Loading field map
Why do firms hire the number of workers they do, and what determines the wages they offer? For more than a century, economists have given sharply different answers, and the debate has never settled into a single orthodoxy. The central tension runs between two intuitions: that hiring and pay are governed by the same forces of supply and demand that clear any other market, and that labor markets are fundamentally different—shaped by institutions, power, incomplete information, and the fact that workers are not commodities. Labor demand theory is the story of how successive frameworks have tried to reconcile or choose between these two visions.
Neoclassical Labor Demand Theory, built from the marginalist revolution of the late nineteenth century, gave the first rigorous answer. A profit-maximizing firm hires workers up to the point where the value of the marginal product of labor equals the wage. The labor demand curve slopes downward because each additional worker adds less to output than the previous one, given fixed capital. In this framework, wages and employment are determined simultaneously by supply and demand; there is no involuntary unemployment except as a temporary disequilibrium. The theory replaced the classical labor theory of value—which held that labor embodied in goods determined their prices—with a subjective, marginalist account. For the neoclassical economist, a minimum wage above the market-clearing level would reduce employment because it priced low-productivity workers out of jobs. The framework assumed perfect information, homogeneous labor, and frictionless adjustment. It became the benchmark against which all later theories defined themselves.
Keynesian Labor Demand Theory, emerging from the Great Depression, rejected the neoclassical claim that wage cuts could restore full employment. In John Maynard Keynes's 1936 General Theory, the causal arrow ran from aggregate demand to employment, not from wages to employment. Firms hire only as many workers as needed to produce the goods that consumers, investors, and the government are willing to buy. If aggregate demand is deficient, workers can offer to accept lower wages, but that will not raise employment because lower wages reduce spending power and thus further depress demand. The concept of effective demand—the idea that actual purchases, not notional supply, drive production—became the core of the Keynesian alternative. Involuntary unemployment was not a temporary glitch but a normal outcome of a market economy. For policy, this meant that fiscal stimulus, not wage flexibility, was the remedy for recessions. The Keynesian framework coexisted uneasily with the neoclassical one for decades, but by the 1970s it came under pressure from microfoundations critiques: how could rational firms and workers systematically fail to coordinate on a mutually beneficial wage cut? That question opened the door for new theories that tried to explain involuntary unemployment without abandoning individual rationality.
While the neoclassical and Keynesian traditions debated aggregate mechanisms, Old Institutional Labor Economics took a different starting point. Emerging in the early twentieth century, it argued that labor markets are governed by custom, law, and organizational routines that cannot be reduced to supply and demand. Firms do not set wages by marginal productivity alone; they follow internal pay scales, seniority rules, and norms of fairness. The labor contract is incomplete, and workers are not interchangeable commodities. This tradition documented the rise of internal labor markets—career ladders within firms where wages are attached to jobs, not to individuals—and argued that these structures reflected historical accident and worker power as much as efficiency. The Old Institutionalists were skeptical of the neoclassical claim that observed outcomes are optimal. Their work provided a rich descriptive alternative but lacked a unified theoretical engine, which left it vulnerable to the charge that it was merely cataloguing exceptions.
New Institutional Economics of Labor, which took shape from the 1970s onward, shared the Old Institutional focus on rules and contracts but reversed the explanatory logic. Where the Old School saw institutions as constraints imposed by power or history, the New Institutionalists asked whether those same institutions could be understood as efficient responses to transaction costs. Firms exist, in Ronald Coase's famous argument, because using the price mechanism is costly. Within the firm, incomplete contracts, monitoring difficulties, and firm-specific skills create problems that internal labor markets solve. The New Institutional framework absorbed the Old Institutional observation that wages are attached to jobs rather than workers, but reinterpreted it as an efficient adaptation: job ladders provide incentives for workers to invest in firm-specific human capital. This created a sharp disagreement: are observed institutions efficient or are they path-dependent and potentially inefficient? The question remains unresolved, and both positions have active defenders.
Efficiency Wage Theory, launched by a cluster of papers around 1979–1984, offered a new microfoundation for involuntary unemployment. If firms pay a wage above the market-clearing level, workers have more to lose if fired, so they work harder and are less likely to shirk. The higher wage also reduces turnover, attracts better applicants, and improves morale. From the firm's perspective, paying above the market rate can be profit-maximizing because the productivity gain offsets the higher wage bill. The result is a labor market in which some workers are unemployed—they would gladly work at the prevailing wage but cannot find jobs—yet firms do not cut wages because doing so would reduce profits. This explanation of involuntary unemployment did not rely on Keynesian demand shortfalls; it worked even in a fully rational, microfounded model. Efficiency Wage theory thus preserved the neoclassical assumption of optimizing firms while rejecting the conclusion that markets clear. It coexists with the neoclassical framework as a modification rather than a wholesale replacement: in sectors where monitoring is easy and turnover costs are low, the neoclassical model may apply; where effort is hard to observe, efficiency wages dominate.
