1.. l . . J ‘V. LAND ' “arse and the .. regions. The two lctcrs which he firteen regions. :‘ie original ti 355’ was used t fixated 0r re‘ Land yak {35519 in the a: firm.” Prices, in» ABSTRACT CHANGES IN LAND VALUES IN THE UNITED STATES, 1925-1962 By Arne Larsen During the last two-three decades the value of land has increased in the United States both relatively and absolutely. Land values have increased rapidly in the face of fluctuating or even declining net farm income and the increase in land values has varied greatly among areas and regions. The two main objectives of this study were: (1) to delineate factors which have caused a major part of the variation of and increase in land values, and (2) to investigate the causes of regional variations in land value increases. Time series correlation analyses for the period 1925-26 were carried out by state for the 48 states in the conterminous U. S. Combined cross-sectional and time series analyses were used for thirteen regions. Due to strong trends and multicollinearity problems in the original time series data, a model employing differences of first order was used throughout the study. The use of first differences also eliminated or reduced autocorrelation problems. Land value per pasture acre equivalent was the dependent variable in the analyses. Independent variables were: (1) index of expected prices, (2) government expenditures on agricultural conser- vation, (3) government expenditures on the conservation reserve part fine soil bank prfi :ggef'x'e part Of thf grime, (6) {61' rare, (8) pOpuI l'ne stans: s.g:s which wculc' :2;i::a1tlrne seri agitance of the 52:25 analys e s. éflerally decreas 533-5heter3genei' The indice ~ . ‘.l .A 45195 in the a Tia-43E 1 e an: Go‘v-ernme Hcan) mos :55“. ”ea-He lne Out“ e V‘Qtlon‘ “Pi‘ 2". 9 "6:1“; a . .8, L815 AR NE LAR SEN of the soil bank program, (4) government expenditures on the acreage reserve part of the soil bank pregram, (5) output per man-hour in agriculture, (6) fertilizer use per pasture acre equivalent, (7) output per acre, (8) population density, and (9) personal per capita income. The statistical analyses generally yielded the coefficient signs which would be expected from an economic point of view. The regional time series and cross-sectional analyses increased the significance of the estimated coefficients as compared with the time series analyses. However, inclusion of cross-sectional variation generally decreased the coefficient of multiple determination, indicating some heterogeneity among the cross-sectionally combined states. The indices of expected prices were among the most important variables in the analyses. The indices were mainly important in explaining the annual variation, while they had little to do with the increase in land values. Government expenditures on conservation were highly correlated with land value changes. The conservation expenditures were associated with a substantial part of the relative increase in land values. Of the soil bank variables the conservation reserve variable was clearly most important. The coefficients of the conservation reserve program were larger and generally more significant than those of the acreage reserve program. The output per man-hour variable had generally a high simple correlation with land values. However, due to intercorrelation problems, this variable was in many cases replaced by fertilizer use or output per acre variables. v" T:- ir‘e la. .04 .’ V . - 3:21- 3.9“: ....~g seen”; #.,...~ .‘J v T?- l ..E «of- a: 9‘!- at o—IoO'~" I I‘D-A nb‘ ,.-~v u,» :7. .5 ‘ .ld - o. .- ~ H. ..oa~’; ‘ya . I 6‘..C’A. -W‘; AR NE LAR SEN The largest coefficients for a unit change in p0pulation density were in areas with large relative increases in population. Personal income seems to have affected land values in only a few regions. The importance of conservation practices in the analyses implies that large increases in land productivity are gained through conservation. Therefore, a main part of the conservation subsidy program is in conflict with government programs intended to reduce or stabilize the supply of agricultural products. The relative changes which have occurred in the explanatory variables during the examined period indicate that the regional differences in relative land value increases are in large part caused by government subsidy programs. The capital gains or losses occurring from initiation or termination of government agricultural programs differ widely among regions. LAN D l CHANGES IN LAND VALUES IN THE UNITED STATES, 1925-1962 by Arne Larsen A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1966 Tlis th 3, Band H. 1 ---‘c‘s...S, an .:-:' :articularl , . ;!'=' ..;~‘| a Dr. L. .. ‘1. ‘- I 33.35? 33.1 {REILLY aCKZ 312:5!le C11: 2;;rec1atec. I am p :3: tie financ :' 1.4!. ‘!~;~- -- «tween 5 mm 311 11:13 1 .fl-A' I ~vv::mlc s a! a F;“-“ '4’. u‘d.‘ 1“? "-531 Stu u.‘; ACKNOWLEDGEMENTS This thesis was written under the general supervision of Dr. David H. Boyne, my major professor. Dr. Boyne's guidance, criticisms, and encouragement throughout my graduate studies and particularly with regard to this thesis are gratefully acknowl- edged. Dr. L. V. Manderscheid undertook careful review of earlier drafts of this thesis. His comments and suggestions are gratefully acknowledged. Dr. G. L. Johnson's comments and criticisms during the development of this thesis are greatly appreciated. Discussions with a number of faculty members and fellow graduate students in the Department of Agricultural Economics assisted in the development of this thesis. Assistance was also extended by several individuals in the USDA, who sent comments or were helpful by sending unpublished materials. I am particularly grateful to the W. K. Kellogg Foundation for the financial assistance which made possible my graduate studies at Michigan State University. For the financial support during the work on this thesis, I am grateful to the Department of Agricultural Economics and Resources for the Future, Inc. Finally I wish to thank the students and faculty members at Michigan State University, who, through our discussions, had a part in making my graduate studies a stimulating and rewarding experience. Arne Larsen ii Enter 1 INT} 1A} '” "."1 4 SI." T: L 0‘. 5! Y‘ 1 Q Chapter I II III IV V TABLE OF CONTENTS INTRODUC TION ............. . The Problem . ...... . ..... OutlineoftheWork. . . . . . . . . . . . LAND VALUES AND RELATED FACTORS . Research on Land Values in the U. S. Warranted Values. . . . . . . . . . . Studies ona State Level . . . . . . . Price Support, Allotments and Land Values Conservation and Land Values . . . . STATISTICAL METHODS: 0 EVALUATION AND INTERPRETATION . Supply Function for Land. . . . . . . The Model . . . . The Use of a Difference Model of First Order................ Features of Time Series Data . . . . Selection of Time Period. . . . . . . Aggregation Problems . . ...... Summary ............... THE VARIABLES O O O O O O O O O O O O O ValueofLand........... ..... Expected Prices . . . . . . . . . . . Conservation............. Influence of Technological Changes . TheSoilBank............. Population Density . . . . . . . . . . Personal Income . . . . . ......... ANALYSIS OF LAND VALUES BY STATES AND REGIONS. . . . . ....... southeast 0 O O O O O O 0 O O O O O O O Appalachians.............. Northeast. . . . . ...... . . . . LakeStates............ ..... Corn Belt . . . .......... . . Delta States. . ........... . Southern Plains. . . . . . . . . . . . NorthernPlains . . . . . . . . . . . . MountainStates. . . . . . . . . . . . PacificStates. . . . . . . . . . . . . Conclusions.............. iii 15 19 20 21 24 25 26 28 38 39 41 45 47 47 52 55 59 65 68 70 72 7'7 83 88 94 98 102 105 109 113 118 121 Chapter Page VI COMPARISONS AMONG REGIONS . . . . . . . . 124 Regional Differences in the Estimated Coefficients . . . . . . . . . . . . . 124 Price Expectations ............ 125 Conservation Expenditures . ....... 128 Soil Bank Variables ............ 132 Efficiency and Input Variables. . ..... 135 POpulation Density and Income . ..... 138 Constant Terms .............. 140 Magnitude of Change in Variables ..... 142 Sources of Increased Land Values. . . . . 146 VII CONCLUSIONS AND IMPLICATIONS . . . . . . . 151 BIBLIOGRAPHY . . . ............... 156 APPENDIX A ............ . . ..... 161 APPENDIX B ................... 163 APPENDIX C ................... 172 APPENDIX D . . . . ........ . . ..... 183 iv ,.. .g. ' .3u1c Table 10 11 12 13 14 15 16 LIST OF TAB LES Variables Used in the Regression Analyses. . . . . Southeast: Results of Land Values Regressions, 1925-62....................... Appalachians: Results of Land Value Regressions, 1925-62...................... Northeast 1: Results of Land Value Regressions, 1925-62................. ..... Northeast 2: Results of Land Value Regressions, 1925-62. O O O ..... O O O O O O O O O O O O Northeast 3: Results of Land Value Regressions, 1925-620 0 o o o o o o o o o o o o o o o o o o 0 Lake States: Results of Land Value Regressions, 1925-62O O O O O O O O O O O O O O O O O O O O O Corn Belt: Results of Land Value Regressions, 1925-62 0 o o o o o o o o o o o o I o o o o o o 0 Delta States: Results of Land Value Regressions, 1925-62O O O O O O O O O O O O O O O O O O O O O Southern Plains: Results of Land Value Regressions, 1925-62....................... Northern Plains: Results of Land Value Regressions, 1925-62....................... Mountain 1: Results of Land Value Regressions, 1925-62....................... Mountain 2: Results of Land Value Regressions, 1925-62....................... Pacific States: Results of Land Value Regressions, 1925-62....................... Regional Coefficients for Price Expectations . . . . . Regional Coefficients for Conservation Expenditures . Page 73 78 84 89 90 91 95 99 103 106 110 114 115 119 126 129 u Table Page 17 Regional Coefficients for Soil Bank Variables. . . . . 134 18 Regional Coefficients for Efficiency and InputVariables................... 137 19 Regional Coefficients for Population and Income Variables. . . . . . . . . . . ....... 139 20 Regional Constant Terms (time). . . . . . . . . . . . 141 21 Relative Changes in and Size of Selected Variables by Region, 1925-62 . . . . . . . . . . . 144 22 Estimated Sources of Land Value Increases by Region, 1925-1962 . . . . ....... . . . . 147-148 23 Value per Pasture Acre Equivalent by States 1925-62 0 o o o o o o ccccc o ..... o o o o 0 164-171 24 Indices of Price Expectations by States, 1925-62............. ..... .....l75-182 25 Relative Changes in and Size of Selected Variables by States, 1925-1962 . . . . . . . . . . 184-185 vi .. :-«"[‘"\' C..‘Aa .u.-I .- TU a” T\'T 1.\; (3 LJ CI) ['1 LIST OF APPENDIC ES Appendix Page A PASTURE ACRE EQUIVALENT UNITS ..... 161 B SERIES OF LAND VALUES BY STATES ..... 163 C INDICES OF PRICE EXPECTATIONS BYSTATES..................172 D RELATIVE CHANGES IN AND SIZE OF SELECTED VARIABLES BY STATES, 1925-1962 . . . ...... . ......... 183 vii This st agat'flmal c 722 eifect of 1 2::szc'ered by , en‘- “ ‘eé-vhal v2 titates, We #5.. 5-1.,3e :‘v.‘ “- 5 ‘- 1 7‘ . 11m 1 L: 'e F0111 tie ; “Pact c CHAPTER I INTR ODUC TION This study focuses on an explanation of changes in land values. It is concerned directly with two government programs-- agricultural conservation payments and the soil bank program. The effect of government price support programs is also considered by the inclusion of price expectations in the analyses. The primary objectives are: (1) to explain changes in land values at state and regional levels, and (2) to investigate the causes of regional variation in land value increases. As the first objective indicates, we wish to study the structural relationships underlying land value determination. The second objective is to compare and analyze the differential impact among regions of the factors found important in explaining land value changes. Secondary objectives include: (1) testing the explanatory power of Lerohl's price expectations as related to land values on a regional level, (2) testing the impact on land values of different- conservation practices, and (3) testing the impact of soil bank programs on land value changes among regions. This study is part of a larger project sponsored by Resources for the Future, Inc. The objective of the entire project is to evaluate the impact of selected U. S. agricultural policies and programs on IESJL‘JCE use and 3317-62. Four s :cizpletec‘ -- one expectations by 1 1:“; me by Chen: studies by C. 1.. rescurce flow 5 : Tzearcblem K During t " in: real 85 “ 11'.) :. .1 n+5); ‘drm r‘ 1.135 re [11 (I) H ”A (D _,4 vs. u‘r‘JQA‘“ IL , \AL‘es .ILC,5 “ , “593 :91( resource use and allocation in U. S. agriculture for the period 1917-62. Four studies contributing to the parent study have been completed -- one on labor by Jones, 1 one on product price expectations by Lerohl, 2 one on farm real estate by Rossmiller, and one by Chennereddy4 on labor. Presently in progress are studies by C. L. Quance and Francis VanGigch on capital and resource flows respectively. The problem During the period 1925-1962, the index of value per acre of farm real estate in the U. S. increased almost twice as fast as the wholesale price index (WPI) of all commodities. From 1950 to 1962 farm real estate values increased more than five times as fast as the WPI. lBob F. Jones, "Farm-Non-Farm Labor Flows, 1917-1962.“ (Unpulished Ph. D. dissertation, Michigan State University, 1964.) 2Milburn L. Lerohl, "Expected Prices for U. S. Agricultural Commodities, 1917-1962. " (Unpublished Ph. D. dissertation, Michigan State University, 1965. ) 3George E. Rossmiller, "Farm Real Estate Value Patterns in the United States, 1930-1962. " (Unpublished Ph. D. dissertation, Michigan State University, 1965.) 4Chennareddy Venkareddy, "Present Values of Expected Future Income Streams and their Relevance to the Mobility of Farm Workers to the Non-Farm Sector in the United States, 1917-62. " (Unpublished Ph. D. dissertation, Michigan State University, 1965.) The relativ Emgeneous amOI argest increases “:2 Southeast and estate values in t1 fiertheastern re gi For the 1( rea‘ estate value: 2::11exas to th. laeani in the st strease per acr The relative increase in farm real estate values has not been homogeneous among states and regions throughout the country. The largest increases per acre in real estate values have occurred in the Southeast and the Pacific regions. 6 From 1940 to 1962‘7 real estate values in these two regions increased about 365%. In the Northeastern region the increase in the same period was only 188%. For the 10-year period 1952-62 the largest increases in real estate values took place in the southeast quarter of the country from Texas to the Atlantic Seaboard and south of the Mason-Dixon line and in the states on the Pacific Seaboard. The percentage increase per acre in the Southeastern region was almost three times as large as that occurring in the Corn Belt and the Northern Plains. 6The ten regions referred to throughout the thesis are defined as follows: Northeast--Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, New Jersey, Pennsylvania, Delaware, and Maryland. Corn Belt--Ohio, Indiana, Illinois, Iowa, and Missouri. Lake States--Michigan, Minnesota, and Wisconsin. Appalachians--Virginia, West Virginia, North Carolina, Kentucky, and Tennessee. Southeast-~South Carolina, Georgia, Florida, and Alabama. Delta States--Mississippi, Arkansas, and Louisiana. Southern Plains --Oklahoma, Texas. Northern Plains--North Dakota, South Dakota, Nebraska, and Kansas. Mountain--Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, and Nevada. Pacific--Washington, Oregon, and California. 7Comparable data for the regions are not available previous to 1940. For indiVl 525 to 1962 0“" Eezrgia (251 " pi: was onlY 39! fit with onlY a 6 Many 331' gephenomenon < :iez'ace of fluctu cthe suggested :rresults obtain expiairing the re received little at Although ;:::es are the re 35am: real est The real Price of land are ‘4'; agricultural ‘,:5§n : ‘ “andn 01 Certal For individual states the larges land value increases from 1925 to 1962 occurred in Florida (305%), Mississippi (261%) and Georgia (251%). In South Dakota the average value per acre in 1962 was only 39% higher than the 1925 figure. Maines was also low with only a 61% increase. Many agricultural economists have been concerned with the phenomenon of rapidly increasing farm real estate values in the face of fluctuating or even declining net farm income. Most of the suggested explanations for this phenomenon have been based on results obtained from aggregate U. S. data. The problem of explaining the regional differences in land value increases has received little attention. Although some of the increases occurring in farm real estate prices are the results of inflation, the real wealth gains experienced on farm real estate have been very large. The real gains accruing from the change in the relative price of land are of specific interest to agricultural economists and agricultural policy makers. Are the gains due to the capital- ization of certain government subsidized agricultural programs ? fl 8See David H. Boyne, Change inpthe. Real. Wealth Pos‘ition...'of Farm Qperators, 1940-1960, Technical Bulletin 294, (T964), Michigan State University, Agricultural Experiment Station. Boyne found that in the period 1940-59 farmers had experienced real wealth gains on farm real estate amounting to 36. 47 billion dollars (1960 prices). The wealth gains differed widely among regions, being lowest in Northeast and Lake States and highest in Southeast, Delta States, and Souther Plains (p. 59). This led Boyne to suggest that regional analysis Inight show whether the differences are due to varying conventional income. streams (p. 66). ‘-‘ -~ then they ‘. :4, :seme of th L. :land value ID. . e "F +1 321'16..1€1' .118 J. J 1 3221.435 .11 a GE e“ F " v; ‘ ' "it :1: mill ‘I:ch;! 1A” ‘ modVoOLE Adoac 1‘. . .‘ n r, r, J dEsMal (l, ;3C:“rarns. '“st LA. eat ‘15: e‘ -.‘ ' ~~ 159315 15 "4'84 . V “““12 lame J: 3.;‘CinE Riva \— i"‘~.. ; “94.2112 lar a V q M‘J. . , we 0: t" \i v If so, then they could be controlled by policy makers. Assuming that some of the benefits from agricultural programs are capitalized into land values, then the policy makers are faced with the question of whether the benefits are distributed among farmers and among regions in a desirable pattern. In this thesis one of the major concerns will be the regional patterns in factors believed to influence land values. Therefore, we will also be concerned with the regional differences in benefits obtained through agricultural programs. Instead of studying the variation in farm real estate values, this thesis is limited to a study of land value changes. By omitting building values we avoid concern with the changes in the proportion of building values to farm real estate values and concentrate on examining land value changes. Outline of the work Chapter II will review other studies which have investigated the price of land and related factors. A section is devoted to studies concerning the influence of conservation on land values. Chapter III is a discussion of the method used in the study and the features and limitations of the tools used in the statistical analysis. The reason for using first differences instead of original values is outlined, and the time aggregation carried out for the regional analyses is discussed. Chapter IV deals with the variables. Data for some of the variables are not directly available, and their derivation is explained. presented a: :zezzzczents ..,.' .1 ' .:5..:al cna :scussec. T1 116! The economic relationship between dependent and independent variables is discussed. Shortcomings of available data are pointed out. In Chapter V the state and regional regression results are presented and discussed. Differences within regions are evaluated, and statistical problems are mentioned. Chapter VI presents interregional comparisons. The coefficients obtained in the different regions are compared, and the regional changes in the variables over the 1925-62 period are discussed. The conclusions are given in Chapter VII. as I it , . l-Hf' ”’ flfifi" SES- 0 an. f‘ 3., F'F‘W“ J..‘......C r» b n t'» Tries ha a . es 1: . h C “h t‘ . Wt Ly..“‘l‘1“ I: f Ilse 24 pp, CHAPTER II LAND VALUES AND RELATED FACTORS Through time, the problem of how to determine the value of land has received considerable attention. Much of the interest in assessing the value of land has been for taxation and mortgaging purposes. Few studies have been concerned with quantifying the economic forces underlying land value determination. Even fewer studies have looked into the interregional heterogeneity in the land market. In this chapter we discuss studies about and factors related to land values . Research on Land Values in the U. S. In 1924 Chambers1 found a close relationship between past rates of change in cash rent to land and computed expected future changes in cash rent to land. Using a capitalization formula, and assuming the mortgage rate to be the proper discount rate, he computed the annual rate of change in the income to land that would justify present land prices. The underlying assumption was that the computed income reflected the expected future changes in rent. The method also 1C. R. Chambers, ”R elation of Farm Land Income to Farm Land Value, " American Eggnomic Review, 34 (December, 1924) pp. 673-698. 1:265 7'5; ex; is itself arr .2: realize 7 .n :s::g price Size proce ‘ . 9",". v- a H. .....-n- £6134: rate Cashi] s for 1 9 ‘9 -g“ 1‘, fl 9‘3. "I‘l‘fi 3“»; assumes that ex post information about changes in rent is used in forming expectations about future rent, and in fact they are extra- polated into the indefinite future. The factors influencing the returns to land (rent) are not brought into the study. Furthermore, the past cash rent figure is not an exact expression for income to land, but is itself arrived at by some expectational procedure, and it is likely that realized errors in former expectations would lead to revision of present and future expectations. In 1935, Thomsen2 related farm real estate values to whole- sale prices of farm products and to real estate taxes per acre of land. An expression for price expectations at one point in time was found by using prices for the past ten years, giving them weights inversely related to distance (number of years) from that point in time. The same procedure was used to obtain an expectation of land taxes using the tax rates of the past five years. Thomsen found a curvelinear relationship between real estate values and agricultural product prices for the period 1912-1933. The curvelinear relationship, which gave sharper changes at extreme values, was rationalized by asserting that land received a proportionately higher or lower part of total income for ”extremely” high or low product prices. The fixity of many agricultural inputs was suggested as the cause of the 2F. L. Thomsen, ”Factors Affecting Farm Real Estate Values in the United States, " Journal of Farm Econornicg, 17 curvelinearity. Comparing two periods, 1912-1923 and 1924-1933 respectively, Thomsen obtained a good curvelinear fit for each period. However, the two curves were at different levels and diverging for increasing values. This phenomenon was explained by changing tax rates. Inclusion of the tax expectation variable for the period 1912-1933 gave a squared correlation coefficient of . 986. No constant term or significance levels were given. Bean3 found that neither Chambers' nor Thomsen's method was satisfactory when applied to periods other than those for which the studies were made. Bean regressed the index of land values for the United States on the index of current prices received by farmers and the index of prices received by farmers lagged from one through six past periods; 1. e. Bean estimated the weights of influence of past years' prices. In the period studied, 1912-1937, Bean found that about one half of a given rise in land values was associated with prices of farm products in the current year. Using prices from the last two years explained about seventy-five percent of the changes in land values. Increasing the number of price variables, he found that the coefficient of multiple correlation increased. In a model such as Bean's the price variables would most likely be highly intercorrelated and, thus, it would be difficult to get 3Louis H. Bean, ”Inflation and Price of Land, " Journal of Farm Economics, 20 (February, 1938), pp. 310-320. p :fcoefzicier ship would 1 tzzc med 1 r,“ (U 11 significant coefficients. This problem was touched upon by Bean in his article, but the estimated equation and the standard deviations of coefficients were not given. Furthermore, this type of relation- ship would not catch the turning points in land values, and serial correlation in the disturbances would be expected. Actually, the graphs shown indicate serially correlated residuals. Bean was concerned with the “land boom and its aftermath of rural distress. "4 He found, like Thomsen, a curvelinear relationship between land values and product prices, the relationship being steeper for ”extreme" values. This led him to the conclusion: Our study indicates how a land boom arises out of the current and anticipated profits that go with price inflation. It suggests that if we want to avoid a land boom we must avoid monetary price inflation or inordinate price advances for any other reason. This conclusion might be valid for the period considered by Bean, in which product prices appeared to be the main determinants of land values. However, the "land booms" which occurred after World War II, and especially those of the fifties, cannot be explained by inflation alone. 4Ibid, p. 320. SIb‘d 31 1 , p. 9. 6See Boyne, Table 17, p. 55. For 3:55 inco I! T 5Lye, . 1.: . more :or as much o 3:55 inc< 3) (1. a ‘;m boaAi' LEIILCLET. 3 Kg; ‘1 3‘ s- r ‘a‘ 12 For the period 1920-53, Renshaw7 computed an “expected" gross income per acre and related it to farm real estate value per acre. The "expected" gross income was found by weighting gross income for the preceding ten years, the weights being derived by regressing gross income on itself lagged one through ten years. He also used a five year moving average of gross income which explained as much of the variation in real estate values as did the "expected” gross income (R2 was about . 80). Inclusion of the mortgage rate and a time variable in the analysis increased the squared correlation coefficient to . 97. The trend variable had a large negative coefficient. Renshaw's predictions of real estate values for 1955 were much too small, and he concluded that farm real estate prices in 1955-1956 were too high. Hoover8 studied land prices in the United States, 1911 -1958, using a model which postulated the change in expected returns to land as a function of the deviation between the previous year's actual returns and expected returns. Variables of returns to land included the price of crops, crop production, the rent-income ratio, expected price level changes, and the dividend-price ratio. The relationship was assumed to be linear in logarithms. Hoover obtained large multiple correlation coefficients which were, however, mainly due 7Edward F. Renshaw, " Are Land Prices Too High: A Note on Behavior in the Land Market, " Journal of Farm Economics, 39 (May, 1957), pp. 505-510. 8Dale M. Hoover, "A Study of Land Prices in the United States, 1911-1958," (Unpublished Ph. D. thesis, University of Chicago). u =2 11‘. For r. grI Ir : 1 "EV, <12 “32221;; 10 l Iyazsron VV 8 56.18 1 o g -’ and he ars 1911- see 'rplainec a between re. .o‘ D . b,» Jul. “”1“,“o." v.93»...- 11. 12 13 to the inclusion of lagged land prices among the explanatory variables. Omitting the correlation due to the lagged variable, Hoover's model explained a much larger portion of the land price variation for the years 1911 -1941 than it did for latter years. For the 1940-1957 period Scofield9 found a closer relationship between gross national product and the price of farm real estate than between real estate prices and net farm income. He also pointed out the "belief that land offers safety and protection of capital from loss of purchasing power during periods of inflation. "10 Scofield also refers to the impact on land values brought about by conservation and soil improvement practices. He mentions the effect of an increase in population and the eventual land shortage it might bring about. In 1964, Scofield suggested three main reasons for rising land prices: (1) keeping up with inflation, (2) capturing the gains from new tech- nology, and (3) economics of scale. 11’ 12 The importance of the expansion buyer in the land market is stressed, and references are made to the non-farmer investor buyer, credit availability, and impact of programs like the federal highway programs. Much 9William H. Scofield, "Prevailing Land Market Forces, ” Journal of Ffiarm Economics, 39 (December, 1957), pp. 1500-1510. loIbid, p. 1501. 11It should be noted that economics of scale has not been revealed in cross-sectional Cobb-Douglas studies. 