Toward an Interdisciplinary Political Economy of Wages

Abstract

This article reviews how some social scientists have transcended disciplinary boundaries in their scholarship on wage formation and proposes specific pathways for further trespassing. It contends that an interdisciplinary political economy should connect power to prices. This means that economists should bring power—including political influence and social status—into the heart of their analysis of wage formation, and sociologists and political scientists should extend their analysis of social structure and power dynamics to their endpoint in wage levels. The article concludes with suggestions for how to integrate political and/or sociological perspectives with economic models, both in theory and in empirical research.

The fundamental struggle between capital and labor over the returns from production offers fruitful terrain for interdisciplinary political economy. Some economists have ventured beyond their home turf to embrace the study of power, and other social scientists have stretched into the economists’ realm to examine wage formation. And scholars in these various disciplines have collaborated productively in research projects and policy institutes across the globe. Yet social scientists could go further toward a truly interdisciplinary political economy of wages.

This essay reviews some of the ways in which social scientists have transcended disciplinary boundaries in their scholarship on wage formation and proposes pathways for further trespassing. Specifically, it argues that economists should bring power into the heart of their analysis of wage formation and expand their conception of power to include political influence and social status. Most critically, they should begin their analysis from a presumption of employer power, as in a monopsony model, rather than a perfect market. Labor market monopsony refers to a situation in which a single employer dominates a market, such as in a given geographical area. The firm faces an upward-sloping labor supply curve, meaning that it can reduce wages without losing all of its workers, unlike in a perfectly competitive market. And firms with greater monopsony power will tend to face a lower elasticity of labor supply, meaning that fewer workers will be likely to quit if the firm cuts wages. ¹ Economic models predict that monopsonists will employ fewer workers and pay them less than in a competitive market. ² The concept of monopsony can also be usefully applied when there are multiple employers and the employers have a substantial advantage in power relative to workers. ³ Monopsony in this broader sense is especially amenable to interdisciplinary scholarship.

Some economists have been moving in this direction in recent years, with a substantial increase in the literature on monopsony after 2000 and a veritable boom after 2020. ⁴ Monopsony models have the potential to reframe policy debates because they strip the status quo of an unwarranted positive valence, including the assumption that wages reflect marginal productivity (the amount a worker contributes to additional output) or the “just deserts” from skill and effort. ⁵ They challenge the presumption that policies to strengthen labor power generate a “deadweight loss”—a loss of societal welfare due to market distortions—and demonstrate how such policies can enhance welfare. And they facilitate integration with other social sciences that focus more on power relations in labor markets.

Meanwhile, political scientists, sociologists, and other social scientists have stressed the power dynamics and the social context of labor relations, yet they have rarely extended that analysis to wage levels. They explore how employers exploit their workers, but not how much that costs the workers. Political scientists’ studies of labor policies and sociologists’ studies of labor practices have major implications for wages, yet these scholars often stop short of specifying these effects. In doing so, they obscure how their research applies to policy debates. They could engage more directly with economists and economic models both critically and constructively. That is, they could demonstrate how their research challenges standard economic analysis, and they could propose how to refine economic models to incorporate political and social factors.

In terms of the causal chain illustrated in Figure 1, this means that economists would stretch backward to power and politics, and other social scientists would extend forward to wages. The goal would not necessarily be convergence on a common model but rather more crossover between disciplinary research paradigms and greater integration of political, social, and market forces in the study of wage formation. The figure presents a causal chain that runs from politics to government policies, to business practices, to power balances, to supply and demand, and finally to wages. In reality, of course, the causality does not run in only one direction, and all these variables (and more) are interrelated in complex ways. The “circularity” arrows indicate elements of these interaction effects, suggesting, for example, that wages and power balances also shape politics.

Figure 1.

A schematic model of wage formation.

The figure also provides a simplified rendition of the core research agenda of three disciplines. Political scientists tend to focus primarily on explaining how politics affects policy outcomes. Sociologists study how social context shapes business practices, or how business practices affect worker well-being. And economists examine how market variables affect wage levels. Yet scholars in these disciplines often stretch beyond the confines of their home turfs, and they could do so even more.

