OF NQN CIRCULATING CHECK FOR UNBOUND CIRCULATING COPY UNIVERSITY OF ILLINOIS Agricultural Experiment Station BULLETIN No. 269 THE MARKETING OF MILK IN THE CHICAGO DAIRY DISTRICT BY H. A. Ross URBANA, ILLINOIS, JUNE, 1925 FOREWORD If marketing conditions for agricultural products are to be improved, plans for improve- ment must be built on a solid foundation of ascer- tained facts, not upon opinion and guesswork. This bulletin describes many of the significant conditions surrounding the Chicago milk market, the present agencies for distribution, and the trend of consumption, with special reference to the problem of surplus milk. It is the purpose of the Agricultural Experi- ment Station to collect and publish from time to time similar facts concerning other agricultural products. It is hoped that these will constitute a sound basis for the improvement of marketing facilities. H. W. MUMFORD, Director CONTENTS PAGE INTRODUCTION 461 THE CHICAGO MILK MARKET AND MARKETING AGENCIES 463 Types of Middlemen 465 The Milk Dealers' Bottle Exchange 467 THE CHICAGO DAIRY DISTRICT 469 Regional Production of Milk and Transportation Costs 469 Factors Tending to Retard the Extension of the Chicago Dairy District 473 Intensity of Dairying in the Inner and Outer Parts of the Chicago Dairy District 478 MILK PRODUCTION IN THE CHICAGO DAIRY DISTRICT 479 Variation from Year to Year 480 Seasonal Variation in Production 483 Seasonal Production and Price 487 CONSUMPTION OF DAIRY PRODUCTS IN CHICAGO 489 Consumption in Different Sections of Chicago 490 Long-Time Trend of Consumption 494 Seasonal Variation in Consumption 495 Variation in Consumption by Days of the Week 498 Consumption of Milk as Affected by Temperature 500 Consumption of Milk as Affected by Price 503 THE PROBLEM OF SURPLUS MILK 510 Variation in Surplus from Year to Year 511 Relation Between Size of Business and Amount of Surplus 513 Seasonal Variation in Amount of Surplus 515 Daily Variation in Amount of Surplus 516 The Utilization of Surplus Milk 519 Relation Between Prices of Sweet Cream and Whole Milk 523 THE COURSE OF MILK PRICES 526 Alternative Markets for Milk 527 Price Determination 532 CONCLUSIONS 533 APPENDIX (Tables) 535 ACKNOWLEDGMENT To the many agencies that furnished the mass of data embodied herein producers' organizations, milk dealers, condenseries, creameries, cheese factories, railroads, meat packers a word of acknowledgment is due. Special mention should be made of the courtesy of several milk distributing companies that supplied most of the detailed data. These companies not only opened their books to the investigator but were also generous in giving the time both of their execu- tive and their accounting forces. To the Illinois Agricultural Association, the Experiment Station is indebted for a part of the funds for this inquiry. The Association, realizing the need for a study of milk mar- keting conditions, made provision for a half-time research assistantship in dairy marketing at the University of Illinois for the year 1922-23, and the funds so provided were used to supplement those of the Experiment Station in carrying forward this investigation. THE MARKETING OF MILK IN THE CHICAGO DAIRY DISTRICT BY H. A. Ross, Formerly Associate in Dairy Economy INTRODUCTION Almost all city milk markets at some time or other have exper- ienced serious difficulties over the question of price. Producers usually believe that the prices they receive for milk are too low, particularly in relation to the prices paid by consumers. This belief has resulted in the formation of many cooperative marketing organizations, some of which have proved successful, while others have failed. The Chicago market has not escaped the difficulties arising from price disagreements. Dissatisfaction among producers has been mani- fested by four milk strikes in the last nine years. Altho no strikes occurred during 1921 or 1922, dissatisfaction was especially strong at that time because the margin between the price paid by the consumer and the price received by the producer was unusually wide (Fig. 1). For many years the wholesale price to the producers amounted to about 40 percent of the retail price. From 1917 to 1920, wholesale prices rose more rapidly than retail prices (Fig. 2) so that the producer received about half the consumer's dollar. For the next two years, however, wholesale prices declined much more than retail prices and the farmer's share fell to about a third. Higher prices during 1923 and 1924 have brought back the ratio of earlier years. The present investigation was undertaken with the twofold aim of analyzing the Chicago milk situation and of determining some of the fundamental facts underlying milk marketing. It was thought that past failures of cooperative marketing organizations were due in many instances to lack of knowledge of the economic principles involved. Much thought has been given in the past to the form of organization but, unquestionably, too little attention has been paid to obtaining the economic facts of milk marketing upon which ^to base sound marketing practice. This study, therefore, intentionally passes over the question of organization and deals only with the specific facts of marketing. It may be stated at the outset that no all-sufficing remedy for the ills of the situation is offered. Certain principles, however, are presented, based on a careful statistical analysis of the data, which may assist in the solu- tion of the problems. The data, obtained for the most part during the year October, 1922, to September, 1923, are of two kinds: (1) survey data obtained by the 461 462 BULLETIN No. 269 [June, issuance of questionnaires to all of the milk dealers in Chicago, and (2) detailed data taken directly from the books of a few dealers. Months of work in going over ledgers and other records yielded a large amount of material for study. All this has been analyzed and the results are presented in the various tables. Because of the keen competition between milk companies in Chicago, the dealers required that many of 190X1904 J905 190 fi / o ON ^o 10 vo -so ^ *v O * * co -so o sc ^ . .j^r^r-. -co -m Tti Tf CO CN CN CN o l> o u o S '33 ffi ^ o -ocor- .-H -^ *^ *x c> r^- CN *o *o oo CO * * ^O '^ CO OO W> 15 * : : "" " : " ' W", < *X O i CM^- CN Ov O O CN 00 -OS ft, oo oc CO ^o CN cO CO CN 1^ OO 00 00 SO I-~ so Number of farms O\ co < i CN Tf o O CN *-^ ON 1^ O la a a ct -a CN CN ON ON O\ ON ON ON ON ON s ON .7925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 483 year old in 1918 was 23.3 percent less than in 1913, while the average under one year old was 34.1 percent below. The number of heifers continued to decrease even thru the period of high milk prices (1918, 1919, and 1920), and only a slight increase is shown for heifers under one year old, in 1922. It is to be expected that a certain amount of lag behind milk prices would be found, but the long continued decrease in the raising of heifers indicates other restraining factors. The very poor market for dairy cattle during 1921 and 1922 permitted the purchase of excellent cows at low prices and this undoubtedly tended to discourage the raising of young stock. It is possible that this period may mark a permanent change in the method of maintaining the herds in the inner part of the Chicago district, and that the ratio of heifers to cows will never again be as large as it was in 1916. It is quite probable, however, that the ratio will increase until it is somewhat greater than it was in 1922. The effects of the heavy production of milk in 1921 and 1922 were made worse by the demoralized condition of the condensed milk market. In October, 1920, many condenseries closed and others made drastic cuts in the prices paid for milk. Naturally, some of the milk that had previously been condensed sought the Chicago market and the existing surplus was still further increased. SEASONAL VARIATION IN PRODUCTION In addition to the variation in production from year to year which results from the combination of numerous economic factors, there is the much greater seasonal variation. This variation is an ever present problem to the milk distributor and, because of the maladjustment of supply and demand, a highly important factor in determining the prices received by the producer. The seasonal variation in the production of milk is shown by Table 9 #nd Fig. 5. The dairy district has been divided into sections in order to show some important differences between the various regions. For example, Cook and Lake counties show the least variation, with Kane and McHenry counties next. It is worthy of note that the production in these counties, which represent the most intensive district, is very little greater during the months of May and June than it is during January, February, March, and April. On the other hand, in the more distant parts of the Chicago dairy district the production during May and June is much higher than during the other months. This is particu- larly true of the Swiss-cheese region which is represented by Green and LaFayette counties. The winter production in this region is extremely low, the amount produced in December being but 44 percent of the June production. The greater seasonal variation in production in the surplus-milk region is another factor which tends to retard the extension of the Chicago dairy district and gives the producers in the inner district an 484 BULLETIN No. 269 [June, I < S! Q < o . - 52 S5 II Q cu u: E 5 I j* ' O co J> u JJ O. I c > 83 u,v O ,*K " c c C i i 1 