I/I E> RAR.Y OF THE. UNIVERSITY* OF ILLINOIS" cop- 2. NON CIRCULATING CHECK FOR UNBOUND CIRCULATING COPY What Determines SOYBEAN PRICES By G. L JORDAN Bulletin 546 UNIVERSITY OF ILLINOIS AGRICULTURAL EXPERIMENT STATION CONTENTS PAGE Soybean Meal Prices 1 72 Soybean Oil Prices 180 Soybean Prices as Related to Value of Meal and Oil. . 183 Seasonal Pattern of Soybean Prices 1 85 Summary 1 87 Tables. .189 Urbana, Illinois March, 1951 Publications in the Bulletin series report the results of investigations made or sponsored by the Experiment Station WHAT DETERMINES SOYBEAN PRICES By G. L. JORDAN, Professor of Agricultural Economics THE PRICE which farmers receive for soybeans in peacetimes is determined by several factors, including: (1) the price users will pay processors for soybean oil; (2) the price users will pay processors for soybean meal; (3) the processing margin, or the amount that processors are able to obtain for their services; and (4) transportation and other handling costs on beans and products. This analysis assumes no effective price supports and no effective price ceilings. One fact that makes the analysis of soybean prices so complicated is that there are several _substitutes_ for both the oil and the meal. Users are interested in a number of edible fats and oils, not just soy- bean oil, and they are interested in a number of protein supplements for livestock feeding, not just soybean meal. On both the demand and the supply sides we must therefore consider .total available edible fats and oils and total available protein supplements. While soybean oil can be used as a drying oil, and is so used when price relationships justify it, there are other drying oils that compete with it, so that the demand for it for this purpose is not large and we can ignore it. We can also forget nonfeed uses of soybean meal, as feed use sets the price and only small amounts are used for nonfeed purposes. We shall see how the prices of soybean oil and soybean meal are determined, then note the relation of soybean prices to prices of soy- bean oil and soybean meal. Some consideration will also be given to the seasonal movement of soybean prices. It is recognized that a large number of forces affect prices of soybeans and that these forces vary in degree of influence over a period of time. It is also recognized that the period since soybeans have become a major crop and since the end of World War II, with all its disturbances to economic relationships, is so short that the relationships shown in this study may very well change during the next few years. The post-World War II period has been characterized by inflationary pressures, abnormally large con- sumer buying power in this country, exports facilitated by United States financing of importing countries, and disruption of world trade in fats and oils. These have all influenced soybean prices. 171 172 BULLETIN No. 546 [March, Soybean Meal Prices We shall first examine the factors which affect the prices of soy- bean meal. It is a commonplace that in a freely competitive economy supply and demand determine prices. In the absence of free compe- tition, dependence is placed, on arbitrary decisions made by some person or agency regarding the supplies that will reach the consumers or the prices which may be charged. This latter condition prevailed during World War II; therefore the years 1942-43 to 1945-46 are ig- nored in this study. But we have to have some measures of demand and supply. In laymen's language, demand refers to both the ability and the willing- ness to buy. Supply refers to the total quantity which people are willing to sell. Obviously a person's ability and willingness to buy is related to his financial resources, and for most people that means their current income after taxes. Disposable personal income will therefore be used in this study as a measure of domestic demand. The demand for meal is derived from the demand for meats and livestock products for the production of which the meal is used, and the demand for oil is derived largely from the demand for margarine and shortenings. The principal use of soybean meal is for protein supplements for livestock feeding, but other protein supplements are available for the same use. It is therefore necessary to take into ac- count the total quantity of oilseed meals and animal protein supple- ments available. 