ELASTICEW OF SPRENG WHEAT ACREAGE RESPONSE [N SWEDEN UNDER A FGRWARD PfiICMG $YSTEM Thesis for fine Degree .0? M. S. MICHEGAN STATE UNIVERSITY Erik Eliasson 1958 THESIS ELASTICITY OP SPRING WHEAT ACRENGE RESPONSE IN SHEDII'UNDIR A FORWARD PBICING SYSTEM by EBIKIELIASSON AN ABSTRACT Submitted to the College of Agriculture of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Agricultural Economics 1958 W 4:22. g é mtg, ABSERACT During the period from l9h3 to 1956, Swedish farmers operated under a pricing system in which the price to be paid for certain crops was announced before the farmers planted their acreage. This situation pro- vides a unique opportunity to study the effect of different years' prices on the allocation of the crop acreage in a given year. The obJectives of the study are mainly to investigate: l. The extent announced price for the present year and previous years' announced prices have influenced the acreage allocated to growing spring wheat. 2. The extent other selected factors have affected the acreage planted to spring wheat. 3. Whether or not different growing areas in Sweden responded in a similar manner to the various factors affecting the acreage allocated to the production of spring wheat. Due to time limitations this study has been limited to the consider- ation of only one crop, spring wheat. The relationship between crop acre- age and the various determining factors have been investigated using the techniques of multiple regression. Theresults of this investigation indicate: 1. The acreage of spring wheat is influenced not only by announced price in the current year but also to some extent by previous years"prices. 2. The acreage of spring wheat during the period l9h3—l956 has been influenced by price, trend, last year's yield of spring wheat, and the extent to which winter crops were destroyed during preceding winter. The results suggest that an increase in the price of spring wheat of 10% would 5 on the average, give an increase in spring wheat acreage of about 37,000 hectares if the new price level holds for at least three years. The trend, which includes mainly the effect of the technological development during the period, was very significant. The acreage of spring wheat had, on the average during the period 19h3-1956, increased about 5,000 hectares (3.5% of the average acreage) per year due to the trend. 3. Different factors are of importance in different parts of the country. The results obtained suggest that the farmers in the Southern part of Sweden are less responsive to price in growing spring wheat than agricultural producers in Middle Sweden and more responsive to last year's yield. The trend has been important in each of the investigated areas but exerts influences of about twice the magnitude in.Middle Sweden as compared to Southern Sweden. ELASTICITY 01' SPRING WHEAT ACRRAGE RESPONSE IR SWEDEN UNDER A FORWARD PBICIM SYSTDI by ERIK ELIASSOH A THESIS Submitted to the College of Agriculture of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Agricultural Economics 1958 / / - - , '1: (L:- 7 1', I) ,1.) ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Dr. Dean E. McKee for his guidance throughout the study of this thesis. His interest in the study, advice when particular obstacles arose and patience during the many hours of consultation were a constant source of encouragement and help to the author. The burden of the language corrections which have been necessary because English is a foreign language to the author has also been carried‘by him. Special thanks are also due to Dr. Lawrence Witt who guided most of the author's graduate program.at Michigan State University The author has benefited from.many discussions with graduate students of the Department of Agricultural Economics, especially kyle rettig and Lee Schrader. The author wants to express his sincere appreciation to Mrs. Joann Prendergast who has done all the typing and to the personnel in the statistical pool of the Department of Agricultural Economics who have perfermed the statistical calculations. Sincere appreciation is also expressed to the W. K. Kellogg Founda- tion, USA, and Helge szson Johnson's Stiftelse, Sweden, whose generosity made possible the whole experience of studying and living at an American university. The author's deepest debt of gratitude is to his wife for her encouragement and patience during the work with this thesis. CHAPTER I. II. III. VII. TABLE OF CONTENTS PAGE Immon O O 0 O O O O O O O O O O O O O I O O O O O 1 ummmeeeeeeeeeeeeeeeeeeee5 CHARACTERISTICS OF CROP PRODUCTION IN SWEDEN . . . . . . 21 m Clmto. 0 O O O O O O O O O I O O O O D O I O I I 21 'rhe Sons. 0 O l 0 O O o O O O O O O O O O O I O I O O 21 AcreageofDifferentCrops..............22 Planted versus Harvested Acreage . . . . . . . . . . .. 26 Average Yield. 0 O O O I O O O O O O O I O O O O O O O 27 Mechanizatlon.....o.o.............29 crOpBOtationeeeeeeeeeeeeeeeeeeeee31 Some Geographical Characteristics. . . . . . . . . . . 32 m AGRICULTURAL PRICE POLICY IN SWEDEN, who-1956 . . . 35 The Principles for the Price Setting System. . . . . . 5 Fomrd Prices 0 O O O O O O O O O O O O O O O O O O O hl DIFFERENT FACTORS AFFECTING THE SIZE 01' THE ACREAGE OF A CROP O O O O O O O O O O O O O O O I O O 0 O O O O O I O 1‘3 The Effect on Acreage Allocation of: 1. Changes in the Prices of Alternative Crops . . . Ah 2. Changes in the Size of the Yield from Various Cropseeeeeeeeeeeeeeeeeeeeee1‘7 3. Changes in the Price of Various Inputs . . . . . 1+8 1}. ChangesinTechnology.............. A9 TEEDATAANDHOWTEEIAMOBTAINED...........51 AcreagoData.....-oo.............51 Pficenataeeeeeeeeeeeeeeeeeeeeee52 PriceIndexes........... . ......53 Selection of Index to Deflate the Prices With. . . . . 55 SIATISTICALANALISIS..................6O l. The Influence of Different Years' Prices on Acre- agonlocationeeeeeeeeeeeeeeeee61 2. The Extent to Which Different Factors Affect the AllocationofAcreage.............. 73 3. Differences in Acreage Response Within Various Parts of Sweden and Between USA and Sweden . . . 79 SW m CONCLUSIONS. 0 O O O O O O O O O O O O O O O 90 mm 0 O O O O O O O O O O O 0 O O O O O O O O O O O 96 BELIW O O O O O O O O O O C C O O C O O . . C O O LIST OF TABLES TABLE PAGE 1 Summary of the Relationships and Response of Acreage of Various Crops to Various Price Factors. . . . . . . l2 2 A Comparison of Two Methods (Special and General) for Estimating the Elasticity of Supply as Measured by the Response of Acreage to Expected Price for Cotton, Wheat ”acorn‘19w-1932)eeeeeeeeeeeeeeeee 18 3 Acreage of Important Cash Crops, Fellow and Total Ara- ble Land, 1918-1956, Thousands of Hectares . . . . . . 22 h Wages for Farm Labor During 1953 in West European Comtries....c..................25 5 Percentage Acreage on which the Crop was Destroyed mnetmwmtereeeeeeeeeeeeeeeeeee27 6 Average Yields of Important Cash Crops Dm'ing 1913-1956 in swdono O O O O O O O O O O O O O O O O O O O O O O 28 7 Annual Consumption of Artificial Fertiliser in Pounds per AC” 0 O O O O O O O 0 O O O O O O O O O O C O O 0 a 8 The 'Number of Tractors Used on Swedish Farms and the Prices of Tractors and Farm Machinery Relative to the wage Rate for Hired Farm Labor, Every Second Year, 19h3'19560-eeeeeeeeeeeeeeeeeeeeee 30 9 Million Man-Hours in the Swedish Agriculture . . . . . 30 10 Acreage Planted in the Fall of Oil Crops (Winter Rape and Winter Rybs) , l9h2-l956, in Three Different Dis- tricts of Sweden, Thousands Hectares . . . . . . . . . 33 ll The Development of the Production of Bread Grains and the Prices of Wheat 19h8-h9 - 1952-53. . . . . . . . . no 12 The Development of Milk Production and the Price of Butter 1911-8-19 "' 1952-53-e e e e e e e e e e e e e e e h]- 13 Price Indexes of Some Production Factors in Agricul- h tm O O O O O O O O O O O O O O O O O O O O O O O O O 9 lh Announced and Final Price of Spring Wheat 19143-1956, on ”r m0' 0 O O O O O O O O C O O O O O O O C O O O 53 LIST OF TABLES - continued TABLE PAGE 15 TheProducerPriceIndex............... 55 16 Sumary of the Statistical Results Obtained at Increas- ing Influence of Previous Years' Prices. . . . . . . . 68 17 Acreage Response to Various Factors Within Different Parts Of swdon. O O O O O O O 0 O O O O O O O O O O 0 8h 18 Price Elasticities Obtained from Swedish and American Data 0 O C O C O O O C O C 0 O O O O O O O O O O O O O % Data Series Used in the Analysis I (Appendix). . . . . 97 Data Series Used in the Analysis II (Appendix) . . . . 98 DIAGRAM LIST OF DIAGRAMS PAGE Acreage Winter 011 Crops (5-Year Moving Average) in Shane, Gotaland except Shane, and Svealand . . . . . 31* Relationship Between the Estimated Elasticity and the Number of Years Included in the Moving Average. Price. 0 O O O O O O O O O O O O O O O O O O I O O O p 67 Actual and Calculated Acreage of Spring Wheat: meI 12 O O O O O O O O O O O O O O O O O O O O O O 76 UnexplainedBesiduals................ 83 INTRODUCTION Many of the difficulties in agriculture arise due to the low elasticities of both supply and demand for agricultural products. The demand for most agricultural products has a very low price elasticity, especially in highly developed countries. Consumers buy and consume the basic foods with little consideration to price. when production increases and supply tends to exceed demand, large price reductions are necessary in order to equate supply and demand. Higher production is, therefore, often synonymous with lower aggregate income for the farm population. The opposite is also true. During periods of shortages the price of food tends to be very high if not regulated. Agriculture is also faced with a low income elasticity for its products. Increased incomes for consumers will only to a very limited extent go to agriculture . Consuners in developed countries are already consuming as much food as they want and they prefer to use their increased 1ncomes for other things. Mainly other parts of the econow will benefit frcm in- creased consumer incomes and the agricultural population will share in the higher living standard only to a limited extent. Low price and income elasticities are two important characteristics which to a large extent determine conditions in agriculture. These, how- ever, consider only one side - the demand side. The nature of supply is equally important. This is a fact which sometimes has been overlooked in the agricultural policy. In order to help the agricultural population to overcone the handicap of the low price and income elasticities of its products and to get a "fair" share of the increase in the living standard, agricultural prices have been supported in most countries. Higher and more stable prices have, together with technological development, increased supply over time. The main agricultural problem in many countries is one of overproduction because too little attention has been given to the effect of higher and more stable prices on the supply side. For several coun- tries it has become obvious that all of the support to agriculture cannot be given via prices without production control. Resources which other- wise would have been employed in other industries will be attracted to agriculture and, hence, give rise to loses both for agriculture and the whole of society because of wasted resources in agricultural overproduction. In order to determine how much support can be given via prices, it is necessary to know what supply can be expected at different prices. In- creased tnowledge about the nature of the supply response must, therefore, be considered as important. It is likely that the great agricultural problems in the future will have their solution on the supply side rather than on the demand side if rational solutions shall be attained. Supply response in agriculture has received comparatively little attention in earlier research. Much more has been done on the demand side. This is a result partly because of the difficulties in estimating and measuring supply response in agriculture. Supply response in agriculture has long been a subJect of controversy due to this fact. The greatest controversy exists regarding the total agricultural supply but there is disagreement even at the single conodity level. During recent years several conendable studies have been published, within the area of supply response to price, for single agricultural coalsodities. One of the problems in relating price to supplied quantity (acreage) is what price to use. The normal approach has been to relate last year's price to the present year's acreage of the crop which is studied. Herlovel argues that the farlmrs do not normally assume that last year's price will prevail for the coming year. Instead he suggests that the farmers respond to an expected price which is dependent not only on last year's price but also to some extent on previous year's prices. This study deals with acreage response to price and other factors in Sweden. In connection with expected prices as discussed by Herlove, it should be particularly interesting to study this response in Sweden, where the prices of most crops were announced prior to planting during the period under study, 1918-1956. This announced price is assused to be largely synonymous with farmers' expected prices. This circumstance will mahe it possible to study the influence from different years' prices on acreage allocation in the present year. Obggctive . The obJectives of the study are mainly to investigate: l. The influence of different years ' prices on the acreage allocation in the present year. 2. To what extent non-price factors affect acreage allocation in Sweden. 3. Acreage response in different areas of Sweden. The study is intended as an introduction to a more extensive study of the agricultural supply in Sweden. The calculations are limited to only one crop - spring wheat. Chapters I - V present the background for the statistical analysis in Chapter VI. Chapter I contains a review of the literature concerning 18ee Chapter I . supply response in agriculture. Some characteristics of importance for supply response in crop production in Sweden are described in Chapter II. A survey of the background of Swedish price policy and some of the effects of the pricing system of mnounced prices is given in Chapter III. Chapter IV contains a short analysis of how different economic factors in- fluence the allocation of spring wheat acreage. The data, how they are obtained, their limitations, and problems in connection with the selection of the index for use in deflation of the price data are given in Chapter V. Chapter VI contains the statistical analysis made in the study. CHAPMI LITERATURIREVIEW In this review of previous research in this area are included mainly those works that have contained motheses, observations, and results, which, according to the writer's opinion are of interest in analyzing the Swedish data. In 1925, Smith published an article about how cotton acreage was related to economic factors.l He studied the relation between absolute changes in cotton acreage to prices during the five months (November- March) before the planting of cotton. The prices for each month were de- flated by a wholesale price index of agricultural comedities for the same month. When Smith made this study no index of prices received for farm products was available. Instead he computed an index of prices based on data from the New York and How Orleans markets. By deflation with this index, Smith intended to remove (1) the effects of changes in the general price level, and (2) the influence of changes in the prices of other crops which might compete for acreage with the crop he was studying. ’ Although anith worked with the absolute changes in acreage as the dependent variable in his regressions, he did adJust for trend changes. On the basis of his estimates, Smith recanended that the January price and the lagged first difference of cotton production should be used as independent variables for forecasting purposes. In this regression he 1B. B. hith, "Forecasting the Acreage of Cotton," Journal of American Statistical Association, new Series, Vol. 20, (1&5), pp. 3147. found an R2 I 0.98. In including lagged production, Smith wanted to remove the effect of that part of the price which was considered abnormal. Smith continued to work with this problem and in 1928 he published a work dealing with both the supply and the demand for cotton.1 In this analysis Smith used the technique of graphical curvilinear correlation. An index of multiple correlation of 0.95 was obtained when the percentage change in cotton acreage (1903-1926) was related to the average Decaaber price of March futures in How York. The prices were deflated by the Bureau of Labor Statistic's index for farm product prices. Also, the per- centage change in cotton acreage was related to cotton price lagged one year and to the percentage change in cotton acreage during the year pre- ceding the harvest year. As a Justification for including the change in acreage which took place the preceding year, Smith says: Yet another factor which is desirable to enploy in analyzing the change in acreage that takes place in am given year is the change that took place in the preceding year. For, other things being equal, a reaction from the change taking place in the pre- ceding year is to be expected. Chiefly because agricultural pro- duction is practically never in precise adjustment to demand. The coefficient for lagged average change obtained by Smith was, however, negative. This appeared to be due mainly to the choice of a future price rather than a prevailing price. Bean3 published, in 1929, an article in which he gave the results of a study‘on changes in the acreage of potatoes, sweet potatoes, cabbage, 3's. B. Smith, Factors meetinithe Price of Cotton, USDA Tech. Bul. Ho. 50, (1928). 2‘1'.bid., p. 21. 