O - o y>. c OJ JJ S -C A < i n cs cs cs cs vo m ON oo vo n T-H-^v 5; OO '-' i -OOOOOOOOOOON -> 'fior^-r^r^soo^o^ OCN a 1925} MARKETING MILK IN CHICAGO DAIRY DISTRICT 485 advantage because of their greater seasonal uniformity. It must be remembered, however, that the newer regions can adjust their pro- duction to the changed conditions which a fluid-milk market imposes. The most striking feature of the seasonal variation is the low pro- duction during the months of July, August, and September. At this Group I - Cook f Lake Counties, ////no/3 Group 2- Kan* fMc/fenry Counties. Illinois Group 3 - Will. Deftalb. Ogle. 9 Boon* Counties. Illinois Group 4 - Kenosha. Waukesha if Rock Counties. Wisconsin Group 5-Grte/i $ Lafayette Counties. Wisconsin 120 110 100 90 60 70 60 50 40 30 20 10 Jan. fth Mar. Apr. Hay June July Aug. Sept. Oct. Nov. Dec. FIG. 5. SEASONAL VARIATION IN PRODUCTION OF MILK IN DIFFERENT PARTS OF THE CHICAGO DAIRY DISTRICT The graph represents the amounts of milk sold from an average of 3,051 farms during the three-year period 1920 to 1922. The average monthly sales for thirty-six months equal 100 percent. The least seasonal variation is found in the inner district (Groups 1 and 2). Altho the variation in the Swiss-cheese region (Group 5) is very marked, the proportion of milk produced during the shortage period in the late summer is greater than for any of the other districts. Some dealers draw on this territory to supplement their supply at that time. time flies are bad, pastures are short, few dairymen have sufficient silage for summer feeding, and the fall- and winter-freshening cows are well advanced in the lactation period. If a dealer's regular supply of milk from the country is sufficient for his needs during this period, it means that his surplus is very large during the season of heavy production. In order to avoid this surplus he may depend on jobbers or other dealers for a part of his supply dur- ing the summer, but the number of companies which can follow this practice is, of course, limited. In addition to the preceding figures, which are based on detailed data taken from the books of three large companies, supplementary 486 BULLETIN No. 269 {June, figures showing the range in the seasonal production for 1922 were obtained by a survey of 64 dealers. During the season of low production (August and September) the combined daily purchases of these com- panies amounted to 1,476,978 pounds of milk. Their purchases during the season of heaviest production (May and June) totaled 2,128,086 pounds per day, or 44.1 percent more than the purchases during the shortage period. Table 10 shows the differences between the purchases for May and June and for August and September for companies of various sizes. As would be expected, the smaller companies have in general the least variation in purchases, because they get practically all their supply from the inner district, where the seasonal production is most nearly uniform. Group 4, in which are classed dealers purchasing from 20,001 to 30,000 TABLE 10.- -DlFFERENCE IN AMOUNTS OF MlLK PRODUCED DURING SEASON OF HEAVY PRODUCTION AND SEASON OF LEAST PRODUCTION Ratio of May Group Volume handled daily per dealer, May and June Number of dealers Total daily purchases, May and June, 1922 Total daily purchases, Aug. and Sept., 1922 and June pur- chases to Aug. and Sept. pur- chases Ibs. Ibs. Ibs. perct. 1 Less than 5 000 23 53 312 43 942 121 3 2 5 001 to 10 000 10 79 650 66 114 120 5 3 10 001 to 20 000 12 207 641 149 829 138 6 4 20 001 to 30 000 5 115 580 94 060 122 9 5 30 001 to 40 000 8 279 325 194 872 143.3 6 Over 40 000 6 1 392 578 928 161 150 Total... * 64 2 128 086 1 476 978 144.1 pounds of milk daily, shows, however, less variation in the seasonal purchases than does Group 3, which handles less milk. The more uniform supply of Group 4 is easily accounted for by the chance grouping of five dealers who receive from the inner district 93 percent of all the milk which they purchase. Group 3, on the other hand, includes twelve dealers, some of whom purchased milk in Wisconsin, where the seasonal production shows greater variation. The six large dealers who purchase milk from all parts of the dairy district had the greatest seasonal variation. During the flush season (May and June) they received one and one-half times as much milk as during the shortage season (August and September). In view of the large volume of milk handled by these 64 dealers it appears safe to conclude that in 1922 the May and June production in the Chicago dairy district as a whole was approximately 44 percent greater than the August and September production. This proportion varies, of course, for individual companies and in different years. A few dealers who obtain the greater part of their supply from the MARKETING MILK IN CHICAGO DAIRY DISTRICT 487 inner district have found it advantageous to establish one or more plants in the cheese districts, where spring-freshening cows predominate and where the bulk of the year's production comes during the pasture period. Altho this still further increases the May and June surplus, it supplements the short supply from the inner district during the summer months. Green and LaFayette counties, Wisconsin, (Fig. 5) are illustrations of counties which show a much higher production in July, August, and September than any of the other districts. SEASONAL PRODUCTION AND PRICE Since 1919, dealers and producers have attempted to obtain a more uniform production of milk thruout the year by changing the relative TABLE 11. SEASONAL VARIATION IN PRICE OF 3.5-PERCENT MILK AT COUNTRY STATIONS IN CHICAGO DISTRICT BY FIVE-YEAR PERIODS, 1904 TO 1923* Average monthly price for each 60-month period = 100 percent Month 1904 to 1908 1909 to 1913 1914 to 1918 1919 to 1923 January pent. 123 perct. 119 5 perct. 113 4 perct. 106 8 February 118 3 116 1 107 1 101 8 March 109.6 106.2 99.9 94.8 April 99.3 92.9 99.2 92.8 May 75.7 76.3 81.8 87.8 June 67 68.4 70.1 87.0 July... 74 9 82.3 87.1 103.0 August 89 9 91 6 98.7 110 September 97 8 95 6 99 5 102.9 October 108.8 109.5 111.2 106.0 November 115.9 118.8 f!6.0 105.1 December 119.8 122.8 116.0 102.0 Average 100.0 100.0 100.0 100.0 ''Corrected for secular trend. levels of winter and summer prices. This is shown by Table 11, which shows the seasonal variation in milk prices by five-year periods from 1904 to 1923. Until 1919, the winter price of milk was very high as compared with the spring and summer prices, and winter dairying was stimulated. The result, as previously discussed, was too much winter milk and a sharp decline in production in the late summer. Beginning with 1919, milk prices for the months of May, June, July, August, and September have been relatively much higher than before, while the winter prices have been relatively lower. This change in the relative prices for summer and winter has apparently had little effect on the seasonal production of milk in the inner part of the dairy district, as may be seen by Table 12. In spite of the greater seasonal uniformity in prices, which has been maintained since 1919, the high winter production continues. Tho this may be 488 BULLETIN No. 269 [June, TABLE 12. SEASONAL VARIATION IN AMOUNT OF MILK PRODUCED IN INNER PART OF CHICAGO DAIRY DISTRICT BY FIVE-YEAR PERIODS, 1908 TO 1922* Average monthly sales for each 60-month period = 100 percent Month Average daily sales of milk per farm 1908 to 1912 1913 to 1917 1918 to 1922 January Ibs. 395 403 401 395 412 419 324 269 292 305 326 368 perct. 110.0 112.2 111.7 110.0 114.8 116.7 90.2 74.9 81.3 84.9 90.8 102.5 Ibs. 390 406 395 391 404 417 327 283 111 290 316 355 perct. 110.1 114.6 111.5 110.4 114.1 117.7 92.3 79.8 78.2 81.9 89.2 100.2 Ibs. 391 407 419 403 415 405 310 281 287 316 343 383 perct. 107.6 112.0 115.3 110.9 114.2 111.5 85.3 77.3 79.2 86.9 94.4 105.4 February. March .- April May June July... August September October November December Averaee. . 359 100.0 354 100.0 363 100.0 "Based on sales from an average of 544 farms to seven plants in Lake, Kane, and McHenry counties, Illinois. TABLE 13. SEASONAL DISTRIBUTION OF FRESHENING Cows AS INDICATED BY RECORDS FROM FIFTEEN COW-TESTING ASSOCIATIONS IN NORTHERN ILLINOIS* Number of cows freshening each month expressed as percentages of total for year Month Group 1 Cook and Lake counties Group 2 Kane, McHenry, and DuPage counties Group 3 DeKalb, Ogle, Boone, and Winne- bago counties All January No.* 274 289 290 191 171 128 133 124 258 340 329 294 perct. 9.7 10.2 10.3 6.8 6.1 4.5 4.7 4.4 9.1 12.1 11.7 10.4 No. 407 349 370 244 217 168 205 295 443 499 438 394 perct. 10.1 8.7 9.1 6.0 5.4 4.2 5.1 7.3 11.0 12.4 10.9 9.8 No. 163 97 86 100 49 42 49 87 234 295 228 192 perct. 10.0 6.0 5.3 6.2 3.0 2.6 3.0 5.4 14.4 18.2 14.1 11.8 N. 844 735 746 535 437 338 387 506 935 1134 995 880 perct. 10.0 8.7 8.8 6.3 5.2 4.0 4.5 6.0 11.0 13.4 11.7 10.4 February March April May June July... August September October November December Total 2821 100.0 4029 100.0 1622 100.0 8472 100.0 "Acknowledgment is due Mr. C. S. Rhode, Assistant Professor of Dairy Husbandry Extension, for his courtesy in permitting the use of these records. partly due to the fact that changes in farm practice naturally lag behind changing price levels, the principal reason is the fact that summer prices are not yet high enough to induce farmers to produce summer milk in competition with crop growing in regions where a very large proportion of the land is tillable. 