability in the relationships after 1950.

Possibly a high income elasticity of demand existed for broilers
over the period studied. This probably resulted from the fact that
chicken was once considered to be a luxury and was consumed almost

entirely on Sunday. If the influence of preferences could have been

84
separated from the income coefficients in this analysis, it is believed
by the author that the analysis would have still indicated a high
income elasticity of demand for broilers.

Rask0pf2 found in Tennessee that per capita consumption of broilers
for families living in the county averaged 2.5 pounds lower than for
families living in cities or suburbs. Since urban families eat more

broilers than families living in the country, this would indicate that

 

the national trend toward urbanization caused an increase in the demand

for broilers.

 

Although increased income and urbanization may have increased the
demand for broilers, availability and quality of broilers are believed
to be the most important factors causing the increased preferences
for broilers. However, in the future it is expected that income will
become a more important factor affecting the demand for broilers as
quality becomes stablized. '
Some Problems and Suggestions for Future Studies

Each of the models presented in this study contains sets of inde-
pendent variables that were highly correlated with the dependent variable;
in each model the independent variables were associated with 85 percent
or more of the variance occurring in the dependent variable. Although
each of the models presented in this study estimated the observations
within the period studied, 1950 through 1958, with a high degree of
accuracy, reliability of future predictations is questionable and will

depend largely upon whether the independent variables remain as highly

 

2RaskOpf, Factors Affecting Per Capita Consumption of groilers in
Tennessee, p. 21.

 

85

intercorrelated in the future. High intercorrelations existing between
the independent variables give sufficient reason to question the degree
of reliability which one might place on the various regression co-
efficients and the elasticities computed from these coefficients; this
reduces the usefulness of the regression models and especially the

regression coefficients for those variables that were highly correlated

 

with other independent variables included in the same equation. Should
the independent variables included in the models become less inter-
correlated in the future, predictions obtained from these models can-
not be expected to exhibit as high a degree of accuracy as the estimates
obtained during the period studied. The correlation between income and

time in the original models was the highest of the intercorrelations

 

existing in the models calculated for the analysis. Since income and
tastes or preferences of the consumers were highly correlated over tflme,
the income coefficients in the models after time was deleted would be
expected to include the influence of changes in tastes because it was
not possible to separate the influence of the two variables. If the
regression coefficients of other variables do not represent the real
influence of the variables on the level of demand due to the inter-
correlations, then a change in the intercorrelations may result in poor
estimates of the dependent variable.

Results of the graphic analysis indicate that data prior to 1950
are of little value in a demand study for broilers which uses the single
equation approach. If a longer period of analysis is deemed necessary
by future researchers in this area, then they will need to use

simultaneous equations or delay their analyses until additional obser-

vations are available. Perhaps extension of this analysis to include
the simultaneous equations approach with models for other meats would
yield more realistic results in terms of the elasticities compared to

the elasticities obtained in this study.3

This study does provide use-
ful information on the timing of shifts which have occurred in the relation-
ships and should be of interest to future researchers constructing
economic models in this area.

Since broilers have a shorter production period than most agricult-
ural commodities and three or four batches of birds can be produced in
one year with the same fixed facilities, future work on the demand for
broilers could be extended to include an analysis of the demand for
broilers for shorter time periods than the year used in this study.
Shortening the length of the time period considered should reveal
additional relationships that exist but have been lost in the analysis
because annual estimates have been used. ‘

Distributed lags were not considered in any of the models used in
this analysis, but lagged variables may be considered to be relevant
in future research. Perhaps there is a tendency for changes in con-
sumption patterns to lag behind changes in income or to be based on
anticipated income or perhaps a tendency for consumption in one period
to be dependent on consumption in the past period. However, with the
time period being one year it is possible that the effect of any lags

may be quite small.

 

3 The study by Nordin, Judge and Wahby discussed in Chapter III is
a possible beginning point for such a study.

 

 

 

87

BIBLIOGRAPHY

Agricultural Marketing Service, United States Department of Agriculture,
Chickens and Eggs: Farm Production1,Disposition,,Cash Receipts and
Gross Income By States, Revised Estimates 1950-54, Statistical
Bulletin No. 183, June 1956.