An interdisciplinary political economy of wages would begin from the recognition that labor markets, like all markets, are embedded in politics and society. ⁶ And this brings power, broadly defined, to the center of the analysis. Wages are driven by balances of power, which are in turn shaped by market institutions, including government laws and regulations, private sector standards and practices, elite ideologies, and societal norms. ⁷ John R. Commons articulated such an interdisciplinary perspective in his brand of institutional economics in the early twentieth century. ⁸ He identified the inequality of bargaining power between employers and workers as a defining feature of labor markets:

If [the laborer] has no property of his own sufficient to fall back upon, he is under an imperious necessity of immediately agreeing with somebody who has. This peculiar relation between a propertyless seller of himself, on the one hand, and a propertied buyer on the other, coupled as it is with equal suffrage of both in the politics of the country, has gradually acquired recognition as something sufficiently important for the government to take notice of. ⁹

Commons argued that the government should enhance labor power by passing laws and enforcing regulations on union formation, collective bargaining, minimum wages, and health and safety standards. ¹⁰ Joan Robinson first introduced the term monopsony, and demonstrated how imperfect competition in the labor market shifts bargaining power from workers to employers, dragging down wages and suppressing employment. ¹¹ And she advocated for strong labor unions as a source of countervailing power. ¹² “For without Trade Unions,” she stressed, “monopsony is more powerful, the ratio of real wages to real profits is smaller, thriftiness is greater and the level of employment with a given rate of interest consequently lower.” ¹³

Nonetheless, most contemporary scholarship in labor economics begins with the assumption of perfect markets and proceeds from there. This assumption can provide a useful reference point for pedagogical or heuristic purposes, but it can also be profoundly misleading. For example, a standard economic analysis might compare a labor market in which unions exercise power to one in which they do not. And it would show that union power generates higher wages and less employment than perfect competition, transfers some surplus from firms to workers, and thereby creates a deadweight loss. ¹⁴ Yet there is no market without power relations between employers and workers, and there is no market wage in the absence of struggles over power. So there is no reason to assume that union power would generate a deadweight loss relative to the status quo. And if we adopt a more realistic assumption, that employers enjoy some monopsony power, that leads us to the opposite conclusion: that union power should improve welfare overall. ¹⁵ As Alan Manning notes, a monopsony model of labor markets quickly unravels some of the most enduring puzzles in labor economics, such as why minimum wages do not necessarily reduce employment. ¹⁶ In fact, standard economic analysis shows that in the case of monopsony, a minimum wage can increase both wages and employment. ¹⁷

Figure 2 illustrates how core market relationships can be viewed as balances of power. ¹⁸ If we imagine a spectrum of power balances from employer domination to worker control in labor markets, there is no point along that spectrum (such as the points marked A and B) that represents the power-free or balanced or neutral market. Instead, there are only an infinite array of market governance possibilities, each reflecting a distinct balance of power. So policy measures designed to shift that balance do not constitute interventions in a free market.

Figure 2.

Market governance and market power relationships.

This is not merely a question of semantics. It suggests that we should view power as endogenous to labor markets. That means viewing power as a primary driver of wage outcomes, rather than as a market failure or some other sort of anomaly, and defining power broadly to include political and social as well as market power. A political or sociological perspective offers a way to situate labor markets within a broader institutional context, including laws, practices, and norms. ¹⁹ This could allow scholars to adapt economists’ analytical tools—such as supply and demand models or indifference curves—to address a broader array of dimensions of power. And it could foster empirical models that combine economic, political, and/or social variables and explore the interactions among them.

This essay identifies some fruitful directions in the literature that could be explored further: (1) embracing a more expansive view of the social and political context of wage formation; (2) opening the black box of the firm to investigate how firm-specific features affect wages; (3) exploring variable-sum models of employer-worker relations that address how labor market institutions can foster cooperation (and mediate conflict) between management and labor; (4) recognizing that firms do not simply maximize profits but also pursue other goals such as control (over workers), stability (of prices), market power, market share, and market expansion; and (5) developing a dynamic model of wage formation that captures the interaction between firms’ political and business strategies. And in the concluding section it turns to some specific ways scholars might integrate political and/or sociological perspectives with economic models and apply those models in empirical research.