1925} MARKETING MILK IN CHICAGO DAIRY DISTRICT 489 The preference for fall-freshening cows in this region is clearly shown by Table 13, which gives the months of freshening of 8,472 cows during the years 1917 to 1922. It is significant that the least number of spring-freshening cows is found in DeKalb, Ogle, Boone, and Winne- bago counties, where dairying is of minor importance and crop growing of major importance. It is an open question whether it is more economical to raise sum- mer prices and lower winter prices until a more uniform seasonal production is obtained, or to draw upon the summer dairying regions of Wisconsin to supplement the short summer and fall supplies. Either course, however, is preferable to the present situation, which requires the carrying of a large surplus during the winter months, when pro- duction costs are highest. The present tendency, as has been indicated, is both to change the seasonal price levels and to draw on the summer dairies of Wisconsin. CONSUMPTION OF DAIRY PRODUCTS IN CHICAGO When the present study was undertaken, it was hoped that accurate information on the per capita consumption of milk in Chicago could be obtained. The difficulty of securing complete enough figures, however, made impossible anything but an estimate. Figures were obtained from retail routes showing the average daily sales of milk per family. The data included sales to 7,743 families in one section of Chicago and to 8,511 families in another part of the city. The first section embraces one of the better residential districts, where the people have incomes sufficient for the luxuries of life; some may be classed as wealthy, while many are prosperous business or professional men. In the second district the incomes are smaller, being derived largely from clerical and factory employment. Altho not one of the poorest sections of the city, this district ranks somewhat below the other. During January, 1923, the average daily retail sales of milk amounted to 1.65 pints per family in the wealthier section of the city. In the poorer district, the sales were almost 20 percent greater, or 1.96 pints per family. The larger consumption per family in the poorer section may be due, on the one hand, to the fact that a family with a small income exercising the greatest care in its expenditure feels that it gets full value from the purchase of milk, and on the other hand, to the number of children per family. By making proper allowances for the amount of milk and cream purchased at wholesale (Table 14), it is possible to compute the per capita consumption for the city as a whole if it is assumed that these 16,254 families are typical. Such an assumption, of course, is open to question, but under the circumstances it is the best approach to an estimate. Assuming further that these families were of average size (4.33 persons each) 1 and consumed the average amount of dairy 'United States Bureau of the Census. 490 BULLETIN No. 269 [funt, products, the daily per capita consumption of milk would be only 0.62 pounds, and of milk and cream in terms of 3.5-percent milk, 1.01 pounds. Altho the foregoing figures on the per capita consumption are some- what unsatisfactory because of the relatively small proportion of the population which is included, a very large mass of data was obtained which gives accurate information on the aggregate demands of this market. Ths following data cover the entire city and because of the large amount of business represented may be taken as correctly por- traying the demand for dairy products in Chicago. TABLE 14. RELATIVE CONSUMPTION OF CERTAIN DAIRY PRODUCTS IN CHICAGO Estimate based on the combined sales of seventy-two milk dealers'* operating ap- proximately 60 percent of the retail wagons in Chicago Product Unit Total sales per day Retail Whole- sale Total Equivalent of 3.5-percent milk Milk.. Quarts Pints Half-pts. Quarts Gallons Half-pts. Half-pts. Gallons Gallons 363 445 95 602 "4 182 71 899 29 957 26 425 435 344 125 559 26 425 4 182 14 093 120 221 9 616 5 098 671 Ibs. 935 990 134 976 14 203 8 991 121 200 400 456 45 859 271 723 51 197 perct. 47.2 6.8 .7 .5 6.1 20.2 2.3 13.7 2.5 Milk Milk Certified milk Bulk milk 14 093 23 439 Cream 22-percent 96 782 9 616 32-percent 22-percent, bulk. . . 32-percent, bulk . . . 5 098 671 Total 1 984 595 100.0 a lncluding 17 strictly retail dealers, 5 wholesale dealers, and 50 dealers doing both retail and wholesale business. The relative amounts of milk and cream consumed in Chicago are indicated by Table 14. A limited number of products are included in the table, but these comprize by far the greatest part of the milk con- sumption. It may be noted that fluid milk makes up 61.3 percent of the total consumption and that 77.5 percent of this amount is purchased in quart bottles. On the basis of the retail and wholesale sales of these 72 dealers, it may be concluded that about 29 pints of milk and 28 half-pints of "coffee" cream are purchased for each 100 quarts of milk. CONSUMPTION IN DIFFERENT SECTIONS OF CHICAGO The foregoing estimates apply to the consumption of milk in the city as a whole. As may be seen from Table 15, these proportions vary in different parts of Chicago and in the suburbs. For purposes of comparison, the city was divided into sections based on the territory served by different distributing stations. The 1925-} MARKETING MILK IN CHICAGO DAIRY DISTRICT 491 relative consumption of dairy products in a few representative sections is shown. Altho the population of the various districts is not exactly homogeneous from sociological and economic standpoints, the pre- ponderance of certain classes is sufficiently great so that the sections may be taken as typical of the various standards of living. Thus, Section A represents a district of comfortable homes and moderately large incomes, and includes some districts made- up largely of the Semitic race. Section B is a region of high rentals and high property values, and includes many apartment hotels. Section C includes a some- what mixed native and foreign population, while Section D includes the Negro district, a part of the foreign district and a small corner of TABLE 15. RELATIVE CONSUMPTION OF CERTAIN DAIRY PRODUCTS IN DIFFERENT PARTS OF CHICAGO Based on average retail sales to more than 200,000 families during the three-year period 1920 to 1922 Section of Chicago 8 Amount consumed for each 100 quarts of milk Pints of milk Half-pints of 22-percent cream Half-pints of 32-percent cream Quarts of buttermilk A.. 16.5 24.3 23.5 39.2 23.7 18.9 15.9 18.3 41.5 18.9 39.4 26.0 20.7 39.7 1.8 2.8 1.4 3.0 1.9 1.8 3.4 1.6 2.7 2.8 8.7 2.1 2.1 3.6 B C D E Oak Park Evanston All of Chicago and suburbs 28.9 28.2 2.4 3.7 *See key map, Fig. 6, for selected sections of the city, and Table F, Appendix, for monthly consumption data. the better-class white section. The great south-side residential district is represented by Section E. Oak Park and Evanston are shown as representative of the suburbs. The consumption of dairy products in these districts shows some interesting relations. For instance, it has ordinarily been supposed that a pint of milk was the usual purchase of poor families. The high pro- portion of pints sold in Section D (39.2 pints for each 100 quarts) would appear to substantiate this belief. As a matter of fact, however, the heavy sale of pints is due largely to the great factory district included in this section. During the summer, many factory workers drink milk with their lunches and the proportion of pints is thereby greatly increased. During the winter, many of the workers change to coffee or other hot drinks and the sale of pints falls off markedly. With the exception of the district just mentioned, Section B pur- chases a greater proportion of pints than any other district shown, altho it has practically no factories and comprizes a region of high property 492 BULLETIN No. 269 {.June, Drainagt Cana. values and high rentals. Far more wealth is represented here than in Section A, but the ratio of pints to quarts is 50 percent greater than in the less affluent district. In all probability, the higher consumption of milk in Section A is partly due to the large number of Semitic people who are noted as being, heavy consumers of milk. Several milk dealers have voiced the opinion that milk in pint bottles is too expensive for the poor and that the quart is their usual purchase. The data in Table 15 and the figures on the consumption per family (page 489) would tend to confirm the view, but a more detailed study than is here possible should be made before extending the above generalization to include the more needy families. The sales of cream, in comparison with the sales of quarts of milk, vary greatly in the different dis- tricts. It is not surprising that Section A, which was classed as a region of com- fortable homes but of no great wealth, consumes only 18.3 bottles of 22-percent cream for each 100 quarts of milk, while the wealthier and more fashionable Sec- tion B consumes more than twice as much (41.5 bottles). Similarly Oak Park, which has a somewhat smaller proportion of wealthy residents than Evanston, purchases only 20.7 half-pints of cream for each 100 quarts of milk, as compared with 39.7 for Evanston. Section C, which includes one of the poorer sections of Chicago, also shows relatively small purchases of cream (18.9 bottles). On the other hand, Section D, which includes the poorest district in the city and is made up of a large Negro and foreign population, purchases 39.4 half-pints of cream for each 100 quarts of milk, or practically as much as Section B and FIG. 6. MAP OF CHICAGO SHOWING SECTIONS OF CITY SELECTED FOR STUDYING CONSUMPTION OF DAIRY PRODUCTS The demand for dairy products is different in the various sections of the city because of economic and racial differences in the population. 