Division of Agricultural Sciences UNIVERSITY OF CALIFORNIA ECONOMIC ADJUSTMENTS FOR CALIFORNIA'S MANUFACTURED-DAIRY-PRODUCTS INDUSTRY Olan D. Forker r~5 9 4{1££T 6,1 iHrO CALIFORNIA AGRICULTURAL EXPERIMENT STATION BULLETIN 816 The dairy industry finds it necessary to adjust and readjust continually to the changing economic environment. Following the recent short-run increase in the supply of milk available for manufacturing it is almost certain that a long-run decrease in this supply, now starting gradually, will bring about profound changes before 1975. The purpose of this study is to provide information which will help those in control of the dairy industry in California to make decisions for immediate and future ad- justments. Special attention is given to: • methods of determining the probable situation in 1975; • reallocations in the distribution of market milk; • economies of scale, with details of costs for manufacturing different volumes of various products; • shifting most of the manufacturing capacity to cottage cheese and ice cream mix in preference to other manufactured products; • changes in the size and locations of manufacturing plants; and • increased efficiency of operation and reduced costs of both processing and ship- ping. CONTENTS Summary 3 Introduction 4 The industry in California 9 The industry projected to 1975 18 A case study 27 The nature of plant operations and costs 29 Analysis of variable costs at six cooperative plants in 1960 . 36 Possible short-run reallocations for six cooperative plants 45 Long-run and short-run adjustments 54 Literature cited 57 Appendix A. Analysis of the demand for Class I milk in California 59 Appendix B. Analysis of the demand for cottage cheese in California 61 Appendix C. Product yields and conversion data 62 AUGUST, 1965 The Author: Olan D. Forker is Economist in the Agricultural Extension Service and Associate on the Giannini Foundation of Agricultural Economics, Berkeley. ECONOMIC ADJUSTMENTS FOR CALIFORNIA'S MANUFACTURED- DAIRY-PRODUCTS INDUSTRY 1 Summary Projections of significant trends indi- cate that the total numbers of milk cows on California farms and the average yearly production per cow can both be increased considerably — as long as milk production is sufficiently profitable. Thus, production can undoubtedly meet the de- mand for Class I milk products, whether this is 15 or 50 per cent greater in 1975 than in 1965. Two important manufac- tured dairy products, cottage cheese and ice cream, are bulky and perishable and relatively profitable to make, and they can probably be supplied from milk pro- duced within the state for years to come. Supply and demand forecasts indicate that these will be the principal manufac- tured dairy products in California by 1975. The quantity of milk available for manufacturing other products can be ex- pected to decline slowly as the expanding population requires more land and more Class I milk products. The less perishable products — butter, nonfat dry milk, and evaporated whole milk, as well as other products not analyzed here, such as hard cheeses — are less expensive to ship and can more readily be imported to Califor- nia from other states. Relative prices and the amount of competition by other uses for California's resources will determine the timing and the balance of importa- tions of dairy products. Per unit manufacturing costs are lowest in plants of relatively large capacity. Analysis of scale economies associated with raw-product assembly and process- ing indicates that such economies would be significant for a cottage cheese-ice cream mix plant processing up to 100 million pounds of milk a year. About 23 such plants instead of the 65 manufac- turing plants in California in 1962 could supply the state's requirements for these two products under the projected condi- tions of 1975 and also could process small amounts of storable or specialty products to utilize all of the milk that might be available for manufacturing. Considerable economies might be ef- fected immediately by shipping products for which transportation rates are highest — in this case, fluid milk and cream — from the plant or plants that are nearest to each market and by concentrating manufacturing activities in plants that are nearest to the raw-product source. Of course, the consideration of any such re- allocation requires that the operations of several plants be viewed jointly, and the realization of these economies requires that these plants be controlled as if they were a single entity. A case study of the operations and costs of six cooperative firms in the San Joaquin Valley illustrates possible ap- proaches and methods in the study of short-run and long-run economic adjust- ments and indicates the magnitude of po- tential savings. In the short run the firms studied could take advantage of scale economies as well as shipping economies by consolidating some of their many ac- tivities. For the long run these firms will need to change most of their manufac- turing facilities over to the processing of the two important products — cottage cheese and ice cream — or else have con- 1 Submitted for publication September 16, 1963. [3] tracts to supply other firms with the raw ingredients for their manufacture. To achieve economic efficiency under such conditions the cooperatives will have to operate fewer plants and keep them op- erating near capacity. The next few years will be a time of transition. Foresight, skillful manage- ment, and close study of changing condi- tions will be needed to maintain efficiency in the industry and normal profits for in- dividual firms. I. INTRODUCTION This study concerns problems of the portion of the dairy industry that con- verts milk available for manufacturing into butter, nonfat dry milk, cottage cheese, ice cream, and other manufac- tured dairy products. Continual and rapid changes in the economic environment of this industry force its entrepreneurs to adjust and readjust, not only to the en- vironment but also to the adjustments being made by their competitors. Such adjustments, necessary for business sur- vival, require detailed information on the present conditions of the industry and the expected direction and magnitude of future changes. This report presents the available information and offers guide- lines as to what adjustments should or could be made. Milk production, the second-largest agricultural activity in California, re- turned to its 8,500 dairy farmers $397 million in cash farm receipts in 1962. On this basis, California ranked third among the nation's dairy states. In 1962, 40 per cent of the 8 billion pounds of whole milk produced in Cali- fornia was available for manufactured dairy products. The other 60 per cent was processed and sold as fluid milk and fluid- milk products. Scope and Objectives of the Present Study Three lines of inquiry have been fol- lowed : 1 . Analysis of the current nature of the industry and of recent changes. 2. Projection of the future nature of the industry based on the probable amount of milk available for manufac- turing in 1975 and the market require- ments for the most important manufac- tured products. 3. A case study of the marketing prob- lems of six existing dairy cooperatives, with empirical models based on two ac- tual situations, to determine what savings would be possible if the actual production were reallocated most efficiently among the six plants. The case study is limited to the feasible reduction in costs of manufacturing and handling specified products in the interests of the cooperatives. The possibilities of manipulating supply or demand or of measuring the extent of imperfections in the marketing system are beyond its scope. The ideas and methods presented are applicable to any firm with more than one plant. Milk Production in California Milk produced on California farms is of two grades. Market milk is produced under more rigid sanitary regulations than is manufacturing milk. Market milk is the only milk legally eligible for Class I use — i.e., to be sold to consumers as fluid milk, or skim, or cream, or concentrated milk, or any other fluid-milk product "which is not sterilized and packaged in hermet- ically sealed containers" (California Agri- cultural Code, 1959, sec. 4226). Large quantities of milk are produced as market milk — more than is sold for Class I uses. The extra amount is available for manu- facturing and is often called surplus mar- ket milk. Manufacturing milk is eligible for use only in manufactured dairy products. The supply of milk available for manufactur- ing is an aggregate of all manufacturing milk and all surplus market milk. Al- though the production of manufacturing- grade milk has decreased steadily for several years, the quantity of surplus mar- ket milk has increased more than enough to offset the decrease — at least through the year 1962. [4] Since early days, the manufactured- dairy-products industry has been concen- trated in the large milk-producing coun- ties of the San Joaquin and Sacramento valleys and in Humboldt County. Manu- facturing plants in the valleys have long functioned as receiving stations for fluid milk — called country plants — by as- sembling supplies from nearby producers for transportation to the more distant markets. Thus they regulate the flow of milk to market according to the demand and divert surplus milk to its most profita- ble use, within the restrictions of available manufacturing facilities and markets. Variations in the Quantities of Milk Available for Manufacturing Annual variations. The quantity of milk available for manufacturing is re- ported in figure 1 in terms of milk fat. Over the past 30 years, this quantity has varied widely from year to year. Before 1953, the variations were mostly below the 95-million-pound level; since then they have been well above that level. Al- though the quantity of skim milk available for manufacturing is not always in pro- portion to the quantity of milk fat, the two quantities rise and fall in similar patterns. Seasonal variations. Normally, milk production is low from September through February and high from April through August. Conversely, retail sales of Class I milk are low from May through July and high from September through November, not only in percentage but also in actual volume. Thus, in summer the increased amount of market milk available for manufacturing augments the naturally high supply of manufacturing milk, and in winter the greater demand for market milk reduces the naturally low supply (fig. 2). For fluid cream the yearly pattern is similar, although additional demand in summer — largely for ice cream manu- facture — somewhat offsets the drop in bulk sales of fluid whole milk and makes the surplus of market-milk fat somewhat more uniform throughout the year than the surplus of market-milk skim. Over the 10-year period 1952-1961, the seasonal variation in production of manufacturing-milk fat was relatively 1930 1935 1940 1945 1950 1955 I960 Fig. 1. Total amounts of milk fat available for manufacturing in California, 1930-1962. Data from California Crop and Livestock Reporting Service (1 961 a-1 963a). Tables 15, 23, and 26 in the report for 1960 (published in 1961) give data for 1930 through 1960. [5] O m m o (spunod uoi 1 1 ! IAI ) W.w ^I^S o m o m o m o ^ _ |^ oj ro ro co o CVJ 00 in (M CO CN *o t> r~ CD (T> Os D D '- CO £ *o o o* «+- 1— "5 A O r- "O CO .E os CD o> C — .E © »- u 5 > *■> !r o © «£ to ? » o .E 0") E t u o O D- 00 m CD m CD in CD m in CD o s E o E a ^ 2 "8 1 °£ 5 "5 1 § 3 Q o E < CN oS CO ^ (XI O 00 CO (spunod uoj | |i|AI ) W >1I!IAI CO uniform from year to year. It is repre- sented in figure 3 by a single curve, with indexes of seasonal variation calculated by the method of moving averages. How- ever, the seasonal variations in the amount of market-milk fat available for manufacturing were not uniform over time. Therefore, moving averages were calculated for four separate and over- lapping four-year periods, with similar magnitudes of annual variation during each period. All the curves peaked in March, April, or May and troughed in September or October, but the magnitude of variation between successive time peri- ods became less through the decade. Data on the total amount of milk fat available for manufacturing during the decade 1952-1961, combined in a single curve (fig. 4), indicate that, on the aver- age, there was 20 per cent more in May and 16 per cent less in October than if the total annual quantities had been dis- tributed evenly. Because of this variation, some of the processing facilities required for peak production are used to capacity only three months in a year. In future years, however, a continuation of the x 0) c Jan. Feb. Mar. Apr May June July Aug. Sept. Oct. Nov. Dec. Fig. 3. Seasonal variations in the quantities of milk fat available for manufacturing in Cali- fornia, 1952-1961. Indexes computed by using a 12-month-centered moving average of the average daily quantities for each month. 100 is the mean of the monthly averages for each index series. Data from California Crop and Livestock Reporting Service (1 951 a-1 962a). [7 Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Fig. 4. Seasonal variation in the total quantities of milk fat available for manufacturing in California, 1952-1961. Indexes computed by using a 12-month-centered moving average of the average daily quantities for each month. 100 is the mean of the monthly averages for the decade. Data from California Crop and Livestock Reporting Service (1953a-1962a). trend toward less seasonal variation in the amount of surplus market milk may diminish the need for extra processing facilities for the peak production months. Related Research Several studies on the manufactured- dairy-products industry have been made at the Giannini Foundation of Agricul- tural Economics. Hassler (1953) 2 studied pricing efficiency in this industry and con- cluded that, for the products studied, "the pricing mechanism was remarkably com- patible with a competitive system." Boles (1958) studied evaporated milk plants and developed short-run and long-run cost functions for the evaporating proc- ess. Simmons (1959) included both the fluid-milk portion and the manufac- tured-dairy-products portion of the in- dustry in his study on locational rela- tionships in the 11 western states. He concluded, among other things, that the entire production of milk in California in 1975 would be just adequate to satisfy the state's market requirements for fluid milk. Sosnick and Tinley (1960) studied the marketing problems of five of the six cooperatives discussed in the present study. Dean and McCorkle (1961) pointed out that alternative agricultural production will compete quite strongly with the dairy industry for California's agricultural resources in 1975. The various manufactured dairy prod- ucts yield different returns to the process- ing firms. Clarke, Forker, and Johnson 2 See literature cited for citations referred to in the text by author and date. [8] (1964) give estimates of the comparative net margins obtainable from alternative uses of milk available for manufacturing and evaluate the feasibility of regulating prices paid for manufacturing milk. John- son, Forker, and Clarke (1964) give esti- mates of the total costs of processing manufactured dairy products. II. THE INDUSTRY IN CALIFORNIA Discussion of the manufactured-dairy- point where the combined costs of as- products industry is limited, in this study, sembling the raw product, processing it, to plants that process butter, nonfat dry and distributing the final product are the milk, cottage cheese, unsweetened con- lowest. Two important aspects of this densed skim milk, unsweetened evapo- theory are: rated whole milk, and ice cream. This i. it explains why facilities for manu- definition excludes a small number of facturing dairy products are located near plants that manufacture only hard cheeses the point of supply. Since processing re- and specialty products. duces both weight and bulk of the raw Total output of the six products in Cali- product and also protects it from de- fornia in 1962 was as follows: terioration, it is usually cheaper and easier Butter . 