1 The relation between the price of corn and the price of supplements also affects feeders' willingness to buy supplements. This price relationship is affected by the quantity of corn available as well as by the quantity of protein supplements available. So the size of the corn supply must be considered. Also, regardless of the size of consumers' incomes, the demand for protein supplements is affected by the number of grain-consuming livestock to be fed. In the present analysis of the factors affecting soybean-meal prices, a reasonably high correlation between actual and estimated weighted composite prices of all protein supplements for the periods 1935-36 to 1941-42 and 1946-47 to 1948-49 was obtained by basing estimates on the influence of three factors: (1) disposable personal income 2 ; (2) the production of protein supplements (in terms of meal equiva- 1 These include soybean meal, cottonseed meal, linseed meal, copra meal, peanut meal, gluten feed and meal, tankage, meat scraps, and fish meal. Wheat- milling by-products were not used. 2 Income for October to September was used in order to correspond with the crop year. Prior to 1939 calendar-year data were interpolated. 1951} WHAT DETERMINES SOYBEAN PRICES 173 o E c o xj !I! o c tfl O 11 o h > " E j_i ** S >> rt O - en *! T3 & s^ 2^ ji to O 'to o m s 'O co Sg ^2 E , O.O) ft-t .5 s '5 o 1*5 I s| o 8i3 174 BULLETIN No. 546 lent) per grain- consuming animal unit (except horses and mules) ; and (3) the corn supply (production plus carryover) per grain- consuming animal unit (except horses and mules) (Figs. 1 to 4). Several other combinations of factors were tried and discarded. Logarithms of the data were used for the dependent variable, the weighted annual composite price per ton of protein supplements, and for the independent variables 1 and 2 described above. However, numbers of bushels were used instead of logarithms for the independ- ent variable 3. The use of logarithms facilitates comparisons of pro- portional (percentage) changes between two variables rather than changes in absolute terms. Where there are such large differences in values over a period of years as occurred in this study, the relation between the percentage changes of two variables would seem to be more significant than just the relation between changes in actual values. Also a higher correlation was obtained by using logarithms (in some of the charts actual values are indicated on the scales to cor- respond with the logarithms used). The relation between supplement prices and corn supplies per animal unit did not appear to be propor- tional; hence the use of numbers of bushels instead of logarithms for that factor. With this method a large fraction of the year-to-year variation in the weighted average composite price of protein supplements has been explained. It remains to explain the variations between changes in the price of soybean meal and the changes in the composite price of the protein supplements. Actually from 1939-40 to 1948-49 the price of 41-percent soybean meal at Chicago varied less than $2 a ton from the computed weighted composite price of all the protein supplements. Interpretation of analysis. A price (in logarithms) that corre- sponds with the anticipated disposable personal income (also con- verted to logarithms) is read on the left-hand scale of Fig. 2. Let us say that disposable income is expected to be 199.5 billion dollars. The logarithm on the lower horizontal scale is 2.30. The logarithm of price that corresponds to that (intersection of diagonal and vertical lines) is 1.992 (or $98.18). If we assume that protein-supplement sup- plies are equal to 1948-49 supplies (0.0636 ton per grain-consuming animal unit), we deduct from our original logarithmic value of 1.992 the fraction 0.064, which is the vertical distance from the diagonal line to the zero line (at the logarithm of 0.0636 on the horizontal scale, or 0.80) in Fig. 3. This leaves 1.928 (or $84.72). If now we assume a bumper crop of corn amounting to 25 bushels per animal unit, we deduct 0.045, the vertical distance from the diagonal line to the zero 1951} WHAT DETERMINES SOYBEAN PRICES 175 line in Fig. 4, from the 1.928, leaving a logarithmic value of 1.883 (or $76.38) as the estimated composite price of all the protein supplements when converted to a meal-equivalent basis. The slope of the line in Fig. 2 indicates that a