1925} MARKETING MILK IN CHICAGO DAIRY DISTRICT 493 Evanston. In addition to the Negro and foreign districts, Section D includes one corner of a moderately prosperous white district, but since this district occupies less than one-tenth of the total area of Section D, it appears improbable that the purchases in such a relatively small area could be sufficiently great to raise the level of the whole district to the height indicated. It seems evident, therefore, that the Negro population has a taste for cream in spite of the apparent lack of wealth in the region. Similar variations in the sale of 32-percent cream were found, altho the price of "whipping" cream 1 keeps the sale of this commodity at TABLE 16. RELATIVE CONSUMPTION OF MILK AND CREAM IN CHICAGO AND SUBURBS Amount consumed per 100 quarts of milk Computed from: Pints of milk Half-pints of 22-per- cent cream Half-pints of 32-per- cent cream Retail sales of 2 large dealers; 3-year average, 1920-1922 Table 15 Table 14 28.9 28.8 28.2 27.6 2.4 2.2 Retail and wholesale sales of 72 dealers; March, 1922 Difference .1 .6 .2 about one-tenth that of "coffee" cream (22-percent fat). The three leading sections of the city in the consumption of the higher testing cream are, again, the fashionable and prosperous Section B and Evanston and the Negro and foreign Section D. Both Evanston and Section D show a decided liking for pure- culture ripened skim milk, known by the trade as buttermilk. In the suburb, the proportion is 3.6 quarts of buttermilk for each 100 quarts of milk. The Negro and foreign populations of Section D, however, are by far the largest consumers of buttermilk, purchasing 8.7 quarts for each 100 quarts of milk. The foregoing statements, relating to the consumption of milk in various sections of Chicago, are based on the average daily retail sales of two large companies over a period of three years (1920 to 1922). It is evident from Table 16 that the consumption represented by these sales is representative of the consumption of the city as a whole; that is, the relative amounts of milk and cream sold at retail by the two large dealers are practically the same as the relative amounts sold at wholesale and retail combined by the 72 dealers included in Table 14. The following data on consumption are also based on the three-year average sales of the two large companies. J Table B, Appendix. 494 BULLETIN No. 269 [June, LONG-TIME TREND OF CONSUMPTION Several very significant changes in the secular, or long-time, trend of consumption are indicated by the wholesale and retail sales of milk and cream during the years covered by these data. The annual increase or decrease in sales of each product is expressed in Table 17 in percentages of the three-year average. Since these data are based on the total sales of certain dealers, the increase during the period under study may have been due to one or more of the following factors: (1) increased per capita consumption; (2) increased population; or (3) increased business for these particular dealers at the cost of customers for other dealers. Hence, figures of the secular trend in sales are signi- TABLE 17. TREND* IN SALE OF CERTAIN DAIRY PRODUCTS FOR THE THREE-YEAR PERIOD 1920 TO 1922 Yearly increase expressed in percentages of average sales for the period Product Unit of sale Annual increase in sales Retail Wholesale Milk.. Gallons perct. perct. 1.1 1.0 12.3 68.7 4.3 19.5 3.0 14.3 Milk Quarts 6.4 20.1 ( : -9J (-.8) Milk . . Pints Milk Certified milk Half-pints Quarts Buttermilk Quarts Cream 1 8-percent Gallons 20-, 22-, 24-, 26-, 30- and 32-percent . . 22-percent 32-percent Gallons Half-pints Half-pints 7.3 3.7 All milk Bottled and bulk Bottled and bulk Bottled and bulk 7.8 8.4 C 8.0 C All cream All milk and cream b "Lines of secular trend fitted by method of least squares. b ln terms of 3.5-percent milk. c Retail and wholesale. ficant only in showing relative changes. For instance, during these three years (1920 to 1922), the annual increase in the sales of all milk by these dealers was 7.8 percent of the average of the period, but which of the three factors causing this increase was the most important can only be surmised. By comparing the percentage increases of the various products, however, very significant changes in consumption are apparent. Retail sales are based on the purchases of over 200,000 families and differences in the rate of increase are undoubtedly indicative of the trends of consumption. Wholesale sales, however, are based on pur- chases of a relatively small number of buyers, some of whom take only one grade of cream or, perhaps, only bulk milk. The transfer of their trade from one dealer to another under the existing competitive condi- .7925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 495 tions necessarily affects these figures in a way not possible in the case of the retail sales. This is particularly true in the case of bulk milk and cream, and very much less so for bottled goods because of the larger number of purchasers of milk in this form. The increase in the sale of pints of milk at retail has been three times as rapid as the increase in the sale of quarts. At the same time the sale of certified milk actually decreased. Buttermilk also failed to hold its own. Retail sales of "coffee" cream increased twice as rapidly as the sales of "whipping" cream. Of special interest is the rapid growth of the school trade in half-pints of milk. From practically nothing at the beginning of 1920, this business grew to large proportions by the end of 1922. The sale of pints of milk at wholesale also grew rapidly, the rate of increase being twelve times that of quarts or of bulk milk. SEASONAL VARIATION IN CONSUMPTION The consumption of dairy products by individuals varies consider- ably in different seasons, and this, together with the migration from the city during the summer vacation months, results in significant TABLE 18. SEASONAL VARIATION IN CONSUMPTION OF MILK AND CREAM IN CHICAGO AND SUBURBS* Based on daily retail b and wholesale sales of milk and cream equivalent to 865,000 pounds of 3.5-percent milk. Average monthly sales for 36 months (1920 to 1922) 100 percent. Month Bottled and bulk milk Bottled and bulk cream in terms of 3.5-percent milk Total milk and cream in terms of 3.5-percent milk January perct. 98.2 perct. 100.6 perct. 99.0 February 100.2 104.6 101.8 March 101.3 105.3 102.7 April 100.1 106.6 102.4 May 99.8 104.3 101.4 June 102.1 99.0 101.0 July.., 98 4 90.5 95.6 August 96.3 88.0 93.4 September 101.2 93.5 98.4 October 103.2 99.0 101.8 November 101.1 102.6 101.6 December 98.1 106.0 100.9 Average 100.0 100.0 100.0 "Corrected for secular trend. b Of the total sales included in the table, the proportionate amounts sold at retail were as follows: milk, 73 percent; cream, 56 percent; milk and cream combined, 67 percent. seasonal variations in the total consumption of milk in the city. Table 18 shows the extent of this variation. From January until June there is a somewhat irregular increase in the amount of milk consumed. With the close of school, the exodus of 496 BULLETIN No. 269 [June, people from the city begins. Sales fall off rapidly, and the low point is reached in August. The public schools open immediately following Labor Day, and a large proportion of the vacationists return at that time. The others follow shortly, and practically all are back in the city by October, when the peak of consumption is reached. During this month, 7 percent more milk is consumed than during the low month, August. Quarts of milk Pints of milk fa pints of 22 ptrant- crtam It pints afil pertent crtam Quarts of butttrmilk Jan. fib. Uar. Apr. May Junt July dug. S*pt Oct. NOK Dtc, FIG. 7. SEASONAL VARIATION IN CONSUMPTION OF CERTAIN DAIRY PRODUCTS IN CHICAGO AND SUBURBS Sales of cream and quarts of milk decrease in the summer, when people leave the city for their vacations. Sales of buttermilk and pints of milk increase, because hot weather stimulates their use as bever- ages. This graph is based on retail sales to more than 200,000 families during the three-year period 1920 to 1922. (Secular trend not removed. Average monthly sales for thirty-six months equal 100 percent.) The advent of cold weather again cuts the sales, which decline for the next two or three months. The seasonal variation in the consumption of cream is still greater. August is the low month, while the heaviest consumption comes during the winter and early spring. A fifth more cream is consumed in April than in August. The combined consumption of milk and cream expressed in terms of whole milk shows peaks in the spring and fall, with a slight decrease in mid-winter and a decided falling off during the summer. A more detailed analysis of the seasonal variation in consumption is shown by Table 19 and Fig. 7, which give the fluctuations in the retail sales of certain dairy products. During the months of May, June, 7925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 497 July, August, and September, the sale of quarts of milk is below the average of the year. The sale of pints, on the other hand, is greatest during these months. Greater use of milk as a luncheon drink by workers and lack of ice in many households explain the summer demand for pints, while the decrease in the sale of quarts is accounted for by vacations and week-end trips. The effect of this temporarily decreased population is better illustrated by the retail sale of cream. As may be seen from Table 19, the consumption of "coffee" cream (22 percent fat) falls off markedly during July and August, the sales during the latter month amounting to only 81 percent of the December peak. The consumption of "whip- ping" cream shows a much greater decrease, the sales in August amounting to only 44 percent of the April peak. It may naturally be TABLE 19. SEASONAL VARIATION IN RETAIL SALES OF CERTAIN DAIRY PRODUCTS* Average monthly sales for 36 months (1920 to 1922) = 100 percent Month Quarts milk Pints milk Half-pints 22-percent cream Half-pints 32-percent cream Quarts buttermilk Quarts certified milk January pent. 103.4 perct. 