39 515 793 pounds to sn *P the manufactured products than Nonfat dry milk, both to shi P an equivalent amount of fluid spray and roller milk the same distance. processes 81,454,304 pounds 2. It explains why the relatively bulky Cottage cheese curd.. 86,294,277 pounds or perishable dairy products are manu- Unsweetened con- factured in those plants that are nearest densed skim their particular markets. Thus, in 1962, (bulk) 90,818,421 pounds cottage cheese and condensed skim were Evaporated whole the most important products processed milk (canned) 193,819,194 pounds in the southern part of the San Joaquin Icecream Valley, nearer the Los Angeles market, (wholesale) 55,803,658 gallons while plants in the northern part of the valley made more nonfat dry milk and The existing structure of the industry evaporated whole milk. Cobb and Clarke is the result of economic forces and of the (i960) found a similar situation in the institutional framework, including labor New York marketing area, contracts and sanitary regulations, within which the industry operates. Structure in- Numbers and Locations of volves many dimensions. This section of Bl .- - _ . the paper will discuss it in terms of four p, o"ts Manufacturing Dairy important factors: (1) the location of Products in 1962 plants manufacturing dairy products, in The California Crop and Livestock Re- terms of the geography of California; (2) porting ServicCj in COO p er ation with the their number during a given year; (3) California State Department of Agricul- relative plant size, referring not to plant ture and the United States Depa rtment of capacity but to the quantity of each spe- Agriculture, collects a monthly report of cific product manufactured by a plant in pro duction from almost every plant in a given year; and (4) product combina- California that manufactures a dairy tions, referring to the number of different product Annual reports group these sta . products manufactured at a particular tistics for dght geographical districts, plant, the output of each, and the seasonal whose locations and output are shown in variations m this production. fieure 5 -■ , .. £ m- - . In 1962, according to the Reporting The Location Ot Manufacture Service, there were 157 separate plants The theory of location defines the tend- that manufactured one or more of the six ency to locate a processing plant at a dairy products considered in this study. [9] 1. North Coast II. North Central IV. Central Coast J 80 o o 80 o o 80 p ? 60 z 60 60 U - 40 o 40 p 40 U o *o 'o F c 20 c 20 20 f- 1 i 5 o ■ ■ u (1) D- u a. oEi iJ 12 3 4 5 12 3 4 5 6 1 2 3 4 5 6 Product Prodi JCt Product V. Sacramento Valley N-VI. Northern San S-VI. Southern San Joaquin Valley Joaquin Valley o 80 ~o 80 Products: 1. Butter 2. Nonfat dry milk 3. Cottage cheese curd 4. Unsweetened condensed skim 5. Unsweetened evaporated wholft milk 6. Ice cream (wholesale) Fig. 5. Crop-reporting districts of California and the geographical distribution of the output of six manufactured dairy products in 1962. The Northeast District, III, and the Sierra Mountain District, VII, had no manufacturing plants. Data from California Crop and Livestock Reporting Service (1963a). Butter was processed in 32 plants, non- fat dry milk (spray process) in 20, cot- tage cheese curd in 35, condensed skim in 23, evaporated whole milk in 7, and ice cream (wholesale) 3 in 103. Ninety-two of these plants processed nothing but ice cream and will be called specialized ice cream plants. The remaining 65 plants manufactured one or more of the other five specified products and will be called manufacturing plants. Figure 6 shows the locations of the 65 manufacturing plants in 1962, including 1 1 plants that made ice cream along with some other dairy product. Three areas of concentration are worth noting. The most important is in the northern portion of the San Joaquin Valley — one of the state's major production areas for manufactur- ing milk. The second — Fresno, Kings, and Tulare counties — is in one of the largest, most commercialized agricultural areas in the United States. The third — in and around metropolitan Los Angeles — is located in the major population center of the state and its major milk-consuming area. Other plants are scattered in dif- ferent parts of the state. Relative Importance of the Various Products in Different Geographical Districts Plants in the San Joaquin Valley, northern and southern regions, are the most important makers of all products except ice cream. In 1962 they made 47 per cent of the butter produced in the state and 71 to 90 per cent of the other four products. Butter is important in city plants, also — partly to utilize the fat from salvage products, especially the milk fat from returns and from wash water. The result is a fairly wide distribution of butter pro- duction, not only in the major producing areas but also in the population centers. The greatest numbers of specialized ice cream plants are in Los Angeles and Ala- meda counties, and there are several fairly large plants in San Francisco and in the Sacramento area (fig. 7). Smaller ice cream plants are distributed throughout the state. Some of the country plants pro- duce the basic ingredients of ice cream — manufacturing cream and condensed skim. They may sell these ingredients separately, or combined into ice cream mix. Actual freezing most often takes place in or near the retail centers. Changes in Numbers and Average Output of Plants, 1950 through 1962 Using 1950 as a base year, an index of annual output (fig. 8) indicates that production trends for cottage cheese and ice cream have been generally upward — even in 1952, when the quantity of milk available for manufacturing was at its lowest level. Correspondingly, the most pronounced changes in numbers of plants and in their average annual output have occurred in the production of both cottage cheese and ice cream (table 1), where plant numbers decreased about half but total volumes increased about half and the average plant size (annual volume) was roughly trebled. Trends for nonfat dry milk and for condensed skim have been generally up- ward but with large variations from year to year. Nonfat dry milk, especially, fol- lowed and exaggerated the smaller varia- tions in the quantity of skim available for manufacturing. Total manufacture of condensed skim increased about 50 per cent between 1950 and 1962, as did the manufacture of ice cream, and the trends for those two products have been similar, though condensed skim varied more widely — especially before 1955. In gen- eral, the average plant output of con- densed skim has increased without major changes in the number of plants. Butter production ranged from 52 per cent below the 1950 base to 15 per cent above it, also in relation to the curve for the quantity of milk fat available for manufacturing. In 1962, butter output almost reached the peak of 1954, although the number of plants processing butter had decreased steadily. 3 Not including the large number of retail ice cream-manufacturing plants which produce mostly the so-called soft ice cream by processing purchased ice cream mix through counter freezers. [ii] u -D O k. Ol 5 X *0 t/> £ M- I °o o "n D D U D O _ as -° -« *- b 00.O OrtOI-HNlO-HinCMSMrtN 3 03 r~.ir>ococM**-0)cot£S^HOir>M- •2-a. o"ro rtCMcororococorftminiotn E 3 1— 00 o ** ai o a> u o ~~ l_ v> Nmrvoo^oi-oirvwoooro Nio^oomnNrtHoooo CM—I^Hrt,-l-H^Hr-<^^H,-«.-l.-l 2 as ■0 ,„ B J2 JC coooinwocnoooomTtiotMoo -.- b 3 CTJ £■0. m^ 3 mpsin^-i-iivinMinuirsoioo E _as 00 c ii vo)T(-MNOO)in(<)* co" r-T o 3 j= a. CMCMCMCMCMCMCMCMCMCMCMCMCM -B O + - 5 ■o as IS o O a. > as ■£: rt-H— .rt^-HooO'— oo o) r-. 2 as "o .« Q -<-> =■ = rtrtU)t<)lDO)N010N(000) -w B ooaiONoomcoroo^oorfrt E 3 ta 3 ^rocooos^rsooowomoi 12 £& cm" co" CM cm" co" cm" cm* co* co* co* ^r ** co" ■o O ■*■* as CO B a> •o B O i_ u> o I" VOOMfOMNlfllOftNrO CMCMCMCMCMCMCMCMCMCMCMCMCM i* 2 as •o Q -*- (D-HP0«Orttff0CMrtO0>-<<0 wiorvm«3-inmfl- as ■B U as bo re "o o as -ti rtUJNcoc\jwrsin(j)-H»inin O toirsioirstcirsirsinco^a-cococo 2 as =-3 Q -+- oooHNininiHOiiDO)N>H^ — B 1COO)NOON>hO)0)-h(000* 3 CTJ S-o. 3 (O^inoOMN-HtfOOWONM CM i-T «-" CO* CO* CO CNJ* CM CM CM* CO* CO* CO* O *" ^a "co >> e 2 "o o a. 2 «s 11 ifiinminNOirtNHNNoo rtrtrtH HrtCMNN(MNCMN 2 as MfifflNOfOM B ootnrs-Hinoo«)rvoiu^(0(o 3 «J 3 N^ONOOOmOOIBOOONN ~* -«* -." —" as O *" *2 3 OQ l_ 00 as-j; rs»t- < CM mminiflinminmminuuus 0)009)0)0)010)010)0)0)0)0) E-3 (30 Q. oco Q.W) EZi >>T3 "E£ OO 1 = "-o 7 . • % } • L • I • > •• 1 # — >— \ 1-»^ \ • • • Fig. 6. Locations of California's 65 manufacturing plants in 1962. Data from California Crop and Livestock Reporting Service (unpublished). The total production of evaporated whole milk has declined consistently since 1950, whereas that of four other products has increased, and butter pro- duction is again above the 1950 level. Two evaporated milk plants closed down dur- ing 1959 and two during 1961. In the re- maining seven plants, average production in 1962 was almost as high as the 1950 average. Product Preferences Per capita consumption of individual products is relatively stable from year to year, and aggregate consumption changes in proportion to the total population, with fairly systematic seasonal variations. However, certain products have prece- dence over others in the sense that firms prefer to manufacture the most profitable. [13 ••! Fig. 7. Locations of the 92 specialized ice cream-manufacturing plants in California in 1962. Data from California Crop and Livestock Reporting Service (unpublished). Production patterns for the favored prod- ucts will be the most systematic because they are consistent with changes in de- mand. When more milk is available than is needed for the preferred products, it will be diverted to the less profitable uses. Products that are least profitable to manufacturers — e.g., butter and nonfat dry milk — are made primarily to utilize residual quantities of cream and skim after all other market requirements are satisfied. As a result, their production pat- terns amplify the magnitudes of the sea- sonal variations in milk supply. [14 Multiproduct Plants Of the 157 plants manufacturing dairy products in 1962, 16 processed two of the six dairy products and 20 processed three or more, in various combinations. Multi- product plants comprise only 23 per cent of the total number of plants, but in volume processed they represent an im- portant part of the industry. The San Joaquin Valley had 18 of the 36 multi- product plants. The greatest numbers of single-product plants were in the Southern California and Central Coast districts. ] c O v_ i E •♦— 3 — «2 = 3 E s " — aj ja E o en a e CM O CM o o CO CM 3 ' r-» o a a r~ a U3 a a o c _03 O 3 •a o w'E CO ■a B 3 O CO CT> e=>" oo CM to oo" CM CM CO oo CO 35 a. ■o b > JT &s| «o to 3 O en o a to o co a LO CO CM CM o a to o o o> -E «*• £ co" to" CO ^f Q. to i ** OS -i er cn CO a CM 1.1 o x3- a O LD cu ■«* lo" 'S- co oo" 3 o CM Cn CM **- b CO ca E 03 -Q •a e 3 o o a. Cn CO CM — < r^ r^ CD LT ^ o |i CO LO oc 00 LO oc CD LC cr 0) TJ ^ 11 CJ en en r- c CO re LO V) co-— e ** o O r-T co" ** CO ^J-" CN CM i— 1 ** u 3 ■a o j= o to 3 O -E T-t C* tn CO CL M T co S5 CM ^ a> £ 03- E a) LO •* c a ^-" Cs r-. «? 5 O 3 u *S- i r»* c co a o *a- *s- o TO CO co** 09 Q. 3 E ^j o> r-> co a o r»« lo o «* LT to £ o °r= 1— E 5 r~ cm CO CO CM CM a CO t^ o _o cn oo rt JD re > CS? CO 're "S'^Sb o CO c a r~» lo i— i ^ LT LO t= — -o o ^H CO CO V OC Cs c; r~- ^ E g+s CT> *f •* 0C 2 CO T3 E ""■ ~ "* •£ = r— lo o a o o en a C J3 C 3 re <^ •IE O to a> o~ o Q. "* CSJ r~« oo ifl CO to > ,_ E < o _o £6 E 1 a E++ 3 E CO o e '.G o o w t m o oo a- CO ** c 00 O CO oc LT to Cs oc LT to re r- oc cr U3 oc r> LP oo m CM CT r> ^*- o- CO O'JS — ' -h CV CNJ "*• LO CL cnj" cm" "^ o" > < => » a> a> re re , ■ > > .CJ E E I I .3 >> a re re c 00 ■S ° c re = JA E E ~ > 2 re re re n: S o i o CO co CJ "u- EEC E C re c a "<7 £ ai iiS a 2 CJ CJ re re — -!= -= 2 2 £ -C x: to i; = = c " a re .— O o c a c c C 2 CT c^ CJ CO C^ 2 Z 2 £ E re" o c • — =i2 CO O CJ ■o ca-z: E .S = E-S2 O-o E o astsz *-SEi2 ^=c5« = _ co ^^ re m£ aj*. *" ** .ii ° E E > <^ 5 5 « . o i: cj.ee co o j?"S-i-2-S re £ ro Sen* > a) — E »o-s .£ > o> = = >o > E E .Eg-E 3 re re S" ° « 2 re ° ECO re . E QJO 111 32 c 5 £_ — .O 3 ~ <" e E-2 o .2*- c 5 -^ E° = Eo *J < "O *- ^5 -O <= ""Li. 2°> — _. t^ C^O^feS^E Pco^grS^ OUJ E £ 0J,/ > e^°S S eS- ~o •- LO ocn uigto E 3 E e «o a> 51 .21^.2 -E t= >l E CO — F re E o E ID JT •a c co "a re E CO -. 5 03 5 o tr "re > E o ■a -* 03 re in i 5 a c QJ re re o o CO 03 to ■o JO J5 cr 03 re «,5S^ (O T3 E i 2 "ii 3 t is 100 Id use in the roducts ream m cts. Sm >resent. — > o ». o> o re a s _^ qj o ^_ ■u j= >»•-£■" a re TO— X ° E 03 M - jx: y of e n Va m mi some ese a tooth oduct F pacit oaqu crea ture e che rely us pr _Q3 5 o CO t The projected the Southern San tage cheese and i ley would manuf one plant for cott y plant devoted e all amounts of var B 3 o CL x' F F rej 03 o * 03 CJ E o.™ o = E show sufficient diseconomies to greatly offset the economies of scale in process- ing." Even where milk production was not concentrated near a plant, the dis- economies of collection were not large enough to result in total diseconomies at any plant size considered. Simmons studied hard-cheese plants, also, and ar- rived at similar conclusions. The combined cost functions for col- lecting and processing that Simmons (1959, pp. 247-255) developed for Cali- fornia conditions show continuous though decreasing economies of scale throughout the range studied, i.e., with annual fluid- milk supplies as high as 164 million pounds for evaporated milk plants, 92 million pounds for butter-powder plants, and 89 million pounds for hard-cheese plants. Plants that produce ice cream mix are similar to those that process evaporated whole milk or butter and nonfat dry milk. All have receiving operations, separators, evaporators, and storage facilities. Plants that process cottage cheese are similar in many aspects to those that process hard cheese. In the major surplus-producing dis- tricts production is concentrated heavily, and it seems unlikely that costs for col- lection of raw product up to 100 mil- lion pounds would be high enough to off- set scale economies in processing. How- ever, assembly at a single plant of milk from more than one district would result in higher per unit collection costs and would come closer to developing dis- economies in combined costs of assem- bling and processing. Thus we can assume scale economies in collection and processing at least through the 100-million-pound level for a cottage cheese-ice cream mix plant within any district. An economical plant of this size would receive annually 100 million pounds of milk for manufacturing and produce 11.48 million pounds of creamed cottage cheese and 22.38 million pounds of ice cream mix. The scale economies discussed above are possible today, but other factors limit the size of plants. The present nature of the industry has been determined, in part, by its past structure plus the past and present competitive environment. Many manufacturing plants now re- ceive more than 100 million pounds of milk a year but few of them utilize more than 100 million pounds in manufactur- ing. Some cottage cheese plants are in- stalling larger vats and machinery for moving the curd and are developing tech- niques to shorten the length of time necessary to make a vat of curd. Between 1950 and 1962, the average size of cot- tage cheese plants increased by 185 per cent while the number decreased by 43 per cent (table 1). The size of facilities for condensed