12William H. Scofield, ”Dominant Forces and Emerging Trends in the Farm Real Estate Market, ” (Paper prepared for seminar on land prices, North Central Regional Land Economics Committee, Chicago, Illinois, November 12, 1964). ,. o finiaflCE : ' .1“ gweazt... Sco i525: 5716‘; 5:31:21: :1: eLargeme :erxeen :1; q -w- u- _..:IICE C per acre t . ' ‘ IMF-a 1"- ..:.. g ‘4 3.1.29 2C 8 l wees tc \- 14 importance is attached to the desire to accumulate land as a form of wealth. Scofield attaches highest research priority to a comprehensive depth study in the area of the "economics of farm enlargement, ” and although he rec0gnizes the connection between economics of farm enlargement and technological advances, he finds little connection between his technology measure and land prices. In testing the influence of changing technology on land values, he uses net return per acre to labor, land, and capital as an expression for techno- logical change. This does not seem to be an adequate test of the influence of technology. We would expect many technological changes to reduce labor and/or capital requirements per unit of land, leaving a relatively larger part of the returns to land. Only in the cases where the technological changes have brought about increased output, but unchanged labor and capital requirements, would Scofield's test be appr0priate. The influence of the soil bank is also discussed by Scofield. The benefits derived from the conservation reserve part of the soil bank, which are more closely associated with ownership of land, are suggested to have had more influence on land prices than did the acreage reserve part of the program. The conservation reserve program might encourage farmers eligible for payments to retain ownership of some land that might otherwise be sold, i. e. increase the demand for land. '2'.u—"‘Or ”1.. 6"" ~ ...;O~’I; ;51,-LG.AZ€ .. ,- .. . ... .: 1.4-.ZEE firears s-é- ~. 9 a, w”. '~ F I... , V . .l".. l— ...e .‘ HA F ‘ 15 Warranted Values In 1948 Larsen13 computed "warranted” values of land for the years 1910-1948. The warranted value for any year was the capitalized actual net rents for the succeeding ten years, plus the capitalized average (1910-48) net rents for the entire period beyond ten years to infinity. The capitalization rate was approximated by the average mortgage rate plus one percent. The one percent was added to reflect the additional risk an owner has as compared with a mortgage holder. For actual land values to be equal to the warranted values we would have to assume that farmers had perfect foresight of net rents for the next ten years and that beyond ten years to infinity, they would expect the average 1910-48 net returns. A poor relation- ship between warranted values and actual values indicates that the rent expectations were different from those which actually occurred. From the figures shown it appears that a change in warranted values precedes a change in land values by three to four years. This is not too surprising since a change in actual returns would probably precede a change in expected returns. Actual land values, as compared with Larsen's estimates, indicated that an extremely high capitalization rate was used during the war years. This indicates a changing risk expectation, and questions the feasibility of using the mortgage rate 13Harold C. Larsen, "The Relationship of Land Values to Warranted Values, 1910-1948, " Journal of Farm Economics, 30 (August, 1948) pp. 579-588. real estate {hear in l was the vs 1.3:. rea :bianed fr :zeteen z Vv'. ('"H‘v up fiaot. 8.1Le . u 1 _ ll I~ “t“ iEQ ‘1 V -\ ~.,.‘1\E ‘ ¥“ 1 . LI‘] ‘i‘h‘ - N s.‘p,‘ ,. M 16 or the mortgage rate plus some constant as the capitalization rate. Rossmiller14 computed the marginal value product of farm real estate by fitting a restricted form of the Cobb-Douglas function (linear in logarithms) for the yearsl930-1962. The dependent variable was the value of total farm output, and the independent variables were farm real estate, labor, operating expenses, and time. Data were obtained from representative type farms in 19 different areas, and nineteen zero- one variables were included to account for area differences. For each area, an ex post farm real estate value series was computed by capitalizing the marginal value products of farm real estate for 34 years ahead. In cases where less than 34 MVP terms were available, the average from 1958 to 1962 was substituted for the remaining terms. The capitalization rates used were the rates charged for new loans by the Federal Land Banks in the re3pective areas. An ex ante farm real estate value series was computed for each area by capitalizing an average of the MVP's for the past five years. The ex post series were consistently below the ex ante series in the production function model. The ex post series indicated that in most areas market prices are below what could be paid to real estate under the assumptions of the production function model. 4George E. Rossmiller, "Farm Real Estate Value Patterns in the United States, 1930-1962. " (Unpublished Ph. D. dissertation, Michigan State University, 1965.) It at ante Se ——-——'- "v- . u‘. “ . I “S‘T‘I‘v c‘»' I: f“. 17 It is not surprising that the ex post series were below the ex ante series. This is inherent in the model used. As mentioned by Rossmiller, the capitalization of the 34 first terms of an infinite series, using a 5 percent capitalization rate, only accounts for 81 percent of the ”true" capitalized value. Also, there should be no long run trends in the MVP's. However, the downward bias in the ex post series would put them consistently below their comparable ex ante series, which were capitalized for infinite MVP series. An important point in relation to Rossmiller's MVP derived series, is that he did not subtract taxes and depreciation from the MVP's before they were capitalized. It might be difficult to find a reasonable figure for depreciation, but the fact that no allowance was made would tend to lead to overestimated farm real estate values. The average tax rate per $100 of farm real estate value in the U. S. has varied considerably over time, but would for the period 1930-62 average close to $1. 00.15 Thus, using a capital- ization rate of 5 percent, the capitalized value of taxes would be close to -$20 per $100 farm real estate value. Therefore, the upward bias in the ex post series because of failure to adjust for taxes would on the average be offset by the downward bias caused by using 34 terms only. However, this applies to the ex post series only, the ex ante series are still biased upwards. 15U. S. Department of Agriculture, “Land Values and Farm Finance, " Mam Statistical Series thhe USDA, 2, Agricultural Handbook No. 118. (Washington: U. S. Government Printing Office, 1957) p. 34. . ‘. t '. :lrael. przt C I. s- n .0 T54 ,- 50‘.‘ A. ‘00 ‘ . l’:'t‘ " _ 1'::.C2 ‘il 9 u 0""... A,“ FA” in -'~.u\.slv.. . o “‘ strain the *4: Lie on 3.. “5‘ «'16 r» at. ‘1“ “ §:'V‘-1, :‘Nd; ‘fiirnl 18 Since taxes vary considerably among states, 16 Rossmiller's regional comparisons would probably need some revaluation. Rossmiller's conclusion, based on ex post series, that market prices are below what could be paid to real estate is valid only if marginal value products of real estate do not fall below the 1958-62 average and if the coefficients derived by the aggregate production function model are reasonably reliable. Different ex ante and ex post series were computed in a similar manner using residual returns to farm real estate in place of marginal value products. The residual return is net farm income less an imputed return for operator and family labor and non-real estate capital inputs. The series derived from the computed residual returns have, since early post World War II, with few exceptions, had the Opposite trend of the market value series. Rossmiller argues that the imputed returns to labor and capital are probably too high. Hence, since these are subtracted from net farm income in order to obtain returns to land, the returns to land are underestimated. Rossmiller suggests that the main reason for increasing marginal products of farm real estate is due to the technological revolution going on in agriculture, and that "current land prices are below what expansion buyers could afford to pay for farm real estate. "17 For instance, in 1957 the tax per $100 farm real estate was . 46 in Arkansas and 2. 32 in Maine. (From photostats of USDA worksheets on farm real estate.) 1 71bid, p. 136. V. ,‘JC’WI‘S "'-l:b$$ ‘ , n V .anfi‘4‘ Osqo logaobob ‘L 'l reporte cl. O-fi. ¢-. ‘Z‘QA edeLe 111‘.- Year's “i, A 4 41.21 uLLD‘ ‘1 .~ e6 uh“ L as s‘r‘ 19 Like Scofield, Rossmillerl8 discussed several factors which might affect land values. These included conservation, soil bank programs, price support, and farm consolidation. However, no quantitative relationships between the factors and land values were reported. Studies on a State Level In a Minnesota19 study, covering the period 1939-1957, farm real estate value per acre was regressed on net income per farm, man years of labor available divided by labor requirements, total farm output divided by total farm inputs, debt as percentage of farm assets, voluntary farm sales, and security yields divided by the farm mortgage interest rate. Only the output/input and the voluntary farm sales variables showed any reasonable significance. In a cross-sectional study of Indiana land values Scharlach and Schuh20 regressed value of land and buildings per acre on population density, farm expenditures, distance from Chicago, farm wage rate, property tax, land capability, fertilizer, and average size of farms. With the exception of the coefficient for fertilizer and average size, all the estimated coefficients were significant lSIbid, Chapter III. 19R eport on the Governors Study Commission on Agriculture, Minnesota, (St. Paul: Office of the Governor, 1958) pp. 194-195. 20Wesley C. Scharlach and G. Edward Schuh, “The Land Market as a Link Between the Rural and Urban Sectors of the Economy,‘ Journal of Farm Economics, 44 (August, 1962), pp. 1406-1411. 1" ‘ V a 1.. l" . a “C I -' ‘ \fi‘ 1. VA. as: dolla ‘7‘ .-. 'R A, iciuaca Via. :terev‘“ ~ ~....; ‘ - O‘ . . .11"? a ~10“ "Fg ~oV;o 0 A “fin . . fess," f‘ " . m A. JI‘EH‘KJ‘. V «n l 2‘33 Gib S 0: :e pe:. ”I; Auk m M....,_en_t h \/ ' VA '5’ :r‘lg. 20 at the . 05 level. The multiple correlation coeffient was . 89. Some of the coefficients seem unreasonable. For instance, a one dollar increase in farm expenditure per acre increased the real estate value by twenty dollars. A one dollar increase in taxes per acre decreased the real estate values by two and one half dollars only, indicating a capitalization rate of about 40%. It is interesting to note the significant, though small, influence of population density and distance from a large city on real estate values. Price Support, Allotments, and Land Values A few studies report on the influence of price support and production control programs on land values. In a study of flue-cured tobacco allotments Maier, Hedrick, and Gibson21 found, using a multiple regression analysis, that for the period 1954-57 the market value of an acre of flue-cured tobacco allotment (without associated land and buildings) could be as much as $2,500. Regressing sale value of farms on several variables including acres of peanut allotment, acres of crOpland, and acres of non-cropland, Boxley and Gibson22 found that an acre of peanut allotment, independent 21F. H. Maier, J. L. Hedrick, andW. L. Gibson, Jr., _'I_‘__he Sale Value of Flue-Cured Tobacco Allotments, Technical Bulletin No. 148, Virginia Polytechnic Institute, 1960. 22R. F. Boxley, Jr. andW. L. Gibson, Jr., Peanut Acreage Allotments and Farm Land Values, Technical Bulletin 175, Virginia Polytechnic Institute, 1964. a: eke x" a. U0 5‘ Us m , . ._. t o. C) ’1 I F or: ‘ .":3'..‘ ‘uo.~§. 7 11:: so a most 1 aszrazn procuct; V..- ‘9 h .g k. ll‘ 11 21 of the value of the associated land and buildings, was estimated to sell for $565 during 1956-60. Conservation and Land Values Many agricultural conservation practices not only conserve land so as to retain its production potential for the future, but almost immediately increase land productivity. Practices such as drainage, irrigation, and treatments with fertilizers give rapid production responses. Increased productivity of land would be expected to increase the relative price of land. While a number of publications deal with the profitability of carrying out conserva- tion practices, little has been done to quantify the impact of different practices on land values. Since most conservation practices are heavily supported with federal money, the differential gains which occur between individual farmers as well as between various regions become highly important for policy determination. Cotner found that, although the maximum benefits from the conservation program would involve a shift in the allocation of funds, the allocation of funds among states and counties was practically fixed over time. Hathaway found that "a very high proportion of the expenditures [ conservation payments] go for inputs which increase farm outputifin Z3Melvin S. Cotner, "The Impact of the Agricultural Conser- vation Assistance PrOgram in Selected Farms Policy Problem Areas, " (Unpublished manuscript, Michigan State University, 1962), p. 15. ggshort r1 he discus SE 0139111 and 1 sussequentl :cnservatic result will The sec ond raj.- lead to 111 h: :5 the real 1 be due to a CCCU.’ as a 1 In a "We. eutereaeio. O. “ ’22 1 1 the short run and will sustain the increase in the long run. "24 Further he discussed how the subsidized conservation practices may increase output and thereby bring about lower farm commodity prices and subsequently reduce returns to other agricultural inputs. Since the conservation practices are built into the land input, the immediate result will be higher returns to land, at least in physical output. The secondary effects, however, namely lower product prices, may lead to lower returns to all agricultural resources including land. In his study of real wealth gains, Boyne25 suggests that some of the real wealth gains shown in the farm real estate sector might be due to a change in the quality of land. Such quality changes could occur as a result of conservation practices. In a discussion of land values and government programs Rossmiller refers specifically to the Agricultural Conservation Program and states: "Effects of this program[conservation] then to a greater extent than effects of others in this category tend to be directly capitalized into land values. "26 However, he made no attempt to quantify the relationships. Scofield has on several occasions made references to the relation between land values and conservation practices and states: 24Dale E. Hathaway, Government and Agriculture (New York: The Macmillan Co. , 1963) p. 312. 25Boyne, p. 46. 26Rossmiller, p. 88. .\'c- s accc pub} mar The set time i ieSoil Ba influencing- As ships betw :arried on Statistical values. I test. The 23 No systematic attempt has been made to deve10p a capital account for land. . . . that recognizes both private and public investmefis that have become capitalized into market values. The problem of obtaining data for conservation investments over time is also pointed out by Scofield. 28 As mentioned earlier, the Soil Bank conservation programs are also stated as factors influencing land values. As already pointed out, little quantification of the relation- ships between land values and conservation practices has been carried out. Therefore, one of the tasks of this study is to test the statistical relationship between conservation investments and land values. There exists a problem of getting adequate data for such a test. This problem will receive further attention in Chapter IV. 27W. H. Scofield, "Investment in Farm Real Estate," Journal of Farm Economics! " XLV (May, 1963), p. 405. 28mm. , p. 404. STAT :0: seem 1031-5 invc els ewhe r M w ccmpa CHAPTER III STATISTICAL METHODS: EVALUATION AND INTERPRETATION Referring to a distinction made by Pigoul between tool makers and tool users, this thesis is essentially concerned with the use of well-known tools in a particular area of research. Thus, it might not seem necessary to devote much space to a discussion of the tools involved since detailed discussion of these has been carried out elsewhere. However, in order to appraise the results and to be able to compare them with those obtained by other methods it is important to have a clear understanding of the method which is used. Also a discussion of the Los and fl inherent in using the given method is essential for an evaluation of the results. Therefore, the chapter includes a discussion of the problems and benefits arising from using a given set of tools for a specific problem. The chapter includes the following sections: Supply Function for Land, The Model, The Use of a Difference Model of First Order, Features of Time Series Data, Selection of Time Period, and Aggregation Problems. The reader who is familiar with the use of first differences and general statistical problems may prefer to read 1A. C. Pigou, ”The Function of Economic Analysis, " Sidney Ball Lecture, 1929, reprinted in Economic Essays and Addresses, p. 3, quoted by J. Robinson, "The Economics of Imperfect Competition, " (London: MacMillan, 1961) p. 1. 24 cf as an in idenificat simultanet ‘ . 9... -‘ LC. an ax 25 the first two sections and then move to Chapter IV. Supply Function for Land The value of land at a specific point in time can be thought of as an intersection between a demand and a supply curve. The identification problem thus arising would call for the use of a simultaneous equation system. However, supply of land is such that an alternative approach seems justified. In the period considered, the estimated quantity of land2 used in agriculture has changed very slowly. Only in a few states has the quantity of land been halved or doubled during the entire period. In most states the difference between the lowest and highest quantity of land is less than twenty percent. More important than the long-run change is the year to year variation. Since the variation in quantity is distributed over a long time period, in many cases over the entire time period, the yearly variation is indeed very small. It seems reasonable to conclude that, since the supply of land is almost perfectly inelastic at a given point in time and only changes slowly and rather constantly over time, the influence of price of land on supply of land is negligible. Thus, with a constant supply of land, we assume that the change in land value is caused by shifts in the demand for land. This approach lacks some of the appeal of the simultaneous equations approach, but it is considerably easier to handle empirically as well ZSee Appendix A. o—h as 5.4»- 1 .Y . s::::*..-ta: 1:.e, the set: with TI . .r m, Ar ..- . n.‘ .. t} r‘51-01'13 26 as statistically. It is believed that the gains in using state and regional data will be larger than the possible losses from not using simultaneous equations. Given the data, resources and time avail- able, the simultaneous equations approach could not have been carried out with as much disaggregation. The model The capitalization formula is the basis for the model used in this study. (In a perfect knowledge situation the value of a capital stock is equal to the discounted value of the perfectly known income streams: 00 R1 (1) Land Value (LV) = E31 ———i- 1" (1+r.) 1 R1 = Net returns to land in year i. r. = The capitalization rate in year i. 1 For a fixed income stream and a fixed capitalization rate the formula reduces to: (2) LV = htISU This formula is used, assuming that most patterns of income can be expressed in terms of a fixed income stream, and that a steady capitalization rate is applicable for future time periods. The land value (LV) expressed in (2) is nonlinear with respect to the capitalization rate. Instead of entering r as an individual variable, we rearrange (2): (3) The 1m :5 earlables. :ttcerned wit: 1 :coel of first (4) A 11:13 azA t:y 5:1 .here 27 (3) (va r) = R The income stream to land (R) logically depends on a number of variables. Letting these variables be X1, X2, . . . , Xn , and being concerned with the changes in land values, the following difference model of first order was tested: (4) A (rxLV)ts=f(a, AXI, AXZ, ..., Axn)ts+uts u = Disturbance term a = A constant t = Year: 1925-1962 s = l, 2, . . ., n (number of states in a given region). The relationship is assumed to be linear. The variables are described in Chapter IV. Modified model--Using the average farm mortgage rate as a proxy for the capitalization rate and equation (4), ordinary least squares regressions were carried out for a sample of states. The results were not encouraging. Compared with analyses using value of land alone as the dependent variable, the analyses using the farm mortgage rate times value of land as the dependent variable came out short. The coefficients of multiple determination were generally lower, and the significance of the estimated coefficients was not increased. Also the Durbin-Watson statistics generally decreased, indicating serial correlation in disturbances in some cases. This occurred in spite of using first differences. The average farm mortgage rate used fluctuated relatively little, but in a model using first (ii: rather e1. .'.. 28 first differences a fluctuation of . 2 or . 3 of a percent brought about rather large changes in the dependent variable. Given the relatively small variation over time (but relatively big variation from year to year) in the farm mortgage rate, and the lack of estimates concerning the varying risk attached to ownership of land, it was decided to avoid the difficulties of approximating the capitalization rate by using the following model: (5) A(LV)ts = f(a,AX1, AX AX) +u 2’ "" nts ts This model relates the changes in land values directly with the factors assumed to express the expected returns to land and assumes no effect from changing mortgage rates. Time series data are obtained for each of the 48 states in the conterminous U. S. for the period 1925-1962. Using ordinary least square regression, the model is fitted to the time series data for each state. Combined time series and cross-sectional analyses are carried out for regions as defined in fn. 6, p. 3. However, due to dissimilarities among states within regions, some of the regions are subdivided for the combined analyses. Since intraregional heterogeneities are small as compared with interregional heterogeneities, the state time series data are not combined cross-sectionally for the entire nation. The Use of a Difference Model of First Order When variables are expressed as changes from the preceding year, we say that the variables are differences of first order. It ispossfble t0 2&2. actual da enables are is of more in‘ ifierence nit ihmkeduse important whe :ffirst differe 1f l‘WO 1 as the causal wfllty] probab repres hths 1 581'30 “V 1031 29 is possible to work with first differences of some variables, combined with actual data of other variables. However, in this study all the variables are entered as first differences. 3 When year to year change is of more interest than deviation from a long-term average, a first difference model is appr0priate. The pre-World War II average is of limited use for the post-war period. This reason is most important when working with undeflated data. With respect to the use of first differences instead of original values Ezekiel and Fox state: If two or more economic time series are intercorrelated as the result of trends which may not reflect logical or causal relations between them, the use of first differences will typically reduce intercorrelation and increase the probability that the regression coefficients obtained will represent meaningful relationships. In this study some initial analyses indicated that the original values would not yield lOgical results. A model using original values, linear in logarithms, yielded highly significant coefficients of the independent variables, but more than half of the signs were opposite those expected from an economic point of view. The model was also tested using first differences of logarithms, but this functional form did not yield results which were any better than when first differences of original values were used. Also, first differences of logarithms are cumbersome to work with when observations for some of the variables 3The yearly government investments in agricultural conservation are treated as being first differences in themselves. 4M. Ezekiel and K. A. Fox, Methods of Corregtion and Regression Analysis, (New York: John Wiley and Sons, Inc. , Third Edition, 1963), pp. 340-342. are 250' Multicoll is referred to as pra‘slem when ti ispresent the V itis difficult to 1 sends, as time enables to be i rec'uce multicoll substantially. 5 'ezils" of multic technological ch ccllinearity bec conservation pat ithas been kept I Such as output p been deleted 01- 312.6 analyses, 3:18! and with I: deleted becalISe nrrelated Varia 30 are zero. MulficollineariLy- -‘Corre1ation among the independent variables is referred to as multicollinearity. Multicollinearity is a general problem when time series are considered. When multicollinearity is present the variances of the estimated coefficients increase, and it is difficult to get significant coefficients. In data showing pronounced trends, as time series frequently do, we would expect the independent variables to be intercorrelated. Using first differences will generally reduce multicollinearity; in this study the multicollinearity was reduced substantially. 5 However, use of first differences does not cure all the "evils” of multicollinearity. Some of the variables used to express technological changes show similar second order trends, and multi- collinearity becomes a problem even in first differences. Government conservation payments is one variable affected by this problem, but it has been kept in the analyses consistently. However, other variables such as output per acre, fertilizer use and output per man-hour have been deleted or included in the analyses according to their significance in the analyses, and according to their intercorrelation with each other and with the conservation payments. When some variables are deleted because of intercorrelation, the estimated coefficient for a correlated variable will be biased. The estimated coefficient for a 5A rather typical example of the reduced intercorrelation between independent variables is the following found in Minnesota data. The simple correlation between the index of price expectations and the index of output per man-hour was . 70. Using first differences the correlation fell to -. 03. given Variable part Of the effe 535E653 that t} and otiier varia shat biased. using first diffe effects of time. change, even if The constant me rate over time. differences (Whe sad their combir Erosion losses, variable of this l Serial co fine series are I. the unexplaine among successii Eincerned with t new: - - . ~a.o.lit1es for 31 given variable will include, beside what it is designed to estimate, part of the effects due to a deleted but correlated variable. This suggests that the estimated coefficients for conservation practices and other variables reflecting technological advances may be some- what biased. Constant term--If a constant term is included in a model using first differences, the constant term will represent the linear effects of time. This implies that the dependent variable would change, even if there were no change in the independent variables. The constant measures the effect of variables changing at a constant rate over time. Such variables cannot be included when using first differences (when using actual data, one such variable can be included), and their combined effect will be measured by the constant term. Erosion losses, for which linear estimates are available, is one variable of this kind. Serial correlation--Another problem which often apears when time series are used is that of serial correlation. Serial correlation in the unexplained disturbances violates the assumption of independence among successive disturbances. A study by Cochran and Orcutt 6 concerned with this problem was published in 1949. They discuss possibilities for avoiding serial correlation by changing some of the 6D. Cochran and G. H. Orcutt, ”Application of Least Square R egression to Relationships Containing Autocorrelated Error Terms, " Jgurnal of the American Statistical Associati_o_n, 44 (March, 1949). pp. 31-61. fsized to be est usir. iI‘ Li. (1:) (IV) If we term then' will t steps This Shoes using 30 amid the impened sc 16l'recalm assumptiOn r 533d incete: E5 estl'l'nates Elatisu'cal n the Period regressive l 32 variables, adding additional variables, or modifying the form of the relationship. However, this is not a way out if one has arrived at a form of relationship and a selection of variables which are judged to be a reasonable choice. Where this is the case, they suggest using first differences: If we prove to be right about the nature of most error terms in current formulations of economic relations, then the residuals of the first difference transformation will turn out to be sufficiently random and no further steps will be necessary. This paper led to a large increase in the percent of statistical studies using first differences. The joy of having found an easy way to avoid the problem involved in serially correlated disturbances was dampened somewhat by the findings of Hildreth and Lu in 1960. 