by Ibanca Anand
The years between 1935 and 1965 marked a second “Agricultural Revolution” for the United States (Rasmussen 1962). The introduction and widespread adoption of farm machinery, genetically modified crops, artificial insemination, chemical fertilizers and pesticides, and other emergent technologies fundamentally transformed what it meant to farm in midcentury America. By the 1960s, the productivity of “the American farmer” had tripled. With labor demand so drastically reduced, much of the agricultural labor force struggled to stay gainfully employed. Many farmers left for cities: between 1920 and 1970, rural communities underwent a dramatic outmigration of at least 40 million former farmers (Shideler 1973, 291). Those who successfully modernized became newly dependent on a ballooning system of agricultural corporations, including tractor manufacturers, seed companies, and processing facilities. These potential social costs, however, were outside the purview of those who saw in this transformation the marrow of progress.
While some agricultural thinkers in the twentieth century envisioned, analyzed, and preached a story of technological success, we need not endorse their interpretations. Political economy can bring what they excluded back into view. If intellectual history allows us to examine the thoughts and emotions of historical actors, political economy equips us with tools to critique them. Actors’ ideas become less significant than the implications of those ideas. On the terrain of their own intellectual blueprints, we can challenge their guiding ideologies, revealing them as historical specters: ghosts that continue to haunt the intellectual landscapes we inhabit today. In attempting to explain the ‘miraculous’ productivity growth in midcentury American agriculture, for example, techno-optimistic economists tackled the emerging problem of a measurement “residual” in ways that embedded a bias for capital over labor into subsequent growth theory. To understand these conceptual maneuvers, I turn to the intellectual contributions of one of the residual’s main theorists, Lithuanian-American agricultural economist Zvi Griliches (1930–1999).
***
Now remembered as a period of material prosperity, the postwar American economy was in fact perceived by many living through it as incredibly fragile and destined for depression (Galbraith 1993). In response to a series of forecasts predicting a massive spike in unemployment, the U.S. government passed the Employment Act of 1946. As originally conceived, the act considered economic stability to be the chief goal of federal economic policy (Collins 2023, 16). By the early 1950s, the goal had shifted: economic growth was now deemed the most important safeguard of the national economy. This shift solidified and radicalized by the 1960s, as policymakers came to consider even unemployment secondary to the higher end of growth.
Though not previously a core concern of analysis, “growth theory” became a cornerstone of postwar economics. Early researchers, however, were faced with what seemed to be an intractable problem. In the 1950s, study after study confirmed an apparent paradox: the causes of a significant portion of the sources of economic growth resisted explanation. The Cobb-Douglas production function, introduced in the late 1920s and refined over subsequent decades, was foundational to American growth theory, with Robert Solow’s (1956) formulation quickly becoming the consensus. This function separated labor and capital inputs, each raised to a variable factor, to explain overall GDP growth. But measurements of growth in labor or capital inputs did not correspond to measurements of overall output growth when economists began empirically testing the function. One study calculated that, while real national income had grown at a rate of 1.9% annually between 1870 and 1950, the annual growth rate of the combined index of capital and labor inputs had been only 0.3% (Fabricant 1954). In other words, economists could not account for more than three quarters of the economic growth rate. After coming to similar empirical conclusions, American macroeconomist Moses Abramovitz (1956, 11) famously wrote in 1956, “the indicated importance of this [residual] element may be taken to be some sort of measure of our ignorance about the causes of economic growth.” To avoid this aura of ignorance, the concept of the residual was often repackaged as “total-factor productivity,” “multi-factor productivity,” or “technical change.”
Agricultural economists were among the first to note the phenomenon of large residuals. As early as 1950, D. Gale Johnson had shown that the growth of agricultural output in the United States had far outstripped the growth in conventionally measured inputs in recent decades. Johnson’s colleague, Theodore Schultz (1956), similarly noticed a residual and made the study of changes in relative inputs in agriculture a core element of his research program in the early 1950s (Krueger and Taylor 2000, 181). For Schultz, much of the residual could be explained by public and private investments in education and research. By the early 1960s, Schultz (1961) would pioneer the concept of human capital in agriculture and question the sharp distinction between labor and capital categories entirely. But it was Schultz’s graduate student, Zvi Griliches, who theorized the concept of “technical change” in ways that would have lasting consequences on how productivity statistics are calculated across all sectors to this day (Heckman 2006).
Griliches’s training took place in the heady environment of the University of Chicago’s economics department in the mid-1950s. He wrote his dissertation on the diffusion of hybrid corn and in 1957 began his research on fertilizer use—two early demonstrations of his interest in agricultural (and commercial) technologies. These studies allowed him to empirically ground the abstract and elusive concept of “technical change.” Like his mentor Schultz, Griliches became captivated by questions of growth and productivity in agriculture, a sector where the labor population was decreasing and the ratio of capital to labor was rising apace. Yet unlike Schultz, who emphasized the role of the public sector, Griliches located the productive forces in the private sector. Dissatisfied with the prospect of a residual where productivity rose out of the ether to produce economic growth, Griliches attempted to demystify this thorny variable.
Griliches (1963) attributed large multi-factor productivity results—the euphemistically renamed residual—to a combination of specification, measurement, and weighting errors. After correcting for these differences in the statistical data for agriculture, Griliches turned what previously looked like an increase of 48% in multi-factor productivity for agriculture into a decrease of 6% for the period between 1940 and 1960. This did not mean that Griliches found agricultural productivity to have declined, but rather, that much of the phenomenon of productivity could be endogenized to the labor and capital variables.