skim has increased by 59 per cent. Number and location (table 8). If the input of all California plants ap- proached 100 million pounds per year, 22 plants specializing in cottage cheese and ice cream mix and one plant manu- facturing other products could process all the state's projected quantity of milk for manufacturing. The San Joaquin Valley would manufacture 88 per cent of the cottage cheese and 92 per cent of the ice cream mix; three districts would have cream and skim which could be used to make other products. Nine plants in the Southern San Joa- quin Valley District would specialize in cottage cheese and ice cream mix and ship their products locally and to south- ern California. Ten plants in the Northern San Joaquin Valley District would spe- cialize in cottage cheese and ice cream mix and would have enough churning and drying capacity among them to process the residual cream and skim into butter and nonfat dry milk. Two plants in the Sacramento Valley would specialize in cottage cheese and ice cream mix and between them would process their re- maining cream and skim into other prod- ucts such as frozen desserts, butter, and nonfat dry milk. Only in the North Coast District would one entire plant be devoted to other products in addition to the one plant specializing in cottage cheese and ice cream mix. Specialized plants to com- plete the ice cream manufacturing process would exist in or near the population centers, in addition to the 23 manufac- turing plants in the producing areas. [26] IV. A CASE STUDY The Six San Joaquin Valley Dairy Cooperatives There are six cooperative firms in the San Joaquin Valley. They handle 25 per cent of the milk produced in the valley. As part of the network of country plants they supply large quantities of fluid mar- ket milk to Los Angeles and the Bay Area. In addition, they make large quan- tities of manufactured dairy products. These six firms provided the data for this portion of the study. Their legal names and their locations from south to north are as follows: 1 . Dairymen's Cooperative Creamery Association, Tulare, Tulare County 2. Kings County Creamery Associa- tion, Inc., Lemoore, Kings County 3. Danish Creamery Association, Fresno, Fresno County 4. Los Banos Dairymen's Association, Los Banos, Merced County 5. East-West Dairymen's Association, Newman, Stanislaus County 6. Milk Producers' Association of Cen- tral California, Modesto, Stanislaus County The Danish Creamery Association op- erates a second plant, No. 3A, at Chow- chilla, Madera County — used primarily as an evaporated milk plant. Data for Plant 3A are included in this section, to give a picture of the operations of the six firms. Plant 3A is omitted from the detailed analyses in sections VI and VII because it is the only plant with canning facilities and therefore cannot be consid- ered in the reallocations. To make the distinction clear, data given for six firms cover seven plants and data for six plants do not include 3A. As a group, these cooperatives are faced with increasing costs, seasonal ex- cess capacity in some phases of process- ing, chronic excess capacity in some, and insufficient capacity in others. The five firms (all but No. 6) that handle market milk strive continually to increase Class I sales of their milk, so as to obtain higher returns for their producer-owners and higher margins of profit for their plants. Sosnick and Tinley (1960) studied the marketing problems of firms 1, 2, 3, 4, and 6. Some of the following discussion will repeat or parallel portions of their work, to make the present report com- plete. However, this report is based mainly on a special survey, late in 1960, of the operations, characteristics, and problems of the plants. All plants receive manufacturing milk and manufacture dairy products. Five plants receive mar- ket milk and make bulk shipments of Class I fluid whole milk to other fluid- milk processors. Two of these five estab- lished fluid-milk bottling operations after the study was completed. In 1959 the seven plants produced nearly one- third of the state's total output of butter, one-third of the nonfat dry milk, and more than one-fifth of the condensed skim (table 9), but relatively small pro- portions of cottage cheese and ice cream mix. Milk Supplies and the Producing Units From 1957 through 1960 the coop- eratives increased their receipts of market milk nearly 37 per cent and their share of the state's total 25 per cent (table 10). They increased their Class I sales nearly 30 per cent and their share of the Class I market nearly 25 per cent. The number of producers supplying market milk remained fairly constant dur- ing the four years, but their average pro- duction increased 35 per cent. On the other hand, the number of producers sup- plying manufacturing milk decreased one third, while their average production in- creased only 15 per cent and their total production decreased 24 per cent. Al- though the average of receipts from individual market-milk producers has remained nearly five times as great as receipts from manufacturing-milk pro- ducers and although the total amount of milk received in 1960 was greater than that in 1957, the amount available for manufacturing in 1960 was slightly less, [27 Table 9. Output of the Seven Cooperative Plants in 1959 Number of plants Output from all plants Proportion of total output Product California San Joaquin Valley 7 6 7 2 6 1 1 pounds 6,143,331 8,891,965 21,635,948 8,255,496 19,825,634 7,312,389 431,009 per cent * 30.3 33.7 3.8 22.2 2.7 1.6 per cent * Butter 67.3 43.0 Cottage cheese (creamed and uncreamed) 4.1 24 5 Unsweetened evaporated whole milk (canned) 2.9 * * Estimate not available. SOURCES OF DATA: Output figures from the six cooperative firms. Statistics based on figures of the California Crop and Live- stock Reporting Service (1960a). because of increased Class I sales. In spite of these changes, the amount of manu- facturing milk received by the coopera- tives was considerably more than the amount of market milk available to them for manufacturing. The shift from manufacturing-milk re- ceipts to market-milk receipts affects supply conditions for the plants. Most market-milk producers operate on a large scale and use bulk tanks. Most manufac- turing-milk producers have smaller op- erations and some still ship milk in cans. As receipts are shifted to market milk or as manufacturing milk is handled in bulk, management can eliminate the can-receiv- ing facilities with the attendant extra equipment and the duplicate labor and utility requirements. Possibly, also, the firms can then deal with fewer producers and thus reduce record-keeping. In the four-year period studied the total number of producers declined 25 per cent while average and total production increased, but even then manufacturing-milk pro- ducers made up more than two-thirds of the entire membership. Variations in Quantities of Milk for Manufacturing Seasonal variations in milk receipts and in Class I sales, explained in section I, make the greatest impact on capacity re- quirements at the plants. During the four-year period studied the quantity of milk fat available for manufacturing by the cooperatives was highest in July, 1957, and lowest in October, 1958 (fig. 10). The difference was so great that capacity for handling the July 1957 supply would have to be 1.86 times that needed in October, 1958. The quantity of milk received is rela- tively constant from day to day. On the other hand, the rather extreme day-to-day variations in sales of fluid whole milk cause variations in the quantity of milk available for manufacturing. Figure 11 shows the more or less sys- tematic weekly variation in out-of-area Class I shipments, related to the pattern of consumer purchases. Retail stores usually stock heavily on Friday and Satur- day to meet the week-end demand. Allow- ing time for processing and transporta- tion, Wednesday and Thursday shipments of fluid milk from the plants are the larg- est; Saturday shipments are the smallest. On Thursday, October 16, 1958, when the cooperatives made the month's larg- est shipments of market milk, there were only 30,966 pounds of milk fat available for manufacturing. Near the other ex- treme, 75,756 pounds became available on Saturday, July 4, 1959 — nearly 2.5 times as much as on October 16. Such wide variations in supply create problems of adjustment and storage and a need for some additional manufacturing capacity. Although milk can be stored for only about three days before processing, manu- facturing activity can be spread out over the week in a more or less uniform man- ner. The real problem is that the total and [28] average amounts of milk for manufac- turing are so much higher throughout the V. THE NATURE OF PLANT Stages of Operation The total operation of a dairy-products processing plant may be analyzed by stages, as shown in figure 12. Each stage consists of all the productive services involved in a single operation or in a group of closely related operations in the manufacture of one basic product. Some of the basic products, shown as output in the diagram, may be combined to make additional products, e.g., creamed cottage cheese and ice cream mix. Nonfat dry milk and sweet butter can be reconstituted and used in making cottage cheese and ice cream. entire month of July than they are in Oc- tober. OPERATIONS AND COSTS Butter-making and cottage cheese- making are batch processes. Output is dis- continuous and its rate is measured in units per churn or per set. Each of the other stages is a continuous-flow process, and output rate is normally measured in units per hour. For all practical purposes the output rate is constant, because each piece of equipment has a particular rate of operation that minimizes costs. How- ever, under certain limiting conditions, when the daily market requirements ex- ceed the quantity that can be produced at this rate within the operating day, it may be profitable to increase the operating rate. Table 10. Milk Supplies of the Six Cooperative Firms, 1957-1960 Factor 1957 1958 1959 1960 Market milk Manufacturing milk Total Utilization Class I sales Market milk available for manufacturing All milk available for manufacturing. . . . Number of producers Market milk Manufacturing milk All producers Average supply per producer Market milk Manufacturing milk All milk Share of California total Market milk receipts Manufacturing milk receipts Class I sales Milk available for manufacturing 6.8 19.5 5.3 17.5 thousand pounds of fat 14,269 12,956 15,091 11,104 17,202 9,900 19,500 9,844 27,225 9,156 5,113 18,069 26,195 10,193 4,898 16,002 27,102 10,469 6,733 16,633 29,344 11,876 7,624 17,468 number 428 1,599 429 1,321 438 1,119 432 1,054 2,027 1,750 1,557 1,486 pounds of fat 33,338 8,103 13,431 35,177 8,406 14,969 39,274 8,847 17,407 45,138 9,340 19,747 per cent in terms of milk fat 7.0 19.3 5.8 16.4 7.5 18.3 5.8 15.9 8.5 16.7 6.6 16.0 SOURCES OF DATA: Supply figures from the six cooperative firms. Proportions based on figures of the California Crop and Livestock Reporting Service (1958a-1961a). [29] "* ! >"N S> ^^ - \ - \ # 1 \ > "v > * Z / available : or manufacturing 1 1 1 1 1 1 1 Market milk fal 1 1 1 1 1 1 3 U \ o - § ^~ "5 E ^ *• >v E ^ o / / — \5 (S ••* -* ....•"' E • .E *•• '^- • -i / i Manufacti 1 ^ .* " */> in . o> O) D c c 'Z. o 3 E ■G _ O c en in ^-5 5 a CD o o "a m- a> *- O D Q ♦- o • M- O ^ ^ Os 00 m E J m hs *t: ""> O Os *"" fl) V I I c *- o <*= O.I +- d E •- o a .2" 8 U. u N- U) 0^ o o o o CO m (spunod puDsnom) jdj >mi/\| in T3 C O Q. T3 C o to O 860 752 645 537 429 322 215 108 '•:'•','• Wed. 8 Wed. 15 Wed. 22 Wed. 29 Wed. 03 O 645 — July 1959 __ — 537 __ — :•:• •:■: : : : ; ;t: 429 — : : : • £• — m ;•:• ;•:• JTT : : : T^! 32? : : : ::": T-! J-^k- J-ji — - : : : : *: : :• ■ ;•:■ ;•:« T^l _ F :: «:•:• --. 215 :•:• : : : S!i| : ■:•: :•:::■ ':':'•': :i — !••• !■ : :'■: J«J 108 n ••;: :): ':':■': ;.;. • |:j in I Wed. 8 Wed. 15 Wed. 22 Wed. 29 Wed. Fig. 11. Out-of-area shipments of fluid market milk from five cooperative plants. Plant 6 did not handle market milk. Data supplied by plant managers. Variable Operating Costs Variation in the quantity of a product processed on any one day changes the total plant costs and usually changes also the per unit cost of processing the par- ticular product. Decisions that cause changes in volume processed must there- fore take into account the associated im- pact on costs. Costs incurred only on days when an operation is started or that vary system- atically with the time of operating or the quantity processed are considered variable costs. As these costs are the only ones in- volved in short-run reallocation decisions, it is important to isolate the relevant vari- able-cost factors and to determine the re- lationship between these and the volume processed. The actual variable costs of producing different volumes of each product at each plant are the basis for re- allocation and are considered in sections VI and VII. Downtime. On a given day the opera- tion of any stage requires that equipment be set up, operated long enough to yield a specified output, and cleaned in place or dismantled and cleaned. The time re- quired to assemble, service, dismantle, [31 Whole milk 4 Receiving Storage O- — £) Fluid whole milk A Separation and pasteurization 40% Storage r , ;: Cream Churn- ing o~ Butter Skim 5> A Cottage cheese- making o~ A Storage L i 2. Evapo- rating a condensing o~ 5 O Dr y inq ^ 6 Cottage Condensed Nonfat Skim cheese skim dry milk curd milk Fig. 12. General flow of milk through the processing stages in a dairy manufacturing plant. and clean the equipment — called down- time — is not directly productive; only the operating time yields any product. The total cost of labor, utilities, and cleaning agents used in downtime is the same for any given day regardless of the volume processed. However, these costs are con- sidered variable costs because they are not incurred when the particular stage is not operated. Figure 13A illustrates the effect of volume variation on costs in a hypotheti- cal processing stage that is operated by one man, requires four hours of downtime per day, and has an output of one unit per hour of operating time. The capacity of such a stage is 20 units of output per 24-hour day. Utility costs per unit of output may be constant during the operating time, as [32] specified for this hypothetical stage (slope of line segment mn). In some other in- stances they may either increase or de- crease when larger volumes are processed. For this hypothetical stage, trace mn rep- resents the total variable cost of utilities at various units of output. Labor costs are discontinuous, be- cause of institutional arrangements, and do not vary proportionately with volume changes. For all practical purposes labor is employed in eight-hour increments. Wages for the first shift plus utility costs are the basic input costs for volumes of output up to four units per day. How- ever, for volumes over four units per day, management has the alternatives of using overtime labor at time-and-a-half or of hiring a second full shift. In the hypo- thetical case it is cheaper to employ over- time labor for volumes between 4.0 and 9.3 units per day (trace tv compared to trace uv). For volumes between 9.3 units and the two-shift capacity of 1 2 units it is cheaper to employ a second shift. The situation is similar for processing volumes from 12 to 17.3 to 20 units. This explanation assumes a very simple situation. When an operation involves more than one man, management may stagger the arrival and departure times of individuals on a shift so that the right numbers of men will be present when needed during the various phases of processing. If management can organize its labor force so as to use overtime labor when it is most economical, then additional pro- duction in any day usually results in lower per unit variable