8 They recalculated seventeen time series equations in which the assumption of independence in disturbance terms was rejected or found indeterminate by the Durbin-Watson test. The estimated coefficients of original values and of first differences were compared to estimates found by the maximum likelihood method using a statistical model in which the random disturbances in successive time period were assumed to be generated by a first order auto- regressive process. In cases where the estimated autocorrelation 7Ibid, p. 54. 8C. Hildreth and J. Y. Lu, Demand Relations with Auto- correlated Disturbances, Technical Bulletin 276, (November, 1960) Michigan State University, Agricultural Experiment Station, East Lansing, Michigan. coefficients Zkeli'nood If than that be‘ was in agre. autocorrela' m estima' as a routine clearly dam diferences istirbance “trend e‘f Net‘efihdes hj~l3li63is ‘ oasis-Wat correlation SinC favor of 115i m the distu :easibilit . C Cerrelation. \ I 52’ / 33 coefficients were negative, the agreement between the maximum likelihood method estimates and original estimates was closer than that between maximum likelihood and first differences. This was in agreement with expectations. However, in cases of positive autocorrelation the performance of first differences, as compared with estimated coefficients by maximum likelihood, could not be judged better than that of the original analysis. The conclusion was: “To use trend terms, first differences, or lagged variables as a routine precaution against autocorrelation [ italics mine] is clearly dangerous. "9 As pointed out earlier in this chapter, first differences are not used as a precaution against autocorrelated disturbances but (1) as a remedy for illogical relationships caused by trend effects and (2) as a means to decrease multicollinearity. Nevertheless, in the test analyses carried out on original data, the hypothesis of independent disturbances could not be accepted. Durbin-Watson statistics on the residuals indicated positive serial correlation in the disturbances. Since other reasons, as stated above, led us to decide in favor of using first differences, the indication of serial correlation in the disturbances did not cause us to investigate (empirically) the feasibility of other methods used to avoid the problem of serial correlation. Such other methods as the iterative procedure discussed Ibid, p. 43. rmfldvery l satlsfactory fieprocedut “as size of s correlatnin correlation senalcorre ofhnear tre tierences, 3.3%” of One EIIDrs in ill EIIOIS. If SweperlO< 3“r'ers of 1a TEE}; ’35 to 1 10 53118001, C 11H 34 by Johnston10 or the "crude" method suggested by Hildreth and Lu11 would very likely solve the problem of serial correlation in a more satisfactory manner than do first differences. The variation between the procedures is due to different assumptions about or estimates of the size of serial correlation. For the sake of correcting serial correlation the first difference approach assumes original serial correlation of l. 0, while other methods estimate or assume the serial correlation to be between -1. O and l. 0. Thus, the effects of linear trends in the variables will be best handled by the first differences, since the first differences eliminate at least as large a part of the linear trend as other methods. One of the reasons for serially correlated disturbances, when the original data were used in this study, could be the problem of errors in the data. There is little doubt that the dependent variable (land values) in this study is particularly subject to estimation errors. If the land value assessors tend to be overly Optimistic in some periods and too pessimistic in other periods, this could lead to serially correlated disturbances. The same would happen if the buyers of land followed cycles of Optimism and pessimism about returns to land. The problem of timing is also of importance for 103'. Johnston, Econometric Methods, (New York: McGraw- Hill Book Company, Inc., 1933) pp. 193-194. 11Hildreth and Lu, pp. 13-14. ..:.. E $.‘ O 6111 L... fin -.A vAA Y'fif‘: ARV, ...L $5.. 6 CC. . ‘ P. ..- P” " Lu. \J. 3.- ...at a... 35 serial correlation. Land values are estimated March first of each year. Other variables are not measured on this date, but may be lagged some months relative to land values. If perfect timing is required in the relationship among the variables, the lag introduced, when the explanatory variables are measured 0 - 6 months ahead of or behind the dependent variables, could cause serial correlation in the disturbances. Even though the first difference approach might not be the best tool to solve the problem of serially correlated disturbances, the other reasons stated for using this approach appear to be important enough to justify its use in this study. Multilale correlation coefficient--It is generally observed that correlation analyses utilizing first differences tend to have a lower multiple correlation coefficient than when original data of the same variables are used in the analyses. A main reason for this is that first differences eliminate correlation due to time trends of first order. Cochran and Orcutt give the multiple correlation coefficients, using original data and first differences respectively, of eleven demand studies from the time period 1920-38.12 Only in one case 12Cochran and Orcutt, p. 55. The analyses were carried out by Richard Stone, who published the analysis using original data in: "The Analysis of Market Demand, " Journal of the Royal Statistical Society, 108 (1945),. pp. 286-391. 71* of e 3" [062115] :v‘m a .. 'e in The tir -. n. P<§Q CHETE‘ . a ma. 36 out of eleven did the first differences give a higher multiple correlation coefficient. In the other cases the multiple correlation coefficient fell from a few percentage points in some cases up to more than twenty percentage points in other cases. Other examples are provided by Foote, who shows the simple correlation between prices of cotton seed oil and prices of butter, lard, and other fats and oils respectively.1 The time period was 1922-40. Changing from actual data to first differences decreased the simple correlation by at least .17, and as much as . 49 in one case. When deflated data were used the simple correlation between actual values fell, while it increased when first differences were used, thus leaving less difference in the correlation coefficients between the two methods. However, the estimates using actual data still showed the highest correlation coefficients. Almost no literature is available to explain why a conversion to first differences should yield a lower multiple (correlation coefficient. However, since a main reason for using first differences .is to eliminate spurious correlation due to long-run trends, it seems obvious that this must be one reason for the reduced correlation 13Richard F. Foote, Analytical Tools for Studying Demand and Price Structures, Agricultural Handbook No. 146. ERS, U. S. Department of Agriculture (Washington: U. S. Government Printing Office, 1958), p. 34. 1: coefficient. field a highe correlation i serial corre trends in the correlation correlation series are ccrrelatior 37 coefficient. 14 Heifner found that first difference analysis would yield a higher multiple correlation coefficient only if the serial correlation in the disturbances exceeds a weighted sum of the serial correlations in the independent variables. 15 With strong trends in the time series, it seems most likely that serial correlation in the independent variables would exceed the serial correlation in the disturbances. Serial correlations in time series are often of the magnitude . 8 to . 9, while serial 16 correlations of disturbances are in most cases below . 5. l4Let Yt be the dependent variable in year t and Y‘ = ~ Y - Y. Also ut values the squared multiple correlation coefficient (Ri) is 2 (YE2 - {Tb/Z Yéz. For first differences the squared multiple = unexplained residual in year t. For original . . . Z . , , 2 ~ .... Z correlation coeff1c1ent (RB) 18 Z[ (Yt - Yt-l) - (ut - ut-l) ]/ E (Y; - Y't_1)z. Now, 2 (Y‘t - Yi-l)2 < Z? sz if the autocorrelation in Y exceeds . 5. Zfii't - flit-NZ = 2313: , given an autocorrelation of . 5 in residuals. Thus, assuming the autocorrelation in Y to exceed . 5 (which is most likely in a time series), and the auto- correlation in residuals to be less than or equal to . 5, the R; < R:. 15Richard Heifner, "A Note on the Relationship Between Coefficients of Determination for Regressions Computed in First Differences and in Original Values, " (Unpublished paper, Michigan State University, 1965). 16Hildreth and Lu, p. 17. Out of 17 estimations of serial correlation in disturbances only 3 exceeded . 5. The lower degr original ya proud feeli is success; sign, a re: Features 0 A n thesis will involved in chapter, a; Work with j Whe Universe fr Ol’er the Se 30 the first Zead t0 aVOi Ol'er time, t5 remain C which haVe is donbtful 1 land has inc deve10p e d. :a‘ 301" has 1'. 38 The indications are, that first difference analysis yields a lower degree of multiple determination that would be found by using original values. However, we shall be happy to give up some of the proud feelings with which one can point to high Rz's, if the approach is successful in giving coefficients with an economically meaningful sign, a reasonable degree of significance and good predictive power. Features of Time Series Data A major part of the statistical analyses carried out in this thesis will be based on time series data. Many of the problems involved in using time series have been pointed out earlier in this chapter, and it has been explained why these problems caused us to work with first differences of time series data. When we use time series, we make the assumption that the universe from which the observations are drawn remains stable over the selected time period. This assumption is carried over to the first differences of time series. The first differences do lead to avoidance of the problems caused by first order trends over time, but we do assume the relationship between the variables to remain constant over the selected time period. With the changes which have taken place in agriculture during the last forty years, it is doubtful that this assumption is realistic. The total output of land has increased rapidly. Many substitutes for land have been developed. The productivity of complementary factors, such as labor, has increased. Is it reasonable to assume that the relation- gfipbenxeel orice expec‘: I they were it have becorn and it does price expec the questior time, but it assumption 561ection oj \ It h; from data c inagivent Stud}- annuE iiscuSSed l QBSer.'.atiOI Stat CEOlce of ti his desira ES‘ . , . sumpthr-l 39 ship between, for instance, land values and agricultural product price expectations has remained constant over this time period? Obviously, changes in price expectations have not been nearly as closely related to changes in land values during the last decade as they were in the thirties. However, this is because other factors have become more prominent in the determination of land values, and it does not mean that the relati.0nsh'ip between land values and price expectations have changed. No firm answer can be given to the question about the assumption concerning stability through time, but it is believed that the results will indicate that the assumption is not too unreasonable. Selection of Time Period It has been mentioned earlier that the estimates are obtained from data covering the period 1925-1962. The number of observations in a given time period depends on the choice of time unit. In this study annual data are considered to be satisfactory. However, as discussed in the next section, the influence of using biennial observations is tested in the combined analyses. Statistical considerations are conflicting with respect to choice of time period. In order to obtain a "good statistical fit", it is desirable to have a large number of observations. The assumption of a stable structural relationship during the time period and the problem of getting reliable and consistent observations call for a short time period. The tlm silt-e parent P were Carried O the earliest Ye reasons for thi years, several interpolations ante periods a er‘vlaining the since other sta the period 192 The sel :onditions suc War and vario right have affe econometriciar structural rela war years are Wars are large T . .here 18 no si Certain time p .‘1 wbably no m \ 17 98’10d193tif1‘ 40 The time period of this study is related to the time period of the parent project, namely 1917-62. 17 Some initial analyses were carried out for this period, but the observations for a few of the earliest years did not give a good fit. There could be several reasons for this, but a major one is probably that, for the earlier years, several of the independent variables had to be obtained by interpolation. Other evidence suggests that the data of earlier time periods are less reliable. Since we are more interested in explaining the more recent deveIOpment in the land market, and since other stated considerations call for a shorter span of years, the period 1925-62 was decided upon. The selected period includes a variety of general economic conditions such as the depression period, World War II, the Korean War and various business cycles. The different economic conditions might have affected the structural relationships. Indeed, many econometricians have omitted war years on the ground that the structural relationship was changed for these years. However, if war years are omitted, we must ask if the structural changes during wars are larger than those occurring from the business cycles ? There is no simple argument by which we can decide to omit certain time periods whether these are well defined or not. It is probably no more heroic to assume that the structure remained 7See Chapter I, p. 2. Rossmiller's study uses the time period 1930-1962. ;:ts:ant dui scarred in finally up u able from v is be incluc Aggregatio 80:: a simplifie iata gener. means that often-ids: On. the state 1 less than I on a State the aggre ECOaniC between 1- min dual 41 constant during the war periods than it is to assume that no change occurred in the structure from 1925 to 1962. The question of whether or not to include war years is finally up to the researcher's discretion. The information obtain- able from war years is, in this study, considered important enough to be included. Aggregation Problems Some degree of aggregation is used in any model attempting a simplified representation of the real world. Aggregate economic data generally involve an averaging of heterogeneous data. This means that the simplification sought in a model involves an averaging of non-identical relationships. One dimension of aggregation is Space. The use of data at the state level means that the space aggregation in this study is less than that of other studies using national data. However, data on a state level represents a large degree of space aggregation. The aggregation is justified since we are interested in the micro economic effects. We are interested in the differential effects between regions, but we are little concerned with effects on individual farms. The use of one year as the time unit means that the obser- vations are aggregated in time. It also means that seasonal variation cannot be accounted for. The prit aver commodit: with the weig‘: index, changes should not give states the aggr : mmodity mi; related to land Extens: mestimation : piblished in l 5’ relationships, and showed tl: assumption of "in practice". Grundi te specify mic Empiric ally t} 18 _l l.‘lmSterdar;1: l 9Y e . ~essar11y Drual'y’ 1; p ”EC (i, 42 The price expectation indices used in this study are aggregated over commodities. The enterprise combination within a state, for which the weights are obtained for construction of an aggregate state index, changes little from year to year. Thus, commodity aggregation should not give serious problems in state analyses. However, among states the aggregate indices could be based on quite different commodity mixes and the coefficients of price expectations as related to land values could vary substantially. Extensive attention to the problem of aggregation of data used in estimation of economic relationships was given by Theil in a work published in 1954. 18 In his comparisons of Inicro and macro economic relationships, he assumed the micro equations to be perfectly specified and showed the errors which could occur from linear aggregation. The assumption of perfectly Specified micro equations is hardly applicable "in practice". Grundfeld and Griliches argued that in general it is not possible to specify micro equations perfectly because of lack of knowledge. 19 Empirically they proved that aggregation would not necessarily produce an error, but there might be an aggregation gain. They concluded that: 18H. Theil, Linear Aggregation of Economic RelationshipsJ (Amsterdam: North Holland Publishing Company, 1954). 19Yeheda Grundfeld and Zvi Griliches, "Is Aggregation Necessarily Bad?",, The Review of Economics and Statistics, 42 (February, 1960), p. l. 1. It L his ve: eQ’ 2. CC as. x be” mi is They0. ponmmaxo on no: @150 pm. a GM :3 aqua-woman; m than “53°20.“ .34.?“ new votes can 2: «ESE .Sh m .mo .u usgmnsjsa .Nw . "mmmnooO .oo J "MGSOHMU fisomaumoomhm sure—30h no woman— ohm JOE? magmas? mafia-030w may mafia: vosmpnmvnmum one Saudi-«five onus mush—use 05 nootfimnd wonmnaou 23 an h . mmo mbms< “0233500 nun—3m 33.383 N oVUUQHUNQ mm MUMHOVW m we . _ S .N ms . a No . _ nunnsfim nofin?-n3§q 5. ms. 2.. we. 3.. we. Nm 3m e SN .2 :n .v 3v . an .v :e e on . No . Z .N on . ow .- we .- ransom nonunion :5 .V 3o 4 SN e as e E e 8 _ e om . a. . we . 2 . om . S . 28¢ 8a 3&5 so .V as .V - Co e A: .V 8o .V we . co . .. 8 . mo . wc . o>uomom owmoho< :3 .V A? .2 E. .t as e E. 4 gm e 3 .N 3 .N i. .3 E .m a ._ co . n 383m 533335 a: a: $2: :3 8m; am; no . so .2 8 .N 3 . cm .2 am 4 managedaxm nondidnnoo so .V as e an e 3o e as e :b e f . om . so . mm . 3 . a. a . 3338st dumps AS ..V an .v an . 3 3e 4 3m e tom O Na .7 g .- so .m- on .- R. .- co .- «sensuous .SGO NGSOHmU muse? sou-H undo? :4 nmvmnoah mamnmfiafi mmwuooo gun-om ofifimmudxr Ncummod .msmemoumom 03m> wand mo mud-mom Hummus—«50m .N sun—sh. 79 value of a pasture acre equivalent by Ella. 49. The coefficients Obtained in the other analyses of the Southeast and in all other regions are interpreted in a similar manner. However, in the analyses using biennial Observations, the constant term measures not the annual, but the biennial change in the pasture acre equivalent value. The structural relationship based on Florida data differs from that in the rest of the area. Florida was, therefore, not included in the combined cross-sectional and time series analysis for the region. When all years were included in the regional analysis, the coefficient of multiple determination fell below the level for individual states. This indicates some structural differences between the states. Aggregation of the first differences over two year periods raised the multiple determination coefficient by . 14 indicating consider- able random variation in the variables. As pointed out earlier, most Of the decrease in random variation is probably related to random variation in the dependent variable. Value and quantity of land--At the outset of this thesis it was noted that some of the largest increases in land values had occurred in the Southeastern region of the U. S. This also holds when we express the quantity of land in terms of pasture acre equivalents. Florida had a much larger absolute change in land values over the period 1925-62 than did the other states in the region. Georgia had the largest percentage increase due to a low initial unit value. In Georgia the area of cr0pland has decreased rapidly and brought about a decrease in the quantitative measure Of land, while in Florida the quantity of land has more than doubled from 1925 to 1962. South Carolina and 80 Alabama had, like Georgia, decreases in the number of pasture acre equivalents (30-40% between the endpoints of the period). Constant term--The constant was negative and significant at the . 05 level in all the analyses except for Alabama. Price ex1Lectations--The indices of price expectations for the Southeast (excluding Florida) increased somewhat more than in most other regions. Cotton production is of considerable importance for South Carolina, Georgia, and Alabama, but cotton price expectations increased by about the same amount as the aggregate national index. The tobacco price expectation index, which increased about three times as fast as the aggregate national index, has influenced the indices for South Carolina and Georgia substantially (see Appendix D). In South Carolina, where the price expectation index increased most, tobacco accounts for about twenty percent of total farm sales. For Florida, where oranges account for about forty percent of farm income, the long run changes in the state index of price expectations were much smaller than in the other states. The sizes of the coefficients indicate that cotton prices have more influence on land values than do tobacco prices. Other regions will provide further evidence on this point. Except for Florida the coefficients were significant at the . 05 level or better. Conservation expenditures--With about six to seven percent of farm land in the U. S. , the Southeast received eight to nine percent of the conservation payments over the 1936-61 period. 81 The intercorrelation between output per man-hour and conser- vation payments was high (about . 8) in the Southeast, and the high coefficients for conservation expenditures are likely to be measuring some influence of technological changes as well as the impact of conservation expenditures. The index of output per man-hour in the cotton industry has increased more than most of the other eleven enterprise groups from which the aggregate indices Of output per man-hour were computed, and the impact Of mechanization in the cotton industry would to some extent be measured by the coefficients estimated for conservation expenditures. The conservation expen- ditures explained more of the land value variations than did the output per man-hour indices, which were not retained in the analyses. The differences in size of the estimated conservation expen- ditures coefficients could be due to differences in the kind of conser- vation carried out. The conservation data suggests that the practices carried out in South Carolina and to a lesser extent in Georgia are of a more enduring kind than those carried out in Alabama. In all three states a major part of the conservation payments are Spent on measures to establish permanent cover. In addition, in South Carblina substantial amounts are spent on drainage and in Alabama a substantial part of the conservation measures are for temporary protective cover. Conservation reserve--The Southeast has received a substantial part of the payments made for contracted acreages under the conservation reserve program (in 1960 it received 8. 2% of U. 5. total). Georgia and South Carolina received most while Florida received relatively little. 82 Compared with the national average rental rates per acre under contract, the rates for the Southeast (with the exception Of Florida rates) have increased faster Over time. 6 This has led to a relative increase in the acreage contracted for the conservation reserve program in the Southeast. In 1960, the year with the highest national expenditures on the conservation reserve program, the payments to the Southeast amounted to 26. 7 million dollars. The coefficients show that the conservation reserve program has a large influence on land values. For example, using the estimated coefficient (2. 2) for the combined analysis would indicate that from 1955 to 1960 the cons ervation reserve program has caused an increase in land values in the Southeast amounting to 583 million dollars, or about 25 percent 0f the total increase in that period. The Florida coefficient is out of line with other estimated coefficients. This could be due to Spurious correlation associated with the relatively small area in the conser- vation reserve. AcregggreserveuThe acreage reserve programs had little influence on land values in the Southeast. This is not because of small Participation in this program. In the last year of the program 101 million dollars, or 15% ot total expenditures under this program, were Spent in the Southeast. Most of these payments (85%) were Spent on decreasing the active cotton allotments. The insignificant coefficients 6U. 8. Dept. of Agriculture, Conservation Reserve Program and WUSeAdjustment Programs, Statistical Summary 1963, Agricultural abllization and Conservation Service, Washington, D. C. , p. 14. 83 could indicate that the acreage reserve payments on cotton allotments did not increase returns to land, at least not in the Southeast. Output per acre--The output per acre variable was significant at the . 05 level or better, indicating rather homogeneous crOp production conditions among the states in the region. As discussed in the previous chapter, 7 it also suggests little change in the ratio of cropland to other land. Population densityuThe population density coefficient for Florida suggests that a large part of the extremely large increase in land value is due to increased pOpulation. The population density in Florida increased from 23 persons per square mile in 1925 to 101 in 1962, indicating an increase of more than 150 dollars in value per pasture acre equivalent due to pOpulation increases. This is probably an over estimation of the impact of pOpulation density, since the variable most likely measures some Of the impact due to technological advance. The intercorrelation between population density and output per man- hour was . 5. Also, the very high negative constant could be associated with overestimation of other rather constant variables. Nevertheless, there is little doubt that influx of pe0ple to Florida has had a large impact on land values. Appalachians The statistical results for the Appalachians are given in Table 3. A good fit is indicated, and the D. W. statistics give no indication of 7Supra, p. 63. .momoaunonom cw sou-“w one magnet-op whopsmumm .wm .uoommonnoh. was .ho .fi KMoBGoM 69 J "msfiouoo aunoz .mh . newswmnfiw «no? .HN .d "owswwnwkrnumoozm o>32ou no woman. one JOE? 84 muamwose mafia-030m ofi magma nonfloumwsfim one musofiourgwo onus ougmom 9.3 .momufidnm Moo-£9800 on? :H H mm. 4 3 .N ew .H S .N do . a nonnndnm noundkéfina 3. me. 3. S. E. 2.. 3. mm 8m; 2.: am; E: as .V 8M; 3.: mm . 1.. . M: . om . R .. No . me . Sands perineum :3 e so 4 E. e am e as .V am .V 3.: mm . S . om . mm . 2 . E . 2. . Suzanne 2: a so e A: e A2 4 3H 4 so 4 an e E . mm . m H . mm . me . 2 . mm . 98.4. use 3350 a: 8: am; See a: STE 5.; mm. mm. em. ms. mm. mums... mm. uiuuum pudenda. 8.: 3... .v as: 2%.: 33.3 8a.: 23.8 m~.~ 3d and 22m 22m 3..- mu; 383m sore/$200 a: 2: .2 :Ne 3s .v SN .v a: as; ndufineqdaxm no . md . om. S . am .2 2. . aw ._ 8398300 so; as; ace 2: .v A: .2 so; 8: S . mm . em . R . mo . mm . 2 . unonanddmxm durum Sp .V :m e :n .v as .V 3w .v a: 1:; 3.? on .7 3.7 3.7 S .7 no.7 3.7 383 2388 3:0 msSoumO 35m§> endow. no>m muoow :4 oomossoH 532.com finoZ «no? oficmmnfir o33nm> Hmomnfimsd- 6939.500 monoum E56335: "HR monmmom .msowmmonwom o3m> mucosa mo muadmom "msoEooHsmmasa .m 3an 85’ serially correlated disturbances. The estimated coefficients suggest some heterogeneity among the states'and this is in agreement with the fall in the coefficient of multiple determination when all years are included in the combined analyses. The significance of the estimated coefficients increased considerably in the combined analyses. Time aggregation increased the coefficient of multiple determination and generally decreased the significance of the coefficients. Time aggregation is likely to increase multicollinearity, and the rather large changes in the price expectations and the fertilizer coefficients indicate intercorrelation. Value and quantity of 1and--The percentage increase in land values in the Appalachians during the 1925-62 period was/much lower than the increase in the Southeast, but it was still somewhat above the U. 5. average. The Virginias had small relative increases, especially West Virginia which was far below the U. S. average. North Carolina had the largest increases relatively as well as absolute. The quantity of land, as measured in pasture acre equivalents, declined in all states. The decline in West Virginia was almost 50 percent, while other states showed 19-30 percent decreases. Constant term-~The constant terms were negative and significant at the . 