How Labor Economists Study Power

Let us begin by examining how some economists have incorporated power into models of wage formation. A standard economic model of wages begins with supply and demand. How much do workers have to compete for jobs? And how much do employers have to compete for workers? In a perfect market, supply and demand would be seamlessly matched via the price mechanism, and wages would reflect the marginal productivity of labor (shaped by human capital, such as education and experience). But economists recognize that markets are not perfect, so they examine various deviations, such as unemployment or labor shortages, and they evaluate policy remedies. They also investigate other anomalies (market failures), such as imperfect information or unreliable enforcement—and, of course, power.

Yet labor economists tend to study power in fairly circumscribed ways. They focus particularly on two elements of power: how employers can benefit from market concentration (monopsony in labor markets and/or monopoly in product markets), leading to lower wages; and how workers can benefit from the inability of employers to monitor their effort, leading to higher wages to induce that effort. ²⁰ For example, a dominant firm might account for a large share of a local labor market (monopsony), leaving workers in that region with a narrow range of job options. That would give the employer an edge in bargaining power because it would have more options (for other workers) and workers would have fewer (for other jobs) than in a more competitive labor market. This allows the firm to pay lower-than-market wages, other things being equal. ²¹ Scholars have produced a wealth of evidence documenting monopsony effects on wages. ²² Azar and Marinescu measure labor market power via “markdowns,” referring to the gap between wages and marginal productivity. They report that studies generally find markdowns from 15 to 50 percent, implying that wages would increase by 15 to 50 percent if firms did not have monopsony power. ²³

Economists examine “labor discipline” as another facet of power in labor markets. That is, to what extent can employers control workers’ effort and performance? And to what extent can workers shirk without fear of reprisal? Employment contracts are “incomplete” in that they cannot specify everything the worker must do, monitor that performance, and enforce it. There are some jobs that are amenable to payment by piecework, such as retail sales, call center transactions, garment production, or warehouse work. But in most professions, employers have limited ability to measure effort and performance or tools to induce maximum effort (to “discipline” labor in the language of the subfield). So scholars have hypothesized that employers must pay workers above their “reservation” wages, the minimum wages they would accept to take jobs, to motivate greater effort. ²⁴ This bonus payment will lead workers to prefer their jobs to others, so they will make more effort to ensure that they are not fired.

A sociologist might question the assumption that employers would favor wage-based strategies of retaining workers over non-wage-based ones, because the latter could be more effective and less costly. That is, employers might try to cultivate a favorable work environment to reduce turnover and increase work performance—and possibly to retain workers at a lower wage level. ²⁵ The sociologist might challenge the economists’ propensity to focus on exit mechanisms, such as laying off workers or quitting a job, rather than on voice mechanisms, such as motivating workers or complaining to management. ²⁶ In practice, it is not so easy for employers to replace workers or for workers to switch jobs. This involves substantial transaction costs, both monetary, such as severance costs or moving costs, and less tangible, such as gathering information or negotiating contracts. ²⁷ And as Albert Hirschman notes, the more difficult it is to exercise exit, the more likely actors are to favor voice and to cultivate channels for the expression of voice. ²⁸ Economists are less prone to explore voice: how employers deploy social rewards to foster worker loyalty; how they leverage social hierarchies to intimidate workers; or how they mobilize political channels to enhance their control over workers.

Some economists have focused more on the labor power side of the equation. In a classic work on the topic, Freeman and Medoff explore the role of unions in labor markets, and find that unions contribute to firm productivity and promote a more equal distribution of wages. ²⁹ In fact, they leverage Hirschman to demonstrate how unions exercise voice to improve working conditions. Stansbury and Summers contend that the decline in worker power in the United States, more than the increase in employer market power, explains much of the decrease in the labor share of income. The decline in union power has undermined labor's bargaining power relative to employers, and this has driven down wages over time. ³⁰ Suresh Naidu engages with perspectives from sociology and political science to examine how social networks, workplace conflict, and the dynamics of collective action shape union power in the United States. ³¹

Other economists have explored a broader array of dimensions of power. Blanchflower, Oswald, and Garrett examine what they call “insider power,” which includes an employer's financial position as well as its market power. ³² Their approach directly challenges the classic theory of wage determination and incorporates power into its model. Specifically, they stress that if wages were determined by the external (to the firm) market and not by internal power relations, then the employer's financial status should not affect wage rates—because employers would have to pay and workers would have to accept the “market” rate.