One of Griliches’s (1960) key observations was that labor’s share in growth accounting had been overstated, whereas the contribution of capital had been underestimated. This followed in the footsteps of Schultz, who had been making an argument about “efficiency” in agriculture for some years. Schultz (1947) had claimed that the marginal product of labor in agriculture was substantially lower than existing agricultural wage rates, whereas the marginal product of capital was substantially above conventional bank or mortgage rates. Accordingly, Griliches recommended a reweighting of these two inputs in the production function for agriculture.
Griliches (1963, 337) doubled down on the devaluation of labor by emphasizing the relative elasticities of substitution between the inputs. He challenged the Cobb-Douglas function’s assumptions of a ‘unity’ of substitution, in which labor and capital inputs are easily substitutable:
From 1950 to 1961, the ratio of machinery inputs to labor inputs increased by 74 per cent. In the same period farm wages rose only by 9 per cent relative to machinery prices, implying a substantial decline in the ‘share’ of labor in agriculture. This, of course, could be explained by an elasticity of substitution that was larger than unity.
Griliches then further tweaked the function to account for capital’s primacy in the production process. Claims that the USDA had overestimated the rate of depreciation for machinery were added to bolster the capital share (Griliches 1960). The role of fertilizer was emphasized; in one instance, Griliches identified the coefficient on the marginal product of fertilizer as 200 in 1949 and 110 in 1959. The coefficients on labor’s marginal product, meanwhile, were estimated at 1.17 and 2.02, respectively (1964, 969).
Griliches also innovated, albeit problematically, the idea of “quality change.” His argument was that both labor and capital inputs took insufficient account of the dramatic improvements in quality. For labor, this meant an increase in education levels (i.e., human capital theory) and, for capital, this meant improvements in technology. The idea of quality change, however, is not exactly apolitical, either. At one point, Griliches (1964, 970) surmised that any “differences that remain may be due to the lower-than-average quality of the workers (the hired ones)” compared to “family workers,” by which he meant white middle-class farmers of the imagined “American family farm.” Quality then became a slippery concept for Griliches, especially when he considered the demographic changes in American agriculture.
Griliches’s quality index also problematically assumes that increased prices, whether in the form of higher wages or higher prices for products, are genuine reflections of quality. This adopts a market-based approach to value, in line with theorists like Friedrich Hayek and Milton Friedman, who consider prices as signals of consumer preferences. What it does not account for, however, is that prices (whether in the form of low wages or inflated prices) could be vehicles of exploitation in a structurally unequal economy. ‘Quality,’ for Griliches, thus became a composite measure that collapsed the complexity of price politics into one seemingly objective measure (Banzhaf 2001, 336).
Quality adjustments were covertly political tools. They explained rising prices by attributing them to the increase of innovation rather than the spread of monopolies. Moreover, quality-adjusted indexes and functions offered a convenient and ostensibly neutral justification for dismissing workers’ demands for higher wages and farmers’ demands for higher profits: the coefficient on their contributions to economic growth was simply too low. In this way, production functions not only explained, but simultaneously prescribed (Callon 2006). Viewing them in this way opens up urgent new questions for intellectual and economic historians alike. To what extent did employers, for example, consider labor’s supposedly meager contribution to output share as a reason to invest in machinery and pay workers less? How might the maximization of sectoral growth—rather than farmers’ economic security, for example—have served as a new criterion of agriculture’s well-being? The political stakes and material implications of these production functions may not have been interrogated by theorists of growth and productivity at the time, but their diffuse and potentially consequential repercussions certainly warrant our examination today.
Griliches’s manipulation of growth theory’s residual was far more than a mere technicality. By discounting labor in the production process and emphasizing the returns to capital, Griliches provided an intellectual justification, even if unwittingly, for industrialized agriculture. For Griliches and his colleagues, productivity led to growth, and productivity was downstream from capital investment in technologies provided by the growing infrastructure of agribusiness. Values like self-sufficiency, sustainability, or small-scale production were occluded from a calculus that viewed success in one dimension: output. Moreover, the maximization of output and the maximization of profit correspond conveniently under capitalism (Chambers 2022). If the farmer or farmworker had interests, they simply fell outside the scope of analysis.
Yet we can transcend our historical actors’ blind spots. Economic life can be viewed through prisms beyond the constraining one of neoclassical economics; this is, after all, the promise of our (re)turn to political economy. Approaching the thought-worlds of bygone days with more emancipatory analytical frameworks can help pry loose ideas that no longer serve us. And in clearing the cobwebs of ideas past, we may bring into sharper focus elements and agents long overlooked, yet haunting economic analysis from the periphery to this day. Postwar agricultural economics may have conveniently occluded the experiences of farmworkers, small landholders, or peasants worldwide from their analysis; but conjuring them remains a live possibility for us as we consider and construct a more worthy agricultural system in the present and beyond.
This think piece is part of a JHI Blog forum: “The Return of Political Economy in Intellectual History.”
Ibanca Anand is a PhD Candidate in modern American intellectual and economic history at Johns Hopkins University. Her dissertation examines the political transformations within twentieth-century U.S. agriculture through the lens of economic ideas and interventions. At JHU, she serves as a co-coordinator for the Modern America Seminar and as a steward for TRU-UE Local 197.
Edited by Minke Hijmans.
Featured image: Digital montage of “A Holt Harvester threshing barley at the Tule Lake Segregation Center, 1944” and Table 7 of “Research Expenditures, Education, and the Aggregate Agricultural Production Function” by Zvi Griliches, 1964.