3L. H. Bean, ”The Famrs' Response to Price," Journal of Farm Economics, Vol. 11, (1929), pp. 368-385. strawberries, watermelons, flax, rye and cotton. In the summary of the article Bean says: (1) There appears to be a general type of production re- sponse to price, cannon to each of the cases analyzed. (9.) In each case the price received for the production of the preced- ing season is the dominant factor in the change in production in any given year. In most cases the price received during the season two years preceding is also an important factor particu- larly if the price has been low. (3) Although there is a general similarity in the nature of the response, of production to the price one year preceding, the extent of that response for differ- ent prices differs by regions and by comedities, some showing a greater response to high prices; others to low prices. (A) The response to price two years preceding does not follow a general type, but differs markedly for several of the commodities. (5) For each comodity the analysis reveals that there is, under or- dinary conditions, a definite national average price which tends to maintain acreage unchanged from that of the preceding year. For arw one comedity there are different regional prices which tend to maintain acreage unchanged in the different areas. Where competing crops are involved these equilibrium prices tend to vary with the prices of competing crops. (6) Prices only moder- ately above or below this equilibrium price (3 10%) tend to be followed by about the same percentage increase or decrease in acreage but very high prices bring forth no materially greater increase in acreage than do moderate prices. (7) In some cases the effect of high prices on subsequent production lasts only one season, in others, at least two seasons. The effec of low prices in most cases extends over at least two seasons. Bean used the absolute change in acreage Inflected as the dependent variable and prices for the two preceding years as the independent vari- ables. The prices were deflated by an index of the general level of farm prices, for most crops. Bean was probably the first to use prices of competing crops to deflate the prices of sweet potatoes and flax. He also fitted the trends but this work was done in a somewhat arbitrary way Just as were some other procedures by which he obtained his results. Some of his results are, therefore, considered to be less reliable. His 11hid., pp. 38h-385. greatest contribution was that he introduced many interesting modifications in the method of analysis. In a book published by Black)" in 1929, we find: Farmers are not likely to Judge two series of high prices alike, for the reason that no two such series will be altogether alike - they will vary in duration, level of prices, sequences of changes and other accompanying circumstances. The matter of duration is of prime importance ..... The time of response will vary for different comuodities. The acreage of an mnual cash crop that requires little special equipment or skill can be in- creased quickly. But if the shift requires reducing the acreage of feed crops needed for an established livestock industry it will proceed much more slowly. If it requires largely revising a system of crop rotation, it will proceed more slowly. It is of‘special interest to notice Black‘s recognition of the different time of response to price for various products and that the size of this response depends on the length of time the price remains at this level. In 1933, T. B. Manny2 used an alternative approach in the form of an interview study in which he tried to determine: (1) Recent changes in acreages of important crops, (2) the reasons given by farmers for planting the specific acreages they did, and (3) the premises and sources of information upon which these farmers' decisions were based. The study was based on interviews with 83h farmers. A third of the farmers based their acreage decisions on price considerations. About half of this third (17.5%) said that last year's price bad influence on the acreage planted, while the rest (16.0%) said that the price they expected during 1John D. Black, Agricultural Reform in the United States, New York, 19$) PP 0 112-116 a 2r. B. Manny, rarner Qpinions and Other Factors Influencing Cotton Production and Acrgaggfltrmnts in the South, USDA Cir. Ho. 2 , 1933). the coming harvest influenced their decisions. Twenty-eight percent of the farmers claimed that their long-run plans were the dominant factor in making the acreage allocation, while the rest planted their acreage more or less haphazardly. The long-run plans were probably to a considerable extent dependent upon price relations. This result seems to indicate the importance of long-run response of acreage to price. In other words, the acreage allocation is a function not only of expected price for the coming year but also of prices during some previous years. The results also support Black's opinion mentioned above. Cassels1 took up the problem of the duration of the price and dis- cussed "short-run" and ”long-run" supply curves: Time is required for the organization of extra shifts, for the renovation of old machinery, for the augmentation of the labor force and for the assembling of additional supplies of the irnrut elements. More tire is required for new producers to come into the field and still more for the efficiency to be introduced into all the new arrangements. Cassels also discussed the possibilities that the "short-rm” supply curve was irreversible. He said: ...... that each supply curve mat be regarded as relating to an established level of output and should be recognized to have two distinct parts, one representing expansion beyond that output and the other representing contraction below it. Kuznets2 made, in the early 191+O's, a study of the supply of sugar hosts in California. He worked with a model in which the acreage of sugar beets was a function of the yield per acre of sugar beets in the 1J. M. Cassels, "The Nature of Statistical Supply Curves ," Journal of Farm Economics, Vol. 15, (1933), pp. 378-387. 26. M. Kuznets, Statistics Port to the Acre e of Beets in California, 1222-1252, University of ali ornia, ollege of Agricul- ture, Ag. Expt. Sta., Bulletin, Berkeley, 191:3. 10 preceding year and the lagged ratio of the price of sugar beets to the price of dry edible beans. In this model the acreage of sugar beets was deflated by an index of the capacity of the sugar beet processing industry. Through this procedure he obtained acreage figures which were more compar- able to the prices received during different years than by using-unadJusted figures. A preferable procedure would perhaps have been to include data on the capacity of the sugar beet industry as a separate variable. He also constructed an alternative model in which the prices of sugar beets and beans in the preceding year were separate variables. Of special in- terest is that Kuznets introduced the price of a competing crop explicitly and that he took account of the farmers' changing notions due to the size of the yield the previous year. In 191m, Walshl published a study on the supply of cotton and cottonseed. He studied the acreage response to price during the period 1910-1933. The result showed that there had been a shift of the supply schedule between 1921+ and 1925. This shift was due to the fact that cotton had expanded into new areas and the bell weevil had been brought under control. As a consequence of this shift, Walsh divided the data into two periods, l9lO-l92h and 1925-1933. In the analysis, acreage was related to the price of cotton, the price of cottonseed and to the composite price of cotton and cottonseed. The prices were deflated by an index of prices paid by farmers and lagged one year. A larger correlation was obtained for the later period than for the earlier period. The elasticities lay 1R. M. Walsh, "Response of Price in Production of Cotton and Cottonseed," Journal of Farm Economics, Vol. 26, (19%), pp. 359-372. 11 between 0.22 and 0.25. welsh also studied the effect of the corn price on the cotton acreage but did not find a significant coefficient for corn price lagged one year. In 1950, Kohls and Paarlbergl published a study on the price re- sponse of a large number of commodities. As the dependent variable, they used the percentage change in acreage from the preceding year. In most of the analyses they used data for the period l922-l9h2 including the years of depression in the 1930's. The elasticities, and for some products, the correlations, were as given in Table l. Kohls and Paarlberg cemented won the results in the following m: In the analysis of the different crops there was some evidence that farmers as a group do, or intend to, respond to changing relative crop prices from year to year by changing the acreage planted. However, the amount of variations in either intended or harvested acreage that could be explained by price changes were in litany instances quite snall...... In showing slight acreage response to year-tonyear price changes, farmers are behaving intelligently and in their own interests. A high price for a crop in a given year is little indication that the price will be high in the following year. Consequently, a close response of acreage to price from year to year for crops would mean unwise allocation and lower income for farmers. D. Gale Johnson2 has developed a theory of aggregate supply response in agriculture which rests upon the assumptions that the farmers attempt to maximize profit and that (l) the supply of labor shifts with business activity, (2) the supply of land is inelastic, and (3) the supply of capital has a low elasticity in contraction and higher in expansion. According 13. L. Kohls and n. Paarlberg, The Short-Tine Regponse of M- tural Production to Price and Other Factors, Purdue University, Ag. Expt. Sta., Bul. 50. 555, Lafayette,Tnd., 1950. an. Gale Jolmson, ”The Nature of the Supply Function for Agricultural Products," American Economic Review, Vol. ho, No. It, 1950, p. 539. 12 to his opinion agriculture is price responsive. He also points out that fixed assets and resource availability are more important in the explana- tion of output than fixed costs. Table 1. Smary of the Relationships and Response of Acreage of Various Crepe to Various Price Factors.a Percent increase in acreage with 10 percent in- b crease in factor Percent of acreage varia- tion explained Most important Crop factors explaining acreage change (10 x elasticity (32 ° 100) of ac Cabbage Price preceding season 513 2.9 Cantaloup Price one and two pre- ceding seasons 51* 3.5, 2.5 6”” P033 None found e e e e e e e e Onions Price preceding season 27* 1.9 Snap Beans Em found e as . eeee e Tomatoes None tom eee eeeee Watermelons Price preceding year 23* 2.1 Strawberries Price 1, 2, and 3 pre- ceding years 52* 1.5, 2.9, 2.9 Potatoes December price two pre- ceding years; August price preceding year 57* 0.65, 0.78 Spring Wheat March price planting year 23* 3.1 Pall Wheat Average acre value same and preceding year 29* 1.9, 2.h Corn (in- Non-Jan. price two tentions) preceding years 25* 0.7 Oats Rom fond eee eeeee *Significant. aAll prices are corrected for price level fluctuations. bWhen two or more factors are involved, not effect is given in the order of the factors mentioned; otherwise the gross effect is given. Glenn Johnsonl published a study in 1952 on the effects of govern- ment programs on burley tobacco. He used lagged price deflated by costs 101cm 1.. Johnson, Burl Tobacco Control Pro , Bul. 580, Ken- tucky Agricultural Expt. ta. , versity of K's-mucky, Lexington, 1952. 13 of production, acreage allotment minus previous six-year average acreage, and overplanting penalty to estimate underplantings of burley tobacco. Bis yield model included lagged price (in logs) , prices paid, trend, acreage allotment minus three-year average acreage harvested and a weather index computed from test plot yields. Hathawayl made an analysis of the effects of the price support pro- gram on the dry bean industry in Michigan. In order to explain three important phases in producing and marketing beans, be constructed and tested three econometric models: One for planted acreage, one for yield, and one for price received for pea and medium white beans. In the acreage model he calculated planted acreage as a function of the percentage of the previous years acreage abandoned before harvest, index of expected income from corn and wheat, cost of production, and log of price re- ceived the previous year. In his yield model he considered weather (test plot yields), log of price received the previous year and current acreage. The price model included the following independent variables: The support price, supply of pea and median white beans in USA, disposable consumer income in USA, supply of Great Northern beans in USA, and the wt of beans delivered to the government under the support program. Coefficients of determination were 0.80 for the acreage model and 0.87 for tie yield model. The 813W reveals a willingness on the part of farmers to supply more products at the same average price after the reduction in uncertainty through price support programs. 1"Dale R. Hathaway, The Effects of the Price Mort Program on the Bean Induster Michi an, Michigan State College, Ag. xpt. Sta. , April 1955. 1h Nerlovel has, in his dissertation on supply elasticities of corn, cotton and wheat, made an advanced and extensive work in the formulation of price expectation models to be used in the estimation of acreage re- sponse to price change. Earlier studies of the acreage response to price change had all suggested that farmers responded very little to price in planning their acreage. Valsh had estimated the elasticity of acreage response with respect to last year's deflated price to be of the order of only 0.2 and Kohls and Paarlberg had got about 0.07 for corn and 0.2 for wheat. Herlave maintains that these low elasticities are contradicted by experiences under the price support programs. If these very low elasticities are correct, then the support prices must have been much in excess of the equilibrium prices. To substantiate his argment, he cites the experience in 19h? and l9h8 with price support programs. In 191:8, 36% of the cotton crop, 28% of the wheat crop and 19% of the corn crop was placed under loan while in 191:7, less than 2% of the total production in USA of corn, cotton, and wheat was placed under loan. The support prices of the three comedi- ties increased by only 11-1: for cotton and wheat and not at all for corn from 19%? to l9h8. The price the farmer got in 19k? exceeded the loan rate by 36% for corn, 17% for cotton and 19% for wheat. In 19h8, corn price was 11% below the loan rate, cotton price 5% above the loan rate, and wheat price was at about the loan rate. At the above mentioned low elasticities of supply the support price of corn should have exceeded 11hrc Nerlove, Estimates of the Elasticities of m of Corn, Cotton and Wheat, Dissertation at elm Hopkins varsity, altimore, Maryland, 1955. 15 the equilibrium price by about 15%. For cotton, should the corresponding figure have been about 70%, and for wheat 50 - 60%, in comparison with the period following World War I, these figures imply, according to Nerlove, changes in the equilibrium prices of corn, cotton and wheat of an unreasonable magnitude between 19%? and 19h8. Instead Horlove finds it far more reasonable that the elasticities of supply are higher than previous measurement would suggest. The law elasticities of acreage response to price is, according to Nerlove, due to a great extent to the fact that price lagged one year has been identified with the price to which farmers react. As have Kohls and Paarlberg, Nerlove points out that farmers would not be acting in their own interest if they based their production plans only upon the previous year's price. Nerlove attempted to take this into accorurt in his model in which he assured that farmers react to an gected price which depends only to a limited extent on last year's price. This ex- pected price cannot be observed but Nerlove assrmes that it depends in some way on what prices have been in the past. Also, such factors as support programs (CCC) and acreage allotments are likely to affect the expected price. In order to avoid these later effects , Nerlove has limited his analysis to the period before 1933. In estimating the expected price, Nerlave seemed that the influence of more recent prices is greater than the influence of less recent prices. Thus, the expected price can be calculated as a weighted moving average of past prices in which the weights decline as one goes back in time. The earlier practice of representing expected price by price lagged one year is a special case of this more general Impothesis. 16 In deriving a farm for the weighting system, Nerlove made the hy- pothesis that farmers revise the price they expect to prevail in the coming year in proportion to the error they made in predicting price for this year. This mpothesis can be expressed mathematically as follows: at * i- (1) P1 - JP.“1 = fluid - Pm) where P: = expected price this year P:_1 expected price last year Pt-l I actual price last year the proportion of the error by which farmers revise their expectations. This is called the coefficient of eggctation and is asslmed to lie between zero and one. If the acreage devoted to a crap is a linear function of the ex- pected relative price of this crop then acreage this year It can be expressed as: (2) xt = a0 * 31?: + “t where P: still is the expected price this year, and Ut a random residual. P; which cannot be observed, can now be expressed as a function of acreage. Also, the expected price last year, (P:_1) can be expressed in the same way. With help of equation (2) for two years, P: and P:_1 can be eliminated in equation (1) and we will get this year's acreage (It) expressed as a function of last year's observed price (Pt-1) and last year's acreage, (xi_1) which‘both.canfibe observed. (3) Xt - T); + TEFL]. + Wext-l 4: V,c where Vt is a random residual different from Ut and/[7.2.1-fl. 17 It can be shown that the hypothesis that the farmers revise the price they expect in proportion to the error they have made in prediction is equivalent to one in which expected price is represented as a weighted moving average of past prices where the weights are functions solely of the coefficient of expectation. (h) P: -,3Pt-l + (1 -,3) #2th + (1 «,3)2 fr,” + . . .‘ Because 04/ i: 1 the weights will decline toward zero as one goes back in time. After some years back, the weights are so small that they can be ignored. The number of years which must be included in the estima- tions depends on the size offl. Iffl" 0.5 the prices from the five past years must be included in order to come within 5% of the expected price. Nerlove has mde estimates for three crops: Cotton, wheat and corn for the period 1909-1932. The main results of his study are summarized in Table 2. In Table 2 are shown results obtained from two methods: The "special" and the "general" method. The one which Nerlave calls the special method is the same as has been used earlier with the assumptions that farmers, in the planning of their acreage, react only to last year's price. The coefficient of expectation (’6) is equal to one in this method. The general method is the one developed by Nerlove under the aselmption that the expected price (the price to which the farmers are seemed to react) is a moving average of all past prices with declining weights at increas- ing distance in previous time. The coefficient of expectation, which determines the weights, is, in this method, estimated from the data. There are considerable differences in the results obtained by the two methods. The elasticities of acreage to expected price are two to three times as large when the estimates are based on the general method 18 compared with estimates based on the special method. These higher estimates are more in accordance with practical observations, support programs, etc. than those obtained when K is arbitrarily assured to be one. Also the fact that a larger part of the variance of acreage (R2) is explained by the general method supports this method. Table 2. A Comparison of Two Methods (Special and General) for Estimating the Elasticity of Supply as Measured by the Response of Acreage to Expected Special thod ._. -_._ ;.._n_._...i ~—_’.