05 level throughout. The absolute value of the terms tend to vary among states in agreement with the variation in the value of pasture acre equivalents. Thus, it would appear that as expected, the total annual depreciation of land is related to the value of land. Price expectations--Increases in the indices of price expectations during the 1925-62 period have been considerably higher in the Appalachians 86 than in the country as a whole. The largest increase occurred in North Carolina where the average percentage increase has been twice that of the entire country. The rapid increase in North Carolina was due to the large influence of tobacco prices. West Virginia, where dairy is the most important enterprise, had the lowest increase in the index of price expectations for the region, but it was still appre- ciably above the U. S. average. While the coefficients for West Virginia, Kentucky, Tennessee and the combined analyses are significant at the . 01 level, the coefficients for Virginia and North Carolina are not significant and are considerably lower than those for the rest of the region. In both North Carolina and Kentucky tobacco is the main enterprise, but Kentucky produces burley tobacco while North Carolina produces flue-cured tobacco. Conservation expenditures--The Appalachians receive a relatively high proportion of the total government conservation subsidies. In the period 1936-61 the Appalachian payments amounted to about 13. 0 percent of all payments made in this period. 9 The conservation investments were primarily spent on establishing cover and cover protection measures which includes fertilizing and and liming. In North Carolina a large part of the subsidies were 8Virginia also produces flue-cured tobacco, but the impact of tobacco is much less than in North Carolina. c)Farm land in the Appalachians constitute about 6. 5% of total U. S. farm land. \‘ww. :.- 87 spent on drainage. The coefficients were significant at the . 01 level. Conservation reserve--The soil-bank variables were not significant in the Virginias, but they were important in other states of the region. Participation in the conservation reserve program is relatively low in the Virginias and North Carolina, but above average in Kentucky and Tennessee. Participation was low for the entire region until 1959 when the rates increased for some states up to almost 100 percent while the national average rate increased 30 percent only. Acreage reserve--The acreage reserve program had much participation in North Carolina and little in the Virginias. In North Carolina it mainly decreased the utilization of cotton and tobacco allotments, while in Kentucky it mainly influenced corn production. Fertilizer and output per acre--In the Appalacians both output per acre and fertilizer were included in the analyses since, for some states, they both had significant influence. Generally the two variables did not show high intercorrelation, indicating that the output per acre is influenced by many other factors than those approximated by the fertilizer variable. Weed and insect killing chemicals would be important factors in increasing output per acre. Population and income--Population density did not have any significant influence on a single state basis, but was significant when cross-sectional variation was introdiced. Income was not retained in the analysis due to its high intercorrelation with price expectations. 88 The relatively low increases in land values in West Virginia are related to generally lower price increases on farm products, below average conservation expenditures, little participation in the conservation reserve program, and small effects of variables related to technological changes. Northeast Due to much heterogeneity among states, the Northeastern region was split up in three sub-regions10 according to statistical similarities such as multicollinearity and significance of coefficients. The split turned out to be reasonable also from a geographical point of view. Tables 4, 5, and 6 give the statistical analyses of the three Northeastern sub-regions. The picture given from these tables is not too encouraging. As anticipated earlier, the first differences model did not perform well in the small northeastern states. The larger states and Delaware in Northeast 1 gave a good fit and reason- able significance of the variables. The results for the small states in Northeast 2 and 3 show a poor fit and little significance of the variables. The D. W. statistics for the smallstate‘s give in some cases reason to assume positively autocorrelated disturbances, and 10The three sub-regions are: Northeast l-—New York, Pennsylvania, Delaware and Maryland. Northeast 2--Massachusetts, Rhode Island, Connecticut, and New Jersey. Northeast 3--Maine, New Hampshire, and Vermont. 89 .momonfisohmm an So>wm ohm mfiofimgop pudendum N no 4 "seminars was .oe .H nousBmHoQ .oo 4 "mason-Hamssonm .NN. . 320.». BoZIumooEm o>EoHou no woman. one so??? musmflok- mug/030m oau mnwmfi posmpummonfim one muGonZd—uo ohom ohgmom ofi momafimns poGEEOo on? a: we .N we .fi ho .N mo .N moflmfioum acmumguswnudfl on. mo. 5. he. 2.. Nb. mm $0.9 $0.3 Sui 52$ A833 Agra; mm .m. am .n om . 7 om .m Nw .Nu om .fi oEOoGH Hmsomaom so .V :5 .v a: as; See ad; 3 . mo . mo .. 2 .- mo .. so . ransom moraines 2:; 8o; :2; ANN; 31v 2:; on . mm . mm. mm . em . a. . 33-5.2 8a 33:0 :3 Am: 3: ANN; as; can o“. mum. mm; om. me. we. o>aomom ommononfi AM: is 5%: 3a.: an; 5: mm . Nb . am .w me .v ww . Hm J o>somom sofiournomsou a: .v A: 4 :me a: new; as. .v nonfienudxm am. Ne. we. mum. mo. mm. con—okraomsoo so .v 3: 8o; 8: so; as; a. . 3 . f . em . om . 3 . 20338me SE as .v Sm; 8: E; 5.; Name mm . 7 Nb ... we . 7 mo . 7 C. .u hm .... AoEEV unsumsou 3:0 mask-3m endow sou-H memo? :4 Oceanus” ondBoHoQ ussonm v30? 3oz ozownmsr moon» on o5 So no m on 5- ~ H 4 p .O. O 9. «M H p. .94 NOImN®~ .mQmemOHMOm 05Hd> VEGA HO mudsmom u." umMOSHHOz 0* OHQMH. lll lll 90 .uouoapnondm an Go>nw ond unonudfiwop pndpndum N .oo 4 unwounoh #62 end .oH .2 unvofloofinoo .PO .2 "deHuH opoam .wo . “unnouHEOduude/Huuuoonnm o>3d~on so poudn. ond 2333 3:395 wank-020m on? menus pounpndpndnu ond upsofidzfivo onod onfiuudm on» uouufidad ponnaaoo on» 2.2H 2 on .N 3 . em .2 mm .2 nunusfiu douudaéfina om. S. 3.. cm. mm. mm. mm 23 .8 3m .3 $2 .022 an .222 32 .32 $2. .222 S .e um .... mm .3 2n .9. on .e S .u- quuS 382$ :22 .2 so .2 $22 .2 a2 .2 82 .2 so .2 22. mo . om . 2222. no .. no .- bands persuade 8.2.2 32.2 23.2 :m.2 3.12 am; 8. n2. 2:. .2: mm. 2.- ends. 3.2 5390 A322 3m.22 A322 Anni 22932 Sun: me . mm . 3. .. S .2: we .3... mm .2 diuudm sessionnoo as: am .2 22d .2 852 2312 El $332.5me 3. .2 312 E . ~w.~ 22.2 5.2 dosnidudoo 32.2 822 .2 $2 .2 sum .2 SN .2 82 .2 2. 2 . mm . d2 . S . oe . mm . usosfiumumxm 82.5 3.2.22 8.2.2 $222 36.: 8a.: N252 me .~. N2. .- mm .- 2s .- em . om .- $82: 2328 nSGO >ounoh «.903 pndHuH uuuousgo undo? Gourm— undo? S< 3oz ..ooEHOO opoam ..duudz oandnnd> 2 u o ubdflxw posnnflno D H nduduu 2333252 ili Lllllllllll NQImN®~ .mGmemmhwom 05Hfl> UGNWH HO muflfimom "N umdoguhoz om QHQNH. 91 .mmmmficmhmm 222 Go>2w 0.2m 222202233022 wnwwzmum N do . 222208.232. 22am .oo .2 “magmagmm 52672 .hm .2 afimzunmmofium 9523292 220 vmmmn «am 22033 muawEB wank/0220.2 @222— wc2m5 anfluhmvgum mum munm2m>2d2uo whom mhgmmm $2.22 m0m>2m222w 252222222200 @2222 N22 2 2o .2 .222 . 2.2 .2 8232232.. nofimBéBSQ mm. A22.. 20. 222.. om. Nm 22.2.2 22.N.N2 3.2.N2 2222.22 €2.22 E .2 2.2 .- 22 . cN . N2 .m- 2:822 2982mm SN .2 22.2. $2 .2 EN .2 2mm .2 mm . 2.2. N2. N0 . Ne . 52...qu 2202322292 22222 .2 so .2 so .2 N 2 .2 22 .2 N2. oo . 8. oo . N2 . 98... 222 2222220 2N2.2 2mm.2 2mo.2 2N2.N2 2No.2 2o .2 2n . pm .2 No .I mo . o>nomwm Go2wm>u®mnoo 2: .2 222 .2 $2 .2 SN .2 SN .2 m2. . Fm . vM . mp . No . mmhgwwcomxm Go2um>nm mac 0 22.22 .2 No .2 No .2 3.22 .2 2S .2 mo . No. 22. N22 . 2.2 . 302238me 825 Cm .2 32.2 :2 .2 2S .2 2N2. .2 cm .I pm .1 cm .I >2. ... 02. .n 2082225 “22.32200 .225 mnmow 220>H 92di Sam 2:08.32/ mHEmQEmE 32672 $222.34 o2nm2nm> 2 m 0 94.92.22. 2352,2220 D 33% 232.5252 Noummm; .mno2mmmnwmm 2223/ 2:812 20 mu2dmom 2m ammosuhoz .0 m2an m 92 in a single case (New Jersey) there is indication of negatively correlated disturbances. The high D. W. statistics for New Jersey is due to two consecutive large errors with opposite signs. Increased time aggregation alleviates such errors. The combined analyses indicate that there is no more homoge- neity among states in the small northeastern regions than there is in other regions. The use of cross-sectional analysis did increase the significance of the variables, but time aggregation did not increase the multiple correlation coefficient more in the Northeast than it did in regions with a better statistical fit. Therefore, the random variation in the dependent variable in the Northeastern states does not seem to be much different from that in many other states. The poor statistical fit could be caused by cyclical variation in estimates 01" land value or it could be due to exclusion of important variables. Value and quantity of land--The larger Northeastern states gene rally showed a little less than average relative increase in land Values during the 1925-62 period. The northeastern part of the North- east Region had the lowest land value increase in the region, while the sElitnaller states around New York City had relative increases which were somewhat above the U. S. average. New York state was substan- t’ lally below the national average. The decrease in number of pasture acre equivalents ranged fr cm 12 percent in Delaware to 62 percent in New Hampshire. C ompared with other regions, the decreases were large. Among the 31113 ‘- regions, Northeast 1 had the smallest decrease in quantity of 1a 116- input. The rate of decrease has increased toward the end of i?— ‘V'; r‘” .. l" 93 the period. The USDA indices of average value per acre for the North- east give substantially lower relative increases in land value than do the series of land value per pasture acre equivalent which are used in this study. This indicates that the pasture acre equivalent series have been more sensitive to the decrease in farm land than have other indices. Constant term--With the exception of Rhode Island, all the constant terms were negative. The Rhode Island coefficient is not significant and the positive sign could be due to intercorrelation with conservation. Reasonable significance was obtained in Northeast 1 states and in the combined analyses for Northeast 3. Price expectations--The price expectations were generally quite significant in the analyses. Increases in the indices of price expectations in the Northeast over the 1925-62 period have generally been much below increases in other regions. Connecticut and Maryland have above average increases in indices of price expectations due to the importance of tobacco in these states. The dairy industry is an important enterprise in all the states, but it is the reliance on poultry WhiCh has kept the price expectation indices down. The importance 0“- Potatoes in Maine (more than 50% of total farm sales) has kept its index on a particularly low level. Conservation expenditures--Compared with farm land area in the region, the conservation payments are relatively large. However, con“pared to the level of land values the conservation payments to Northeast 2 were relatively small. The estimated coefficients are, except for Northeast 2, lower than the coefficients for other regions. 94 A very large part of the conservation expenditures were spent on measure to establish permanent cover such as the use of lime, fertilizer, etc. Except for New York, the coefficients were significant at the . 05 level or better. Conservation reserve-~The participation in the conservation reserve program is relatively low in Northeast 2, but rather high in Northeast 1 and 3. The many insignificant coefficients of the variable seem to be due to little participation in the programs. Acreage reserve--Maine and Rhode Island did not participate in the acreage reserve program at all, and for the entire region the participation was low. Coefficients are estimated for Northeast 1 only. Output per man-hour and output per acre--The output per man-hour variable was significant at the . 05 level or better in Northeast 1 but was not included in the other sub-regions due to intercorrelation with other variables. The output per acre variable did not show any significance in the state analyses of Northeast 2 and 3. However, output per acre was significant at the . 05 level in the combined analysis of all years for Northeast 3. Population and income--Population density showed some Significance in the combined analyses, but was generally insignificant in the time series analyses. Personal income was not significant. “Lake States Table 7 gives the statistical results for the Lake States. A good fit and no serial correlation problems are indicated. Heteroge- 95 6222303 mug/0220.2 9.22 mag-9 pong-S2253 mum 32822232225 whom 0.23922 ofi .momrmom consacoo 92» n2 .momofibohmm 222 22®>2m ohm mco2um2>o2u vnvaMum N .02” .~ "NHOWQGCAE .wo . "GmeOUmwg NCO ow "GMMMJUMEIIWOUMHQ 0>MHMHOH «HO mummdfi. 2 222. .2 8 .N N22 .2 82323.5 cements-2222225 222. N2.. 2:. mm. 2.. N22 22 .2.2 :3 .N2 2222 .222 22.2 .N2 22 .N2 NN.22 SN 2.N.2.2 £2.22 m2..N 2.28222 29.8332 22N .2 $2 .2 22: .2 SN .2 22N .2 mm . 2N . N2. . .22 . «N . .222..ch 282222292 222 .2 22222 .2 $2 .2 8o .2 $2 .2 ON . 0N. mm . NN . .NN . Son-:32 .222 222226 SN .2 :2 .2 2mm .2 8m .2 2NN .2 mo . mo . mo .2 3. . M2. . o>uomom owmouor. 222m .2 2222. .2 2mm .2 2%. 822 .2 mm .N 2.22 .N $.N NN . 2.N .N 3.88m 2823:0980 SN .2 $2 .2 23 .2 2222 .2 :N .2 3.23.22.5me we . vw . wv .2 2:. . co . Goth-enomnou 2S .2 2.12 .2 2.22 .2 22.22 .2 2:2 .2 N2. 2. 2 . 2.2 . 2.2. 2. 2 . 202238me 322m 22$ .2 :N .2 8m .2 8m .2 N22m .2 N2. .N- 25 .2- 2.m .N- 8 .N- cm .2- 2.852 33300 3.5 whom-w Go>m~ mums? 224 3022022222224 222m220om2>> 22mm222022>2 o2£o2am> mo 322-9222. 220222222220 0 2 ill 1" «.395 2333262 No-mNS .mqo2mm3mmm 222m > @284 20 E22532 "m83m 823 N “.2an 96 neity among the states is indicated by the fall in the multiple correlation coefficient when the states are combined. Some random errors in the data, mainly in the dependent variable, are indicated by the increase in the multiple correlation coefficient when every other year is omitted. With the exception of population density, the variables were generally significant in the analyses. Value andguantity of land-~The percentage increase in land values in the Lake States region during the period 1925-62 was substantially below national average. However, Michigan showed an increase which was only a little below the national average. The number of pasture acre equivalents was very stable over the period. Michigan showed the largest change with an 18 percent decrease. Minnesota and Wisconsin had small percentage changes between the first and last year in the period. However, there were first increases and later decreases in the quantity of land, and the difference between the lowest and highest year in the entire period was about 10 percent. Constant term--The constants were negative and significant at the . 01 level in all the analyses. When every other year was omitted in the combined analysis, the constant did double as would be expected. Price expectations--The price expectations variable was significant at the . 01 level. The indices of price expectations, which were highly influenced by the sale of dairy products, 11 beef,and hogs, 11In Wisconsin dairy products constitutes about 50% of total farm product sale. 97 increased over time with about the national average. In the analysis utilizing every other year only, the price expectations became quite intercorrelated with income (. 72), and its significance decreased somewhat. Conservation expenditures--C ompar ed with farm land area, the Lake States received a relatively large part of total government conservation expenditures. In Michigan and Minnesota a large part F i: of the conservation aid was spent on drainage, while in Wisconsin i’ the major part was spent on establishing permanent vegetative covers. I Michigan and Wisconsin received, relatively, the largest amounts (see Appendix D). The coefficients were significant at the . 01 level. Conservation reserve--The conservation reserve program has had much influence on land values in the Lake States. In 1960 the payments to the Lake States for areas in conservation reserve amounted to almost 42 million dollars, which is more than 12 percent of the total payments made that year. Next to Texas and North Dakota, Minnesota has been and is the state receiving the highest conservation reserve payments. The estimate suggests that 36 percent of the increase in Minnesota land values during 1955-62 is caused by the conservation reserve program. The relatively low coefficient and low significance level of the Wisconsin estimate indicates that the opportunity costs there are closer to the conservation reserve rates offered than they are in Michigan and Minnesota. Acreage reserve--There was considerable participation by the Lake States in the acreage reserve program. While a considerable part of Michigan's participation consisted in reducing wheat allotments, the 98 major crop reductions for the region occurred by decreasing the corn allotments. The only explanation for the higher Minnesota coefficient is that the rates offered there were relatively more attractive to farmers. Except for Wisconsin, the coefficients were significant at the . 05 level or better. Ouiquer man-hour--The output per man-hour variable worked well as it gave significant results in most analyses without giving inter- L‘ correlation problems. Income and population--The general level of economic activity “1.: Inc-“hm..- _ _. as measured by personal income has considerable influence on land values. This is especially true in Wisconsin where the income variable was one of the most important. All the states have land which is very usable for recreation and the proximity of large industrial towns is also a factor of importance. Except for Michigan, the income coefficients were significant at the . 05 level or better. The high level of significance for the Wisconsin variable is most likely connected with the proximity of Chicago. Population density was not significant in the analyses. Corn Belt The statistical results for the Corn Belt, which are presented in Table 8, do not show as much uniformity among states as expected a priori. The low multiple coefficient of determination for Ohio is mainly due to four observations, namely a low negative residual in 1950, followed by a high residual in 1951. The same thing happened, with opposite signs, in 1957-58. This also explains the high D. W. 99 I; It!” ,. .‘Eflfllfltfli .momofiunohmm 222 220>2w 0.2.2.2 22220222232222 pumps-mam .wm . "2.2509222 2222.22 .om .2 222302 .om .2 "m2o222222 .oo .2 2222222222222 .vw . 202220unmoo2fi2 92222-2292 so pom-ma. 0.2m 2202223 m2£w2o3 wag/02202 222.22 w222m22 25222222222222.2222 ohm 322022252525 whom 92259222 922 2202222222222 2202222222200 222 222 2 22 .N 222 . 2 m2. . 2 2: .2 22.. .N 82222322 2832.25-23.25 2222. N2.. 222. 5. N22. 2222. mm. Nm 2NN .2 2222222 .2 2pm .2 25 .22 2Nm .2 222. .2 2NN .2 mm. N2222 . 2.N .- 22m .- 22m . 2.22 .2 222 .- 2222...me 28222222222022 2NN.2 22.2.2 222N.2 2222.2 23.2 2222.2 222.2 2.2.2 3. 2.2.. No.2- 2222. N2.. 222.. 8222232 22.N .2 2222 .2 2222.2 2222 .2 23 .2 222. .2 22.. .2 22m .2 2222. m2. . 2.2 .- 222- . mo . 3 . 2,2332 omega... 2222.22 2222.22 2NN.22 2NN.N2 23.222 2NN.N2 22.2.N2 22. .2 2222 .22 mm .2 3 .2. 22. .N2 3 .m NN .2. 3.5322 282.23.82.20 222 .2 2222 .2 2NN .2 2N2. .2 2222. .2 28 .2 22.2 .2 m332222822xm 2N . to. . 2m .2 2.2 .2. 222. .m S .2 222. .2 28223.23qu 2222 .2 2:2 .2 2:2 .2 2222 .2 2NN .2 22.2.2 222. 2N . N2. . 2.2. 22. . 2..., . 2.N . NN . 3022882252 82222 22.22.22 2.1..2 22.2.2 222N.22 2222.22 2222.22 N22222.22 2.22 .- 2.2. 222 . 2- N2. .2.- S .N- 2222 .N- :2 . 2- 2222222 283280 222:0 mums? 220229 whom? 22% 2222092222 .9502 m2o222222 2252222222 022.20 022262-2222? 2mom>2wn< 2202222222200 moumum 3522222222222 .v .No-mmo2 .mno2mmonmom 0522222 2222224 20 3225022 222022 22.200 , .22 2.222222% lOO statistic which implies negative autocorrelation. The modest coefficients of multiple determination were mainly due to a few big errors. The model apparently did not catch the 1951 land market boom, and especially for Illinois, a stronger business cycle indicator might be useful. The combined analysis brought about a large decrease in the multiple correlation coefficient, but the significance level of the coefficients increased substantially for all variables except the constant. Value and quantity of land--The increase in relative value of land in the Corn Belt was somewhat below the national average. How- ever, there were wide differences between the states—-from a less than 80 percent increase in Iowa (using the endpoints in the period 1925-62) to more than 180 percent increase in Indiana. Due to the high absolute level of land values in the Corn Belt, the absolute increases were very large. The number of pasture acre equivalents for the Corn Belt was as stable as in the Lake States. The largest decrease (19 percent) occurred in Ohio. Constant term--The constants were negative and significant at the . 05 level for the individual states except for Ohio. However, the combined analyses yielded constant terms that were not significant. In the combined analyses large declines in the conservation expenditure coefficients occurred simultaneously with an increase in the constants. This suggests some interaction between these two variables. Also the individual state estimates suggest some relation 101 between increasing negative constants and increasing conservation coefficients. Compared with other regions the constant and the conservation coefficient in the combined analyses of the Corn Belt would appear to be too high and too low, respectively. Price expectationsuThe price expectation indices increased over the 1925-62 period with about the same as the U. S. index of agricultural price expectations. The coefficients were significant at the . 05 level or better except for Ohio. The main enterprise in the region is corn, but it is for the most part converted into beef, hog, and dairy products from which a main part of gross farm receipts are obtained. The estimated coefficients vary, as would be expected, according to the value of a pasture acre equivalent in the respective states. Conservation expenditures--The government conservation payments from 1936 to 1961 to the Corn Belt constituted about 18 percent of total U.S. payments. The conservation measures consisted mainly of lime and phosphate treatments and drainage. The coefficients were significant at the . 05 level or better. Conservation reserve-~Acreage participation in the conser- vation reserve program is close to the U. S. average when compared to total number of acres of farm land. However, due to the higher Payments per acre, the region received 14-15 percent of total U. S. Payments in 1960. All coefficients were significant at the . 05 level or better and were larger than in other regions. Acreagg reserve--The participation in the acreage reserve Program was very high in the region. In 1958 more than one fourth 102. of total acreage reserve payments went into the Corn Belt. Iowa alone received more than 7 percent of the total payments. The program mainly affected corn allotments. Also the wheat acreage was reduced, and in Missouri substantial payments were made to reduce cotton acreages. The coefficients were significant at the . 05 level except for Illinois and Iowa. Fertilizer--Fertilizer use was significant in the combined analyses at the . 01 level and some of the single state coefficients were significant at the .10 level. However, the Iowa coefficient had a negative sign. This might be due to intercorrelation problems. For example, the simple correlation between fertilizer and conser- vation expenditures was . 49. Population and income--Population density was significant in the combined analyses, but the coefficient was very low. Due to intercorrelation problems with price expectations, personal income was not retained in the analyses. Delta States The statistical results for the Delta States are given in Table 9. The goodness of the statistical fit varies among states, and the usual decline in the multiple correlation coefficient for the combined analyses is observed. For the combined analysis including even years only, the multiple determination increased more than in all other regions, except for the Southern Plains. Value and quantity of land--The Delta States are among the states in the nation with largest increases in land values. Among 103 .momofiooumm 222 220>2w 0.2m 92022322622 pudendum-n. .mo .2 222333.904 2222a .00 .2 "mmmomv224 .22. . 2222222mm2mm22>2unmoo2nm 222/22.392 220 pom-2222 22.2-m 22022.23 32232225 mag/02202 02.22 w222mo 22223222322222: 0.2m 3222229225220 0.20.22 22.23.22.222 @222. £0922.me 2222222222022 @222 222 2 e2. .2 N22 .N E .2 8222.22.26 segue-2222222222 me. 222... N2. 2222. 2.22. N22 2222 .222 22222 .22 2.222 .2.2 22N .N 22 222. .222 N2. .2.2 2.22 .m- 22. .- mo .N 3 .22- 2:82.22 2982.322 222. .2 22.2 .2 22.2 .2 22222 .22 2m2. .2 222 . 222 . A22. . N2. . 22m . 222.com 282.22.322.22 22222 .2 22222. 22.22 .2 2222 .2 2222 .2 222 .- 2222. 2.22. 222 . 222 .- 2.8.2.. $22 222.26 222.2 22.2.2 22.2.2 2NN.2 2N2.2 N2. N2. 2...: 2.2.. N22. 92.88222 $8.822 23.22 2222. .2 22.2.N2 222m .22 2222.N2 2.2 .N mm .N 222 .222 2N .N 2.N .2 9;..QO 222.223.23.80 SN .2 2222 .2 22m .2 222. .2 222N .2 $2322.522xm NN.2 NN.2 2222.2 2.N.2 .NN.2 2822928280 22.22 .2 2222 .2 222.2 2N2 .2 22222 .2 2.22 . NN . 2.N . 2.N . 222 . 3022882252 32.22 22.2. .2 2222. .2 2222 .2 2222 .2 N223 .2 2.N .Nn 2.2V ... 2N ... mm ... mm .n 2222222 22239200 22:0 made-x- Go>m undo-W 22222 922.223.9012 moms-92.222- 222222mm2mm222 o2£m22o> 2momfl2ms< 282222222200 moumum 292222322222 { N2.-322.2 2 222223222322 222232 223. 222 32222222 item 22.22.26 .m 2232K 104 individual states, Louisiana is second only to Georgia with respect to relative increase during the 1925-62 period. Arkansas had the lowest relative increase in land values in the region, which was, nevertheless, substantially above the national average. The number of pasture acre equivalents had an increasing tr end in the first part of the period and a decreasing trend in the latter years. However, the differences between 1925 and 1962 were Within ten percent. Taking the lowest and highest years of the period gave differences up to 25 percent. Constant term-~Constant terms were negative but not significant in the analyses except in the combined analysis of even years only. Price expectations--The price expectations were significant in the analyses at the . 05 level or better except in the combined a-11a.1ysis utilizing time aggregation. The increase in price expectation indi ces, mainly reflecting cotton and oil prices, was less than that for 111CDist other regions. Conservation expenditures--Conservation payments to the r e gion over the period 1936-61 were, relative to farm land area, g1- e ater than to most other regions. Most of the conservation practices De 3:‘tained to establishment or improvement of permanent covers. A sub stantial amount was also spent on drainage. The coefficients were 3 5‘ gmificant at the . 01 level. Conservation reserve--The conservation reserve payments Ina-(is to the region are about average in the sense that the percentage Qf total payments under this program received by the region is close 1: Q the region's share of total farm land. However, within the region 105 Arkansas receives relatively most of the conservation reserve payments. While the Mississippi coefficient is not significant, other conservation reserve coefficients are significant at the . 05 1e vel or better. Acreage reserve--The Delta States participate heavily in the acreage reserve program. In 1958 alone they received more than 11 percent of all payments made under the program. Mainly the production of cotton and rice was affected by the program. The A rkansas and Louisiana coefficients are significant at the . 05 level. Other coefficients are not significant. Output per acre--The output per acre variable was not 3 5- gnificant in the analyses. Output per man-hour was highly correlated With land values, but was not retained in the analyses because of inter- C Q r relation with other explanatory variables. Population and income--Population density was significant in the combined analyses, but not in the individual state analyses. Income Wa 8 significant only in the combined analysis with time aggregation. I‘Ic>\7vever, in this analyses correlation between income and price expectations was high (. 66). sflljhern Plains Table 10 gives the statistical results for the Southern Plains, and a reasonably good fit is indicated. The D. W. statistics for 01‘1 ahoma is so large that there is reason to suspect some negative 3.111: Qcorrelation. The combined analysis gave a modest decrease in t he multiple correlation coefficient, but where every other year was 106 32230.5 9222262202 @222 92232 22022222232523 v.33 322032222220 whom 0.233222 2.2222 .mom222m22m 22022222222022 222.22— 222 .momoaunohmm 222 22o>2w ohm moon—2222622 22.2-222222.95 N .wo . $38.2. 