According to the competitive model, an employer is a wage-taker and must set that wage rate which gives workers the market level of utility. There is no scope for bargaining; employees are unable to appropriate any of the returns to an improvement in their firm's prosperity; there are no rents; insider workers and outsider workers are equal. ³³

The authors find in their empirical study that more profitable firms pay their workers more. While this result may appear intuitive to the untrained observer, the authors are right that it violates the assumptions of the classic model. Moreover, they claim more broadly that the classic model cannot explain a substantial range of variation in wages, so they conclude that market forces only set the outside limits within which wage bargaining occurs.

Yet we could take this critique further. We could explore the extent to which power relations could explain some of this residual, or the “indeterminacy” left by the competitive model. And we could reconsider whether it even makes sense to think of (external) market forces as setting the outward bounds of wages and power relations as shaping the bargaining within those bounds. Wouldn’t it make more sense to view power relations as shaping those market forces in the first place?

The parsimonious economic models of labor supply and demand have the virtue of being amenable to the incorporation of other factors, from market fluctuations to policy measures to power imbalances, on a case-by-case basis. And scholars have produced an impressive body of research doing just that. For example, they study how unions transform the bargaining power of employers and workers, how specific changes in laws and regulations affect this balance of power, and how these balances in turn affect wage levels. Or they investigate how gender or racial identity affects bargaining power and wage outcomes. ³⁴ Typically, they introduce these factors one at a time to keep the models and the research strategies manageable, and to isolate specific causal relationships. ³⁵

A political or social perspective shifts the emphasis from specific exit opportunities for employers and workers to a more holistic assessment of the environment in which negotiations take place, and how this affects market outcomes such as wages. This is consistent with a Marxian approach in that it focuses on the fusion of market power and other forms of power and recognizes the propensity for capital to rig the rules in its favor. ³⁶ But it differs in that it highlights the variations on that theme. It puts less emphasis on the common propensities of all capitalist systems, and more on the particular varieties among them. It is less universally pessimistic about the prospects for a fairer and more productive market economy, and more amenable to research on the specific forms of labor market governance.

Some heterodox economists have integrated a Marxian perspective with empirical research on variations over time and across nations, sectors, or firms. For example, McDonough, Reich, and Kotz situate wage bargaining within broader Social Structures of Accumulation (SSA), and analyze how those structures evolve in long cycles over time. ³⁷ Boyer and Saillard and others in the French Régulation School identify the wage-labor nexus as one of five spheres of the political-economic system, along with the financial system, the competition regime, the state, and the international economy. ³⁸

Consistent with the argument advanced here, Dev Nathan proposes that degrees of buyer power (monopsony) should be incorporated into the more standard supply-demand analysis of labor markets. He explicitly links monopsony relative to workers to monopsony relative to suppliers. In global supply chains, dominant multinational buyers can exercise monopsony over workers directly via labor arbitrage, taking advantage of the differential wages for the same tasks performed in different locations. But they can also exercise monopsony indirectly by pressuring weaker supply firms for cost reductions, which then squeeze their workers with lower wages and harsher working conditions. Nathan then explores how the features of specific sectors affect the levels of monopsony power in those sectors. For example, he suggests that multinational buyer firms will have more monopsony power in sectors in which supplier firms are less able to switch to other buyers or other products. ³⁹

How Other Social Scientists Study Wages

Political scientists and other sociologists have naturally focused more on the role of power in labor relations than economists, yet they have mostly stopped short of quantifying the impact of power balances on wages or demonstrating how their insights could be incorporated into economic models of wage formation. Scholars of power resource theory assume that employers have a structural advantage over workers because they control the means of production, but they emphasize that workers can partially offset this advantage by mobilizing their own power resources, and they unpack power into multiple dimensions, such as organizational capacity or political coalitions. ⁴⁰ Offe and Wiesenthal examine how business organizations and labor organizations adopt different associational practices, or logics of collection action, that systematically favor business over labor. ⁴¹ And Fligstein and Fernandez define the power balance between firms and workers in terms of firm control over labor demand and worker control over labor supply. And they stress how state, firm, and worker strategies and organizations interact to set the balance. ⁴²