, Chickens and Eggs: Farm Production,
Disposition, Cash Receipts,_Gross ngomeLBy States, 1954-55, AMS-lla,
April 1956.

, Chickens and Eggs: Farm Production,
Disposition,,Cash Receipt§,,Gross Income By_States,Al956-57, April
1958.

, Chickens and4§ggs: Farm Production,
Dispositon, Cash Receipts, Gross Ipcome by States, 1957-58, April
1959.

, Egg and Poultry Statistics Through 1957,
Statistical Bulletin No. 249, May 1959.

, Livestock and Meat Statistics, 1957,
Statistical Bulletin No. 230, July 1958.

, Supplement for 1958 to Livestock and
Meat Statistics, Supplement for 1958 to Statistical Bulletin No.
230, JUne 1959.

, §upplement for 1956 to Consumption of
Food in the United:§£ates, 1909-52, Agriculture Handbook No. 62,
September 1957.

, Supplement for 1957 to Consumption of
Food in the United4§tates, 1909-52, Supplement for 1957 to Agric~
ulture Handbook No. 62, August 1958.

, The Poultry and Egg_Situation, November
1958, March 1959 and May 1959.

Agricultural Marketing Service and.Agricultural Research Service, United
States Department of Agriculture, Agricultural Outlook Charts,,1959,
November 1958.

Bain, Joe 3., Price gheopy, Henry Holt and Company, New York, 1952.

Baker, George L., Jr., and Paul L. Farris, An Empirical Estimate of
Broiler Price Elasticity and Crosswglasticities at Retail, Paper
presented at the Joint Annual Meeting of the American Farm Economic
Association and the Canadian Agricultural Economics Society,
Winnipeg, Canada, August 20-22, 1958.

 

 

 

88

Beanblossom, Floyd 2., and Marshall‘M. Miller, Market Demands on Broilers,
B- -9lO, Texas Agricultural Extension Service, College Station, Texas,
JUne 1958.

 

Bluestone, Herman, Delmarva-Mapyland-Delaware Broiler Statistics and
Related Data, 1934-57, issued cooPeratively by the Agricultural
Extension Services of the Universities of Maryland and Delaware
and the Maryland-Delaware CrOp Reporting Service, Agricultural
Marketing Service, United States Department of Agriculture,
August 1958.

 

Boulding, Kenneth E., Economic Analysis, 3rd Edition, Harper & Brothers,
New York, 1955.

Branson, Robert E., The Consumer Market for Beef, Bulletin 856, Texas
Agricultural Experiment Station , College Station, Texas, April 1957.

 

, and George J. Mountney, Consumer Attitudes and Pref- ‘
erences Regarding Chicken, Bulletin 895, Texas Agricultural Experi-
‘mental Station, College Statinn, Texas, March 1958.

 

Bureau of Agricultural Economics, United States Department of Agriculture,
Farm.Production,_Disposition and Income from Chickens and Eggs,_1909-
44, Revised Estimates, June 1952.

 

, Farm Production, Disposition and Income from
Chiclens and Eggs, 1945-49, Revised Estimates, June 1952.

 

Bureau of Labor Statistics, United States Department of Labor, ”Weekend
Prices of Food," Retail Food Prices by Cities (mimeographl September
1956 and October 1956.

Christensen, Raymond . and Ronald L. Mighell, Competitive Position of
Chicken and Egg:Production in the United States, Technical Bulletin
No. 1018, Bureau of Agricultural Economics, United States Department
of Agriculture, August 1950.

Clawson, Marion, "Demand Interrelations for Selected Agricultural Products,”
Qparterly Journal of Economics, Vol. 57, February 1943, pp. 265-302.

 

Croxton, Frederick E. and Dudley J. Cowden, Applied General Statistics,
2nd Edition, Prentice-Hall Inc., Englewood Cliffs, New Jersey, 1955.