rise or fall of 0.100 in terms of logarithms of disposable personal income (in billions of dollars) was associated with a rise or fall in the same direction of about 0.119 in terms of logarithms of prices of all the protein supple- ments (in dollars per ton) . That means that for every rise of 1 percent in disposable personal income, the composite price of the protein sup- plements (combined on a meal-equivalent basis and each weighted according to actual production) rose almost 1.2 percent. The declines were in the same proportion. SUPPLEMENT COMPOSITE PRICE DOLLARS LOGS 100.00 89.10 79.40 70.80 63.10 56.20 50.10 44.70 39.80 35.50 31.60 28.20 25.10 2.00 1.95 1.90 1.85 1.80 1.75 1.70 1.65 1.60 1.55 1.50 1.45 1 Art /, '**/ / / / '48 / / X /& } / /, \y / / k / / '41* * 36 . / / / ^ 3 . 5 '3 f ^ 2 1.75 1.80 1.85 1.90 1.95 2.00 2.05 2.10 2.15 2.20 2.25 2.30 LOGARITHMS 01 CO <0 >o d K> tO f- * - O 04 0> O O CM N CO O = BILLIONS OF DOLLARS DISPOSABLE PERSONAL INCOME (U.S.), (OCT.- SEPT.) Relation between disposable personal income, United States, and composite price of protein supplements (meal-equivalent basis) : 1935-36 to 1948-49. (Fig. 2) 176 BULLETIN No. 546 [March, PRICE RES LOOS .110 .100 .090 .080 .070 .060 .050 .040 .030 .020 .010 -.010 -.020 -.030 -.040 -.050 -.060 -.070 -.080 -.090 -.100 -.110 -.120 -.130 _ IAA DUAL: I ^6 *S5 ^>^ k --\ "^ ; "\ S7 46 ^>v. ^ 47 > ^^ ^^ ^^ """-^ "41. ^^ '40 ^s . ^45 N / \ / 'I 8 / *43 / / \ &2 / ^ 44^. .57 .59 .61 .63 .65 .67 .69 .71 .73 .75 .77 .79 .81 LOGARITHMS .037 .039 .041 .043 .045 .047 .049 .051 .054 .056 .059 .062 .065 TONS PROTEIN SUPPLEMENT PER ANIMAL UNIT Relation between composite prices of protein supplements and supplies of protein supplements (meal-equivalent basis) per grain-consuming animal unit except horses and mules: 1935-36 to 1948-49. (For convenience, the characteristic 8 minus 10 for the logarithms of the supplies has been omitted.) (Fig. 3) In Fig. 3 the deviations of prices from those explained by the size of the disposable personal income (the vertical differences between the plotted values and the corresponding values on the diagonal line in Fig. 2) are related to the production of all the protein supplements. This chart indicates that as the quantity of protein supplements (meal- equivalent basis) increases 1 percent, the price of protein supplements declines approximately 0.43 percent. Thus the influence of a change of 1 percent in the production of protein supplements on changes in the price of the protein supplements was about one-third as large as the influence of a change of 1 percent in disposable personal income. In Fig. 4 the unexplained portion of prices (measured as vertical distances between the actual points on Fig. 3 and corresponding values 1951} WHAT DETERMINES SOYBEAN PRICES 177 PRICE RESIDUA LOGS .110 .100 .090 .080 .070 .060 .050 .040 .090 .020 .010 -.010 -.020 -.030 -.040 -.050 -.060 -.070 -.080 -.090 - .100 .1 10 - .120 . ISO LS 36 X X. ^x \ X, ^ X 47 X 46 X ?v \ \ T ^ k x ' 38 * X x 0 1.70 1.80 I.9O 2.00 2.10 2.20 2.30 LOGARITHMS >.9 50.1 63.1 79.4 IOO.O 125.9 158.5 199.5 BILLIONS OF DOLLARS DISPOSABLE PERSONAL INCOME (U.S.), (OCT. -SEPT.) Relation between composite prices of four edible fats and oils and dispos- able personal income, United States: 1929-30 to 1948-49. (Fig. 6) protein supplements, we obtain the estimated composite prices for these fats and oils. It is then necessary to estimate the price of soybean oil. 1 In recent years soybean oil has made up an increasing share of the total supplies of the four edible fats and oils: soybean oil, cottonseed oil, lard, and butterfat (Fig. 8) until the composite price has become rather heavily weighted by soybean-oil prices, and the actual soybean- oil prices have come closer and closer to the composite price. In recent years soybean-oil prices have averaged approximately 55 percent as high as the composite price of the four edible fats and oils. The composite price has been higher because it has included prices of butterfat. If, compared with the most recent five-year aver- 1 Prices used in this study refer to crude soybean oil in tank-car lots at midwestern mills, weighted by volume of sales. 