A COMPARISON OF STATIC AND DYNAMIC INCOME-EXPENDI'WfiE RELATIONSHIPS FOR SELECTED MEAT ITEMS Thais fat the Dear» a? M S. MICHIGAN STAYIE UNIVERSITY ': Robert J. Bachfaéa ‘ :95? "I!“ LIBRARY (fit. A COMPARISON OF STATIC AND DYNAMIC INCOMEPEXPENDITURE RELATIONSHIPS FOR SELECTED MEAT ITEMS By Rdbert J. Bachleda .A THESIS submitted to the College of Agriculture Rfichigan State University of Agriculture and Applied Science in.partial fulfillment of the requirements for the degree of NMSTER OF SCIENCE ‘Department of Agricultural Economics 1959 ACKDUWLEDGMEN’I‘S Sincere gratitude is expressed to the Department of Agricultural Economics for the academic instruction ani financial assistance provided during the author‘s graduate studies . - The author is indebted to Dr. H. M. Riley for his guidance, patience, and constructive criticisms offered throughout the preparation of this manuscript. Special thanks are extended to Dr. G. G. Quackenbush and Dr. J. D. Schaffer for their excellent cooperation in supplying the essential data which made this study possible. Appreciation is also expressed to Dr. L. Manderscheid for his helpful suggestions. Special thanks is expressed to Mrs. Judy Leach for her aid in tabulations arri typing the original manuscript. Similar appreciation is extended to Mrs. Shirley Goodwin for typing the manuscript in its present form. The author will accept the responsibility for any remaining errors in the manuscript. Wfi A.COMPARESON OF'STATIC AND DYNAMIC INCOME-EXPENDITURE RELATIONSHIPS FOR.SELECTED MEAT ITEMS By Robert J} Bachleda AN ABSTRACT Submitted to the College of Agriculture 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 Year 1959 ABSTRACT The primary purpose of this study was to examine the validity of the implicit assumption involved when the relationships derived from cross-sectional income-expenditure studies are used in making predictions. The assumption is that families experiencing an increase in income will adjust their meat expenditure patterns to be similar to the expenditures of families in the income group they are entering. A secondary ob- jective was to determine the relationships between family size and adjustments in meat expenditures associated with increased income. The data were provided by a ninety-nine family sub-sample of the Michigan State University Consumer. Panel. During the time period , 1953-57, annual family expenditures for eight meat items were "observed" before and after families received an income increase. Per capita income per family served as the criterion for placing families into two income groups. Income groups were then sub-classified by two family sizes . Static cross-sectional income-expenditure elasticities were com- puted between income groups. The "expected" expenditure adjustments were computed by applying these cross-sectional relationships to actual income changes. The “observed" adjustments in family expenditures per capita associated with increased income served as the basis for comput- ing dynamic income-expenditure elasticities for each income group. iv "It ‘fl, “I“ I - ilk A comparison of the "observed“ and ”expected" expenditure adjustments served as a basis for judging the validity of the assumption involved when cross-sectional.relationships are used for making predictions. Student's "t" test was used to evaluate differences in expenditure levels between families grouped by income and family size. This test was also used to indicate if the "observed" expenditure adjustments 'were significantly different from zero. The results of this study indicated that the validity of the assumption that families experiencing an increase in income will adjust their meat expenditure patterns to be similar to the expenditures of families in the income group they are entering can.be seriously questioned. It was found that per capita meat expenditures were sig- nificantly related to family size among low income families. It was also found that high income families were smaller in size than low inr come families. Predictions of expenditure adjustments based upon cross-sectional relationships are therefore likely to be greater than actual dynamic adjustments associated with a rise in income. There was also some evidence that year to year expenditure adjustments are less than "expected" adjustments based upon cross-sectional relationships. It was further observed that dynamic income-expenditure adjustments are probably affected by inflationary income increases and shifts in demand associated with changes in tastes and preferences. "a! TABLE OF CONTENTS ' CHAPTER Page I IMWUCTION...‘....'...C..‘C....-.....0..........C........... Purpose Of Study....u......u......o...o................ PreVious StUdieSooooooo00.000000000000000...oooooooococoa Time-Series WBSCOQOOOQoOQOQOOOQ0.0000000000000000. CI‘OSS‘SeCt-ional AnalyseSoooooooooo0.0000000000000000... Sumary 0f Pre'ViOUS StIldieSoooooooooooo000000000000.cocoa 15 USGfulness 0f Result-8.00000000000000000.0000000000000000. 16 OD-I-TUJI-J I--' II mmom OF MEI—3000000on...coooooooooooocooocooooooooooo 17 The GemraJ. ApproaChoooooooo00.00.00000000000000000000000 17 Source 8311 Nature Of Data.............oo................. 18 Selection Of Families......................u............ 19 Units 0f Measuremenbouu...“"no“.................... 20 GrOUPin-g 0f FamjlieSooooooooooooococooooooooooooooooooooo 21 SeleCtion 0f Meat CUtSoooooooooooooooooooo0.000.000.0000. 22 Selection Of Time P6ri.od..............ouu.............. 23 Hem InveStigationoo0.000.000.0000.cognac-cocoa... 30 He'I-JhOdOlOgoooooooooooooooooooooocooo00000000000000.0000. 3O Static Cross-Sectional Relationships................... 30 W0 RelatiODShiPSooooooooooooo0.000000000000000coco 3h Comparison of the Static and Dynamic Income-Expenditure Relationshipsouuu....an.......o.n......o......... 36 Limitatiom 0f mysisooooooooococo0.0000000000000000... 39 III Games IN BEEF CUT EXPENDITURES ASSOCIATED WITH ImREASED ImOEOOCOOOCOOO'OOOOCOOOOOOOOOOOOOOOOOOOOOODOOOOOOOOOOOO h]- Introduction...”................o...o................... ’41 Results from a Static Cross-Sectional Analysis of Beef R0831} arri Steak Experfiitures........onuo.o.o......... All Results from a Dynamic Analysis of Beef Roast and Steak Expenditures.............o....o......................... AIS Comparison of the Observed and Expected Changes in Beef R0851.) am Steak hWMitureSoococo00.000000000000100... I48 continued vi ' 'i "' ‘fiu '5‘ u . not I O ‘ \ r, . ~ Q I -. t . ‘ t 1 O l v 1 I ‘ '5 - n O I I ‘ v 0 Q y '\ c v '\ Q .— C O O O r. 1 0 .0 TABLE OF CONTENTS - Continued CHAPTER Page IV CHANGES IN SELECTED PORK CUT EXPENDITURES ASSOCIATED WITH Immm IMOMOOOOCOOOOOOOOOO0.0.0.0....OOOOOOOOOOOOOOO 51 IntrOduCtiOnoooo000000000000000.00000000000000000.00.0000 5].. Results from 3 Static Cross-Sectional Analysis of Selected Pork Item ExpeIfiitureSOOOcoooooooo0.0000000... 52 Results from a Dynamic Analysis of Selected Pork Item Experditures...u.o."nun“...nuo...“............ 55 Comparison of the Observed and Expected Changes in Pork Item mmmtureSoooooooooooooooooooooooooooooooooooooo 59 V CHAIRS IN BROILER AND COLD CUT EXPENDITURE ASSOCIATED WITH ImREASED WOWOOOOCO0.0COOOCOOOOOOOOC.0.0.0.000... 62 IntrOdUCtiono.00000000000000.000000.00000000000000000000o 62 Results from a Static Cross-Sectional Analysis of BrOiler kaditureSooooooooooooQOOQ00000000000000.9900 6’4. Results from a Dynamic Analysis of Broiler Experditures.. 66 Comparison of the Observed and Expected Changes in Broiler ExpeIflitUreS.....................a.o...o....... 68 Rewlts from a Static Cross-Sectional Analysis of Cold CHI; Expeniitures.c."no............................... 69 Results from a Dynamic Analysis of Cold Cut Expeniitures. 71 Comparison of the Observed and Expected Changes in Cold GUI: Ebtperflitures....u.....o.u........................ 73 VI smm COMLmIOEOOOOOOOOOOOOOOOOOOOOOOOOO...00...... 75 Emomr....‘OOOOOOOOOIOOOOOC...0....0..COCOOOOOOOO;OIOOOOOOOO 8h mmn.’COOOOOOOOOOOOOCOCOOCOOOOOOOOCOOCOOOOCOOOOOOOCOO...0.00.. 86 vii T! 'f-_ 5'; 's avw ‘. E ' I I ,. ~ * . a A n. > 1 ~ " fl " I I O I: . , a a . 9 - , - s , q a ’\ . c a r . r x a ,t , .. ' - a l -. t e . A I y .. t - g ' I I f‘ u 1 - Q ' 1 Q . .. 4 . O C 1 Q 9 O 1 . t O I I n 1 " \ u a 4 ~I ‘5 0 O x k. ii 'u |Q an o. I h I ‘.\Q \fl. I.’ 5 I (I K Q .A V I” TABLE 8a 10 LIST OF TABLES Inc ome-Quantity Elasticities for Selected Meat Cuts , Urban Families in the United States, Spring, l9h8........... ...... Weekly Expenditures for Selected Cuts of Meat by Per Capita Incqme Groups, MarCh and June, 19h800000000000000000.0000... Income-Quantity Elasticities for Meat Items by Size of Family) LanSing, Spring, 19500000000000.0000...ooooooooooooo Family Characteristics by Per Capita Income Groups, Sub- Sample of Ninetyhhtine Families, Michigan State University Consumer Panel, 1953‘195700000000600000oooooooooooooooooocoo Average Annual Prices for Selected Meat Cuts , Michigan State University'Consumer PanBl, 1951-19570.0000000000000000000000 Per Capita Disposable Income of a Ninety-Nine Family Sub- Sample, Michigan State University Comumer Panel, 1953-1957. Average Annual Per Capita Income, by Income Groups Sub- divided by Family Size, Ninety-Nine Family Sub-Sample, Michigan State University Consumer Panel, l953-l957......... Per Capita Expenditures for Selected Beef Cuts, by Income Groups, Michigan State University Consumer Panel Sub-Sample, 1953‘57000000000000000.000.000.000...00000000000000.00000000 Per Capita Experditures for Selected Beef Cuts, by Income Groups, Michigan State University Consumer Panel Sub-Sample, 1953-570000000000000.000000000000000000000000000000000000000 Average Level of Family hpenditures for Selected Beef Cuts, All Urban Families in the North Central. Region, A Week in .April-June, 1955000000000000000.0000...00000000000000.oooooo Per Capita Expenditures for Selected Beef Cuts, by Income Groups Subdivided by Family Size, Michigan State University consumer Panel SUb-Sample, 1953‘570000oooooooooooooooooooooo Annual Average Per Capita Beef Cut Expeniitures Associated With Increased Income, by Income Groups, Michigan State University Consumer Panel Sub-Sample, 1953-57............... Page 10 22 26 29 32 311 11,2 12 11.5 In. "W- J‘ n“ I C'o's b Q ~ 1 ‘ c _ u A o m I 5 O . " I t f’ ‘ I . . I U -- ‘ ‘I l . . C I a C D Q a a A A‘\ LIST OF TABLE - Continued TABLE 12 16 17 18 19 2O 21 22 Page Annual Average Per Capita Beef Cut Expenditures Ass ociated with Increased Income, by Income Groups Subdivided by Family Size, Michigan State University Consumer Panel Sub- Sample, 1953‘570000000ooocoo-00.00.00000.0000000000000000... AL? Observed am Expected Beef Roast and Steak Adjustments Associated with Increased Income, Low Income Group, Michigan State University Consumer Panel Sub-Sample, l953-S7......... LIB Per Capita Expeniitures for Selected Pork Guts , by Income Groups, Michigan State University Corsumer Panel Sub-Sample, 1953-5700000006000000000OOOOOOOOOOOOOOOOOOOOOOO000.000.0000o 52 Average Level of Family Expenditures for Selected Pork Cuts, All. Urban Families in the North Central Region, A Week in Aprfl‘JIme 1955.00.000000.00.00.0000000000000.000000000000... 53 Per Capita Expenditures for Selected Pork Cuts, by Income Groups Subdivided by Family Size, Michigan State University commer Panel Sllb‘sample’ 1953-57.... 0.00000000000000000000 SLI- Annual Average Per Capita Pork Cut Expenditures Associated with Increased Income, by Income Groups, Michigan State University Consumer Panel Sub-Sample, 1953-57 . . . . . . . . . . . . . . . 55 Dynamic Inc ome-Expenditure Elasticities for Selected Pork Cuts , by Income Groups , Michigan State University Consumer Panel SUb-Sample, 1953-5700. ooooo'ooqoooooooooooo 00000000000. 57 Annual Average Per Capita Pork Cut Experditures Associated with Increased Income, by Income Groups Subdivided by Family Size, Michigan State University Consumer Panel Sub-Sample, 1953’s7oooooo0000000000000cocoa-cocooooooooooooooooooooooooo 58 Observed and Expected Pork Cut Expenditures Ass ociated with Increased Income, Low Income Group, Michigan State University COIBUJIIBI' Panel Sub-Sample, 1953‘s7009000000000 oooooooooooooo 59 Average Level of Family Expeniitures for Broilers and Cold Cuts, All Urban Families in the North Central Region, A Week In April-June, 19550000000000.00.00.0000a...0000000000000000 63 Per Capita Broiler Expenditures, by Income Groups, Michigan State University Consumer Panel Sub-Sample, l953-S7......... 6h "fl LIST OF TABLES - Continued TABLE 23 2h 25 26 27 28 29 Page Per Capita Broiler Experditures , by Income Groups Sub- divided by Family Size, Michigan State University Consumer Panel Sub-Sample, 1953‘s.]...00000000000000...00000000000000. Annual Average Per Capita Broiler Expeniitures Associated With Increased Income, by Income Groups, Michigan State University Consumer Panel Sub-Sample, l953-S7............... Annual Average Per Capita Broiler Experxiitures Associated with Increased Income, by Income Groups Subdivided by Family Size, Michigan State University Consumer Panel Sub-Sample, 1953'570000000000coco-0.0000000000000000.ooooooooooocooooooo Per Capita Cold Cut Expenditures, by Income Groups, Michigan State University Comumer Panel Sub-Sample, l9S3-57......... Per Capita Cold Cut kpenditures, by Income Groups Sub— divided by Family Size, Michigan State University Consumer Panel SUb-Salnple, 1953-5700.00000000000000.0000...0.00.0000. Annual Average Per Capita Cold Cut Experditures Associated with Increased Income, by Income Groups, Michigan State University Consumer Panel Sub-Sample, l953-57............... Annual Average Per Capita Cold Cut Experxlitures Associated With Increased Income, by Income Groups Subdivided by Family Size, Michigan State University Comumer Panel Sub-Sample, 1953‘5700000000000000000000cocoaooooooooo0.0000000000000000. 65 66 67 7O 71 72 73 l . m — ' I I . f g . . ‘0 . .. .u .‘Aosrru' o."'|’t"'|l"h .. '. , 'lb' u\".\'l ’0 'nf 7 ’ ‘,, -- a r- L F- FIGURE LIST OF FIGURES Page I United States Consumer Price Index, for All Items, Total Food II and Total Food Eaten at Home, Department of Labor, Bureau of Labor Statistics, 1953‘57000000000000000000000000000000.0000. 28 Changes in Per Capita Beef Steak Ebcperditures Associated with Income Increases, Michigan State University Comumer Panfil Sub-Sample, 1953-57000000000000000000000000000000.0000. 31 xi rfl‘ CHAPTER I INTRODUCTION Purpose of Study There have been several cross-sectional studies which have investi- gated the relationship between meat expenditures and family income for specified time periods. In using these results for prediction purposes, there is usually an implicit assumption that families experiencing an increase in income adjust their meat purchases so they are similar to the purchases made by the income group the families are entering. The primary purpose of this study was to examine the validity of this implicit assumption. The data utilized in this study were obtained from the Michigan State University Consumer Panel which was organized and operated under the direction of Quckenbush and Shaffer.1 The Consumer Panel was in operation for approximately eight years, beginning in February 1951 and concluding its operations in December 1958. During this time approxi- mately 250 family food purchase diaries were received weekly. The panel was one of the few conducted in the United States which provided weekly information on consumer food purchases. 1J. D. Shaffer and G. G. Quackenbush, "Cooperation and Sampling in Four Years of M. S. U. Consumer Panel Operation," Quarterly Bulletin, Agricultural Experiment Station, Michigan State University, East Lansing, August, 1955, pp. 85-103 nl‘ Consumer panel purchase data were suited to both time-series and cross-sectional analyses.1 Therefore, it was possible to examine con- sumer meat expenditures before and during an increase in income. Expenditure adjustments were then compared with the "expected'changes in expenditures obtained from a conventional cross-sectional analysis of the same data. This comparison was the basis for accepting or rejecting the validity of the implicit assumption used in.making pre- dictions from the income-expenditure relationships derived from cross- sectional studies. The second objective was to determine the level of expenditures for eight meat cuts for different levels of family income. These relationships were then.compared with the results Obtained in other cross-sectional studies. The third objective was to estimate the relationship between family size and per'capita meat expenditures among families with similar levels of per capita income. If per capita meat expenditures differed among families of different sizes with similar levels of income, then, income-expenditure estimates would be subject to error where family size is not homogeneous in the group from'which the relationships were derived and the group whose responses to an increase in income we are attempting to predict. 