Dixon, W. J. and F. J. Massey, Jr., Introduction to Statistical Analysis,
2nd Edition, MCGrawbflill 3001: Company, Inc., New‘York, 1957.

Fisher, Malcolm R., "A Sector Model -- The Poultry Industry of the U.S.A.,”
Econometrica, Vol. 26, No. 1, January 1958, pp. 37-66.

Foote, Richard J., Analytical:Tools for Studying Demand and Price
Structures, Agriculture Handbook No. 146, Agricultural Marketing
Service, United States Department of Agriculture, August 1958,

89

 

and Karl A. Fox, Analytical Tools for Measuring Demand,
Agriculture Handbook No. 64, Agricultural Marketing Service, United
States Department of Agriculture, January 1954.

Fox, Karl, A., "Factors Affecting Farm Income, Farm Prices and Food Con-

sumption," Agriculturaligconomics Research,Vol. 3, No. 3, July 1951,
pp. 65-111.

, The Analysis of Demand for Farm Productp, Technical
Bulletin No. 1081, United States Department of Agriculture,
September 1953.

Friedman, Joan and Richard J. Foote, Computational Methods for Handling
Systems of Simultaneous Eguations, Agriculture Handbook No. 94,

Agricultural'Marketing Service, United States Department of Agriculture,
November 1955.

 

 

Gerra,‘Martin J., ”Do Broiler Prices Rise When Supplies Fall?",
Agricultural Situation, Vol. 42, No. 7, Agricultural Marketing

Service, United States Department of Agriculture, July 1958,
pp. 12-130

Gieseman, Raymond R., Changes in.Meat Purchases During_1952 and 1953
Related to Family Characteristics, Unpublished M.S. Thesis,

Department of Agricultural Economics, Michigan State University
1956.

Gray, Leo R. and William L. Mitchell, Marketing Spreads for Eggs and Frying

Chickens in the United States and Selected Cities, ANS-296, Agricultural
Marketing Service, united States Department of'Agriculture, January 1959.

Hood, Hm.C. and Tjalling C. K00pmans, Studies in Econometric Method,
John Wiley & Sons, Inc., New York, 1953.

Karpoff, Edward and John J. Scalan, "Eggs and Poultry,” Marketing, The
Yearbook of Agriculture - 1954, United States Department of
Agriculture.

Kennerly, A.B., "The Broiler Business in Transition,” Texas Agricultural

Klein, Lawrence R., A Textbook of Econometrics, Row, Peterson and Company,
Evanston, Illinois, 1953.

Kohls, R. L., H. G. Diesslin,‘M. I. Bevins and R. D. Halleck, The Broiler
Industry in Two Areas oprndiana, Station Bulletin 608, Agricultural

Experiment Station, Purdue University, Lafayette, Indiana, September
1954.

 

Larzelere, H. E., "What Needs To Be Done? The Poultry Industry,” Proceedings
of the Iowa State College Feedjhivestock Workshqp, February 16-18, 1959,

Ames, Iowa, Center for Agricultural Adjustment, Division of Agriculture,
Iowa State College, Ames, Iowa.

 

 

9O

Lebrun, Edmond J., Consumer Purchase and Use of Poultry Meat, Bulletin
524, Maine Agricultural Experiment Station, University of Maine,
Orono, Maine, January 1954.

 

Leftwich, Richard H., The Price System and Resource Allocation, Rinehart
and Company, Inc., New York, 1955.

Manderscheid, Lester V., ConsumerLExpenditures As A Measure of Demand
For Food, Unpublished M. S. Thesis, Iowa State College, 1952.

Marshall, Alfred, Principles of Economics, 8th Edition, Macmillan & Co.,
London, 1920.

‘Meinken, K. V., A. S. Rojko and G. A. King, ”Measurement of Substitution
in Demand from Time Series Data -- A Synthesis of Three Approaches,”
igurnal of Farm Economics, Vol. 38, No. 3, August 1956, pp. 711-735.