222222 .00 .2 "maoflm2xollmoo2nm 922.202 no woman 0.2.22 2202223 2 2.22 .N .22. . 2 3222.22.25 22823-522225 2222 . 2222 . S . N22 . N22 2N2.2 222N.2 SN .2 2.22.2 2222 .2 NN . 2.22 . 2.2 .2 222223 8222.29.22.22 22.2 22.2.2.2 2N2.2 2N22.22 222 . 2.22 . NN .2 N2. . .2.-2.222232 2222 .2 22N .2 22.N .2 2S .2 2o ..- no .u we . MN ... o>nomom 093202.. 22.2. .2 22.22 .2 22.2. .2 22.2 .22 N22 . 2 22 2 .N 3 .N N2 .2 9:332 28223.22...qu 2222 .2 2N2, .2 2pm .2 22.... .2 massages-.22 N2. .2 mm .2 222. . N2. . 2 222.223.28qu 22.22 .2 2.222 .2 22.22 .2 2222 .2 w 2 . o 2 . o 2 . 2 2 . m22022m2oo22xm 922.222 2Nm .2 2mN .2 222m .2 N222. .2 om ... 2.N .- m 2 .u 2.2m ... 22222223 «22.32200 .2225 mums-w 22o>m2 .2220? 24 2222202222220 933.2. m2202um22m> 2 m 0 922-2244.. 25222222220 O 83% 232.2325 I will n 2 .22 222.2 2222222222QO 22222222 22223 22. 222222.22 2222222222 2222222222222. .3 22225 107 omitted the coefficient increased more than in any other region. Value and quantity of land--In the Southern Plains, land values increased during the 1925-62 period by a percentage which was above the national average. The absolute level of land values as well as the relative increase in land values is uniform for the tw 0 states. The quantity of land was fairly stable in Oklahoma with the maximum reached in 1934. Texas had an almost continuous increase in number of pasture acre equivalents throughout the period, and the total increase amounted to 32 percent. Constant term--The constant term for Texas was negative and significant at the . 05 level and for Oklahoma it was negative, 8 I"ITIaIl and insignificant. Price expectations-4n the examined period, price expectation i"CI-dices increased about the same as the national average index. The Iuru-aain enterprises influencing the indices are cotton for Texas and wheat for Oklahoma. Beef production is very important for both 3 tastes. The higher coefficient for Oklahoma could indicate that W11 eat price expectations influence land values more than do cotton 1)::- i ce expectations. More evidence on this proposition will be presented in the next chapter. The coefficients were significant at the . 05 level. Conservation expenditures-~The government conservation pay- htl ents to the Southern Plain states in the period 1936-61 were compared Wi- th the land area in farms below national average. A large part of the Q Q’Ili'iservation expenditures in Texas were for ”control of competitive allIrubs on range and pasture land. " In Oklahoma there was much a «when!!! ’2 2 108 emphasis on measures to establish permanent cover and large amounts were spent on fixtures for watering of livestock. Water conservation was quite important in both states. The coefficients, which were all significant at the . 01 level, indicate that the conservation practices carried out in Texas had a large impact on land values. Conservation reserve-~A1though the conservation reserve payments made to Texas are the largest for a single state (39 million dollars in l 9 60) the payments are less than average when compared with land in The rates fa. arms. The payments to Oklahoma are relatively larger. paid per acre are fairly similar in the two states, but Oklahoma has a. larger percentage of farm land participating. The estimated (2 oefficients and their significance indicate that the conservation r e s erve program has a considerably larger influence in Oklahoma th an in Texas. Acreage reserve--The acreage reserve program did not yield The participation was below average. Cotton 8 5— gnificant results. Wa s the main crop affected by the acreage reserve programs, but in O1<1ahoma considerable participation came from wheat areas. Fertilizer--The fertilizer variable was significant at the . 05 1 e Vel for Oklahoma, but was not significant for Texas. The high Q Q hservation coefficient in Texas might include some effects of t . e thological changes. Population and income- —A relatively large increase in the I)Q’pulation in Texas was reflected in a significant population density Q eff1c1ent. Income was not retained 1n the analysm as it was inter- Q :rrelated mth price expectations. 109 Alorthern Plains Table 11 gives the statistical results of the Northern Plains, and a good fit is indicated. At the . 01 level, the assumption of independent disturbances cannot be rejected for any of the time s e ries analyses. In the combined analysis, utilizing every other year only, multicorrelation became more of a problem than in most other r e gions. Many of the coefficients changed substantially, and the income coefficient changed sign. Thus, the increase in the multiple correlation coefficient brought about by time aggregation was accompanied by less desirable features related to this procedure. Value and quantity of land-~The Northern Plains region had a s mall increase in land values both in percentage terms and in absolute te rms during the 1925-62 period. In percentage terms, South Dakota had the lowest increases in land values for the entire country (41%). I{a-I'lsas and North Dakota, even though they were far below the U. S. a"re rage increase, had increases of more than 100%. The number of pasture acre equivalents increased for all the s . . tates. For Kansas and Nebraska the increases have been continuous th 3": ough the time period, while some small decreases occurred for SQ11th Dakota and North Dakota in the latter years. The increases f O 1‘ the entire period were from 19 to 32 percent. Constant term--The constants were negative and significant at the . 01 level for all analyses. 3 110 .momoeooonmm 222 220>2w 0.2m 92022222222222 pumps-mum N .N 2 .2 "moms-22M 2222s .oo .2 22222222222232 .mm . 230me 222-Dom .om . 2.30me apnoZ-ImooEm 22222222292 220 woman. one. 2202223 322N203 wn2302202 0222 N229: 220-2222232233 0.2m 32202222252222 who-m 22.2.222me @222 mom>2m22m 2202222222200 @2222 222 2 2222 .N 222 .N 22... . 2 222 .2 822222.25 53.222-223.222 2222. 2.2. .22. N2. 2. 2.2. N22 2M2..N2 222.22 222.3 22.2.22 222.22 222222 22N .2.. 2.22 .N- 22. . 222 .2- 2.22 .N- 22. .2- M.62.“...2 2228.22.22 22.2 .2 22.2 .2 222.2 .2 SN .22 22m .22 222.2 .2 2.N .m 2222 .2 22.. 22... ... ms . 222 . 222222.22 222.222.222.202 2N2 .2 222. .2 2222 .2 22.2 .2 2222 .N2 2222 .22 t... 2.22. 222 . 2222. Nm .N 22. .2 $222232 22N .2 22N .2 2.222 .22 22.2 .22 2222 .2 222... .2 222.2 2.2-:2 22.2 2.N.2 222.2 222 . 2.32.322 222.223.3280 22N .2 22N .2 2222. .2 222.2 .2 222. .2 2222 .2 2225.22.22.20. 3 .2 2222 .2 2.2. .N N .N 2.2. .N 222. .2 222.222.53.80 22.22 .2 2.222 .2 2222 .2 2222 .2 22.22 .2 2N22 .2 2.2 . 2.N . 222 . 2222 . 22 2 . >22 . 222.223.22.22 82.2.2 2:. .2 2NN .2 22.... .2 2222. .2 22m .2 N22.2 .2 N22 .N- Nm .2- N2. .- 22.2 . 2- N22 .2- mm .- 22.8222 25.22.222.22 222:0 3822222 32.23 undo-w 2205.222 memo-x- 224 22.22.2532 232992222272 floom 222.2072 22022622m> 2.2822292... 2.8222800 $2.3m 22.222.232.222 H 1.1....111 A} 3222.2 2222222222222 2222222 22223 22. 2222222222 2222222222. 2222222222222 .2 2223 111 Price exgectationsuThe estimated price expectation coefficients are significant at the . 01 level. The indices of price expectations increased during the entire period, a little more than the national average except in North Dakota where the increase was less. Wheat is the dominant enterprise in North Dakota and Kansas. Beef produc- ti on is important in all four states, but more so in South Dakota and Nebraska where corn and hog production is also rather large. Even t110ugh the pasture acre equivalent prices in this region vary from state to state and the coefficients therefore cannot be directly compared, th e coefficients seem to indicate that in states where wheat is the Pr edominant enterprise the price expectations have had less influence on land values than in states with mixed agriculture. Conservation expendituresuFrom 1936 to 1961 the government 8 111:) sidies to conservation in the Northern Plains as compared with the a. :1: ea. of farm land, were much below average. Even when we consider th e relatively low land values in the Northern Plains, the conservation 8 11b sidy per dollar value of land is below that paid in the southeastern pa- rt of the country. The conservation expenditures appear to be the Si hgle most important explanatory variable in the analyses of the N O :rthern Plains. In North Dakota a large part of the conservation 3.1 <1 is used on temporary protection from erosion, and we would ea":IJect a lower coefficient. In South Dakota a large part of the Q thervation consists of measures to provide water for livestock. In Kansas and Nebraska the major emphasis is on measures to Q ' . QItiserve or dispose of water; terraces and sod waterways are the thrlQst important single measures. 112 Conservation reserve-~Participation of the Northern Plains in the c onservation reserve program was high with North Dakota having particularly high participation. However, the estimated coefficient for North Dakota is rather low. This indicates a small absolute effect of the conservation reserve on the income stream to a pasture acre equivalent. However, considering the low value of land in North Dakota, the relative increase is large. The estimate suggests that about 25 percent of the land value increase in North Dakota during the 1 955-62 period was due to the conservation reserve program. Acreage reserve-~Participation in the acreage reserve program was very large in 1957 but declined rapidly in 1958. Wheat was the main product affected, and Kansas received larger total payments in 1 957 than any other single state. The variable was not retained in the final analyses since it was not significant but did give inter- co :- r elation problems. Fertilizer--The fertilizer variable was significant at the . 05 level only for North Dakota and the combined analysis for all years. Population and income--Population density was significant at the - 1 0 level in the combined analyses. Personal income was signi- ficant at the . 05 level with negative signs for three of the states. This indicates a decline in demand for land associated with increased personal income. The relatively high coefficient for Nebraska might suggest that Off-farm employment is easier to obtain there than in North and South Dakota, 113 Mountain States The large Mountain region was divided into two sub-regions12 according to geographical and statistical differences. Tables 12 and l 3 give the statistical results. The Mountain 2 region and Idaho show signs of positively autocorrelated disturbances and the dependent variables indicate some cyclical expectations not accounted for by the model. A strong business cycle indicator might be of value to explain some of this variation. However, the land values indicate low exPectations during W. W. 11 and subsequent high expectations of the post war years. This cannot be explained by any of the usual business indicators. In the combined analyses personal income and population density became significant where they were retained in the final analyses. Value and quantity of land--The percentage increase in land values during the 1925-62 period were generally above the national averages. The absolute increases were low since the money value of a Pasture acre equivalent is very low. Arizona and New Mexico had the largest relative increases, while Idaho and Nevada were somewhat below the national average. All the Mountain states had increases in the quantity of land. The increases were continuous throughout most of the time period, but lz'The two regions are: Mountain l--Montana, Idaho, Wyoming, Colorado, and Utah Mountain 2--New Mexico, Arizona, and Nevada 114 .momoauoonmm 222 22o>2m mum 920222222522 22.2-32222225 N .mo .2 22.22322 222222 .N2 .2 2022-98200 .222. . 29222202225 .mm .N 2022.222 .00 .2 “mom222021-moo2222 222/2230.2 220 pom-2222 N2.2-m 2.322225 m2£w2o3 922302202 02.22 N222m22 22222222222223? 9% muse-2222222225 whom 022229222 0222 2222922922... 2222222222200 @222 222 2 2.2. .2 2N .N 2.2 .2 N2. .2 N2 .2 822.2228N ....BmB-cfiSo 2.N. 2..2. .2.2. E. 222. N2. 2.N. N22 2.22.2.2 22.2.N.2 22.22.22 2NN2.22 22222.22 22NN.N2 222.2..2 2222 .2 2.2..2 . .222 . 22N. N22... . $2.. .2. 22N . 288.22 29.8.32 2222.2 222.: 222N.2.2 2N22..2 22222.22 2N2.2..2 2oNN.N2 .222. .2 N2.2. 2.NN. NNN .2 NON . .222. 2.222 .N ...-22222232 2:2. .2 222 .2 22.2.2 .2 222.2 .2 22.2.2. .N2 2N2» .2 22.N.2 .2 22.22. N2N .- 2N2 .- .2.22.- N22 .2- 22.N .- NNN .- 2:332 2N22.22 2N2N.2 2NNN.2 22NN.22 2NNN.2 2NNN.N2 202N.22 2222.22 2222 .2 322 .N 22.2. .2 N2.2. .2 2222. .2. 222 .N 2mm .2. m$2.522 822322200 22.NN .2 2.2: .2 222.22. .2 2222. .2 :3 .2 22.2..2 .2 22222 .2 2.2..-222.....22xm N22.2 .22.2.2 N22.2 NNN.2 .22.2..2 2N2 .N 2.2.2 .2 8223.23.80 222222 .2 2.22222 .2 2NN22 .2 2.2222 .2 2N222 .2 22.N22 .2 2.2222 .2 NN22 . S22 . .2222. N222 . 2.N22 . .2222 . .2222 . 28223....me 82E 2N222 .2 222.22 .2 22.3.2 222 .2 2.22.22 .2 SN .2 2NN2 .2 N 2N .- N2.2 .- NNN .- 22...N .- N2.2 .- NN... .- N22 .- 228222 22.23.2822 222220 22.N-mow oo>m2 whom-x. 22¢ grub ops-20200 92222202225 022N222 92.322032 220222.25» 2222 922222222. 22222222222222 0 2 $222225 222.323.4224 Nolmmmm 2m2202mm02wom 0522222- Ugmq .20 mud-50m .2 33.2262 .N2 m2.2.2.2. 115 ... . v - [I‘lliui 1.... tutu-F 2.2. .momoeuooumm 222 22o>2m 22.2-2... moo22m2>op 22.2-momma N .2.o .N 22222222672 2222.22 .mo . 222220N22< .oo .2 2002232212 BoZuumoo2222 o>22m2o2 220 pom-mo. ohm £022.23 NEWER. 922302202 222.22 M22232 22022222232233 who? muoo2m>2fivo who-m 22.252.22.22 0222. 2392222222. 2202222222200 222 222 2 m mmzmcdw 220222222220 0 meadow 2w5222>222~22 No-mNmL 62202322322 222222.222 2.22.2.2 20 32262. NN . 2 2.22 .2 2.2 .2 822.2235 cots-2273.22.20 2.2. 2.N. 222.. 2.2. 2.2. N22 :2. .2 SN .2 EN .N2 222.2. .2 22. 2N .2 N2.2 .2 NN22 .2 2.NN .2 222.22 . N222 .- 222.com 2822.29.28 2223.2 228.2 22.8.2 2.28.2 2N222; 2.222 . 22.22. moo . .22.N . N2.22 . ham-.322 2......2 222.2225 2N22.22 2m2N.2 2222.22 2omo.2 2.NN. 222.2 - 23.2 N22.2 9.22.322 omega...- 2N2v.22 222N.2 2Noo.ow2 222N.2 NN.2 .N- 22.N .N- - SN .22 NNN .N- 8.3322 2822388022 22.2 .2 22.N2. .2 2N 222 . 22 2NN2. . 22 2.2 2N .2 222322.:o22xm EN .2 N2.2. .2 .2222 .N N2.2 .N .222 .2 822.25.28.80 2222 .2 22.2222 .2 22.222 .2 2NN22 .2 22.222 .2 2222 . N2222. 2.2222 .- N2222 . 2.8 . 22.02.389.22 82.2.2 38 .2 2.2222 .2 NS .2 2N.22 .2 22.22 .2 2.22. .- N2.2 .- 2NN .- N 2N .- N 22 .- 2.8222 2.28.2822 22:0 mums? 220222 whom-x- 2222. 2222222672 moon22< 00223224 .2672 22022522222/ 22262223022. .M2 0.2an 2N. 116 in the latter years increases were smaller and in some instances decreases actually took place. While Colorado had only a 17 percent increase for the period as a whole, other states had substantially higher imreases. Idaho, Arizona and Nevada had increases between tvvo and three hundred percent. Constant term-~The constants were negative in all the analyses and except for Montana and New Mexico, the coefficients were significant at the .10 level. Price expectations--The price expectation indices were signi- fic ant at the . 05 level in the Mountain 1 analyses, except for Idaho. In Mountain 2 only New Mexico had a significant coefficient. Beef is a main enterprisein all the Mountain states. In Mountain 1, wheat is an important enterprise, and cotton is very important in New Mexico and Arizona. The increases in price expectation indices over the time period were generally substantially larger than the increases in the U. S. agricultural price expectation index. Conservation expenditures--Conservation subsidies for the region during the period 1936-61, considering the quantity of farm land, where relatively small. The main part of the conservation aid was Spent on water conservation and a substantial amount was spent on measures to provide water for livestock. Some of the effects on land Values brought about by irrigation measures would be accounted for In the measure of pasture acre equivalents, but such measures as terrating and leveling of land for efficient use of irrigation water would not be accounted for. The coefficients obtained show a strong correlation _ .1 ‘23-‘- NIP .1 117 betwe en land values and conservation subsidies. Conservation reserve--Participation in the conservation reserve program is generally low in the Mountain region, and practically zero in Nevada. The highest participation is in eastern Colorado. The estimated coefficients are significant at the .10 level in Mountain 1 except for Idaho. The unreasonable results obtained for Mountain 2 state 8 must be attributed to spurious correlation in connection with the small magnitude of the variable. Acrgigireserve--The acreage reserve variable was not significant in the analyses (except a negative significance in the combined analyses for Mountain 1). The participation in the program was low, both in acres and especially in money terms. The low acreage participation indicates that the rates which were offered in the Mountain states were not too attractive. In the Mountain 1 region, wheat was the only crop affected, while in the southern states mainly cotton acreages were reduced. Fertilizer and output per man-hour--The fertilizer variable inc luded in Mountain 1 was significant at the . 05 level in Colorado and in the combined analyses. The output per man-hour which was included in Mountain 2, was significant at the . 05 level in the combined analysis Of all years and in the time series for New Mexico and Arizona. Population and income-~As mentioned earlier, the personal i“Confle and population density variables were significant only in the combined analyses. Due to intercorrelation problems, population density was not retained in the analyses of Mountain 1 and income was not retained in Mountain 2. w T; .' U‘¢(.I.2f_ 2‘ 1‘er ..2 ‘5 I v . 118 Pac ific States The statistical analyses of the Pacific States were presented in Table 14 and a reasonably good fit is indicated. However, the D. W. statistics for California indicates serially correlated disturbances at the - 01 level of significance. Like some of the Mountain and North- eastern states, California land values seem to be influenced by expectations generated outside the state itself. For California this is further underlined by the fact that the population density is the mo st important single variable. 13 The behavior of land values during the years of and immediately following World War 11 makes the business Cycle indicators insufficient as explanations of the cyclical variation. Neither of the combined analyses showed a higher degree of multiple determination than the lowest single state analysis. The modest increase in the multiple correlation coefficient due to time aggregation suggests small random errors in the variables. Value and quantity of land--In the Pacific region, the per- centage increase in land values was about equal to the U. S. average incr ease. However, on a state basis the increase in California was substantially above the increases in Oregon and Washington. If we consider the period 1940-62 only, the differences between California and other states would become even more pronounced, indicating 1a1'8'31' falls in the California land values during the depression of the thirties, When population density was deleted from the California analysis, R?- fell to . 53. 119 .momoauoonmo 222 220>2w ohm. 320222222622 ohmoomumN .vw .2 22222220222220 2222s .wo . Homo-HO .oo .2 22202w22222mm2$n|moo2222 02222-222222 so pom-mo. N2.2-m 2.202.223 RENE? 9222502202 @222 mo2m22 22oN2222m22222m2m v.33 322322522225 whom 92.25922 @2222 £092,293 2222222222200 222 222 2 2N .2 2.N .2 2.N .2 Erasmsm magma-.2325 2.2. 2.2. NN. N2. N2. N2 2.22.2 222.2 222. .2 2.22 .2 222m .2 2.N . 2.2. . 22 .N m2- ..- om . >22m220n2 22022-252202 2.22 .2 SN .2 2S, .2 2222 .2 22.N .2 mo. 2.2.. 2m. 2w .2 Nm ... 20N22222oh 22.22 .N2 2N2 .2 2N.2 .22 N22 .22 22.2.. .2 mm . 22N ..- mo . mo .Na 5. ... 222/202.2022 owns-No.22. 2.2N.N2 2N2..22 2N22.2.22 222.22 2222.22 NN. N. 2.2. .... NN .2- 2.2 . 2:332 2.022.252.3022 2N2 .2 222. .2 22.22 .N2 2N2 .2 22.2.2 222322.525 222. .N Nm .N Nm .N N .N S .N 28229.8..qu 2.222 .2 2N2. .2 22.22 .2 2N22 .2 22.22 .2 N2. N2 . 22 . 2.22 . 2.2 . 302.22.282.22 82.22 2222 .2 22N .2 222. .2 SN .2 N2NN .2 2222 .N- 2.2. .- .25 .N- NN .- Nm ... 2.28222 2:32.80 .2226 memo? 2222262 mums-W 224 22222-20230 cows-HO 22392222935 Go2um2hm> $2,392.22 222.222.28.20 mossm 232.232.:2 Nonmmmfi 2mno2mmmhmom m52m> 22222212. .20 .325th "moo-mum 02.22UonH 4.2 02992... 120 The number of pasture acre equivalents in California and Washington increased over the entire period about 50 percent, while the increase in Oregon was 26 percent. These increases were c ontinuous . The USDA average value per acre indices for the Pacific states showed increases in land values which were substantially above the 1 U2 S. average increase. The disagreement between these indices and the value per pasture acre equivalent series used here is mainly due to the large increase in irrigation. Increased acreage of irrigated land increases the number of pasture acre equivalents and subsequently affects the change in their average value. Constant term--The constants were negative and significant at the . 05 level. For California the negative constant takes a relatively large value. This might inflate the estimated population density coefficient, because the population density has been continually increasing, though not at a constant rate. Price expectations--The price expectation indices increased over the 1925-62 period a little less than the U.S. index of expected agri cultural prices. Beef and dairy are important enterprises in all three states. Wheat gives a large part of total farm revenue in W~"3-Sl'1ington and Oregon, while fruits, vegetables and cotton are important for California agriculture. The coefficients were signi- fiCant at the . 01 level except for California. Conservation egpenditures--Conservation payments during the 1 925-61 period when compared with farm land area, were some- what below average for the U. S. The main part of the conservation 121 aid was used for measures to conserve or dispose of water. As for the Mountain states, the quantity measure of land does not take into account all the different conservation measures related to irrigation, and there is generally a high correlation between land values and conservation expenditures. However, the California coefficient is not significant. Soil bank--The soil-bank variables were not significant in the analyses of the Pacific region. The participation in the program was relatively small for the region. Fertilizer-—The fertilizer variable is significant at the . 05 level in all the analyses. The Washington coefficient deviates by having a negative sign, but it contributes little to the analysis. Population and income--Population density, which was highly irnPCDIP‘tant in the California analysis, was also highly significant in the C: ombined analysis. Due to intercorrelation problems, income was not retained in the analysis. W The selected variables used in the statistical model generally explained a major part of the variation in land value changes. The hr St difference model gave, for the significant coefficients, the signs which generally would be expected from an economic point of View- The assumption of independent disturbances could be accepted in meat of the analyses. Positive autocorrelation was indicated for a. few states, of which most had a low multiple correlation coefficient. A 10W Durbin-Watson statistic, indicating positive serial correlation a 122 of disturbances, was in many cases caused by the World War II and postwar years. The pessimistic expectations, indicated by the small increases in land values, during the war years were followed by high expectations in the postwar years. The combined time series and cross-sectional analyses gave a decrease in the coefficient of multiple determination. This implies some heterogeneity in the variables among states. The significance of the coefficients increased substantially in the combined cross- sectional and time series analyses as compared with time series alone- A large increase in the coefficient of multiple determination was brought about by aggregating the first differences over two year Periods. This indicates some random variation in the variables. At the outset of this chapter it was argued that random variation in the independent variables would be alleviated by the number of The random variation eliminated by time aggregation The time Vari able 8. must therefore originate mainly in the dependent variable. agg r egation gave in some cases multicorrelation problems and Subs equently less significance of the coefficients. All the independent variables listed in Table 1 were significant at one point or another in the explanation of land value changes. Price eFPe ctations and conservation expenditures were significant in almost all til'le analyses, and they explained a major part of the changes in land Value 8. The conservation reserve variable was generally significant in the areas where substantial payments were made through the program, w - hlle the acreage reserve variable had less impact on land values and m 123 did not show any significant impact in some areas where large pay- ments were made under the program. The output per man-hour variable was highly significant where it was retained. Fertilizer and output per acre gave best results in the combined analyses. Population density and personal income also gave the best results in combined analyses, but personal income was generally of little importance. The regional discussions suggested that among regions there is a rather close relationship between the relative increase in land values and the relative change in or the size of the explanatory variables. This will be analyzed and discussed in the next chapter. The r egional difference in the estimated coefficients will also be dis Cu ssed in the following chapter. X "r.l"-Q u CHAPTER VI COMPARISON AMONG REGIONS The analyses presented in the previous chapter focussed on the significance of several variables as related to the variation in value of land. Differences among states within the defined regions were discussed. In this chapter the major emphasis is on inter- regi onal differences. The chapter consists of three parts; one in which the estimated coefficients are compared among regions, another in which the magnitude of change in the variables is compared among regions, and a third part which combines the effects of coefficient size with magnitude of change. Regional Differences in the Estimated Coefficients Coefficients would be expected to differ among regions for the same reasons as they differed among states within a region. The aggr egate variables are built from dissimilar components, and these C01Illznonents do not enter into the aggregate variable in constant Proportions among states or regions. For example, the most impor- tant influence on agricultural price expectations in one state might be tobaC: co prices while cotton prices would have a major influence in a neighboring state. Similarly, the kind of conservation practices Carri ed out differ among areas. Interregional differences which will b - e d1 scussed in the following pages are generally larger than the int . r a- regional differences. 124 ‘ 125 Pr i c e E3er ctations The price expectation indices are obtained by aggregating the individual commodity indices. The weight of a given commodity index varie s with the commodity's importance, in relation to gross farm sale 8, in the different areas.1 For areas close to each other the vecto rs of weights will generally be much alike. More distant areas have, in general, substantial different weight vectors. 2 This hetero-i gene ity among regions is likely to result in different coefficients for pric e expectations. For instance, we would expect a more direct relationship between wheat prices and land values than between dairy pric es and land values. The coefficients for price expectations are given in Table 15. The coefficients are given for 13 regions and 11 selected states. In each of the selected states, a single commodity price index plays a IIla-j or role in calculating the aggregate price expectation index for the 8'33-":e. The importance of a single commodity price expection index is give :11 in parantheses behind the commodity (Column 4). For example, in Maine the revenue from potato sales constitutes 54 percent of the total revenue from the 13 commodities for which price expectations are available. In computing the aggregate index for Maine the potato 1See Appendix C. in 2The heterogeneity is not a function of distance alone. For 1n(Bi-"antics, due to climatic changes a distance along long’itudes Will be a]. e1y to give more heterogeneous vectors than would the same distance 0113 latitudes. 126 Table 15. Regional Coefficients for the Price Expectations Estimated 1940 value Adjusted Products yielding R egion coefficient of a pasture coefficients a major part of acre equiv. col.lx100 farm income col. 2 (1) (Z) (3) (4) Southeast (excl. Fla.) . 205*403" 15. 53 l. 32 cotton, hogs, tobacco Appalachia . 231 *** 28. 59 . 81 tobacco, dairy Northeast 1 .160**>¥< 21. 8O . 73 dairy, eggs Northeast 2 . 217“? I 49. 54 . 44 dairy, eggs Nor theast 3 . 080>1i< 8. 24 . 9? dairy, potatoes Lake States .17O>5<*>‘3< 22. O3 . 7? beef, dairy, hogs Corn Belt . 425>i<>i<>t~ 39. 