1For further information on the Objectives and usefulness of this panel, see: G. G. Quackenbush, "Demand Analysis From the NBC Consumer Panel,n Jburnal of Farm Economics, Vol. 36, No. 3, l95h, pp. hlS-h27. ni‘ Previous Studies Engel was the first to make a statistical analysis of the relation- ship between consumption and income. As a result of this study, Engel concluded, "The poorer a family, the greater the preportion of total spending which goes for food purchases.1 The generalization that ins comes are first devoted to the necessities of staying alive and that luxury spending or saving occurs at higher income levels was deveIOped from the above study. Therefore, it was expected that consumers in the lower income bracket would be more responsive in adjusting their food purchasing patterns after experiencing an increase in income than families at higher income levels experiencing similar increases in income. "Income elasticity“ is a quantitative measure of the responsiveness of consumers in adjusting their purchases after experiencing a change in income. Consumer responsiveness can be measured either as changes in quantities or expenditures. For purposes of empirical measurement, income elasticity is usually defined as the percentage change in either quantity or expenditure associated with a one per cent change in income, 2 all other things remaining the same. The variables which may influence income elasticity estimates are assumed constant in this study. These 1George J. Stigler, "The Early History of Empirical Studies of Consumer Behavior," Journal of Political Economy, Vol. 62, No. 2, l9Sh, pp. hB-lOO. 2Ibid., The Theory of Price, (rev. Ed., New York: Macmillan Compam’: 1952): PP- E953. 1 variables may be classified into four groups: 1) conditions which affect the desires of consumers for commodities, such as, changing styles, advertising, and living customs; 2) changes in the quantities available, and prices of competing or substitute products; 3) changes in the composition and prices of all other items in the budget; and h) availability of unused sources of purchasing power over and above current income . Previous studies have been usually classified as one of two types, time-series or cross-sectional analyses. The type of data, methodology, results am limitations of each classification are discussed in the following section. In analyzing data, some researchers prefer to adjust income and expenditures for price level changes by means of an index ._ and other researchers prefer to leave the data unadjusted. It is im- portant to understand that these differences 'in handling and analyzing data result in income elasticities which differ and in some cases are not comparable . Time-Series Analyses ’ Time-series analyses have been largely limited to broad food cate- gories such as, all meat, beef, or pork due to the limiting nature of the data available. Several of these studies were based on annual data for the United States, with observations beginning in the late 1910's 1F. L. Thomsen, "Measuring Changes in the Demand for Farm Products ," Journal of Farm Economics, Vol. 21, No. l, 1939, pp. l32-lh23 see also, Marguerite C. Burk, "Changes in Demand for Food From l9hl to 1950," Journal of Farm Economics, Vol. 33, No. 3, 1951, pp. 291-295. nl‘ am extending into the 1950's. The war years were generally excluded because the effects of price controls and other variables which influ- enced the economic system of the country. These data provide information on income, food consumption, and expenditures over a period of time with the resulting income quantity or expenditure elasticities represent- ing an average responsiveness to an increased income for all families in the United States. These income elasticities are usually computed by regression techniques . However, the relationships between quantities or expenditures, and income are not derived directly from the equation an! some manipulation of the coefficients is usually required to obtain the elasticity estimate. One of the earliest time-series analyses was done by Schultz who used annual average observations for the years 1922-1933.1 Quantity and income data were reduced to a per capita basis to eliminate the variability due to different family sizes. Quantity data represents carcass weight with veal‘being included with the beef data. Prices were deflated by the Bureau of Labor's Consumer Price Iniex. Schultz used a single equation multiple regression technique with the dependent variable being quantity, ani the independent variables being the prices of all meat, beef, pork and income. The estimated income-quantity elasticities for all meat, beef and pork were .18, .h8, and .514 res pectively . “Henry Schultz, ‘Ble Theory and Measurement of Demand, (Chicago: University of ChicagoPress, 1938), p. 5L1. and Shepherd, using a multiple regression analysis estimated the income-quantity elasticity for all meat to be about .75 .1 United States per capita annual average observations for all meats for the period 1920-}41 were used. Fox estimated the income-consumption elasticity for all meat to be .56. Income elasticity estimates for beef and pork were about .7.2 Annual per capita data were used for the period 1922-hl with income elasticities being computed by the single equation method. Linear relationships were assumed between per capita consumption am disposable income per' person. A single New Englard community was the source of two and one-half years of weekly data used in computing short and long-term income- expeniiture elasticities .3 Aggregate data for the community were used rather than data for a sample of households because mpenditure decisions of the household may be formed in such a manner that a budgetary study would not determine the impact of an income change on expenditures. Income payments to consumers. were obtained from the major firms in the area. Bank balances were obtained directly from each bank to determine if not saving or dissaving occurred during the period. The Sales 2‘-(.‘1eoffrey Shepherd, Changes in Demand for Meat and Baggy Products g the United States Since 1210, Res. Bul. 368, (Amos, Iowa: Iowa Agr. hp. Stan, 19 9 , p. 3 10 2Karl A. Fox, “Factors Affecting Farm Income, Farm Prices, and Food Consumption,“ 5gp. Econ. Res., No. 3, 1951, pp. 65-Dl. 3George K. Brinegar, “Short-Run Effects of Income Change on kpenditures,“ Journal of Farm Economics, Vol. 35, No. l, 1953, pp. 99-109 . n“ records of 10 retail firms were used to estimate the purchases of goods by consumers. The analysis of income and expenditure data consisted of fitting linear regressions of data for: 1) six month periods within the two and one-half year period, and 2) the two and one-half year period. The income-expenditure elasticities were computed at the income means. Brinegar calculated confidence limits for the elasticities at the 95 per cent level. In adjusting for price changes, price elasticities were assumed which corresponded to the findings of price elasticity studies. he short-term elasticities calculated for six month periods were characterized by relatively wide confidence limits which prohibited detailed generalizations about meat expenditures . However, the analysis of the two 3111 one-half period yielded an income elasticity for fresh meats unadjusted for price changes and iseasonality of .30 ard .h6 adjusted for seasonality. 'llhe income-expenditure elasticity adjusted for price changes and seasonality was .27 and .30 adjusted for seasonal- ity. Since the short-term elasticities were computed at the mean income, it lies possible that increases in meat experditures for low income fmdlies were offset by decreases in expenditures by large income families, which resulted in the wide confidence limits for the six month periods. Income elasticities based on time-series analyses have been computed by utilizing annual average data for a long period of time. These aggre- gate data .1251 result in an income elasticity which excludes the evidence of trerxis in consumption of meat groups or cuts of meat over a long period of time. Income elasticities based on data from an earlier time period for a product which was of minor importance in the consumer food budget will result in an inaccurate estimate of consumers responsiveness to an increased income if consumers increased the relative importance of this product in their food budget without the benefit of an increase in income. Cross-Sectional hlalyses Gross-sectional or budgetary studies have utilized both survey and panel data for the United States, Western Europe, and specific geo- graphical. areas in computing income elasticities for all meat, beef, pork am selected cuts of beef and pork. Arc elasticity is usually used to determine the responsiveness of consumers in adjusting their meat expenditures to an increase in income. Data collected by the survey method usually represents a short period of time whereas panel data represent a long period of time, usually one year, or average yearly data if more than one year's data were utilized. These studies attempted to analyze the effects of an increased family income on consumer meat purchasing patterns. For prediction purposes, it was assumed in some studies that fanilies adjusted their purchasing patterns so they were similar to the economic class they were entering. Other studies did not make this assumption. Attempts to determine the relationships between socio- economic variables and levels of meat consumption were the objectives of ethos studies . A survey of 2200 families was made in Minneapolis in 19311,, a time of large unemployment and depressed incomes.1 It was found that meat consumption was greater among the high income families. The study illustrated that size ani composition of families had minor influence upon per capita meat consumption but more recent studies2 have indicated that family size was an important variable influencing per capita meat consumption. During both 19148 and 1955, the United States Department of Agri- culture corducted surveys of families in order to determine the average weekly food expenditures. Classification data such as income arxi size of family were obtaimd so similar families could be grouped. Waite and Trelogan used these data to compute income-quantity elasticities for specific meat cuts .3 Table 1 summarizes these results. The elasticities (shown in Table 1 on the following page) are simple arc elasticities using the accord income group, with an average family income of $1,555 and the sixth income group with an average income of $5,861. The wide variation in meat purchases. among the 1w. c. waite and R. w. 0011:, A Study of the Consumption of Meats in Minneapglis, 1931., Minn. Agr. Exp. Sta., Bul. 321, 1935. 2For further details, see: J. D. Shaffer, G. G. Quackenbush and T. N. Moss, The Consum ion of Meat and Related Products in Lanai Michi an S l O, M.S.C. Agr. Exp. Sta., Techn. Bul. 239, 19 pp. 11-1 3 also, W. M. Simone, Consumer Meat Purchases in Syracuse M York 19148 ani Comparisons with 191?, Bul. 359,21thaca, New York: Cornell University Agr. Exp. Sta., 1951), pp. 6-7. _ 3?. C. Waite ani H. C. Trelogan,Aicu1tural Market Prices, 2d. ed. , (kw York: John Wiley & Born, 1951) p. ’ 10 TABLE 1 INCHES-QUANTITY ELASTICITIES FOR SELEIJTED MEAT CUTS, URBAN I‘M IN THE UNITED STATES,.SPRING_19}48 y W ===== Meat Cut Income Elasticity Beef steak other than rouni 1.037 Smoked ham (cooked) .820 Poultry total .636 Beef steak round .553 Bacon .102 Pork chops -.212 Source: Computed from Food Consumption of Urban Families in the United States, l9h8 Food Consumption Survey USDA, Report No. 5. families with the same income irflicates that the reliability of these 1 estimates may be rather low. Simmons conducted a cross-sectional study of 726 families living 2 in Syracuse, kw York with the following objectives: 1. To determine the level of meat consumption in Syracuse, lbw York in March and June 1918: to determine variations in consumption ani expeniitures for meat among different groups of consumers ; and to ascertain the factors associated with or affecting these variations. 2. To determine how changes in economic status for a given group of comumers affect the total quantity of and expenditures for various kinds and cuts of meat. ReSUlts showed that there was a direct relationship between per capita income ani expenditures for beef roasts, beef steaks, bacon, poultry, and cold cuts. Results for pork steaks and chops, pork roasts, am ham were inconclusive. See Table 2,- 1Bur. of Hman mtr. and Home Econ., U. S. Dept. of Agr., Meat: Eariations in Constmption and Interrelationships with Other Foods, Commodity Summary No. ll, 195$ p. 3. BSimmoh's, gp. 33:11., p. 30. TABLEZ KEELI EXPEDDITURFS FOR 3mm CUTS OF MEAT B; PER 0mm Imoma: GROUPS, MARCH AND JUNE l9h8 ‘7 i. Per Capita Inc omeb 0 Heat Cut Low Low-Medium High-Medium High fleekly Expenditures per capita, dollars) Beef ' Steaks .13 .18 .21; .30 Roasts .16 .18 .30 .31; Pork Steaks and chops .10 .ll .09 .09 Rowlb .12 .10 01-]- .10 Bacon .011 .07 .10 .12 Ham .16 .17 .21 .07 Poultry .ll .18 .26 .36 Cold Cuts .22 021 .19 all]. aSource: H. M. Simmons, Consumer Meat Purchases in S acuse New York in 12g8 am Comggisons with 12E2, Bul. 869 EIthaca, New York: Cornell University Agr. Exp. Sta., 1951), pp. 1h,20,23. b Income Group Bourriaries Mean Income Low ‘ 1110-1471: 369 Low-Medium 1175-6714 562 High-Medium 675-9514 79h High 955-3690 1311 Other variables which influenced per capita expeniitures were size of family, nationality, religion, and the activity of the wage earner. Simom concluded that a definite association existed between rates of change in per capita income between 19112 and 19148 ani changes in per capita experditures for meat. In the group whose 1911.2 income was below the median, a third of the families with a low rate of increase in income purchased about 10 per cent less meat per capita in 19118 than in 19h2 in.contrast to the third of the families with the highest relative increase in.income by l9h8 who purchased about 65 per cent more meat per capita than.in.l9h2. (Per capita income in l9h8 was adjusted to 19h2 dollars by dividing per capita income by the cost of living index for Buffalo, Nsw York, with March am June 19112 I- 100.) Those families which.had income below the median.in.l9h2 decreased their beef roast expenditure by 10.h per cent after experiencing an increase in.disposable income. FamilieS'which had income above the median.in l9h2 decreased their expenditures for beef steak and roasts, pork steaks, chops and roasts. The decrease in.expenditures for these meat cuts was greater than.10 per cent with the exception of pork roasts which decreased by 8.8 per cent between.l9h2 and 19h8. Simmons concluded that in the short-run, consumers tended to spend about the same amount for meat. Consumers adjusted to changing market- ing conditions and prices by increasing or decreasing the quantity of meat purchased Or by shifting to other kinis of meat considered "superior" in nature. In the spring of 1950, l885fami1ies in Lansing, Michigan were contacted and interviewed by individuals employed by the Department of Agricultural Economics, MichiganHState University; The objective of this surVGy'was to obtain the quantities and expenditures for food and socio—economic data.by which families were classified. Analysis of the information obtained determdned whether or not differences in family size, income, age of housewife, education of housewife, and occupation of householder were related to significant variations in per capita experriitures. Moss, using cross—sectional tables, illustrated that family income ani size of family appeared to be the most important factors affecting per capita meat consumption.1 Family income appeared to have little ifany effect on poultry expenditures. However, by size of family, a downward trend in per capita poultry expenditures was evident as family size increased. Riley, using data in Moss's thesis, computed a series of simple arc elasticities.2 The results indicated that the income-quantity elasticities varied by size of family and that the low-medium income group had a higher income-quantity elasticity than did the medium-high income group. 1116 results of this analysis are presented in Table 3. west used Michigan State University Corsuner Panel data for the period 1951-55 to compute arc income-expenditure elasticities for broad food groups, including all meats:3 Included in the all meats category were meat, poultry, fish ani eggs. Families were classified into 3 per J’fhomas Neil Moss, "Some Relationships of Selected Socio—Economic Factors to Food Consumption and Expenditures, Lansing, Spring 1950," (unpublished Ph.D. thesis, Department of Agricultural Economics, Iflchigan State University, 1952). zHarold M. Riley, "Some Measurements of Consumer Demand for Meats ,“ (unpublished Ph.D. thesis, Department of Agricultural Economics, Michigan State University, 1951;) . 