 

Miles, James F. and Wendell H. Thomas, Marketing Poultry Meats in Retail
Food Stores in South Carolina, Bulletin 455, South Carolina Agri-
cultural Experiment Station, Clemson Agricultural College, Clemson,
South Carolina, April 1958.

Morrissett, Irving, l”Some Recent Uses of Elasticity of Substitution --.
A Survey,” Econometrics, Vol. 21, No. 1, January 1953, pp. 41-62.

 

Mortenson, W. P., Trends in the Poultry Industry ... Effects in The
Midwest, Bulletin, 532, Agricultural Experiment Station, University
of Wisconsin, Madison, Wisconsin, June 1956.

Nadin, K. D., and G. A. Jackson, Jr., Financing Western Broiler Production,
Bulletin 753, California Agricultural Experiment Station, University
of California, Berkeley, California, May 1956.

Nerlove, Marc, Distributed Legs and Demand Analysis for Agricultural and
Other Commodities, Agriculture Handbook No. 141, Agricultural
Marketing Service, United States Department of Agriculture, June 1958.

Nordin, J. A., George G. Judge and Omar Wahby, Applications of Econometric
Procedures to the Demands for Agricultural Products, Research Bulletin
410, Agricultural Experiment Station, Iowa State College, Ames, Iowa,
JUIy 1954.

Nybroten, Norman, Consumer Preferences for Poultry Meat, Bulletin 389,
West Virginia Agricultural Experiment Station, Morgantown,‘West
Virginia, May 1956.

 

Parsons, Merrill M., Newppaper Advertising of Meat Products and lgs Relation
to Consumer Purchases in Lansing,'Michigan,_During 1956, Unpublished
M. S. Thesis, Department of Agricultural Economics, Michigan State
University, 1958.

 

91

Pullen, W. E. and W. E. Savage, Sources of Main Poultry Meat and Market
Outlets, Bulletin 531, Main Agricultural Experiment Station, Univer-
sity of Main, Orono, Maine, June 1954.

Raskon, B. D., Consumer Rating of Broilers in Tennessee, Bulletin No.
255, Agricultural Experiment Station, University of Tennessee,
Knoxville, Tennessee, November 1956.

, Factors Affecting Per Capita Consumption of Broilers in
Tennessee, Bulletin No. 288, Agricultural Experiment Station,
University of Tennessee, Knoxville, Tennessee, August 1958.

Riley, Harold M., Some‘Measurements of Consumerpggmand for‘Meats, Unpub-
lished Ph.D. Thesis, Department of Agricultural Economics, Michigan
State College, 1954.

Saunders, Richard F., Consumer Poultry Meat Studies in the Northeast,
Bulletin 536, Maine Agricultural Experiment Station, University of
Maine, Orono, Maine, January 1955.

, Supermarket Sales of Poultry Meats, Bulletin 537,
Maine Agricultural Experiment Station, University of Maine, Orono,
'Maine, February 1955.

, Contract Broiler Growingpin Maine, Bulletin 571,
Maine Agricultural Experiment Station, University of Maine, Orono,
Maine, May 1958.

Schultz, Henry, The9gheory and Measurement of Demand, University of
Chicago Press, Chicago, Illinois, 1938.

Shaffer, J. D., G. G. Quackenbush, and T. N. Moss, The Consumption of Meat
and Related Products in Lansing, Michigan, Spring, 1950, Technical
Bulletin 239, Agricultural Experiment Station, Michigan State College,
March 1954 .

Shepherd, Geoffrey, Changes in the ngand for Meat and Dairy Products in
the United States Since 1910, Research Bulletin 368, Agricultural
Experiment Station, Iowa State College, Ames, Iowa, November 1949.

, Agricultural Price Analysis, 4th Edition, Iowa State
College Press, Ames, Iowa, 1957.

: J. C. Purcell and L. V. Manderscheid, Economic Analysis
of Trends in4§eef Cattle and Hog,Prices, Research Bulletin 405, Agri-
cultural Experiment Station, Iowa State College, Ames, Iowa, January
1954.