Jake Rosenfeld addresses the drivers of wage levels more directly. He challenges analyses of wage formation that center on worker attributes, including models of marginal productivity and theories of human capital. ⁴³ He charges that the human capital model builds on three false assumptions: that there is such a thing as a worker's marginal product; that this can be calculated; and that paying workers based on it is a good idea. ⁴⁴ He argues that real-world wages are driven by four things: power, inertia, mimicry, and equity. He defines power to include its institutional, social, and ideological forms as well as bargaining leverage more narrowly conceived. Managers often fall back on inertia: they take for granted the legacy pay scales for certain jobs and only make incremental adjustments to those rates. They also mimic practices elsewhere by surveying pay practices in their industry and following suit. They may claim that they are simply paying the “market rate,” but this rate may reflect common practices more than the iron forces of supply and demand. And they are strongly bound by prevalent norms of fair pay. So even if they can identify differences in performance across workers, they are likely to maintain relatively uniform wages so as not to violate the workers’ sense of fairness. And they may actually favor layoffs over pay cuts because reducing pay would undermine morale for workers whereas layoffs could improve morale for those who remain. Rosenfeld gives us a richer understanding of how employers set wages, but he does not fully flesh out the implications of his research for how much they pay.

Nathan Wilmers makes that final link (from Figure 1) by tying buyer power to wage levels and specifying the mechanisms that link the two. ⁴⁵ He finds that a 10 percent increase in reliance on dominant buyers is associated with a 1.2 percent decrease in supplier wages. He argues that dominant buyers can restrain profits among their dependent suppliers, reducing the rents that might be available to workers. They can circumvent wage norms within firms because they enjoy social distance from the suppliers’ workers. And they do not confront the collective action challenges that more dispersed buyers face when cutting labor costs.

Social scientists also probe inside the firms that set them. Firms are complex arenas of coordination and conflict, with contests over control within firms as well as across firms. ⁴⁶ And those governance structures shape the distribution of financial returns, including wages. For example, corporate governance systems that feature greater incorporation of labor interests, such as the German co-determination system, foster more employment stability and less variation in wages. They also facilitate collaboration between management and labor. Thus a more institutional perspective on labor markets shifts the model from one that presumes a zero-sum relationship (the division of rents) between employers and workers to one that can account for more positive-sum relations (collaboration to raise productivity), and to account for variation over time and across space in the balance between conflict and cooperation. For example, it can explain why German and Japanese firms have been more inclined to develop manufacturing systems that rely on management-labor coordination to enhance productivity and why US firms have been more likely to opt for the “low road” of cutting wages to boost profits. ⁴⁷

Neil Fligstein argues that managers of incumbent firms often deviate from profit maximization. ⁴⁸ They try to stabilize markets by exercising power within the firm, across firms, and in the political arena. They develop business strategies to moderate price competition, including control over labor markets. And they cultivate political strategies to alter laws and regulations or the enforcement of those laws and regulations. ⁴⁹

We can view this as a snowball effect, as depicted in Box 1, in which market power and political power reinforce each other over time. ⁵⁰ This perspective produces a more dynamic model of wage formation. Managers not only seek to maximize returns at a given moment in time, but they deploy business and political strategies to enhance their advantages in the future. For example, US firms have actively fought off proposals for labor market reform to preserve their power advantage relative to labor. And they have lobbied to allow noncompete agreements and mandatory arbitration of labor disputes. Meanwhile, they have taken advantage of these political victories in their business strategies. And they have consolidated their power relative to labor with aggressive union-busting activities, outsourcing, offshoring, franchising, and worker misclassification. Yet this employer snowball dynamic is not inevitable, as can be demonstrated by comparisons across time, such as the United States in 1950–80 versus 1980–2020, or cross-nationally, such as the United States versus Germany. ⁵¹ Countries with stronger labor incorporation at the firm and political level are less vulnerable to this dynamic. ⁵² And even countries with weaker labor representation can combine policy reforms and corporate adjustments to reverse this snowball at key moments, as the United States did in the New Deal period.

Box 1. The Labor Inequality Snowball.

Incumbent firm owners and managers seek both stability and profits.

They can pursue these goals by raising productivity and/or by extracting higher returns at the expense of workers and other stakeholders.

They deploy political strategies, such as lobbying, and business strategies, such as collaboration and collusion, to extract higher returns.