Search and Matching Theory, developed from the 1970s onward by Peter Diamond, Dale Mortensen, and Christopher Pissarides, introduced a different friction. Workers and firms do not find each other instantly; they must search, and the process is costly and stochastic. The matching function—an aggregate relationship between unemployment, vacancies, and hires—captures this friction. Firms post vacancies, workers apply, and matches occur at a rate determined by the thickness of the market. Unemployment and vacancies can coexist because the matching process is not instantaneous. The theory generates involuntary unemployment from the fact that workers and firms cannot costlessly coordinate. It differs from Efficiency Wage theory in its mechanism: unemployment arises from search frictions, not from incentive problems. But the two frameworks are complementary rather than rivals. Modern models often combine them: search frictions explain why unemployed workers do not instantly find jobs, while efficiency wages explain why firms do not cut wages to attract them. Search and Matching theory also provides a natural account of the Beveridge curve—the negative relationship between unemployment and vacancies—and has become the dominant framework in macro-labor for analyzing business-cycle fluctuations in unemployment.
The Design-Based Approach, which gained momentum in the 1990s, was not a theory of labor demand in the same sense as the earlier frameworks. It was a methodological revolution that changed what counted as evidence. Instead of estimating structural parameters from a theoretical model, design-based researchers used natural experiments, instrumental variables, regression discontinuity, and difference-in-differences to isolate the causal effect of a specific policy—a minimum wage increase, a payroll tax cut, a change in unemployment insurance—on employment. The approach was explicitly agnostic about which theoretical framework was correct; its goal was to estimate a treatment effect, not to test a structural model. This created a productive tension with the structural traditions. Design-based studies could show, for example, that a minimum wage increase had little or no effect on employment in certain settings, challenging the neoclassical prediction of job loss. But they could not easily say why the effect was small—whether because of monopsony power, efficiency wages, or search frictions. The Design-Based Approach thus coexists with the older frameworks as a source of empirical discipline: it constrains what theories must explain, but it does not replace them.
Behavioral Labor Demand Theory, emerging around 2000, imported insights from psychology and experimental economics into the study of firm behavior. Workers care about fairness: they reduce effort if they perceive their wage as unfair, even if they cannot be caught shirking. Firms, anticipating this, may pay above-market wages not because of a monitoring problem (as in Efficiency Wage theory) but because a wage perceived as unfair damages morale and productivity. Reference dependence—the idea that workers evaluate outcomes relative to a reference point, such as their previous wage or the wage of a coworker—can generate downward wage rigidity: firms avoid nominal wage cuts because workers treat them as losses and retaliate. Behavioral theory extends and challenges Efficiency Wage theory by adding psychological mechanisms that do not require the assumption that workers are purely self-interested or that effort is unobservable. It also challenges the neoclassical assumption that firms set wages based solely on marginal productivity and market conditions. The framework is still young, and its main contribution so far has been to show that laboratory and field experiments reveal systematic deviations from the predictions of both neoclassical and standard efficiency-wage models.
Today, no single framework dominates labor demand theory. The neoclassical model remains the default benchmark, especially in textbook treatments and in settings where frictions are small. Search and Matching theory is the workhorse of macro-labor and is used to analyze unemployment dynamics, vacancy posting, and the effects of labor market policies. Efficiency Wage theory is invoked to explain wage rigidity and involuntary unemployment, often in combination with search frictions. The New Institutional Economics of Labor continues to inform the study of internal labor markets, contracting, and firm boundaries. The Design-Based Approach has reshaped empirical practice: most applied papers now prioritize causal identification, and the burden of proof has shifted onto structural models to show that their assumptions are consistent with the reduced-form evidence. Behavioral Labor Demand theory is an active frontier, though it has not yet produced a unified alternative to the older frameworks.
The major disagreements today are three. First, are observed labor market institutions efficient adaptations to transaction costs, or are they shaped by power and historical contingency? Second, should unemployment be understood primarily as a search friction, an incentive problem, or a demand deficiency? Third, what is the right relationship between structural modeling and design-based empirical work—should the two be integrated, or do they answer fundamentally different questions? These debates are unlikely to be resolved by a single knockout argument. Instead, the field has settled into a productive pluralism in which different frameworks are deployed for different questions, and the best research often combines insights from several traditions.