182 BULLETIN No. 546 [March, age, the supplies of soybean oil plus cottonseed oil increase much more or decrease much less than the supplies of lard plus butterf at, soybean- oil prices will drop until they are around 50 percent as high as the composite price. With opposite supply relationships, soybean-oil prices may go as high as 60 percent of the composite price. In 1948-49 soy- bean-oil supplies plus cottonseed-oil supplies made up 50.6 percent PRICE RESIDUALS FROM FIG. 6 uOGS .20 .10 -.10 -.20 .1 3. X 38 X 30 J9 : ^r ^ <<5s ^s X '? 50 .65 .70 .75 .80 .85 LOGARITHMS 33 4.47 5.01 5.62 6.31 7.08 BILLION POUNDS TOTAL SUPPLIES OF FOUR EDIBLE FATS AND OILS Relation between the composite price of four edible fats and oils (soybean oil, cottonseed oil, lard, and butterf at) and total supplies: 1929-30 to 1948-49. (Fig. 7) of the supplies of the four edible fats and oils and soybean-oil prices were 52.8 percent of the composite price. In 1947-48 figures were 48.0 for supplies and 63.4 for prices. In 1946-47 the comparable figures were 44.0 and 61.3. For forecasting purposes it would be necessary to estimate carry- overs and production of all four edible fats and oils. Some knowledge of probable imports would help, but for these fats and oils imports are not large enough to exert much influence on soybean prices under present-day (1951) conditions. Their influence has therefore been ig- nored in this study. It would also be necessary, for forecasting purposes, to have an 1951] WHAT DETERMINES SOYBEAN PRICES 183 PRICE OF SOYBEAN OIL AS PERCENT OF COMPOSITE PRICE SUPPLIES OF SOYBEAN PLUS COTTONSEED OIL AS PERCENT OF FOUR FATS AND OILS 30 20 10 '35- '36 '37 Supplies of soybean oil plus cottonseed oil as percentage of supplies of four edible fats and oils; and price of soybean oil as percentage of composite price of four edible fats and oils (soybean oil, cottonseed oil, lard, and butterfat) : 1935-36 to 1948-49. (Fig. 8) estimate of the level of disposable personal income. Such estimates are made from time to time by agencies of the government and can be obtained from that source. Soybean Prices as Related to Value of Meal and Oil As processing techniques have improved, the number of pounds of oil obtained from a bushel of soybeans has increased and the number of pounds of meal has decreased. A reasonable figure to use at present is 10 pounds of oil and 48% pounds of meal. Some processors get more oil, hence correspondingly less meal, by using new equipment and the latest methods of extraction. Soybeans are worth the value of the oil and meal less all transpor- tation, handling, and processing costs from the local bean market until the oil and meal are processed and sold by the processor. 184 BULLETIN No. 546 [March, The weighted crop-year Illinois farm prices of soybeans and the differences between the value of the products meal and oil and the price of soybeans are shown in Fig. 9. 1 A logarithmic, or ratio, DOLLARS PER BUSHEL VALUE OF MEAL AND OIL >i / \ ILL. FARM / / \ PRICE OF ~r 1929-3031 Relation between the value of soybean meal and soybean oil and the Illinois farm price of soybeans: 1929-30 to 1948-49. (Fig. 9) 1 The margins shown here are a little larger than actual margins because soy- bean and oil prices are weighted (seasonally) by sales, but the annual price of meal is a simple average of monthly prices. Processors have sold large quantities of oil and meal directly to users in early winter to be delivered late in the year. The annual price should therefore give more weight to October, November, and December prices. Because spot sales of oil instead of forward sales were some- times used as weights, and because meal prices were not weighted by sales of any kind, the margin is too large. This does not, however, affect the usefulness of the data used for the purpose of forecasting the prices of the soybean meal, oil, or beans. The bias tends to be uniform throughout. 