35 1. O8 corn, beef, hogs, dairy Delta States . 225>t>7=* 21. 64 l. 04 cotton 50111: hern Plains .161>‘<** ll. 96 l. 34 wheat, cotton, beef Nor thern Plains . 237*** 15.11 1. 57 wheat, beef Mountain 1 . o37>:<>3<>:< 3. 24 1.14 beef, wheat, dairy Mountain 2 .008 3.00 .27 beef, cotton Pac iiic .123~»:w:<>:< 12.19 1. 01 all commodities §1 62 cted States: Maine .17o>r=>t>i 12. 97 1. 31 potatoes (54%) Nor th Dakota . 074:::<-:<>:< 5. 87 1. 24 wheat (47%) Arkansas . 267** 21. 64 1. 23 cotton (45%) WYoming . (137-totsl 3.12 1.19 beef (63%) V? rrIiont . 086>:<>:<>:< 7. 94 1. 08 dairy (72%) 1 S S issippi .156>fi‘4< 1 5. 38 l. 01 cotton (53%) IOWa . 471 49.40 . 95 hogs (42%) New Hampshire . 075* 8. 24 . 91 eggs (36%) ILEIentucky . 269...... 30.65 . 88 tobacco (32%) N or lda . 070 23.19 . 3O oranges (40%) Orth Carolina . 052 28. 59 .18 tobacco (50%) \ $9,: :13: 9.0;: Significantly different from zero at the one percent level. Significantly different from zero at the five percent level. Significantly different from zero at the ten percent level. 127 price expectation index has a weight of . 54 (out of l. 00). The regional coefficients in column 1 are those estimated in combined cross-sectional and time series analyses including all years. 3 The state coefficients in column 1 are those estimated from time series as presented in Chapter V. The coefficients are not directly comparable since they are obtained by regressing series of land values at very different levels in terms of price per pasture acre equivalent on price expec- j... tation indices at the same level. The 1940 land values, given in column 2, are used to adjust the estimated coefficients. 4 The adjusted coefficients given in column 3, express the impact of a one point change in price expectations on a unit of land worth $100 (1940 values). The adjusted regional coefficients show that changes in price indices of wheat, cotton, and beef tend to have a larger influence on land values than do changes in egg and dairy prices. This would be expected, since the latter group requires relatively larger amounts 0f rather fixed non-land inputs. Changes in returns, which would a'ffeczt the MVP's of fixed inputs, would be distributed over all such inPuts. Thus the change in returns to one such input would tend to be The analyses including all years were chosen in favor of those With increased time aggregation, because in the latter, multicorrelation Problems were encountered in some cases. The method is essentially the same as that used in order to czombine states in regional analyses. In the combined analyses the Ilumbers of pasture acre equivalents were adjusted according to relative land prices, thus changing the value of land series before the regression. Here the coefficients are adjusted after the regression. 128 smaller with) an increasing ratio of fixed inputs to total inputs. The adjuSted coefficients from the selected state analyses give essentially the same picture as do the regional analyses, namely that wheat, cotton, and beef coefficients are larger than egg and dairy coefficients. However, potato prices, which do not have a large influence on any single region, have the highest coefficient in the comparison among states. Tobacco and cotton were represented by two states each because the coefficients estimated for North Carolina and Mississippi, respectively, deviated from the general level of the coefficients estimated for other states where these enterprises were Prominent. Nevertheless, the results would seem to suggest that tobac co prices have a smaller influence on land values than do other cr0p prices. 92118 e rvation Expenditures The conservation expenditure variable consists of the govern- ment subsidy payments for agricultural conservation. The conservation Inea-8ures vary among regions, and to the extent that the varying practices TESult in different returns to land, the estimated coefficients would be expe cted to vary among regions. The estimated coefficients are given in Table 16. Assuming that a dollar spent on conservation of low value land adds as much to land value as does a dollar spent on high value land, 5 the coefficients This is implicitly assumed in the linear relationship over time. However, the variation in land values over time is much less than the Variation among areas. For price expectations the linear relationship Could hardly be extended over regions with widely varying land produc- tivity, but the conservation payments are expressed as dollars per land unit and an extension of the linear relationship seems reasonable. 129 Table 16. Regional Coefficients for Conservation Expenditures —__ L _— r i Estimated The main conservation Region Coefficient practices carried out Southeast (excl. Fla. ) 1. 041 establishing permanent cover Appalachia . 95 establishment and improvement of cover; drainage Northeast 1 . 42 Northeast 2 1. 39 estabhshment and improvement of permanent cover, drainage Northeast 3 . 57 ' Lake States . 84 establishment of permanent ‘ cover; drainage Corn Belt . 37 lime, phosphate, drainage Delta States 1. 28 establishment and improvement of permanent cover; drainage Southern Plains l. 35 controlling shrubs; permanent cover Northern Plains 1. 86 water conservation and disposal; temporary cover Mountain 1 1. 62 water conservation and disposal Mountain 2 l. 94 water conservation and disposal PFicific Z. 52 water conservation and disposal \ _— All the coefficients are significantly different from zero at the one Percent level. 130 are directly comparable. The coefficients indicate that a dollar spent on water conservation in the Pacific, Mountain or Northern Plains region increased land values more than if spent on the typical conser- vation practices in the eastern and southern states. For example, 10 cents spent on conservation increases land values in the Northern Plains by $1. 86, while in the Southeast it increases land values by $1. 04. Apart from the suggestion that water conservation increases land values more than other conservation practices, no general conclusions about the conglomerate of conservation practices can be derived from Table 16. The high coefficient in Northeast 2, as compared with other northeastern regions, does not seem to have any apparent explanation. Most of the coefficients are much higher than what can reasonably be expected to be the influence of government subsidized conservation practices on land values. If land values increased with the full amount of the conservation investments and the government subsidy amounts to 50%, we would expect the coefficient to be . 2. 6 Since most of the estimated coefficients are much above . 2, some other explanation is necessary. It is likely that the MVP of expenditures for conservation practices is somewhat above the cost of conservation, and thereby 6The conservation variable is entered as $5. l/pasture acre equivalent. If the value of land increases with the amount of govern- ment expenditures on conservation, the coefficient'should be .1. 131 increases the estimated coefficients. Another, and probably more important reason for the high coefficient, is that the amount of conservation carried out with government funds is highly correlated with the total investment in agricultural conservation practices. Also, as has been mentioned previously, there was generally a high correlation between the conservation variable and the output per man-hour variable. Since, for this reason, the output per man-hour variable was in most cases omitted from the final analyses, the estimated coefficient for conservation has most likely measured some of the impact of technological change. Also in some analyses, intercorrelation between conservation expenditures and the constant term could lead to overestimated conservation coefficients. In the Corn Belt high negative constant terms were generally associated with high coefficients for the conservation variable. However, in the regional comparisons there was little correlation between coefficients for conservation expenditures and constant terms. The effects of eliminating the constant term were tested, and most of the coefficients, including the conservation coefficients were depressed somewhat. However, simultaneously the coefficients of multiple correlation fell, the significance of the coefficients fell, and the sum of the residuals was negative. In summary, there is :no doubt about the strong relationship between conservation payments and land values. However, the conser- vation coefficients might be biased upward because they estimate effects of conservation expenditures not measured directly by the variable, and because they might to some degree estimate effects of technological Change. 132 Soil Bank Variables The conservation reserve part of the soil-bank program has, as indicated by the estimated coefficients, had the major impact on land values. The coefficients for the conservation reserve part of the program were generally much larger than those for the acreage reserve payments. This could be due partly to the longer term contracts for the conservation reserve program, the different nature of the programs, or it could be because conservation reserve program payments, relative to opportunity costs, were greater than acreage reserve program payments. Some of the difference in performance of the soil bank variables might be associated with specification problems. The specification of both soil bank variables is based on the assumption that the rates paid per acre and the number of acres contracted for the soil bank influence land values. An alternative specification would be to let land values be a function of the level of average rates paid. The underlying assumption would be that the MVP of all land changes proportionately with the change in average rates paid. A third way to enter the soil bank variables would be to specify the variables as being the contracted acreages. This specification, whichwas tried in some initial analyses, explained less of the land value variation than did the specification used throughout the study. The chosen specification is a combination of contracted acreages and rates paid per acre. This specification seems superior to using '5‘}! III.“ 133 the contracted acreage alone, but has not beencomparatively tested with rates paid per acre. Conservation Reserve--The regional estimates of the soil bank variables are compared in Table 17. The conservation reserve coefficient for the Corn Belt is more than double those for other regions. The data show that the Corn Belt had a large participation in the program and the rates paid per acre were substantially above the national average rates. 7 High rates per acre would be expected in the Corn Belt to attract high value land into the program. However, an understanding of the high coefficient would seem to require that the difference between rates paid and opportunity costs must have been higher in the Corn Belt than in other regions. The Northern Plains also had relatively high participation in the program, but the estimated coefficient is low, apparently indicating a much smaller difference between conservation reserve payments and opportunity costs for the land input. Low, but less significant, coefficients were also obtained in the Northeast and the Pacific regions. The negative coefficient for Mountain 2 must be attributed to spurious correlation in connection with the very low participation in the program. Except for the Corn Belt and the Northern Plains the positive coefficients significant at the . 01 level, were in the range 2. O to 3. 0 indicating a rather stable change 7In 1960 the average rates paid in the Corn Belt were 17-18 dollars per acre. The U. S. average was 11. 85 per acre. 134 Table 17. Regional Coefficients for Soil Bank Variables Estimated Coefficients: Main cr0p Region Conservation Acreage allotments Reserve Reserve reduced by the acreage reserve program Southeast (excl. Fla. ) 2.15*** . 06>!< cotton Appalachia 2. 46*** . 28’50-‘0':< tobacco, corn Northeast 1 . 72* . 29** tobacco, corn wheat Northeast 2 . 85 -- tobacco, corn Northeast 3 . 71 ’1‘ -- tobacco Lake States 2. 10*“< . 63*** corn, wheat Corn Belt 6. 06*** . 80*** corn Delta States 2. 85*** . 12* cotton, rice Southern Plains 2. l 8*** -. 07 cotton, wheat Northern Plains 1. 34*** -- wheat Mountain 1 2. 56*** -. 22** wheat Mountain 2 -3. 35*** l. 03 cotton Pacific . 57 -. 26 cotton, rice, wheat ** ** :4: Significantly different from zero at the one percent level. Significantly different from zero at the five percent level. J: Significantly different from zero at the ten percent level. 135 in land values (among regions) for a dollar spent in the conservation reserve program. For example, a coefficient on 2. 5 would mean that land values increased 25 dollars per dollar change in conservation reserve payments. The stability implies that land value increases due to the conservation reserve program are almost pr0portionate to the payments received by states or regions. Acreafi Reserve-~The payments of the one year contracts for land in the acreage reserve have had less impact on land values. Since the program was in effect three years only, it turned out to be wise not to capitalize much of its benefits into land values. The coefficients do suggest that the payments made for reduction of corn allotments were high relative to alternative opportunities for land, since they had the largest significant influence on land values. Efficiency and Input Variable s Outpgt per man-hour--The output per man-hour variable was included in the analyses because it was expected to provide a better measure of technological advances in agriculture than would the usual forms of a time variable. As pointed out earlier, the variable measures changes in efficiency associated with increased output per acre and with the use of labor-saving machinery. The output per man-hour variable was generally highly correlated with land values, but the variable was in several instances deleted from the final analyses because of intercorrelation problems with conservation expenditures. Where the output per man-hour 8Some losses have also occurred due to decline in payments during the latter years. 136 variable was deleted, either fertilizer use or output per acre was included. The variable which caused least multicorrelation problems and was most significant was retained. Generally, the substitute variables had a lower simple correlation with land values than did output per man-hour. Table 18, column 1, gives the output per man-hour coefficients for the regions in which the variable was retained. For comparison among regions, the coefficients are adjusted (column 2) in a manner similar to that for the price expectations variable. The adjusted coefficients suggests that a change in the index of output per man-hour of one index point would have relative more effect on land values in the Mountain 2 states and less in the Lake States. The enterprises with heaviest weights in the index for Mountain 2 are beef and cotton, while in the Lake States index dairy and feed grains weight heaviest. This could imply that efficiency changes in the beef and cotton production have relatively larger effects on land values than do efficiency changes in dairy and feed grain production. Outputper acre--The estimated and adjusted coefficients for the output per acre variable are given in Table 18, columns 3 and 4. The variable was retained in five regions and was significant at the . 05 level for the Southeast, Appalachia, and Northeast 3. The variable fluctuated less violently in the eastern regions than in other parts of the country. Some of the fluctuation in the indices of output per acre was alleviated by using a three year moving average. The significant coefficients suggest that changes in output per acre have a relatively small effect on land values in Appalachia. The Table l 8 \\ Region Southeas (excl. Appalac} Northeas Northea: Northea; Lake St; Corn Be Delta St Souther: Norther MC’ul‘s'tai Mouhtai patific Table 18. 137 Output per Regional Coefficients for Efficiency and Input Variables W; 1 Output per Region man-hour coefficient acre coefficient Fertilizer Estimated Adjusted1 Estimated Adjustedl use . (1) (2) (3) (4) coeff1c1ent Southeast . 21 *** 1. 35 (excl. Fla. ) Appalachia . 22*** . 77 . 17*** Northeast 1 . 58**>1< 2. 66 Northeast 2 .16 . 32 Northeast 3 . 09*)? l. 09 Lake States . 29*"6'k l. 32 Corn Belt . 63*** Delta States . 06 . . 28 Southern Plains . 84*** Northern Plains . 84** Mountain 1 . 79** Mountain 2 . 1 0’10“< 3. 33 Pacific , 44*>:< 1The adjusted coefficients give the change per $100 land value (1940 values) brought about by a one index point change in the independent variables. expectations in Table 15. a»: ** The adjustments are similar to those used for price 1: Significantly different from zero at the one percent level. Significantly different from zero at the five percent level. fertilize but this the two ‘ I coefficie entered pasture The V‘arfi Corn Be level in l ”impact c Mountai: 7 varied g I- ~ hid 2: a1 138 fertilizer use variable was also included in the Appalachian analysis, but this has probably not affected the output per acre coefficient as the two variables had a low intercorrelation (-. 07). Fertilizer use--Table 18, column 5, gives the regional coefficients for the fertilizer use variable. Since the variable was entered in the analyses as tons of commercial fertilizer per thousand pasture acre equivalents no adjustment among regions was required. The variable was significant at the . 01 level in the Appalachian, the Corn Belt, and the Southern Plains. It was significant at the . 05 level in the Northern Plains, Mountain 1, and the Pacific. The coefficients indicate that fertilizer use has had the largest impact on land values in the Southern Plains, Northern Plains and Mountain 1 regions. Poplilation Density and Income The coefficients for population density and personal income varied greatly between regions (Table 19). Population density-J31 unit change in population density has had least influence in heavily populated states such as those in Northeast 1 and 2, and in the Corn Belt. The coefficients for less densely populated areas are higher. The highest coefficients were obtained for Florida and California, both states with large absolute and relative increases in population. The varying sizes of the coefficients might suggest that the relationship between land values and population density is nonlinear. However, since the differences in population density among regions are generally much larger than the differences occurring within regions during 139 Table 19. Regional Coefficients for Population Density and Income Region Estimated coefficients: Population Density Pe r s onal IncomeT Southeast (excl. Fla. ) Appalachia Northeast 1 Northeast 2 Northeast 3 Lake States Corn Belt Delta States Southern Plains Northern Plains Mountain 1 Mountain 2 .02 .31** .05 .08** .17 .21* .OOZ*** .28* 1.03*** -3, 04>:<>:< . 65** Pacific . 77:;<>:<>:< __ Florida 2.l7*** __ California 2. ll*>:<* __ l A two sided significance test is used for personal income. *** Significantly different from zero at the one percent level. ** Significantly different from zero at the five percent level. Significantly different from zero at the ten percent level. the inve does no not imp respect in the N 140 the investigated time period, the assumption of linearity within regions does not seem unreasonable. Personal income--In general, the personal income variable was not important in the analyses. However, as discussed under the respective regions in Chapter V, it was a rather important variable in the Northern Plains and in the Lake States. The income variable was in several instances deleted from the final analyses because of its intercorrelation with the price expectations variable. In such cases, the price expectation coefficients might include some of the effects due to changing income. As discussed earlier, income would probably have a positive effect on land values in areas with large urban communities. In other areas income increases in the non-farm economy might be associated with increased migration away from rural areas and, thus, have a negative effect on land values. It is possible that in some areas, the two opposite effects have offset each other and resulted in an insigni- ficant income coefficient. Constant Terms The constant term in an analysis of first differences estimates the changes in land values over time given no change in the other indepen- dent variables. The regional constant terms and adjusted constant terms are given in Table 20. The adjusted coefficients show the yearly change in land values per $100 of land value measured in 1940 dollars. All the significant coefficients are negative and range, when adjusted, from -3. 00 in the Delta States to -8. 75 in the Northern Plains. The reason for a Table 2( Region Southeas Appalaci Northeat Northea‘ Northea Lake Sta Corn Be Delta St Southerr Northerfl MOuntaii MOuntaiq Pacific \ 1.... ihe ad‘ Value (1 . . Signi ‘. 'P 141 Table 20. Regional Constant Terms (time) Region Constant term Adjusted Coefficient1 Southeast (excl. Fla. ) -. 87*** -‘5. 62 Appalachia -1. 59*** -5. 58 Northeast 1 -. 72*“< -3. 30 Northeast 2 -. 42 -. 84 Northeast 3 -. 37*“< -4. 55 Lake States -1 . 71 *** -7. 92 Corn Belt .14 .35 Delta States -. 67* -3. 00 Southern Plains -. 29 -2. 44 Northern Plains -1. 32* * * -8. 75 Mountain 1 -. 1 5*** -4. 50 Mountain 2 -. 1 9*** -6. 07 Pacific -. 99*** -8. 05 1The adjusted coefficients give the yearly change per $100 land value (1940 values) with no change in other independent variables. *>k* Significantly different from zero at the one percent level. ** Significantly different from zero at the five percent level. Significantly different from zero at the ten percent level. positive be due ti: Corn Be the Corr -- t individuz constant values 0 differ“c That diSt Value pe: The Seri Show ger fer the F 1853 and average . land usec increaSe regmn it toward t] land "alt land Corr simmer 142 positive and non-significant coefficient for the Corn Belt seems to be due to intercorrelation problems in the combined analyses of the Corn Belt. The constant terms in the individual state analyses of the Corn Belt were negative and generally significant. The constant terms measure the effect of variables which individually or combined have a constant effect on land values. The constant terms would probably estimate some of the effects on land values of erosion and soil depletion. Magnitude of Change in Variables In the introductory chapter some attention was given to the differences in relative land value increases among regions and states. That discussion was based on indices of average farm real estate value per acre as published in Farm Real Estate Market Deve10pments. The series of land value per pasture acre equivalent used in this study show generally the same trends as those indices. However, the series for the Pacific and the Northeastern regions give, respectively, much less and much more increase in land values than do the indices of average value per acre. This is due to the quantity measure of farm land used in this study, namely pasture acre equivalents. The large increases in irrigation of previously unirrigated land in the Pacific region increased the number of pasture acre equivalents rapidly toward the end of the time period and subsequently reduced the relative land value increase. In the Northeast, the decrease in acreage of the land component with the lowest per acre value (other land) was relatively 9 Smaller than the decreases in acreages of other land components. _— c)Se e Appendix A. This can relativel relative per pastt the indic '1 1925-62, in and re table. 10 143 This caused the series of pasture acre equivalents to decrease relatively faster than the acres of farm land. Subsequently, the relative increase in land value, as portrayed by the series of value per pasture acre equivalent, is larger than the relative increase in the indices of average value per acre. The relative increase in the value of a pasture acre equivalent 1925-62, is given by region in Table 21, column 1. Relative changes in and relative importance of related variables are also given in the table. 10 The relationship between land values and the included variables was discussed in Chapter V. The estimated coefficients varied for different parts of the country, indicating that even equal long run changes in the independent variableswould lead to differences in land value increases. However, Table 20 shows that there are rather large quantitative variations among regions in the changes in or the size of a given variable. The regions are arranged in two groups according to relative land value increases. Generally the largest numerical values of column 2 through 6 are found in the upper five rows which contain the regions with largest land value increases. A crude test of this proposition is supplied by‘ the sum of ranks. The average sum of ranks for the upper five rows is considerably lower 10The output per man-hour variable is, as the most successful, taken to represent the three efficiency and input variables. The acreage reserve payments are not included since the program is terminated and its total effect in the analysis therefore is neutral. Personal income was excluded because it was not found to be important in this study. ‘Fn——r\ufiu¢.\ u-A— . .73 ”mXCde m3 ..CZTJ a: uC‘JCZ‘ZQV newswhvxjox Ki umV$~\N03~ mCCRE NCA HImNA u IRAs—IE \A m. a t,. t . . . z 2 m. ...3.-_H3> 3.193790. H0 mem. EXSHQ>< 0:1 p: $71.33: 3 U 0 . . at N h. H>.~..~NW~.U~N .NN GUNNINQN 144 .womdouocw moan.» puma Headache 5:» .356 on: was .momdouocw 93d». puma swung €53 oco .mcowwou mo masonw 025 HOH exam." mo 55m omdno>< .033“ @3550 who mxcdu .305 .momhdmcm coammouwou 05 5 pocmfinoumo mm .3438.an cofimczomcoo was mcoflmuoomxo scrum mo meson—nomad newnma on: o» osfl .ohsmfl amok/0H cc» 0» OH 352 was umomnmfi 04.3 8. H Mama m5>ww oH Ow 5.80.3, poxcmn mm? GEsHoo comm .o amdoflfi N 48530 mo mxcmn ecu mo Esm goofing owe: 35m.» pddfi m0 003 mom Saccfiuoommou .mucoE>mm grammes cowumc’uomcoo pad mohsuwpcoaxo cowud>uomcoo Ho ”—5588 owduo>< .. w c8530 :5 GEDHOU .. o was m 4.85500 “moqu o .ov wwo . mm . NH 4 0H .m mo 4 on. .H mcwmfinm Gavan—H02 ofim wmo. we. we .H 906 we; owé men—mum 934 “.34 o .mm nmo . em . mv In ww .N mo 4 vm .N 30m ch00 mév mmo. co. 5‘; NH .m ow; wed ammocuaoz 06m .20. mm. 2 .m mod ooé No.m won—mum 330mm m .Nm 4.8. 2.. 3. 4 Ne .N as 4 E .m emfiomfiga. m .om vmo . om . Ma .N op .N em .4 3‘ .m GMSGDOE o .Hm m Am owe . om . No 4 NH .m co 4 om .m 83.3% Gnocusom o Sm $0 . No A on .4 3 .m mm .H mp .w moumam SEQ o .2 m2. mm 4 a... 4 mm .m S 4 2. .m 72m 483; sawmnasom a: E e: 5 a; E E E «crammed mousfip H50: omdno>< cowmom Snomcou inseam G03 Sofia: ncda Mom .uoomxm ooflm> cowmom QSOHO am am owcdco nm>nomcoo ismom 39.30 063m 984 "Baum mo Esm $3 «5554 smouocfio‘ umwswmoofi moflmm Noodnmmou .cowwom >0. mozmwhmxr wouooaom mo exam omdnocfiw was Gm mowfionU o>fid~om AN @139 fimnthe merease have be: South E increas ranksi behnee the De expect notrn Vahae hOur Subs; Cons expe Wit} larg this 145 than the average sum for the group with lower relative land value increases. This implies that the relative changes in the variables have been largest for the group with large land value changes. For individual regions, the very low sum of ranks for the South East indicates that this region has had rather large relative increases in all the variables. Some of the variation in the sum of ranks is due to the inability of the ranking method to distinguish between large and small changes. Thus, the high rank sum for the Delta region is much influenced by its high rank for price expectations. However, the increase in its price expectations is not much below the U. S. average. The two regions with the largest relative changes in land values also had the largest relative changes in the output per man- hour variable and received the largest government conservation subsidies. Both regions also have high participation in the conservation reserve program. The largest changes in price expectations occurred in the Appalachian and the Mountain regions. With the exception of Florida, relative population changes were largest in the Pacific region, but other variables had high ranks in this region. The relatively low sums of ranks for the Lake States and the Northern Plains need some explanation. In the Lake States the land values have not increased as much as in other regions, with a similar sum of ranks, because the price expectation and conservation variables had relatively low coefficients (as statistically related to land values). In the Northern Plains the stagnating population density and the had darr by regic and spe to the c3 price e) exPectat in the v; 1 acre eq1 the tren $123. 6, the BSti The inc price E factors, ReServ Output 1 interpr I The M0 mCrEaS 146 and the negative relationship between land values and income have had dampening effects on land value increases. Sources of Increased Land Values Table 22 lists the estimated sources of land value increases by region. The estimated changes in land values due to trend factor and specified factors are obtained by applying the regional coefficients to the changes in the variables. For example, in the Southeast, the price expectation coefficient was . 