 

Smith, Harold D.,‘Market Area For Delmarva Broilers, Bulletin A-64,
iMaryland Agricultural Experiment Station, University of Maryland,
College Park, Maryland, August 1951.

92

, Consumer Preference and_§uying Habits for Chicken,
Bulletin X-S, Department of Agricultural Economics and Marketing,
Maryland Agricultural Experiment Station, University of Maryland,
College Park, Maryland, September 1951.

, Consumer Demand for Chicken As Related toégemand
for Total Meat, Miscellaneous Publication No. 112, Department of
Agricultural Economics and Marketing, Maryland Agricultural Experi-
ment Station, University of Maryland, College Park, Maryland,
October 1951.

, and Francis Dallavalle, A_§tudy of the Merchandising
of Poultry and Eggs in Retail §tores, Miscellaneous Publication No.
144, Department of Agricultural Economics and Marketing, Maryland
Agricultural Experiment Station, University of Maryland, College

Park, Maryland, JUIy 1952.

Smith, R. C., Factors Affecting Consumer Purchases of Frying Chickens,
Bulletin No. 298 (Technical), Agricultural Experiment Station,
University of Delaware, Newark, Delaware, July 1953.

Staehle, Hans, ”Relative Prices and Postwar Markets for Animal Food
Products”, Quarterly qurnal of Economics, Vol. 59, February 1945,
pp. 237-279.

Stigler, George J., The Theory of Price, Revised Edition, Macmillan
Company, New York, 1952.

Szatrowski, Zenon, "Time Series Correlated with the Beef-Pork Ratio,”
Egonometrica, Vol. 13, No. 1, January 1945, pp. 60-78.

United States Department of Agriculture, Contract Farming and Vertical
Integration in Agriculture, Agriculture Information Bulletin No. 198,
JUIy 1958.

Wallace, Thomas D. and George G. Judge, Econometric Analysis of the Beef
And Pork Sectors of the Economy, Technical Bulletin T-75, Agricultural
Experiment Station, Oklahoma State University, Stillwater, Oklahoma,
August 1958.

Waugh, Frederick.v. (editor), Readings on Agricultural Marketing, Iowa
State College Press, Ames, Iowa, 1954.

, Graphic Analysis in Economic Research, Agricultural
Handbook No. 84, Agricultural Marketing Service, United States
Department of Agriculture, June 1955.

 

, "A Partial Indifference Surface for Beef and Pork,”
qurnal of Farm.§gonomics, Vol. 38, No. 1, February 1956, pp. 102-112.

 

, Graphic Analysis in Agricultural Economics, Agri-
culture Handbook No. 128, Agricultural Marketing Service, United
States Department of Agriculture, July 1957.

 

Weintraub, Sidney, Price Theory, Pitman Publishing Corporation, New York,
1949.

Working, Elmer J., "What Do Statistical 'Demand Curves' how?,” Quarterly
Journal of Economics, Vol. XLI, No. 2, February 1927, pp. 212-235.

, Demand for Meat, Institute of Meat Packaging,
University of Chicago, Chicago, Illinois, 1954.

 

 

 

APPENDICES

95

APPENDIX A
DATA
Appendix Table 1
Commercial Broiler Production, Farm Chicken Meat Production

and Gross Income from Commercial Broilers, United States
1934 to 1958

 

Farm Chicken
Year Meat Production
(million pounds)

Commercial Broiler
Production
(million pounds)

Gross Income
fromngrpilers
(million pounds)

1934 2,105 97 19
35 2,210 123 25
36 2,410 152 31
37 2,032 196 42
38 2,185 239 46
39 2,338 306 52
40 2,158 413 72
41 2,586 559 103
42 3,005 674 155
43 3,679 833 238
44 3,009 818 235
45 3,315 1,107 327
46 2,715 884 289
47 2,668 936 302
48 2,289 1,127 405
49 2,643 1,570 443
50 2,310 1,945 533
51 2,312 2,415 689
52 2,025 2,624 756
53 2,046 2,904 786
54 1,948 3,236 747
55 1,634 3,309 834
56 1,668 4,270 838
57 1,336 4,683 886
58 1,516 5,431 1,002