1951] WHAT DETERMINES SOYBEAN PRICES 185 scale is used. If the margin had increased in proportion to the price of beans, the two lines would have had the same slope; they would have remained equidistant regardless of whether prices were rising or falling. Except for 1946-47, the first year after the removal of price ceilings, the margin has been rather stable in dollars but has declined as a percentage of the price of soybeans. Doubtless "sticky" freight rates, improved methods of processing, and competition have tended to hold down the margin in this period of rising prices. For the period studied, excluding the war years, the marketing and processing margin has tended to fluctuate around a figure equal to 25 cents plus 25 percent of the value of meal and oil combined, as used in this analysis, less 2 cents a year beginning with 1935-36 as year 1 and counting the war years. The agreement between actual margins and margins estimated by this formula is fairly close. There appears to be a gradual decline in margins compared with actual values of the products. This decline is represented by subtracting a value that in- creases 2 cents a year beginning with 1935-36 as year one. There also seems to be a base below which margins do not fall. The base used here is 25 cents. There also appears to be some adjustment necessary for changes in costs of handling and processing caused by inflation or deflation, which is accounted for by using 25 percent of the combined value of the products. The above formula cannot, however, be used over a long period of time because it includes a constant decline, which would eventually mean no margin. The soybean processing industry is in a dynamic stage of development, so the factors affecting margins should be re- viewed frequently. The windfall profits or losses resulting from rapid changes in prices of the products after harvest are received by the owners of the product. If the farmers sell their beans at harvest time and the processors sell the meal and oil for future delivery to actual users, then the product users make the gains or incur the losses. Seasonal Pattern of Soybean Prices For the period 1930-31 to 1940-41 the lowest prices received by Illinois and United States farmers for soybeans tended to be in Oc- tober. The highest prices were in May and June. The monthly figures, calculated as percentages of the average yearly price after making ad- justments for changes in demand and year-to-year changes in supplies, were as follows: October, 82; November, 89; December, 96; January, 186 BULLETIN Xo. 546 [March, 102; February, 103; March, 107; April, 110; May, 115; June, 114; July, 105; August, 90; and September, 88. This pattern has changed as production has increased and market- ing procedures have changed. From 1931-32 to 1935-36 the October low was about 78, from which there was a steep rise to about 122 in June. July was only a little lower than June. After July there was a very rapid decline. From 1936-37 to 1940-41 the seasonal fluctuations were somewhat less pronounced, and the peak was reached earlier in the season. The October price was about 85 percent of the crop-year average, and the peak was reached in May at approximately 114 percent of the yearly average. No consideration is given to the war years because the OPA did not permit the usual forces to operate then and little seasonal variation occurred. Since abandonment of price ceilings, the price movements have been erratic, but the amount of seasonal spread has been about the same on a percentage basis as in the period from 1936-37 to 1940-41. The timing of the peak in prices appears to have advanced a little to April or May, and a larger frac- tion of the rise has occurred by January than was true before the war. The postwar seasonal price pattern would be approximately as follows: October, 85; November, 100; December, 100; January, 105; February, 100; March, 110; April, 114; May, 105; June, 105; July, 102; August, 92; September, 82. The heavy penalty on October sellers has persisted into the postwar period. Since there may be some tend- ency to return to prewar seasonal patterns, the figures for the two periods are averaged here: October 84 January 103 April 112 July 103 November 94 February 102 May 110 August 91 December.. ... .98 March 108 June 109 September 85 The reasons for this very large seasonal price spread have been explained. 1 The spread is believed to be related to and largely caused by the heavy marketings by farmers at harvest time, combined with the inability of processors to hedge their purchases to advantage in the "futures" markets. Processors have been obliged to carry the price risk, which they are not prepared to do, or to sell oil and meal to actual users for later delivery. These users have demanded and obtained substantial price concessions for carrying the risk. The seasonal move- ments of soybean-meal and soybean-oil prices have been small com- pared with the seasonal movements of prices for the beans. 