205 and the increase in the price expectation index was 41. 7 index points resulting in a $8. 6 increase in the value of a pasture acre equivalent due to price expectations. In the Southeast, the actual increase in value per pasture acre equivalent amounted to $91. 5. The estimated change due to the trend factor and the specified factors amounted to $-32. 2 and $123. 6, respectively. Due to errors in the estimates, the sum of the estimated changes deviates slightly from the actual change. The increase due to specified factors is distributed as follows-- Price Expectations: $8. 6 or 7. 0% of total change due to specified factors, Conservation Expenditures: $82.4 or 66. 6%, Conservation Reserve: $16. 4 or 13. 3%, Population Density: $. 4 or . 3%, and Output per Acre: $15. 9 or 12. 9%. Data for other regions are interpreted similarly. The actual land value increases per pasture acre equivalent were largest in Northeast 2, the Corn Belt, and the Delta States. The Mountain States and Northeast 3 had the smallest land value increases. —l 147 Table 22. Estimated Sources of Land Value Increases by Region, 1925-1962 ? Estimated Chan e ‘ Change Actual Trend Specified ($ figure on Region chan e Factor Factors Price Conserv. $/pae $/pae $/pae Z Expect. Expend. Southeast (Excl. Fla.) 91.5 -32.2 123.6 8.6 82.4 7. 0 66. 6 Appalachia 82.1 -58. 5 140. 8 ll. 6 94. 3 8. 2 67. 0 Northeast 1 66.6 -26.8 93.6 5.5 30. 5 5. 8 32. 6 Northeast 2 145. 1 -15. 4 160. 6 6. 8 120. 5 4. 2 75. 0 Northeast 3 23. 2 -l3. 3 36. 8 2. 3 28. 5 6. 3 77. 4 Lake States 40.5 -63. 3 103.6 6.5 43. 5 6. 3 42. 0 Corn Belt 123. 9 5.1 116.1 19. 7 36.1 17. 0 31. 1 Delta States 103. 7 -25.1 128.4 8.0 104.4 6. 2 81. 3 Southern Plains 43. 3 -10. 7 53. 9 6. 7 35. 4 12. 4 65. 7 Northern Plains 29. 7 -48. 5 78. 2 9. 5 61. 2 12.1 78. 3 Mountainl 11.1 - 5.5 16.5 1.7 11.9 10. 3 72.1 Mountain 2 15.3 - 7.0 22.4 .4a 10. 7 l. 8 47. 8 Pacific 33.0 -36. 7 69.5 4.8 31.9 6. 9 45. 9 1pae = pasture acre equivalent. zEstimated change from specified factors = actual change - trend factor - residual. 8’The coefficient used to estimate this figure was not significant at the . 05 level. 148 attributal to specified factors. top, ‘70 of total change due to specified factors on bottom) Conserv. POpulation Personal Output per Output per Fertilizer Reserve Density Income Man-Hour Acre Use 16.4 .4a 15.9 13.3 .3 12.9 8.7 8.1 11.6 6.6 6.2 5.8 8.2 4.7 2.5a 5.6a - .4a 49.8 2.7 6.0 - .4 53.2 1.3a 18.4 6.2a 7.4a .8 11.5 3.9 4.6 1.3a 1.7a —1.0 4.0 3.5 4.6 -2.7 10.8 9.9 7.2a 11.9 24.7 9.6 6.9 11.5 23.8 29.8 .1 30.4 25.7 .1 26.2 10.1 9.5 -6.2a 2.5a 7.9 7.4 -4.8 1.9 4.8 3.1a 4.0a 8.9 5.8 7.4 6.4 1.8a -5.4 4.7 8.2 2.3 —6.9 6.0 1.1 1.1 .7 6. 6 6. 6 4. 2 -1.2 5.9 6.6 -5.4 26.3 29 5 .3a 28.7 3.8 .4 41.3 5.5 A and perc This is r high lanc' expectati mainly d C associat the incre exceed t existed Values d~ Ontput pg the belie impact (5 That the Belt Whi is due tc expendit- 149 Among regions the price expectations had the largest absolute and percentage influence on land value increases in the Corn Belt. This is mainly due to the high coefficient which seems related to high land productivity. The relatively large influences of price expectations on land values in the Southern and Northern Plains is mainly due to the high price expectation coefficients for these regions. Generally, the major source of land value increases is associated with the conservation variable. Only in Northeast 1 did the increase in land values due to the output per man-hour variable exceed the increase due to conservation. Some intercorrelation existed between these variables and the percentage of increased land values due to conservation seems to be depressed somewhat when the output per man-hour is retained in the analyses. This substantiates the belief that the conservation coefficients measures some of the impact due to efficiency changes. As discussed earlier, it appears that the relatively small part of land value increases in the Corn Belt which are explained by the conservation expenditure variable is due to intercorrelation problems between the conservation expenditure variable and the constant term. The conservation reserve program has, according to the estimates, had substantial impact on land value increases in most of the regions. The largest influence was in the Corn Belt and in the Southeast. The conservation reserve had little impact on land value increases in the Northeastern and Pacific regions. The negative impact estimated for Mountain 2 is probably due to spurious correlation in connection with little participation in the program. l inmeIJ for, and east 2 h. populati large irl value 1! Plains. hav e h Output hay e f impat vfilm Cans With Con: and Pr- \3:} 150 Increased population density had a major impact on land values in the Pacific region. 11 The Pacific region had a large coefficient for, and a large change in population density. Mountain 2 and North- east 2 had lower coefficients for population density, but with the large population increases, population changes are estimated to have caused large increases in land values. Personal income is estimated to have caused substantial land value increases in the Lake States and some decreases in the Northern Plains. Output per man-hour is estimated, where it was retained, to have had a large impact on land value increases. The changes in output per acre and fertilizer use variables are also estimated to have increased land values. The fertilizer use variable had a large impact in the Corn Belt. In summary then, the analysis of sources of increased land values suggests that a large part of the increase which occurred was caused by conservation expenditures or factors strongly correlated with the conglomerate of expenditures defined as being related to conservation. The change in land values associated with the efficiency and input variables was substantial. Price expectations also appeared to cause land value increases in all regions. The conservation reserve program, pOpulation density and personal income were other variables which had a definite impact on land values in particular regions. 11The impact of population density on land values in Florida has been pointed out earlier. Florida was excluded from the combined analysis of the Southeast. CHAPTER VII CONCLUSIONS AND IMPLICATIONS There were two major objectives of this thesis. One was to delineate factors, including government programs, which have caused a major part of the variation of and relative increase in land values. Another objective was to investigate the causes of regional variation in land value increases. The difference model of first order errployedin the analyses generally yielded coefficient signs consistent with the usual economic expectations. Also, multicollinearity and serial correlation problems were reduced by using first differences. With respect to the use of original data, initial analyses indicated that the strong trends in these data would produce many "wrong" signs. Combined time series and cross-sectional analyses for regions increased the significance of the estimated coefficients as compared with individual state time series analyses. However, inclusion of cross-sectional variation generally decreased the coefficient of multiple determination indicating some heterogeneity among the cross- sectionally combined states. Time aggregation of first differences over two year periods gave large increases in the coefficients of multiple determination, indicating considerable year to year random variation in the data. It is argued that the random variation is likely to originate mainly in the dependent variable. Time aggregation I produced increased intercorrelation among the independent variables. 151 152 Even though the combined analyses of data with time aggregation yielded higher multiple correlation coefficients than the analyses of data based on yearly observations, the results of the latter analyses are believed to be more reliable on the regional level. In several instances, time aggregation caused increased multicollinearity which led to less significant coefficients. In analyses where intercorrelation did not increase appreciably with increased time aggregation, as in the Southeast, the estimated coefficients did not show much change from those based on yearly observations. The indices of expected prices were among the most impor- tant variables in the land value analyses. The commodity price expectation indices, from which the aggregate state indices were derived, were estimated in another part1 of the Resources for the Future, Inc. project. The price expectation indices were mainly important in explaining the yearly variation, while they had relatively little to do with the increase in land values. Very large increases in land values have occurred since the early 1950's, a period in which the price expectations have shown slight downward changes. The largest estimated impact of price expectations on land values was found in the Corn Belt. Government expenditures on conservation were highly correlated with land value changes. Conservation expenditures 1M. Lerohl, op cit. 153 were found to be associated with a substantial part of the increase in land values. The variable also seems to be of some importance in explaining the short-run variations in land values. Conservation expenditures were cut substantially in 1948 and in 1953-54; the following years gave decreases or small increases in land values. Part of the estimated effects attributed to conservation expenditures were undoubtedly due to technological changes. This is implied by rather high correlations between conservation expenditures and output per man-hour. The estimated impact of conservation expenditures on increased land values was largest, in absolute terms in Northeast 2 and Delta regions. However, in the Northern Plains and Northeast 3, about 78 percent of the increase in land values due to specified sources, was associated with the conservation variable. Thus, also in regions with comparatively low land value increases a large part of the increases were associated with conservation expenditures. Of the soil bank variables, the conservation reserve part was the most important in the land value analyses. Coefficients for the conservation reserve variable were larger and generally more significant than those for the acreage reserve variable. Since the acreage reserve program was terminated in 1958, its impact on land value increases over the examined period was zero. Large payments are still being made through the conservation reserve program, and the estimates suggest that these payments have been a substantial source of increased land values. The largest increases in land values 154 associated with the conservation reserve variable occurred in the Corn Belt, the Southeast and the Lake States. The output per man-hour variable had a high simple correlation with land values. Due to intercorrelation problems, this variable was deleted in many cases and fertilizer use or output per acre substituted. There is little doubt that large efficiency changes (technology) have caused some of the relative increases in land values. Population density was significant in many of the combined cross-sectional and time series analyses. The largest coefficients of the population density variable were in areas with large relative increases in population, such as Florida and California. A consider- able part of land value increases in Mountain 2 and Northeast 2 were also explained by population increases. Personal income has had increasing effects on land values in the Lake States and in Mountain 1, but led to decreased land values in the Northern Plains. The large differences in relative land value increases among regions appear to be caused in large part by the distribution of government program payments. Agricultural conservation payments have been shown to be strongly correlated with land values, and the distribution of these payments as well as the conservation reserve payments have had substantial impact on regional differences in land value increases. Relative changes in price expectations and indices of output per man-hour were also important in explaining the regional differences in land value increases. The price expectation series are greatly influenced by government regulations and subsidies, and govern- ment pro: in price 6 increases to influer A governm changes, payment varying capital ] waSter: COQSQIV Values Practic r9138 at Subsi d Which Produc affeCt the di 155 ment programs have caused some of the large regional differences in price expectation increases. For example, the largest relative increases in price expectations occurred in the Appalachians due to influence of tobacco price expectations. Although some of the impact on land values measured by the government conservation payments might be due to other technological changes, there is little doubt about the importance of conservation payments, soil-bank payments, and price subsidies as causes of varying real estate capital gains among regions. Similarly, the capital losses which occurred when the acreage reserve program was terminated and the losses occurring through liquidation of the conservation reserve program differ widely among regions. The importance of conservation practices as related to land values implies that large increases in productivity are gained through practices defined as conservation. Therefore, we must once more repeat the argument that a main part of the government conservation subsidy program is in direct conflict with other government programs which are aimed at a stabilized or reduced supply of agricultural products. In decisions concerning government agricultural programs affecting the value of real estate, policy makers should be aware of the differential capital gains among regions due to the distribution of program benefits. Of at least as much importance are the differential regional impacts of capital losses occurring with the termination of agricultural programs. Bot 80: Ch E2 F1 :4; BIBLIOGRAPHY Books, Bulletins, Thes es Boyne, David H. Changes in the Real Wealth Position of Farm Operators, 1940-1960. Technical Bulletin 294. Michigan State University, Agricultural Experiment Station, 1964. Boxley, R. F., Jr. and Gibson, W. L. Jr., Peanut Acreage Allotments and Farm Land Values, Technical Bulletin 175, Virginia Polytechnic Institute, 1964. Chennareddy, Venkareddy. Present Values of Expected Future Income Streams and their Relevance to the Mobility of Farm Workers to the Non-Farm Sector in the United States, 1917-62. Unpublished Ph. D. dissertation, Michigan State University, 1965. Ezekiel, M. and Fox, K. A. Methods of Correlation and Regression Analyses. New York: John Wiley and Sons, Inc. , Third Edition, 1963. Fraser, D. A. S. Statistics: An Introduction. New York: John Wiley and Sons, Inc., 1960. Hathaway, Dale E. Government and Agriculture. New York: The Macmillan Co. , 1963. Hildreth, C. and Lu, J. Y. Demand Relations With Autocorrelated Disturbances. Technical Bulletin 276, Michigan State University, Agricultural Experiment Station, 1960. Hoover, Dale M. A Study of Land Prices in the U.S., 1911-1958. Unpublished Ph. D. dissertation, University of Chicago, 1961. Johnson, D. G. Forward Prices for Agriculture. Chicago: University of Chicago Press, 1947. Johnson, G. L. et a1. (edit. ). A Study of Managerial Processes of Midwestern Farmers. The Iowa State Press, Ames, Iowa, 1961. Johnston, J. Econometric Methods. New York: McGraw-Hill Book Company, Inc., 1963. Jones, Bob F. Farm-Non-Farm Labor Flow, 1917-62. Unpublished Ph. D. dissertation, Michigan State University, 1964. 156 157 Lerohl, Milburn L. Expected Prices for U. 3. Agricultural CommoditiesJ 1917-1962. Unpublished Ph. D. dissertation, Michigan State University, 1965. Maier, F. H., Hedrick, J. L., and Gibson, W. L. Jr., The Sale Value of Flue -Cured Tobacco Allotmentg, Technical Bulletin No. 148, Virginia Polytechnic Institute, 1960. Nerlove, Marc. The Dynamics of Supply: Estimation of Farmers Response to Price. Baltimore: The John Hopkins Press, 1958. Petit, Michel J. Econometric Analyses of the Feed-Grain Livpstock Economy. Unpublished Ph. D. dissertation, Michigan State University, 1964. Report of the Governor's Study Commission on Agriculture. Minnesota. St. Paul: Office of the Governor, 1958. Robinson, Joan. The Economics of Imperfect Competition. London: Mac Millan Co. , 1961. Rossmiller, George E. Farm Real Estate Value Patterns in the United States, 1930-1962. Unpublished Ph. D. dissertation, Michigan State University, 1965. Theil, H. Linear Aggregation of Economic RelationshiLs. Amsterdam: North Holland Publishing Co. , 1964. Journal Articles Bean, Louis H. "Inflation and Price of Land, " Journal of Farm Economics, 20 (February, 1938) pp. 310—320. Chambers, C. R. "Relation of Farm Land Income to Farm Land Value, " American Economics Review, 34 (December, 1924) pp. 673-698. Cochran, D. and Orcutt, G. H. "Application of Least Squares Regression to Relationships Containing Autocorrelated Error Terms, " Journal of the American Statistical Association, 44 (March, 1949) pp. 36-61. Durbin, J. and Watson, G. S. "Testing for Serial Correlation in Least Squares Regression, Part II, " Biometrica, 38 (1951), pp.159-l77. Grundfield, Y. and Griliches, Zvi, ”Is Aggregation Necessarily Bad?" The Review of Economics and Statistics, 42 (February, 1960) pp. 1—10. Ibach, D. B. ”Role of Soil Depletion in Land Valuation, " Journal of Farm Economics, 22 (1940) pp. 460-472. 158 Johnson, G. L. Book Review of Marc Nerlove, “The Dynamics of Supply: Estimation of Farmers' Response to Price, " Agri- cultural Economics Research, 12 (January, 1960) pp. 25-28. Larsen, Harcold C. "The Relationship of Land Values to Warranted Values, 1910-1948, " Journal of Farm Economics, 30 (August, 1948) pp. 579-588. Lave, Lester B. "Technical Change in U.S. Agriculture: The Aggregation Problem, " Journal of Farm Economics, 46 (February, 1964) pp. 200-2170 Renshaw, Edward F. "Are Land Prices Too High: A Note on Behavior in the Land Market, " Journal of Farm Economics, 39 (May, 1957) pp. 505-510. Scharlach, Wesley C. and Schuh, G. Edward. "The Land Market as a Link Between the Rural and Urban Sectors of the Economy, " Journal of Farm Economics, 44 (August, 1962) pp. 1406-1411. Scofield, William H. "Investment in Farm Real Estate, " Journal of Farm Economics, 45 (May, 1963) pp. 396-405. . "Prevailing Land Market Forces, " Journal of Farm Economics, 39 (December, 1957) pp. 1500-1510. Scott, Anthony. "Conservation Policy and Capital Theory, " Canadian Journal of Economics and Political Science, 20 (November, 1954) pp. 504-513. Theil, H. and Nagar, A. L. "Testing the Independence of Regression Disturbances," Journal of the American Statistical AssociationJ 56 (December, 1961) pp. 793-805. Thomsen, F. L. "Factors Affecting Farm Real Estate Values in the United States, " Journal of Farm Economics, 20 (February, 1938) pp. 310-320. Government Publications Foote, Richard F. Analytical Tools for Studying Demand and Price Structures. Agricultural Handbook No. 146, ERS, U. 5. Depart- ment of Agriculture, Washington: U. S. Government Printing Office, 1958. United States Bureau of Census. Census of Agriculture, 1910, 1920, 1925, 1930, 1935, 1940, 1945, 1950, 1954, and 1959. 159 Historical Statistics of the U. S. , Colonial Times to 1957. Washington: U. S. Government Printing Office, 1960. United States Department of Agriculture. Agricultural Conservation Program: Summary by States 1961, Agricultural Stabilization and Conservation Service, U. S. D. A. Washington: U. S. Govern- ment Printing Office, 1964. Agricultural Conservation Program: 25 Years Summary, 1936 through 1960, Agricultural Stabilization and Conservation Service, U. S. D. A. , Washington, D. C. Changes in Farm Production and Efficiency: A Summary Report 1964. Statistical Bulletin No. 233. Washington: U. S. Government Printing Office, 1964. Conservation Reserve Program and Land Use Adjustment Program. Agricultural Stabilization and Conservation Service, U. S. D. A. Washington: U.S. Government Printing Office, 1964. ERS, Farm Real Estate Market Developments, Series CD-24 through CD-66, 1949-1964. "Land Values and Farm Finance, " Major Statistical Series of the U. S. D. A. , 2, Agricultural Handbook No. 118, Washington: U. S. Government Printing Office, 1957. . Statistics on Fertilizer and Liming Materials in the United States. Statistical Bulletin No. 191, U.S. D. A. Washington: U. S. Government Printing Office, 1957. The Soil Bank Program: How it Operates, How it will Help Farmers. Washington: U.S. Government Printing Office, 1956. United States Department of Commerce. Personal Income by States Since 1929. Office of Business Economics. Washington: U.S. Govern ment Printing Office, 1956. . Statistical Abstract of the United States. Washington: U. S. Government Printing Office, 1933 through 1963. Survey of Current Business, 44 (August, 1964). 160 Cotner, Melvin L. “The Impact of the Agricultural Conservation Assistance Program in Selected Farm Policy Problem Areas. " Unpublished manuscript, Michigan State University, 1962. Heifner, Richard G. "A Note on the Relationship Between Coefficients of Determination for Regressions Computed in First Differences and in Original Values. " Unpublished paper, Michigan State University. Johnson, Glenn L. “An Evaluation of U. S. Agricultural Programs, 1956 to 1960. " Report for the Committee on Economic Develop- ment. East Lansing: Mimeo. , 1961. Scofield, William H. "Dominant Forces and Emerging Trends in the Farm Real Estate Market. " Paper prepared for seminar on land prices, North Central Regional Land Economics Committee, Chicago, Illinois, November 12, 1964. United States Department of Agriculture. Photostats of U. S. D. A. work- sheets on farm real estate made available by W. H. Scofield, Leader, Farm Real Estate Group, Farm Production Economics Division, ERS, U. S. D. A. . Photostats of U. S. D.A. worksheets on funds obligated for the acreage reserve program made available by the Conservation and Land Use Division, Agricultural Stabilization and Conservation Service, U. S. D. A. APPENDIX A PASTURE ACRE EQUIVALENT UNITS In order to obtain the value per unit of farm land, series of pasture acre equivalents are established. The pasture acre equivalent series quantify the larger changes in land quality. The quality changes are brough about by reproducible capital investments in land. Five categories of land are defined: Z1 2 Acres of pasture land not irrigated. zz = Acres of cropland irrigated. 23 2 Acres of pasture irrigated. 24 = Acres of cropland not irrigated. Z5 2 Acres of other land. Pasture acre equivalents for year t in state s (paets), is then: 5 : E paets 1:1 ris zist where ris = For state 5: Price per acre of land in category zi/Price per acre of pasture land. The method was suggested by Hoover, who worked out the ris coefficients using price data supplied by the USDA.1 Hoover considered three time periods, 1929-31, 1939-41, and 1949-51, but found little Variance between the periods. The period 1939-41 was used to compute 1D. Hoover, Appendix A. 161 162 the ris's. Since 1939-41 is close to the center of the period studied here, these coefficients are considered suitable for the entire period. The five categories of land are obtained from census data. However, in earlier censuses, agricultural land was not listed in the given five categories. A guide for reconcilliation of census data was given by Hoover2 and has been used in this study. Data on categories of agricultural land are available for census years only. For years between censuses, linear interpolation has been used to get yearly numbers of pasture acre equivalents. '5’ id. , Table 4. APPENDIX B SERIES OF LAND VALUES BE STATES The land value of a pasture acre equivalent in state s is found by, for each year, dividing the number of pasture acre equivalents for state s into total value of farm land in state s. The series of pasture acre equivalents were derived as explained in Appendix A. Value of farm land is found by subtracting value of farm buildings from value of farm real estate. Details about the value of farm real estate and farm building series were given previously.1 Table 23 gives the state series of land value per pasture acre equivalent, 1925-1962. 1 Supra, pp. 47 - 51. 163 164 Table 23. Value per Pasture Acre Equivalent by States 1925-62 New Massa- Rhode Con- Year Maine Hampshire Vermont chusetts Island necticut 1925 19.28 11. 51 11.16 40.40 50.45 45.95 1926 20.08 11.86 11.29 42.00 53.92 50.66 1927 20.19 11.91 11.21 42.04 58.43 55.86 1928 20.68 12.04 10.77 43.20 61.53 61.19 1929 20. 85 12. 09 11. 20 44. 31 64. 20 66.12 1930 21.03 11.60 11.21 43.20 64.32 64.88 1931 20.84 11.46 11.02 43.11 63.95 66.37 1932 18. 87 10. 55 10. 25 40.17 64. 03 64. 08 1933 16.21 9.10 9.28 37. 87 56.11 60.99 1934 15.99 9.36 9.15 37.76 55.83 60.98 1935 16. 20 9. 52 9.27 38.35 59.51 66.73 1936 15.55 9.31 9.01 37.31 59.93 65.75 1937 15.24 9.16 8. 85 36.60 60.59 64.65 1938 14.39 8.74 8.43 36.05 59.33 61.55 1939 13.51 8.44 8.16 34.85 58.31 59.18 1940 12.97 8.24 7.94 33.90 53.20 54.79 1941 13.38 9.23 8.18 36.15 56.25 58.51 1942 14. 89 10. 45 8. 46 38. 62 61.10 63. 06 1943 16.43 11.91 9.48 42.01 64.05 65.75 1944 20.18 14. 09 10. 55 49. 07 72. 86 73. 08 1945 22. 89 17. 89 12. 40 57. 59 84. 48 81. 67 1946 25.96 19.34 13.78 61.58 94.22 91.62 1947 29.04 22.16 15.84 63.01 104.66 104.57 1948 30.74 23.61 16.28 67.11 110.13 108.71 1949 33.78 24.53 17.53 69.30 114.04 111.01 1950 32.15 23.00 16.00 64.20 107.25 105.94 1951 37.66 23.71 17.11 68.55 122.90 114.04 1952 37.74 24.43 18.47 77.50 131.80 118.12 1953 35.34 24.60 18.67 80.25 135.21 120.56 1954 34.92 23.68 17.98 79.16 136.14 118.70 1955 33.09 23.27 17.81 82.02 145.60 123.27 1956 33.97 23.88 18.22 81.23 138.09 133.48 1957 35.55 24.54 19.04 85.94 139.05 152.83 1958 38.20 25.56 20.25 89.29 142.96 173.69 1959 41.49 27.68 21.57 94.11 148.47 195.04 1960 44.57 32.47 24.02 102.18 160.12 216.37 1961 48. 45 36.61 25. 38 109. 24 172.90 232. 95 1962 51.28 44.24 27.60 116.63 187.00 258.63 1% Table 23 . Continued HHHHH HHHHH P HHHHH HHHHH HF“ FHHH WOW 0000 New Penn- New York Jersey sylvania Ohio Indiana Illinois 22.03 54.40 29.99 52.14 60.29 108.69 21.99 59.98 30.56 50.00 56.12 102.20 22.13 63.02 30. 44 47.10 50.61 91. 54 22.08 65.42 30. 52 45. 37 48.44 87. 53 22.36 68. 99 30. 67 44.80 47.62 85. 21 21.85 68. 63 29.97 42.12 45.59 80. 85 9 20. 47 65.26 28.14 37.98 40. 86 71.08 932 19.35 60. 94 26. 27 32.09 33.95 58.18 933 17.30 55.35 21.49 26.44 30.00 47.47 934 17.28 53.07 21.02 27. 82 31.54 50. 37 93 5 17.65 52.36 21.82 28.82 31.43 51.65 93 6 17.57 52. 86 22.52 31.23 36.57 53.85 937 17.34 52.65 22.44 32.69 38.36 57. 29 93 8 17.00 52.09 22.58 32.21 39.48 58.78 93 9 16. 43 51.65 22.06 32.92 39.10 57.65 940 15.61 49.54 21.80 32.97 39.35 50.07 9 41 15.63 54. 47 22.04 34.44 40. 26 60.17 942 15.88 59.95 23.46 38.02 45.37 68.42 943 17.59 65.75 25. 81 41. 87 49.43 72.43 944 18.03 68.98 27.97 47.90 56.29 83. 83 9 45 19.87 76. 61 31.85 52.94 63.11 90.66 9 46 21.70 86. 61 33. 97 60. 89 73.50 101.34 947 25.05 101.14 37.99 68.43 80.32 116.74 948 25.78 105.05 41.27 72.24 87.42 125.12 9 49 27.81 112.54 44.60 75.50 89.70 132. 28 950 27.17 111.38 42.71 71.50 88.49 135.51 9 51 28.91 118.53 48.40 84.71 106.35 159.12 952 31.93 134.16 53.13 93.96 116.95 172.58 9 53 32.05 145.31 52.49 93.11 120.25 176.14 9 54 30.49 146.42 52.28 94.32 119.19 179.42 9 55 32. 47 163.14 53.88 100.79 127.90 183.46 56 34.62 163.14 58.00 109. 89 136.02 197. 51 5'7 38.88 173. 07 62.17 129.11 149. 47 220. 90 58 41.47 179. 92 66.25 127. 95 158.10 230.70 59 46.43 187.13 70.22 135.20 168.65 256.16 60 48.59 192.40 76.40 139. 03 175.07 259.35 61 49. 84 199. 54 80. 85 138. 69 169.12 250. 37 62. 53.57 202. 99 87.33 145.76 173.62 257.42 166 Table 23. Continued North Year Michigan Wisconsin Minnesota Iowa Missouri Dakota 1 925 37.04 48.25 57.52 105.21 48.01 14.65 1 926 35.54 45.55 55.40 99.40 44.25 13.84 1 927 34.73 43.44 51. 01 91.24 41.98 12.87 1 928 33.99 41.76 48.53 87.67 40.83 12.55 1 929 33.36 40.43 47.20 85.66 40.30 12.09 1 930 31.63 38.52 44.45 82.90 39.11 11.79 1 931 29.95 33.93 38.17 71. 50 34.19 10. 56 1 932 25. 00 29.57 31.69 57.68 29.07 9.02 1 933 20. 66 25.83 24. 94 41.58 24.19 8.12 1 93 4 20. 99 25.27 25.64 44.22 25.20 8.32 1 93 5 21.15 25.36 25.13 45.10 25.28 8. 22 1 93 6 21.23 24.98 25.46 48.85 25.14 8. 28 1 93 7 22.63 25.37 25.82 49.32 24.78 8. 03 1 93 8 22.33 24.14 25.95 49.39 23.62 7. 57 1 93 9 22.00 22.56 25.13 48.98 22.19 6.70 1 9 40 22. 03 21. 63 25.45 49.40 22. 67 5. 87 1 941 22.54 21.33 25.92 50. 20 23.30 5. 99 1 942 25. 14 23.13 27.38 53.68 23.66 6.49 1 9 43 27.63 24. 60 30. 47 58.90 28.93 7. 00 1 944 32.08 27.62 33.90 68.82 32.30 8.62 1 945 35.32 30.21 35.70 73.83 36.60 9. 61 1 946 40.32 32.90 39. 90 83.19 41.06 10. 60 1 9 47 46.26 37.03 44.02 92. 65 45.74 11.70 1 948 46.72 39.68 48.32 104. 07 46.67 14.10 1 949 47.41 41.33 50.09 107. 69 49.98 15.07 i 9 50 45.82 39.18 51.26 109.90 49.92 14.72 1 9 51 52.86 43.56 60. 02 129.08 58.65 16.03 1 9 52 56. 55 46.07 65.77 138.08 66.05 18.45 1 9 53 58.78 46.83 67.45 135. 