costs, even with over- time wages. If overtime cannot be used effectively, total variable costs increase sharply at any point where an additional shift is required. In such a case, the per unit cost of producing five units (fig. 13B, cost Oe) may be greater than that for four units (Od). Lowest per unit costs (fig. 13B, cost 0c) occur at the capacity operation of 20 units per day, because this makes full use of regular shifts and spreads the costs for downtime over the maximum output. Conversely, it is costly to operate any stage at a low output. For producing two units, the per unit cost is Of. Thus, it would be efficient to combine a low-out- put stage at one plant with the same stage in another plant. Cost differences occur among plants for many different reasons. In fact, the cost functions for a given stage would be the same at two plants only if each manu- factured the same volumes of the same products with the same types of equip- ment and paid the same prices for utili- ties and all other inputs. For example, cost differences can exist between plants on the basis of size of equipment. If the dryer in Plant A can process 1,200 units of nonfat dry milk per hour with only one man on duty and the dryer in Plant B can process only 600 pounds per hour, and if both dryers are operated at capacity, the per unit labor cost for the operating time at Plant B would be twice that at Plant A. Prices paid for utilities are also a source of cost differences. The rates charged depend on the quantities used during a month; the quantities used at any plant depend on the amount of fluid milk handled and the amounts and relative pro- portions of all the products manufactured. Thus, when the price of utilities used at one stage is a function of the quantities used at other stages, there is a certain de- gree of interdependence among the stages. The method of computing rates for the utilities is explained in section VI. Nonvariable Plant Costs Overhead plant costs, which are in- curred whether or not the various stages are operated, are considered nonvariable or fixed. In the study that follows we assume that each plant will continue to receive its same quantity of milk regardless of the proposed alternative allocations in its dis- tributing and manufacturing activities. Therefore, the cost of operating the re- ceiving stage is considered nonvariable (except for the possible effect of variation in utility rates) because the volume of ac- tivity at this stage is not affected by any reallocations. On this basis the present study ex- cludes nonvariable costs as follows: ad- [33] c/) o o jo o o "D £b 2 8 10 12 14 16 Hours of operation J I I I L 20 22 24 2 4 6 8 10 12 Units processed 14 16 18 20 B •*— to o o a> o 'of > i\ -t= e c Q. C -l-\ 1 _ j N^ 1 | 1 1 1 i III! 2 4 8 12 Units processed 16 20 ministration and management; a basic plant crew and many of the overhead labor force; plant heating and lights; labo- costs of plant operations. The basic crew ratory, storage, and warehouse costs; de- is the labor that does not vary as the preciation on buildings and equipment; product allocations are altered. Johnson, repairs and maintenance; and insurance Forker, and Clarke (1964) give further and taxes. information on total processing costs, The quantity of milk received at each both fixed and variable, for the manu- plant determines in large part its basic factured-dairy-products industry. Fig. 13. A. Accumulation of variable-cost components for a hypothetical processing stage. Although the four hours of downtime are shown as a block at the start of the operation, the teardown and cleaning part of this segment comes at the end of the last shift of each day and is done by first-shift labor only when there is no later shift. In illustrating a day's costs, the full cost of downtime labor and utilities must always be counted, however small the output. The cost of processing a specified quantity of product is indicated at the intersection of the cost trace (bstvwyz) with the vertical axis (ordinate) corresponding to the number of hours of operation or the number of units required. Alternative costs for labor between 8 and 13.3 hours and between 16 and 21.3 hours are shown at overtime rates (tv and wy) and also at full-shift rates (uv and xy). B. Per unit variable costs — highest for one unit, lowest for 20 units. The smooth curve represents all variable costs when overtime labor can be used for greatest economies; the upper traces show the discontinuity in costs when full shifts are used instead of overtime labor. The scale of one cost unit on the vertical axis in diagram B is three times that in diagram A. [35] VI. ANALYSIS OF VARIABLE COSTS AT SIX COOPERATIVE PLANTS IN 1960 This section of the paper presents the cost factors that are variable when the manufacturing activity at any plant is altered. These costs are analyzed as a basis for planning shifts in productive activity among the plants, so as to minimize costs for all. It is specified that these six plants would continue to exist, that each would receive its former volume of milk, and that the out-of-area market for each prod- uct would remain the same. The method of cost synthesis, often called the building-block technique, is used rather than an accounting method for developing the variable-cost estimates. In this technique, the first step is to calcu- late input-output relationships — the kinds of cost factors (labor, utilities, supplies) and the quantity of each required for a specified output. The second step is to find the appropriate cost rates and apply them to the input factors, to arrive at a total cost estimate. Table 11 presents daily processing capacities of the seven cooperative plants in terms of raw-product input, so that outputs of different manufactured prod- ucts can be compared or aggregated in skim-equivalent or cream-equivalent units. These capacities are based on a 24-hour operating day, with appropriate adjust- ments for downtime. Appendix C gives quantitative relationships between the raw products and the manufactured products. Six plants were studied late in 1960. Plant managers and superintendents sup- plied data for determining hours of labor. Engineering specifications of the actual equipment, data from engineering manu- als and from technical journals, and the plant personnel's knowledge of each op- eration provided a basis for calculating input requirements for the utilities. Labor Requirements. A labor schedule, based on one, two, and three shifts, was calculated for all productive operations at each plant, to determine the volumes over which it would be more economical to use overtime labor than to hire an ad- ditional shift. Tables 12-16 show the re- lationship between amounts processed and variable labor requirements for five proc- essing stages at six plants, on the basis of the existing equipment. For example, the Table 11. Operating Capacities of Cooperative Plants, 1960 Amounts of milk or cream that could be processed in a 24-hour operating day, allowing for downtime Processing stage Raw product Plant number 1 2 3 3A 4 5 6 Whole milk Cream Skim Skim Skim pounds of raw product Separation and pasteur- ization 510,000 24,516 324,000 266,200 561,000 40,860 70,000 119,808 952,000 76,630 486,000 479,196 612,000 288,000 179,694 561,000 34,328 240,000* 263, 454 f 408,000 36,450 120,000 285.000 119,808 714 000 Churning 81,720 426 000 Cottage cheese-making. . Evaporating and con- densing* Evaporating and dryingf. 210,767 * Evaporation specified at an input-output ratio of 3:1. t Drying is independent of the condensing process only at Plants 2 and 5. At the other plants all milk is condensed before drying, and this uses up condensing capacity. For example, if Plant 1 makes 266,200 pounds of skim into powder, only 57,800 pounds of capacity remain for production of condensed skim. This assumes that the dryers at Plants 1, 3, 3A, and 6 receive 3:1 condensed skim from the evaporators. At Plant 4 the drying capability of the dryer is comparatively greater than that of the evaporator. Drying the maximum capacity at Plant 4 would involve coordinating the evaporating and drying operations so that the evaporator would remove less moisture and thus handle a larger input than that indicated for the 3:1 ratio. SOURCE OF DATA: Figures from plant managers and superintendents. [36] Table 12. Variable Labor Required for Separation and Pasteurization Six cooperative plants, 1960 Plant number Input, whole milk Day labor Night labor Overtime labor hundred pounds per day 0- 675 man- per hours day* man-hours per hundred poundsf 1 8 675-2,358 8 0.0033 2,358-3,075 12 4 3,075-4,521 12 4 0.0039 4,521-5,100 12 12 2 0- 880 8 880-2,328 8 0.0038 2,328-2,970 12 4 2,970-4,832 12 4 0.0030 4,832-5,610 12 12 3 0-2,240 8 2,240-4,224 8 0.0028 4,224-5,040 12 4 5,040-8,173 12 4 0.0018 8,173-9,520 12 12 4 0- 880 8 880-2,328 8 0.0038 2,328-2,970 12 4 2,970-4,832 12 4 0.0030 4,832-5,610 12 12 5 0- 960 8 960-1,802 8 0.0066 1,802-2,160 12 4 2,160-3,536 12 4 0.0041 3,536-4,080 12 12 6 0-1,680 8 1,680-3,141 8 0.0038 3,141-3,780 12 4 3,780-6,232 12 4 0.0023 6,232-7,140 12 12 * Applies to the entire range, t Applies to amounts in excess of the lower volume. SOURCE OF DATA: Calculations based on information from plant managers and superintendents. manufacture of 20,000 pounds of cream into butter at Plant 1 requires 1 2 hours of day labor, 4 hours of night labor, and 2.1 hours of overtime [(200 - 184) x 0.1305]. At Plant 5, 24,000 pounds of cream can be processed in two shifts without over- time. Wages. Wages are specified by union contract. 7 All six firms had the same con- tract in 1960. All labor in this study is in Bracket I, which includes receivers, testers, butter and cottage cheese makers, ice cream mix makers and mixers, and cold- room men. The effective wages for Bracket I labor consisted of a basic hourly Table 13. Variable Labor Required for the Churning Stage Six cooperative plants, 1960 Plant number Input, 40% cream Day labor Night labor Overtime labor hundred pounds per day 0- 82 82-153 man- per hours jay* man-hours per hundred poundsf 1 8 8 0.0782 153-184 12 4 184-227 12 4 0.1305 227-245 12 12 2 0-61 61-104 8 8 0.1305 104-245 24 8 245-360 24 8 0.0978 360-409 24 24 3 0-460 460-766 32 40 4 0- 98 8 98-200 8 0.0543 200-245 12 4 245-314 12 4 0.0815 314-343 12 12 5 0- 96 8 96-196 8 0.0554 196-240 12 4 240-328 12 4 0.0644 328-365 12 12 6 0-245 16 o 0.0489 245-472 16 472-572 24 8 572-745 24 8 0.0652 745-817 24 24 * Applies to the entire range. f Applies to amounts in excess of the lower volume. SOURCE OF DATA: Calculations based on information from plant managers and superintendents. 7 Manufacturing-milk-products agreement between Locals affiliated with the International Brother- hood of Teamsters, Chauffeurs, Warehousemen, and Helpers of America, approved by Western States Dairy Employers' Council, Joint Council 38, and Milk Products Manufacturers' Association of California, effective January 1, 1960. [37 Table 14. Variable Labor Required for Cottage Cheese-Making Two cooperative plants, 1960 Plant number Input, skim milk Day labor Night labor Overtime labor hundred pounds per day 0-400 400-678* 678-700 0-500 500-847 t 847-1,200 man-hours per day* man-hours per hundred poundsf 2 16 16 24 16 16 20 20 5 0.0040 0.0032 * Applies to the entire range. t Applies to amounts in excess of the lower volume. X This figure does not imply that a fraction of a vat is proc- essed on any one day. Instead, it refers to a breaking point over an average of several days' production. SOURCE OF DATA: Calculations based on information from plant managers and superintendents. Table 15. Variable Labor Required for Condensing Skim (Evaporating Stage) Five cooperative plants, 1960 Plant number Input, skim milk Day labor Night labor Overtime labor hundred pounds per day 0- 720 720-1,470 1,470-1.800 1.800-2.815 2,815-3,240 0-1,080 1.080-2.580 2.580-3,240 3,240-4,382 4,382-4,860 0- 480 480-1,147 1,147-1.440 1,440-2.117 2,117-2,400 0- 450 450-1.283 1,283-1.650 1.650-2.496 2,496-2,850 0- 852 852-2.035 2.035-2,556 2,556-3,757 3,757-4,260 man-hours per day* man-hours per hundred poundsf 1 3 4 5 6 8 8 12 12 12 8 8 12 12 12 8 8 12 12 12 8 8 12 12 12 8 8 12 12 12 4 4 12 4 4 12 4 4 12 4 4 12 4 4 12 0.0074 0.0055 0.0037 0.0049 0.0083 0.0083 0.0066 0.0066 0.0046 0.0046 * Applies to the entire range, t Applies to amounts in excess of the lower volume. SOURCE OF DATA: Calculations based on information from plant managers and superintendents. rate of $2.66 plus fringe benefits as fol- lows: 1 . Welfare and retirement benefits (based on 160 hours' work) Welfare plan $10.30 Retirement plan 17.30 Uniforms and licenses 10.00 Total, per month $37.60 2. Vacation and holidays Days worked 243 Vacation days 10 Paid holidays 8 Days paid, per year 261 3. Social Security payments by the em- ployer. In 1960 these were 3 per cent of the first $400 earned by each worker in any month, with a maximum payment of $ 1 2 per month for each worker. Although some workers earned more than $400 a month, some earned less, so the em- ployer's cost is estimated at 3 per cent of the adjusted regular-time rate, to include vacation and holiday pay at the basic rate but to exclude overtime pay. The labor cost per man-hour for Bracket I labor was computed as follows : Basic time rate $2.66 Vacation and holiday factor (261/243) x 1.074 Adjusted hourly rate, for regular time only $2,857 Social Security 0.086 Welfare and retirement benefits 0.235 Effective hourly rate, for regular time only $3,178 One firm's payroll accounts showed that annual payments for fringe benefits were 16.5 per cent of the total payroll. The above result is 19.5 per cent over the regular straight-time rate, but the per- centage would be less if the total payroll included much overtime, which is calcu- lated without fringe benefits. Labor beyond eight hours in any day received one and a half times the regular hourly rate. In addition, men working be- tween 6:00 p.m. and 6:00 a.m. were paid a night bonus of 10 cents an hour. Effec- [38] Table 16. Variable Labor Required for Drying Skim Milk* Six cooperative plants, 1960 Plant number Input, skim milk Day labor Night labor Overtime labor hundred pounds per day 0- 533 man- per hours Jayt man-hours per hundred pounds! 1 16 533- 720 16 0.0075 720-1,272 16 0.0149 1,272-1,470 20 4 0.0074 1,470-1,597 24 8 1,597-1,800 24 8 0.0075 1,800-2,348 24 8 0.0130 2,348-2,815 24 16 0.0055 2,815-3,662 24 24 2 0- 266 8 266- 636 8 0.0150 636- 799 12 4 799-1,080 12 4 0.0200 1,080-1,198 12 12 3 0-1,065 24 1,065-1,080 24 0.0085 1,080-1,864 24 0.0122 1,864-2,580 32 8 0.0037 2,580-2,928 36 12 2,928-3,240 36 12 0.0085 3,240-4,242 36 12 0.0134 4,242-4,382 36 28 0.0049 4,382-4,792 36 36 4 0- 555 16 555-1,228 16 0.0164 1,228-1,525 24 8 1,525-2.307 24 8 0.0144 2,307-2,635 24 24 5 0- 262 8 262- 636 8 0.0150 636- 799 12 4 799-1.080 12 4 0.0200 1,080-1,198 12 12 6 0- 444 16 444- 852 16 0.0103 852- 983 16 0.0149 983-1,220 20 4 0.0046 1,220-1,846 20 4 0.0137 1,846-2,035 20 12 0.0046 2,035-2,107 24 16 * One process at plants 2 and 5. Man-hours at the other plants include labor required for separate evaporating. t Applies to the entire range. | Applies to amounts in excess of the lower volume. SOURCE OF DATA: Calculations based on information from plant managers and superintendents. tive hourly wage rates for 1960 were as follows: Straight time Day $3,178 Night 3.278 Overtime (without fringe benefits) Day 3.990 Night 4.090 Regular rates were paid for regular shifts on Saturday and Sunday. Overtime rates were paid only when a man worked more than eight hours in a day or more than 40 hours in a week. Electricity Requirements. The amount of elec- tricity required for each installed motor at each stage was calculated from its horsepower specifications. Horsepower requirements were converted to kilowatt- hours (kwh) at the normally accepted rate of a gross input of 1 kwh of energy per horsepower per operating hour. This allows for inefficiencies in the motor and in the system. Power requirements for a mechanical refrigeration system vary considerably with operating conditions, including the condition of the equipment. An adequate estimate for milk-plant operations — de- termined from engineering data (Farrall, 1953) — is that 1 ton of refrigeration re- quires 1 motor horsepower, roughly equivalent to 1 kilowatt of electrical energy, again allowing for motor and sys- tem inefficiencies. Assuming a specific heat of 1, the cooling of 1 pound of milk through a temperature range of (t x - t 2 ) degrees Fahrenheit involves refrigeration at the rate of (f x - t 2 )/\ 2,000 tons per hour, which can be converted to elec- trical requirements by the formula: kwh = (f 1 -f s )/12 > 000. Table 17 gives estimates of the elec- trical requirements for each stage of op- eration at each plant: those associated with downtime, which are constant for each day of operation, and those for processing, in terms of the quantities of cream or skim processed. For example, Plant 5 requires 24 + (200 x 1.383) kwh of electricity to process 20,000 pounds of skim into nearly 6,000 pounds of con- densed skim. [39] Table 17. Amounts of Electricity Required for the Processing Stages'* Six cooperative plants, 1960 Stage Plant number Downtime: kilowatt-hours per operating day Separation and pasteurization Churning Cottage cheese-making Evaporating and condensing.. Evaporating and drying! Separation and pasteurization Churning Cottage cheese-making Evaporating and condensing.. Evaporating and drying! 