Rational Expectations in American Agriculture, 1867-1914

Abstract

DESPITE historical importance of Populist uprising in late nineteenth century, causes of agrarian unrest which culminated in that movement remain obscure. Economic historians have become increasingly dubious of justifications advanced by Populists themselves, partly because many of Populist programs (such as Free Silver plan) were little more than schemes for an involuntary redistribution of wealth in favor of farmers, and partly because many of farmers' stated grievances fail to appear in aggregate economic statistics of period (Bowman, 1965; DeCanio, 1974a; Bowman and Keehn, 1974; North, 1974). The most comprehensive recent study concludes that the agrarian protest of late nineteenth century was not a simple, straightforward consequence of economic factors as many economic historians have believed (Klepper, 1974, p. 285). There is some evidence that cyclical economic fluctuations coincided with upsurges of protest (Bowman and Keehn 1974; Klepper 1974), but agrarian spokesmen of time (as well as subsequent historians) attempted to identify long-standing structural problems of agricultural sector as ultimate explanation of farmers' distress. One recurring theme in historical explanations of agrarian unrest locates source of farmers' difficulties in their perceptions of and responses to economic requirements of market. Thus, Mayhew suggests that protests of Grangers and Alliancemen were a reaction to commercialization of agriculture. This commercialization may have increased farm incomes, but it also made farmers subject to impersonal market forces (Mayhew, 1972). Farmers' failure to understand operation of markets for their products is featured in textbook accounts of Populist period (Davis, Hughes, and McDougall, 1969, p. 368; Gray and Peterson, 1974, p. 320; and North, 1974, p. 134). Econometric studies of price-responsiveness of late nineteenth and early twentieth century American agriculture have shown that sector as a whole responded properly to market prices in both choice of crop mix and choice of technique (Nerlove, 1958; Fisher and Temin, 1970; Hayami and Ruttan, 1971; DeCanio, 1973). These investigations of farmers' responses to output and input prices, however, do not indicate whether estimated agricultural response parameters were in any sense optimal, nor can their fixed-parameter estimation techniques reveal whether farmers' behavior changed in an appropriate manner as underlying market conditions changed. This paper goes beyond previous studies by testing directly a rational expectations hypothesis for American agriculture during Populist period. Using a varying-parameter estimation methodl it is possible to trace changes in supply response parameters over time, and to compare those parameter variations with variations implied by a model of rational price expectations. We will show that changes in farmers' price expectations were indeed consistent with theory of rational expectations. Since our estimates are based on same statewide aggregate data used in previous supply response studies, it is not possible to conclude from our results that all farmers formed rational expectations, but existence of rational expectations in aggregate leads us to reject notion that farmers as a group were unable to Received for publication October 10, 1975. Revision accepted for publication March 15, 1976. ' The helpful comments of Paul David, Stanley Engerman, Jacob Metzer, Joel Mokyr, Marc Nerlove, William Parker, Merton Peck, Edward Prescott, Joe D. Reid, Jr., Peter Temin, and participants in seminars at University of Pennsylvania and Columbia University are gratefully acknowledged. Responsibility for errors remains ours. The research was supported in part by NSF Grant GJ-l 154X3 to National Bureau of Economic Research, and by NSF Grant SOC75-08056. I Recent developments in theory of models with varying parameters are discussed in Cooley (1971) Rosenberg (1973) and Cooley and Prescott (1973a, 1973b, 1973c, and 1976).