1 Hieronymus, T. A., and Jordan, G. L., Farm Storage of Soybeans Carries Small Price Risk. Illinois Farm Economics, University of Illinois, October, 1949. Reprinted in The Soybean Digest, January, 1950. 1951] WHAT DETERMINES SOYBEAN PRICES 187 The farmer is in a better position than the processor to carry this seasonal price risk and should store more beans on the farm rather than sell such a large fraction at harvest time. Unless interfered with because of war influences, increased participation in the soybean "futures" markets, particularly by speculators who are willing to buy when processors wish to sell, and vice versa, will, it would seem, facilitate hedging by processors and thereby reduce the seasonal price spread. More farm storage would have the same effect. But for the next few years the forces which have caused this wide seasonal spread will continue to operate in the absence of government price controls such as price ceilings or price supports above the freely competitive market price. It should be emphasized that the seasonal price movements re- ferred to here are movements after the influence of changes in demand have been eliminated. A rise in demand for meal or oil would cause prices to rise more than seasonally or fall less than seasonally, and a decline in demand would have the opposite effect. Likewise, a very large crop of soybeans could seriously depress prices at harvest be- cause of lack of storage space. The seasonal rise would tend to be larger than normal. Summary Prices received by farmers for soybeans depend on the prices that consumers will pay for meat and other animal products, margarine, and vegetable shortenings; on the size of the livestock population; the supplies of protein supplements, corn, and edible fats and oils; and the costs and profits of handlers, transportation agencies, and proces- sors. The dominant causes of year-to-year variations in soybean prices are changes in the ability of the United States consumer to buy these related products and in the quantities of protein supplements and edible fats and oils available. In the absence of price ceilings or effective price supports, prices of soybean meal can be expected to change during the next few years about 1.2 times as much (in percentage) as disposable personal income, and in the same direction. Meal prices can be expected to move in the opposite direction to changes in supplies of the supple- ments used in this analysis (converted to a meal-equivalent basis) and about 0.40 to 0.45 as much (in percentage). An increase of one bushel per animal unit in the corn supply would tend to drive down the price of meal about 2.5 percent. During the next few years the actual price of soybean meal will probably be about the same as the 188 BULLETIN No. 546 [March, weighted composite price of all the supplements used in this analysis when converted to a meal-equivalent basis. Soybean-oil prices, in the absence of price controls or effective price supports, will change in the same direction and to about the same extent (percentage) as disposable personal income changes. They will also tend to change about 1.4 to 1.5 percent in the opposite direc- tion with every 1 percent change in supplies of the four edible fats and oils soybean oil, cottonseed oil, lard, and butterfat. The actual price of soybean oil will be approximately 55 percent as high as the weighted composite price of the four edible fats and oils. From the combined value of the oil and meal must be deducted the marketing and processing costs. These costs fluctuate rather widely. They have become somewhat stabilized, however, since 1934-35 around a figure equal to 25 cents a bushel plus 25 percent of the value of the meal and oil minus 2 cents for every year since 1934-35 (cumulative). Up to World War II, as soybean production expanded rapidly and the market became a little better organized, the seasonal spread in soybean prices from the October low to early summer high declined somewhat. Since World War II the October price has continued to be very low compared with the yearly average price (after corrections are made for changes in demand and production) ; but the peak has come sooner in April or May and much of the rise has occurred by January. More farm storage until January or May will probably pay growers for several years unless we get into a period of price controls. In case price controls are applied, some consideration will have to be given to the matter of orderly marketing, either by allowing seasonal variations in prices or by paying farmers or others a fee for storing. 