To the extent that they succeed in these strategies, inequality increases between capital and labor, dominant firms and challengers, and top executives and other workers.

Political and market power reinforce each other to produce a snowball effect.

Politics and policy can accelerate, decelerate, or reverse this dynamic.

Political scientists naturally focus more on power in the political sense, but they also recognize that market power and political power can interact and reinforce each other over time. Charles Lindblom shows how we might take this further with his concepts of circularity in markets and circularity in politics (illustrated in Figure 1). ⁵³ Firms not only benefit from market power, but they reinforce it via marketing strategies (circularity in markets). They not only benefit from political power, but they reinforce it via indoctrination strategies (circularity of politics). They seek an ideological hegemony that forecloses options for greater worker power, especially those that constrain private capital and its pursuit of profit.

The United States is at the extreme among industrial countries in the degree to which wages have become unhinged from productivity increases and wage inequality has soared since 1980. ⁵⁴ Labor productivity and average wages increased in tandem from 1945 to 1980. Labor productivity has continued to increase since then, but average real wages have barely increased at all. This has been driven primarily not by globalization and technological change but by politics and policy. ⁵⁵ Congress failed to pass measures to preserve union strength; political leaders took other steps to undermine it; state governments passed laws hostile to labor; and employers increasingly deployed practices to discourage union organization. ⁵⁶ Hence an understanding of wages requires an analysis of politics.

Rosenfeld and other sociologists provide a more thorough depiction of the wage-setting process by adopting a very different research strategy from that of economists, conducting surveys of managers and workers and field research on firms, unions, and other organizations. ⁵⁷ This leaves us with some critical methodological questions. Does this greater richness justify the sacrifice of parsimony in the model? Does it lead to better explanation or prediction of real-world wages? And can we incorporate this broader environmental context without overwhelming the analysis with complexity?

Recognizing this broader context does not necessarily imply that a more complex depiction of the wage formation process is always better. But it does prompt the analyst to look at wage formation more holistically, including power relations, social factors, and more specific market dynamics. And it situates more specific dynamics, such as wage bargaining, in a broader political-economic context. This opens the possibility of evaluating the relative weight of different types of factors empirically, and probing the interaction among them. We might begin this exercise with a basic scheme of labor market governance that includes three broad categories: (1) laws and regulations, (2) standards and practices, and (3) norms and beliefs. ⁵⁸ This typology covers a spectrum from formal to informal and from public to private. It could then be applied to the specific features of labor markets in a particular country, time, or sector.

This brings us back to the fundamental question of what really drives wages. Is it the broader environment or the attributes of the worker? Is it the (macro) power relations that affect the bargaining leverage of the two parties? Or the more specific (micro) factors? The standard economic models assume away power to focus on more immediate market dynamics, and that can make sense as a research strategy to isolate certain variables. Yet shouldn’t we really be conceiving of power as endogenous to labor markets? To address these questions adequately, we must first review some of the key assumptions underlying economic models.

The “Market Wage” and the Distribution of Rents

We can grasp the theoretical implications of bringing power to the center of the study of wages by reviewing two key concepts from economics: the market wage and rents. Like other subfields in economics, labor economics begins with the assumption of perfect markets and then gradually introduces imperfections (market “failures” or anomalies). Economists present the perfect market as a simplifying assumption, not as a claim to an accurate portrayal of reality. But it does become the reference point, sometimes in the form of the “market wage.” But what is a market wage? In the realm of theory, it refers to the wages that reflect marginal productivity that would be paid in a perfectly competitive marketplace. In empirical research, economists typically compare actual wages in a given realm, such as a local or sectoral market, to average wages over a broader area, such as the national economy. For example, a local condition of monopsony represents one type of market imperfection. Employers with substantial market power can pay wages substantially lower than the market wage. For practical purposes, economists then compare those (local) wages to average (national) wages to assess the impact of that market power. ⁵⁹

We should keep in mind, however, that there is no such thing as a market wage absent the particular balance of power between employers and workers. The average wage within a geographic area represents the average wage in that area given that balance of power. There is no market without governance, and governance is never neutral. The government must provide a legal system to enforce contracts, but it cannot do so in a purely neutral manner. For example, the government can allow noncompete clauses or it can forbid them. The former favors employers and the latter favors workers. But the government cannot opt out of that choice. It cannot just leave that for the employers and workers to figure out for themselves. Likewise, it can allow mandatory union dues or it can forbid them. The former favors workers and the latter favors employers. But it cannot escape this choice. Government policy structures the balance of power, and that balance powerfully shapes the “market wage.” So the power balance is not exogenous to the market: It defines the market.