___..__..__ .—_. p—m ._ - General thod Crop and Magnitude Corpared (Restricted : (fl: 1) (Unrestricted ,3 ) (1) (2) (3) Cotton: . Elasticity 0.20 0.67 Coefficient of expectation (/3) 1.0 (9.51) -.17 n? 0.59 0.7h Trend 0.’+8 0.18 (1.10) (i.12) Wheat: Elasticity 0J+7 0.93 Coefficient of expectation (/9) 1.0 (0.52) :.1h 32 0.6h 0.77 Trend 1.03 0.53 ( i.17) ( f.17) Corn: Elasticity 0.09 0.18 Coefficient of expectation {/e) 1.0 (2.3:) R2 0.22 0.35 Trend 0.21 0.16 (1.10) (2.11) (The figures in parentheses below the estimates are the standard errors of the estimates.) Nerlove's idea to estimate the farmers' expected price will be of particular interest to study in connection with analyses on Swedish data because the prices during the period under study were announced 19 before the farmers planted their acreage. The announced price in the Swedish data can be considered as the price the farmers expect to get and , consequently, correspond to what Nerlave intends to obtain with his expected price.1 Cromarty2 has analyzed the supply of a nmer of agricultural products and includes in his models: (1) Legged prices of the product, (2) prices of alternative crops, (3) costs of production, (h) weather, and (5) technology. For some of the crops he studied the following price elasticities of supply were found for the period 1929-1953: Wheat 0.129 Food Grains 0 . 36h Cotton 0 .361 Flue Tobacco 0.516 Johnson3 suggests in a recent publication that supply in agriculture is affected, not by a single, but many factors, as (1) business activity, (2) technology, (3) resource movements between regions, firms and enter- prises, (h) changes in risk and (5) changes in asset holdings. He defines a fixed asset as one in which its marginal value productivity in its present use neither Justifies acquisition of more of it nor its disposi- tion. According to this fixed asset theory, the aggregate supply function is only partly reversible. 1'lerlove's assumptions about the farmers' expected price will be discussed in more detail in connection with the presentation of the results from the Swedish data in Chapter VI. 2William A. Cremarty, Eggnomic Structure in American gouture, Ph.D. thesis, Michigan State University, East Lanai-18; 9 . 3Glenn L. Johnson, "Supply Function - Some Facts and Notions," Agricultural Adlustment Problems in a Grow, North Central Farm Management Committee, Iowa State College Press, Amos, 1958. 20 The above review is by no means exhaustive. It presents in very brief outline a variety of opinions and methods which constitute the back- ground to the present situation within the area of supply response. Most of the empirical studies cited generally support the notion that farmers respond very little to price. New methods such as Nerlove's have given higher price elasticities which appear to be more in accordance with other experiences. In this situation it is hoped that the analysis of the Swedish data with announced prices will give additional infome- tion in this area. CHAPTER II CEARACTERISTICB OI CROP PRODUCTION IR SWEDEN l The Climate. As a result of Sweden's length from north to south and of its changing topograplv, the variations in climate within the country are quite considerable. In the south, where the country lies open to the sea, the warm winds from the Atlantic, which is warmed up by the Gulf Stream, can sweep across the country. In the southern and middle part of the country, which include the main agricultural areas, the annual mean temperature is ever #00 1'. Plant cultivation is also greatly influ- enced by the length of the sunsner. The southernmost part of the country has an average summer of 21:0 days with a mean daily temperature of more than 50° F. and the middle part about 200 days. These differences, are, however, equalized to some extent by the fact that the nights are far shorter and lighter in the north than in the south. The amount of precipitation also varies but no where is the average annual rainfall too great or too sparing for agriculture. The correlation between rainfall and annual growth is, however, quite evident for most crops. The Soils.2 Similarly to the climate, Swedish soil type varies a great deal. All types are to be found, from recalcitrant moraines to easily cultivated sand and clay, and from barren, but easily tilled sandy J‘Hugo Osvald, Swedish Agriculture, The Swedish Institute, Stockholm, pp. 10‘19e 2 Hugo Osvald, 92. 335., pp. 19-28. soils, to fertile, but heavy clays, very difficult to work properly. The main part of the important agricultural areas consists of rather fer- tile clay. Some of the soils have a surplus of lime and potassium but most soils must be lined or fertilized with nutrient substances of one kind or another. Acreage of Different Crops. It is a characteristic of Swedish agri- culture that a great diversity of crop production exists. Most farms grow both feed crepe and cash crepe. At least 80$ of the farmers had dairy as an enterprise in 1951. Feed crops were grown in 72% of the arable land and cash crops on the remaining 28%. The acreage of important cash crops, fallow and total arable land, are shown for the period 1918-1956 in Table 3. Table 3. Acreage of Important Cash Crops, Fallow and Total Arable Land, 1916-1956, Thousands of Hectares. Fallow Tear Wheat Rye Oil Sugar Potatoes and Total Crops Beets Unused Arable Arable Land Land 19h3 269 220 30 50 1&3 205 3720 19hh 273 201 hh 55 138 205 3720 19h5 291 168 38 55 155 205 3720 19% 303 157 27 55 lh3 205 3720 19h? 292 115 h3 us ‘ 1h2 208 3722 19h8 316 160 79 1+8 11:8 182 3721+ 1919 307 135 11*2 1‘9 135 177 3727 1950 339 127 159 5h 130 172 3728 1951 32h 95 192 5h 121 177 3728 1952 328 123 1h9 5h 127 17h 36h? 1953 387 132 87 51 127 197 351‘? 195h h32 lh9 100 59 120 185 36h? 1955 353 95 101 53 123 225 36’?! 1956 397 123 30 50 122 20h 3598 Source: Arsvaxten, Central Bureau of Statistics, Stockholm. 23 All acreage figures shown in this table are those derived from the annual acreage investigations which are done during smer when all spring crops are planted. These are the only data on the distribution of acreage which are estimated from spring crops. They are, thus, considered to represent both the planted and harvested acreage for spring crops. There is normally no abundance of spring planted acreage. For winter crops, however, the data obtained in the summer investigations can represent only harvested acreage because the acreage of these crops is sometimes .re- duced to a considerable extent during the winter. Planted acreage of winter crops is obtained by a special investigation during the late fall. (See Chapter V). 0f the acreage planted in the fall , that portion which winter kills is normally replanted in the spring with spring crops. It is, therefore, not possible to use planted acreage of winter crops in this Chapter where the distribution of the total acreage by crops is discussed because the acreage which winter kills would then be double counted. One characteristic feature of the crop production in Sweden is that practically all arable land except fallow is planted every year. The figures for fallow and unused land vary comparatively little from year to year. Reclamation of new land is insignificant, and not even enough to compensate for transfer of arable land to building grounds and other uses. The total acreage used for crop production is, thus, fairly constant. Increasing use of acreage for one crop must, therefore, be offset by de- creasing the acreage of other crops. The acreage of wheat has, on the average, increased by about 10,000 hectares per year during the period from 1913 to 1956 (the coefficient for trend is 10,191) . This increase in wheat acreage has largely been 2h in spring wheat. This is mainly due to the fact that new varieties have made it possible to obtain good yields further north than has previously been the case. The acreage of rye also shows a definite trend, but in the opposite direction, with a decrease on the average of 6,5h0 hectares per year. The change from rye to wheat is mainly the result of a smaller demand and lower price for rye especially during recent years. The increased use of fertilizer has made it profitable to grow the higher yielding wheat on the poorer soils where befbre only rye could be grown. Also, wheat is less susceptible to lodging under heavy fertilization than rye due to its stiffer straw and, hence, more easily harvested by combine methods. The increase in the rye acreage during the years following 1951 was caused by a severe attack of wheat stem.rust in.195l. The oil crops (winter rape, winter "ryhs", spring rape, white mustard and linseed) are comparatively new in.Sweden. They were grown only to a limited extent in the southeast part of the country during world Whr II but expanded rapidly toward the north during the years following the war, reaching their highest acreage in 1951. They compete for the acreage par- ticularly with bread grains and sugar beets. The low acreage of sugar beets in l9h7-h9 may have been caused by the rapidly expanding oil crop production within the sugar‘beet growing area during that time. Also, the expansion of the wheat acreage after 1952 may, to some extent, be explained by the contraction of the oil crop acreage. The large variations in acreage of oil crops during recent years has been due to bad winter conditions (especially during the winter of 1955-56), plant diseases and decreasing prices. 25 The acreage of sugar beets is comparatively stable. They are grown in a rather limited area in the southeast part of the country. One of the main objectives for the price policy of sugar beets has been to offer a price such that enough sugar beets will be grown for dosestic consumption needs. Therefore, they have tried to set the price of sugar beets so that the net income (profit) in growing sugar beets will change in the same way as the net income in growing competing crops. The price, there- fore, to some extent, reflects changes in the costs of growing sugar beets. The highest acreage of sugar beats in Sweden was grown in 1951+. The high yield the year before, which was the highest average yield ever obtained in Sweden (see Table 6), may have had some influence on this high acreage. Potatoes are produced for food, manufacturing of starch and alcohol, and for feed purposes. The part of the potato production used for feed is decreasing due to relatively increasing production costs in growing potatoes. Potato production requires comparatively much labor, (1933;; intensive) the price of which has increased to a greater extent than have prices of other production factors. Only human consumption has been able to pay the increasing production costs. Feed crops were , in 19113, grown on 71% of the arable land in Sweden but had, in 1956, decreased to 72%. In particular, the acreage of such feed crops as fodder beats and hay, which have gained comparatively little from mechanization, have decreased while other highly mechanized feed crops such as feed grains have increased in acreage (by 17% between 19113 and 1956). The relative decrease in total acreage for feed crops is also due to the fact that dairy production is a labor intensive industry. 26 There is a tendency in Swedish agriculture to divert from labor intensive production to more capital intensive and more labor extensive production because Sweden has the highest wages for farm labor in Europe (Table 15). Labor intensive production is, therefore, in a relative disadvantaged position as conmared with other countries in Europe. Italy $ .19 3h The Netherlands .28 50 France .31 55 Switzerland . 32 57 Great Britain .38 68 Norway .38 68 Denmark .110 71 Belgium .141: 79 Sweden .56 100 Source: The data obtained from Jordbrukets Utredningsinstitut, Stockholm 1. Planted versus Harvested Acreage. For spring planted crops harvested and planted acreage is considered to be approximately the same. Winter crops, are however, sometimes destroyed to a considerable extent during the winter. This is especially true for oil crops such as winter rape and winter "ryhs", but also the acreage of winter wheat and winter rye is normally somewhat reduced due to winter killing. The percent of planted acreage on which the crop was destroyed during the following winter is shown in Table 5 for important winter crops in Sweden. There is a very high variation for oil crops. of the winter oil crops were destroyed. percent destroyed than rape. tive to hard winters. in areas with colder climate, which explains the higher percent destroyed. The acreage on which crops were destroyed is normally replanted in spring preferably with spring cash crops, e.g. spring wheat. the percent destroyed of the winter crops, therefore, may explain some of the variation in the acreage of spring wheat. 27 In the winter of 1955-56 more than 80% On the average, rybs has a higher This does not mean that ryhs is more sensi- The opposite is true and, therefore, rybs is grown The variation in Table 5. Percentage Acreage on Which the Crop was Destroyed During the Winter. Year winter Winter Winter Winter Wheat Bye Rape Ryhs - - - - - - - - - é‘EETEEEf'HSETEEiSE::' -------- 19h2-h3 3 5 o - 19h3 4h 2 h 9 - 19hh-h5 0 5 11 19 l9h5-h6 3 3 17 29 l9h6-h7 22 1h 73 h8 l9h7-h8 1h 6 10 ll 19h3-h9 3 3 h 9 19h9-50 3 5 1h 7 1950-51 10 12 6 13 1951-52 0 3 9 h 1952-53 8 8 37 51 1953-5” 7 h 2 13 195h-55 3 13 5 13 1955-56 6 L 8». 80 Average 6 6 2O 25 Source: Computed of data from Arsvaxten, Statisti Average Yield. The average yields in Sweden for'important cash crops are shown in.Table 6. The very low yields in.l9h7 of most crops was due to the severe drought that year. ska Centralbyran, Stock- In 1951, both winter and spring wheat 28 suffered from an unusually serious attack of stem rust. The variations in the other years are more nearly normal and influenced mainly by normal weather variations. There is, however, a trend toward increased yield for most of the crops, due mainly to better varieties and increased use of fertilizer. The annual consumption of artifical fertilizer in Sweden during the five-year period 1911-145 and the year 1950-51 is shown in Table 7 . Table 5o 2 -2” 13mm CahsCros 1' - ', SW. “0er Rectum Year Winter V Spring Winter Sugar winter Wheat Wheat Rye from Rape Sugar Beets (18% water) 19h3 23.3 1h.5 18.7 63.h 23.8 l9hh 21.8 1h.8 17.8 55.9 .lh.9 l9h5 22.3 15A» 16.6 58.1 1h.3 19h6 2h.7 17.h 18.6 56.9 17.3 19h? 12.5 lh.7 12.5 55.2 10.h 19h8 2h.8 19.2 20.1; 66.1: 21.0 l9h9 25.h 19.5 20.8 63.5 20.3 1950 21+.h 18.5 19.5 61A 18.0 1951 16.9 12.6 18.1 57.6 18.3 1952 28.5 18.9 22.9 88.3 21.1 1953 27.8 23.3 22.8 70.1» 16.6 195h 26.h 20.7 20.h 52.9 21.7 1955 23.8 16.3 18.3 50.1 19.7 1956 2h.o 23.9 22.1 58.0 12.2 Mean 23.1; 17.8 19.3 58.2 18.1 Source: Arsvaxten, Central Bureau of Statistics, Stockholm. The very low use of phosphorous during the earlier period was partly due to the lack of this fertilizer during World War II. An attempt has been made to estimate, on the basis of fertilizer experiments in various crops, the effect of artificial fertilizers on the yield increase. The experiments indicate that in the period 1915-52 29 the fertilizers probably accounted for at least 20% of the total crop.l Table 7. Annual Consumption of Artificial Fertilizer in Pounds per Acre. Nitrogen Phosphorous Potassium Year N P K 19111-19115 8 .118 3 . 32 7 .19 1950-1951 13 .70 10. 30 10.1w Source: Hugo Osvald, 92. _c_i_t., p. 57. The success of plant breeding in the increase of yield, the baking quality, the hardiness, the straw stiffness and the resistance against diseases has been especially great for wheat. This explains to some extent the increase in the wheat acreage. Good results from breeding have also been obtained in the improvement of sugar hosts and rape. Mechanization. Mechanization has probably affected the crop pro- duction in Sweden more than alvthing else during the period under study. Machinery of all kinds have been procured by the farmers but most important has been the tractor. The number of tractors used in agriculture is, therefore, a good measurement of the extent of mechanization. Table 8 shows the number of tractors on farms and the prices of tractors and farm machinery relative to wages of farm labor. The rapid increase of the tractor stock is easily explained by the relative prices of tractors to labor; tractors being substituted for labor as a result of the increase in relative price of labor. Also improvement in the quality of tractors has increased the flow of services llingo Osvald, gp. $3., p. 58. 30 from a tractor. The marginal value product of the flow of services from a tractor have thus increased and made the use of tractors more profitable. Table 8. The Nrmher of Tractors Used on Swedish Farms and the Prices of Tractors and Farm Machinery Relative to the Wage Rate for Hired Farm Labor, _ ... 