22 70 12 11 t t t t t + + t t X 28 17 24 t 28 17 * Operating time: kilowatt-hours per hundred pounds of raw product 0.983 1.077 t 1.261 2.014 1.051 1 043 1.014 1.006 1.016 0.474 0.627 0.587 0.243 X X 0.243 X 1.270 1.358 1.383 1.810 1.515 1.899 1.810 0.998 832 X 1.277 2.043 * Including all refrigeration requirements, t Electricity is required only during operation. % No facilities. § Electrical requirements for the drying process in plants 1, 3, 4, and 6 include those for the independent evaporating stage. The Swenson dryer in plants 2 and 5 uses skim directly, because its wet collector serves as a preconcentrator. See footnote t. table 11 SOURCE OF DATA: Engineering specifications for the actual equipment at each plant. Rates. The maximum demand at a plant for a given month is defined as the maximum average quantity of electrical energy used during any 15-minute in- terval in that month. The maximum de- mand and the total quantity used during Table 18. 1960 Rates for Electric Power Maximum demand 200 kilowatts Power company and terms of billing Cooperative Monthly billing rate Southern California Edison Company plant number 1 2, 3, 4, 6 5 dollars Basic meter charge 203 75 All power used 0.008/kwh Pacific Gas and Electric Company First 6,000 kwh 0.0264/kwh 0.0091/kwh 019/kwh Additional units Minimum charge $150.00/ month Modesto Irrigation District First 13,400 kwh Next 13,400 kwh.. „ Next 40,200 kwh 0.012/kwh 008/kwh Additional units 0.006/kwh Minimum charge $168.30/ month SOURCE OF DATA: Rate schedules of the power companies. the month, both expressed in kwh, are the factors that determine power rates for each plant. The plants studied usually operated at or near a maximum demand of 200 kilowatts. Table 18 shows the 1960 schedule of rates charged at this demand level by the three power companies that served those plants. Reallocations of productive activity could result in extreme uses of electricity. For example, the dryer is the largest user of electrical energy at Plant 1. If the dryer was operated five full days each week, that operation alone would use 118,000 kwh per month. If all other power used in the plant amounted to 100,- 000 kwh per month, the full-time produc- tion of nonfat dry milk would double the use of power and thus reduce the effective rate for the month by 1 1 per cent — from $0.01 to $0.0089 per kwh. The effective price for electricity is determined after the total power requirements for the plant have been estimated. Natural Gas Requirements. According to a for- mula explained by Farrall (1953, p. 127), a boiler operating at 70 per cent efficiency [40 produces 1 pound of steam from 1.558 cubic feet of natural gas containing a heat value of 1,100 BTU per cubic foot. At milk-plant-operating pressures and temperatures, approximately 1,000 BTU of heat energy are available per pound of steam. If heat-transfer efficiency were per- fect, it would take 1/1,000 pound of steam to raise the temperature of one pound of product (specific heat = 1 ) 1° F. On this basis, the following formula was used to determine the steam requirements per pound of product at each stage : 1,000 S - pounds of steam; Q - pounds of prod- uct; (t 2 - t x ) is the temperature increase specified for a given product at a given stage. Although the heat transfer is not perfectly efficient, heat regenerative equipment used at all plants offsets losses to some extent. A regenerator partly heats the incoming cold product and partly cools the outgoing hot product by bring- ing the two into thermal contact. The ef- fect is difficult to measure, but the above approximation is adequate for this analysis. Steam requirements for the evaporation process were determined from manufac- turers' specifications. Steam requirements for cleaning at each stage were estimated by the plant personnel. Steam require- ments are converted to gas requirements (table 19) by using the factor 1.558, ex- plained above. For plants 2, 5, and 6, with steam- heat dryers, gas requirements per pound of nonfat dry milk produced were de- termined on the basis of manufacturer specifications. For plants 1, 3, and 4, that used gas for direct-heat drying, gas re- quirements per pound of nonfat dry milk were estimated from readings of a gas meter installed at the dryer in Plant 1. Rates. Table 20 presents the 1960 rate schedules of the two companies from which the six plants purchased natural gas. If the quantity purchased from P. G. & E. was 1 million cubic feet per month, the effective price was $0.0546 per hun- dred cubic feet. If the quantity was 10 million cubic feet, the effective price was $0.0489 per hundred. As the effective price is a function of the quantity used during the month, it is determined after total gas requirements for the plant have been estimated. Table 19. Amounts of Natural Gas Required for the Processing Stages' Six cooperative plants, 1960 Stage Plant number Downtime: cubic feet per operating day Separation and pasteurization Churning Cottage cheese-making Evaporating and condensing.. Evaporating and drying Separation and pasteurization Churning Cottage cheese-making Evaporating and condensing.. Evaporating and drying 935 1,090 1,558 1,090 779 1,558 t t t t t t t t t t t t 1,558 t 15,580 15,580 4,144 9,519 1,558 15,580 15,580 11,856 Operating time: cubic feet per hundred pounds of raw product 20.25 t 24.88 63.59 20.25 t 192.59 20.25 t 92.70 131.41 20.25 t 69.33 108.04 20.25 10.59 192.59 20.25 t 75.94 184.48 * For generating steam and for operating the dryers at Plants 1, 3, and 4. Calculated for gas with an average heat value of 1,100 BTU per cubic foot. t A negligible amount required to produce steam and hot water for cleaning. j No facilities. SOURCE OF DATA: Calculations based on engineering data for the actual equipment at each plant. [41] Table 20. 1960 Rates for Natural Gas* Gas company and terms of billing Cooperative Monthly billing rate Southern California Gas plant number 1,2 3, 4, 5. 6 dollars per thousand cubic feet (Mcf) First 200 Mcf 0.522 Next 800 Mcf 0.462 Next 2,000 Mcf 0.447 Next 3,000 Mcf 0.437 Next 4 000 Mcf 0.427 Next 10,000 Mcf 0.405 0.395 Minimum charge $50.00 per meter per month Pacific Gas and Electric Company First 1 000 Mcf 0.546 Next 2 000 Mcf 0.503 Next 3 000 Mcf 0.486 Next 4,000 Mcf 0.471 Additional units 0.407 Minimum charge $80.00 per meter per month * The standard heat value of natural gas is expressed as an average of 1,100 BTU per cubic foot. SOURCE OF DATA: PUC Schedule G-50, effective August Water Water requirements (table 21) for making steam, for cleaning equipment, rinsing butter, and making cottage cheese were estimated by plant personnel. Water requirements for operating the evapo- rators were estimated from the manu- facturer's specifications. Extreme dif- ferences existed among plants because some did not have regenerating systems to recover water from the steam used in evaporation. Rates. The average rate for water in the San Joaquin Valley is approximately one cent per 1,000 pounds. This rate is used for convenience, instead of making a detailed calculation of the cost of elec- tricity used in pumping the required amounts of water from wells owned at the plants. Any difference between the two rates would be minor. It would have little effect on the cost estimates for any stage and would not alter any proposed reallocation. Cleaning Agents Many different types and combinations of chemicals are used to clean equipment, in compliance with sanitary regulations. Table 21. Amounts of Water Required for the Processing Stages* Six cooperative plants, 1960 Stage Plant number 1 2 3 4 5 6 Downtime: pounds per operating day Separation and pasteurization Churning , 15,000 4,178 t 5,000 5,833 834 4,178 t t 4,168 10.000 5,001 t 8,000 9,600 834 1,250 t 5,000 6,667 800 834 t 6,650 4,168 834 1,250 Cottage cheese-making t Evaporating and condensing 3,750 Evaporating and drying 4,170 Operating time: pounds per hundred pounds of raw product Separation and pasteurization Churning 139 t 38 38 140 200 t 100 t 888 888 140 t 833 833 140 200 2,500 51 Cottage cheese-making f Evaporating and condensing 10 Evaporating and drying 10 * Including water used for cleaning equipment and that converted into steam. f No facilities. t Amounts of water used for cleaning vats included under operating time, below. SOURCE OF DATA: Manufacturers' specifications for the actual equipment at each plant and estimates by plant personnel. [42] Table 22. Costs of Cleaning Agents Used at the Processing Stages Six cooperative plants, 1960 Stage Plant number dollars per operating day Separation and pasteurization Churning Cottage cheese-making Evaporating and condensing. . Evaporating and drying 2.50 1.00 3.51 4.04 2.90 1.00 0.005t * 2.38 2.00 1.94 0.25 2.25 2.90 3.20 * 5.34 7.80 2.00 1.00 0.005t 3.64 2.38 1.50 0.72 * 4.80 6.40 * No facilities. t Dollars per 100 pounds of skim processed. SOURCE OF DATA: Calculations based on specifications for each process at each plant. The chemical mix and the cleaning pro- gram are specified for each plant. Table 22 gives estimates of the dollar costs in 1960 for cleaning agents at each stage on each operating day. Aggregate Costs Table 23 compares per unit variable costs at the different plants, computed at three different volumes for each of the four products. Because each product was considered independently in this table, the cost calculations for the utilities are based on fixed prices, corresponding closely to average prices paid at the plants. The result is that costs are overestimated slightly for large volumes and under- estimated slightly for low volumes. Ac- tually the per unit variable costs for any product are affected somewhat by the total volume and number of all products proc- essed at a plant, primarily because of the Table 23. Per Unit Variable Costs of Processing Four Manufactured Dairy Products* Six cooperative plants, 1960 Product Raw-product input Plant number Butter. Cottage cheese. Condensed skim. Nonfat dry milk. hundred pounds per day 120 240 480 350 700 1,200 1,200 2,400 3,600 600 1,180 2,000 dollars per hundred pounds of cream processed 0.35 0.34 t 0.87 0.44 t 0.86 0.44 0.27 0.29 0.23 t 0.28 0.23 t dollars per hundred pounds of skim processed t X X 0.064 0.055 t 0.149 0.121 0.112 0.15 t t 0.12 X X t X X t 0.098 0.110 X 0.091 0.094 t 0.087 t 0.20 0.234 0.198 0.18 0.168 0.178 t 0.160 0.156 0.15 0.09 0.12 0.09 0.08 t 0.20 0.18 t 0.44 0.22 0.22 076 074 0.231 0.192 0.172 * The calculations include only those labor and utility costs that would vary if production were reallocated among plants and do not include costs for receiving or for separation and pasteurization. To simplify this one table, an average price of 1 cent per kwh for electricity was used for all plants and a price of 0.5 mill per cubic foot for gas. t Above plant capacity. X No facilities. SOURCE OF DATA: Preceding tables in this section. [43] Table 24. Shipping Costs, 1960 Six cooperative plants to two markets Product and market area Fluid milk or cream Los Angeles Bay area Butter Los Angeles Bay area Condensed skim Los Angeles Bay area Cottage cheese Los Angeles Bay area Nonfat dry milk Los Angeles Bay area Los Angeles Bay area 0.435 0.445 0.231 0.285 0.128 0.131 0.043 0.043 173 234 Plant number dollars per hundred pounds of milk or cream processed 0.455 0.455 0.245 0.284 * * 099 0.099 0.043 0.040 0.465 0.400 0.254 0.239 0.137 0.118 0.043 0.038 0.635 0.330 0.285 0.220 0.187 0.097 0.043 0.033 0.660 0.330 0.333 0.201 0.194 0.097 0.116 0.068 0.061 0.033 miles from plant to market 205 230 219 188 289 124 313 94 0.665 0.330 0.333 0.201 0.195 0.097 0.061 0.033 312 101 * Not shipped from this plant. SOURCE OF DATA: Rate schedules of commercial carriers. graduated utility rates. This relationship is taken into account in the calculations in section VII. As would be expected, the per unit costs from these computations take on the na- ture of the hypothetical cost function in figure 13. Per unit variable processing costs vary significantly at different vol- umes in one plant; they may vary also in different plants operating at the same volume. For example, Plant 3 has a large churn that requires four or five men to operate it, so it requires large volumes to bring unit costs of butter- making down to a level with those of plants with smaller churns, that can be run by one or two men. In Plant 6, with a smaller churn, unit costs decrease rapidly at low volumes but level off quickly. Plant 2, with low-capacity equipment, has high per unit costs but Plants 4 and 5, with slightly lower capacities, have low per unit costs. In cottage cheese-making, production is increased either by setting more vats or by setting some vats more than once in a day. Plant 5 can process the 700 cwt of skim per day into cottage cheese by using a combination of regular and overtime labor from a single shift. An average daily volume of 1,200 cwt of skim would re- quire hiring an additional three men (see table 14). This raises the payroll sharply, so that even at the capacity operation of 12 vats per day the unit costs are higher than at the seven-vat level. Unit costs of condensing and of drying both decrease with higher production and vary considerably among the plants, de- pending on the type, size, and age of equipment, the utility requirements, and the rate and efficiency of operation. Double-effect evaporators require less steam per unit of output than do single- effect evaporators. Transportation Costs Transportation-cost estimates for ship- ping products from the six plants to the two major markets are based on 1960 rate schedules of commercial carriers, with the following specifications: [44] 1. All products are transported by truck, with minimum loads of 20,000 pounds. 2. Fluid milk, condensed skim, and fluid cream are shipped in bulk in single tankers, with an average load of 22,000 pounds. 3. Butter is shipped by van in 60-pound containers. 4. Nonfat dry milk is shipped by van in 100-pound containers. 