 

 

*Source: Chicken Meat Production:

1934 to 1956, Egg_and Poultry Statistics

Through 1957, Statistical Bulletin No. 249, AMS, USDA, May 1959, p. 80;

1957 and 1958, Chickens and Eggs, AMS, USDA, April 1959, pp. 7 and 9.
Commercial Broiler Production and Gross Income from Broilers: 1934 to 1956,

Egg and Poultry Statistics Through 1957, Statistical Bulletin No. 249,
1959, pp. 16 to 17.

AME, USDA,

May

96

Appendix Table 2

Broilers: Farm Price, Retail Price, Farm Value
and Farm Retail Spread, United States, 1940 to 1958*

 

Farm-Retail

 

Year Farm Price Retail Price Farm Value Spread
(cents) (cents) (cents) (cents)
1940 17.3 28.1
41 18.4 30.2
42 22.9 36.1
43 28.6 41.5
44 28.8 41.7
45 29.5 43.1
46 32.7 48.7
47 32.3 51.1
48 36.0 56.7
49 28.2 53.8
50 27.4 58.5 37.4 21.1
51 28.5 61.3 39.0 22.3
52 28.8 61.6 39.7 21.9
53 27.1 60.0 37.0 23.0
54 23.1 54.2 31.6 22.6
55 25.2 56.2 34.6 21.6
56 19.6 49.0 26.9 22.1
57 18.9 47.9 25.9 22.0
58 18.5 47.4 25.5 21.9

 

*Source: Farm Price: 1940 to 1957, Egg and Poultry Statistics Through
1957, Statistical Bulletin No. 249, AMS, USDA, May 1959, p. 22; 1958,
Chickens and Eggs, AMS, USDA, April 1959, p. 17.

Retail Price: Poultry and Egg,Situation, AMS, USDA, November 1958, p. 41.
Farm Value and Farm Retail Spread: L. R. Gray and William L. Mitchell,
Marketing Spreads For Eggs and Frying Chickens;;n The UnitedpStates and
Selected Cities, AMS-296, AMS, USDA, January 1959, p. 21.

97

Appendix Table 3

Per Capita Consumption of Total Chicken Meat and Broilers,
Ready-to-Cook Basis, United States, 1909 to 1958*

 

 

Total Total
Year Chicken Broilers Year Chicken ggrpilers
(pounds) (pounds) (pounds) (pounds)
1909 14.7

10 15.5 1935 13.1 0.7
11 15.6 36 13.7 0.8
12 14.9 37 13.6 1.1
13 14.5 38 12.7 1.3
14 14.5 39 14.1 1.6
15 14.4 40 ' 14.1 2 0
16 13.8 41 15.4 2.8
17 13.3 42 17.7 3.2
18 13.3 43 23.0 4 1
19 14.2 44 20.4 3.9
20 13.7 45 21.6 5.0
21 13.4 46 19.4 4.1
22 14.2 47 18.1 4.3
23 14.6 48 18.3 5.5
24 13.7 49 19.6 7.1
25 14.3 50 20.6 8.7
26 14.2 51 21.7 10.4
27 15.2 52 22.1 11.7
28 14.6 53 21.9 12.3
29 14.3 54 22.8 13.7
30 15.7 55 21.4 13.9
31 14.1 56 24.6 17.5
32 14.4 57 25.6 19.2
33 14.7 58 28.5 21.7

34 13.5 0.5

 

*Source: Total Chicken Meat: 1909 to 1955, Supplement for 1956 to Con-
sumption of Food in the United States, 1909-52, Agriculture Handbook

No. 62, AMS, USDA, September 1957, p. 13; 1956 and 1957, Supplement

for 1957 to Consumption of Food in the United States, 1909-52, Agriculture
Iiandbook No. 62, August 1958, p. 16; 1958, Poultry and Egg Situation, AMS,
USDA, November 1958, p. 39.

lBroilers: Data for per capita consumption of broilers were not available
prior to 1934; 1934 to 1939, Calculated from data in chmarva-Maryland-
ggglaware Broiler Statistics and related data ... 1934-57, p. 45; 1940

t:o 1957, ggg_and Poultry Statistics Through 1957, Statistical Bulletin
ISO. 249, AMS, USDA, May 1959, p. 8; 1958, Calculated from the data in

Eigultry and Egg Situation, AMS, USDA, November 1958, p. 39.