2 - “a .7 _ ._ _ A M- m, Tctros Farms 0 ram and Farm figurine Thousands Relative to Wage Rate 19113 20 100 19115 20 85 19h? 27 1+8 1919 113 M 1951 6h #5 1953 85 38 1955 105 32 1956 115 30 aindex for*prices of tractors and farm.machinery deflated with index for wages of hired farm labor. Source: Sven Holmstrom, Jordbrukets Utredningsinstitut, Stockholm 1. The steady increase in the number of tractors on farms has contributed to the substantial reduction of man-hours used in agriculture. Table 9 shows the development of the total number of man-hours used in agriculture during the period from 19113 to 1956. Table - Million Han-Hours in the Swedish . - a ' 1" a ”8 W ‘ " ' an” 19h3 1170 100.0 1950 1000 85.5 1956 790 67.5 3These estimates of the total labor volme employed in agriculture have been made in calculating the total costs for price setting purposes. Government Bill No. 180, 1956, Stockholm. These figures include all labor in agriculture, both the farmers' own, and hired labor. The annual decrease of the labor volume is 31 estimated to be one percent of the previous year's volume during World War II up to 19h? to be three percent at the end of the 19h0'e and to be four percent after 1950. The whole decrease during these 1h years is one-third of the labor volume in 19h3. The figures shown here refer to all of agriculture but a substantial part of the decrease of the labor volume is a result of the mechanization of the crop production. Another result of the mechanization of crop production is the de- crease of the number of horses. In 19h3, the total number of horses in agricultule and forestry was about 600,000 but had up to 1956 decreased to less than half of that figure. The horses now left serve minly for- estry. The large reduction of the number of horses has decreased the acreage vhich is needed for feed production for the horses, Other crops, probably cash crops, are nov grown on this acreage. Crop Rotation. Crop production is usually organized in a typical crop rotation such as the English Norfolk rotation or modifications thereof. Three typical rotations are: In Southern Sveden In Middle Sweden Winter grain Winter rape Black fallow. Beets Winter grain Winter rape or ”rybs" Spring grain Beets Winter grain Clover Spring grain Clover and grass Clover Clover Clover and grass Clover Spring grain Spring grain Winter sheet is comonly chosen for the heavy and ordinarily good soils, rye for light and poorer soils. The black fellow in middle Sweden has decreased as the acreage of oil crops has increased. Oil crops now occupy about 5-101: of the arable land in the central provinces. Although there are new variations of crop rotations, the farmers generally follow a rather restrictive pattern in their acreage allocation 32 in order to maximize the yields. This circumstance can be expected to limit the farmers' response to price and cost changes. Some Geomhical Characteristics. Practically all cormnercial crops are grown in the southern and middle part of Sweden. The best conditions for crop production are to be found in the southeast province Skane. In this province which has 15% of the arable land in the country about 80% of the total acreage of sugar beats are grown (82% in 1951). Also some other crops which require good growing conditions are concentrated to the southern part of the country. This is especially true for commercial crops which occupy a greater proportion of the total crop acreage in the south than in the north. Due to the more favorable climate the farmers in the south have greater freedom in choosing between different crops. There is also reason to believe that the farmers in the south are more responsive to changing conditions and new opportunities in crop production. This may be illustrated by the rate at which a new crop - the oil crop - was adopted in different parts of the country during and after World War II. The curves in Diagram 1 are five-year moving averages of the acreage of oil crops grown in the southeast province, Shane; the southern part of the country except Skane, Gotaland; and the middle part of Sweden, Svealand. The data are shown in Table 10. The oil crop production started in Shane and up to 19h? its largest acreage was in this province. After 191W, the oil crop acreage expanded very fast in the rest of southern Sweden (Gotaland) where the acreage reached its peak somewhat later than in Skane. In the middle part of the country the oil crops were not grown to am greater extent until in the 1950's and stil1 appear to be expanding. 33 Acreage Planted in the Fall of Oil Crops (Winter Rape and winter Rybs), 19112-56, in Three Different Districts of Sweden, Thousands Hectares. Table 10. Svealand S-Eear Moving Average Actual Data Gotaland (except Skane) Actual Data Skane Actual Year 5- ear Moving Average 5- ear Moving Average Data 112nw2nUe7e-Oe/8e/Oe53 ccc137u1mssn oeloelencnlenwenlefiVHhVH/Oenmnde 000 00 Leo 3 8h. 7 O 0 ...-End 2283 06 71:"... 5321.1.“ 28 235hx3h®®M2 355156 9Rwaesflw sale—“ha o11staunemsnhnnn 7503213567.“.2 eeeeeeeeeeee 67n5afifl58mmw h519729831392170 k.%ki&&&.im.%i$& n eamwel 1112 23 5 7. 3 5 ssmswswwmmnsWsflm 1111111111..”11111 3h Diagram 1. Acreage Winter 011 Crops (5-Iear Moving Average) in Shane, Gotaland except Skane, and. Svealand. Hectares Gotaland except Skane 50 0/3‘. / \ 1955 1957 cam III AGRICULTURAL PRICE POLICY IN SWEDEN, 19110-1955 The unique characteristic of the Swedish price data used in this work is that the prices were announced before the crops were planted. In order to evaluate the significance of this circumstance it is necessary to know how the system of forward prices worked. The background and some of the important effects of this price system which was in oper- ation in Sweden during the period 19140-1956 will, therefore, be described and discussed in this chapter. This discussion will be limited to the effect of the price policy on the allocation of production and no attempts will be made to evaluate the adequacy of the Swedish agricultural price policy in general. The Princilles for the Price Setting m} The main goal of the Swedish ayicultural policy has been that of income equality between the farmers and other ”comparable groups". One way to reach this goal is to set the prices of the agricultural products in such a way that income equality will be attained. During the period under study it was assumed that income equality existed during the crop year 1938-39. The efforts, therefore, aimed at maintaining the insane-cost relation 1 Sven Holmstrom, cultural Policy in Sweden, A of Govern- Insult Measures, 19111-195 , ord‘brukets Dthngshstitute: tockholm. Ingvar Lindstrom, Jordbruhets inkomstlikstalligget, (Income Parity of the Swedish Agricultural Population , ordbrukets tredningsinstitut No. 7, 1956, Stockholm. 36 unchanged from 1938-39, which became the base year, 1.0., the prices of agricultural products should be set so that the aggregate income of agri- culture changed 2922‘. 2:35:11 to costs changes. If the aggregate cost in agriculture had increased by say 10% since the base year, the prices should be adjusted so that the aggregate income also inreased 10%. The great difficulties involved in a complete estimate of aggregate income and cost in agriculture was also a reason for using this relative comparison instead of an absolute comparison. In the method of absolute comparison it is necessary to know all incomes and all costs in order to be able to estimate what product prices should be required to balance the budget for the whole agriculture. But it is very difficult or even impossible to estimte all costs and incomes in an industry such as agri- culture with so many small enterprises. In using the method of relative comparison this was not necessary. It has not even been known how large a percentage of the total cost or income in agriculture which was included in the calculations. What was estimated was that the known part, say 90%, of the cost had increased by a certain rate since the base year. Con- sequently, the prices should be adjusted so that the part of the total income which was included in the calculations increased by the same rate. How the prices developed in the part of the cost not included in the cal- culations was unknown Just as were the prices of excluded part of the incane. The estimate of aggregate cost and income in agriculture were, in most cases, based on surveys. For estimates of aggregate cost, firms selling machinery, fertilizer and other inputs to the farmers had to make statements about the value of everything delivered for agricultural 37 purposes, not including delivery for personal purposes or for other industries such as forestry and horticulture. Included in the aggregate cost were also wages not only to employed laborers, but also to cover the work performed by farmers and their families. Farmers and their families were rated at the same wage as the groups of enployed labor to which they most closely correspond. Interest cost was also included, both on loans and on the farmer's own capital that they had invested in their farms. In this way all expenditures which could be estimated were in- cluded in the calculations. Estimates of income were made from statistics of harvests, produc- tion, sales, etc. Certain items could be determined almost exactly from final statistics. This applies, for instance, to cereals, sugar beets, potatoes for distilleries, oil seed crops, and so on. Special investiga- tions had to be made on certain details, as for example, the consmption of farm products by farming households. It is, of course, inevitable that certain estimates were only approximations. This applies particularly to income from eggs, market garden products, etc. In spite of the un- certainty as regards the absolute level, the change compared with the basic year should generally be correct, if the estimates were comparable from year to year. Due to the fact that new prices were announced before the farmers planted their crops, nothing was known about the size of the yield at. the time when the calculations were made. All estimates had, therefore, to be based on normal harvest. The above described system of forward prices increased the price certainty and stability for the farmers. This does not necessarily mean 38 that the income stability was increased. The Swedish agricultural prices are normally far more dependent on the world market prices and, conse- quently, 0n the size of the yield in the major production areas of the world than on the supply in the domestic market. There exists, however, at least some tendency toward lower prices at higher yields even in Sweden when the prices are not set by the government. Especially the fact that a larger part of the Swedish production has to be exported at a high yield contributes to lower prices. This has a stabilizing effect on the income which is removed when the prices no longer are dependent on the yield in the same year. Instead, the income of the farmers has become more dependent on the size of the yield. In order to reduce the income variations for the farmers, a new estimate was made Just before the harvest (in August) when data from the preliminary harvest estimate were available. If this fall estimate showed that the aggregate income of agriculture deviated more than 11% from an income required to keep an unchanged relation between aggregate income and cost cornered with the base year, the prices should be ad- Justed. This rule, called the four percent rule, thus limited the varia- tions in the total income to that of the ”parity income" (the cost index times the base income). AdJustments according to this rule have been made only twice (in 19h? and 1955), both times owing to the fact that aggregate income was more than 11% lower than the aggregate cost. The circumstance that aggregate income was increased at the same rate as the aggregate cost does not mean that all prices were increased equally. The aggregate income for agriculture was given and could not be changed but the allocation of the income between various products 39 was determined by the government after negotiations between farmer and consumer organizations. If it was considered important that production of one item should be increased the price of this item could be raised more than that of others. This influence over the allocation of the income between different products has, however, also the power to allocate the income between different groups of farmers. This power was also used u: some extent to favor the support of smaller farmers. These attempts to support special farm groups via the price system had significant effects on the direction of production. Especially during the years following World War II considerable effort was made to create ”Justice" in the income allocation between farmers of different size and different parts of the country. The small farms, which according to the existing opinion should be supported, obtained most of their incanes from milk production. Therefore, most of the increase in the income for agri- culture was given as increased price of milk. 0n the other hand, the bread grains were produced mainly by large firms. The price of wheat etc. was, therefore, kept comparatively low. This price policy rather soon resulted in an increased production of milk and a decreased production of bread grains (as is shown in Tables 11 and 12). From 1938-39 to 1951-52 the planted acreage of bread grain had de- creased about 79,000 hectors or 16%. This was mainly a result of the price Policy which was followed up to 1951-52. During these years the export surplus was converted into an import need , as is shown in Table 1. At the same time as "social" prices were set in Sweden in order to support the small farmers the world market prices moved in an opposite to Table 11. The Development of the Production of Bread Grains and the Prices 1&8 urplus 4 Price of -. Planted 8c Deficit - Winter Wheat wheat from USA Price Index Crop Acreage of for Domestic in Sweden Within Outside in Sweden Tear Bread Grains NOOdB in m a m Sweden 1000 hectors 1000 tons Sw. cr/dt Sw. cr/dt Sw.cr/dt 19h9-100 19h8-h9 A75 +130 30 37 37 100 l9h9-50 hh2 .115 31 he as 101 1950-51 h65 - 22 31 h5 57 117 1951-52 #18 -283 h2 ha 60 126 1952-53 h50 - 20 55 an 5h 128 aInternational Wheat Agreement, started 19119-50. Source: Jordbrukseconomiska Meddelanden, Stockholm, 1953. direction to the Swedish prices. In 19h8-h9 the Swedish price of winter wheat was 7 cr/dt. (one dt. . 100 kilo) lower than the world market price. This difference had more than doubled two years later. Instead of exporting and being able to'gain from the high world market prices, which should have been the case if the Swedish prices had followed the world market prices, Sweden must now import bread grains at higher prices. In 1951, the idea of social price setting had to give way to economic reality and the Swedish wheat price was increased up to the world market level, which led to a strong increase of the wheat acreage. The price policy in the milk production shows a different picture. Here production expanded steadily from the very low production in 19111-112 up to the years around 1950 when it reached its peak. This increase was caused by steadily increased dqmstic prices at the same tine as the world market prices were decreased or unchanged. The export surplus increased as is shown in Table 12. From 1951 the milk production has decreased mainly due to increased prices of wheat and oil seed. 1+1 Table 12. The Development of Milk Production and the Price of Butter Surplus Domestic Export Crop Milk which must Price on Prices Consumer Tear Production be Exported Butter Including Price Expressed Including Export Index in Milk Governmnt Costs in Sweden Subsidy 1000 tons 1000 tons Shcrlkilo Sw.cr[kilo 191198100 19118419 11516 +125 5.111 6.12 100 1919.50 11875 +376 5.56 5.63 101 1950-51 W72 +523 5.91 5A0 117 1951-52 11538 +3118 6.82 5.1111 126 1952-53 M92 411011 6.99 6.20 128 Source: Jordbrukseconomiska Meddelanden, Stockholm, February 1953. For the society this "social" price setting became expensive be- cause it increased the production of butter, which could be exported only at a loss, and prevented the Swedish agriculture from taking advantage of the good world market prices of wheat. The main reason for this failure in the agricultural price policy was not necessarily the system of set prices but that this system had been used in order to manipulate the prices to support special farm groups. The pricing system has since been modified in an attempt to maintain the goal of income equality for the farm population and at the same time minimize the loss of efficiency. Forward Prices. When the system of announced prices started at the beginning of World War II, Sweden had to rely practically completely on her own production to supply her population with food. In this situ- ation it was often of great importance to induce the farmers to increase the production of some comedities very fast. After a few years of ex- periences with price regulations, it was thought that the farmers at 2&2 a given price change should respond faster and to a greater extent if the new prices could be announced before the farmers made their decisions for the allocations of acreage to different crops in spring. In 1913, the method of forward announced prices started for spring planted crops and continued to 1956. Normally, the prices were announced in February. The farmers favored this method of announced prices and in 1919 it was extended also to crops planted in fall. Since 1956, the prices have not been announced except for bread grains. CHAPTERIV DMFACTORSWCTINGTHESIEGFTEACREAGEOFACBOP The factors affecting acreage allocation can be divided into economic and non-economic factors. Non-economic factors are such as weather, soil, tradition, skill and interest of the farmer, etc. Even if the influence of these factors on acreage allocation may be considerable they are of less interest 'for the purpose of this study, Changes in weather are randomly distributed, soil does not change and there are no data to measure the relationship between acreage and tradition, skill, etc. Economic factors are of more significance. Crop production is an economic enterprise and it may, therefore, be reasonable to assume that the farmers try to allocate their acreage so that they maximize profit. Profit is maximized when the mrginal value product of any input used in producing any product equals the corresponding marginal factor cost of that input. (1) MVP REEL-'o....- x1y1=.....mxny1_ WE MFC x1371 1“1171 3931 MVny MVP y ...°' 1 'eeeee= nm -1 EA SEC—— ’1’; Wm where 11 . . . :1 . . . rhare different inputs, and 71 e e e VJ e e e ym an different produCtS. The last unit of each input devoted to each crop should give an additional return equal to its cost. This is true for all inputs and, thus, also for land. The total amount of land is, however, sometimes fixed for the in- dividual farmer. Therefore, he cannot always reach the high profit point where MVP = MFG but combines inputs in the least cost combination for production. He, however, can always allocate the acreage he manages so that the additional returns from the different uses of the acreage are proportional to the additional land costs for the different uses: 2 MVP () Lyle....:mLyJ: sfl W— . O O O . Ly]. WCLyd MFCLym where I. is land, and y1 . . . ya . . . ym are different products. The equilibrimn conditions for maximizing profit which are given in equation (1) will be changed 1: any of the variables are changed. Changes in the prices of the products or of the amunt produced for one unit input will affect the MVP and changes in the prices of the imuts or in the use of inputs (technology) will affect MFG. Any such change will change the equilibrium system and, thus, also the allocation of acreage if profits are to be maximized. The acreage allocation will be affected by changes in (l) the prices of alternative crops, (2) the yield per unit of land from various crops, (3) the prices of different inputs, and in (h) the state of technology. 