An expanded version of the present paper is available in limited supply as No. 74–22 in the Discussion Paper Series, the Graduate Program in Economic History, University of Wisconsin, Madison 53706. The author wishes to thank Nathan Rosenberg, Theodore W. Schultz, Jeffrey G. Williamson, two anonymous referees, and the economic history seminar at the University of Michigan for helpful comments on earlier drafts. The financial support of the Population Council and the Rockefeller Foundation is most gratefully acknowledged.

¹ See Yasuba, Yasukichi, Birth Bates of the White Population in the United States, 1800–1860: An Economic Study (Baltimore: Johns Hopkins, 1961)Google Scholar; Easterlin, Richard A., “Does Human Fertility Adjust to the Environment?” American Economic Review, LXI (May 1971), 399–407Google Scholar; and Forster, Colin and Tucker, G.L.S., Economic Opportunity and White American Fertility Ratios, 1800–1860 (New Haven: Yale University Press, 1972).Google Scholar

² E.g. Spengler, Joseph J., “The Economist and the Population Question,” American Economic Review, LVI (March 1966), 16–19Google Scholar; and Meadows, Donella et al., The Limits to Growth (New York: Universe Books, 1972), chs. 2 and 3.Google Scholar

³ Principles, 1920 edition, as cited in Schultz, Theodore W., “Land in Economic Growth,” in Modern Land Policy, ed. Halcrow, Harold G. et al. (Urbana: University of Illinois Press, 1963), p. 23.Google Scholar

⁴ Schultz, Theodore W., “The Declining Economic Importance of Land,” Economic Journal, LXI (December 1951), 725–740CrossRefGoogle Scholar; idem, Economic Growth and Agriculture (New York: McGraw-Hill, 1968); idem, “Land in Economic Growth.”Google Scholar;

⁵ In what follows it will be necessary to use the phrase “economic growth” as a shorthand for the proximate sources of past growth in U.S. product per member of the labor force: rising supplies of human and nonhuman capital per member of the labor force, rising total factor productivities in various sectors, and a bias toward substituting capital and skills for unskilled labor in various sectors. Clearly these different sources of growth in incomes per member of the labor force have different implications for land prices. The text below goes to some length to sort out these differences. Yet it remains convenient to speak of the effects of “economic growth” when referring strictly to the effects on land prices of the combination of capital accumulation, skill accumulation, productivity advance, urbanization, and factor bias of technological change that characterized American economic growth in the period under discussion.

⁶ Chambers, Clyde R., Relation of Land Income to Land Value, U.S. Department of Agriculture Bulletin No. 1224 (Washington, D.C.: G.P.O., 1924.CrossRefGoogle Scholar

⁷ There are two reasons why the trend in residential site rents has been misstated by the site component of the cost of shelter index. One is that the rise in the share of multi-dwelling housing units has probably reduced the average acreage per household. That means that the available figures on the site share of rent per household have steadily declined relative to the more relevant series on site value per acre. The second bias is a book-value bias. Both for owned and for rented housing, the Bureau of Labor Statistics gathers data on the prices actually paid. The price paid for a house and lot is typically a few years out of date, and thus underestimates the current opportunity cost of ownership. For rented dwellings as well, rents are typically those fixed about a year earlier. This lag would not bias the index if it were consistent. But with each revision of the cost of housing index, the BLS apparently splices an up-to-date price onto the previous obsolete one, and allows the book value drift to begin anew as the new sample of units is re-surveyed over the years. Over a period of decades this procedure underestimates the rise in both housing costs and site costs.

BLS officials seem aware of a downward bias in the shelter index when they speak of “new bias” and “aging bias” in the rent index. (U.S. Department of Labor, Bureau of Labor Statistics, The Consumer Price Index: History and Techniques [Washington: G.P.O., 1966], p. 52.) Yet they have in mind a somewhat different bias, and adjust for it only for rented dwellings. Their concern seems to lie with the fact that the units surveyed are progressively more depreciated in quality than the average of all dwellings. Adjusting for this quality bias, however, only mixes a more rapid quality improvement in with price changes, still obscuring the latter.Google Scholar

⁸ The value of farm land in constant dollars per acre would have nearly tripled between 1850 and 1910 even if the value of farm buildings in 1850 had been zero. Attempts have been made by Tostlebe (Tostlebe, Alvin S., Capital in Agriculture: Its Formation and Financing since 1870 Princeton: National Bureau of Economic Research, 1957])Google Scholar and Primack (Primack, Martin L., “Farm Formed Capital in American Agriculture, 1850 to 1910,” unpublished doctoral dissertation, University of North Carolina, 1962)Google Scholar to estimate the shares of land and buildings in the total value of farm real estate over the second half of the nineteenth century. Their assumptions differ, and the true shares for this period remain unclear. For a discussion of these estimates, see Bowman, John D., “Trends in Midwestern Farm Land Values, 1860–1900” (unpublished doctoral dissertation, Yale University, 1964), ch. 3.Google Scholar Even extreme movements in these shares, however, would not upset the present conclusions.