5. Cottage cheese is shipped by van in 100-pound containers. The estimates of transportation costs in table 24 are in terms of raw-product equivalents. This makes it possible to com- pare the cost of shipping fluid cream or skim with the cost of shipping the more concentrated manufactured products. For example, the cost of shipping the butter made from 100 pounds of cream from Plant 2 to Los Angeles is almost 50 per cent less than the cost of shipping 100 pounds of fluid cream. Although, in general, shipping costs in- crease with distance, Plants 4, 5, and 6 all have the same per unit shipping costs to the Bay Area for each available product except butter. Therefore, reallocation of quantities among these three plants would have no effect on transportation costs un- less butter was involved in the realloca- tion. VII. POSSIBLE SHORT-RUN REALLOCATIONS FOR SIX COOPERATIVE PLANTS In this study, the economic time con- cepts long run and short run are nar- rowed in reference to a suggested re- organization among the cooperatives. Here the short run is defined as the cur- rent time period. Short-run adjustments are the reallocations that can be made in the kinds and quantities of products manufactured with existing facilities and transported to existing markets. The long run is defined as a period of time during which adjustments can be made in the type of equipment and in the number, operating capacity, and location of plants. Such adjustments — discussed in section VIII — would be determined in relation to changes in supply and demand conditions, with the purpose of providing the highest possible returns to the producers col- lectively. This section is restricted to the possible alternatives that would minimize the total costs of shipping fluid milk and cream and of processing and distributing four specified products. It treats the six firms as one entity but omits Plant 3A from most of the reallocations. The objective is to determine the nature of the optimum reallocation and the source and magni- tude of any cost reduction that might result. Evaporated whole milk and ice cream mix are not considered specifically in the reallocation study. Evaporated whole milk is omitted because only Plant 3A has canning facilities and no reallocation is possible. The cooperatives do not manu- facture ice cream, but they do supply its principal ingredients — manufacturing cream and condensed skim — both con- sidered in the reallocation study. Short-Run Adjustments Optimum short-run adjustments were determined by means of cost-minimiza- tion calculations based on the specifica- tions of present ownership, equipment and facilities, and locations. Each calcu- lation specified the physical conditions at the six plants and the actual quantities of milk received at each plant in the month under consideration. Also it specified and presented as daily averages the actual amounts of fluid milk and cream and of the four manufactured products that were shipped from the plants during that month. Three factors restrict the possible re- [45] allocation of manufacturing activity: the quantity of milk available for manufac- turing at each plant, the processing ca- pacities of each plant, and the market outlets for each product. The fact that the first and last items vary seasonally af- fects the extent and the type of realloca- tion that are possible. Therefore the actual data for a month of low manufac- turing activity, with unutilized capacity — October, 1958 — and the data for a peak month — July, 1959 — were used as a basis for studying reallocation possibilities. Re- stricted reallocation alternatives were found only in the peak months, when full capacity levels for some products were allocated to some of the plants. The reallocation is essentially a trans- portation problem, as discussed by Dorf- man, Samuelson, and Solow (1958, pp. 106-129). It specifies only the amounts shipped, not the amounts sold locally or processed for storage. The solution is con- sistent with location theory. In determining a minimum-cost solu- tion for a given month, the first step is to reallocate shipments of fluid whole milk among the plants so as to minimize ship- ping costs. As fluid milk is the most ex- pensive product to ship, it is most eco- nomical to ship it from the plant or plants that have the lowest rates for transporta- tion to the given market, especially as the per unit cost of handling fluid milk is low and does not differ much among the plants. The cost matrix for this step (table 25) contains only the transportation costs for fluid milk in bulk from each plant to each market. Milk that was not required in either of the two principal markets was assigned to a residual market, which rep- resented the quantity of milk available for manufacturing at each plant or for local sale as manufacturing milk. The milk for manufacturing is sepa- rated into cream and skim. A tentative reallocation of the shipments of manufac- turing cream and other products was calculated on the basis of minimizing total transportation costs. As per unit processing cost is a function of the vol- ume of product processed at each plant, a new cost matrix was then computed for each reallocation, using appropriate prices for the utilities in relation to total plant requirements. As reallocation altered the per unit costs of both labor and utilities, this matrix sometimes led to further re- allocations. In each matrix, 1960 rates (section VI) for labor, utilities, and ship- ping were used as a basis for comparing the total costs of the various allocations. The minimum-cost solution for each new allocation — determined from the combined-cost matrix — was calculated by the indirect-cost method described by Dorfman, Samuelson, and Solow (1958, pp. 106-117). After several successive iterations the cost matrix would become stable and the last allocation would be considered optimum in that any change from this allocation would result in higher total costs of processing and transporta- tion. The residual quantities of cream and skim were considered available for allo- cation at any plant. In the optimum al- location, the cost of having residual cream or skim at any one plant was considered extremely high, making it the least de- sirable alternative, indicated in the matrix by the dummy market with an arbitrarily high cost rate. The assumption is that residual quantities are to be destroyed and that the loss from such disposal is the same at each plant. In actual practice the quantities not shipped were probably sold locally or else processed into butter and nonfat dry milk. Such processing might require a somewhat different allocation because of capacity limitations, and this would affect the calculations on savings. The particular approach used in this study gives an indication of reallocation possi- bilities and limitations. Actual and Optimum Shipment Patterns for October, 1958 Fluid milk. Table 25 gives averages of the actual receipts and shipments of fluid whole milk at six plants for the 31 days of October, 1958. Five plants shipped some quantity of fluid Class I milk to Los Angeles and two shipped some to the Bay Area. [46] Table 25. Fluid Whole Milk: 31 -Day Averages of Actual Receipts and Shipments in a Month of Low Activity, with Suggested Reallocations Based on Minimum Shipping Costs Six cooperative plants. October, 1958 Total milk supply Market Shipping- cost matrix Fluid whole milk* Plant number Shipments Residual amounts Actual Reallocated Actual Reallocated hundred pounds per day 3,633 1,473 5,583 2,629 2,592 1,702 L. A. Bay area L. A. Bay area L A. Bay area L. A. Bay area L. A. Bay area L A. Bay area L. A. Bay area dollars per hundred pounds 0.435 0.445 0.455 0.455 0.465 0.400 0.635 0.330 0.660 0.330 0.665 0.330 hundred pounds per day 1 2,567 711 1,437 653 805 131 486 * * 3,402* 872* 1,377 1,139 * * 1,066 762 3,493 1,824 1,975 1,702 231 2 601 3 4,206 4 2,629 5 1,453 6 1,702 Totals 17,612 5,651 1,139 5,651 1.139 10,822 10.822 * All shipments were of market milk, with the exception of 200 pounds a day of manufacturing milk from Plant 1 to Los Angeles' Plant 6 received no market milk and shipped no fluid milk. In the reallocation, residual amounts for Plants 1 and 2 represent manu- facturing milk only. SOURCES OF DATA: The manager of each plant supplied estimates from plant records of actual receipts and shipments. Shipping costs are fnm table 24. Under the optimum pattern of realloca- tion, the two plants nearest to Los Angeles would ship all their Class I milk to that market and Plant 3 would fill the remain- ing Los Angeles requirement. Either Plant 4 or Plant 5 could supply all the Bay Area demand. Plants 4, 5, and 6 had the same costs for shipping fluid milk to the Bay Area, but Plant 6 handled only manufacturing milk. The reallocation as shown would reduce transportation costs for fluid milk by $234.76 per day— 7.5 per cent of $3,- 127.57, the calculated cost for the actual shipment pattern. Also it would allocate a larger proportion of the milk for manu- facturing to Plants 3 and 4, which are farthest from either market. Manufacturing cream (table 26). During October, 1958, all six plants shipped cream to the Los Angeles market — most of it for the manufacture of ice cream. Plant 6, the most distant, shipped the most. Two plants shipped cream to the Bay Area; two plants shipped butter to Los Angeles. The reallocation that would minimize costs would be to ship fluid cream from the four plants nearest to Los Angeles and from one plant near San Francisco, leav- ing large residual amounts in the three northern plants rather than smaller amounts scattered among six plants. Plant 4 would satisfy the small market re- quirement for butter and Plants 4, 5, and 6 would have to churn their residual quantities of cream, either for local sale or for storage. Skim milk. Table 27 shows how the six plants actually shipped the skim por- tion of their available milk. In October, 1958, four plants shipped two products each and two plants shipped one each. Reallocations were made to minimize combined processing and shipping costs. In some instances, large-volume opera- [47] Table 26. Utilization of Manufacturing Cream: 31 -Day Averages of Actual Shipments in a Month of Low Activity, with Suggested Reallocations Based on Minimum Combined Costs Six cooperative plants. October, 1958 Available cream Product and market Plant number 40% Cream Butter Residual amounts LA. Bay area LA. Bay area Actual shipments hundred pounds of raw product j 99 71 324 169 183 158 89 11 33 150 24 184 3 68 39 9 10 2 18 3 282 4 19 5 91 6 - 26* 1,004 491 71 48 394 Combined-cost matrix for final allocations dollars per hundred pounds of cream processed and shipped t, 1 0.4351 0.455 0.465 635 0.660 0.665 0.445 0.455 0.400 0.330 0.330 0.330 0.831 1.081 1.104 0.882 0.940 0.940 2.00§ 2.00 2 00 2 3 4 2 00 2 00 2 00 5 6 Reallocation hundred pounds of raw product 21 56 390 244 135 158 21 56 390 24 71 48 172 135 87 2 3 4 5 6 1,004 491 71 48 394 *The apparent overshipment from Plant 6 could indicate either high butterfat content of the milk supply or an error in this estimate. t Excluding costs for receiving and for separation and pasteurization. t Bold numbers indicate an active cell in the final solution. § Arbitrary, high cost figure used for the undesirable dummy market. SOURCES OF DATA: The manager of each plant supplied estimates from plant records of actual shipments. Cost data from tables in section VI. tions would reduce unit processing costs at one plant sufficiently to offset some cost disadvantage for transportation to one of the markets. Plant 3 would supply all of the condensed skim shipped in this month and Plant 4 would supply all of the nonfat dry milk. Only these two plants had sufficient skim to meet the entire re- quirements, and the large volumes proc- essed gave each one the lowest combined [48] Table 27. Utilization of Skim Milk: 31 -Day Averages of Actual Shipments in a Month of Low Activity, with Suggested Reallocations Based on Minimum Combined Costs Six cooperative plants. October, 1958 Plant number Available skim Product ancTmarket Cottage cheese L. A. Bay area Condensed skim LA. Bay area Nonfat dry milk LA. Bay area Residual amounts 1 2 3 4 5 967 691 3,169 1,655 1,792 1,544 9,818 1 2 3 4 5 6 1 2 3 4 5 209 545 3,817 2,385 1,318 1,544 9,818 407 407 5.00J 233 5.00 5.00 0.236 5.00 Actual shipments hundred pounds of skim processed * 277 287 203 * * * 475 268 * 823 372 181 228 * 496 895 143 575 2,252 228 1,450 143 Combined-cost matrix for final allocations dollars per hundred pounds of skim processed and shippedf 5.00 0.233 5.00 5.00 0.165 5.00 0.248 5.00 0.227 0.337 0.344 0.344 251 5.00 208 0.257 0.247 0.247 243 225 213 202 241 241 0.243 0.222 0.208 192 0.213 0.213 Reallocation hundred pounds of skim to be processed * * 407 * * * * 2,252 228 * * 1,450 143 575 * * 407 575 2.252 228 1,450 143 403 81 2,426 832 1,011 10 4,763 200t§ 2 00 2 00 2 00 2.00 2 00 209 138 1,337 792 743 1,544 4,763 * No facilities. f Excluding costs for receiving and for separation and pasteurization. t Arbitrary, high cost figures are used for plants without the needed facilities and also for the undesirable dummy marketi § Bold numbers indicate an active cell in the final solution. SOURCES OF DATA: The manager of each plantsupplied estimatesfrom plant records of actual shipments. Cost data from tables in section VI. cost for its product for both markets. Each of the cottage cheese plants would ship to its nearest market, as before. Slightly more than half of all the skim available for manufacturing at the six plants was proc- essed and shipped to the two markets. The other half was residual, and all six plants would have a quantity to dry for storage unless they could develop local sales out- lets for fluid skim. [49] Actual and Optimum Shipment Patterns for July, 1959 Fluid milk. Table 28 gives averages of the actual receipts and shipments of fluid whole milk at the six plants for the 31 days of July, 1959. The shipping pat- tern of the different plants was similar to that for October, 1958, as each firm would tend to retain its markets throughout the year. However, as the daily receipts of whole milk averaged more than 25 per cent higher in July than in the preceding October and the amounts shipped were nearly 50 per cent lower, the daily residual quantities of fluid whole milk available for manufacturing were actually greater in July than the daily receipts of all milk in October. In the actual pattern, four plants shipped Class I fluid milk to Los Angeles and two to the Bay Area. In the minimum- cost reallocation, Plant 1 alone would supply all the fluid milk required for Los Angeles and Plant 5 all that for the Bay Area. This one reallocation would save $229.91 per day— 13 per cent of $1,- 717.43, the calculated cost for the actual shipment pattern. Manufacturing cream (table 29). In July, as in October, six plants shipped manufacturing cream to Los Angeles and two plants shipped to the Bay Area. Los Angeles used nearly twice as much cream as in October, but Plants 1, 2, and 3 would supply it all in the reallocation, and Plant 6 would supply the slightly re- duced Bay Area demand. The demand for butter was nearly seven times as great in Table 28. Fluid Whole Milk: 31-Day Averages of Actual Receipts and Shipments in a Month of Peak Activity, with Suggested Reallocations Based on Minimum Shipping Costs Six cooperative plants. July, 1959 Plant number Total milk supply Market Shipping- cost matrix Fluid whole milk Shipments Actual Reallocated Residual amounts Actual Reallocated Totals hundred pounds per day 4,218 2,427 7.222 3.570 3.005 2.058 22,500 L A Bay area LA. Bay area LA. Bay area LA. Bay area LA. Bay area L A. Bay area L A. Bay area dollars per hundred pounds 0.435 0.445 0.455 0.455 0.465 0.400 0.635 0.330 0.660 0.330 0.665 0.330 hundred pounds per day 736 356 762 479 832 2,686 967 2,686* 967 2,686 967 3,482 2,071 5,981 2,738 2,517 2,058 18,847 1,532 2,427 7,222 3,570 2,038 2,058 18,847 * Los Angeles received 18,600 pounds a day of manufacturing milk, supplied by Plant 1. All other shipments were of market milk. Plant 6 received no market milk and shipped no fluid milk. SOURCES OF DATA: The manager of each plant supplied estimates from plant