1. The Effect of Changes in the Prices of Alternative 010M We Allocation. Other things being equal, the rational manager would change the allocation of acreage between different crops according to changes in the prices of the various crops. In determining how to respond to price changes he must consider (A) how the price of a crop has changed relative to prices of other crops which may compete for the :45 acreage and (B) the expected duration (stability) of a new price level. A. Relativegice. It is obvious that the acreage response to a price increase of, say, wheat also depends upon changes in the prices of other crops which compete for the acreage. If the wheat price re- mains the same the acreage of wheat is likely to decrease when the prices of other crops increase and increase when the prices of other crops de- . crease. Thus, in studying acreage response to price, relative prices must be used, i.e., the price of a crop should be deflated with a price index of all other crops which compete for the acreage. This procedure will also transfer the prices for different years to the same money value, which also is necessary in order to get comparable prices. For further discussion of the purpose of deflation, see Chapter V. B. The Duration of a Price is important mainly because (a) more costs are fixed in the short run and because (b) it takes time before the information of the price change has reached the farmers and the signi- ficance of it has been recognized by them. (1). A reallocation of the production will affect not only the income but also the costs. It may be necessary to buy new machinery and some earlier investments may be mused. It may be necessary to develop new skills. A change in the production nearly always involves some cost and the more suddenly the change is done the higher will be the cost. Therefore, it may not be profitable to change production even at a comparatively large increase of the price if the new price is expected to last only for a short period, while it may be profitable to adjust to a smaller price increase if there are reasons to believe that this price will prevail for a long period. 1.5 A price increase which makes e.g. wheat production more profitable than eariier compared with the production of other crops will at first attract the acreage which is not bound to the present production with high fixed costs. Some of the acreage of closely related crops whose production requires the same machinery and other equipment will first be diverted to wheat production. At the same time as these short run adjustments are taking place, the farmers also make long run decisions which will effect the wheat acreage in the future if the price remains on this higher level. Pmduction of some crops requires special machinery and equipment which cannot be used for other crops. Most of the acreage of feed crops is bound to its present use by the high fixed costs of dairy production such as buildings, cattle, etc. which may have no alter- native use. However, in the long run, most fixed costs will disappear. If the increase of the wheat price caused a substantial part of the farmers to decide not to invest more in those fixed assets which prevent them from diverting more acreage to wheat production, gradually more acreage will be released for wleat production. The response of acreage to price may, therefore, be expected to increase over time. In other words, some of the response will occur with a time lag. The whole accumulated response is, therefore, expected to be larger than the response during the first year of a price increase. Also, the growing technique makes a rapid change costly and diffi- cult. Some crops must be preceded by certain other crops in the crop rotation and changes in the acreage of these crops must be planned more than one year ahead. An increased acreage of a crop requires also in- creased use of different production factors such as seed, fertilizer, etc. , which must be ordered some time before the planting of the crop. h? (2) It takes some time before the information of a price change reaches the farmers and the farmers have been able to Judge the importance of this change. Only the best farmers will evaluate the effect of a price change on their enterprises innediately. Some poor farmers may not be able to make such estimations at all. Many of these instead react to how successful farmers in their neighborhood change their production. This procedure will probably delay some of the aggregate response to a price change with one or several years. 2. The Effect of Changes in the Size of the Yield from Various Crops on Acreage Allocation. At a given price the total income from production of a crop will vary with the quantity and the quality of the yield. When, as the situation has been in Sweden during the period under study, prices are set one year ahead, the size of the yield will become of greater im- portance because a larger yield will not result in lower price which may be true in a free price system. Stable prices may, thus, tend to make income more unstable and more dependent won the yield. Both the quantity and quality of the yield vary randomly from year to year due to the influence of the weather. The rational farmer cannot take this random variation into consideration in his acreage decisions because it is not possible to predict if a coming year will give a high or low yield of a special crop. A response to last year's yield must, therefore, be considered as irrational if there is no trend in the yield. However, the size and quality of the yield is also affected by factors other than weather. New varieties of a crop, new machinery and increased use of fertilizer, insecticides, fungicides, etc. may increase the quantity as well as the quality over time. Thus, there may be a h8 trend toward higher yield and/or quality. If the trends for the different crops are unequal the relative profitability in growing different crops will be affected. It may be profitable to divert more acreage to a crop because the yield of this crop has increased more than the yield of other crepe. A response to last year's yield may in this case be rational. 3. The Effect on Acreage Allocation of Changes in the Prices of Various Inputs. A change in the price of an input will change the MFC of this input in equation (1). Profit will be maximized at a new equili- brium and, thus, also at a new allocation of acreage. In other words, changes on the cost side, also, will affect the allocation of acreage. If one production factor increases faster in price than other factors this will di sfavor that production which requires comparatively much of this factor. At unchanged incomes from different crops the rational manager should, therefore, try to divert acreage from those crops which employ relatively more of the production factors that have increased in relative price. During the period under study, labor wages in Sweden have increased more than the prices of other production factors as is shown in Table 13. Farm wage rates were four times as high in 1951!- as in 1939 while the interest rate had even decreased a little. Machinery and fertilizer had about doubld in price during this time. It is obvious that these developments in the relative prices of the various production factors must have affected the production costs of different products in an unr- equal way. A labor intensive crop requires a much faster increase in price in order to have unchanged profitability than a crop which is labor extensive. The acreage of a labor extensive crop can, therefore, ’9 be expected to increase at higher wages if other profit affecting factors such as price of the product, adaption of new technology, etc. have developed at the same rate as in the labor intensive crop. There should, in other words, be a positive correlation between acreage of a labor ex- tensive crop and the wages of labor and a corresponding negative correla- tion between acreage of a labor intensive crop and wages. Table l}. Price Indexes of Some Production Factors in Aggiculture. Price Indexes in 1951+ with 1939 as Base Year Wages of Farm Labor #18 Machinery (tractors, etc.) 212 Fertilizer 177 Interest 97 Weighted Average of all Production Factors 321+ Source: The data obtained from Sven Holmstrom, Jordbrukets Utrednings- institut, Stockholm. h. The Effect on Acreage Allocation of Changes in Technolgg. During the period under study (1913-1956) tremendous changes in tech- nology have occurred. During this time tractors and other machinery have replaced most horses and a considerable part of the labor in agri- culture. The extent of this substitution in the Swedish agriculture has been discussed in Chapter III. New and improved machinery for culti- vating the soil and for harvesting the crops such as the combine have also com with the advent of the tractor. Other changes of great significance for the crop production have occurred. The use of fertilizer has increased several times over since 19:3, and preparations against pest and plant diseases have become widely 50 used. All these changes in technolog have greatly changed the condi- tions of crop production in Sweden. The production of some crops has been more favored by the technological changes than others. For example , grain production has during this period become highly mechanized while the production of beets and potatoes have benefited less from the technological development. Labor has increased more in price than other inputs. These labor intensive crops require a relatively larger increase in price than for grains if the profitability should remain the same. The feed crops are in the same situation as labor intensive crops, be- cause the extent to which they are grown is also influenced by the tech- nological development in the labor intensive dairy production, which has not been able to replace labor in the same proportion as grain pro- duction. All the changes discussed above on both the income and the cost side should be taken into account if the acreage response for every year is to be related to the profit in growing the crop studied the corresponding years. Unfortunately, data are not available to estimate the development of the profit for different crops. Therefore, the study must be limited to including into the regression function as many of the above discussed factors as data are available for. The most important is price but as has been indicated in the above discussion, it cannot be expected that the rational farmer should respond only to the price of the product. Some of these Other factors of economic importance such as technolog and wages of labor have, however, changed rather regularly over time and will, therefore, be taken care of fairly well by the trend variable. va mmmmwrmmomrm The data upon which the estimates are based in this tbsis are official data collected and comiled by various goverment agencies and far- organisations in Sweden. There are three types of data mainly, which have been of special importance for the calculations. Those are acreage, price, and price indexes. Item Data. Data on acreage in Sweden are collected by the Central Bureau of Statistics which is a goverment agency. During Horld War II and up to 1950 these acreage collections included all fans with an acreage of two hectares and more. After 1950 , the agency returned to the use of the eagle lethods used before World War II. This techique consists of stratified sam- ling in commotion with a systematic selection in strata, where the farms are arranged according to the size of the arable land and natural agricultural districts. Within each district the farms were divided into size groups in such a way that the same acreage should be included in every group. The eon-pro selection is randou and includes about 10¢ of the total me:- ef farms above two hectares. This procedure makes it possible to measure the exactness of the estimate. The war of fares from which data were collected in 1951 anointed to 28,015. The final estimate that year was based on 99.21 of these forms . 52 The above described estinate of acreage is ceniled for all crepe on the 15th of July every year. The data obtained represent both planted and harvested acreage for all spring planted crops. For winter crops planted in fall the result fren the survey a July 15 will be useable only as harvested acreage because the acreage of these crepe my be considerably reduced during the winter. A special survey for fall planted acreage is cheted late in the fall. The sane techique is used in the fall as in the our survey. Price Data. During tb period studied, 19153-56, the prices of nest agricultural products were set by the goveruent. These set prices were anounced in February or early in the spring before the farnrs planted their crops. Thus, the farners knew the prices they would re- ceive before they allocated their acreage ueng the various crops. The annouced prices nomlly have deviated only little fren the final prices received. Announced and final prices of spring wheat are snared in Table 1h. Changes in these two price series nove in the sa- direction although mood prices often show a tendency to be lower than the final prices. Because the calculations in this work deal .inly with data fru spring wheat the comarison has been linited to this crop. The variations between announced and final price for nest other crops, however, show about the sane pattern. The fact that the announced price for a few years differed rather significantly fru the price the farners actually received can, accord- ing to the writer's opinion, be expected to have had little effect on the farners' response to the announced prices. At the tins of planting there were no reasons to believe that the prices should be other than 53 a-ounced. Such things which would alter the announced prices, libs ertrenely poor or good yield (the four percent rule), disturbing events in the international trade etc. , were Just as unhne‘ for the farners as to those nahing the price calculations.. The annouced prices, I therefore, can at the tins of the planting be considered to be the best possible predictions of the run price and the best price to base the acreage decisions on. Even if the deviations hd been still larger there should not have been any reason for the rational far-er to expect other prices than those announced. Table 15. Mud and Final Price of Spring meat 1913-1956, 0re per Kilo. crop Your Announced Final 0”? “‘1' Amced Final 1 30t- ' Price Price 1 Sept. " Price Price 2 A28. __ __ _ 31 mo __ 1913-“ 28.50 29.0? 1950-51 32.50 33.15 l9hh-h5 28.50 28.55 1951-52 #3 .00 #3 $5 19h5-h6 28.50 27.88 1952-53 55.10 55.35 191.647 28.50 28.11 1953-5h Irma l9.91 1957-53 31.50 33.51 195%” I$5.10 $.37 19‘849 3h .50 3h 050 1955‘56 h3 .33 l‘3 020 1959-50 35.10 35.19 1956-57 I8.50 *3 .85 Sources: Final Pfices: even fiolnstron, ”Producentpriser och Prem- tionsvardsn", 1938-39 - 1956-57, Jordbruhets Utredningsinstitut, NCMe Announced Prices: Iannart Porsberg, "Prognos och Blutligt Beeultat i Jordbruhshalkylan' , Inntbruhshogsholan, Uppsala 7. Gover'nnent Bill '0. 165, 1956, Stockholn. The final prices are averap prices of the daily narhet prices in sone specified areas of certain grades. Price Indexes. All prices used for estinting regressions in this thesis have been deflated by the Producer Price Index for the total agricultural production (corresponding to the Index of Prices 5’4 Received by Parners in 0310.1 rm: index is published in two we. divisions: (l) The Price Index for Vegetable Produce which is de- signed to neasure the price changes for crepe sold in the narhet (cash crops) and (2) the Price Index for Aninal Produce which nsasures price changes for aninal products. The Preduoer Price Index is also counted as a Real Producer Price Index. In this case, the Preduoer Price Index has been adjusted for cm. in the noney value by division with the Consuer Price Index, which is considered to unsure changes in the noney value. For cal- culations in this thesis, tb unadjusted Producer Price Index hasrbeen used. The Producer Price Index is Med only fron the crop year 1950- 1951.2 Its been is the starting prices (nittpris) for the new price systen which started in Beptenber l, 1956. Theee prices are approxi- . nately the sane as the average prices in August and smarter, 1956. The 1953-55 production of the following co-odities as'weights have been used (the figure in brachets indicates the total production, .111. kg.): Winter wheat (96), spring wheat (328), rye (226), household potatoes (665), factory potatoes (228), sugar beets (1,8h8), oil seeds (173), fluid mt, 3$ fat content (890.9 nill. litres), fluid nilh, 3.5; fat content (55.5 nilLlitres), thin crean (25.3), thick crean (19.8), Bvecia-cheese (35.8), lanyards-cheese (15.10 butter (86.7), 1The reasons for the selection of this index are discussed in the following section. e'l'his new index is coquted in Btatens Jordbruhsnmd, Stockholm. 55 .111: powder, whole (L5), nilk powder, ski-ed (6A), fluid .111: sold directly fron farners to mrs (106 nill. litres), fluid nilk canon-ed at the farurs' households (uo7 1:111. litres), age (73.5), beef (96.7), veal (35.7), horse-neat (11.3), nutton (1.9), and pig-mat (191.9). For the years before 1950-51, the old Producer Price Index has been used} This index has the year 1939 as base and the weights con- sist of the production in 1935-37. Because the prices which are de- flatedarecoquted for cropyear(8ept. l -1)“. 31) it hasbeen necessary to transfer the calendar year index figures in the old Preduoer-Price Indexto cropyear. This hasbesndone sothat one- third of the no calendar year index figure plus two-thirds of the following calendar year index figure establish the crop year index. figure. ThenewandtheoldProducerPrioe Indexes havethanbeen linked together with the crop year 1950-51 as the base year. The different indexes are shown in Table 1.5. Selection of Index to Doflate the Prices With. In order to detenine the extent to which acreage allocations are affected by price changes, tenet hnowhowthe acreageandthepriceof a crep have changed during the period under study. The acreage is neasured in a unit which does not change over tine. Therefore, the data of planted acreage can be used unadjusted. The value of the noney unit, homer, isnot stable, but varies fronyeartoyear. Parnersarenot lihely‘ to increase their acreage of wheat when wheat price has increased lThis old index 'as counted in Jordbrnhets Utredningsinstitut, ”Owls Table l . The Producer Price Index. The Old Producer The lew Producer Year Price Index Price Index Tear Preduoer Price Index (1232:100) (Lemme-100) (igmaoo) 1939 100 1939-#0 55 19» 122 ”to-hr 66 1951 119 19h142 7t 1912 161 19h243 75 1916 158 198-“ 76 19M 160 1941;.” 76 1915 160 191.545 77 165 19‘t647 80 19k? 173 19117-118 86 19m 187 19118-159 89 1959 188 1919-50 91 1950 195 1950-51 100 1951 ‘ 221 1951-52 115 1950-51 Th5 1952-53 125 1952-53 93.2 1955-55 121 1953-53 89. 1955-56 133 1955-55 90 s 1956.5? 