⁹ Another basic question about the effects of land scarcity is treated elsewhere, but cannot be taken up here. It is natural to wonder whether the availability of extra land is not a factor making unskilled labor more scarce and incomes more equally distributed. The answer is clearly yes if by the “availability” of land one means its availability to small owner-cultivators rather than to large estates alone. In this sense the availability of land has surely been an equalizing force in America. The effects of the aggregate physical availability of land are less obvious. The extra land may still be an equalizing force, but as I have argued in another paper (“Fertility and the Macroeconomics of Inequality,” University of Wisconsin, Institute for Research on Poverty, Discussion Paper No. 219–74, August 1974),Google Scholar its influence on the income distribution cannot have been great, since the supply expanded more rapidly before 1900, when income inequality was widening, than it did during the twentieth-century equalizing of incomes.

¹⁰ Tostlebe, Capital in Agriculture, Appendix A.

¹¹ Goldsmith, Raymond W., The National Wealth of the United States in the Postwar Period (Princeton: National Bureau of Economic Research, 1962), pp. 94, 95, Tables A-5, A-40 through A-43Google Scholar; Deane, Phyllis and Cole, W. A., British Economic Growth, 1688–1959, 2d ed. (Cambridge University Press, 1969), pp. 270–271.Google Scholar

¹² See, in addition to the discussion paper cited in Footnote 9 above, Williamson, Jeffrey G., “The Relative Rental Price of Men, Skills, and Machines: 1816–1948,” University of Wisconsin, Graduate Program in Economic History, Discussion Paper EH 74–25, July 1974.Google Scholar

¹³ Gallman, Robert E., “Changes in Total United States Agricultural Factor Productivity in the Nineteenth Century,” in Farming in the New Nation: Interpreting American Agriculture, 1790–1840, ed. Kelsey, Darwin P. (Washington: Agricultural History Society, 1972).Google Scholar

¹⁴ U.S. Department of Agriculture, Agricultural Production and Efficiency, Major Statistical Series of the USDA, vol. 2, USDA Agricultural Handbook No. 365 (Washington, D.C.: G.P.O., 1970), p. 9Google Scholar; USDA, Agricultural Statistics—1972 (Washington, D.C.: G.P.O., 1972), p. 536Google Scholar; USDA, Economic Research Service, Farm Real Estate Historical Series Data: 1850–1970 (Washington, D.C.: G.P.O., 1973), Table 1Google Scholar; Kendrick, John W., Productivity Trends in the United States (Princeton: Princeton University Press, 1961)Google Scholar; Tweeten, Luther, Foundations of Farm Policy (Lincoln: University of Nebraska Press, 1970).Google Scholar

¹⁵ These sectoral patterns of factor use are documented using the 1963 input-output table in the extended discussion-paper version of this article. It appears likely that the sectoral contrasts in the relative use of unskilled labor were stronger for earlier years.

¹⁶ The results that follow have also been derived by applying parameter values from the United States in the twentieth century to a general-equilibrium model in the appendix to the discussion-paper version of this article.

¹⁷ The conclusions of this paragraph contradict the spirit of the conclusions recently reached by Eilenstine, Donald and Cunningham, James P., “Projected Consumption Patterns for a Stationary Population,” Population Studies, XXVI (July 1972), 223–231.CrossRefGoogle Scholar They rightly point out that discussions of changes in population arising from changes in fertility overlook the fact that the age distribution of the population will also adjust to the change in fertility. The shift to a younger population after a rise in fertility would entail age-distribution effects on patterns of consumption that tend to be of opposite sign to the family-size effects. They then produce estimates suggesting that the shifts in consumption resulting from a stationary population would not be significant. This last conclusion may mislead. Their own estimates show the share of food in total consumption would be cut by 1.2 percent of itself if population had been stationary rather than growing in the pattern of the late 1950's. This shift would have a greater percentage effect on the terms of trade for agriculture, since both supply and demand are relatively price-inelastic in the agricultural sector. And the shift in the terms of trade would be likely to have an even greater percentage effect on the returns to farm land. In addition, their estimates make no allowance for the ultimate labor-market effects of population growth. Whatever affects labor supply and product per capita, as noted in the text, shifts the demand for food through Engel's Law in a way not captured by their estimates.