 

Appendix Table 4

Data for Variables in Regression Equations*
Relative Retail Prices

Retail Prices

 

 

Beef Pork Beef
Year Broilers Broilers Pork Beef Pork
Y3 Y4 Y5 XZ .43
(cents) (cents)
1925 .882 30.7 34.8
26 .842 31.4 7.3
27 .940 32.8 34.9
28. 1.137 37.4 32.9
29 1.163 39.2 3.7
30 1.117 36.2 32.4
31 1.128 30.0 26.6
32 1.412 24.9 17.6
33 1.378 21.5 15.6
34 1.110 23.3 21.0
35 1.010 30.5 30.2
36 .960 28.6 29.8
37 1.062 32.5 30.6
38 1.051 28.7 27.3
39 1.194 29.5 24.7
40 1.050 .769 1.366 29.5 21.6
41 1.043 .901 1.158 31.5 27.2
42 .970 .911 1.064 35.0 32.9
43 .872 .817 1.068 36.2 33.9
44 .820 .765 1.072 34.2 31.9
45 ..777 .745 1.044 33.5 32.1
46 .873 .848 1.029 42.5 41.3
47 1.209 1.188 1.018 61.8 60.7
48 1.328 1.088 1.220 75.3 61.7
49 1.271 1.037 1.226 68.4 55.8
50 1.289 .942 1.368 75.4 55.1
51 1.439 .966 1.490 88.2 59.2
52 1.406 .933 1.506 86.6 57.5
53 1.152 1.058 1.088 69.1 63.5
54 1.264 1.196 1.057 68.5 64.8
55 1.201 .975 1.232 67.5 54.8
56 1.347 1.063 1.267 66.0 52.1
57 1.474 1.257 1.173 70.6 60.2
58 1.709 1.367 1.250 81.0 64.8

 

*Data for variables Y1 and X7, the per capita consumption of broilers, and
variables Y2 and X1, the retail price of broilers, were given in previous
tables (per capita consumption in Appendix Table 3 and retail price in
Appendix Table 2) and are not included in this table.

Source: Retail Price of Beef: 1925 to 1957, Livestock and Meat Statistics,
1957, Statistical Bulletin No. 230, AMS, USDA, JU1y 1958, p. 271; 1958, Suspic-
ment for 1958 toyLivestock.and.Meat Statistics, Supplement for 1958 to

99

Appendix Table 4 - Continued

Data for Variables in Regression Equations
Relative Per Capita Consumption

 

 

Beef Pork Beef
Year Broilers Broilers Pork
X4 X5 X6

1925 .891
26 .941
27 .805
8 .687
29 .714
30 .730
31 .711
32 .661
33 .728
34 .991
35 1.099
36 1.098
37 .989
8 .935
39 .845
40 27.450 36.750 .747
4 21.750 24.429 .890
42 19.125 19.906 .961
43 13.000 19.244 .676
44 14.256 20.385 .699
45 11.88 13.320 .892
46 15.024 18.488 .813
47 16.186 16.186 1.000
48 11.473 12.327 .931
49 9.000 9.535 .944
50 7.287 7.954 .916
51 5.394 6.914 .780
52 5.316 6.188 .859
53 6.309 5.163 1.222
54 5.847 4.380 1.335
55 5.899 4.806 1.228
56 4.880 3.851 1.267
57 4.406 3.203 1.376
58 3.710 2.797 1.326

 

Statistical Bulletin No. 230, AMS, USDA, JUne 1959, p. 133.