1951] WHAT DETERMINES SOYBEAN PRICES 189 TABLES Table 1. Illinois Farm Prices of Soybeans, and Amount of Disposable Personal Income in the United States, 1929-30 to 1948-49 Year Illinois farm price of soybeans* Disposable personal income b Year Illinois farm price of soybeans* Disposable personal income 1 " 1929 per bu. . . $1.61 billions $75.9 1939. . . per bu. $ .77 billions $ 74 1930 1.18 65.7 1940 85 86.9 1931 .35 51.6 1941 1.54 108.8 1932 . . 47 45.8 1942 1 60 131 9 1933 71 50.0 1943 1 . 80 141.3 1934 .86 56.4 1944 2.04 151.8 1935 68 64.1 1945 2 08 154.8 1936 1.18 69.8 1946 2.52 166.2 1937. .80 66.9 1947 3 39 184 1 1938 64 69.0 1948 2.40 189 4 From Illinois Agricultural Statistics, No. 445 (p. 130), Illinois Cooperative Crop Reporting Service. Crop years 1929-1944 begin in September; 1945-1948 begin in October; all are weighted averages. b From U. S. Department of Commerce, Bureau of Foreign and Domestic Commerce, Survey of Current Business (National Income Number), July, 1950 (pp. 30-31). Crop year is computed by inter- polation, except for 1939-1948. Table 2. Average Price per Ton of Protein Supplements, Bagged, in Wholesale Lots at Leading Markets, 1935-36 to 1948-49" Peanut Year Soybean 35- P t < cSio Linseed b meal, Minne- apolis Cotton- seed meal, 41 percent, Memphis meal, 45 percent, f.o.b. ' south- eastern milling Gluten feed, 23 percent, Chicago Copra meal, b Los Angeles Tankage digester, 60 per- cent, Chicago Meat scraps, Chicago Fish meal, 67 percent, San Fran- cisco points 1935. $28.66 $31.04 $24.33 $24.17 $23 . 52 $24.95 $48.15 $36.21 1936. 40.61 40.80 32.88 35.69 31.57 34.81 55.69 44.73 1937. 27.71 39.71 22.01 24.98 22.58 28.73 46.01 42.48 1938. 25.98 36.90 22.50 22.06 18.78 24.74 52.57 42.57 1939. 28.90 29.89 27.99 29.36 22.22 22.95 49.66 51.20 1940. 30.49 29.65 28.50 25.56 23.17 29.32 53.14 61.46 1941. 41.87 37.56 36.33 40.57 28.69 47.33 74.15 74.61 1942. 42.80 44.79 40.10 40.57 32.39 52.30 74.55 79.50 1943. 51.91 45.50 48.50 53.00 40.94 50.98 74.55 79.50 1944. 52.00 45.50 48.54 53.00 41.27 49.92 74.55 $72.90 79.50 1945. 62.39 55.57 60.08 61.32 48.85 53.58 82.75 81.77 83.46 1946. 81.10 81.54 75.83 74.01 58.28 70.85 110.63 101.80 159.52 1947. 91.61 81.27 83.20 83.19 74.47 85.33 119.39 110.82 163.19 1948. 76.39 67.24 63.12 65.48 53.78 67.34 120.65 116.54 193.45 Figures are from U. S. Bureau of Agricultural Economics, Feed Statistics, 1946 Annual (p. 60), and 1949 Annual (pp. 71-76). The percentages refer to protein content. b Variable protein content. 190 BULLETIN No. 546 [March, j,-. o-*moot-ci(NOfOooasc<3>fla> o g| 8o{2SSoSSioE5?o?[2 1 1>"S ooooooooomoswo-* a o Qt t |I| ** S * & CO <-l OJ 00 * 1-1948 are rei oi OOOOO-H-* S O VO CO dp if (NOO^COO i vrj co O\ f o> 3 H o * c^ OJiOtOt-00 c? a & a|s ziiiii to ~ to < <* 3 8 ^ 5 o O s! 1 nual (pp. 59 d Statistics, Economics). "- 1 *H CO ^ 0) fl EOS CO i a s Il!!!!i 03 ^ y * T3 . *J 1 , g S5gSooS252S22 ^ g < 9< -*j *""a tO) OS tO5 * rf S I 3 ~ w g "C a" OQ 3 -O a Ills 3* S-g so O <-M O c J| S ral Economi 3 estimated s al United St t scraps (U. 'roductio J; lO l OlNOO>OOOOCNCO^OO 'o . jj g *T 1? il** g m O O IN 00 O OS >O CO CO CO O CO CN CO >O t- 1~ Tf ??i 2 1 i IS 2-a . . 2 fp >^* K^ ^^ ** o **" o> o t. p~co g J3 c/a 2 U J3^ a go^ M . 3 _ T3 "S ol> *> IN 00 1-1 r-< CO O CO O 1-1 IN O i-i **" O 00 M< IN t- CN t .i||| U eo * a -- t>. CO O 00 1-1 1^ I-H IN O> CO t~ CO t~- O! O O CO * IN CN CO tN CN T-I CN IN CO CO O m lO t 00 CO ?2a III 1 1 jsg 3 >-i . 2 %lll 2tf 11 ssasssssss&sssssssss 'StO g 2^1 1 nj C J3 3 C^OcOcOt^C^WO5t^cDOOcO"5iCiOXt*COt* a& S 2^ ^r5 X i^^H i trH^H f-Hr-tr I-^^HC^C^ ( M * k< ^ "? w *^ cu S3 2 ^H* ^ O 5 ^ y m rt 3 K a S J3 ""8 s 'C .2 fe CQ 111 " '-S r * r2 | T) ^lfe% ^0 Oi *- ( O5 f-i t* 00 CO O O CD ^ ^ t iO iO 1C iO 00 iO (N .* J3 22 o oljj- | 'i * W-o f 1 ** *5 O -gft S t>- CD CO CC * 00 X O5 CD ^ iO O (N N (N M W Tf O 1C Hi-tr-tr-ii-lOKN! ( ||| fcn O - ^ 3 "S ^ S C3 ^ 2^^ '-co X. > -rt- "73" ^ O> '"ft ~o*> -o . D.-OCN S -o ^ c . > S "oO w 'O |g1 *o Oi CO ^ O CO CD CM O C^ CO O ^ (N iO CO I-H CD CO O "^ O CO Tt* CD 00 t>- Tf O t^- 00 OS O (N t t>- 1> OS 00 1^ i-t c: 02 c a a-- 1 3^ o S "C g O . -i i-" O 1 " ^^Hr-I^^CSWrH l|||l| I o V M lib ~ B 3 K? S"8 PH MO