3Jerry Glenn West, "Estimates of Income Elasticity From Consumer Panel Data,“ (unpublished H1.D. thesis, Department of Agricultural Economics, Michigan State University, 1958) . n“ TABLE3 Imam-QUANTITY ELASTICITII'B FOR MEAT ITEMS BY SIZE OF mar, LANSING, SPRING 1950a A Elasticity Between Income Groups Kind of Meat Size of Family Low-Medium Medium-High Beef 1‘2 068 .20 3'14 01.1.1 016 5 or more .28 .06 Pork 1‘2 01.18 -0 3-14 008 .00 5 or more -.16 _ .15 Beef Steak ' l~2 .81 .52 B-h .52 .30 5 or more .80 .115 a‘Source: Harold M. Riley, "Some Measurements of Consumer Demand for Meats ," (unpublished Bl.D. thesis, Department of Agricultural Economics, Michigan State University, 19511), p. 51. capita income groups: less than $ll00, $1100 to $1900, and more than $1900. Income elasticities obtained for all meats corresponding to the above income classification were .07, .21 and -.03. West attributes the negative elasticity of -.03 for the more than $1900 group as the result of families having increased their expenditures for meals away from home rather than their expenditures for food served at home. A E951: reasoning irriicated that the elasticities for all meat were lower than if the elasticities had been computed for red meats alone because the income elasticities for poultry, fish, ani eggs are probably lower than the elasticities for red meats. West's study was different from earlier research in that expenditures by a particular group of 15 families were examined before arxi after increases in family income. Khan these income elasticities are used for prediction purposes, there is no need for the assumption used in conventional cross-sectional. analysis. Vest examined the differences in per meal expenditures for three income groups, when families were classified according to family income. Statistical analysis indicated that there were no significant differences in per meal per person expenditures between income groups even at the fifty per cent level. Similar results were obtained when computing the per meal expenditures per capita per family year for the three per capita income groups. The analysis used by West to examine expeniitures for food groups before and after changes in per capita income is essentially the same method used in this study to determine the validity of the assumption implicit in predicting the effects of increased income on consumer meat expeIfli'bureSo Summary of Previous Studies Results from cross-sectional analyses have generally iniicated that a direct relationship existed between family income and expeniitures for most meat items. When family income increased, it was assumed for prediction purposes, that families receiving increases in income adjust their expenditures so that they are similar to the expenditures of the income group the families are entering. The primary purpose of this thesis was to test the validity of this aSsumption. West observed family meat expenditures before and after increases in income. Estimates of income elasticity obtained reflected the actual changes consumer panel families made in their expenditure patterns when their income increased. These estimates of consumer response in adjust- ing their meat purchases after receiving an increased income were much lower than the estimates which have been derived from other cross- sectional studies. These results tended to indicate that families did not adjust their meat expenditures as much as was predicted by cross- sectional elasticities. The study being uniertaken will determine if the above results are true for selected meat items. Usefulness of Results Some of the uses which might be made of the resu1ts of this study are as follows: 1) Improve the estimates of consumer respomi‘veness to imreases in real income ani thus the accuracy of forecasts on the national level. These results will also be useful in analyzing the expected effects or Government food subsidy programs; 2) Provide infor- mation on the detailed consumption patterns for selected meat cuts by income groups mich is of considerable value in planning, marketing, promotionfl, and merchandising programs, and 3) The results obtained will add to the funiamental knowledge concerningthe validity of the implicit assumption used in conventional cross-sectional analyses for prediction purposes for specific cuts of meat. CHAPTER II METHODS OF ANALYSIS The General Approach Comumer panel families who experienced an increase in income between two consecutive years were included in the study. These families were aggregated into two per capita income groups.) Cross- sectional. analysis was used to determine the relationship between income ani meat expenditures. Cross-sectional. income-expenditure elasticities were then computed between these income groups. Families in each income group were regrouped according to size of family. This was done to determine if family size affected the level of meat expenditures. Meat expenditures of the families in the two per capita income groups were "observed" before and during an increase in income. Dynamic income-expeniiture elasticities were computed to determine the responsive- ness of families in adjusting their meat expenditures to an increased income. Families in each income group were then grouped according to family size to determine if this variable affected expenditure adjust- ments associated with increased income. Static cross-sectional ard dynamic imome-experxiiture relationships were compared. This comparison was the basis for accepting, modifying 1? 18 or rejecting the assumption underlying the use of conventional cross- sectional relationships for prediction purposes. Source and Nature of Data The Michigan State University Consumer Panel consisted of approxi- mately 250 families who kept weekly food ptirchase diaries which showed the price, quantity am total expenditures for each food item purchased. During the panel's existence, there was a constant flow of families entering and leaving the panel. The primary reasom families left the panel were: moving out of town ard not enough time to cooperate:L An attempt was made to obtain and maintain an unbiased probability sample. The original sample was obtained by conducting a sample survey of 1885 families in the Lansing area. From this group a sub-sample of 300 families was selected on the basis of; income of the household, number in the household, age of the housewife, ani education of the housewife.2 As panel families dropped out, new members were recruited from the list of families in the sample survey. The old family was replaced by a new family which had similar characteristics. Provisions were also made for selecting a proportionate number of newly formed 3 families so as to maintain a representative sample over time. 1Shaffer ani Quackenbush, pp. _c_:_i_._t_., p. 91. 2J. D. Shaffer, "Methodological Bases for the Operation of a Consumer Purchase Panel," (unpublished H1.D. thesis, Michigan State University, 1952). 31bid., "A Plan for Sampling a Changing Population Over Time,“ Journal of Farm Economics, Vol. 36, No. 1, 195b,, p. 156. _ 19 The weekly food purchase diary listed approximately 1,000 food items. These food items were categorized into product groups. Meat items were categorized into six groups: beef, pork, lamb-mutton, veal, other meat and meat mixtures, and poultry. Size of family and income were reported weekly as well as the number of meals eaten away from home. At the end of the year, a questionnaire'was sent to each participating family asking several questions, one of which concerned the income received during the year. The summation of income reported in the weekly diaries were compared with the income estimate on the eni-of-year questionnaire. If these two figures were comparable, it was assumed that the income figure represented a valid estimate for that family. If the two income esti- mates did not coincide with each other, an attempt was made to reconcile the differences by interviewing the family. Selection of Families Michigan State University Panel families who met the following criteria were included in the sub-sample utilized in this study. 1) Families who participated in the panel for two consecutive years and experienced an increase in income. ,2) Families who maintained a constant size or decreasing size during the period 1953-57. This selection was done to eliminate fluctuations in family size due to birth of children which resulted in lower per capita meat expenditures for that family, even though members of this family consumed the same or greater quanti- ties of meat than before the increase in family size. 3) Families of 20 two or more persons were included in the sub-sample. In previous studies, it was noticed that one-person families were primarily responsible for the extreme variability of expeniiture data. These families were not included in the study in an attempt to reduce the variability of expenditures for different income groups. 1;) Families who sent in at least 140 diaries per year. These diaries were not required to be sent in consecutively. For those families who partici- pated less than 52 weeks, an adjustment of their expenditures was made to make these expenditures comparable with those families who partici- pated for 52 weeks. This adjustment brought all expeniitures to the sane level. A family~ observation was considered to consist of a family who participated in the Michigan State University Consumer panel for two comecutive years and who met the above criteria. Family meat expenditures included only meat consumed at home. Expenditures for meals eaten away from home were excluded as participants in the consumer panel only reported the price of the meals and not its comems 0 Units of Measurement The measurement of expenditures can be per family, per meal, or per capita. To be comparable per family or per meal expenditures must be reduced to, a per capita basis in order to eliminate the consumption variability due to different family sizes. In this study, income and empe‘niiture data were reduced to a per capita basis even though the per capita method tends to place may of 21 the large families in the low income group while placing small families in the upper income groups. large sized families purchase more meat than do small sized families, but reducing family experriitures to a per capita per family basis, the per capita expenditures of large families were smaller than those of small families. The optimum situ- ation would be to determine the meat expenditures associated with increased income for different family Sizes within income groups. When the sub-sample data were grouped according to family size by income groups, it was four! that there were very few families in several of the fuily size categories. To have an adequate number of observations in each category, families were aggregated into two groups; 2-3 person and h or more persons per family. Income can be measured as gross or disposable income. Disposable income was more significant than gross income because of the influence of size of family and other tax deductable items. In the remainder of this analysis the term M will refer to per capita disposable income unless specified differently. Grouping of Families There were ninety-nine families who met the criteria used in select- ing families for the sub-sample used in this study. These families were divided into two per capita income groups: less thani$l700 and $1700 ani above. 'me less than $1700 and the $1700 and above income groups were referred to in this study as the lgw‘ and ma income groups. 22 The relationship between family characteristics am disposable income are important for correct interpretation in analyzing consumer responsiveness in adjusting their purchasing patterns after experiencing an increase in income. 1319 higier income families are smaller in size 3111 the homemaker was older and had lower levels of education than homanakers in low income families. See Table 14. TABLE )4 FAME! CHARACTERISTICS Br PER CAPITA INCOME (moms, SUB-SAMPLE or NINETY-MNE mums, MICHIGAN STATE UNIVERSITY. consumes mm, 1953-57 Average Average herage Number herage Age of Education of Income Groups of Persons Income Householder Householder (dollars) (years) (years) 1.653 than $1700 h.15 1218.93 140.6 11.9 81700 and above 2.61; 26%.58 h8.6 12.3 Selection of Meat Cuts 'llie meat cuts included in this study were primarily selected by their importance in the families meat purchasing patterns. The limited number of observations on other meat cuts and their small percentage of expenditures places severe limitations on their usefulness. A secondary criteria used in the selection of meat cuts was that low and high income families have different meat cut expeniitures. If the income groups‘ meat expeniitures were similar, the “expected" expenditure adjustments 23 computed from cross-sectional relationships would be close to zero. Since the primary purpose of this study was to determine the validity of the assumption unierlying cross-sectional predictions, it was necessary for the "expected" adjustments to be significantly different from zero. The eight meat cuts studied were: beef roasts and steaks, pork chops, roasts, bacon ani ham, broilers, and cold cuts. Bacon included Canadian bacon while cold cuts excluded weiners and frankfurters. Selection of Time Period An attempt was made to use. a maximum amount of data from the Michigan State University Consumer Panel, but to avoid maj or price fluctuations. Therefore, data were used only for those years where meat prices were relatively stable. It was also deemed desirable to select a period during which the general. price level was relatively stable. If we use money income as one variable and do not introduce the general price level separately, the income expenditure elasticity estimate will be accurate for pre- diction purposes only if we are dealing with a period when both the price level and real income are changing in about the same ratio or if the price level is stable. If, however, there is a large change in the price level and a shift in the relationship occurs between the price level an real income, we cannot expect the previously observed relation- 1 ship between money income and expenditures to remain the same. 1E. J. Working, "Agricultural Demard During Rearmament,“ Journal of Farm Economics, Vol. 3h, No. 2, 1952, p. 210. 214 Price elasticity is defined as the percentage change in the quantity taken associated with a one percentage change in price.1 Previous studies have indicated that many of the meat cuts have price elasticities centering arouni unity. Taeoretically, when a meat item has a price elasticity of one, a one per cent increase or decrease in {rice results in a one per cent increase or decrease in the quantity demanded but expenditures remain constant. For those meat cuts which have a price elasticity around unity, expenditures will not change because of price variations. Those meat cuts which have a price elasticity greater or less than one, expenditures will increase or decrease, depending on the direction of the price change. However, if the price changes are small consumer expenditure adjustments due to price variations are likely to be negligible. Purchasing patterns will not be adjusted unless the magnitude of the price change will insure a significant increase in satisfaction derived from a different bundle of goods. Thus, a fandlies' purchases of a given item will vary according to the magnitude of thehprice change. Bilkey generalized that the smaller the price change, the smaller the probability of families adjusting their purchasing patterns and conversely, the larger the price change, the greater the probability of families 2 . adjusting their purchasing patterns. Thus we would expect that lstigler, pp. _<_:_:_i._t., p. 32. aWarren J. Bilkey, The Basic Relationshig In Consumer Expenditure Bflavior, Harvard Studies in Marketing Farm Products, Number L—H, Cambridge, 1951, p. 39. 25 sub-sample families made few adjustments in their purchasing patterns due to price variations. The Michigan State University Consumer Panel operated from February 1951 througx December 1958 but significant price variations occurred during this period which limited the years to be used in this study. Data for 1958 was unavailable. Table 5 indicates that during the period 1951-52 beef roast and steak prices were higher than during the period 1953-57. Steak ani roast prices decreased during the 1952-53 period by 114.5 and 23 .0 per cent respectively. During this period the supply of slaugiter cattle increased, depressing beef prices. The period 1953-57 exhibited relative price stability for these two‘beef cuts. As stated earlier, the prices of the meat cuts analyzed must be stable in order to minimize the effects of price on the purchasing patterm of consumer. For this reason, the period 1951-52 was dropped from the analysis. Pork prices were more variable than beef prices during the period 1953-57. See Table 5. The variability of pork prices was primarily due to the differences in the production cycles of pork and beef. For cattle this cycle is 114-16 years, whereas the production cycle for h0g3 is 3-5 years. The production of hogs reached its peak in the last part of 1953 and the-first part of 195).; and 1957 which depressed pork prices during these periods. Production of hogs decreased during the period 1955-56 which stimulated pork prices. 