87.6 perct. 105.4 perct. 121.9 perct. 70.9 perct. 98.7 February 103.7 89.7 105.9 121.9 79.2 102.0 March 103.1 98 2 104.4 120.5 87.7 103.6 April 101.4 99 3 105 1 127.6 93.2 100.6 May 99 5 102 5 102 1 112 4 103 1 101.3 June 98 8 110 8 97 3 85 3 142 9 99.8 July... 93.7 107.3 89.4 59.9 155.3 93.7 Aucust 91.6 107.3 86.1 56.6 127.7 95.1 September 97.5 105.2 93.5 65.1 103.5 100.1 October 101.9 103.5 101.3 92.1 89.4 102.4 November 102 2 99 4 103 3 113 1 76.7 101.7 December 103 2 89 2 106 2 123 6 70.4 101.0 Average 100.0 100.0 100.0 100.0 100.0 100.0 "Corrected for secular trend and changes in the relative prices of pints and quarts. concluded that a larger number of the people who buy the more expen- sive "whipping" cream are financially able to spend at least a portion of the summer away from Chicago. Buttermilk is a popular summer drink, over twice as much of this product being consumed in July and August as is consumed in December and January. Wholesale sales (Table 20) show a seasonal variation which differs from that of retail sales. The demand for bulk milk, quarts, and pints, is highest during the summer. Sales of half-pints of milk correspond very closely to the school year, altho a certain amount of office trade gives some outlet for milk in this form during the summer vacation. The effect of the Christmas holiday on the half-pint trade is easily discernible in the December demand. The wholesale sales of cream show about the same seasonal variation as do the retail sales, altho 498 BULLETIN No. 269 [June, they are somewhat more irregular, owing probably to the greater effect of competition. In the foregoing discussion of the seasonal consumption of dairy products, the aggregate demand of the city rather than individual demand was considered. Altho the total sales of milk decline in the summer, in all probability the purchases of people remaining in the city are higher per person during this period than at any other time, but the data used here are not in a form to prove or disprove this belief. However, it is the total demand of a city or of a district, and not the per capita demand, that is of greatest interest to both producers and dealers. TABLE 20. SEASONAL VARIATION IN WHOLESALE SALES OF CERTAIN DAIRY PRODUCTS* Average monthly wholesale sales for 36 months (1920 to 1922) = 100 percent Month Gallons bulk milk Quarts milk Pints milk Half- pints milk Gallons 18-perct. cream Half- pints 22-perct. cream Half- pints 32-perct. cream January. . . perct. 91.3 perct. 94 2 perct. 84 7 perct. 119.2 perct. 93.4 perct. 96.6 perct. 108.3 February . . . March 98.6 101.8 95.4 95.7 84.7 85.4 120.8 136.3 103.5 106.7 99.0 98.8 113.3 116.6 April 99.7 94.3 89.0 120.3 105.8 103.7 128.4 May 98.8 96 9 97.7 115.9 98.6 107.7 147.4 June 103.6 106 5 117.2 90.9 92.2 102.8 109.4 July... 103.8 110.7 120.6 28.6 85.4 99.3 76.7 August 102.8 107.4 121.4 26.0 94.4 100.3 70.5 September . . October November.. . December. . . 101.1 103.0 100.0 95.5 109.7 102.4 94.6 92.2 120.8 108.6 92.6 77.3 88.6 136.6 132.7 84.1 95.2 105.4 108.0 111.4 103.0 99.6 94.1 95.1 73.6 77.7 85.9 92.2 Average. . 100.0 100.0 100.0 100.0 100.0 100.0 100.0 "Corrected for secular trend. VARIATION IN CONSUMPTION BY DAYS OF THE WEEK Not only does the consumption of dairy products vary with the season, but, as may be seen by Table 21 and Fig. 8, it also varies with the days of the week. In so far as the sale of quarts of milk is con- cerned, Sunday and Monday are the only days that show significant variations from the normal. On Sunday, many housewives purchase an extra quart of milk, while others for whom a pint a day is the usual purchase, buy a quart. The result is an increase of approximately 5 percent in the sales. Apparently, however, there is a slight tendency to overstock, because on Monday the sale of quarts is about 3 percent below normal. 1 As would be expected, pints of milk are purchased in the smallest numbers on Sunday, altho Saturday also shows a low consumption. TTie average consumption for the five days, Tuesday, Wednesday, Thursday, Friday, and Saturday, has been arbitrarily designated as normal because the Sunday and Monday purchases show the greatest variation for most dairy products. 19251 MARKETING MILK IN CHICAGO DAIRY DISTRICT 499 The chief reason, of course, is the fact that with the Sunday holiday and the Saturday half-holiday, fewer people are lunching away from home and the demand for milk as a luncheon drink is greatly decreased. The fact that they are at home accounts to some extent for the increased sale of quarts on Sunday. faroit 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 Averagt of Tut, Wtdi Thu. FrijSat 100 pirctnt Percent 300 260 240 220 200 180 160 140 120 100 80 60 40 20 FIG. 8. VARIATION IN CONSUMPTION OF CERTAIN DAIRY PRODUCTS IN CHICAGO AND SUBURBS BY DAYS OF THE WEEK Saturday and Sunday are poor days for the sale of pints of milk, but the sale of quarts on Sunday is a little above the average. Sunday is the big day for cream, especially for whipping, but the housewives tend to overstock and on Monday the sales fall below normal. The greatest variation in the consumption of dairy products is to be found in the case of cream. The Sunday sale of "coffee" cream is 54 percent above normal, while the sale of "whipping" cream on that day is three times as great as during the rest of the week. In both 500 BULLETIN No. 269 [June, cases, the tendency to overstock is shown by the subnormal consump- tion of Monday. The consumption of cream on Thursday, as shown in Table 21 and Fig. 8, is slightly exaggerated because of the large amounts that are purchased for Thanksgiving Day. Sales of "coffee" cream on that day TABLE 21. CONSUMPTION OF CERTAIN DAIRY PRODUCTS IN CHICAGO AND SUBURBS BY DAYS or THE WEEK. Based on average retail sales to more than 200,000 families during the three-year period 1920 to 1922. Average sales for the five days Tuesday, Wednesday, Thursday, Friday, and Saturday = 100 percent. Day of week Quarts milk Pints milk Half-pints 22-percent cream Half-pints 32-percent cream Pounds butter Sunday pent. 104.7 perct. 64.3 perct. 153.7 perct. 311.9 perct. 87.0 Monday 97.2 100.8 93.7 76.4 78.1 Tuesday 99 9 104 2 100 7 88 1 91.1 Wednesday .... 99 3 104 9 99 3 96 3 97 6 Thursday 100 4 104 4 103 1 107 3 90 8 Friday 100.4 105.2 97.6 93.0 95.6 Saturday 100.0 81.3 99.3 115.3 124.9 are almost double the regular Thursday sales, while six times the usual amount of "whipping" cream is required to supply the demand. As three of the 157 Thursdays which are included in the three-year period, 1920 to 1922, were marked by the heavy Thanksgiving Day sales, the average for Thursday is raised a small amount. Even if the three days of abnormal purchases are omitted, however, the Thursday consump- tion is still a little above normal. Other holidays affect sales of milk and cream also. The consumption of milk in Chicago not only increases as summer approaches, but it also fluctuates from day to day with fluctuations in temperature. The location of the city on Lake Michigan makes the temperature during the summer months subject to abrupt changes: when the wind blows from the lake, it is usually cool; a shift to the west or south ordinarily brings warmer weather and an increased demand for milk. The degree of correlation between the daily fluctuations of tem- perature and the corresponding fluctuations in the sale of milk is shown by Table 22. The correlation for retail milk is higher than for whole- sale milk. This is to be expected, because for retail milk the relation is between temperature and purchases by actual consumers, while for the wholesale milk the relation is between temperature and estimates by hotel and restaurant managers of the amount that will be consumed. 