136 1955’ 99 03 57 101.2 Linked Together if all other prices have increased at the sane rate. But other things such as production costs, etc. being equal, the retinal farners should increase their acreage of wheat if the price of wheat has increased nore than other prices. Therefore, the price of wheat should be con- puted relative to other prices. This is done by deflating the price of wheat with a price index for other crops which are carpeting with wheat for the acreage. This was done for two purposes, to: 1. Transfer the prices to the sane nonay value as in the base period, 2. Tabs account for effects of changes in the prices of other products whiohnight haweheenproduced instead ofthe oneweare dealing with. 57 The deflated price of wheat should increase when the observed price of wheat has increased nora or decreased less, than the prices of cupeting one, and decrease when the observed wheat price has increased less or decreased nore, than the prices of cometing crops. If other things have ruined equal, the rational far-pr should be expected to adJust his acreage of wheat in accordance to variations of this deflated price of wheat. During these carditions we could ex- pect to find a positive correlation between deflated price and acreage. There is, however, great difficulties involved in constructing such an index which will give true relative (deflated) prices. By deflating with an index which does not give the true relative prices to which the rational farner should respond tln price elasticity con- puted will be affected. lost probable is that an inparfect index will lower the price elasticity because it is not so lihaly that a price index should vary inversely to the variation in acreage. This will happen, however, if the index includes the price of the product which is deflated and this product has high weights in the index. The varia- tims in the price of this product will then be decreased by deflation. The result will be the sue if tin price of another product with high weights inthe indaxbut not M138 for the sane acreage as the product under study varies in price sinilar to the product under study. The value of the price elasticity we will get fron a statistical analysis is, thus, to can extent, dependent m the index used for deflation. This linits the possibilities of conparing the elasticities between different countries with different noney hits and with differ- ently castructed indexes. Also, price elasticities within a comtry are ceweerehie only if the prices are deflated with the sane index. 58 An ideal index for this purpose should include prices of all crops which conpate for the acreage of the crop under study and the weights should be proportional to the carpetitive ability for the acreage of the different crops. A crop which occupies acre or less the sane acreage regardless of price should, therefore, have a low weight. The prices should be average prices which the farners can get by selling the crop or by utilising it in the aninal production. In practice, the indexes of prices received by farners are cur- structed for other purposes. The weights are nor-ally the Mities of different products produced during a chosen period. These weights are probably not proportional to . the conetitive ability for acreage ‘ of the respective crops. Acreage of the different crops would probably mks better weights. Tor nost of the crops which are nainly utilised in the aninl industry the narket price is based on such a little part of the produc- tion that this price is a poor nasurenent of the average value of the entire production. For sons feed crops such as silage, fodder beets, etc., a narbet price does not even exist. . The Swedish index for prices of crops, the Price Index for Vege- table Produce, includes only crops sold in the nrhet (cash crops). is sore than 70$ of the acreage is used for feed crops it nust be con- sidered inadequate to deflate with an index curtaining mly cash crops. However, the capetitive ability of the feed crops is, to sue extent, a function of the prices of aninal products. Therefore, there say be reasons for using an index including the prices of both cash crops ad ani-l products. But also such an index has so- ilportant 59 weaknesses. Increased prices for aninal products do not alwus neon increased return to the feed crop enterprises. The increase in the prices of aninl products say be offset by increased prices of labor, buildings, etc. which is used to a large extent in aninsl production. If labor cost increases more than other costs which has been the case during the actual period, the profitability of the labor intensive aninal production will be decreased to a greater extent than the labor extensive cash crop production. A relative increase in the prices of aninal products can, therefore, not always be expected to lead to an expanded feed crop acreage. This difficulty, however, can be overcome, to sons extent, by adding a variable in the regression equation for changes in labor cost. Also, a trend variable will take care of these changes if the wage rate has increased rather linearly over tine. The above considerations and the fact that neither tine nor data have been available here to custruct a new index constitute the back- ground for the selection of the Producer Price Index for the total agricultural productiar which include prices both of cash crops and aninal products. CHAPTER VI STATISTICAL ms The statistical analysis is divided into three parts. The problsn of the extent to which acreage allocation is influenced by different years' prices will be investigated in Section 1. In Sectim 2 the analysis will be extended to stub how variables other than price affect the allocation of acreage. In the third Section comparisons are nade between different price elasticities obtained from various parts of Sweden and between price elasticities obtained in USA and Sweden. The analyses are based on Swedish data for spring wheat relating to the period fron l9h3-1955. All prices are deflated with Producer Price Index for the total agricultural production. The relations be- twean dependent and independent variables are estinated using the nethod of least squares regression. The test of significance is designed for use with data having the following characteristics:1 (a) the population must be honogeneous, (b) the variables not be nor-ally distributed, (c) observations lust be independent, and (d) the sample lust be selected at randon. Since the data used in the study only approxintely fulfill these conditions, it follows that the results not be evaluated as to reasonableness and correspondence with experience. 1e. 3. Shepherd, Agricultural Price Lflzsis, Fourth edition, The Iowa State College Press, s: 19 7. . 61 To facilitate evaluation and comparison certain statistical tests of significance are applied. The student "t” test is applied to test the lupothesis that the regressiar coefficient tested does not differ significantly from zero. One of the asst-ptions nde in the estination process is that cat serial independence of the residuals. The Durbin-Hatson test1 is applied to deternine the degree of independence in the unexplained residual. Results of the test are reported as indeterninate, acceptance or rejection at the .05 level of significance of tb Whesis of serial independence. Elasticity of acreage response to price is defined as the relative change in acreage (I) oveAr'Ithe relative change in price (I) or g n '5— : r . 4r ‘2'!— T 21‘ T. _§_§_ n b s the price coefficient. Section 1. The Inflmnce of Different Iears' Prices on Acreage Allocation. As has been developed in Chapter IV, the acreage allocation can be expected to be a function of several factors, price being one of the nore ilportant factors. In this first part of tie statistical analysis the studies are linited to the influence of price on acreage alloca- tion. The obJectiva is to investigate to what extent prices in nore or less distant years influence the allocation of acreage in the current 1.7. Fried-an and B. J. Toots, manual llethods for Handling fistus of Silultaneous tions, Agricultural Handbook . 9 , m, a 3313M”: 0 or 95- 62 year. In this w it my be possible to clarify the controversial question of what price the farners respond in their acreage allocation. In nest studies of this kind it has been assuned that the farners responded to last year's price. lerloval has constructed a nodel in which he assunes that farnrs respond not only to the last year's price but to an expected price which he derives as a loving weighted average of previous years' prices with declining weights as one goes back in tin. With this nodal lerlove has obtained results which lore nearly correspond to the observed acreage response twider price support procras- In Chapter IV it was shown that there are reasons to believe that the acreage allocation in the current year is influenced not onlyby the price far-rs expect to receive for the current year's crop but also to sons extent by previous years' prices. The faraer is, due to high fired costs, crop rotations, etc. able to reallocate his acreage in a single year only to a linited extent. If, for exanplo, the price of a crop is increased sharply he cannot i-ediately (in a single year) increase the acreage of this crap as mch as the high price should activate. The following years, however, he can transfer nore acreage to the higher priced crop if the price renains high. The acreage which in the current year is head to its present use with high fixed costs will be released over tin. Therefore, it see-s reaemable to motheeise that previous years' prices also have sons influence on the asleep allocation in the current year. According to this mothesis J'larc lerlove, 22. 913. 63 farners respond to the projection of price over a period of future years. The lupothesis cannot be tested unless the price fernrs expect is hem. As earlier described, the Swedish price data consist of announced prices which can be considered as the prices the far-are expect. If it is true that there is a delayed acreage response to price, i.e. that the allocation of the acreage in the actual year is a fraction, to sue extent, of the prices in previous years, then it should be possible to show this through analysis of the Swedish data. It should be possible to separate the effect of announced price in a single year fro- the effect of previous years' prices. The effect of announced price during the current year and dining years prior to the current crop year was studied fron two different approaches: (A) The price variable was used as a moving average of the increasing lumber of previous years' prices, and (B) The prices during different years were included in the nodel as separate variables. A. The Price Variable as a Moving Average of an Increasing Imar of Previous Iaars' Prices. The first approach is based on the following nodal: (l) I Satthb t 1 1"" 2 where It a acreage of spring wheat year t P a different prices with increasing influence of previous years' prices. The following variations are cmsidered: a. announced price the current year (year t) b. noving average of announced prices in years t and t-l c. noving average of enhanced prices in'years t, t-l and t-2 6h d. noving average of annornrced prices in years t, t-l, t-2 and t-3 e. noving average of annotnrced prices in the years t, t-l, t-2, t-3 and t-h. T 3 Trend U 3 Panda residual. In including the prices fron an increasing mater of previous years the noving averages will be ears and nore influenced by the prices in the past. The results obtained in relating these noving average prices with the acreage in year t will give sons conprehsnsion of the influence on the acreage allocation fron previous years' prices. It can Just as wellba supposedthat the influence fro-previous years' prices should cons as well as free final prices as fron announced prices. Wintevsr of these two kinds of prices is influencing the acre- age allocation the result will be about the sans because the differ- ences between then are seen. a. Psannouncedprice 111m (2) r = -39.3 + 3.1a r + 7.80"! (2.33) (1.71») *‘l‘ Significant at .001 level. The price coefficient is significant at .20 level. Standard errors of the b's are shots; in parentheses. Standard error of estinte S I 25.6 Proportion of variance explained fie s .59 Price elasticity of acreage neasured at the nean E- .87 b.’ P = ave of mad prices in the_years t and t-l 65 (3) I = 408.9 + 5.30 P + 7.67“:- (2.88) (1.61.) as Significant at .001 level The price coefficient is significant at .10 level 8 s 2t.5 112 = .62 5 s 1.35 c- Wm 11- thum- h it}. and t-2 (h) ‘I - -302.9 + ro.h6*P + 7.67'lr (3.68) (1.h2) ** Significant at .001 level ‘5 Sipificut at .02 level 3 s 21.3 52 . .71 8 = 2.69 d. P = my avege of annornrced prices in years t, t--lI t-2 and t-fi (5) ‘r a -203.1 + 8.019? 4 7.81'lr (3.97) (1.62) "l‘ Significant at .001 level G Significant at .10 level 8 = 23.9 fie - .51: E s 1.99 e. Wof announced prices in Ears t, t-lL t-2 t-3 and t-h (6) r a 101.8 - .29P. 7.2h“r (8.01) (1.96) as Significant at .001 raver 8 = 28.0 fie: .51 A mry of the results obtained by adding the influence fu- an increasing nwor of years in the price variable is shown in Table 16. low the price elasticity varies is also shown in Diagran 2. When only the current year's announced price is included in the price variable the different para-store are snallor and loss signifi- cant than if previous years' prices are also included. All the para- .tors show: in Table 16 reach their highest value when announced prices fron three years are included in the price variable. The percentage of the variation explained increased fron 59 to 71, the price elasticity was three tines as high and the 3 statistic increased fron 1J6 to 2.8L The inclusion of prices in still noro distant years decreases the signi- ficance of the ostinatod paranetors. The results obtained in using the above nodal do not provide sufficient evidence to reJect the hypothesis that previous years' prices have an influence on the acreage allocation. On the contrary, there soans to exist a convincing influence fron previous years' prices on acreage allocation. Insteadafusing sinplonavingaveraga as hasbosndmointhe above nodal, the prices could have been weighted with maqual weights. Iftb prices, during different years, were weighted with weights 67 proportional to their influence on the acreage allocation still higher elasticities would prebably have been obtained, because it seems m- libely that the different distant years' prices should have equal in- fluence m the acreage allocation in the current year. It is, however, difficult to say at what rate the influence of previous years' prices decreases as the tine span considered increases. is the purpose of the above estinate was neinly to obtain a first approxinetion of the effect of previous years' prices less consideration was given to the weighting problem. Diagran 2. Relationship Between the Estimated Elasticity and the meter of Years Included in the Moving Average Price. Price elasticity an 2. b HHHHHMMM O O O O O O O I C . omrmmomrmmomrm A A A l a b c d e 1 2 3 h 5 1: Number of years included in mving average price. he 68 Table 16. Suary of the Statistical Results Obtained at Increasing Influence of Previous Tears' Prices. The Price Variable Equals: Mam! of Announced Prices D_u}_-_i_n_g: pr“. in Two years Three yearsTouryear-s Five years 3..., t t and t-l t, t-l and t, t-l, t, t-l, t-2 t-2, and t-2, t-3 t-3 and t-h .e b. C. de .e Price coefficient (b1) 3.111 5.30 10.116 8.01 -.29 Test statistic for sipificant differ- ence fron sero (tb) 1.56 1.811 2.811 2.02 .011 Variatim explain-d (fie) .59 .62 .71 .611 .51 Price a elasticity (6) .87 1.35 2.69 1.99 -.07 Trend coefficient 7.&) 7.67 7.67 7.87 7.23!» Test statistic for significant differ- ence fren sore (tb) 11.59 3.68 5.111 11.87 3.59 .Neasured at the neon. The weighting problen can be avoided by including the prices for different years into the nodal as separate variables. This approach is also a nore direct way to estimate the effect different years prices have «1 the acreage allocation in a given year, and will be followed in the second approach, 3. 3. Prices During Different Tears Included as Separate Variables. The nodal: (7) It = .1 +b1Pt the Pt-l +b3 Pt-2 eh, Pt-B‘ b5! f U 69 where It I planted acreage of spring wheat in year t Pt - Pt-3 = annourced prices in year t, t-l, t-2 and t-3 T e Trend variable 0 3 Bandas residual Iquation (7) fitted for spring wheat is as follows (standard errors of the b's in parentheses): + 6.76"T (8) r s 455.8 + 5.92“ P + 1.30 PM + 6.713 P + 0°32 Pt-3 (1.62) (2.57) " (2.89) (3.03) “2 (2.96) a erg-1:16:31: at the .005 level 1» srmncm at the .05 m1 Standard error of estinate = 22.8 32 = .67 The data were exanined for serial correlation by using the Min-- Watson test. At the. five percent probability level the null mothesis (that serial correlation = 0) was not redacted by the statistic ob- tained (1.71). The statistical results frcn using this nethod indicate that there is a "definite” relation, not only between acreage and announced price the sane year, but also between acreage and announced price in year t-2 (next price two years previous). It nay seen surprising that the previous year's annomoed price (t-l) appears to have a very linited effect a the acreage. The reason for this is probably nainly the in- fluence of the crop rotation. “host is not grown after wheat in Sudan. Therefore, if the annomced price and as a consequence, the acreage of wheat are increased in one year, there will be decreased possibilities to grow wheat during the next year because wheat has occupied a larger than nornel acreage the year before. 70 Also, the size of the winter wheat acreage is likely to affect the extent to which spring wheat is grown. The price of winter wheat has, to acne extent, been dependent on the price of spring wheat. If the announced price of spring wheat was increased in the price negoti- ations in Pebruar'y it would be very likely that the winter wheat price also would be increased. When the spring wheat price was increased the farners could, therefore, be expected to respond to this in two wave: (1) by increasing the acreage of spring wheat, and (2) by in- creasing the acreage of winter wheat the following fall. The next spring the farners would find thenselves in a situation where they had nuch land on which spring wheat was grown the previous year and on which winter wheat was planted last fall. Spring wheat could not be grow on this acreage and, thus, less acreage would be available for spring wheat that year. The low response to price in year t-l nay, therefore, be explained by the limitation in acreage for spring wheat that year. The next year, however, after one year with less wheat, the wheat acreage could be expanded again. This nay, in part, explain the correlation between the price in year t-2 and the acreage in year t. The above nodal has given a nore detailed knowledge about the in- fluence on the acreage allocation of different years' prices than the nodal used in the first approach. It was sonewhat surprising that nest of the price influence fron previous years as concentrated in year t-2. But this may be a characteristic for spring wheat. It should also be noted that the linitations in the data (the few years) any effect the results to acne extent. The influence on acreage alloca- tion of the price in year t-2 was, however, even stronger ( E = 1.7)) than front the announced price in year t (g I 1.50). 