1 The price of cold cuts was relatively stable as shown in Table 5. During the years 1953-57, the average annual price of cold cuts varied TmES AWGE ANNIE PRICES FOR SELECTED MEAT CUTS, MICHIGAN STATE UNIVERSITY COBBUMER PANEIL, 1951-1957 26 Meat CutL Price Per Pound Beef Cuts r Pork Cuts Year Roasts Percentage Steak Percentage Bacon Percentage (cents) Change (cents) Change (cents) Change 1951 .77 _3.9 .95 _5.3 .56 ~s.t 1952 0714 __23 .0 090 “114.5 053 214.5 1953 057 _5 .3 077 _3 .9 066 “3.0 19514 .51. -s.6 .71. M .68 4.7 1955 OS]- 2 .0 '73 0 ,0 '56 -8 .9 1956 '52 7 .7 ‘73 8.2 051 27 .5 1957 ~56 , .79 .65 Meat CutL Price Per Pourri J Pork Cuts Chops Percentage Ham Percentage Roasts Percentage (cents) Change (cents L Change (centsL ChangL 1951 ’71 o .o ‘66 -3 .o ‘57 -5 .3 1.952 .71 ' 805 06).}. 708 0514 5.6 1953 .77 0,0 .69 17,3 57 ~1.8 1951‘ '77 ~10.u '72 -8 .3 ‘56 ~17 .9 1955 .69 0.0 .66 0.0 .h6 _2.2 1956 '69 8.7 '66 -1.5 ‘1‘5 13 .3 1957 .75 g .65 .51 Meat CutL Price Per Pound Other Meats TPoulfitIy Cold Cuts Percentage Broilers Percentage (cents) Change _ (cents) ‘ Change 1951 066 1.5 057 _1.8 1952 067 ‘14 .5 ’56 -1.8 1953 ’a‘ -l.6 ‘55 -18 .2 1951‘ .63 “1.6 .145 1.3 03 1955 062 -3 .2 '51 ‘15 .7 1956 .60 6.7 $3 -u.7 1957 .6h .hl 27 only by 7 cents with the largest percentage change in prices between years being 6.7 per cent. lhe price of broilers and fryers exhibited considerable fluctuations, generally downward, during the period 1953-57. This decrease in price was primarily due to the influence of increasing size and efficiency of the broiler iIflustry. Even though there were fluctuations in the prices of the meat cuts discussed, the magnitude of the price change was such that they probably did not greatly influence consumers expenditure patterns for beef or cold cuts. The variation in pork prices may have influenced consumer pork ecxpeniiture patterns. Income-expenditure relationships were definitely affected by the decreasing price of broilers during the period 1953-57. Income expenditure elasticities computed for broilers repre- sent the responses of consumers to both a price and income change. The general price level was relatively stable during the period 1953-57. The Consumer Price Iniex, (United States l9h7'149 . 100) in- creased only 5.77 per cent during the period, an average increase of 1.15 per cent per year. The Consumers Price Index was unusually stable from 1953-56 but during 1957 the iniex increased by 3.914 per cent. See Figure I. Inflationary pressures affected the prices of food eaten at home less than all items purchased by consumers. Food prices during this period remained relatively constant primarily because of the plenti- ful. food supply. The inflation that occurred was experienced in the durable good segment of the economy. 28 . Goa .. $-53 emummma :33..qu .83 no 369m .33 so Eastman .26m 3 needs tees its a5 .38 its .803 a use .525 more gas Steam Bean: .H enemas .5on p 3” wmmd mema 4mm...” mmmd , i l a _ t I...) d. \“"" ’ ' I \\ ““ "'-"""--- l \ ““ \ Il \‘ 1III|IIIII| a J IIBH 030m a.“ flock dwubfi I --| HH Bemetalllll ”mg 003 3588 29 The ninety-nine family sub-sample from the Consumer Panel experi- enced an average increase in disposable income of approximately 13 per cent. Of this 13 per cent increase in income, approximately 6 per cent was due to inflationary pressures and the remaining 7 per cent repre- sented the actual amount of increased purchasing power in the hands of consumers. The upward trend in per capita income over time exhibited in Table 6 was partly the result of selecting only those families who experienced an increase in income. TABLE 6 PER CAPITA DISPosABLE IECMECP man-mm: FAMILY SUB-SAMPLE, MICHIGAN STATE UMVERSITY. CONSUMER PANEL 1953-57 Year , ' Per Capita Income (dollars) 1953 1802.07 195k 1826.67 1955 1962.32 1956 19h5.81 1957 2203.08 The large increase in per capita income in 1957 as compared to 1956 was primarily due to the large proportion of high income families in the sub-sample. were were very few low income families in 1957 who received an increase in income. 30 Preliminary Investigation Adjustments in family beef steak expenditures associated with increases in income during the period 1953-57 are shown.in Figure 11. The apparent variability shown.in the scatter diagram indicated that any attempts to forecast individual family responses to an increased income would result in a large error of estimate. Plotting changes in beef roast expeniitures against income changes produced similar results. It was hypcthesized that there would be a greater stability of expenditure-income relationships when.dea1ing'with averages for groups rather than with individual families. Consequently, a decision was made to compute arc elasticity estimates to measure the responses of consumer groups to increased income. in arc elasticity computed betwaen two income groups represent an average responsiveness of consumers to increased income. This responsiveness is assumed to be similar over the entire range for'whiCh the elasticity is computed. This assumption should be valid in this study since the elasticities were computed over a relatively short segment of the income range. Methodology Static Cross-sectional Relationships Al‘c elasticities were computed to determine the income-expenditure relationship between the two income groups. The income data used in cross-sectional computatiom are shown in Table 7. (LEI-6‘5)" "unfiuht N h ~ I. . 31 . Rimmed .oadscmubsm Hoses 8588 Bauhaus obese. couscous enoncoaofin ofiouaH “at 8930034 assessing Macaw Noam 33.8 Mom 5” 3838 .H 0% season Hem 0835.. 5” cassava” 33 8: com com ooh” 02. com com 1111—1le 4 J1 i — NH 8“ cannon , chose-owes obsessed teem econ 32 TABLE 7 AVERAGE ANNUAL PER CAPITA INJOME, BY INCOME GROUPS SUBDIVIDED BI FAMILY SIZE, NINETY-NINE FAMILY SUB-SAMPLE, NICHICAN STATE UNIVERSITY CONSUMER PANEL, 1953-57 Number of Income Groups Family Years Per Capita Income (dollars) Less than $1700 119 13114.25 2‘3 person 33 131414-53 h or more 86 1302.62 $1700 and above 95 28114.90 2-3 person 83 2902 .118 11 or more 12 2209 .57 These income data were obtained by adding ani averaging family income before and after the family received an increase in income. The results in Table 7 showed that an inverse relationship existed between per capita income and family size. Cross-sectional income-expenditure elasticities were computed by this formula: EB - EA IB + IA EB "' EA In " IA where EA - herage expenditures for income group A 33 . Average expenditures for income group B IA 1- Average income for income group A IB . Average income for income group B 33 The estimates derived from these computations will be referred to in the course of this study as E33332 income-expenditure elasticities. Income groups were divided into two family size categories: 2-3 person and 1; or more person families. This breakdown of families within income groups indicated if there was any relationship between families who differed in size but who had similar per capita incomes. The income data for different family sizes are seen in Table 7. A research study of the corsumer panel relating consumption of meat to various factors concluded that meat consumption per person was more closely associated with size of family than an other characteristic considered, but the effect was due partly to the accompanying inverse relationship between size of family anl per capita income.1 If there was a significant relationship between family size and meat expenditures, this would sug- gest that cross-sectional elasticities used to predict the responses of consumers to increased income would yield inaccurate estimates if the two populationsiwere not homogeneous with respect to size. Student's "t" test was used to determine if the level of experrlitures were signifiCantly different between families differing in size within income groups. The level of expenditures between two income groups were also tested for Significance. The formula used was:2 t .- isH " EA ‘ 8M lshaffer, Quackenbush, and Moss, pp. 33.3., p. 12. 2Wilfred J. Dixon and Frank J. Massey, Jr. , Introduction to Statistical Anal. is, (2d. ed. , New York: McGraw-Hill Book Company, 11”., 19 7 , pl 11-7. ‘ 311 where ii It Mean expenditure for income group or family size A. EB - Mean expenditure for income group or family size B. SAB I Pooled variance for the two income groups or family sizes. Dynamic Relationships Family meat expenditures were "observed" before and after families received an increase in income. Table 8 summarizes the income data for income groups ani family sizes within income groups. TABLEB Am PER CAPITA IICCME HEPCRE AND AFJER AN INCOME IIDREASE, BY INCCME GROUPS SUBDIVIDED BY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CCNSUMER PANEI. SUB-SAMEE, 1953-57 W * ** Average Income Groups by Number of T1 T2 Increase in Family Sizes Family Years (dollars) (dollars) Income (percent) Less than $1700 1l9 1227 .314 11401.15 11.2 2-3 person 33 1238.70 11450.36 17.1 14 or more 86 1222 .98 1382.26 13.0 81700 arrl above 95 2677 .146 2952 .314 10.3 2-3 person 83 2772.31 3032.614 9.14 14 or more 12 2021.142 2396.92 18.6 *:T1 :- Per capita income before the income increase. Although there were only ninety-nine families in the sub—sample, T2 - Per capita income after the income increase. many families participated at least four years, supplying three obserVations per family. The number of observations per family for 35 the ninety-nine family sub—sample account for the 2114 family-years included in the above table. Let us assume that a family experiences an increase in income each year during the period 1953-57. Thus, there would be four family-years for this family. The family year included T1 which represent the families' 1953 income, and T2 representing the 19514 income. Since we assumed angincrease in income, the next family-year included T1 the 19514 income and T3 the 1955 income, etc. The per cent change in income was relatively small for each income group and family size within income groups. The aSSUmption of average elasticity throughout the income change was a. realistic assUmption to make in this analysis. Meat expenditure adjustments associated with increased income were observed for the W0 income groups. Income expenditure elasticities were then computed for each income group by the following forlmlla: EAl-EAJ. . IA2+IQJ where E1 I- Expenditures for income group A in time period 1. E2 . Expenditures for income group A in time period 2. I1 I Income for income group A in time period 1. I; II Income for income group A in time period 2. These income elasticities will be known in this study as the Mg income expeIfiiture elasticities . Sub-sample data were then classified into family sizes within income groups. 'Jllis was done to determine if family Size affected the 36 families meat expenditure adjustments after receiving an increase in income. If there was a significant relationship between family size anl experditure adjustments this would suggest that elasticity estimates used in predicting responses of consumers to increased income will result in an inaccurate estimate if the two populations were not homo- geneous in respect to family size. The changes in expenditures associated with an increased income were tested by the Student "t" test to determine if expeniiture changes were significantly different from zero for income groups and for both 1 family sizes. The formula used was: d-o S/Ffi' ’0: where d is the difference between the mean expenditures before and after an income change; S is the variance of the differences in mean expenditures between families. Comparison of the Static and Dynamic Income-Expenditure Relationships The validity of the assumption that families experiencing an in- crease in income will adjust their purchasing patterns to be similar to the purchasing patterns of the economic group the families are entering was tested by comparing the "observed" and "expected“ meat expediture adjustments. 1Ibid., p. 121. 3? "Observed" expenditure adjustments were derived by: Et ‘ Et-l “Eercted” expenditure adjustments were derived by: (It It-l It- 1 1 (Et-1) where Et - Expenditures for item in time period 1. Et-1.' Expenditure for item in time period t-l. - Static income—expenditure elasticity for item. It I Per capita income in time period 1. It-l." Per capita income in time period 2. The difference in Et - E%_l'was the "observed“ expenditure adjustments made by consumer after experiencing an increaSe in disposable income. ‘ It-l It—l "expected" from an increase in income of“ multiplied by’ yields the percentage change in expenditures R-a- It—l centage change in expenditures multiplied by'!% gives the "expected" It - I t-l changes in expenditures associated withw--i;-i- increase in income. . The nexpected“ per- In this way it was possible to determine if the "observed" expenditure Changes were in the direction and magnitude of the "expected“ expenditure adjustments. The validity of the assumption when the usual cross-sectional analyses are uSed for prediction purposes was accepted if the "expected" and “observed“ expenditure adjustments were alike. However, if consumers either under or overadjusted their meat item expenditures, the validity of the assumptinn as stated would be critically questioned or rejected. 38 If the increased income was primarily due to inflationary pressures, the income-expenditure elasticities computed from consumer responses to increased income would be approximately one. Therefore, using income-expeniiture data for the period 1953-57, where approximately half of the income increase experienced by consumers was "inflationary,“ dymmic income expenditure elasticities were affected. If the dynmfic elasticities computed from expenditure changes resulting from both real and inflationary increases in income was greater than one, the effects of the inflationary increase would tend to reduce the dynamic elasticity. However, most of the real income dynamic elasticities for, meats are probably less than one. Thus, when inflationary income is averaged with real income in computing dynamic income-expeniiture elasticities, the result is to increase the "observed" dynamic elasticity. Simona pointed out that there was a definite association between the rate of change in per capita income and the change in per capita e:xpemi:].tu:r'es.:L Tue results indicated that families receiving small . increases in income did not adjust their expenditures as much as those families who received large increases in disposable income. Families do not adjust their meat cut purchasing patterns immediately after receiving an increase in income. This analysis only covered a period of one year in which family expenditures were "observed." Therefore it is possible that family expenditure adjustments occurring 1Simmons, _o_p. git” p. 3h. 39 at a later date were not included in the "observed“ expenditure adjust- ment. Inflationary increases in income, legs in expenditure adjustments 8111 the magnitude of increased income may influence the ultimate decision in accepting, modifying, or rejecting the validity of the assumption used for prediction purposes in conventional cross-sectional analyses. Limitations of Analysis The method used in checking family income assures that families are placed in the appropriate income group. If families reported an income which was erroneous, their expenditure data was placed in a ‘ “wrong" income group which affected the final results. However, it was assumed that all families were placed in the appropriate income group. Other variables which probably influenced changes in expeniitures associatedwith increased income and not examined in this study are: age, occupation, religious beliefs, credit, savings, temporary or permanent increases in income, education of the housewife, and meals eaten away from home. The Consumer Panel sub-sample data was first divided into four imome groups. The results obtained from this grouping did not conform to the results of the 1955 United States Department of Agriculture Household Food Survey. This was especially evident for pork chops, roasts and bacon experfiitures where these expenditures for the medium income group were lower than expenditures for the low income group. [to The results of the food survey indicated that the medium income families had larger meat item expenditures than the low income families. The findings of the consumer sub-sample data grouped into four income groups can be seen in Appendix A. The static cross-sectional income-expenditure elasticities computed from these data were unreasonable. 