7925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 501 The extent to which temperature affects the retail sales of quarts and pints of milk and the wholesale sales of bulk milk is different in each case, as shown by Table 23. Quarts of milk are least affected, since purchases are largely for household use. The proportion of pints pur- chased by workers to drink with their lunches is greater than for TABLE 22. CORRELATION COEFFICIENTS OF TEMPERATURE AND SALES OF MILK FOR THE SUMMER PERIOD, MAY TO OCTOBER* Product b Correlation coefficient Probable error Quarts, retail +0 523 =t 038 Pints, retail +0 567 =t 035 Bulk, wholesale . . . +0.331 .046 a Based on the daily consumption of over 400,000 people during the summer of 1922. b Corrclation of the number of degrees that the maximum temperature on a given day varies from the average of the maximum temperatures of the previous seven days, and the percentage that the sales of milk on the following day deviate from the average sales of the previous seven days. Sales of milk corrected for secular trend and for seasonal and day-of-the-week variations. From a purely statistical standpoint, a more desirable correlation is that of percentage deviations of temperature from the normal and percentage deviations of sales from the expected sales, both expressed in terms of standard deviations. On this basis, the corre- lation coefficient of temperature and sales of quarts of milk is +0.547 == .026, as compared with +0.523 =*=.038 obtained by the above method. Altho a slightly higher degree of correlation is thus obtained, the method of computing the deviations is much more in- volved. Furthermore, calculations of "normal" temperatures and ''expected" sales require data on the temperature and sales for a period following the day on which the deviations are to be measured. This, of course, precludes the use of such deviations in predicting sales, since a milk dealer has only past temperatures and past sales on which to base his prediction. For this reason, the method involving deviations from previous weeks' averages has been used. quarts, and the sale of pints is therefore more affected by temperature. A large share of the bulk milk is sold to restaurants and hotels, and as a result temperature changes have a very marked effect on the sales. Altho some of this bulk milk was probably used for ice cream, the amount is relatively small, as no sales by jobbers or by dealers doing a strictly wholesale business were included. The most striking feature of the effect of temperature on sales is shown when a sudden drop of fifteen or twenty degrees occurs. The resulting decrease in consumption is very much greater than the increase which follows a sudden rise of an equal number of degrees. It appears that many people who drink milk at lunch find a hot drink more to their liking when there is a sharp fall in temperature. Since milk is very rarely used as a hot drink, they shift to coffee, tea, or cocoa, and the demand for milk is thereby decreased. On the other hand, there are many iced drinks from which to choose if the coffee- and tea-drinkers prefer a cold drink when the temperature climbs suddenly. Since milk gets only a part of this shift in demand, the increase in sales is much 502 BULLETIN No. 269 [June, less than it would be if there were no alternative cold drinks from which to choose. 1 These variations in demand may be anticipated to some extent. Since most of the milk is delivered very early in the morning during the summer, a hot or cold day affects the next day's demand rather than the sales for that day. Hence, the amount of milk that is bottled may be varied according to the temperature. TABLE 23. EFFECT OF TEMPERATURE CHANGES ON SALES OF BOTTLED AND BULK MILK DURING THE SUMMER PERIOD, MAY TO OCTOBER Deviation of maximum tem- perature on a given day from average of maxi- mum tempera- tures of previous seven days Predicted sales for the following day in percentage of average sales of previous seven days* When temperature increases When temperature decreases Quarts retail Pints retail Bulk wholesale Quarts retail Pints retail Bulk wholesale Degrees 1 100.1 100.1 100.2 100.2 100.3 100.4 100.4 100.5 100.5 100.6 100.6 100.7 100.8 100.8 100.9 101.0 101.1 101.1 101.2 101.3 100.1 100.2 100.3 100.4 100.5 100.6 100.7 100.8 100.8 100.9 101.0 101.1 101.2 101.3 101.3 101.4 101.5 101.6 101.6 101.7 100.4 100.7 101.0 101.3 101.6 101.9 102.2 102.5 102.8 103.1 103.3 103.5 103.7 103.9 104.1 104.3 104.5 104.6 104.7 104.8 99.9 99.9 99.8 99.8 99.7 99.6 99.5 99.5 99.4 99.3 99.2 99.2 99.1 99.0 98.9 98.8 98.7 98.6 98.5 98.4 99.9 99.8 99.7 99.6 99.4 99.2 99.0 98.8 98.6 98.4 98.2 98.0 97.8 97.6 97.4 97.1 96.8 96.5 96.2 95.9 99.6 99.2 98.8 98.4 98.0 97.6 97.2 * 96.8 96.4 96.0 95.5 95.0 94.5 94.0 93.5 93.0 92.5 92.0 91.5 91.0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 *From curves fitted to means of arrays by the curvilinear approximation method. Ezekiel, Mordecai: A method of handling curvilinear correlation for any number of variables. Jour. Amer. Statis. Assoc., 19, n. s. 148, December, 1924. By the use of these data, sales for the ensuing day may be pre- dicted with a considerable degree of certainty. For example, let it be assumed that a dealer has sold an average of 30,000 quarts, 10,000 pints, and 1,000 gallons of bulk milk during the last week in May. The average temperature was 70 degrees, but on Tuesday the temperature lr This suggests a somewhat neglected point toward which advertising to increase milk consumption could be directed. For many years soft drink manufacturers have stressed the refreshing effect of their beverages when consumed on a hot day. It might pay to do the same with milk. 7925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 503 falls to 55, a decrease of 15 degrees. From Table 23 it may be seen that the demand for quarts the next day will be 98.9 percent of the previous week's average; the demand for pints, 97.4 percent; and the demand for bulk milk, only 93.5 percent: or 29,670 quarts, 9,740 pints, and 935 gallons of bulk milk. These, however, are averages for the week, and to get the sales for any particular day it is necessary to multiply these figures by the corresponding percentage variations for the day of the week (Table 24). Hence, for the estimation for Wednes- day's delivery 29,670 quarts would be multiplied by 99.5 percent; 9,740 pints, by 111.9 percent; and 935 gallons of bulk, by 106.6 percent. The results, 29,522 quarts, 10,899 pints, and 997 gallons, would be the amounts the dealer could expect to sell. TABLE 24. DAY-OF-THE-WEEK CORRECTIONS FOR PREDICTING SALES OF MILK FROM TEMPERATURE CHANGES DURING SUMMER PERIOD Expressed in percentage of average sales of previous seven days Day of week Quarts retail Pints retail, May to October Bulk wholesale, May to October School period* Vacation period 1 * Sunday 103.2 97.8 99.6 99.5 100.4 100.0 99.5 101.6 98.3 100.3 99.6 100.5 100.3 99.4 59.7 109.7 111.4 111.9 112.3 113.0 82.0 80.3 98.7 99.4 106.6 107.0 101.6 106.4 Monday Tuesday Wednesday Thursday Friday Saturday Average 100.0 100.0 100.0 100.0 "From May 1 to close of public schools and from Labor Day to October 31. b From close of public schools to Labor Day. Since the correlation is not perfect, it is impossible to predict exactly the demand for milk from temperature alone, but the above tables may be used as a guide, if a reasonable margin of safety is allowed. Various factors, such as rapidly growing business, vacation migrations, holidays, and sections of the city where the day-of-the-week variations differ, would introduce errors into these computations which would need to be allowed for by the dealer using them. CONSUMPTION OF MILK AS AFFECTED BY PRICE The foregoing discussion of the consumption of dairy products is based on average figures for the three-year period 1920 to 1922. During that time, however, there were six changes in the retail price of quarts of milk and five changes in the price of pints. Quarts ranged from 12 to 16 cents, and pints from 7 to 10 cents. This naturally had some effect on the amount of milk sold. The extent to which these changes affected the sale of milk was determined as indicated on page 510. 504 BULLETIN No. 269 [funf, EKS AFTER u -t bO 4J 'o bo II -s ^ O O O -H -* ON tN 10 ** ^^ONOOO ON^^ O S O ON O OO ON O O u M H M X E* CO CS OJ *" i* lV S preceding change ^ NO o c^ 10 r^ 10 * < FORE AND ' CO 1 u 1 u ^ I i2|S ii i WEEKS BE 'o % C 60 bC M EXPECT PRICES OF CLn cL t, be J* .s g -O-C 4J cj U 3 u i ^ QONONQON CNOO O\ ^ OONONOON ON ON ON EVIATE FR( IN RETAIL s| ^ TJ* oo co r^- ON ^o co *-* !> OOONOON ON-H ON > OOONOON ONO ON OF MILK D CHANGES V . P r ^_ o v 3 O w 3 w S :. ti ex a 5 2 x * tiSiSii & * CO et t ^ a> a> 'C O CO C "5. sJONONONOONON OOt^ OO CO Id J c2 2 Ut *- u'C O CO j* w f- B5 u tr H i> COUNTS c X u TABLE 25. AJ ' ""o u Q o : : |T!*^ ""^ "^ uiiii ii i 7925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 505 X M E j -* %% U) c E .- oc ^ oow-^-o\t^- ^^ ri < en M M i5* > c*J ^-o S c o w I" 5 J^ CX3OOQO O\fS vo ^ OOOSOSO OO !-~- id H M S: 1>J* to u s % CO ss C u -5 g> l ^j wiTfOOOO >no < ^ 80880 88 8 X 01 <<0 u <-> a Q Z en .u < vo O O tS Ox vo O (d M O C T3 U ^ r-~Qcr-ia\ ori s v oo o oo o os o\ o r~- M PQ a H W 1) V*. O E ^c *J J M T3 U (N " ^; O\ t^ SO ON "- 1 CS t~- ^O ^ 1^-CNOvQO ONtS Tf X ooosoooo o\o r- M u M ac "i H S en "c en " 3 U 1 SB ^ * ^ O oo 10 vi i 10 -^ o\ vj -H OC CS ON CS O\c4 ^ O\ O\ ON OS O ON O oo " ei --. O is < ? OS & CT! e 1 ^; SO 00 O t~~ CO O C4 ^H | 8S888 ;S 8 UJ or/ g u o M EXPECTE PRICE OF < OH a u 1! -OJ3 U U ID ^! O w N < SO-^OOSO SO>^) w iJ O OS O t*> O 00 O , 2^222 ^ ^ ~ .VIATE FRO IN RETAIL J^ -a u -| ^ rJ tN O O^ u~> ON c~l so O f*5 O OS O OS Os ft. 2 2 S ^ 2 2 ^ ^ X < Q u x $ o'o 2 !