08 66.87 19.32 9 54 59. 68 44.07 64.63 132.72 63.23 18.99 i 9 55 62.38 43.99 68.24 143.33 66.48 19.18 1 9 5 6 67. 67 46.25 75.67 148.51 70. 80 20. 39 1 9S7 74. 23 49.60 82. 96 158. 26 77.11 22.69 1 958 78.32 51.79 93.01 166.38 84.22 25.32 9 59 86. 80 55.13 100. 45 180.42 91.68 28.43 i 3 60 91.52 57.30 103.86 187.02 95.76 29.63 1 9 61 94. 01 58. 65 103.19 180.12 98. 86 30. 58 62 98.59 63.31 109.73 188. 64 103.75 31.61 167 Table 23. Continued South Year Dakota Nebraska Kansas Delaware Maryland Virginia 1 925 30. 81 40.06 29.49 43.28 31.10 46.38 1 926 28. 43 39.75 28.74 44.43 31.32 44.56 1 927 25.46 38.30 28.74 43.58 30.46 41.98 1 928 24.71 37.65 28.36 43.95 30.31 42.25 1 929 24.03 37.23 28.20 44.29 30.42 42.58 1 930 23.79 36.31 28.33 44. 43 30.67 42.07 1 931 21.10 33.50 25.76 42.56 29.70 36.74 1932 17.20 28. 43 22.07 37.05 25.99 31.16 1 933 14.29 22.36 17.38 30.79 21.73 27.61 1 934 13.80 22.24 17.55 30.45 21.39 28.33 1935 13.09 21.58 17.67 30.96 21.59 30.14 1936 12.70 21.39 18.09 32.21 22.46 31.32 1 937 12.14 20.32 18.78 33.42 23.52 33.01 1 93 8 10. 99 18.96 18.60 34. 85 24.30 32.75 1939 9.14 17.38 18.09 35.03 24.11 31.96 1 940 8. 02 15.11 16.95 34.89 23.61 31.77 1 941 7. 94 14.16 17.12 36.77 24.55 32.22 1 942 8. 43 15.51 17.86 38.46 25.82 33.09 1 943 9.63 17.28 20.12 43.61 28.73 36.44 1 944 11.98 21.38 22.95 45.02 31.72 40.55 1 945 13.02 23.97 26.50 51.60 34.23 47.79 1 946 14.39 27.21 28.53 58.40 38.54 55.89 1 9 47 16. 44 30. 95 32.99 69. 50 46. 23 62. 92 1 948 19.81 36.47 38.38 71.69 47.69 63.91 1 949 21.36 40.46 39.78 73.39 49.54 69. 85 1 950 21.80 37.79 38.67 71.36 47.79 66.98 1 951 24.04 43.14 43.15 74.13 54.48 75.46 1 9 52 27.44 46.96 47.12 75.37 61.53 83.63 1 9 53 27.81 48. 62 47.98 82. 22 63. 55 85. 22 1 954 27. 44 45.50 46.74 85.95 68.31 81.37 i 955 28.81 47.10 48.19 88.57 68.49 85.64 1 956 29.73 47.34 49.87 93.00 77.33 90.18 1 957 31.64 47.29 51.39 100.94 84.95 95.35 1 9 58 34.69 52.12 54.29 114.91 95.86 99.39 9 59 38.88 56.37 56.74 127.31 103.99 106.83 i 960 40.00 58.55 . 59.16 140.90 114.11 110.91 1 961 40.54 58.56 60.18 148.82 121.93 117.34 962 43.40 61. 55 63.62 159.94 131.36 128.37 168 Table 23. Continued West North South Year Virginia Carolina Carolina Georgia Florida Kentucky 1 925 32.55 41.90 23.96 17. 20 60.01 37.32 1 926 31.55 41.32 21.96 16.90 76.89 37.41 1 927 30.12 39.52 19.33 15.71 61.56 36.31 1928 29.85 38.14 18.94 15.78 57.91 35.54 1 929 29.67 36.46 18.84 15.85 56.14 35.46 1930 28.75 34.42 17.66 15.64 51.50 35. 54 1 93 1 26.67 29.49 15.78 14.41 48.39 32.27 1 932 22.01 25.01 13.00 11.58 41.64 28.06 1 933 19.70 19.03 10. 42 9. 66 36.08 22.41 1 934 20. 59 22.15 12. 27 10. 96 34.82 22.85 1 935 20.49 24.40 14.38 12.34 31.52 24. 55 1 936 21.11 24.91 14.57 12.27 29.83 24.66 1 937 20. 94 26.42 15.28 13.12 28.92 26.87 1 938 21.18 29.21 15.75 12. 87 26.64 28.05 1 939 21. 15 28.77 15.61 12.71 24.42 28. 80 1 940 20.75 28.59 15.53 12.75 23.19 30.65 1 941 22.36 28.17 16.72 13.66 24.67 30. 42 l 942 23.29 31.31 18.07 14.58 28.32 33.45 1 943 26.39 33.49 19. 81 16.18 32.70 37.25 1 944 28.84 40.65 24.15 19.15 37.48 40. 83 1 945 29.98 47.18 29.02 21.40 46.56 45. 89 1 946 35.21 56.16 31.18 23.94 56.29 53.15 1 947 40.70 65.32 35.92 28.80 54.25 63.12 1 948 44.77 68.71 38.83 30.47 49.26 63.02 1 949 47.58 72.97 42.30 33.70 48.46 67.55 1 950 43.76 74.35 39.79 32. 22 49.77 65.64 i 9 51 47.56 81.47 43.93 36.32 60.04 73.97 1 952 49.72 92.27 47.45 43.71 69.74 81.05 1 9 53 48.30 97.22 49.09 47.62 75.08 77.46 954 48.24 95.24 50.17 47.91 84.73 74.21 i 9 55 47.52 99. 81 51. 50 50.12 90.08 75.74 1 956 48. 54 105.34 57.23 54.38 103.96 78.90 1 9 57 49.45 114.62 62.70 61.01 124.36 87.27 1 358 51.24 122.32 67.36 67.63 146.63 94.44 59 53.47 129.79 75.36 75.96 169.92 103.30 i 3 60 55.60 140. 69 85. 94 86.43 181.46 109.63 1 9651 57.95 149.59 91.91 92.52 196.18 114.53 62 65.66 166.48 104.50 109. 26 217.83 122.26 -_ F “u 169 Table 23. Continued Mis- Year Tennessee Alabama sissippi Arkansas Louisiana Oklahoma 1925 34.32 18.79 18.93 31.11 24.02 16. 00 1926 33.61 19.36 19.33 29.72 25.56 16.17 1 927 32.65 18.68 18.73 29.30 24.85 16.31 1 928 32.01 19.24 18.99 28.79 25.26 16. 56 1 929 31.47 19.60 19.55 28.55 26.24 16.79 1 930 31.09 20. 26 20.41 28.07 27.58 17.40 1 931 29.03 18.78 18.85 23.70 25.70 16. 00 1 932 24.62 15.35 15.78 21.06 22.17 13.15 1933 20.31 13.52 12.71 16.17 19.53 10.74 1934 21.99 15.18 13.96 18.07 20.69 11.48 1 93 5 23.93 16.70 15.03 19.13 22.00 11.79 1 936 25.24 16.70 14.78 20.27 21.66 12.32 1937 25.48 16.51 15.19 20.48 21.91 12.12 1 93 8 25.59 16.84 16.15 21.58 23.03 12.41 1 939 25.56 16.17 15.72 21.24 22.39 12.17 1 940 26.71 16.24 15.38 21.64 22.80 11.96 1 941 27.89 16.95 16.39 22.62 23.57 12.37 1 942 30.15 18.04 18.27 24.82 25.87 13.10 1 943 33.52 19.70 20.18 27.31 30.61 14.52 1 944 38.12 23.29 22.42 31.31 33.34 15.82 1 945 41.39 26.41 25.70 36. 60 36.52 17.42 1 946 48.97 31.13 30. 58 38. 81 40.81 20.90 1 947 55.58 38.31 33.67 46.02 48.86 23.54 1 948 58.57 39.64 37.36 50.90 49.06 25. 59 1 9 49 61.22 44.35 39.26 55. 83 57.44 29. 56 1 9 So 60. 57 43.54 39.56 53. 26 56.24 28. 90 1 9 51 67.74 48.12 46.35 62.92 59.85 34.00 i <352 74.06 52.93 51.61 67.60 66.53 36. 60 1 953 74.64 55.88 53.86 68.04 73.23 35.05 9 54 69. 81 52.45 52. 81 64.74 78.57 34. 41 :11 955 73.35 55.60 54.15 67.48 82.91 37.67 1 9 56 78.27 61.87 63.91 73.93 87.44 39.03 1 957 83.41 66.50 71.31 79.50 97.17 40.87 1 9 58 88. 49 71.67 73.92 85.83 107.61 43.56 9 59 99.07 80. 86 80. 59 89. 03 122. 59 48.14 i 360 106.01 87.55 82.69 101.19 133.61 51.80 1 9 61 110. 75 94.15 84.51 106. 43 137.57 52. 02 62 119.59 101.24 92.36 118.04 150.68 55. 89 ‘iuz kn: ... _LJ '3‘ '_.L1! 9' L.,-slang 170 Table 23. Continued New Year Texas Montana Idaho Wyoming Colorado Mexico 1925 16.97 5.45 16.48 4.68 6.35 3.84 1926 17.30 5.12 15.72 4.38 6.09 3.80 1927 16. 84 5.15 16.34 4.30 6.14 3.86 1928 16.90 5.33 16. 90 4.54 6.13 3.88 1929 17.02 5.28 16.86 4.69 6.10 3.95 1930 17.46 5.14 16.60 4.78 6.17 4.17 1 931 15.52 4.51 14. 52 4. 53 5.65 3.93 1932 12.40 3.60 11.89 3.56 4.32 3.17 1 933 10.66 2.94 9.18 2.73 3.44 2.68 1 934 11.16 2.99 9.46 2.76 3.39 2.75 1 935 11.53 3.09 9.24 2.84 3.35 2.78 1 936 11.72 3.24 9.11 3.01 3.54 2.86 1 937 11.88 3.35 9.47 3.17 3.80 2.90 1 938 12.04 3.25 8.68 3.15 3.81 2.98 1939 11.67 3.29 8.11 3.14 3.83 2.98 1 940 11.71 3.24 8.20 3.12 3.78 3.00 1 941 12.03 3.41 8. 52 3.30 3.97 3.22 1 942 13.50 3.78 9.34 3.66 4.41 3.63 1 943 14.63 4.33 10.93 4.20 5.04 4.36 1 944 17.29 4.97 13.16 4. 87 6.06 5.56 1 945 19.69 5.67 15.25 5.52 6.97 6.67 1 946 22.04 6.46 16.55 6.55 8.23 7.82 1 947 24. 51 7.43 18.05 7.35 9.66 8. 59 1 948 28.20 7.97 19.23 8.49 10.60 9.89 1 949 28. 53 8.39 19.77 8.30 10.98 10.42 1 950 28.15 7.95 19.70 8.02 10.69 10.16 1 951 33.77 9.79 23.21 9.37 12.61 12.14 1 952 37.27 10.98 25.05 10.21 13.98 13.99 1 953 37.54 11.29 26.00 10.12 13.80 14.24 1 954 37.56 11.26 25.68 9.64 13.76 14. 58 i 955 38.95 11.72 26.77 9.58 13.92 14.92 1 956 39.49 12.76 27.93 10.16 14.47 14.90 1 957 43.44 13.66 28.58 10.87 15.44 14. 57 1 9 58 43.09 14.80 29.90 12.01 16.81 14.98 9 59 46.69 16.40 31.80 13. 50 18.23 15.38 511 9 60 53.44 17.53 32.48 13.95 19.62 16.29 1 961 56.47 17.81 32.37 14.72 20.09 17.66 962 60.79 19.08 33.37 15. 44 21.84 18.82 1T1 ‘ ' ' I. 171 Table 23. Continued Year Arizona Utah Nevada Washington Oregon California 1925 4.76 8.84 10.63 17.80 12.15 38.18 1926 4.99 8.94 10.18 17.69 11.79 37.91 1927 5. 07 9. 20 10. 24 18. 03 12. 12 38. 51 1928 5.51 9.48 9.96 17.93 11.92 38.15 1929 5.69 9. 66 9.84 17.23 11.91 38.11 1930 5.72 9.66 10.00 17.85 11.70 38.62 1931 5.55 8.74 9.41 15.38 10.51 35.39 1932 5.37 7.86 8.36 12.50 8.46 29.38 1933 5.12 6.65 6.72 10.13 6.61 21.92 1934 4. 86 6. 88 6. 80 10. 88 6. 48 21. 80 1935 4.28 6.75 6.34 11.58 6.76 22.70 1936 4.17 6.71 6.54 11.89 7.26 24.39 1937 3. 96 6. 80 6. 64 12. 96 7. 65 26. 42 1938 3.67 6.71 6.48 12.71 8.22 26.05 1939 3.32 6.53 6.43 12.37 8.16 23.46 1940 2. 85 6. 41 6.12 12.19 8. 30 22. 42 1941 3.26 7.11 6.47 12.50 8.96 22. 88 1942 3. 65 7. 84 7. 08 14.15 9. 80 25. 43 1943 4.39 8.64 8.13 15.81 10.95 30.01 1944 5.30 9.92 9.59 19.57 13.35 37.23 1945 6.25 11.39 11.33 21.34 15.39 44.75 1946 7.14 12.92 12.62 24.95 18.09 51.71 1947 7.85 14.58 13.17 26.98 19.81 55.23 1948 8. 09 15. 74 13. 36 29. 21 20. 88 53. 89 1949 8. 09 16. 21 12. 60 28. 88 20. 41 50. 62 1950 7.71 16. 48 12.33 28.42 20.12 48.59 1951 9. 21 18. 56 14. 62 32. 88 23. 40 57. 22 1952 10.61 20.50 17.01 35.53 25.30 66.19 1953 11.78 20.78 17.51 37.58 26.93 68.36 1954 12.06 18.79 19.13 36.92 26.66 68.12 1955 12.73 20.24 19.72 38.28 27.63 71.75 1956 14.83 21.17 20.86 39.46 28.15 77.88 1957 17.38 22.58 21.60 40.55 28.91 84.95 1958 19.75 23.86 22.59 41.63 29.35 92.18 1959 22. 17 25. 61 23. 55 42. 54 30. 20 100. 97 1960 24. 55 26. 94 23. 58 43. 34 30.17 108. 45 1961 26.50 28.16 24.23 43.69 30.33 116.84 1962 28.30 29.46 24. 50 43.69 31.14 121.97 MWJEILAK ‘fl APPENDIX C INDICES OF PRICE EXPECTATIONS BY STATES The series of ten-year expected prices developed by M. L. Lerohll were for 13 separate commodities. In order to develop aggregate indices of price expectations for each state, it is necessary 1'“ ' p. to assign weights to each of the 13 commodities on a state basis. The method used by Lerohl for aggregation on the national level is followed. However, weights on a state basis are not readily available. The more recent censuses provide data by state on the values of products sold L from farms as well as values of individual agricultural commodities sold from farms. However, these values are estimated on basis of a single year's returns, and could be quite heavily influenced by weather conditions. Also it was found that there is a big discrepancy between the data given in the censuses, and those obtained in the yearly agri- cultural statistics. 3 In order to avoid regional disturbances due to weather, it is preferable to use the average returns of consecutive years. It was therefore decided to use the data obtained on a yearly basis. Weights are being estimated for two time period, 1937-39 and 1947-49. 102. cit., Appendix A, pp. 159-185. 2Ibid., pp. 34-37. 3Agricultural Statistics, U. S. Department of Agriculture, Washington D.C. (yearly publication). The discrepancy between the two sources of data is illustrated by the following example. In 1954 the gross farm income from dairy products for Maine was 38. 9 million dollars according to the Agr. Statistic_s, while the figure was 24. 3 million dollars in the census estimate. 172 173 The weight (w) for the j—th commity4 in state s at time t, is: § Pr. jts ths w. = , JS ZZPr'jtstts S 000,130 '00., 480 (_a H II p—a N.N where: Pr. Price of commodity j in year t. jt th = Quantity sold of commodity j in year t. .. t == Year(47-49cn°37-39) ; Given indices of expected prices and weights for the different commodities, a single regional index of expected prices can be computed. Using constant weights for the entire time period, the 1? index of expected prices for state s (IEs)’ is: I , ..., 38. ESZAfixflzfi' 1,2..H13 where: A = A matrix of expected price series, where the columns are the indices of expected prices for the 13 commodities, 1925-62. zS = A column vector of weights attributed to the 13 commodities for state 8. There was little difference between the aggregate indices estimated by the two sets of weights, and no other weights were calculated. The linkage between the two sets of indices is made in year 1943. The expected price indices for each year 1925-43 are 4{Weights are derived for the 13 commodities. However, the price expectation index for wheat is used as an index for food grains and the weights for food grains comprises wheat and rice. Similarly, the index of soybeans is used as an index for oil crops, and comprises soybeans, cottonseed, and flaxseed. 174 modified to reflect the relationship between the two sets of indices in the linkage year. The state indices of expected prices are given in Table 24. 175 Table 24. Indices of Price Expectations by States, 1925-62 New Massa- - Rhode Con- Year Maine Hampshire Vermont chusetts Island necticut 1925 70.86 67.60 59.43 66.61 66.48 63.64 1926 79.03 79.94 61.98 69.61 69.80 66.21 1927 78.53 69.97 61.53 68.70 69.09 65.72 1928 75.73 70.45 62.46 69.40 69.54 66.37 1929 74.52 69.84 62.50 69.03 69.21 65.91 1930 76.56 68.41 60.24 67.15 67.61 64.26 1931 64.78 59.86 53.94 59.21 59.84- 56.51 1932 49.40 47.47 42.35 47.841 47.72 46.15 1933 48.67 44.27 40.01 44.62 45.07 42.91 1934 51.42 45.45 40.83 45.80 46.12 44.66 1935 52.42 52.97 46.80 52.98 52.43 51.20 1936 59.61 56.65 49.42 56.38 55.85 54.69 1937 63.99 59.01 51.07 58.59 58.06 56.88 1938 61.93 60.40 53.58 60.34 59.75 58.38 1939 59.73 57.08 49.88 56.63 56.24 54.93 1940 60.16 57.46 51.04 57.07 56.91 55.42 1941 61.68 58.60 51.46 58.04 57.74 56.54 1942 74.41 66.15 59.95 66.70 66.17 67.23 1943 85.93 76.92 69.01 77.17 76.19 78.09 1944 91.21 85.34 78.46 85.41 84.35 85.67 1945 87.39 78.67 76.18 79.22 79.47 79.92 1946 88.84 85.83 79.53 85.41 84.47 84.33 1947 95.97 93.36 93.14 93.56 93.83 93.52 1948 100.32 99.70 97.72 99.94 99.15 100.14 1949 104.01 106.64 109.14 106.49 106.71 106.64 1950 97.56 99.21 99.33 100.13 98.77 102.34 1951 100.32 98.11 102.29 99.54 99.18 102.93 1952 105.16 106.55 108.38 107.02 106.35 108.71 1953 104.53 108.34 110.06 109.14 107.94 110.06 1954 96.51 103.80 101.99 105.25 102.10 106.50 1955 96.44 99.93 100.80 102.26 99.59 105.39 1956 95.03 97.00 97.53 99.53 96.77 103.70 1957 95.20 96.54 99.89 99.82 97.18 105.10 1958 93.19 94.75 99.38 98.50 95.74 104.85 1959 90.50 94.69 99.25 98.75 94.98 106.05 1960 92.63 93.15 98.98 97.59 94.52 105.33 1961 89.92 92.99 99.49 97.81 94.16 105.84 1962 90.12 92.84 100.86 97.69 94.32 105.99 Table 24. C ontinued 176 ‘— T New Penn- Year New York Jersey sylvania Ohio Indiana Illinois 1925 61.49 68.73 62.80 58.67 57.60 59.90 1926 64.61 72.30 66.12 63.53 61.64 62.39 1927 63.68 71.27 65.36 62.65 60.93 61.78 1928 64.37 71.53 65.53 62.15 60.11 60.95 1929 64.01 70.88 65.28 62.68 61.17 61.78 1930 62.09 69.49 63.13 59.08 57.01 57.05 1931 54.84 61.05 55.60 51.92 50.01 50.84 1932 42.73 48.56 43.64 41.09 39.78 40.18 1933 40.41 45.65 40.91 38.00 36.45 37.41 1934 41.69 47.08 42.81 40.45 39.07 40.81 1935 47.70 53.96 48.97 45.92 44.40 45.89 1936 50.94 57.83 52.91 50.61 49.48 50.88 1937 52.93 60.24 55.12 53.16 52.14 53.17 1938 54.80 61.43 56.30 53.85 52.43 52.61 1939 51.29 57.96 53.08 50.27 48.90 49.97 1940 52.23 58.46 53.91 50.97 49.55 50.72 1941 52.89 59.59 54.98 52.24 51.19 53.57 1942 61.31 67.70 63.96 62.70 62.01 62.84 1943 71.09 78.31 73.65 71.05 69.60 69.88 1944 80.06 85.95 80.84 75.64 72.83 72.65 1945 77.63 79.24 76.61 73.63 71.84 '72.44 1946 80.81 86.25 81.50 77.18 75.07 76.29 1947 93.43 93.65 92.48 91.35 90.79 91.20 1948 98.17 100.24' 99.94 102.53 103.78 103.84 1949 108.10 106.11 106.98 106.12 105.13 104.96 1950 98.83 99.40 99.85 100.37 99.47 100.00 1951 102.26 97.73 101.32 102.53 101.75 1104.43 1952 107.91 106.09 107.40 105.36 104.09 107.45 1953 109.41 107.71 107.83 104.81 102.82. 104.44 1954 101.51 103.76 101.79 100.14' 98.29 99.18 1955 100.50 100.01 99.99 99.18 96.91 97.85 1956 97.10 97.11 96.86 95.24 92.96 93.88 1957 99.09 96.52 97.79 96.11 93.48 94.11 1958 98.17 94.71 96.69 95.40 92.60 93.16 1959 97.71 94.09 96.50 94.31 91.17 92.17 1960 97.47 92.88 95.55 93.32 90.24 91.39 1961 97.56 92.52 95.36 9.33 90.22 91.71 1962 99.00 92.19 96.62 95.48 92.46 94.70 177 Table 24. C ontinued North Year Michigan Wisconsin Minnesota Iowa Missouri Dakota 1925 60.30 57.21 59.19 53.67 57.02 66.38 1926 64.10 60.64 62.69 57.88 60.71 68.76 1927 63.25 60.20 62.11 57.36 59.80 67.36 1928 63.24 60.63 61.56 56.99 59.64 65.94 1929 63.26 61.21 62.33 58.77 60.61 64.92 1930 60.79 58.48 58.49 54.29 56.43 59.78 1931 53.28 51.83 51.88 47.52 49.57 54.19 1932 41.45 40.84 40.84 37.91 38.93 41.30 1933 39.08 37.86 37.92 34.75 35.91 39.71 1934 41.08 39.14 40.47 37.41 38.70 44.58 1935 46.32 45.37 45.70 43.03 44.30 46.85 1936 50.58 48.91 50.26 48.56 49.03 51.33 1937 52.92 50.88 52.66 50.78 51.44 53.99 1938 53.92 52.80 53.23 50.77 51.92 54.55 1939 50.87 49.06 50.10 48.11 49.20 51.70 1940 51.71 50.19 51.04 49.10 50.28 52.66 1941 52.71 50.80 52.56 51.07 51.57 53.26 1942 62.12 60.32 62.54 61.16 61.46 64.98 1943 71.27 69.02 70.23 68.15 69.19 72.68 1944 78.00 76.06 74.32 70.16 72.47 77.00 1945 75.30 73.95 73.08 69.52 71.23 75.76 1946 79.08 77.13 76.68 72.64 74.88 78.23 1947 92.05 91.74 91.01 89.20 90.02 92.16 1948 100.33 99.86 102.85 104.66 102.98 103.12 1949 107.32 108.38 105.83 106.14 106.70 105.02 1950 100.14 99.81 99.85 100.10 101.07 103.31 1951 103.47 102.16 102.75 102.90 105.20 110.94 1952 108.09 106.84 106.01 104.61 107.57 111.70 1953 108.02 107.17 104.40 101.85 104.97 109.74 1954 101.01 99.62 98.77 95.30 97.90 105.04 1955 100.38 98.63 97.29 94.40 97.64 107.65 1956 96.54 95.54 94.00 90.99 94.18 101.33 1957 98.35 97.67 94.77 92.30 95.90 103.75 1958 97.30 97.28 93.87 91.74 95.20 101.84 1959 96.82 97.09 93.24 91.30 95.66 99.72 1960 95.89 96.78 92.48 90.85 94.76 96.75 1961 95.60 96.88 92.48 90.52 94.23 95.59 1962 98.02 98.66 94.72 93.65 97.57 100.94 178 Table 24. Continued South Year Dakota Nebraska Kansas Delaware Maryland Virginia 1925 57.87 57.18 62.40 68.23 60.81 58.70 1926 61.23 60.48 64.65 70.04 63.31 61.85 1927 60.54 59.81 63.35 69.49 62.72 61.03 1928 60.24 59.73 62.92 69.67 62.76 61.27 1929 61.11 60.76 62.62 69.20 62.61 61.35 1930 56.58 56.03 57.14 66.64 59.64 58.62 1931 50.22 49.82 51.72 57.544 52.49 50.73 1932 39.45 39.11 39.67 46.29 42.04 41.29 1933 36.69 36.71 37.81 42.69 39.12 38.35 1934 39.79 40.18 42.42 43.70 41.62 40.90 1935 44.95 45.13 46.17 51.17 47.21 46.23 1936 49.83 50.31 50.73 54.92 51.29 50.76 1937 51.98 52.21 52.76 57.65 53.53 53.23 1938 52.50 52.42 53.49 58.35 54.43 53.80 1939 50.07 50.83 51.53 56.20 51.24 50.78 1940 51.14 52.12 52.82 55.16 51.72 51.09 1941 52.54 53.69 53.59 57.57 53.18 52.68 1942 62.39 63.11 63.78 64.09 63.76 64.45 1943 69.84 70.30 71.26 74.42 73.24 74.42 1944 72.59 72.80 75.14 81.36 79.97 80.16 1945 71.57 71.37 72.92 77.32 76.42 76.64 1946 74.92 75.33 76.45 82.98 80.34 80.32 1947 89.61 89.53 90.20 93.37 92.63 92.30 1948 3103.97 103.47 102.78 100.68 100.34' 101.39 1949 106.72 106.70 106.72 105.94 106.74 106.01 1950 102.34 102.43 104.28 99.53 101.02 101.42 1951 108.44 108.94 112.18 101.41 103.33 103.52 1952 110.12 111.46 113.76 107.85 108.23 107.63 1953 106.15 106.77 110.29 108.12 108.54 107.81 1954 98.30 98.26 102.82 103.47 103.46 103.42 1955 99.19 99.40 105.92 100.98 102.69 102.80 1956 95.17 95.12 99.93 97.09 99.47 100.04 1957 97.49 97.44 103.40 96.88 100.91 101.62 1958 96.49 96.49 101.87 95.29 100.36 101.21 1959 96.68 96.75 101.36 94.45 100.28 101.59 1960 95.34 95.62 98.64 92.95 99.33 101.03 1961 94.50 94.72 97.17 93.03 99.59 101.35 1962 99.41 100.08 103.71 94.26 101.15 102.76 Table 24. C ontinued 179 West North South Year Virginia Carolina Carolina Georgia Florida Kentucky 1925 60.44 51.33 59.84 61.29 95.55 52.32 1926 64.01 52.98 61.52 63.65 101.74 54.75 1927 62.78 53.15 60.46 62.33 94.14 54.81 1928 63.57 53.35 59.76 61.65 92.74 55.08 1929 63.75 53.45 59.89 62.08 92.93 55.61 1930 61.01 50.89 56.53 58.37 88.16 52.57 1931 53.56 43.41 48.76 50.75 70.39 45.44 1932 42.33 38.11 39.99 41.22 59.08 38.24 1933 39.99 33.96 36.22 37.49 57.00 34.59 1934 41.66 38.05 39.98 40.55 68.02 37.97 1935 47.52 42.48 44.46 45.61 65.09 43.23 1936 51.46 46.76 47.93 49.12 69.44 47.94 1937 53.50 49.65 51.94 52.96 75.58 50.20 1938 54.92 49.80 51.04 52.47 68.50 50.62 1939 51.96 46.60 47.16 48.11 64.14l 47.91 1940 52.81 46.83 47.73 48.75 63.56 48.43 1941 53.73 48.67 49.77 50.54 70.67 50.23 1942 62.42 67.74 64.63 63.52 85.44 65.09 1943 71.73 79.09 73.93 72.03 96.29 75.10 1944 78.31 83.50 77.05 75.47 103.02 79.37 1945 75.10 79.24 75.54 74.05 97.52 75.78 1946 79.49 81.71 79.83 78.31 99.44- 78.37 1947 91.54 93.53 94.55 93.58 91.59 90.96 1948 100.78 101.72 101.03 101.52 97.47 102.58 1949 107.38 104.75 104.42 105.20 110.94 106.76 1950 99.89 105.83 103.50 102.11 110.46 104.45 1951 102.27 108.58 108.25 105.86 115.05 107.82 1952 107.26 110.67 110.80 108.77 105.52 110.09 1953 107.38 110.35 109.56 107.73 105.82 108.22 1954 100.30 109.89 108.22 104.99 103.32 103.23 1955 98.89 111.49 108.79 104.77 109.79 104.38 1956 95.68 110.50 105.36 101.09 109.65 103.10 1957 97.39 112.53 105.70 101.47 110.64: 105.90 1958 96.35 113,82 105.97 101.39 115.75 106.54 1959 96.62 116.48 107.70 102.23 119.97 108.81 1960 96.00 115.94 106.63 101.27 125.11 108.37 1961 95.83 117.09 107.16 101.56 127.67 108.71 1962 97.79 117.61 108.22 102.86 128.95 111.02 180 Table 24. Continued ll Nfis- Year Tennessee Alabama sissippi Arkansas Louisiana Oklahoma 1925 58.33 66.01 70.19 67.30 66.14 62.75 1926 61.08 68.13 71.36 69.33 68.61 65.00 1927 60.34 66.36 68.85 67.29 66.73 63.66 1928 60.05 65.43 67.52 66.11 65.80 63.26 1929 60.58 65.68 67.59 66.36 66.29 63.16 1930 57.12 61.88 63.11 62.33 62.33 58.25 1931 49.82 54.16 55.21 54.27 54.36 52.30 1932 40.25 42.64 42.94 42.41 42.78 40.37 1933 36.77 39.19 39.41 38.93 39.39 38.11 1934 39.66 42.11 43.14 42.15 42.50 42.10 1935 45.15 47.41 47.83 47.08 47.43 46.39 1936 49.32 50.41 50.00 49.96 50.53 50.43 1937 52.23 54.59 55.27 54.70 54.90 52.92 1938 52.30 53.66 53.31 53.27 53.80 53.40 1939 48.85 49.30 48.93 49.15 49.85 50.90 1940 49.55 50.20 50.05 50.05 50.86 52.18 1941 51.29 52.17 52.30 52.10 52.72 53.14 1942 63.28 62.62 63.07 62.78 63.41 63.20 1943 72.04 70.44 70.32 70.40 71.02 70.79 1944 75.98 73.49 72.67 72.95 74.26 74.66 1945 74.22 73.36 73.42 73.33 73.63 72.87 1946 78.32 78.71 79.39 78.79 78.38 76.87 1947 92.81 94.50 95.56 94.74 93.73 90.99 1948 101.49 100.59 100.07 100.49 101.13 102.07 1949 106.00 105.21 104.68 104.47 105.44 106.95 1950 102.31 101.62 102.59 102.49 102.82 103.79 1951 106.42 107.37 110.29 109.93 109.91 111.23 1952 109.58 110.86 113.42 112.62 112.31 113.39 1953 108.09 108.88 110.72 110.17 109.92 110.12 1954 103.80 105.24 107.80 107.17 105.64 102.82 1955 103.77 104.86 107.96 107.69 106.77 105.06 1956 100.47 100.29 103.03 102.43 101.38 99.82 1957 101.51 100.03 102.40 102.29 102.12 102.64 1958 101.38 99.61 101.94 101.60 101.24: 101.46 1959 102.50 100.56 103.95 102.87 101.81 101.85 1960 101.64- 99.44 102.51 101.18 99.86 99.67 1961 101.81 99.45 102.55 101.04 99.30 98.51 1962 103.70 101.18 104.42 103.44 102.60 103.90 181 Table 24. C ontinued Ne“! Year Texas Montana Idaho Wyoming Colorado Mexico 1925 63.74 63.22 64.18 53.18 57.41 55.12 1926 65.64 65.19 68.84 56.31 61.06 57.62 1927 63.95 63.77 67.55 55.95 60.40 56.70 1928 63.78 63.19 66.41 57.24 60.79 58.14 1929 64.22 62.42 65.80 58.58 61.52 59.53 1930 59.60 56.94 63.05 54.20 57.70 54.55 1931 52.87 51.75 55.26 48.50 51.16 48.91 1932 40.99 39.14 42.29 36.68 39.13 37.35 1933 38.30 37.65 41.18 34.86 37.51 35.54 1934 42.01 42.53 44.86 38.01 40.90 38.90 1935 47.01 45.56 47.02 43.54 45.37 44.34 1936 50.28 49.86 52.42 48.00 50.37 48.30 1937 53.50 51.95 55.37 49.13 52.12 49.76 1938 53.01 52.85 55.30 50.31 52.57 50.59 1939 50.52 50.82 52.96 50.89 52.18 50.95 1940 52.09 52.13 53.86 53.23 53.93 53.32 1941 53.74 52.60 54.65 54.02 55.04 54.32 1942 63.04 63.11 66.16 61.29 63.67 61.51 1943 70.35 70.71 75.11 68.54 71.46 68.38 1944 73.26 74.46 79.83 71.56 75.04 70.97 1945 72.87 72.70 77.56 70.41 73.36 70.84 1946 77.94 75.68 79.85 73.31 76.28 75.24 1947 92.65 89.44 92.04 86.48 89.14 89.30 1948 100.76 102.74 101.58 102.50 102.43 101.49 1949 105.99 107.52 106.37 110.42 108.14 108.92 1950 103.55 105.22 101.95 105.10 104.27 105.09 1951 111.91 115.01 108.57 115.83 113.76 116.27 1952 114.39 115.77 110.94 116.97 115.24 118.03 1953 110.71 111.20 108.65 109.37 109.94 111.29 1954 104.73 101.84 100.24 93.85 98.50 99.22 1955 106.29 106.04 102.59 97.53 102.00 102.13 1956 101.51 100.36 98.23 95.41 97.93 99.04 1957 102.98 104.84 101.13 101.16 102.36 103.06 1958 102.27 103.36 99.60 100.58 101.16 102.58 1959 104.17 103.72 98.53 104.71 102.52 107.04 1960 102.51 101.02 97.70 104.04: 101.15 106.03 1961 101.88 99.17 95.87 101.94 99.22 104.70 1962 105.94 106.81 100.41 109.93 106.24 111.28 1.. Year 1925 1926 1927 1928 1929 1930 1931 182 Table 24. Continued Year Arizona Utah Nevada Washington Oregon California 1925 59.87 61.10 51.40 66.34- 62.07 66.76 1926 61.84 64.56 54.59 68.86 65.27 69.82 1927 60.34 63.45 54.32 65.68 64.05 68.44 1928 61.43 63.97 56.04 66.86 64.18 68.83 1929 62.67 63.96 57.74 65.59 63.97 68.94 1930 57.66 60.92 53.45 61.30 60.38 65.74 1931 51.29 53.96 47.84 54.35 53.83 58.40 1932 39.41 41.65 36.38 42.55 41.71 46.09 1933 37.23 38.99 34.33 42.23 39.89 43.43 1934 41.07 41.60 37.00 45.34 342.85 46.21 1935 46.37 48.06 43.25 48.42 47.23 52.05 1936 49.62 52.22 47.49 52.03 51.35 55.60 1937 51.99 54.12 48.39 54.09 53.45 57.98 1938 51.86 55.41 49.91 55.60 54.67 58.78 1939 51.60 52.97 50.47 52.71 52.11 56.11 1940 53.98 54.39 52.97 53.21 53.31 57.47 1941 55.47 55.13 53.69 53.68 53.89 58.63 1942 63.02 63.43 60.43 62.66 63.53 67.17 1943 69.85 72.58 67.61 72.17 72.12 76.04 1944 72.23 78.19 70.52 81.25 77.84 81.41 1945 72.80 74.60 69.50 78.71 75.53 79.02 1946 77.66 79.14 72.69 83.25 78.81 83.17 1947 71.80 90.47 85.41 94.00 91.33 91.94 1948 100.69 101.19 102.57 100.05 101.17 99.62 1949 107.51 108.04 112.02 105.64 107.20 108.13 1950 104.62 102.03 105.82 100.01 101.72 103.45 1951 115.26 106.32 117.38 105.98 107.46 109.46 1952 117.15 110.78 118.81 110.42 110.66 111.18 1953 111.38 108.57 109.82 113.97 109.09 109.04 1954 102.37 99.73 91.79 109.34 100.99 101.57 1955 104.49 99.85 95.44 109.51 102.04 102.94 1956 101.27 96.74 94.52 102.10 97.91 100.40 1957 103.55 99.31 100.97 104.92 100.73 102.24 1958 103.28 98.17 100.70 102.32 99.41 102.41 1959 107.44 99.19 106.56 100.84 99.32 104.71 1960 106.62 98.07 106.39 98.80 98.10 105.06 1961 105.84 97.02 104.23 99.12 97.12 104.88 1962 110.62 100.71 112.73 102.70 101.20 107.75 AND 51 In several re in the exp linked wit how much In increases in Severe are giver APPENDIX D R ELATIVE C HANGE IN AND SIZE OF SELECTED VARIABLES BY STATES, 1925-1962 In the presentation of the statistical results in Chapter V several references were made to actual increases which have occurred in the explanatory variables. The actual increases in a given variable linked with its coefficient as related to land values give an estimate of how much this variable has contributed to land value increases. In order to give some indication of the reasons for land value increases and for the different increases among states, the increases in several variables which explained a major part of land value changes are given in Table 25. 183 T able \ 184 Table 25. Relative Changes in and Size of Selected Variables by States, 1925-1962 Ratios 1962/1925 Average Amount ($)1 3 c: 8’ 8 E” .3 33 § .9 :3 .9 S H g '8’ g g: *8 0-1 79"“ 13 9.8 13‘ > f: "" :> 0 > g "a +1 a: (d H "U O H > ' '0 CL 8 0 7" ' ":1 2,? 8 2:03 :3 Region {5“ 11% +37% 8‘ 5% Egg ..1 8. ‘1 H o 2 o. o m o o 0: Southeast South Carolina 4. 36 l. 81 3. 08 1. 42 l. 06 . 150 Georgia 6. 35 1.68 3.23 1.41 1.56 .171 Alabama 5. 39 1. 53 3. 54 l. 33 l. 46 . 074 Florida 3.63 1.35 2.36 4. 41 .51 .018 Appalachians Virginia 2. 77 l. 75 2. 90 l. 76 . 82 . 016 West Virginia 2.02 1.62 2.83 l. 10 .96 . 027 North Carolina 3 97 2. 29 2. 56 l. 64 . 91 . 029 Kentucky 3. 28 2. 12 2. 42 l. 22 l. 04 . 040 Tennessee 3.48 l. 78 2. 94 1.46 .97 . 056 Northeast 1 New York 2.43 1.61 3.11 1.50 .92 .056 Pennsylvania 2. 91 l. 54 3. 18 1. 24 1. 01 . 055 Delaware 3. 70 l. 38 3. 52 2. O3 1. 01 . 033 Maryland 4. 22 1. 66 3. 02 2. 06 . 71 . 029 Northeast 2 Massachusetts 2. 89 1. 47 3. 06 1. 27 . 49 . 002 Rhode Island 3. 71 l. 43 3. 10 1. 33 . 50 . 000 Connecticut 5. 63 l. 67 2. 76 5. 78 . 38 . 005 New Jersey 3. 73 1. 34 3. 37 l. 72 . 52 . 025 Northeast 3 Maine 2. 66 l. 27 3. 03 1. 27 l. 30 .107 New Hampshire 3. 84 l. 37 3. 33 l. 39 l. 80 . O41 Vermont 2. 47 l. 70 2. 96 l. 10 l. 88 . 040 Lake States Michigan 2. 66 l. 63 3.14 l. 86 . 74 . 063 Wisconsin 1. 31 l. 72 2. 97 1. 46 . 72 . 049 Minnesota 1.91 1.60 3.19 1.40 .46 .063 Corn Belt Ohio 2. 80 1. 63 3. 06 l. 62 . 50 . 034 Indiana 2. 88 1. 61 2. 97 l. 52 . 43 . 030 Illinois 2. 37 1.58 3.15 1.43 . 29 .011 Iowa 1.79 1.74 2.63 1.13 .30 .016 Missouri 2.16 l. 71 2. 69 l. 23 . 76 . 043 Table 25. Region K Delta St Mi 8 s Ark; Loui Tex Okl Nort‘ne Sor Ne‘ Ka 185 Table 25. (continued) Ratios 1962/1925 Average Amount ($)1 (0 H C‘. 35’ 8 g .. a E § 8 SN 9 8‘5 .2 ‘35 E ‘9‘: :> g 1.. 1.» :8 *5 H 'o a) H > R ° '6 ‘1 ‘1’ 8 5 ' ":4 3 8 30$ 3 eg1on a“ .30. 3g 0. go. «1cm «1 q, H >4 :3 o 0 >4 .2: o a) 49 mm 02 m om Dom Delta States Mississippi 4. 88 1 49 3 78 1.18 l. 26 . 041 Arkansas 3. 79 l 54 3 58 l. 01 . 91 . 057 Louisiana 6. 27 l 55 3 47 1. 69 . 88 . 035 Southern Plains Texas 3.58 1.66 3.19 1.91 .52 .035 Oklahoma 3.49 1.66 3.05 1.10 64 .063 Northern Plains North Dakota 2.16 1.52 3.98 .97 .80 .173 South Dakota 1. 41 l. 72 2. 98 1. 08 . 79 . 112 Nebraska 1. 54 l. 75 2. 84 1.11 . 42 . 027 Kansas 2.16 1.66 3.07 l. 21 . 35 .035 Mountain 1 Montana 3.50 1.67 2.91 1.31 .70 .039 Idaho 2.03 1.56 2.89 l. 59 . 37 . 023 Wyoming 3. 30 2.07 2. 37 1. 73 .83 .016 Colorado 3.44 1.85 2.59 1.92 .65 . 072 Utah 3. 33 1.65 2.82 2.01 .52 .033 Mountain 2 New Mexico 4. 9O 2. 02 2. 43 2. 59 . 75 . 093 Arizona 5. 95 l. 84 2. 89 3. 88 . 45 . 000 Nevada 2. 30 2.19 2. O9 3. 94 . 31 . 000 Pacific States Washington 2. 45 1. 55 3. 07 2. 05 . 32 . 018 Oregon 2. 56 l. 63 2. 94 2.13 . 38 . 018 California 3.19 1.62 2.81 3.66 .14 .003 lAverage amount of conservation expenditures and conservation reserve payments, respectively, per $100 of land value (1940 value). pae = pasture acre equivalent.