Retail Price of Pork: 1925 to 1957, Livestock and Meat Statistics, 1957,
Statistical Bulletin No. 230, AMS, USDA, July 1958, p. 272; 1958, Supple-

ment for 1958 to Livestock and Meat Statistics, Supplement for 1958 to
Statistical Bulletin No. 230, AMS, USDA, June 1958, p. 133.

Per Capita Consumption of Beef: 1925 to 1956, Livestock and‘Meat Statis-

tics, 1957, Statistical Bulletin No. 230, AMS, USDA, July 1958, p. 283;

1957 and 1958, Supplement for 1958 to Livestock and Meat Statistics, Supplement
for 1958 to Statistical Bulletin No. 230, AMS, USDA, June 1959, p. 138.

100

Appendix Table 4 - Continued

Data for Variables in Regression Eguations

 

 

 

Per Capita Consumption Per Capita Disposable
Year Beef Pork Personal Income
X8 X9 X0
(pounds) (pounds) (dollars)
1925 59.5 66.8
26 60.3 64.1
27 54.5 67.7
8 48.7 70.9
29 49.7 69.6
30 48.9 67.0
31 48.6 68.4
32 46.7 70.7
33 51.5 70.7
34 63.8 64.4
35 53.2 48.4
36 60.5 55.1
37 55.2 55.8
38 54.4 58.2
39 54.7 64.7
40 54.9 73.5
41 60.9 68.4
42 61.2 63.7
43 53.3 78.9
44 55.6 79.5
45 59.4 66.6
46 61.6 75.8
47 69.6 69.6
48 63.1 67.8
49 63.9 67.7
50 ' 63.4 69.2 1,369
51 56.1 71.9 1,473
52 62.2 72.4 1,520
53 77.6 63.5 1,582
54 80.1 60.0 1,582
55 82.0 66.8 1,660
56 85.4 67.4 1,727
57 84.6 61.5 1,782
58 80.5 60.7 l,786

 

Per Capita Consumption of Pork: 1925 to 1957, Liyestock.and.Meat Statis-
tics,,1957, Statistical Bulletin No. 230, AMS, USDA, July 1958, p. 284;
1958, Supplement for 1958 tongvestock and Meat Statistics, Supplement
for 1958 to Statistical Bulletin No. 230, AMS, USDA, JUne, 1959, p. 138.

Per Ca ita DiSposable Personal Income: 1950 t 1957, Su lement for 1957
to Con umption of Food in the United States, 1 09:52,

to Agricultural Handbook No. 62, AMS, USDA, August—I958, p. 29; 1958,
Poultry and EggpSituation, AMS, USDA, November 1958, p. 41.

101

APPENDIX B
Alternative Regression Equations
Demand for Broilers

Consumption of Broilers

A
log Y1 = -6.22661 - 0.86070110g x1 + 0.2887321og X2 + 0.2332231og X3 + 2.4680110g X0
(.19495) (.11439) (.15013) (.23569)
82 = .989 s; - .014

Price of Broilers

/'
log Y2 = -3.05852 - 0.85963210g x7 - 0.38821 16g x8 - 0.28316 10g x9 + 2.18956310g x0
(.31175) (.25511) (.40976) (1.12356)
82 = .817 s§'= .020

Demand for Beef Relative to Broilers

A 1
Y3 = 4.95471 - 0.23867 x - 0.00779x - 0.001103x

(.05373) 4 (.00891)9 (.0048) 0

82 = .746 s; = .085

Demand for Pork Relative to Broilers

A
Y4 = 4.60063 - 0.209812X - 0.00595x - 0.00125}:O

(.08014) 5 (.00541)8 (.00088)
82 = .707 s"= .084
y
Demand for Beef Relative to Pork
4
Y5 = 1.72476 - 1.006411x6 + 0.01771x + 0.00028X0
(.11833) (.01266) (.00042)~
R2 = .925 s; = .044

 

1Significant at the .01 level.
2Significant at the .05 level.
3Significant at the .10 level.