* O -^ D 'u O 06 ^ en C '5. *2 Os Os Os ^5 Os Os oo r^- oo V. ^-c d seasonal M (d J C/J 'z "* D'U O en at iJw^Tfw-,'sC l ^>'* < ^tS C4 rt T3 H X OS o- cr u H "B M H ft ' UJ hi O o u u vii T so r 3 e* 2 -5 3 a ^ cs r^ o o t^ t^ to co " CO T3 u u D p. V 60 C (4 -C S -J ^j r~~ -- OOCONO^^ ONO so ^ OONONOO ONO ON :rcentage a u 60 C rt 5 1 H^ ON co i to o oo ^ ^ $ ONOOONON ONON ON 0. .8 D, 60 C aj v it* O ON ON O O ON oo s Jj O O^. ON O O ONON ON U U C s"S W O^ H O\OO ONO ON ^ OOONOO ONO ON *o 'oS "O M ^^ co ^^ *^% G> co ^H co r*^ o .2 D ) || |SSS3SS KS S 2 < P 5-c" x D, 5 a. V i-H re I CT! VI ~3 . S|, , ti *2 *s) ^tt* ^"1 \o *^* ^ ^ d fN a -a 1 P H a o- h re 8, 3 u aj c rt U i j j j j I jj ^ Date of price o * ^^ *-^ "Corrected 1925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 507 Ij.JE > a. tt 1.9 Average of 3 eeks followin price change ^ C S u .g rt O o O co O u co u i_ *S u o rt'C.S'C 3 pi a a a, a< t> CO S'c'o.S ai CL. O. ^^^ 2 t> w.y 5 PS fe a- CM ^O O " \O '- +1 1 + CS CO 10 CS O 1 +7 i i + CS CO ^^ O CO > i vo ^-1 -H ' CS + 11 + I r- 1 VO -^ CO CO CN I i +7+1+ i O CO -> ' IsO so SO SO SO VO 10 ' ^ Os CT\ CT\ O ON ON ruary tember 1922 ember - Q, U U U- -yj i, 8. E 508 BULLETIN No. 269 [/MM/-, < O' u CO X i U {2 J3 %z 1- U. i> ex E 60 c 4-1 .S O C (V ex 'ex M.a 3 cx^ J* -H' -^ OS -H TjH -^ CN I + T' +l (N I O -H 1-H CN 717+1 I + 1 1 1 . ~ CN v* O CO O **"> O co *" CO bo-* l ooJO'* r-oo 4> Z u t: u 2 rt * r^ C *r" 3 D co u'c'o.S 05 & & :=: o 2 2 u |= i> u 2 ++ I i 77 i + ++ 7 i ++ i O CN so co CO CN ' -H 111 + CN * NO ' CN O CO O > O CO ' fe O *f O CN O ^ 1^ i c. sc so so SO so so ^o ' so so . ON ON ON O ON ON oo I &!! -2l-i -?, to^ Q EL ^ 510 BULLETIN No. 269 [/"" 6 vj r-~ oo * n c>\ CT\ * 05 J3 D "3.S s a P 2 60 K, D V <# fe (^ * C-2 <^ vc -H S, tM-.tN 00 10 13 sl c o T3 s * vo O CO fN O CO >o ^S S3 g (X, ? tKS W 1 , ^ UO ^H O O tS OO % T3 C _ r T3 s- l^ It S2 *! o = 3 oo r~~ r-~ < *o r^ oc ^O O ^o O O . SO O OO CN O f~~ A O cN * C CN>J-i "* CO VO -* 1 1-- VO O\ ^ ^H VO 10 ^ E "2 * C 60 O, P W <^ CN * o O r-i -H * ^H CN VC !->. r~- . cK > i oo O oo i i :$> co \> o * -* 1 ^ rf v> rfi OS ON !7\ I-H i C3O 00 oo c\ % -3< OS ciS o U 1 T "J3 |2 -a SB >> 3 s^ CN O i O > Os -H ir> T}< 00 CN O . CO \O ^O vri co O >-. Sg> co Os r-~ >^ Os w- ~* w-i t^ o -H r~- Tt< (N tN CO r~- VO 00 O CN ecords of sales Number 2 S JJ S Q CO O cN >o oo to CN i CO VO (U to n! U J5 s n u 4-1 6 u c 3 V 13 I i i ja T R W >, n 3 o 1 it. o rt s tT _u JJ 3 f2 t a ' 2 i _>> '3 d 4? 100 O OO Q =S CN CO Tf O c S .S^ss^o rtj H ^ e company included in ( (X 1 5 i i CN CO Tf vr> vo 9 1925} MARKETING MILK IN CHICAGO DAIRY DISTRICT 515 relatively large because, unlike the small dealers, they cannot depend wholly on jobbers and other dealers in case of an emergency. It is one thing for a small dealer to buy a few hundred pounds of milk to fill a sudden shortage of 5 or 10 percent and another thing for a large dealer to find several thousand pounds to supply a similar shortage. This does not mean that the larger dealers do not buy milk from jobbers or from each other when the need arises, but the amounts available from such sources are, of course, limited. SEASONAL VARIATION IN AMOUNT OF SURPLUS Since the production of milk shows a wide seasonal variation and the consumption of milk shows somewhat less variation, it is to be TABLE 33. SEASONAL VARIATION IN AVERAGE AMOUNT OF SURPLUS MILK HANDLED BY Two COMPANIES DURING THE THREE-YEAR PERIOD 1920 TO 1922 Month COMPANY A COMPANY B Percentage surplus was of total milk purchased Percentage variation when average surplus for 36 months = 100 percent 8 Percentage surplus was of total milk purchased Percentage variation when average surplus for 36 months = 100 percent" January 33.7 36.9 38.7 39.9 45.3 43.5 30.9 25.9 18.1 20.4 20.4 28.9 104.4 122.2 132.6 137.5 170.6 162.9 87.2 64.4 40.7 48.2 49.9 79.4 17.4 20.0 25.9 32.0 42.7 42.5 33.3 30.6 23.5 18.8 12.8 14.1 64.6 74.5 100.7 131.4 198.4 195.3 123.0 105.3 77.2 58.9 32.7 37.5 February March April May June July... August September ' October November December Total 32.8 100.0 27.3 100.0 8 Corrected for secular trend. expected that the amount of surplus will also vary at different seasons of the year. Naturally this variation will depend on the year and also on the region from which any individual dealer may obtain his supply. Table 33 shows the average seasonal variation in the amount of surplus milk handled by two large companies. Company A had over four times as much surplus in May and June as it did in September. Company B, on the other hand, had about six times as much surplus in May and June as it did in November. The latter company obtained a considerable part of its supply from the Swiss-cheese region of Wis- consin, where spring-freshening cows predominate. Altho this still further increased the May and June surplus, it tended to maintain the supply during the late summer, when production- in the inner dis- 516 BULLETIN No. 269 [/', trict was lowest (page 485). As a result, this company had the least amount of surplus during November and December. Data on the amount of surplus milk were obtained from Company A, by months, for a period of seven years. During four of the seven years the heaviest surplus came in June, while for the other three years May showed the greatest surplus. September was the month of least surplus for four years, while August, October, and November each had the least surplus one year. During the three years covered by the data obtained from Com- pany B, May twice showed the greatest surplus and June once. November, December, and January each showed the least surplus dur- ing one year. From the above it may be seen that individual companies may have seasonal variations in the amount of surplus milk which do not coincide with the seasonal variations for the city as a whole. In general, however, dealers expect the least amount of surplus in the late summer and early autumn. The surplus increases gradually until the cows are turned on pasture early in May. For two months, May and June, the surplus is extremely large. In July, a combination of factors such as hot weather, flies, short pastures, and advanced lactation for the fall- freshening cows, brings a sudden drop in production which is only partly offset by the decreased demand (Table 18). The result is a decline in the amount of surplus milk, which is much more abrupt than any of the other seasonal changes. The further decline to the low point of the year is more gradual. DAILY VARIATION IN AMOUNT OF SURPLUS The reason that a certain amount of surplus milk must be carried at all times by the dealers becomes apparent from a study of the day- to-day variations. It was shown (Table 21 and Fig. 8) that the con- sumption of milk varies with the days of the week. Thus, on Sunday the demand for quarts of milk is about 5 percent above normal, while for pints it is far below normal; the demand for pints on Saturday is little better. The school trade in half-pints is entirely lacking on both Saturday and Sunday, and the sale of bulk milk also falls. The result is a decided increase in the amount of surplus milk on these days. For Company A (Table 34), the average daily surplus in 1922 amounted approximately to 37 percent of the total milk purchased for Saturday and 40 percent for Sunday, as compared with an average of approxi- mately 34 percent for the other days. In other words, even had it been possible for this company to have eliminated all surplus on the other days of the week, it would have still been compelled to handle surplus milk amounting approximately to 4.5 percent of its purchases for Satur- day and 6.9 percent for Sunday. A compensating feature of this Satur- day and Sunday surplus, however, lies in the fact that it comes at a 1925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 517 very opportune time, as the milk can be separated to meet the heavy Sunday demand for cream (Table 21 and Fig. 8). These week-end fluctuations in the amount of surplus milk are shown clearly by Fig. 10. In addition, this chart shows that there are TABLE 34. VARIATION, BY DAYS OF WEEK, IN AMOUNT OF SURPLUS MILK HANDLED BY COMPANY A IN 1922 Average daily surplus expressed in percentage of total amount of milk purchased Day of sale a Percentage surplus was of total purchased 15 Range in amount of surplus above Friday minimum Sunday 39.8 6.9 Monday 35.4 2 5 Tuesday 33 9 1 Wednesday 33.9 1.0 Thursday 33 * 6 Friday 32 9 o Saturday 37.4 4.5 a As milk is bottled the day previous to its sale, the surplus comes upon the day preceding the sale. b Simple averages. sudden violent drops in the amount of surplus milk. Thus, the amount of surplus on Sunday, June 4, 1922, amounted to 123.8 percent of the bottled and bulk milk sales. On Sunday, June 11, the surplus was only 97.9 percent, a decrease in one week amounting to over one-fourth of the bottled and bulk milk sales. Similar abrupt declines in the amount of surplus milk occurred in July, August, and September. On several occasions, a single day's decrease in surplus, other than the regular daily variation, amounted to as much as 10 percent of the fluid-milk sales. These declines are due to increases in the demand for milk or de- creases in the supply. Frequently both factors operate at the same time. Thus the drop in the amount of surplus milk which occurred in the early part of June, 1922, was due both to the increased demand and to the decreased production resulting from the unseasonable heat, the temperature at Chicago reaching a maximum of 90 degrees on June 9. The late June decline, on the other hand, was due largely to decreased production. At this time pastures were very short owing to dry weather, only 0.65 inch of rain having fallen in northern Illinois during June. The abnormal decreases occurring in July, August, September, and October, 1922, were due largely to hot weather which decreased the production and at the same time stimulated consumption. The big drop in the early part of September resulted from the combined factors of extremely hot weather (maximum temperature 98 degrees, Sep- tember 2) and the return of most of the summer vacationists for the opening of school. 