71 What conclusions should now be drawn fron the statistical findings in the two models? The study shows clearly that the acreage alloca- tion in a given year is influenced also by previous years' prices. Consequently, we have to consider that the value of the price elasticity will be different depending on the nunber of years included in the price variable. Especially it seems important to separate the price influence from a single year fron that of two or several years. The response to a single year's price could suitably be naned M 213 elasticity and the response to nore than one year's price leg; Leg elasticity. The short tern price elasticity is easily defined and easily as- tinated when the price the farners respond to is known. It is, however, probably inpossible to estinate the response to the current year‘s price (expected price) if the prices are not annotnrced. The response to last year's price, which often has been estinated earlier, is a response to a single year's price and, thus, a short tern response. The long tern elasticity is not so easy to define because it is possible to compute nany different elasticities by varying the nunber of years included and the weights with which the prices for different years are weighted. One of these different long tern elasticities scene, however, to be of greater interest than the others and that is the one which explains the largest part of the variation of acreage, i.e. when the residuals are nininised. In the nodal used in part A, 32 was highest when announced prices for three years were norms in the price variable. Most of the accunulated price influence over tine on the acreage allocation will then be included in the price elasticity. 72 The practical inportance of the above result is the different effect the sane price will have depending on the duration of the price. If the price lasts for a long tin, the effect on the acreage alloca- tion will be larger than 1: the duration of the price is short. A 10$ increase of the real price would, on the average, during the period studied, have increased the acreage of spring wheat with about 13,000 hectares during the first year of the price increase, but if the price renains on the sane level for at least three years the increase in acreage would be 39,000 hectares. The above result must also be tahen into consideration in trying to estinate an expected price when prices are not adninistered. In the nethod Harlove has introduced, the farners' expected price is defined in the following way: 2; Vet“ . (1-,8) #112 . (1-fl)2flpt_3 +. . . .. where P t-1, -2! t—3, etc. and fl is aesuned to be the proportion by which the farners Pt Pt-3 etc. are observed prices in years t-l, t-2, revise their price expectations: * a P: .- Pt']. 8 fl(Pt.-1 '- Pt_1) where * P: and P are expected prices in years t and t-l. t-l According to our findings the acreage allocation in the current year is a function of both the expected price and previous years' prices. But in lerlove's nethod, the expected price is a function of previous year's prices. Previous years' prices will, thus, affect acreage alloca- tion in two rays: (1) Indirectly as the price the farners expect (corresponding to the influence of announced price in year t), and 73 (2) as a delayed influence fron previous years' prices (corresponding to the influence of announced price in year t-l, t-2, etc.). The question is now, is it possible to divide the influence of previous years' prices into these two groups when there are no announced prices? lerlove is trying to estimate the price the farners expect to get in the coning year and, thus, he is intending to estinate the short tern response corresponding to the acreage response to announced price in year t. But 1. it possible to give fl such a value that the price which is estinted as a noving weighted average will be the "true” expected price? lerlove Justifies his nethod with the fact that it gives elasti- cities which better correspond to observed acreage response under support progrsns etc. This observed response is, nor-11y the accunu- lated response over tine. It is, therefore, very likely that the re- sponse obtained by lerlove's nethod corresponds closest to the long tern response. The expected price estinetes in his nethod is, thus, not the price the farners really are expecting to get in the actual year (t) but rather a weighted average of this ”true” expected price and prices fron eone previous years. Section 2. The Extent to Which Different Factors Affect the Allocation of Acreage. The objective of this second section is to investigate to what extent various factors have affected the acreage of spring wheat in Sweden during the period with announced prices. It has already been shown in Section 1 that price has a definite influence on the acreage allocation. In order to include, as nuch as 7h possible, of the price influences in the model, not only announced price in year t but also sons previous years' prices, above all, year t-2, should be included in the nodal. Announced prices for different years can be included separately or weighted together in one or another nanner. In this case, it has been desirable to have only one price variable because the number of observations and, thus, the number of degrees of freedon is so snall. The announced prices for the years t and t-z have, therefore, been included as noving averages with each of the years receiving equal weights. There is also a great variety of other factors which can be assured to effect the acreage allocation of spring wheat. Bone of these are changing over tine and, therefore, intercorrelated with the trend. In the first nodel it is asst-ed that these factors are taken into accent by the trend variable. shat prelininnry estinates have shown that besides trend and price, the nest inportant factors which influ- ence the size of the spring wheat acreage are last year‘s yield and the sire of the acreage on which winter crops hve been destroyed during the preceding winter. The nodal now used, therefore, includes these four variables: (11) It - a + hlrtfié) + hand + b3R t brT where Itsacreageofspringwheatinyeart Pt,(t-2) = eimle noving average of announced prices in year t and t-2 ’t-l s average yield of spring wheat (deciton per hector) in year t-l 75 R I reduced acreage of winter crops during the preceding winter (1000 hectares) T a trend The fitted equation is as follows (standard errors of the b's are in parentheses): (12) r - 428.9 f 12.86'*Pt’(t_2) + 2.72%“1 4» .25'3 . 5.36": (8.27) (2.36) (.27) (1.73) H Significant at the .015 level 1» Bipificant at the .30 - .ho level Standard error of estinate - 19.7 r2 . .76 E - 2.56 The Durbin-Vatscn statistic of 1.61 provides insufficient evidence to reject the Methods of serial independence of residuals at the .05 level. A conparieon of actual and calculated acreage is presented in Diagran 3. The largest deviation occurs during acne years at the end of world War II. The effect on the acreage of spring wheat of the average prices in year t and t-2 is highly significant. The price elasticity obtained can be considered to include nest of the aggregate price influence and, t herefore, to closely correspond to the long tern price elasticity. It nay, therefore, be reasonable to consider this price elasticity as the best estinate of the long run effect of a price change. The above obtained result suggests that an increase of the spring wheat price (real price) of 10$ would, on the average, increase the spring wheat acreage about 37,000 hectares if the new price level renains Diagram 3. of Hectares Actual Calculated 170- 16m 190» lio- 130» 120. new I 100- I 90' \ ” 76 Actual and Calculated Acreage of Spring Wheat Model 12 4 ¥ 80 . . . . . . . . . . . 19153 19“ 19h5 19% 19h? 19m l9h9 1950 1951 1952 1953 195’! 1955 19 T? for three years. During the first year with the new price level an increase of only 13,000 hectares can be expected. There are, of course, nany linitations in the above conclusions. The figures only show first has mound during the period under study and under the carditions then prevailing. The response to last year's yield any seen surprising and irrational. nighyieldduringoneyeardoesnot sayarwthingabouttheyieldtha next year. The rational farner should respond only to the average yield. If, however, there is a trend in the yield, a response nay be activated. During the actual years there has been an increase in the yield. The trend coefficient is .39 deciton per hectare and year for the period 19h3-1956. Partly the response to last year's yield any be explained as an indication of the fact that the farners are rather sensitive to changes in net incone (profit) per hectare received fron the different crops. when the price is set the inccne will vary nainly in accordance with variations in yield. These variations are, therefore, becoming rela- tively nore inportant than when the prices are free. It should also be noted that the result obtained is not significant at an acceptable level of probability. The existence of sons relation between reduced acreage of winter crops and spring wheat is only what could be expected. Practically all winter crops grown in Sweden are ccnercial crops. Therefore, when this acreage is replanted in spring nest of it will be planted with co-ercial spring crops of which spring wheat is one of the nest in- portant . 78 The trend variable is highly significant and inportant. Obtained results suggest that the spring acreage in Sweden has, on the average, during the actual period increased about 5,000 hectares per year due to factors which change with trend. The trend includes all changes over tine which are not taken care of by the other variables such as changes in technology, unequal developnent of the prices of different inputs, etc. Trend has nainly historical interest because at least sole of the forces which caused the trend have now ceased. Variables Intercorrelated with the Trend. The trend variable in- cludes the changes which have cone about over tires not included in the other variables. During the period studied large charges which can be expected to have influenced the acreage of spring wheat to great extent have occurred. The wages of labor have increased nore than prices of other inputs. This will favor labor extensive production such as gain growing. Also, the developnent of technology has favored grain produc- tion nore than nest other sgriculttu-al production and has nade grain production relatively still nore labor extensive. The effect on the acreage of spring wheat of these factors are , however, difficult to separate because they are intercorrelated with each other. Here, only an attsnpt shall be nade to study the effect of the one which is be- lieved to be neat ilportant - the wage rate of fan labor. In order to study this influence, the trend has to be dropped. 9.- The nodal considered is: (13) It I a t b1 P*’(t.1)’“_2) + haw where It:acreageofspringwheatinyeart 79 Pt,(t-l),(t-2) - noving averages of announced prices in years t, t-l’ and t-a V I index for real wages of hired far- labor. Base year 1950-51. The fitted equation is as follows: (it) It a -321.6 + 8'u9*Pt,(t-l),(t-2) + 1.68"w (mu) (.38) fl Significant at .001 level * Signficent at .08 level 52 s .61 The result obtained from testing the influence of the wage rate does not provide sufficient evidence to reject the hypothesis that wages have significant influence on the acreage of spring wheat. The result suggests that than the real wage rate is increased it the acre- age of spring wheat has , on the average, increased with about 1,700 hectares. One necessary condition for this result is, of course, a given technologi cal development . Section 3. Differences in Acreage Response Within Various Parts of Sweden and Between USA and Sweden. The study of the farners responsiveness to various factors will in this section be extended to couparisons between different parts of Sweden. The different areas colpared in Sweden are show in the lap. Bus caparisona will also be nade with price elasticities obtaimd fro- Anerican data. The extent to which the farners are able to respond depends a: both the conditions under which they operate and their nsnagerial ability in recognizing and adjusting to changes in price. Different “wofSweden 80 81 results fro- two areas may be caused by either of thee factors. Only ifoneofthencanbe consideredasequelintheareas coaparedis it possible to identify the cause for the difference. The answer a sta- tistical analysis nay give is, therefore, not fully determinate. Omar-isms between different parts of Sweden are based on the sans nodel as was used in Section 2. The nodal considered is: (15) I; = a‘bl Pi’(¢;2) + b2y£_1 +‘b53 4 ht! where It I acreage of springwheat inyeartplantedinthreepartsof Sudan: Shane, Gotaland and Svealand. Pt,(t-2) 8 noving average of announced prices in years t and t-2. ’t-l = average yield previous year (deciton per hectare) B : reduced acreage of winter crops during the preceding winter (1,000 hectares) T s trend Data of average yield previous year and reduced acreage of winter crops have not hem available for the different areas studied. Data regarding the whole Sweden have been used instead. This circuetanoe nay, to sons extent, effect the result obtained. Iquation (15) fitted for M: the sou‘thernnost province and best agricultural area in Sweden, gave the following result: , (16) It = -52.? + 1.00*Pt+(t_2) + 1.35"!“1 + .165: + 1.13”r ( 35) ( .37) (.051!) ( .355) at Sigrificant at .02 level * Significant at .27 level Standard error of estinate = S a 3.9 Proportion of the variation explained £2 : .86 82 . Price elasticity of acreage E = 1.7’4 Durbin-Hatson statistic = 2.03. The twpothesis that serial correlation = 0 not rejected. Gotaland; the southern part of Sweden: an R-zmmoGthflu£)tza5t1+agh+1éfr (2.15) (1.19) (.1h) (.87) 19* Significant at .05 level * Significant at .10 level 8 ' 999 32 : .75 E : .253 Durbin-Hatson statistic I 2.13. The motheais that serial correlation 8 0 not rejected. Svealand: the niddle part of Sweden: (18) r; B -237.0 + 6.97‘Pt*(t-2) + .2131-1 - .06 a . 3.93“! (2.52) (1.3”) L15“) (.98) ts Significant at .005 level * Significant at .02 level 8 = 11.2 12*2 - .n E : h.83 Rubin-Watson statistic I: 1.21. The hypothesis that serial correlation = 0 not rejected. The obtained results are stmarised in Table 17 and the unexplained residuals are shown in Diagram h. The four included variables explain between 715% in Svealand and 861» in Shane of the acreage variation. The results obtained (see Table 2) suggest that acreage response to price is largest in Svealand and 83 Diagram h. Unexplained Residuals Thousands 0 I l I J ' 1 l I l ‘ I I l I I -5 - Gotaland; Model 17 10 - o l [ I g 1 l l l I I [ l -10 ‘ I Svealand; Model .18 10 o I a I , l I I l I l l -10 The Whole Sweden; lbdel l2 we we me E“ mm mm we er en fin en fin e» 5% 81: and snallest in Shane. The price response in Skane is both low and not significant at an acceptable level of probability. somewhat unexpected because Shane is considered to have both good This result is Table 17. Acreage Response to Various Factors Within Different Parts of Sweden. The Skans Gotaland Svealand Whole f Sweden Explained Variation in? .86 .75 .711 .76 Response to Price: The price coefficient, bl 1.00 6.13 6.97 12.86 Test statistic for signi- ficant difference freer zero (tb) 1.17 2.85 2.88 3.01 Price elasticity 1.7!: 2.5!; 11.83 2.56 Response to Last Year's Yield: The yield coefficient, b2 1.35 2.22 .21 2.72 Test statistic for signi- ficant difference from zero (tb) 2.87 1.87 .15 1.15 Response to Reduced Acreage of Winter Crops during the Pre- ceding Winter: The reduced acreage coefficient, b3 e16 e32 -.06 ea Test statistic for si?uficant difference fron zero tb) 2.90 2.30 .39 .90 Trend: The trend coefficient, bk 1.13 1.58 3.93 5.36 Test statistic for signi- ficant difference from zero (tb) 3.26 1.82 I$.00 3.10 Percentage increase per year3 5.2 1.8 7.3 3.7 a The trend coefficient divided by average acreage in the different ”8' e 85 growing conditions and good farners. Farming in this province is less restricted by the climate and should, therefore, have larger opportuni- ties to adjust the production in accordance to changed prices. One explanation to the result obtained nay be that spring wheat is a cap paratively milportant crop in Shane, where high paid crops like sugar beets, oil crops and vegetables are grown to a larger extent ttnn in other parts of the country (see Chapter II). The farmers m respond nere to the prices of these crepe and increase the acreage of spring wheat at decreases in the prices of these crops. There may also be another explanation based on the relationship between price elasticity and trend. Both the price elasticity and the percentage increase in the spring wheat acreage were higher in Svealand than in Gotaland and Shane. It is also a well-known experience from agriculture that expansion is Inch easier than contraction. The price elasticity is normally higher in expansion than during contraction. When the trend is large the greatest part of the changes consists in increasing the acreage. At decreasing prices the acreage remains un- changed or only slightly decreased. There have not been such difficult technical adjustments for the spring wheat growing farners during a price decrease. They have minly postponed the investment decisions for expanding the spring wheat acreage. The unequal trends may, to sons extent , explain the difference in price elasticity between Gota- land and Svealand. In Shane the price elasticity has so low a simi- ficance that no conclusions can be drawn in this respect. ' Thvresponse to last year's yield of spring wheat is significant in Shane and not significant at an acceptable level of probability in 86 Svealand. The high influence from last year's yield on the acreage of spring wheat in southern Sweden is surprising. There does not seen to be any reason for a greater response to last year's yield in one part of the country than in another. The only rational reason for a response to last year's yield should be that the yield of spring wheat has increased more over tine than other crepe and made spring wheat production relatively nore profitable. Even if there is nothing which indicates that this increase has occurred unequally in different parts of the country the result night have been different if the average yield for the different areas had been available. There is also sone intercorrelation with the trend which an effect the result. That sone relationship could be found between spring wheat acreage and the acreage of winterkilled crops is natural. The result suggests that a larger part of the spring wheat acreage in the south is plated on acreage where the crops were winterkilled. Contributing to this result is probably the larger production of the winter sensitive oil crops in the south and also lack of availability of data for each of the areas. The trend is ilportarrt in all the three areas but especially in Svealand. The new varieties of spring wheat ripen earlier in fall and can, therefore, be grown nore to the north. This circunstance er- plains sons of the larger increases in the spring wheat acreage in Svealand. Contributing to this larger transfer to a labor extensive crop is the fact that labor is more costly in the highly industrialised Svealand. 