'Hhen static elasticities computed between the low and medium income groups were used for prediction pur- poses, the low income families experiencing increases in income were expected to decrease their expenditures for pork chops, roasts, and bacon. Previous studies have shown.that low income families increased their pork item expenditures after experiencing an increase in income. Meat item expenditures for different family sizeS'within.income groups showed no consistent relationships. For many of the meat items, large families evidenced larger per capita expenditures than.small families. Both Shaffer and Simmons found that small families had higher per capita expenditure leVels for selected meat items than.did large families. In an attempt to produce results which were internally consistent, the low ani the medium income groups and the medium-high and high income groups were aggregated into two income groups. This was done to increase the number of observations in the remaining income greup and family size categories. .._..__._—. Ir I ‘my CHAPTER III CHANGES IN BEEF CUT EXPENDITURES ASSOCIATED KPH-l IMREASED II‘COME Introduction The purpose of this chapter was to provide information on consumer respomes in adjusting their per capita expenditures for beef roasts and steaks after experiencing an increase in income. Static ani dynamic income elasticities were compared to determine the vflidity of the assumption that families experiencing increases in income will adjust their purchasing patterns so that they are similar to the purchasing patterns of the income group the families are entering. The results of the analyses will be presented in the following order: 1) Static cross-sectional relationships, 2) Dynamic income- expeniiture relationships, and 3) Comparison of the “observed" and "expected" changes in beef roast am steak experriitures. Results From a Static Cross-Sectional Analysis of Beef Roast and Steak Expenditures The high income group had larger per person expenditures for beef roasts and steaks than did the low income group. See Table 8. The level of expenditures for both roasts arrl steak for the two income groups were significantly different at the 5 per cent level. Both income groups spent more for beef steaks than roasts. hl r TABLE Ba PER CAPITA mmmITUBES FOR SELECTED BEEF CUTS, BY INCOME GROUPS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SW, . 1953-57 .Nmnber of Average Annual Per Capita Expenditure Income Group Families Roast Steak (dollars) (dollars) Less than $1700 119 6.29 7.18 $1700 and above 95 10.114 12.30 The static income-expenditure elasticities computed for beef roasts and steaks were .66 ani .714. I The results of the 1955 Food Consumption Survey irfiicated that families in the $5,000 and above income bracket had higher expenditure levels for beef roasts am steaks than families in the less than $14,999 income group. See Table 9. TABLE9 AVERAGE LEVEL or FAMILY EXPENIITUBBS FOR smEGTED BEEF CUTS, ALL URBAN FAMIIIB IN THE, NORTH CENTRAL REGION, AHEEK IN ABEL-JUNE, 1955‘=1 ._ Disposable I ome Roast Steak Per Household Expenditures Expenditures (dollars) (dollars) L638 than $14,999 072 .90 $5,000 ani above 1.21 1.77 8Source: "Food Consumption of Households in the North Central Region," Report No. 3, Washington, D. 0.: U. S. D. L, 1955, pp. 66-67. bOne person families were excluded. )43 Families in both income groups had larger beef steak expenditures than roast expeniitures. Static cross-sectional income-expenditure elasticities computed between the means of the two income groups indi- cated that low income families receiving. increases in income were “expected" to make larger steak than roast expenditure adjustments. Thus far the results of a cross-sectional analysis for beef roasts and steak expeniitures were similar to the findings of the 1955 Household Food Consumption Survey. These similarities were good indicators that an adequate number of observations were obtained in each income group. then the Consumer Panel sub-sample was divided into )4 income groups, beef cut expenditures were not similar to the results of the 1955 Household Food Consumption Study. See Appeniix A. Hhen income groups were subdivided by family size, it was found that small families in both income groups purchased more roasts and steaks per person than did the large families, with one exception. The high income group large-sized families had greater per capita purchases of roasts than did the small families in the same income group. See Table 10. However, the level of expenditures for the family sizes within the high income group were not significantly different from each other at the 5 per cent level. It is interesting to note that the level of per capita expenditures for roasts and steaks by size of family in the low income group tested to be significantly different at the one per cent level, whereas in the high income group, the level of per capita beef cut expenditures for the two family sizes were not significantly different at the 5 per cent level. f—mms'jl“ .m- ran‘ TABLE 10 PER CAPITA EXPENDITURES FOR SELECTED BEEF CUTS, BI INCOME GROUPS SUBDIVIDED BI mar SIZE, MICHIGAN STATE UNIVERSITI COLBUMER PANEL SUB-SAMPLE, 1953-57 Income Groups by Number of Average Annual Per Capita Expezfliture Family Size Families Roast Steak (dollars) (dollars) Less than $1700 2-3 person 33 7.91 8.83 L )1 or more person 86 5.68 6.55 i $1700 and above g 2-3 person 83 10.01; 12.63 h or more person 12 10.37 9.57 L Earlier studies indicated that the small-sized families per person expeniitures for these meat items were greater than the per person experditures of large families. The above analysis indicated that family size was an important factor influencing the level of expenditures for families in the low income group, but family Size as a factor influencing the level of expenditures was statistically non-significant in the high income group. These results have important implications in deriving and comparing static income elasticities from different studies. As a result, income elasticities used to predict consumer responses to an increased income will result in an inaccurate estimate if the populations were not homogeneous with respect to family size. hS Reaults from a Dynamic Analysis of Beef Roast and Steak Expenditures Table ll shows that families in the low income group increased their roast expenditures by 30 cents after experiencing an increase in per capita disposable income. Families in the high income group decreased their roast expenditures by 614 cents. This expenditure decrease was significantly different from zero at the 10 per cent level. Both income groups increased their per capita steak expenditures after experiencing an increase in disposable income. However, these ex— peniiture changes were significantly different from zero only at the ho (per cent level. Consumers in the two income groups adjusted their steak expeniitures similarly after receiving increases in income. See Table 11. TABLE 11 ANNUAL AVERAGE PER CAPITA‘BEEP CUT EXPENDITURES ASSOCIATED WITH INCREASED IMOME, BY INJOI’E GROUPS, MICHIGAN _ STATE UNIVEISITY COLBUMER PANEL SUB-SAMPLE, 1953-57 Number .of a b Change Level Of d Drone Groups Families E1 E2 in E c Significance Roast Less than $1700 119 6.11; 6.1m .30 .20 Steak Less than $1700 95 7.ll 7.21; .13 .hO $1700 and above 12.20 12.39 .19 .hO SE1 It Expenditures before an income increase. cE2 :- Expenditures after an income increase. Change inE-Ez- 133.;l 6 critical level of significance was at the 10 per cent level. he The estimated dynamic income-expenditure elasticities for beef roast fer the low and high income group were .36 and -.61 respectively. The dynamic elasticity for beef steaks for the low and high income groups were .114 and .16 respectively. The static income-expenditure elasticity for steaks and roasts indicated that low income families adjusted their steak expenditures more than roasts expenditures after receiving increases in per capita income. However, the dynamic income ..t-.. "Sir Ifi‘: 17'1—1 elasticity for the low income families showed that these families responied to an increased income by adjusting their roast expenditures more than their steak expenditures. In the next section, dynamic and F'— static elasticities were converted to an expenditure basis and subjected to further analysis. The high income small-sized families decreased their roast expendi- tures by .91 cents after experiencing an increase in disposable income. This group was the only group which decreased their expenditures for either roasts or steaks. The experditure decrease of 91 cents for roasts was significantly different from zero at the 5 per cent level. See Table 12. The large families in both income groups increased their roast and steak. expenditures. However, these changes were not significantly different from zero. Static income expeniiture elasticities derived from previous studies indicated that low-income small-sized families were expected to be more responsive in adjusting their meat expenditures after receiving an ANNUAL AVERAGE PER CAPITA BEER CUT EXPENDITURES ASSOCIATED WITH INCREASED INCOME, BY INCOME GROUPS SUBDIVIDED BI FAMILY SIZE, TABLE 12 MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMILE, LL? 1953-57 Income Groups by Number of a b Change Level of Family Size Families E1 E2 in E 0 Significance .j Roast Leas than $1700 A 2-3 person 33 7.15 8.37 .92 .20 .2 h or more person 86 5.61; 5.71 .07 > .hO i More than $1700 ‘ g 2-3 person 83 10.149 9.58 -.91 .05 "H- h or more person 12 10.09 10.65 .56 .hO Steak Less than $1700 . 2-3 person 33 8.71 8.95 .2D to h or more person 86 6.52 6.58 .06 > .hO More than $1700 '2-3 person 83 12.62 12.614 ' .02 > .hO h or more person 12 9.2).; 9.79 ' .55 .hO aEl . Beef cut expenditures before an income change. C3E2 :- Beef cut expenditures after an income change. dChange in E II E2 - E1. Tue critical level of significance was at the 10 per cent level. increase in income than were the large size families. ‘ Low income small families made larger roast and steak purchases than did” the low income large families. See Table 12. Thus the results from a static cross- sectional and dynamicanalyses were similar. LIB COmparison of the Observed ani Expected Changes in Beef Roast and Steak Expenditures Table 13 shows that the “observed" expenditure adjustments fer the sub-sample of families in the low income group were in the direction of the "expected" adjustments computed by the static income-expenditure relationships for beef roasts and steaks. TABLE13 OBSERVED All) EXPECTED BEEF ROAST AND STEAK ADJUSTMENTS ASSOCIATED WITH IN’JREASED IMOME, LOW IMOME (BOUP, MICHIGAN STATE . ‘ UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-57 . Beef cut Changes in Per Capita Expenditures Expected Observed (dollars) _ (dollars) « Roast _ .57 ' .30 Steak 07,-1- ‘ 013 The lowincome families did not adjust their beef cut expenditures by as much as was "expected." In Chapter II, it was pointed out that the low income group was composed of young families. These families were larger than families in the higher income group. In essence, when a static income elasticity is computed between the low and high income groups, the resulting elasticity indicates that the larger younger families in the low income group receiving increases in income will assume purchasing patterns of the higher income group whose families are older and smaller. I" 19 Thus, the static elasticities are computed between two different populations with respect to size and age of the homemaker. The high income families spent approximately 1.75 times as much for beef steaks ani roasts than did the low income families. Therefore the “expected“ adjustments computed from cross-sectional relationships tended to over- estimate consumer low income families “observed" expenditure adjustments. F— Families experiencing an increase in income do not immediately assume new patterns of Icorsumption. Estimates from dynamic analysis include not only immediate reactions of consumer to an increase in income but also a past period lag reaction. This lag reaction was . in. probably due to changes in income which occurred in earlier years, different family characteristics, increased experience of the homemaker, retirement of the house holder, or purchasing power over and above current income. 8 Low income families experienced a 114 per cent increase in income during the period 1953-57. Six per cent of this increase was attributed to inflationary pressures. These inflationary increases in income tend to make the "observed" expenditure adjustments larger than if the increase in income consisted of an all real income increase. The evidence presented suggests that low income families receiving increases in income do adjust their beef item expenditures in the direction of the higher income group, but lag reactions, and the family characteristics between the two income groups result in low income families underadjusting their beef cut expenditures. 50 Therefore, the assumption.that consumers experiencing increases in income will adjust their beef out purchasing patterns similar to the income group the families are entering does not appear to be valid. #7“ CHAPTER IV CHANGES IN SELECTED PORK CUT EXPENDITURES ASSOCIATED WITH IBEREASED INJOME Introduction The objective of this chapter was to derive static cross-sectional and dynamic income-expenditure relationships for four pork items 3 bacon, chops, ham and roasts. A further analysis will determine the effects of size of family on changes in expenditures for these pork cuts ani the resulting income-expenditure elasticities . Waite arfl Trelogan derived a number of cross-sectional income- quantity elasticities for selected meat item using data obtained in the 19148 Food Corsumption Survey:L In comparing the income-quantity elasticities for beef and pork items, it was generally found that con- sumers experiencing an increase in income increased their quantity of purchases of beef items more than pork items. This evidence tends to indicate that the observed expenditure adjustments for bacon, chops, ham, am roasts associated with increased income would probably be smaller'than the expenditure adjustments for beef roasts andsteaks. The results of the analysis, as in the last chapter, will be pre— sented in the following order: 1) Static cross-sectional relationships, Waite ani Trelogan, _9_p. 33.33., p. 141. 51 52 2) Dynamic income-expenditure relationships, and 3) Comparison of the "observed" and "expected" expenditure adjustments for selected pork cuts. Results from a Static Cross-Sectional Analysis of Selected Pork Item Expenditures Families in the high income group had larger per capita expenditures for the four pork items than did the low income families. See Table 111. 'lhe level of pork cut expenditures for the two income groups were all significantly different at the 5 per cent level. [mm mafimf“: : : Ifii’] TABLE 11; PER CAPITA EXPENDITURES FOR SELECTED PORK CUTS, BY INCOME GROUPS , MICHIGAN STATE UNIVERSITY CODBUMER PANEL SUB-SAMPLE, 1953-57 Nmnber of Avergge Ammal Per Camita mmrflitures Income Group Families Bacon Chops Roasts ~ (dollars) (dollars) (dollars) (dollars) Less than $1700 119 . -h.59 3.12 14.72 2.11 151700 and above 95 5.1m 5.25 . 8.11 3.01 The static income-expenditure elasticities for bacon, chops, ham, ani roasts were .2142, .715, .7112, and .h9h respectively. Table 15 shows the expenditure levels for three pork cuts by income groups obtained by the 1955 United States Department of Agriculture Food Consumption Study. Pork roasts were not reported as a separate category. 53 TABLE 15 AVERAGE LEVEL OF FAMILY EXPENDITURES FOR SELECTED PORK CUTS, ALL URBAN FAMILIES IN 'IHE NORTH CENTRAL. REGION, A WEEK IN APRIL-JUMP. 1955al Disposable Income Bacon Chop Ham per Householdb Expenditures Expenditures Expenditures , ' (dollars) (dollars) (dollars) EF— Less than $h,999 .141 .51 .68 $5,000 and above .57 .58 ‘ .76 .aFood Consumption of Households in the North Central Region, Rept. No. 3, Washington, D. 0., U.S.D.A., 1955, pp. 68-69. One-person families were excluded. [a Families in the $5,000 and above income bracket had greater expeniitures for bacon, chops, and ham than did the families in the less than $11,999 income group. Both of these income groups had larger ham expenditures than bacon and chop expenditures. Although the results of the Food Consumption study and the present analysis are similar, it must be remembered that the former is based on a family income basis while the latter is classified on a per capita basis. The low income small families had larger per capita expenditures for bacon, chops and roasts than did the low income large families. See Table 16. However, large families in the low income group had larger per capita ham expenditures than small families. The, level of ham expeniitures for the two family sizes within the. low income group were significantly different from each other only at the 20 per cent level 0 5h TABLE 16 PER CAPITA EXPENDITURES EORSELECTED PORK CUTS, BY INCOME GROUPS SUBDIVIDED BY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-sums, 1953-S7 W Income Groups by Number of Average Annual Per Capita Expenditures Family Size Families Bacon Chops Ham Roasts (dollars) (dollars) (dollars) (dollars) '5" Less than $1700 2-3 person 33 5 .13 3 .51; h.l9 3 .12 h or more 86 1.1.0314 3002 14.997 1066 $1700 am above 2-3 person 83 5.67 5.61 8.23 2.85 h or more 12 h.97 7.10 7.13 14.15 L._ + The high income small families had larger per capita expenditures for bacon and ham than did the high income larger families. For chops and roasts, the reverse situation occurred; 1.6., the high income large families had the largest per person expenditures. Among the various experditure comparisons between the two family size classifications in the high income group, only the roast expenditures were significantly different at the five per cent level. Shaffer reported that family size was found to be the most important factor in explaining variations in family pork consumption:L In the present study, low income large families tended to have lower per capita pork item expeniitures than the small families in the same income group. There was no such teniency in the high income group. lshaffer, 22. 