records of actual receipts and shipments. Shipping costs from table 24. [50] Table 29. Utilization of Manufacturing Cream: 31 -Day Averages of Actual Shipments in a Month of Peak Activity, with Suggested Reallocations Based on Minimum Combined Costs Six cooperative plants. July, 1959 Available cream Product and market Plant number 40% Cream Butter Residual amounts LA. Bay area LA. Bay area Actual shipments hundred pounds of raw product 1 323 192 555 254 233 191 219 30 234 186 34 208 3 61 50 109 77 32 56 54 2 50 3 244 4 36 5 82 6 - 17* 1,748 911 64 268 56 449 Combined-cost matrix for final allocations dollars per hundred pounds of cream processed and shippedt 1 2 0.435{ 0.455 0.465 0.635 660 0.665 0.445 0.455 0.400 0.330 0.330 0.330 0.571 0.685 0.694 0.534 0.563 0.553 0.625 0.724 0.679 470 0.525 0.542 2.00§ 2.00 3 2 00 4 2 00 5 2 00 6 2 00 Reallocation hundred pounds of raw product 1 142 225 670 331 189 191 142 225 544 64 268 56 2 3 126 4 7 5 189 6 127 1,748 911 64 268 56 449 * The apparent overshipment from Plant 6 could indicate either high butterfat content of the milk supply or an error in this estimate. t Excluding costs for receiving and for separation and pasteurization. t Bold numbers indicate an active cell in the final solution. § Arbitrary, high cost figure used for the undesirable dummy market. SOURCES OF DATA: The manager of each plant supplied estimates from plant records of actual shipments. Cost data from tables in section VI. July as in October. In the actual pattern, shipments were spread among five plants. In the reallocation, Plant 4 again would supply all the butter for both markets and Plants 3, 5, and 6 would have to churn their residual quantities of cream for local sale or for storage. Skim milk (table 30). Except for cot- tage cheese, with a stable demand, the optimum pattern for skim products in July would be quite different from the October pattern. Plant 1 would condense its entire supply of skim for the Los An- geles market, Plant 3 would condense [51] Table 30. Utilization of Skim Milk: 31 -Day Averages of Actual Shipments in a Month of Peak Activity, with Suggested Reallocations Based on Minimum Combined Costs Six cooperative plants. July, 1959 Available skim Product and market Plant number Cottage cheese Condensed skim Nonfat dry milk Residual amounts LA. Bay area LA. Bay area LA. Bay area Actual shipments hundred pounds of skim processed 1 3.158 1,880 5,426 2,485 2.284 1,867 * 410 * * * * 205 * * 410 * 679 * 484 655 263 233 * 595 3,233 967 3.997 763 186 601 - 754 1 2 298 3 945 4 1,830 5 830 6 270 17,100 410 615 2,314 595 8,960 787 3,419 Combined-cost matrix for final allocations dollars per hundred pounds of skim processed and shipped}: 1 5.00§ .231 5.00 5.00 .238 5.00 5 00 .231 5 00 5.00 .161 5.00 .191|| 5.00 233 .311 .308 .305 .194 5.00 .214 .221 .211 208 .193 .161 .178 .187 .261 .275 .193 .173 .173 .177 .214 .247 2.00§ 2 00 2 3 2 00 4 2 00 5.. 2 00 6 2 00 Reallocation hundred pounds of skim to be processed 1,390 2.202 6,552 3,240 1,849 1,867 * 410 * * * * * * 615 * 1.390 * 924 * 595 697 5,6281 2,635 787 o 2 1 095 o 4 . 605 447 6 1 272 17,100 410 615 2 314 595 8,960 787 3,419 * No facilities. t Presumably Plant 1 shipped from storage nonfat dry milk equivalent to 75,400 pounds of skim per day. This market in July is considered available to all the plants. t Excluding costs for receiving and for separation and pasteurization. § Arbitrary, high cost figures are used for plants without the needed facilities and also for the undesirable dummy market. || Bold numbers indicate ai active cell in the final solution. 1 Drying facilities in Plant 3A used for volumes in excess of Plant 3 capacity. Cost estimate includes cost of operating that equip- ment. SOURCES OF DATA: The manager of each plant supplied estimates from plant records of actual shipments. Cost data from tables in section VI. enough to make up the balance, and Plant 6 would supply condensed skim to the Bay Area. Plants 2, 3, and 4 would dry nonfat milk for Los Angeles, Plant 5 for San Francisco. However, Plant 3 would have to use the drying facilities of Plant 3A to handle the allocation that is be- yond its capacity. This would use nearly • In 1959 Plants 3 and 4 sold an unknown quantity of skim to proprietary firms for making cottage cheese. r 52 1 a third of the daily evaporating capacity of Plant 3A. Residual skim at Plants 2, 4, 5, and 6 would be either sold for cottage cheese 8 or dried for storage or sale. Ac- tually the cooperative plants were not equipped for this "optimum" type of re- allocation in 1959, though their capaci- ties were well above the demands of their actual processing and distribution pat- terns. Savings Potential of Product Reallocation The optimum patterns would have re- duced total combined variable costs for all plants by 1 3 per cent in October and by 1 1 per cent in July. About two-thirds of Table 31. Variable Costs in Relation to Product Allocations* Six cooperative plants Table 32. Variable Labor Requirements in Relation to Product Allocations* Six cooperative plants Cost items Actual shipments Reallocation Potential saving dollars per day October, 1958 Variable processing costs: Labor 769.59 183.32 342.65 26.82 48.24 505.49 189.81 329.67 37.76 31.95 264.10 Electricity - 6 49 Gas 12 98 Water -10 94 Miscellaneous 16.29 Total 1,370.62 4,032.98 1,094.68 3,622.74 275 94 Transportation costs . . 410.24 Combined costs 5,403.60 4,717.42 686.18 July, 1959 Variable processing costs: Labor Electricity 1,374.20 382.33 825.10 71.81 62.05 1,150.33 376.55 871.42 76.61 45.49 223.87 5 78 Gas -46 32 Water - 4 80 Miscellaneous 16.56 Total 2,715.49 3,257.34 2,520.40 2,870.33 195 09 Transportation costs . . 387.01 Combined costs 5,972.83 5,390.73 582.10 * Including variable costs at the separation-and-pasteuri- zation stage, and allowing for the interdependence of utility prices at all stages. SOURCE OF DATA: Calculated from tables in section VI, as used in determining optimum allocations. Plant number Actual shipments Reallocation Potential saving man-hours per day October, 1958 1 25 30 32.86 75 50 22.40 40.00 40.20 8.00 24.30 26.60 57.00 27.50 8.10 17.30 2 8.56 3... 48 90 4 -34.60 5 12.50 6 32.10 236.26 151.50 84.76 July, 1959 1 88.90 62.20 115.80 37.40 60.50 43.40 23.80 48.00 124.50 89.80 48.40 17.40 65.10 2 14.20 3 - 8.70 4 -52.40 5 12.10 6 26.00 408.20 351.90 56.30 * Including direct labor required at the separation-and- pasteurization stage. SOURCE OF DATA: Calculated from tables in section VI, as used in determining optimum allocations. the total savings would be due to reduc- tions in transportation costs, and more than half of this would result from the reallocation of the fluid-milk shipments — even in July, when less fluid milk was shipped than in October, both in actual volume and in proportion to other prod- ucts (table 31). The further cost reductions in both months would result mostly from reduc- tions in labor requirements. By manu- facturing only one or two products at any plant — and these in larger volumes — the proportion of production time relative to downtime would be increased and the indivisible eight-hour labor unit would be utilized more completely. This reallo- cation would reduce labor requirements, as calculated for the actual amounts pro- duced, by an average of 84.7 man-hours per day in October and by 56.3 man- hours per day in July (table 32). In the actual shipping pattern for July [53] the variable-labor requirements were al- most twice as great as in the actual pat- tern for October and almost three times as great as in the optimum pattern for Oc- tober. However, the optimum pattern for July would require only about 50 per cent more labor than the actual pattern for October. Even if all the labor needed for operation during the flush season must be held throughout the year, the optimum pattern would save 56 man-hours per day over the actual. However, it should be possible to vary the labor force and save somewhat more by reducing the work week during the slack season and not filling vacancies, and by using overtime labor in the peak season. Total savings might be even greater ex- cept that certain of the utility costs were increased by these particular reallocations — evidently because the different plants have equipment with different utility re- quirements per unit of output. Tables 17, 19, and 21 show these differences and should guide management to investigate relative efficiencies. The two months studied represent the extremes in fluid-milk shipments and in manufacturing activity. Since changes from month to month are gradual and more or less uniform, it is reasonable to assume that an arithmetic average of the estimated daily savings for the two months ($634.14, from table 31) would represent the potential average daily sav- ing throughout the year. The potential annual saving would be $231,461 — almost one cent per pound of butterfat handled, or an average of $155.76 for each pro- ducer who shipped milk to the coopera- tives in 1960. However, this does not allow for the cost of procedures to imple- ment the reallocation. The reallocations calculated in this project involve no change in the quantity ■ of milk received at each plant. Only the out-of-area markets for each product are reallocated among the existing plants so as to achieve the lowest possible com- bined cost of manufacturing and ship- ping. This method of aggregating the supply and product markets assumes that the six firms somehow cooperate by shar- ing profits and losses. The important aspect of the findings is that labor costs and transportation costs both are reduced by the pattern of ship- ping fluid milk from fewer locations and by specializing in one or two products at ■ each manufacturing plant. VIII. LONG-RUN AND SHORT-RUN ADJUSTMENTS The proposed reallocations for existing plants show the economies inherent in consolidating operations in the dairy- products industry. Long-run problems should be considered along with short-run adjustments and with an analysis of liqui- dation values, so that interim decisions and short-run adjustments may be con- sistent with long-run needs. Of necessity the cooperatives, as well as other firms, will change their operating organizations. They will need to abandon some of the existing plants and possibly, later, erect a smaller number of modern plants, incorporating new techniques and equipment and possibly at new locations. However, they can realize current econ- omies if they utilize existing facilities more efficiently and in a way consistent with the long-run outlook for the industry. Long-Run Adjustments , Projections for 1975 point to about 23 manufacturing plants for the entire in- dustry in California. Each plant would process 75 to 100 million pounds of milk a year — the level through which signifi- cant scale economies are most likely to exist. Although actual conditions are not expected to be exactly as projected, the important products manufactured in 1975 will very probably be cottage cheese and ice cream mix. The firms already produc- ing significant quantities of these two im- [54] portant products have a market advantage. Other firms, also, that can gain access to this market in the near future will profit by its increasing importance. To remain competitive in terms of op- erating costs, the cooperatives will need to reduce the number of their plants or increase their share of the raw milk sup- ply, or both. If they merely maintain the 16.6 per cent of the state's milk for manu- facturing that they had during the four years 1957-1960, their actual supply will diminish gradually. Compared with their 485 million pounds in 1960 — an average of 81 million pounds for each of the six firms — their proportional share would be only about 400 million pounds in 1970 and 350 million in 1975 (table 33). The Table 33. Actual and Projected Quantities of Whole Milk Available for Manufacturing Seven cooperative plants Year Quantity 1957 thousand pounds 501,915 1958 444,505 1959 462,018 1960 485,220 1970 400,292 1975 349,795 SOURCE OF DATA: Actual supply figures for 1957-1960 from the six firms. Projections computed as 16.6 per cent of the projected amounts in table 7. cooperatives would need only three or four plants to process their 350 million pounds. Operating the existing plants with reduced milk supplies would increase unit costs and make it difficult to compete against plants that operate at larger vol- umes. Another possibility, of course, is for each cooperative to increase its share of the state's milk available for manufactur- ing. To operate six plants at the level of 100 million pounds a year, the coopera- tives would need 28 per cent of the state's expected milk for manufacturing in 1975 — 33 per cent of the projected supply in the San Joaquin Valley. If this supply could be obtained, it would give the co- operatives a much more important posi- tion in the industry. However, it would involve taking over a share of the milk supply and markets now held by pro- prietary firms. As the other plants will be under the same pressures as the coopera- tives, they might be willing to let the co- operatives take over more of the surplus- handling function. In Maryland, New York, and other eastern states, coopera- tives supply most of the additional milk needed by proprietary firms during the fall months and handle most of the sur- plus during the flush summer months. Adjustments in the direction of the an- ticipated changes should start soon. In 1960, six of the seven plants owned by the cooperatives had manufacturing ca- pacities of 100 million pounds or more, but these facilities were largely for con- densing and drying. The smallest plant, No. 2, was not far below this level and it had some cottage cheese capacity, but it did not have evaporating equipment sepa- rate from the drying process. As a group the cooperative should in- vestigate market opportunities and make long-run plans to develop their outlets. The seven plants now have enough evapo- rating capacity to condense nearly 533 million pounds of skim per year, operating 24 hours a day for 260 days, but in 1975 only 330 million pounds of skim will be needed to satisfy the entire state's ice cream market. On the other hand, Cali- fornia's cottage cheese market for 1975 will probably require about 1.5 billion pounds of skim — more than 20 times the present capacity of the cooperative plants even if they should operate continuously for 360 days a year. An alternative to increasing their cot- tage cheese-making capacity is to provide skim and cream for further processing by other firms. Several of the cooperatives already have contracts to provide pro- prietary firms with skim for cottage cheese, and this is one of their current functions. To produce curd for other firms or to package cottage cheese under the label of others firms would require both 55] new equipment and new outlets and would make the cooperatives vulnerable to some extent to the changing fortunes of com- peting firms. Short-Run Adjustments Because the development of a market requires both time and money, and be- cause relatively large quantities of milk are now available for manufacturing, it is necessary to maintain enough facilities to provide for flexibility in the utilization of current milk receipts. If, when trends be- come more definite, the situation appears to be developing as projected, less flexi- bility will be required and the plants might close down such operations as dry- ing and butter-making. A specific allocation policy to imple- ment the adjustments suggested in section VII would require a central clearing house. Basically the policy is quite simple and straightforward. The following guide- lines may be used to implement realloca- tions by the cooperatives and should be helpful for any entrepreneur with more than one plant to coordinate. 