518 BULLETIN No. 269 *f rt ^ <>* s o U o ^P 6 o ~ c ^ c 3 (U -- a s S .S M g . MARKETING MILK IN CHICAGO DAIRY DISTRICT 519 It is evident from the foregoing that a certain amount of surplus must be carried in order to guard against shortage arising from these sudden changes in supply and demand. Individual dealers, particularly the smaller ones, may shift the burden to jobbers, or other dealers, by depending on purchases from these sources to make up any shortage. In order to maintain a margin of safety, however, for the city as a whole, it appears that a surplus of something like 20 percent of the bottled and bulk-milk requirements should be carried during the sum- mer. Had the margin been less than this during 1922, shortages would have occurred on a number of occasions. It should be noted, however, that during 1922 the surplus handled by Company A fell as low as 20 percent only for a very short period in the late summer. Dur- ing six months of the year the surplus ranged from 50 percent to 125 percent of the bottled and bulk-milk sales. The least surplus on any one day occurred on September 11 and amounted to 9 percent of the next day's sales. Even had the total year's supply been sufficiently reduced so that on this day there was no surplus, the surplus for the rest of the year would still have amounted to 41.1 percent of the bottled and bulk- milk sales, thru the maladjustment of production and consumption. THE UTILIZATION OF SURPLUS MILK It may be seen from Table 14 that the amount of milk which must be separated to supply Chicago's daily demand for cream is approxi- mately 60 percent as great as the amount that is sold as bottled and bulk milk. Naturally, any surplus milk which a dealer has after filling the demand for bottled and bulk milk will be separated. If this surplus does not give sufficient cream to supply his trade, he can make up the deficit by contracting for daily shipments of cream from one of the more remote parts of the dairy district, or by purchases from jobbers or other dealers. If the surplus is more than sufficient to supply his cream needs, it must be utilized in a still less profitable way. Butter is the product that is most generally made. Whether the cream is sold as such or made into butter, there remains the skim milk which can be utilized in various ways. Table 35 shows the methods of utilizing surplus milk as reported by the dealers included in Table 32. Condensed and evaporated milk is ordinarily made from whole milk, while milk powder is almost always made from skim milk. American cheese may be made from whole milk during one or two months of the spring. Some dealers whose plant equipment does not include condensing pans sell their surplus to evaporated milk companies located in the immediate territory. Detailed figures on the utilization of whole milk were obtained from two companies. Tables 36 and 37 show the relative amounts of milk which were sold as bulk and bottled milk or used for other dairy products. During the season of heavy production, a large amount of butter is made from separated cream, but data on the amount used for 520 BULLETIN No. 269 {.June, M C o' f ^ ' ^ O -^ JJ PO e>c .S _^ CH rt' co C .2 S S! 'C 'C 'J < < H rJ 1 g-g U! ' o'w j en 3 S a "c ;~ "- 1 ; ; *< to S a. 89 be C > w fe ctf S !S CD ~ V O 3 6C V S5 C li r~- so cs <-H r^ * Tf N O 13 o! O ^ y Q b. C/> ID o JJ ^D ^^ ^\ i"-* *O ^* ^>l 3 ^ *T- S QQ u s VO r-5 U J CO & E "ols u CO O CN >J-i 00 U-) 3 e u < H c: ^ ^> o 'S u c 3 T: j a S c a c m 6 _c .^ 'i 'c X i jy *e3 /l O C O O o (S r^ Tf >5 c 9 1 n C O O O JS *^ ^^ o I "o a. ^S co o o o o fe "S > ^^S^Ro ~ o o. S 3 ^ 1 e^ v 5 " r<0 '"" XC 55 7925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 521 this purpose are not available. Neither was it possible to compute the amount of shrinkage for Company B, but as the loss is relatively small no deduction has been made from the amount shown as having been separated for cream. TABLE 36. UTILIZATION OF ALL MILK PURCHASED BY COMPANY A DURING THE SEVEN- YEAR PERIOD 1916 TO 1922 Disposal Percentage of total purchases 1916 1917 1918 1919 1920 1921 1922 Bottled milk 56 2 63 68.0 5.9 4.0 60.2 6.7 1 7 1.3 9 63.4 8.2 .4 .5 58.3 7.0 2.5 3.5 58.5 7.0 .3 2.8 Bulk milk 4.9 3.1 5.3 8.0 Sold to manufacturers Evaporated milk Condensed milk Cheese .7 19.6 .8 26^8 .7 2s!o .7 .2 30.7 .5 Separated for cream 34.4 1.4 23.0 .7 21.2 .9 Shrinkage Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 The relative amounts of milk utilized for various purposes are shown for each month by Figs. 11 and 12. These charts show also the great variation in the seasonal production of milk and the large propor- tion of surplus milk which was handled during 1921 and 1922. The large amount of skim milk resulting from separating the sur- plus handled by Company A was utilized. as indicated by Table 38 and TABLE 37. UTILIZATION OF ALL MILK PURCHASED BY COMPANY B DURING THE THREE-YEAR PERIOD 1920 TO 1922 Percentage of total purchases uisposai 1920 1921 1922 Bottled milk 75.9 68.0 68.8 Bulk milk 3.0 2.2 1.7 Separated for cream* 21.1 29.8 29.5 Total 100.0 100.0 100.0 "Also includes shrinkage. Fig. 13. A limited amount of skim milk was taken by bakeries and other industries, but this outlet was of minor importance. "Buttermilk" (pure culture ripened skim milk) offered a method of disposing of a certain amount, as did also cottage cheese. By far the greatest propor- tion, however, had to be utilized in other ways. The manufacture of milk powder was one of the principal methods of utilizing skim milk, but because of the low price of this product during 1921 and 1922 Company A arranged for the sale of most of the skim milk to farmers for feeding purposes. At certain times curd was sold to casein factories, 522 BULLETIN No. 269 {June, and for the first four months of 1920 a small amount of evaporated skim milk was made. | Shrinkage Condensed ewporated '$ cheese Separated for cream Sold to other dealers $ mfars. Bulk, city trade Bottled, city trade JFHAMJJASOHO JTMUJJ/ISOHO JFUAUJJASOND JFUAMJJASOHD JFMMJJASOND JfHAUJJASOND , 1916 1917 1918 1919 1920 1921 1922 FIG. 11. UTILIZATION OF MILK PURCHASED BY COMPANY A DURING THE SEVEN-YEAR PERIOD 1916 TO 1922 Note the large proportion of surplus milk during 1921 and 1922 and the wide variation in the amount of surplus in the different months. During some months less than half of the milk purchased is sold in fluid form. The rest is separated for cream or manufactured. Drawn to scale, with the month of highest purchase (May, 1922) equal to 100. Some plants are not equipped to utilize skim milk to the best ad- vantage, so that occasionally during the flush season of production large quantities are dumped in the sewers. This dumping is more com- monly practiced by companies having no country plants where the TABLE 38. UTILIZATION OF SKIM MILK OBTAINED BY SEPARATING SURPLUS MILK OF COMPANY A DURING THE SEVEN-YEAR PERIOD 1916 TO 1922 Disposal Percentage of total amount of skim milk 1916 1917 1918 1919 1920 1921 1922 Sold in Chicago 33.3 9.2 13.9 16.5 13.7 13.4 34.8 10.7 2.6 30.0 14.8 7.1 32.2 10.5 1.3 41.2 6.3 8.5 21.3 13.5 .4 52.4 6.0 6.4 8.8 11.7 5.1 49.0 16.4 3.6 5.4 6.0 10.1 54.7 15.9 9.5 3.8 4.7 8.5 54.2 24.8 7.8 Buttermilk Sold to farmers Milk powder Cheese Casein. . . . Evaporated milk Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 7925] MARKETING MILK IN CHICAGO DAIRY DISTRICT 523 whole milk can be separated and the skim milk returned to the farmer. This waste is usually decried by newspapers and is sometimes char- acterized as an attempt to destroy a part of the supply for the purpose of maintaining high prices. Waste it undoubtedly is, but with no profitable method of utilizing all the skim milk no other course is open. The low value of skim milk as compared with the high cost of bottling and distribution makes its sale on retail routes uneconomical. RELATION BETWEEN PRICES OF SWEET CREAM AND WHOLE MILK It was stated previously (page 511) that many small and some large dealers purchase a part or all of their cream from jobbers or directly from the country. The advantage of this procedure under normal condi- tions is indicated by Table 39 and Fig. 14, which show the price per gallon for 36-percent sweet cream delivered by jobbers to city plants and the cost of an equivalent amount of 3.5-percent milk purchased at country plants. In the case of the whole-milk equivalent, no allowance has been made for the cost of separating and transporting the cream, or for the value of the skim milk. The comparison is directly between the price of a gallon of 36-percent cream at the city plant and the price of 85.4 pounds of 3.5-percent milk (its equivalent) in the country. Obviously this does not give an exact comparison of the cost of cream obtained in the two different ways, but it does show the change in the relative costs that took place during the summer of 1921. During 1918, 1919, and 1920, the average cost of a gallon of 36-per- cent cream purchased as whole milk was 37 cents greater than the FIG. 12. UTILIZATION OF MILK PURCHASED BY COMPANY B DURING THE THREE- YEAR PERIOD 1920 TO 1922 The surplus during 1921 and 1922 was large for this company also, but the winter surplus was relatively less than in the case of Company A, which did not get so large a proportion of milk from the Swiss-cheese region. Drawn to scale, with the month of highest purchase (May, 1922) equal to 100. 524 BULLETIN No. 269 \_June, - o > CQ 2 U O ^^^ >. to 68 01 ,_, S o II s _ : S 5 V &' OO OO *-H c-i o) 01 cs -H 1-1 r-i c-i cs cs s