87 when with Price Elasticities Obtained in USA. The estinates made in this thesis are all based on announced prices. Time and data have not been available now to expand the estimates to include also a period when the prices were not announced. It is, therefore, difficult to evaluate the effect on the price elasticity of the system with announced prices. Comparison with results obtained in USA nay give sons ideas about the effect of announced prices. There are, however, considerable linitations to such comparisons. Besides the influence on the farners responsiveness from the announced prices, the elasticities obtained from Anerican and Swedish data nay differ for other reasons. The nedels for estimation and the index for deflation are different as well as growing conditions in the two countries. Also, crop acreage has been relatively stable in the USA while there has been a rapid ex- pansion in the wheat acreage in Sweden. Therefore, it is very difficult to evaluate the effect of each of these factors. Price elasticities obtained from American data are conpared with results obtained from Swedish data in Table 18. Obtained response to one year's price as well as to several years' prices are show: for both countries. The response to last year's price based on American data can be considered to correspond to the Swedish response to announced price for the present year. The American results obtained by Rerlove's nethod include influence fron several year's prices and can, therefore, be considered to correspond nest closely to Swedish results including nore than one year into the price variable. Both the short tern and the long tern price elasticities obtained from Swedish data were higher than the American ones. 88 Table 18. Price llasticities Obtained from Swedish and American Data. The Estimates Based on: Swedish Data American Data Response to Response to Announced Last Year's lerlove's Price in rr. Price ”u“ t, t‘l * and t42 Spring Wheat Spring Wheata Wheatb Barleyc llasticity of acreage response to price 0.87 2.69 0.31 0.93 1.32 3'70d 0053 Gel-9 Variation ex- plained by ca and trend R 0.59 0.71 0.23“ 0.72 0.80 ‘Ohteined by rohle and Paarlberg, 93. gig. bObtained by llerleve in using his method. °0btained by Brandow in using lerlm's nethod. See Journal of Fara Boone-ice, p. 719, August 1958. dreroentege increase per year. he trend coefficient divided by the average acreage. °Variation explained only by the price. Sue of these differences can probably be explained by the very un- equal trends. Nerlove's results are based «3 the period 1909-32 1'“ the technological development was colparatively slow. The Swedish re- sults are based on a period 1918-56 with rapid technological changes especially within grain production. The trend in tb Swedish date. from spring wheat production is about seven tines as high as the trend ob- tained by lerlove in wheat production in USA. 89 In the results from different parts of Sweden it was observed that the area with higher trend also had a higher price elasticity. This relationship seals also to be reasonable fron an agricultural technical point of view. The elasticity is normally greater in agriculture during expansion than during contraction. Therefore, it is lihely that a substantial part of the difference in the price elasticities ob- tained from the Swedish data and the American data is due to the differ- ences in the trends. The results obtained do not, however, provide sufficient evidence to reJect the tupothesis that the Swedish acreage response to price m be higher due to the systen of announced prices. It seems reasonable to assme that at least sole part of the difference in price elasticities found in USA and Sweden is caused by the system of amourced prices. mm MAIDCOICIBSIONB When World War II broke out the Swedish narket becale isolated from the world market . In order to prevent mreasonebly high prices for sons cmodities due to a very inelastic amply, price controls were established. The gal for the agricultural price policy was to naintain the relation between aggregate income and cost in agriculture. which existed during the pie-war crop year of 1933-39. In other words, the prices of agricultural products were to be set so that the aggregate incone in agriculture would increase at the sane rate as aggregate cost. In this situation, where unnecessary increases in price mist be avoided in order to prevent inflation and rapid increases in production were ilportant, it was believed that a given increase in the price would give a larger response if the prices were annomoed before the farners planted their acreage. A system of announced prices was introduced in 1913 and operated until 1956. After 1956 a return was made to a freer pricing system and prices were not announced in subsequent years. One of the objectives of this study has been to mks use of the circmstance that the farners knew the prices of the crepe before they made their decisions about acreage allocation. This circunstance has made it possible to separate the effect on acreage allocation of annmrnced price in the current year from the effect of previous years' prices. There are reasons to believe that the farners response to a given price level also depends on the duration of the price. High fired costs in 91 agriculture often make an inediate response irrational. But in the course of a few years the proportion of fixed costs will be lower and acreage which is now bound to its present use with high fixed costs will be released for other uses. The effect of a price change can, therefore, be expected to core partly after sons tine has elapsed. In other words, previous years' prices can be expected to have acne in- fluence on the acreage allocation in the current year even thong: the price for this year is announced. Another objective of the study has been to gain increased knowledge about what factors and to what extent these different factors determine the acreage allocation in Sweden. For this purpose separate investigap tions have been conducted for different areas in Sweden. The studywasbasedononlyone crap- springwheat oforthe period with annornrced prices, 1913-56. These linitations will restrict the useability of the obtained results to this crop and this period with the system of forward prices. The intention of this study was mainly as an introduction to a acre extensive future aim” of supply response in the Swedish agriculture. The reason for the choice of spring wheat was that the price influence on the acreage of this crop was considered to be cooperatively large. The acreage of spring wheat is not affected by the weather to the sale extent as winter crepe and spring wheat is one of the nest important spring cash crops in Sweden. The relationship between acreage as the dependent variable and the various factors influencing the size of the acreage as independent variables has been studied using multiple regression techniques. Different nodels have been constructed in order to study: (1) Short 92 and long tern price response, and (2) factors of importance in the acreage allocation. The prices of spring wheat have been deflated with the Producer Price Index for total production. Results obtained from the statistical analysis suggest that there is an influence from previous years' prices despite the fact that the coning year's price is known. When an increasing umber of prices in previous years were included into the price variable the variation ex- plained and the price elasticity increased Just as the significance of these parameters increased up to the third year included. When the prices for different years were included separately the results obtained indicated that announced prices in the current year and the two years prior to the current year influenced the size of the acreage nest. Also this result suggested that there existed a relationship between previous years' prices and acreage. The practical importance of these findings is the different effect the sane price will have depending on the duration of the price. If a price lasts for a long tine the effect on the acreage allocation will be larger than if the duration of the price is short. The investigation of the various factors that determine the acreage of spring wheat gave the result that the price of spring wheat and the trend explain nest of the variation of the spring wheat acreage. Bone influence was also found fron last year's yield of spring wheat and the size of acreage with winterkilled crops. The variation explained was 76$ when these four variables were included into the nodal. The results obtained suggest that a real price increase of 10% for spring wheat eould, on the average, increase the spring wheat acreage with about 93 13 ,000 hectares during the first year of the price increase and 37,000 hectares after the lapse of three years if the price remains on the same level during this time. The trend which includes all changes over time in technology etc. not included in the other variables is important in explaining changes in the acreage of spring wheat. 0n the average, the results obtained suggest that the spring wheat acreage increased about 5,000 hectares per year during the period considered due to the factors which change with trend. Spring wheat is, like other grain production, labor exten- sive and has, therefore, been favored by the large increase in the wage rate of farm labor from l9h3 to 1956. Grain production has probably also been more favored by technological innovations in the forth of combines, tractors, etc. than most other agricultural production. These two caponents - higher wages and increased mechanization - constitute probably the nest inortant part of the trend. Only the influence of the wage rate on the size of the spring wheat acreage has been investi- gated. When the trend was dropped a significant positive correlation was found between the real wage rate and the spring wheat acreage. The result suggests that when the real wage rate was increased 1% the acre- age.of spring wheat was, on the average, increased 1,700 hectares. As the average increase in the real wage rate has been about 2$ a substan- tial part of the trend may be explained by the wage rate. However, in addition to the wage rate there are several other factors like mechani- zation, yield per hectare, etc. which also have contributed to the in- crease of acreage. The wage rate alone would probably not have had this effect on the size of the spring wheat acreage. 9h Lass important but of some influence on the acreage of spring wheat were last year's yield of spring wheat and the size of the acreage on which winter crops were destroyed during the preceding winter. The comparisons between different growing areas in Sweden suggested that the farners in Middle Sweden were more price responsive in growing spring wheat than producers in Southern Sweden. This result may be affected by different growing conditions, different size of the farms, different nanagerial ability on the part of the farners to adJust the production to changed prices, etc. The result obtained allows one are plauatien which may be of sons interest. There probably exists a res lationship between the trend and the price elasticity. when the trend was high the price elasticity also appeared to be higher. This observes tion can be related to the experience fron agricultural production that price elasticity is higher during expansion than during contraction. The relatively faster expansion of the spring wheat production in Middle Sweden nay explain at least sens part of the difference in the price elasticity between the Middle and Southern part of Sweden. The price elasticities obtained from the Swedish data for spring wheat are about two to three tines as high as elasticities computed from American data. There nay be many reasons for these differences. The index used for deflating the price data and the models for the estimates are different. Also different growing conditions may have had some in; fluence. Especially should be noticed the very rapid increase of the spring wheat acreage which has occurred during the period studied in Sweden. The trend in this data is about seven tines as large as 95 the trend obtained by Nerlove in wheat production in USA. The differ- ence in price elasticity obtained from Swedish and American data nay, therefore, in some part, be due to the differences in trends. Because of the many factors which may cause the differences be— tween the price elasticities obtained in USA and in Sweden for spring wheat it is not possible to draw any certain conclusions about the ef- fect of the system with announced prices. It seems, however, reason- able te assure that the higher price elasticity obtained from the Swed- ish data at least to some extent is due to the system of announced prices. But only an expanded study including investigations for an earlier period when the prices were not administered can give an ac- ceptable answer to this question. Motions for further studies. As earlier has been nentioned, this study is by no means complete. It was intended primarily as an introduction to a nore extensive study of the supply response in the Swedish agriculture. Such a study should include the Iain agricultural products in order to be representative for all of Swedish agriculture. Secondly, the effect of the pricing systen with smeunced prices on the supply response cannot be fully evaluated without comparisons with a period of free prices. Thirdly, it can be expected that the supply elasticity varies between farms of different size. A complete investi- gation should, therefore, include an estimate of the responsiveness within various size groups of ferns. Such a study may also give some information about expected changes in supply response due to the grewth in size of Swedish fares. DATASEBIESUSEDIITEANALYSIS 97 DATASERIEUSEDINTHEAHAHBISI Acreage of Spring Wheat Average ‘Beduced --__The’ field .Acreage Whole of Spring of ‘Tear Sweden Shane Gotaland Svealand Wheat wggtzz ___ - . - - Thousands of Hectares - - - - Jgsrgafl::uei T::::;;:;___ 19h3 115 11.0 76.11 3h.7 1L5 15 19th 83 7.6 60.3 22.11 111.8 26 19h5 87 . 7.9 63.1 23 .9 15.h 17 19h6 91 12.2 6L3 2h.9 17.h 11 19h7 152 27.8 107.0 h3.0 1h.7 '62 19h8 1h3 21.1 86.8 53.9 19.2 39 1919 rho 16 .6 81.h 55.8 19.5 12 1950 161 21.7 90.8 67.3 18.9 22 1951 177 21.8 106.7 70.0 12.6 1+1 1952 168 15.9 89.3 78.8 18.9 7 1953 198 35.1 119A * 78.1 23.3 67 195k 212 37.7 121.2 88.9 20.7 26 1955 155 35.0 103.h 59.7 16.3 27 1956 11m 31.3 88.7 50.2 23.9 86 98 DATASERIESUBEDINTEEAEALISISII W Announced Producer Index for Consumer CM Year Price of Price Index wages of Price Index 1 Sun-31 ““8- 3:22“ £3.32: 1.1:: erases per deciten 1950-51sloo 1950-51-100 1950-51-100 19h3-th 28.50 76$ 52.7 81.7 19M5 28.50 75.5 58.1 81.5 19h5-h6 28.50 76.9 6L3 81.5 l9h6-h7 28.50 80.2 71.5 83.3 19117-118 31.50 85.8 80.7 86.6 1911849 311.50 88.7 86.5 88.9 1919-50 35.10 90.6 89.5 90.2 . 1950-51 32.50 100.0 100.0 100.0 1951-52 13.00 115.2 123.7 110.1 1952-53 55.10 125.1 133.2 11h.0 1953-5h 117.10 120.3 138.1 115.2 19511-55 115.00 121.3 151.8 117.9 1955-56 h3.20 133.3 162.9 122.7 1956-57 I13.85 135.8 173.1 128.0 BIBLIOGRAPHY Bean, L. H. "The Farmers' Response to Price,” Journal of Fern Economics, Vol. 11, 1929. Black, John D. Mtural Reform in the United States, New York: 1929. Cassels, J. M. "The Nature of Statistical Supply Curves," Journal of Pans Economics, Vol. 15, 1933. Cremarty, William A. Economic Structure in American $1th , Ph.D. thesis, Michigan State University, est Lansing: 1957. rersberg, Lennart. "Pregnea och Slutligt Resultat i Jerdbrukskalhylen," Lsntbrukshegshelan, Uppsala 7, 1956. Friedman, J. and Foote, R. J. utational Methods for Systems of Simltaneous muons, Agricultural Handbook No. 9h, .8. Departmnt of culture, Agricultural Marketing Service, Washington 25, me. Government Bill No. 165, Stockholm: 1956. Hathaway, Dale R. The Effects of the Price mm Prom on the M Dean Indust in Michi on, Michigan tats College, Ayicultural Eminent Station, East Lansing: April 1955. Helmstrom, Sven. "Producentpriaer och Produlrtionsvarden", Jordbrukets Utredningsinstitut, Stockholm: 1938-39 - 1956-57. . cultural P011 in Sweden: A M; of Government Measures 1 - , ordbruhets redningsinatitut, Stockholm. Johnson, D. Gale. "The Nature of the Supply Function for Agricultural Products," American Economic Review, Vol. 1+0, No. It, 1950. Johnson, Glenn L. Burlez Tobacco Control m, Bul. 580, Kentucky Agricultural Experiment tation, University of Kentucky, Lexington: 1952. . ”Supply Function - Sons Facts and Notions," ggicultural Adjustment Problans in a 03.91%; “02%: North Central Farm Man- agement Comittee, owa S ate 0 age as, Areas: 1958. Kohls, R. L. and Paarlberg, D. The Short-Tine Response of Agicultural Production to Price and Other Factors, Purdue University, Agri - tural firinent Station, 531-. i0. 555, Lafayette, Indiana, 1950. 100 Kuznets, G. 1!. Statistics Pertaggg to the Acrem of Saga Boots in California -1 , versity of Cal fornia, 0 age of Agri- culture, Kg‘rTculth-E Experiment Station Bulletin, Berkeley: 19h3. Lind-tron, Ingver. Jordbrulnets inkonstlikstalliggt , (Income Parity of the Suedish Agricultural opulation , Jo rukets Utredningsinstitut Stockholm: 1956. Ham, '1‘. B. Famer Opinions and Other Factors Influencing Cotton Pro- duction and Acre e Ad ustnents in the South, . . partnent of Agriculture Cir. No. 2513, Washington, EC" 1933. Nerlove, Marc. Estimates of the Elasticilties of m of Com, Cotton and Wheat, Dissertation at John Hopkins University, t re, Maryland , 1956. Osvald, Hugo. Swedish Moulimre, The Swedish Institute, Stockholm. Shepherd, 6. S. égcultm‘al Price Analysis, Fourth edition, The Iowa State College ss, s: 9 . Snith, B. B. "Forecasting the Acreage of Cotton," Journal of American Statistigal Association, New Series, Vol. 20, 1925. Walsh, R. M. "Response to Price in Production of Cotton and Cottonseed," Journal of Fara Economics, Vol. 26, 19M. ROOM USE ONLY ...J- «919' r m if: rt lkwl'“. l". 2“ 4»!- PL'.‘ '— ”TITI'ITIQHILTIMIMTHfiflilflffllfilifllflljffliifllfl'es