1. On a daily basis, ship fluid whole milk from the plant nearest to the market. When the supply in this plant is exhausted, ship from the next nearest plant and then the next, until the requirements are satis- fied. 2. After requirements for whole milk are satisfied, allocate fluid cream in the same way. Allocate skim to the manufac- ture of cottage cheese and condensed skim in the plant or plants nearest to each market, to the extent of available supplies. 3. Concentrate the production of a given product at a single plant, as far as possible. In other words, have each plant specialize in the manufacture of one or two products. 4. Concentrate the manufacture of butter and nonfat dry milk at plants that have large-scale equipment and large quantities of cream and skim for manu- facturing, principally at plants farthest from the markets. Considerations for Implementing Reallocations The conclusions from this study point to some form of consolidation for the co- operatives. The six firms might merge into one, or form a federation, or make and enact joint decisions through the Cali- fornia Agricultural Processing Coopera- tive Association, which was organized in 1960 for the purpose of labor negotia- tions. Any such procedure would require competent legal counsel, because joint action for manipulating production and shipment patterns might be considered as ■ tending to lessen competition. Without some specified "joint agency in common," the fairly drastic reallocations indicated in the optimum patterns would not be feasi- ble. Moreover, distribution of the in- creased net returns among the individual cooperatives and then to the producer- members would involve problems of equity that are normally difficult to solve. The average gain of $155 per member per year, calculated above for the short- run reallocation, may seem relatively small and might not induce the present members of the cooperatives to give up I the long-standing identity of their organi- zations, except for the prospect of poten- tial long-run gains. However, unless they can pay competitive prices to their pro- ducers, the cooperatives will not maintain their share of the milk receipts. The po- tential cost reduction of one cent per pound of fat is small but it represents an improved competitive position in the mar- ket for milk and milk products. It is important to repeat that these con- clusions necessarily are based on projec- tions and should be used merely as a guide, not as statements of fact. Every entrepreneur who aims to apply them should reassess developments frequently, as to milk supplies, manufacturing costs, and the product markets, so that he may change his operations with the changing times. Adjustment to the continually changing economic environment of the industry is a necessary condition to sur- vival. [56] LITERATURE CITED Appleman, Robert D., and C. L. Pelissier 1961. Relationship of herd size to production. California Agr. Ext. Serv., Davis. 2 pp. (Processed.) Boles, James N. 1958. Economies of scale for evaporated milk plants in California. Hilgardia 27 (21) : 621- 722. California Crop and Livestock Reporting Service 1941 — 1945; 1946a — 1963a. California dairy industry statistics for 1940 [or subsequent year]: Manufactured dairy products, milk production, utilization, and prices. Cali- fornia Dept. Agr., Sacramento. Annual issues, each about 75 pp. 19466 — 19616. Dairy information bulletin. California Dept. Agr., Sacramento. Monthly issues, each about 20 pp. California Department of Finance 1951. Provisional annual estimates of the population of the State of California, 1940-53. California Dept. Finance, Sacramento. 25 pp. 1959. California's population in 1959. California Dept. Finance, Sacramento. 19 pp. 1962. Preliminary projections of California areas and counties to 1975. Special Rpt., Janu- ary 3, 1962. California Dept. Finance, Sacramento. 6 pp. Clarke, D. A., Jr., Olan D. Forker, and Aaron C. Johnson, Jr. 1964. Pricing milk for manufacturing purposes in California. California Agr. Exp. Sta. Bui. 801. 134 pp. Cobb, Fields W., Jr., and D. A. Clarke, Jr. 1960. Class III milk in the New York milkshed: I. Manufacturing operations. U. S. Dept. Agr. Marketing Res. Rpt. 379. 36 pp. Daly, Rex F. 1954. Some considerations in appraising the long-run prospects for agriculture. Long- range economic projections. Vol. 16, pp. 131-89. In National Bureau of Economic Research. Studies in income and wealth. Princeton University Press. Dean, G. W., and C. O. McCorkle, Jr. 1961. Projections relating to California agriculture in 1975. California Agr. Exp. Sta. Bui. 778. 58 pp. Dorfman, Robert, Paul A. Samuelson, and Robert M. Solow 1958. Linear programming and economic analysis. McGraw-Hill Book Company, Inc., New York. 527 pp. Farrall, A. W. 1953. Dairy engineering. 2nd ed. John Wiley & Sons, Inc., New York. 477 pp. Hassler, James B. 1953. Pricing efficiency in the manufactured dairy products industry. Hilgardia 22 (8): 235-334. Jacobsen, M. S. 1936. Butterfat and total solids in New England farmers' milk as delivered to processing plants. Jour. Dairy Science 19: 171-76. Johnson, Aaron C, Jr., Olan D. Forker, and D. A. Clarke, Jr. 1964. Operations and costs of manufacturing dairy products in California. Univ. Cali- fornia, Giannini Foundation Res. Rpt. 272. 72 pp. (Processed.) McAllister, C. E., and D. A. Clarke, Jr. 1960. Class III milk in the New York milkshed: IV. Processing margins for manufactured dairy products. U. S. Dept. Agr. Marketing Res. Rpt. 419. 102 pp. Simmons, Richard Lee 1959. Optimum adjustments of the dairy industry of the western region to economic con- ditions of 1975. Ph.D. Thesis, Univ. California, Berkeley. 352 pp. Sosnick, S. H., and J. M. Tinley 1960. Marketing problems of San Joaquin Valley cooperatives. Univ. California, Giannini Foundation Res. Rpt. 228. 61 pp. (Processed.) [57 United States Department of Commerce 1961. Survey of current business, August, 1961. U. S. Dept. Commerce, Off. Business Econ.41 (No. 8). 75 pp. Walker, Scott H„ Homer J. Preston, and Glen T. Nelson 1953. An economic analysis of butter-nonfat dry milk plants. Idaho Agr. Exp. Sta. Res. Bui. 20. 90 pp. [58] APPENDIX A ANALYSIS OF THE DEMAND FOR CLASS I MILK IN CALIFORNIA The objective of this analysis is to de- velop estimates of the quantity of milk needed to satisfy California's Class I re- quirements in 1975. The analysis is based on the hypothesis that annual per capita consumption — Y — of Class I milk prod- ucts (table A-l) is a function of three se- lected variables (table A-2) : V 19 annual per capita personal income; V 2 , the retail price of milk; and V 3i the proportion of the population under age 15. The relation- ships of these variables to per capita milk consumption were studied by regression analysis of data for the years 1946-1961, using coefficients derived by the method of least squares. The resulting equation can be used to predict the requirements for future years. A trend variable — t = 1 for 1946, 2 for 1947, ... 16 for 1961 — was inserted to explain variations in consumption that might be related to a continuous shift in the food-purchasing habits of the popu- lation. The linear fit of the variables (in which R 2 , the adjusted coefficient of de- termination, was 0.89) provided a better description of past conditions than did their logarithmic transformation (in which R 2 was 0.74). When the three variables were con- sidered, the derived relationship was as in Equation (A-l ) , below. The t ratios appear in parentheses, and R 2 is 0.89. All coefficients are statistically significant except the one based on the price of milk. The magnitude of the price coefficient is consistent with the results of other workers, indicating very little elas- ticity of demand with respect to price (only 0.13, where zero indicates complete inelasticity). Omitting the price variable from the analysis would simplify the prediction equation and would not alter the other relationships appreciably. This omission implies that price changes have no effect on per capita milk consumption or — if they do — that the price level of milk in relation to that of other foods will not change greatly from the averages for 1946 through 1961. When the price varia- ble was omitted, R 2 decreased to 0.87 and the derived relationship was that shown in Equation (A-2), below. This indicates that a $100 increase in per capita income was associated with a 5.7- pound increase in per capita milk con- sumption — an income elasticity of 0.45 — and that a decrease of 1 per cent in the proportion of the population under age 15 was associated with a 13.171-pound decrease in per capita milk consumption — an elasticity of 1.2. The trend variable indicates an opposing trend — independent of price, income, and demographic effects — toward reduction in per capita milk consumption averaging 8.323 pounds per year. Actual per capita consumption in 1953, 1954, 1955, and 1960 was less than the calculated amount by 3 pounds or more and in 1950, 1956, 1957, and 1958 it was greater by 3 pounds or more (fig. A-l). The Durbin-Watson test of serial correla- y = -63.6 + 0.0572F 1 -3.9009Fo + 12.3414F,-7.8668f (5.41) (-0.93) (5.83) (-5.70) Yzz -122.16 + 0.057^+13. 171 F 3 - 8.323/ (5.43) (6.90) (-6.48) [59] (A-l) (A-2) Table A-l. Annual Per Capita Consumption of Dairy Products Table A-2. Variable Factors Studied in Milk-Consumption Analysis Population, July 1 Per capita sales Year Class 1 milk products Cottage cheese Ice cream thousands pounds pounds quarts 1946 9,559 298.7 6.07 23.18 1947 9,832 287.3 6.27 19 54 1948 10,064 284.3 6.29 15.91 1949 10,337 283.1 6.99 14.53 1950 10,609 294.6 7.26 14.54 1951 11,058 301.9 7.44 14.75 1952 11,743 302.5 7.55 15.60 1953 12,168 302.5 7.54 15.12 1954 12,595 299.2 7.40 13.80 1955 13,035 309.1 7.67 13.83 1956 13,594 320.3 8.28 14.28 1957 14,190 320.6 8.30 14.61 1958 14,752 313.4 8.19 14.67 1959 15,280 311.3 8 28 14.61 1960 15,860 302.9 8 18 14.04 1961 16,445 292.8 7.74 13.72 SOURCE OF DATA: California Crop and Livestock Reporting Service (1947a-1962a). Year Vi, Per capita personal income* V 2 , Average retail price of milk*t V 3 , Proportion of population below age 15 dollars cents per pound per cent 1946 2,432 10.60 22.20 1947 2,181 10.41 22 80 1948 2,088 10 55 23 60 1949 2.078 10 67 24.40 1950 2,195 9 87 24 96 1951 2.251 9.93 26 16 1952 2,302 10.72 26 84 1953 2,323 10.56 27.50 1954 2,301 9.97 28 14 1955 2.462 9 97 28.70 1956 2,560 10.05 29.19 1957 2,551 10 26 29 68 1958 2.508 10 29 29 97 1959 2,628 10 39 30 28 1960 2,659 10 51 30 60 1961 2,659 10 53 30.24 * Adjusted to constant 1957-1959 dollars, according to cost of living index (U. S. Departnent of Commerce, 1961). t Arithmetic average of prices in effect each month in the Los Angeles, Sacramento, and San Francisco marketing areas. SOURCES OF DATA: Income data from U. S. Department of Commerce (1961, p. 13); population data and prices from California Crop and Livestock Reporting Service (1946b-1961b). 1946 1950 1955 I960 Fig. A-l. Residual (unexplained) variations in Class I milk consumption in California, 1946- 1961. Base line (0) represents the predicted per capita sales for each year. Quantities above or below the base line show the amount of divergence of actual sales from sales predicted by equation (A-2). Quantities calculated from data in tables A-l and A-2. [60] tion was inconclusive. However, the cor- relation of successive residuals is not critical in a projection equation. If serial correlation did exist it would mean merely that, over time, the actual values would vary in a systematic manner above or be- low the predicted values. Equation (A-2) can be used to provide milk-consumption estimates for future years by substituting the appropriate esti- mates of the independent variables — per capita personal income, age distribution, and trend. The effect of arresting the trend toward reduced milk consumption can be measured (table 5) by holding / at 16. This implies that consumer tastes and preferences will affect sales in 1975 in the same direction and amount as in 1961. On the other hand, we can measure the im- pact of a continuation of this trend by allowing t to progress for future years: t = 25 for 1970 and 30 for 1975. APPENDIX B ANALYSIS OF THE DEMAND FOR COTTAGE CHEESE IN CALIFORNIA Regression analysis of data for 1946 through 1961 indicates a relationship be- tween the annual per capita consumption of cottage cheese — Y cc — and two inde- pendent variables — V lt annual per capita personal income, and trend, t - 1 for 1946, 2 for 1947, etc. Trend, per se, is not an economic variable. However, if changes in the purchasing habits of con- sumers are highly correlated with time, the trend variable will reflect these changes. Equation (B-l), below, in loga- rithmic terms, provided a better fit than did the linear equation. The / ratios appear in parentheses, and R 2 — the adjusted coefficient of determina- tion — was 0.90. With the untransformed data, R 2 was 0.80. When income was used as the only in- dependent variable in a linear equation the income coefficient was significant at the 1 per cent level and R 2 was 0.51 : 1.1 +0.003 V 1 (3.8)** (B-2) Equation (B-l) indicates an income elasticity of 0.17, which is considerably lower than most of the estimates made on other manufactured dairy products. Equa- tion (B-2) indicates an income elasticity of 0.85, which is closer to other estimates. However, the derived relationship in equation (B-2) explains only 50 per cent of the variation in consumption whereas that in equation (B-l) explains 90 per cent of the variation. Therefore equation (B-l ) was used in making the projections for this study. log e Y cc = 0.44273 + 0.17351 (log e VJ +0.11150 (log e t) (1.25)* (7.58)** (B-l) * Not significant at the 10 per cent level. ** Significant at the 1 per cent level. [61] APPENDIX C PRODUCT YIELDS AND CONVERSION DATA 1 Table C-l. Amounts of Raw Dairy Products Used Per Pound of Manufactured Product* Manufactured product 40% cream Skim milk Butter pounds 2.0434 0.0898 0.3683 pounds Nonfat dry milk 11.0927 Cottage cheese curd 6.0557 Creamed cottage cheese Condensed skim (for ice cream mix) 5.5119 3.4027 Ice cream mix (formula used in this report) 0.9476 * According to plant managers, there are slight losses of fat and considerable losses of nonfat solids in processing all products, especially cottage cheese. McAllister and Clarke (1960) give conversion formulas for fat and skim in making cottage cheese and ice cream mix. The butterfat and nonfat solids content of milk from producers varies consider- ably. Typically milk receipts at a plant average between 3.5 and 4.0 per cent fat, by weight. For the purposes of these cal- culations, it was assumed that the whole milk received averaged 3.8 per cent fat, by weight. One quart of 3.8 per cent milk weighs 2.15 pounds. One hundred pounds will yield 9.273 pounds of 40 per cent cream and 90.727 pounds of skim milk, or 4.538 pounds of butter and 8.179 pounds of nonfat dry milk. One pound of milk fat represents 26.32 pounds of 3.8 per cent milk. One pound of evaporated whole milk is produced from approximately 2.1 pounds of 3.8 per cent milk. One pound of ice cream mix makes 1 quart of ice cream. Tables C-l and C-2 give additional data. Table C-2. Usual Content of the Manu- factured Dairy Products Studied in this Paper Constituents Product Fat Nonfat milk solids per cent by weight Whole milk 3.8 40.0* 0.1 80.5 1.25 t 4.0 t 7.9 12.5 8.6 Cream 5 6 Skim milk 9 08 Butter 1 Nonfat dry milk 93 75 Cottage cheese curd 25. 0t Creamed cottage cheese Condensed skim (for ice cream mix) Evaporated whole milk 21.0 32. 0t 18.0 Ice cream mix§ 10.0 * Specified, f Negligible amount. t Total solids, largely nonfat. § Formula used in this report. This mix contains also 15 per cent sugar and 0.4 per cent gelatin. 1 Based in part on data and formulas given by Jacobsen (1936), Hassler (1953), Walker, Preston, and Nelson (1953), and McAllister and Clarke (1960). [62] 6m-8,'65(F4248)JT