232., p. 16. 55 Results from a Dynamic Analysis of Selected Pork Item Expenditures Table 17 shows the "observed" per capita pork item expenditure adjustments made by families in the two income groups who experienced an increase in per capita income. TABLE 17 ANNUAL AVERAGE IER CAPITA PORK CUT EXPENDITURES ASSOCIATED RITE INCREASED INCOME, RI INCOME GROUPS, MICHIGAN STATE UNIVERSITI CONSUMER PANEL SUB-Shun, 1953-S7 . Number of a b Change Level of Income Groups Families El E2 in E 0 Significance Bacon Lass than $1700 119 17.65 17.53 -.12 .30 $1700 and above 95 5.37 5.55 .18 .30 GhoE Less than $1700 ll9 3.11; 3.09 -.05 > .110 $1700 and above - 95 5.25 5.25 .00 > .hO in Lees than $1700 * 119 11.59 h.85 .26 .20 $1700 and above 95 8.29 7.9).; -.35 .30 Roast Less than $1700 119 1.97 2.2L; .27 .05 $1700 and above 95 2.86 3.15 .29 .20 afll - Pork item expenditures before increase in income. bE2 - Pork item expenditures after increase in income. 0 e critical level of significance was at the 10 per cent level. The experxiiture adjustments by income groups were not significantly different from zero in 7 of 8 possible observations. The roast expendi- ture adjustment by the less than $1700 incOme group was significantly 56 different from zero at the 5 per cent level. Large per capita vari- ations in both income groups were primarily resporsible for most of the pork item expeniiture adjustments being non-significant. The low and high income groups adjusted their per capita roast expeniitures approximately the same amount after receiving an increase in income. However, these income groups did not adjust their per person chop expenditures after receiving an increased income. The per capita bacon and ham adjustments indicated that low income families decreased their bacon expenditures and increased their ham expenditures after receiving an increase in income. High income families increased their per capita bacon expenditures and decreased their ham expenditures. Later in this analysis, an examination of expenditure adjustments for the two income groups subdivided by family size will determine if any consistent relationship existed between family size and expenditure changes within income groups . Both income groups! dynamic income-experrliture elasticities inii- cated that families increased their porkroast expenditures more than for the other three pork items. See Table 18. The static income—expenditure elasticities between the low and high income groups were positive for the )1 pork items. Families in the low income group were expected to increase their expenditures for these 1.; pork cuts after experiencing an increase in disposable income. The dynamic income-expenditure elasticities computed from the "observed" pork expenditure adjustments associated with increased income showed 57 TABLE 18 DINAMIC ImOME-EIENDITURE ELASTICITIE‘S FOR SELECTED PORK CUTS, BI INCOME GROUBS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL. SUB-SAMPLE, 1953-s7 Per Capita Income Pork Cuts Less thw11700 $1700 and Above Bacon -.9l7 ' .319 0110138 “0012 0000 Han] ohlé “oh—19 Roast .969 -938 that low income families increased their ham and roast per capita expeniitures, did not adjust their chops expeniiture ani decreased their bacon experditures. Low income large families increased their ham ani roast expenditures 3111 left their bacon and chop expenditures unchanges after receiving imreases in income. Small families in this income group tended to decrease their pork out expenditures as illustrated by negative expendi- ture changes for bacon, chops and ham. See Table 19. High income large families increased their per capita bacon ard chop expenditures and left their ham and roasts expeniitures unchanged after receiving an increased income. High income small families increased their per capita bacon and roast expenditures while decreasing their chop and ham expenditures after receiving an increase in dis- posable income. . Family size was an important factor influencing per capita expendi- ture adjustments for pork cuts associated with increased income. ANNUAL AVERAGE PER CAPITA PORK CUT EXPENDITURES ASSOCIATED WITH TABLE 19 INJREASED INCOME, BI INCOME GROUPS SUBDIVIDED BY FAMILY ‘ SIZE, MICHIGAN STATE UNIVERSITY COIiBUMER PANEL SUB-SAMPLE, 1953-57 58 "' E1. E2 - Pork item expenditures after increase in income. Change in E - E2 dThe critical level of significance was at the 10 per cent level. Income Groups by Number of a b Change Level of Family Size Families E1 E2 in E C Significance Bacon Less than $1700 2‘3 person 33 50,48 1 14.098 “012 030 )4 or more 86 h.3h 11.31; .00 > .110 517%" and above 2‘3 wrson 83 5061 5077 018 030 h or more 12 h.60 5.311 .711 .10 Chow Lass than $1700 2-3 person 33 .63 3.116 --.17 .hO h or more 86 .02 3.01 ~.01 > .hO $1700 and above , 2-3 person 83 5.07 11.92 -. .hO h or more 12 6.171 7.79 1.38 .20 _H___am Less than $1700 . 2-3 person 33 h.h2 3.96 - .116 .20 )1 or more 86 11.71 5.22.51 .05 $1700 ani above 2-3 person 83 8.110 . 8.06 -.3h .30 h or more 12 7.1411 7.h2 -.02 > .hO Roasts Less than $1700 2-3 person 33 3.017 3.20 .16 .NO LL or more 86 1.61 1.72 .09 .30 $1700 and above ' 2-3 person 83 2.67 3.03 .36 .10 b, or more 12 14.111 11.15 .01 > .110 :E1 I Pork item expenditures before increase in income. 59 Large families in both income groups tended to increase their pork item expenditures. Small families on the other hand were primarily responsible for the negative dynamic income expenditure relationships exhibited by bacon, chOps and ham. Comparison of the Observed and Expected Changes in Pork Item Expenditures The low income families: "observed" per capita ham and roast expeniitures were in the direction “expected.“ Low income small fami- lies were primarin responsible for the “observed" per capita chops and bacon expeniiture adjustments being opposite of what was nexpected.n See Table 200 TABLE 20 OBSERVED AND EXPECTED PORK CUTEIPENDIIURES ASSOCIATED RITE INCREASED mR CAPITA DISPOSARLE INCOME, Low INCOME GROUP, MICHIGAN STATE UNIVERSITY COI‘BUMER PANEL SUB-SAMPLE, 1953-57 Changes in Per Capita Expenditures Pork Cut Observed Fbcpected Bacon - "012 .16 011013 . " 005 032 Ham . _ .26 ‘ .h8 Roast .27 .111 Low income families overadjusted their roast experriitures, whereas these families underadjusted their bacon, chop, ani ham expenditures. Ham expeniiture unieradjustments were probably due to the age-size composition of families in the two income groups. These income groups . ill. qt. 60 represented two different populations with the low income group being composed of younger ard larger families than families in the high income group. Static cross-sectional relationships indicated that high income families spent almost twice as much per capita for ham than did low income families. It was reasonable to expect that low income families would not make the "expected" adjustments within one year. Mas-em Family size was found to be an important factor influencing pork cut expenditure adjustments. Low income small families decreased their bacon, chops, and ham expenditures after receiving an increase in income. These responses were primarily responsible for the "observed” bacon and chop expenditures adjustments not being in the "expected" direction. Low income families experienced a 111 per cent increase in income, Of which 6 per cent was inflationary. These inflationary increases in I income tend to make the “observed" expenditure adjustments larger than if the increase in income consisted of an all real income increase. In recent years, consumers have been shifting their tastes am preferences by purchasing more beef and decreasing their pork consump- tion. These changes might also have been responsible for consumers unieradjusting their pork cut expenditures of what was "expected“ based on cross-sectional relationships. . The evidence presented suggests that low income families receiving increases in income did adjust their pork roast ard ham expenditures in the direction expected. However, these families tend to underadjust their “observed“ experriitures of what wll "expected.“ Inflationary 61 increases in income, shifting tastes and preferences and differences in family characteristics between the two income groups. Therefore, the assumption that families experiencing increases in income will adjust their pork item expenditure patterns so as to be similar to the expenditures of the income group the families are enter- ing does nOt appear to be entirely valid. 1 r CHAPTERV CHANGES IN BROILER AND COLD CUT EXPEDDITURES ASSOCIATED WITH II‘CREASED INCOME . Introduction The purpose of this chapter was to derive static cross-sectional am dynamic income-expenditure relationships for broilers and cold cuts. A seconiary objective was to iniicate the relationship that existed between family size within income groups and broiler and cold cut expeniitures. Broiler prices fluctuated the greatest of any of the meat cuts analyzed in this study. There was at least a five cent fluctuation in price between an two Consecutive years during the period 1953-57 with the exception of 1956—57 when the average broiler prices differed by only two cents. Dynamic income-expenditure elasticities derived for broiler are probably affected to some extent by these price changes. Cold cut prices were the most stable of any of the meat cuts studies. The average variation in prices occurring between any two consecutive years during the period 1953-57 was less than two cents. The effects of changes in price influencing the purchases of cold cuts were negligible. Table 21 Shows the results obtained in the 1955 United States Department of Agriculture Food Consumption Study. These data differ 62 63 from the sub-sample data in that chicken expenditures included not only broilers, but all types of poultry products such as eggs, duck, turkey, etc. Luncheon meats included fresh arri canned varieties, but excluded frankfurters . TABLE 21 AVERAm LEVEL OF FAMILY EXPEIDITURES FOR BROILERS AND COLD CUTS, ALL URBAN EAMIIIES IN THE NORTH CENTRAL REGION, A WEEK IN APRIL-JUNE, 1955a - Disposable Income Chicken. Luncheon Meat per Household Expenditures Ebcpenciitures (dollars) (dollars) Less than “1,999 .91 .52 $5,000 and above 1.23 .59 3Source: “Food Consumption of Households in the North Central Region,“ Rept. No. 3, Hashington, D. C., U.S.D.A., 1955, pp. 70-71. bOne person families were excluded. Families in the $5,000 and above income group spent 32 cents more per week for chicken than did the less than $h,999 income group. These income groups spent similar amounts for luncheon meats. . Static income-expeniiture elasticities computed from these data showed that families in the less than $11,999 income group receiving increases in income were "expected“ to make substantial increases in their broiler expenditures and to only slightly increase their cold out expenditures. Cold cut expenditures were expected to increase slightly 6h because of the small difference in the level of expenditures between the less than $11,999 and $5,000 and above income groups. Presentation of analyses was similar to Chapters ELI and IV with the exception that the broiler and cold out analyses are not discussed concurrently. This was done because broilers ani cold cuts are two different types of meat and no direct comparison can be made between [ them. The first section deals with a static analysis of the family sub-sample. The second section discusses the dynamic aspects associated with increased income. The third and final section was a comparison of the dynamic and static income-expenditure elasticities. Results from a Static Cross-Sectional Analysis of Broiler kpeniitures The high income group had larger broiler expenditures than did the low income group. The level of broiler expeniitures for these income groups were signifiCantly different from each other at the 5 per cent level. See Table 22. TABLE 22 PER CAPITA BROIIER mmmITURIS, BY INCOME GROUPS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-57 J _— T :m Number of Average Annual Per Capita Income Groups Families Broiler Expenditures (dollars) Less than $1700 119 g 3.21 $1700 and above 95 6.50 65 A static income-expenditure elasticity computed between the low mm high income groups showed that families in the low income group were "expected“ to increase their broiler expenditures by .9179 per cent after receiving a one per cent increase in income. Small families in the high ani low income groups had larger per capita broiler expenditures than did the large families in these income groups. See Table 23. TABLE23 PER CAPITA BROILER EIPENDITURIS, BY INCOME (ROUTE SUBDIVIDED BY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER. PANEL SUB-SAME. 1953-57 Income Groups by Number of Annual Per Capita Broiler Family Size Families Expenditures (dollars) Less than $1700 2-3 person 33 11.16 h or more 86 2.81 81700 ani above 2-3 person 83 6.89 h or more 12 5.29 The differences in the level of expenditures for two family sizes in the low income group were significantly different from zero at the 5 per cent level, whereas the difference in the level of broiler expenditures for the high income group family sizes was non-Significant. AS it was found for beef ani pork items, family size was an important factor influencing the level of experfiitures for those families in the zfl-k‘nw-I 66 low income group. Static cross—sectional elasticities used to predict the responses of low income families to increased income would result in an inaccurate estimate if the two populations are not homogeneous with respect to family size. Results from a Dynamic Analysis of Broiler Experditures Families average yearly broiler expenditures were "Observed" before ani after a family received an increase in income. Table 2h shows that families in the two income groups substantially increased their broiler expenditures. These expenditure adjustments for both income groups were significantly different from zero at the 10 per cent level. TABLEZh ANNUAL AVERAcm PER CAPITA BROILER EXPENDITURES ASSOCIATED wITH INCREASED INCOME,BY INCOME GROUPS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-S7. Incmne Groups b Change Level of E13 E2 in E c Significanced Less than 81700 2.98 3.104 .h6 .01 $1700 ani above 6.16 6.83 .66 .10 aE1 u- Expeniitures before an income increase. bi”.2 - Expeniitures after an income increase. c Change in E - 11‘.2 - E1. dThe critical level of significance was at the 10 per cent level. The dynamic income-expeniiture elasticity estimates for broilers for the low and high income group were 1.08 and 1.00. Families in 67 both income groups increased their per capita broiler expenditures approximately one per cent after experiencing a one per cent increase in income. The large families in both income groups increased their per capita broiler experxiitures more than small families. See Table 25. The large families expenditure adjustments were also more significant than the adjustments made by small families. TABLE 25 ANNUAL AVERAGE PER CAPITA BR011.ER EIPENJITORIS ASSOCIATED WITH INCREASED INCOME, BY INCOME GROUPS SUBDIVIDED RY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-57 Income Group by Ember of a b Change Level of Family Size Families E1 E3 in E C Significance Less than $1700 h or more 86 2.117 3.111 .67 .005 $1700 and above 2-3 person 83 6.56 7.22 .66 .20 h or more 12 11.85 5.73 .88 .10 3E1 - Broiler expenditures before an income change. bl!)2 - Broiler expenditures after an income change. Change in E :- E2 - E1. e critical level of significance was at the 10 per cent level. C Family size was an important factor influencing broiler expenditure adjustments associated with increased income, especially in the low income group. These results indicated that dynamic income elasticities used to predict. the responses of consumers to increased income would 68 result in inaccurate estimates if the size composition of families in the population from which the elasticities were derived differed from the population whose responses we are attempting to predict. Comparison of the Observed and Ebcpected Charges in Broiler Expenditures Low income families were "Observed“ increasing their broiler r” expeniitures by 146 cents after receiving a 1).; per cent increase in income. Static cross-sectional relationships indicated that low income families were "expected" to increase their broiler expenditures by )10 cents. Thus low income families over-adjusted their broiler expenditures by six cents after receiving an increase in income. The differences in the "Observed" and "expected" expenditure adjustments was the least of any of the meat cuts analyzed. Hmever, during the time period in which data for this study were obtained, broiler prices decreased substantially. 