¹⁸ U.S. Department of Commerce, Bureau of Economic Analysis, Long Term Economic Growth, 1860–1970 (Washington: G.P.O., 1973), pp. 183, 185, 201.Google Scholar For a similar documentation of the rise in the nonfarm share of the value of final output of food, natural fiber clothing, and tobacco products, see Kuznets, Simon, Modern Economic Growth (New Haven: Yale University Press, 1966), pp. 274–276.Google Scholar

¹⁹ Muth, Richard F., Cities and Housing (Chicago: University of Chicago Press, 1969),Google Scholar and his “Permanent Income, Instrumental Variables, and the Income Elasticity of Housing Demand,” Washington University (St. Louis), Institute for Urban and Regional Studies, Working Paper EDA 12 (December 1970).Google Scholar

²⁰ Gold, Neil W. and Davidoff, Paul, “The Supply and Availability of Land for Housing for Low and Moderate Income Families”, in U.S. President's Committee on Urban Housing, Technical Studies, II (Washington: G.P.O., 1969), p. 351.Google Scholar

²¹ Floyd, John E., “The Effects on Farm Price Supports on the Returns to Land and Labor in Agriculture,” Journal of Political Economy, LXXIII (April 1965), 148–158CrossRefGoogle Scholar; Schultze, Charles L., The Distribution of Farm Subsidies: Who Gets the Benefits? (Washington: The Brooldngs Institution, 1971)Google Scholar; Lidman, Russell, “The Distributional Implications of Agricultural Commodity Programs,” University of Wisconsin, Institute for Research on Poverty, Discussion Paper 142–72 (September 1972)Google Scholar; and Boxley, Robert F. and Anderson, William, “Price-Support Programs and Land Values: The Case of Flue-Cured Tobacco,” in Government Spending and Land Values, ed. Harriss, C. Lowell (Madison: University of Wisconsin Press, 1973), pp. 76–103,Google Scholar

Whether government farm support policies tend to raise farm rents depends on how they are administered. Rents, along with values, will be raised by price supports for agricultural products, or by payments for diverting acreage from production if the tenant receives these payments. If the landlord receives the payments for divertment, it is unlikely that rents would be raised by the governments assistance. Current policy reportedly makes an effort to see that the diversion of rented acreage from production is compensated by a payment from the government to the tenant.

²² The same conclusion is supported by the general-equilibrium model given in the appendix to the discussion-paper version of this article.

²³ Kendrick, Productivity Trends; Williamson, Jeffrey G., “Late Nineteenth-Century American Retardation: A Neoclassical Analysis,” Journal of Economic History, XXXIII (1973), 581–607; Gallman, “Changes in Total United States Agricultural Factor Productivity in the Nineteenth Century”; USDA, Agricultural Statistics—1972.CrossRefGoogle Scholar

²⁴ An additional shift that might be thought of as an aggregate land-saving technological change has already been introduced above as a shift in product demand: across this century, and especially before the late 1950's, there has been a dramatic rise in the share of production of finished agricultural products taken by processing and retailing, leaving a dwindling share for the land-intensive farm sector.

²⁵ Bowman, “Trends in Midwestern Farm Land Values”; Davis, Lance E., “The Investment Market, 1870–1914: The Evolution of a National Market,” Journal of Economic History, XXV (1965), 355–393.CrossRefGoogle Scholar

²⁶ Foreign demand for American agriculture may also have shifted up slightly between 1900 and 1915. We know that price ratios shifted in favor of farmers in this period, and foreign demand for American food did jump in 1915 due to the war. On the other hand, the share of food exports in farm output and income had been falling in the first decade of this century, so that it was not higher in 1915 than in 1900. For this reason, I have not emphasized the possible shift in foreign demand as an influence on land rents in this period.

²⁷ We lack data on the course of farm rents during the 1940's. Farm land values rose from 1940 to 1945, but it appears likely that real rents—the marginal product of land—were no higher (or even lower) in 1945 than at the start of the decade. Note that real rents hardly rose at all between the censuses of 1940 and 1950 (Figure 1). Furthermore, the fact that real farm land values rose less than the acceleration in nominal farm (output or land) prices in a period of steady interest rates gives a theoretical presumption that rents fell: the equilibrium formula relating rent to purchase price (see text) implies that if expected increases in nominal land prices () resembled the actual ones for 1940–1945, real rents probably did not rise.

²⁸ The year 1915, while convenient as a peak year for midwestem farm land rents and land values, happens to be a year of fairly high unemployment. Yet none of the conclusions that follow would be altered if the starting year for the period under examination here were 1913 or 1916 or 1918, years of fuller employment.

²⁹ The government also propped up agricultural demand, of course, with pricesupport programs. These are not given much attention here, however, since the price supports have become much less relevant to farm land values than have government restrictions on acreage from the mid-1960's through 1973.

³⁰ The data used here are those for land in the 48 states not used for farms, forest, or grazing. This “other land” category includes “urban, industrial, and nonfarm residentiar areas; military lands; and roads, railroads, ungrazed desert, rock, swamp, and miscellaneous other areas.” (USDA, Agricultural Statistics—1972, p. 506.) About the only thing these diverse uses have in common is the one thing they should have in common for present purposes: they represent the more income-elastic land uses.