3 Consumers have increased their demani for broilers due to this price change by significantly increasing their broiler expenditures and decreasing their pork consumption. _ Tuere has also been an increase in per capita demani (demani curve shifting to the right) due to changing tastes am preferences, greater availability, ani product improvement. The-so price effects intermingled with the income effects resulted in the low income families over-adjusting their "Observed“ brbiler expeniitures . 69 Inflationary increases in income also resulted in the low income families! "observed" expenditure adjustments being larger than if the increasein income consisted of an all "real" income increase. The evidence presented suggested that low income families receiv- ing increases in income do adjust their broiler expenditures in the direction expected. Low income families‘ "observed" expeniiture firm adjustments would probably have unier-adjusted the "expected" adjust— { ments if price and inflationary income increases were eliminated from the data. Therefore, the validity of the assumption that consumer experienc— ing increases in income will adjust their broiler expenditure patterns similar to the expenditures of the income group the Consumers are entering can be critically questioned. Results From 8. Static Cross-Sectional Analysis of Cold Cut Expenditures Families in the high income group had larger per capita cold. out expenditures than did the families in the low income group. See Table 26. The difference in the level of cold out expenditures for these income groups was 83 cents, which is a relatively small difference when compared to other meat items. The levels of expenditures for these two income groups were significantly different from each other at the 10 per cent level. Similar results were obtained in the 1955 Food Consumption Survey. 70 TABLE 26 PER CAPITA COID CUT EXPENDITURES, BY INCOME GROUPS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-57 Number of Annual Average Per Capita Income Groups Families Cold Cuts Experditures (dollars) Less than $1700 119 5.11 $1700 and above 95 5.9).; A static income-expenditure elasticity computed from these cold out data iniicated that families in the low income group receiving a one per cent increase in income were upected to increase their cold out expenditures by .21 per cent. "mpected" responses for low income families in adjusting their cold out expenditures to an increased income were the smallest of any of the meat cuts studied. Results of the Food Comumption Survey showed that low income families were expected to make only a slight increase in cold cut experxiitures after receiving an increase in inCome, the smallest adjust- ment of the seven meat cuts analyzed. Family size did not influence the level of cold cut expenditures in the low income group and was of minor significance in the high income group. The level of cold out expenditures for family sizes within the high income group were significantly different from each other only at the 20 per cent level. See Table 27. Therefore, the static income- expeniiture elasticity computed for cold cuts was unaffected by family size. 71 TABLE27 PER CAPITA COLD CUT EXPENDITURES, BY INCOME GROUPS SUBDIVIDED BY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER NEGEI.SUB_SAMELE, 1953-57 . Income Groups by Number of Amual Average Per Capita Family Size Families Cold Cut Expenditures (dollars) Less than $1700 2-3 person 33 5.10 l; or more 86 5.10 $1700 and above 2-3 person 83 6.22 ); or more 12 )A.7); Results from a Dynamic Analysis of Cold Cut Expenditures Table 28, shows that both income groups enlarged their co1d cut experxiitures after experiencing an increase in income. However, only the low income families expenditure adjustment of 3).; cents tested to be significantly different from zero. The dynamic income-expenditure elasticities for cold cuts for the low air! high income groups were .50 and .31 respectively. These results are similarto those of previous cross-sectional studies which indicated that low income families increased their meat expeniitures more than did the higi income families. A ' Although family size did not influence the level Of cold out expenditures, it was an important factor affecting cold out expenditure 72 TABLE 28 ANNUAL AVERAGE PER CAPITA COLD CUT EXPENDITURES ASSOCIATED WITH INCREASED INCOME, BY INCOME GROUPS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-57 Number of a b Change Level of d Income Groups Families E1 E2 in E 0 Significance Cold Cuts Less than $1700 119 h.9h 5.28 .31; .025 81700 and above 95 5.85 6.03 .18 .30 8E1 - Cold cut expenditures before an income increase. bE2 In Cold cut expenditures after an income increase _cChangeinE-E3-El. dThe critical level of significance was at the 10 per cent level. adjustments associated with increased income. This was especially true for the low income small families whose per capita expeniiture adjust- ments were significantly different from zero at'the one per cent level. See Table 29. The magnitude of the dynamic elasticities were affected (by family size. These elasticities therefore could only be used to predict con- sumer responses to an increased income for populations which are homo- geneoUs with respect to family size. If an attempt is made to predict consumerst responses in areas where family Size composition differed- from the Sub-sample, an inaccurate estimate Of families responses will result, especially in the low income group. ‘i._ 73 TABLE 29 ANNIAL AVERAGE ER CAPITA COID CUT EIPENDITURES ASSOCIATED WITH INCREASED INCOME, BY INCOME GROUIS SUBDIVIDED BY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-57 Income Groups by Number of a b Change Level of Family Size Families E, E,2 in E c Significanced Cold Cuts Less thanfi$1700 2-3 para 0! 33 ).; . 67 5 .53 .83 .01 h or more 86 5.02 5.18 .16 .20 $1700 and above 2-3 person 83 5.17 6.26 .09 .hO h or more 12 h.52 h.96 .hh .30 %1,- Cold cut expenditures before an income change. bE2 - Cold cut expenditures after an income change. dChange in.E - E2 - E1. The critical level of significance was at the 10 per cent level. Comparison of the Observed andExpected Changes in Cold Cut Expenditures Low income families were "bbserved" increasing their cold out expenditures by'3h cents after receiving a 1h per cent increase in income. Static cross-sectional relationships indicated that low income families were “expected" to increase their cold out expenditures by 16 cents. Thus, these families over-adjusted their cold out expenditures by 18 cents. A.factor which might have influenced low income.families to over-adjust was changing tastes and preferences. 7h If the low income families' "Observed" expenditure adjustments were corrected for inflationary income increases, the corrected "observed" experditure adjustments would probably be greater than what was "expected.“ Therefore, the assumption.used when.cross-sectional analyses are used for prediction purposes does not appear to be valid. These over-adjustments might be due to low income families substituting expersive cold cuts such as pressed ham for low cost items such as bologna. CHAPTER VI SUMMARY AND CONJLUSIOBB ‘ In.the past, income expenditure relationships for selected meat items have been computed by either time-series or a cross-sectional analyses. 'Hhen the results of cross-sectional analyses were used for prediction.purposes, it was assumed that families receiving an increase in income would adjust their expenditures to be similar to the expendi- ture pattern of the income group the families are entering. This study was comtructed to test the validity of this assumption. l.secondary objective was to determine the relationship that existed between family sizes within.income groups and meat cut expenditures. The Hidhigan.State University'Consumer Panel was the source of data.for the present study. Only those families who received increases in per capita disposable income and who partiCipated in.the panel at least hO'weeks out of a.year for two consecutive years were included. Onsrperson families were excluded because of the extreme variability in expenditure data fer these families. Some of the wide price Changes in.beef were avoided by omitting 1951 and 1952 Consumer Panel Data, but we still have pork and broiler price fluctuations. Pbrk prices were relatively unstable during the period 1953-57, however, the net price change'was.relatively small and probably exerted a.mincr influence on family purchasing patterns. Broiler priceS'were the most unstable of any of the meat cuts analyzed, 75 76 with the price movements being generally downward. This downward movement was associated with structural changes within the broiler ‘ industry. Per capita income for the Consumer Panel sub-sample exhibited an upward trend during the period 1953-57. The sub-samples' mean per capita income increased approximately 20 per cent during the period 1953-57. - ‘ f The meat cuts studied were beef roasts and steaks; pork Chops, roasts, bacon and ham; cold cuts, and broilers. These meat cuts were selected on the basis of the following criteria: '1) ImpOrtance of the meat cut in the family food budget, and 2) The existence of different levels of experditures among families grouped by level of income. Minty-nine families were first separated into four per capita income groups; less than $1200, 81200—1700, $1700—2h00 and 321;,00 ani above. Cross-sectional income-expenditure elasticities computed for these income groups were unstable and did not bear consistent relation- ships when compared tO previous studies. The results obtained for the eight meat cuts may be seen in Appendix A. These inconsistencies were probably due to the limited number of observations in each income group. The sub-sample was regrouped into two per capita income groups: less than $1700 and $1700 and above. Static cross-sectional income elasticities computed between these income groups were in the direction and approximate magnitude of some of the previous cross-sectional analyses. films by increasing the number of Observations in each income group, stable arfl consistent relationships were obtained. 77 Families within each income group were separated into two family size categories; 2-3 person and h or more person families. The results of this second classification of families indicated the relationship between meat cut expenditures and family size within income groups. Utilizing the MiChigan State University Consumer Panel, changes in meat item expenditures associated with increased income. were "Observed" for each income group. Dynamic income-expenditure elasticities were computed for both income groups. Per capita income and expenditures before and after an increase in income were averaged for each income group. 'flle average per capita meat expenditures and income for the two income groups was the basis for computing a static cross-sectional income-expenditure elasticity between these two income groups. When this type of elasticity is used for prediction purposes, it is usually assumed that families receiving increases in income will adjust their . meat experditures to the purchase patterns of the income group the families are entering. Meat experxiitures were "Observed" before and after the family experienced an increase in income. The "expected" meat expenditures for the low income families were derived by the following procedure. The static income-expenditure elasticity times the low income group‘s percentage change in income times the low income families per capita meat expenditures before the increase in income yielded the "expected" expenditure adjustment. The "Observed" and “expected“ expenditure adjustments were compared‘to determine if low income families adjusted 78 their "observed" expenditures in the direction and magnitude of the static cross-sectional estimate. This comparison was the basis for accepting, modifying, or rejecting the validity of the assumption used in making prediction based on conventional cross-sectional analyses. 0n the average, the high income families per capita beef roast ' and steak expenditures were significantly larger than the expenditures made by low income families. Per capita expenditures for roasts and steaks were significantly higher for the small families in these income groups than larger families. These results indicated that family size was an important factor influencing these beef cut expenditures, especially in the low income group. Therefore, static cross-sectional elasticities should only be used to predict the responses of families to increased income which are homogeneous with respect to family size in the sample from which the static elasticities were computed. Neither of the income groups made significant increases in their "observed“ per capita beef roast and steak expenditures after receiving an increase in income. High income families actually decreased their roast expenditures by 614 cents . Similar results were Obtained by Westl in his study Of all meats. West indicated that high income families may have increased their expenditures for meals away from home and thus reduced their meat expenditures for meals eaten at home. There was a definite tendency for small families in both income groups to be more responsive to an income increase than were large West, _0_Po £312., 1). 72. 79 families. Thus family size tended to affect the values of the dynamic income-expenditure elasticities, especially in the low income group. Low income families urder-adjusted their "observedn beef roast and steak expenditures when compared with the "expected" adjustments computed from cross-sectional relationships. The high income group comisted of smaller and Older families which spent approximately twice as much for these beef cuts than did low income families. These family characteristics affected the static elasticities arxi the subsequent "expected" expenditure adjustments. The "expected" expenditure adjust- ments as well as inflationary increases in income tended to make the "observed“ expeniiture adjustments larger than what was “expected." Based on this evidence, the assumption that families experiencing an increase in income will adjust their beef roast and steak expendi- tures patterns to be similar to the expenditures Of the income group families they are entering does not appear to be valid. The high income families had larger expenditures for pork chops, roasts, bacon, and ham than did the low income families. The low income small families tended to have higher per capita expenditures for these pork cuts than did large families in the same income group. However, in the high income group there was no tendency for family size to affect the pork cut expenditure levels . An examination of the pork cut expenditure adjustments by income groups indicated that there was no consistent relationship between increased income and expenditure adjustments. In the low income group, 80 dynamic income—expenditure elasticities were negative for two of the four pork cuts. When the data were grouped according to family size, it was found that the small families in the low income group were primarily responsible for these negative elasticities. It was evident that family size was an important factor affecting consumer responses. The "observed" chop and bacon expenditure adjustments were not 7“” in the direction of the i‘expected" expenditure adjustments computed from cross-sectional relationships. Low income families over-adjusted their per capita roast expenditures when compared with the "eXpected" expenditure adjustments. Hhereas, these families "Observed“ ham expenditure changes were less than "expected." It was found that inflationary increases in income, family character- istics between the two income groups and family sizes within the low income group were partially responsible for the ulnar-adjustments of bacon, chops and ham with compared with the "expected" adjustments. Therefore, the assumption that consumers experiencing increases in income will adjust their pork item expeniiture patterns to be similar to the expeniitures of the income group the consumers are entering does not appear to be valid. The high income groups' average per capita expenditures for broilers was approximately twice as large as the broiler expenditures made by the low income group. Family size was found to be a significant factor affecting per capita broiler expenditures in the low income group. Low income small families had larger broiler expenditures than did large 81 families in this income group. In the high income group, small families had the largest broiler expenditures, but the difference for the two family size categories was non-significant. The high income groups "Observed" expenditure adjustments for broilers was larger than the adjustment made by the low income group. When the data were regrouped by family size, it was found that large, families in both income groups were more responsive than small families in adjusting their broiler expenditures after receiving an increase in 7 income. These results indicated that the value of the dynamic income— expeniiture elasticities will depend on the composition of family size in the sample. V Low income families over-adjusted their "observed" broiler expendi-v ture adjustments when compared with the “expected" adjustments computed from cross-sectional relationships. The difference in the “observed" and "expected" broiler expenditure adjustments was the smallest Of all the meat cuts analyzed. However, if the "Observed" broiler expendi- ture adjustments were corrected for price changes and inflationary increases in income, the "observed" expenditure adjustments would probably be less than "expected ." These results tend to indicate that the assumption used when conventional cross-sectional analyses are used prediction purposes is not valid. The low and high income groups per capita cold out expenditures were not significantly different from each other. The static income- expenditure elasticity for cold cuts of .21 was the. smallest of any of the meat cuts analyzed. Family size was fourrl not to have a 82 significant influence in determining the level of cold out expenditures. Static elasticities computed for cold cuts in this study can probably be used to accurately predict consumer responses in areas where the composition of family size is different between the two populations. The low income group was the most responsive in adjusting their “observed“ cold out expenditures to an increased income. Only the low income small families cold out expenditure adjustments associated with increased income tested to be significantly different from zero. Thus, only the dynamic income-expenditure elasticities for the low income groups was affected by family size. Low income families over-adjusted their “observed“ per capita cold out expenditures when compared with the nexpectedn adjustments. The assumption that low income families receiving increases in income. adjust their cold out expenditures similar to the expenditures of the income group the families are entering does not seem to be valid. The evidence presented in this study indicated that the assumption used when cross-sectional analyses are used for prediction purposes will result in inaccurate estimates of family responses to increased income unless adjustments for family size variations, lag reactions and family characteristics between income groups are made. The static cross-sectional and dynamic income-expenditure relation- ships-indicated that family size was an important factor influencing the per capita expenditures for most of the meat items analyzed. Family size was of more significance in the low income group than in 83 the high income group. These results suggest that income elasticities used to predict the respOnses of low income families to increased income will result in an inaccurate estimate if the two populations are not homogeneous with respect to family size. Further work in this area is needed to show the adjustments made by families experiencing similar increases in income, but differing in [ family characteristics such as age, education, occupation Of householder, endother factors. The effects of these factors on adjustments made in the purchasing patterns of consumers will probably explain many Of ‘ the differences in the "observed" [and "expected“. adjustments in L”— expenditures examined in this thesis . BIBLIOGRAPHY Bilkey, Warren J ., The Basic Relationships in Consumer Exmnditure Behavior. Harvard Studies in Marketing Farm Products, Number H—H, Cambridge, 1951, 66 pp. Brinegar, George K. "Short-Run Effects of Income Change on Expeniitures,“ Journal Of Farm Economics, Vol. 35, No. 1, 1953, PP. 99‘1090 3*: Bur. of Human Nutr. and Home Econ. U. S. Dept. of Agr. Meat: Variations in Consumption and Interrelationships with Other Foods. Commodity Summary No. 11, 1951, 3lfipp. Burk, Marguerite C. "Changes in Demand for Food From 19hl to 1950," Journal of Farm Economics, Vol. 33, NO. 3, 1951, pp. 291-295. Dixon, Wilfrid J. and Frank J. Massey, Jr. Introduction to Statistical REL—4 Anal is, 2111 ed., New York: McGraw-Hill Book Company, Inc., 1937, 1138 pp- Fox, Karl A. "Factors Affecting Farm Income, Farm Prices, and Food Consumption,“ Agr. Econ. Res., Vol. 3, 1951, pp. 65-lll. Moss, Thomas Neil. "Some Relationships Of Selected Socio-Economic Factors to Food Consumption ani Expenditures .“ Unpublished BI. D. Thesis, Department Of Agricultural Economics, Michigan State University, 1952, 365 pp. Quackenbush, G. G. "Demand Analysis From the M30 Consumer Panel," ‘ Journal Of Farm Economics, Vol. 36, NO. 3, 1951;, pp. h15-h27. Riley, Harold M. “Some Measurements of CorsUmer Demand for Meat." Unpublished Ph. D. Thesis, Department of Agricultural Economics, Michigan State University, 1958, 218 pp. Schultz, Henry. The Theorylam Measurement of Demard. Chicago: University of Chicago Press, 1938, 817 pp. Shaffer, J. D. "A Plan for Sampling a Changing Population Over Time ," Journal of Farm Economics, Vol. 36, No. 1, 19514, pp. Shaffer, J. D. "Methodological Bases for the Operation Of a Consumer Panel.“ Unpublished Ph. D. Thesis, Department of Agricultural Economics, Michigan State University, 1952, 767 pp. 81:, 85 Shaffer, J. D. and G. G. Quackenbush. "Cooperation and Sampling in Four Years of Michigan State University Consumer Panel Operation." Quarterly Bulletin. Agricultural Experiment Station, Michigan State University, East Lansing, Amt, 1955, pp. 85-103. Shaffer, J. 1)., G. G. Quackenbush and T. N. Moss. The Comumption of Meat and Related Products in Lansing, Michigan, Spring, 1950. Mich. Agr. Exp. Sta., Tech. Bul. 239, 1951;, 25 pp. Shepherd, Geoffrey. Changes in the Demand for Meat and Dang Products in the United States Since 121 . Iowa Agr. Exp. Sta. Bul. 368, 1974 s 39 PP. Simmons, Hill M. Consumer Meat Purchases in Syracuse, New York 19MB and Comparison with 19712. New York TCornellTLgr. Exp. Sta. B111. 869’ 1931’ 30 pp. Stigler, George J. “me Early History of Empirical Studies of Consumer Behavior." Journal of Political Economy, Vol. 62, No. 2, 1951» pp. 95-13. Stigler, George J. The Them of Price. New York: Macmillan Compamr, 1952, 310 PP0 Waite, H. C. and R. H. Cox. A Stud of the Consumption of Meats in Minneapolis, 1931;. Minn. Agr. Ep. Sta. Bul. 321,741 pp. Waite, H. C. and H. C. TrelOgan. Aggicultural Market Prices. New York: John Wiley and Sons, Inc., 2nd ed., 1951, 1040 pp. West, Jerry Glenn. "Estimates of Income Elasticity from Consumer Panel Data.“ Unpublished Ph. D. Thesis, Department of Agri- cultural. Economics, Michigan State University, 1958, 122 pp. Working, E. J. “Agricultural Demand During Rearmament." Journal Of Farm Economics, Vol. 314, No. 2, 1952, pp. 206-221;. APPENDJX A Thie appendix is composed of the Michigan State University Consumer Panel sub-sample data divided into four per capita income groups. in attempt was made tO have a similar number Of family observations in each income group. However, when similar observations in each income group were attained, it was fourri that there were only 2—3 person families in the $21400 and more income group. Thus the following tables do not list any family size distinction for this income group. Static cross- sectional and dynamic income—expenditure relationships were com- puted exactly as described in the methodology section Of Chapter II o 86 87 TABLE A PER CAPITA EXPENDITURES FOR SELECTED BEEF CUTS, BY INCOME GROUPS, SUBDIVIDED BY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE 1953-I957 *- ‘1 Income Groups by Number of Average Annual Per Capita Exper‘rliture Family Size Families Roast Steak (dollars) (dollars) I 13”.] Less than $1200 59 5.69 6.12 2-3 person 11:. 7 .119 7 .96 14 or more person 145 5 .143 5.514 7 3120041700 60 6.89 8.23 2—3 person 19 8.33 9.70 14 or more person 111 6.21 7.55 517% 21400 1114 9 .83 1133 2-3 person 32 9.62 12.00 14 or more person 12 10.37 9.52 $21400 and above 51 10.141; 13.26 TABLE B STATIC ncmLE—EYIENDITURE ELASTmITIES FOR SELECTED BEEF CUTS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUBTSAMILE 1953-4957‘a Income Groupsb Roast ‘ Steak Medium to Medium-High 1.000 .912 Medium-High to High .1118 .369 aArc elasticities computed between two income groups. bLow ignome group includes families with per capita income of under 1200. Medium income group includes families with per capita income of time-$1700. ' Medium—high income group includes families with per capita income Of $17004121400. High income group includes families with per capita income of $21400 and above. TABLE 0 ANNUAL AVERAGE PER CAPITA BEEF CUT EXPENDITURES ASSOCIATED WITH ImREASED INCOME, BY INCOME GROUPS SUBDIVIDED BY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-1957 “ v_._-—_' -‘- 88 Income Groups by Number of b Change Level of Family Size Families Ela E3 in E C Significance Roast Less than $1200 59 5.71 5.67 -.014 > .140 2-3 person 114 _ 7.11 7.86 .75 .30 14 or more person 145 5.28 14.99 -.29 .30 $200-$1700 60 6.56 7 .21 .65 .10 2—3 person 19 7.78 8.88 1.10 .20 14 or more person 141 5.99 6.143 .1414 .20 tum-$21400 1414 10.27 9.39 -.88 .20 2-3 person 32 10.3h ‘8.9l ~l.h3 .10 14 or more person 12 10.09 10.65 .56 .140 $2h00 and above 51 10.6h 10.2h -.h0 .30 Steak. Less than $1200 59 ‘ ' 5.90 6.314 .1414 .20 2-3 person 114- 7.90 8.02 .12 > .140 14 or more person 145 5.27 5.81 .514 .20 $12MI700 60 8 032 8 0114 - c 01.10 2-3 person 19 9.52 9.87 .35 .140 14 or more person 141 7.76 7.314 -.142 .20 $17M2h00 1111- 11 053 11012 " o 01-1-0 2’3 wrson 32 12038 11062 ”076 030 14 or more person 12 9.214 9.79 .55 .140 $21400 ani above 51 12.86 13.65 .79 .20 aE1 - Expeniitures before an income increase. 2E2 -- Expenditures after an income increase. 89 TABLE D DYNAMIC INSOME~EIPENDITURE ELASTICITIES FOR SELECTED BEEF CUTS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-1957 Income Groups Roast Steak Less than $1200 -.058 .526 $120041700 .730 - .167 31700421400 - .682 - .295 321“” 8.111 above " 01.1.26 .66; TABLE E PER CAPITA nIENnITURE-s FOR SELECTED PORK CUTS, BY INCOME GROUPS SUBDIVIDED BY FAMILY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER _ PANEL SUE-SAMPLE, 1953-r1957 Income Groups, by Family Size - Number Of Average Annual Per Capita Expenditures Families Bacon Chop Ham Roast (dollars) (dollars) (dollars) (dollars) Less than $12.00 2—3 person 14 or more person tum-$1700 2-3 person 14 or more person 31700—321400 2-3 person 14 or more person 821400 and above 15} K4185 5623 SEE 14065 5.28 11.145 14.53 5.10 14.214 6.21 6.67 14.97 14.70 3 012 14.142 2 .71 3 .ll 2 .67 3.32 4 5.73. . 5 .22 7 .10 11.63 1 1.1032 14.28 11031-41- 5.12 14.10 S .59 8.00 8.21 4 7 .113 8.214 2 .17 3 .91 1.143 2 00h 2 .33 1.90 3 .39 3 .11 14.15 2.61 TABLE F 90 STATIC IICOME—EXIEDDITURE EILASTICITIES FOR SELECTED PORK CUTS, MICHIGAN STATE UNIVERSITY COI‘BUMER PANEL SUB-SAMPLE, _ 1953-1957 W Income Groups Bacon Chop Ham Roast LOW '00 Hedi—m “0069 “0001.1. 01.1.70 “0186 Medium to Medium-High .890 1.683 1.2149 1.1.1.31 HediUIn‘Higl “150 High -0682 "011118 0072 “0638 TABLE G ANNUAL AVERAGE PER CAPITA PORK CUT EXPEIDITURES ASSOCIATED WITH INCREASED INCOME, BY INCOME GROUPS SUBDIVIDED BY EAMJIY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 1953-1957 Imome Groups by Number Of a b Change Level of Family Size Families E1 E2 in E Significance Bacon Lass than $1200 59 14.83 14.146 -.37 .05 - 293 mmon 11.1. 5078 1.1.078 "1000 010 A 1.], or more person 1.1.5 14053 1.1036 “017 030 3120041700 60 14.147 14.59 .12 .140 2‘3 mrson 19 5018 S017 "001 > .140 14 or more person 141 14.15 14.32 .17 .30 317004121400 . 1414 5.90 ' 6.52 .62 .025 2-3 person 32 6.38 6.96 .58 .10 14 or more person 12 14.60 5.314 .714 .10 $21.1!” am above 5]. 1.1.083 1.1057 ‘026 030 ~ Continued TABLE G - Continued 91 Income Groups by Number of a b Change Level of Family Size Families E1 E2 in E 6 Significance Cho 1 Less than $1200 59 3019 3001.1 I"'01; 030 2~3 person 114 14.70 14.114 .56 .30 14 or more person 145 ‘ 2.72 2.70 -.02 > .140 51200-31700 60 3.08 4 3.114 .06 > .140 2—3 person 19 2.56 2.77 .21 .140 14 or more person 141 3.32 3.31 .01 > .140 $17004 21400 1414 5 .1414 6 .01 . 57 .30 2-3 person 32 5.08 5.35 .27 .140 14 or more person 12 6.141 7.79 1.38 .20 . #21400 and above 51 5.05 14.149 -.56 .05 Has Less than $1200 59 14.08 14.56 .148 .10 .2—3 person 114 14.62 3.93 -.69 .20 14 or more person 145 3.92 14.76 .814 .025 $1200-$1700 60 5 .10 5 .114 .014 > .140 14 or more person 141 5.50 5.67 .17 .140 81700-821400 1414 8.00 7.99 -.01 > .140 14 or more person 12 7.1414 7.142 -.02 > .140 $21400 and above 51 8.58 7.89 -.69 .20 RO___a_S__t Less than $1200 59 2.03 2.31 .28 .20 2-3 person 114 3.67 14.15 .148 .30 14 or more person 145 1.53 1.33 --.20 .20 $120Ml700 60 1.91 2 .16 .25 .10 23 person 19 2.140 2.25 -.15 .140 . 14 or more person 141 1.69 2.11 .142 .025 $1700-$21400 14.14 3 .27 3 .51 .214 .30 2-3 person 32 2.914 3.27 .33 .30 14 or more person 12 14.114 14.15 .01 > .140 $21400 and above 51 2.1414 2.78 .314 .20 :- Experxiitures before an income increase. bEiu IExperrlitures after an income increase. 0Change in E- E-2 - E1. TABLEH DIMC-INCOME EXPENDITURE EEASTICITTI'B FOR SELECTED PORK CUTS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE, 92 1953-1957 Income Groups Bacon Chops Ham Roasts Less than $1200 - .662 - .1400 .9214 1.0714 $1200-$l700 .2014 .1148 .060 .9149 111700-321400 .761 .758 - .018 .539 $214.00 am above “06.16 "10310 “0933 101.153 TABLE I PER CAPITA EXPEIDITURES FOR EROILERS, BY INCOME GROUPS SUBDIVIDED BY PANEIY SIZE, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB—SAMPLE, 1953-1957 .___.__ Income Groups by Number Of Average Annual Per Capita Expenditures Fainly Size Families Broilers (dollars) Less than $1200 59 2.87 2-3 person 114 3.214 14 or more person 145 2.75 3120041700 60 3.56 2-3 person 19 5 .08 14 or more person 141 2.85 31700421100 1411 7.140 2-3 person 32 8.18 14 or more person 12 5.29 $21400 and above 51 5.59 93 ' TABLE J STATIC-IEOME EXPENDITURE mAsTICITIES FOR BROILERS, MICHIGAN STATE UMVERSITY COAsUMER PANEL SUB-SAMPLE, 1953-I957 Income Groups Broilers Low to Medium 595 Medium to Medium-High 1.9149 -0685 Medium-High to High TABLE K ANNUAL AVERAGE PER CAPITA BROILER EXPENDITURES ASSOCIATE) WITH ImREAsm INCOME, RI INCOME GROUPS SUBDIVIDED BY PAMELI SIZE, MCHIGAN STATE UNIVERSITI CONSUMER PANEL SUE-SAME, 1953-1957 a . Income Groups by N‘mnber of a b Change Level of Family Size Families El E2 in E c Significance oflers Less than $1200 59 2.63 3.10 .147 .10 2-3 person 114 3.27 3.21 —.06 > .140 14 or more person 145 2.143 3.07 .614 .025 ”200-31700 60 3 .33 3 .78 .145 .05 14 or more person 141 2.50 3.20 .70 .05 31700421400 1414 6.96 7 .83 .87 .20 2-3 person 32 7.75 8.61 .86 .20 1.4 or more person 12 14.85 5.73 .88 .10 821400 am above 51 5.36 5.82 ' .146 .30 :- Expenditures after an increase in income. 8‘1.3.1 I Experrlitures before an increase in income. bEl c 2 Chamge in E I E,3 - E1. 9S TABLEN STATIC INCOME EIPENDITURE ELASTICITIFS FOR COLD CUTS, MICHIGAN STATE UNIVERSITY CONSUMER, PANEL SUB-SAMEE, 1953-1957 Income Groups Cold Cuts Low to Medium .336 Medium to Medium-High .391 ,1 Medium-High to High -.23h 1 TABLE 0 ANNUAL AVERAGE PER CAPITA COLD CUT EXPENDITURES ASSOCIATED WITH INCREASED INCOME, ,BY INCOME GROUPS SUBDIVIDED BY . FAMILY SIZE, MICHIGAN STATE UNIVERSITY - CONSUMER PANEL SUB-SAMPLE, 1953—I957 W Income Groups by Number of a b Change Level of Family Size Families E1 E2 in E 0 Significance Cold Cuts Less than $1200 59 14.55 5.05 .50 .025 2-3 person 114 14.06 5.29 1.23 .05 _14 or more person 145 14.70 14.97 .27 .20 $1200451700 60 5.32 5.51 .19 .20 2-3 person , 19 5.27 5.77 .50 .10 14 or more person 141 5 .314 5.39 .05 > .140 31700421400 1414 6.22 6.22 0.00 > .140 2-3 person 32 6.86 6.69 -.17 .140 14 or more person 12 14.52 14.96 .1414 .30 5.148 5.83 .35 .20 321400 and above 51 8E1 II Experditures before an increase in income. .0152 :- Experxiitures after an increase in income. cChange in E :- E2 - E1. TABLEP 96 DYMC-INJOME EXPENDITURE ELASTICITIES FUR SELECTED COLD CUTS, MICHIGAN STATE UNIVERSITY CONSUMER PANEL SUB-SAMPLE 1953-1957 1 Income Groups w 1 t Cold Cuts 11 Less than $1200 451200451700 $17004521400 $21400 am above .866 .271 0.000 .690 F151 113E CE?” '1’ {a} “‘2' t" 2.. 5‘1: 1.901 fi' 1 7911. 73th:: f-‘vggq ‘7 A ”71174114 1414414111414114114111?!“