A sports analogy may serve to illustrate the point. ⁶⁰ Let's say we were trying to predict how many runs the Red Sox would produce in a game against the Yankees. We would naturally scrutinize the capabilities of the Red Sox hitters and the Yankees pitchers. But our prediction would build on a specific understanding of the governance of the game. This governance would include the formal rules: Hitters are out with three strikes, and they walk to first with four balls. It would include enforcement: Are the umpires generous or stingy with the strike zone? It would include practices: Do coaches take out their best players when they are way ahead in a game? And it would include norms: Do the pitchers believe that putting foreign substances on balls violates fair play? If the rules, the enforcement, the practices, or the norms changed, that would alter the competitive balance between offense (hitters) and defense (pitchers) and affect our prediction for the number of runs. But here is the point: We could not simply say let's imagine baseball neutrality. We could not just forget about the rules, and play ball. The game must have rules, both formal and informal. In fact, the game is defined by those rules. And those rules set the balance of power between pitchers and hitters, just like labor rules determine the balance of power between employers and workers.

The rules that shape the balance of power in baseball are relatively simple and explicit, but this recognition actually helps to reveal the limits of standard economic models and to show the path to deeper analysis. In labor markets, those rules are more complex and obscure, so we can lose track of them and fall into the trap of comparing policy “interventions” to a hypothetical free market. But the baseball case helps to reveal why this makes no sense. We would not view a change in the rules to allow four strikes to an out as an intervention that distorts “free baseball.” It would simply be a rule change that would benefit batters over pitchers. Likewise, there is a structure of rules that defines the balance of power between employers and workers in the labor market: It just is not as simple as balls and strikes. Yet that is precisely the task of the political economist: to unroot that deeper power structure, including laws and regulations, standards and practices, norms and beliefs. Otherwise, we are missing the essence of the market dynamics. Needless to say, this analogy does not fit perfectly, but the differences are illuminating in their own right. For one thing, American baseball has only one major league with one set of rules, whereas labor market governance varies considerably across space and time—and that is in many ways the heart of the action.

Economics uses the concept of “rent” to depict any payment to the owner of a factor of production in excess of the costs needed to bring that factor into production. This is both useful for understanding real-world labor markets and potentially misleading. It allows us to conceptualize labor markets as contests between capital and labor for rents—yet the concept itself builds on perfect markets as a reference point. The rents are the product of deviations from those perfect markets. But how do we make sense of this conceptually or empirically? If there are no perfect markets, then there is no way to set an absolute scale for what constitutes a rent. There is only a spectrum of power balances from employer dominance to worker dominance. It is impossible to identify a free-market or neutral or rent-free point along this spectrum. This is true of other core relationships in the economy, such as incumbents versus challengers or shareholders versus stakeholders, as illustrated in Figure 2. To return to the baseball analogy, we would not describe the rule change from three to four strikes for an out as generating a “rent” for the batter relative to the natural state of baseball. That concept of rent would have to be relative, not absolute.

This still leaves us with room for a productive “theory of relativity” for rents. We could compare one market situation to another one (such as points A and B from Figure 2), or to the status quo, rather than to a perfect market. We could compare two different power balances, rather than an imbalance relative to an imaginary equilibrium devoid of power relations. In fact, economists’ measures of rents, including market concentration, retail markups, and Tobin's Q, would turn out to be quite useful in this world of rent relativity. ⁶¹ They cannot measure rents relative to a state of nature, since that state does not exist. But they can help us to compare two labor market governance regimes, and to evaluate that difference. Likewise, the relativity of rents does not undermine the concept of rent-seeking more broadly, which then simply signifies the effort to shift the rules in one's own favor, such as toward employers or incumbents or shareholders.

Some would argue that the employment relationship inherently advantages employers over workers, so the government should protect workers from exploitation. Yet I would argue that it is more analytically useful to view this relationship as governed by a spectrum of power balances rather than as an absolute imbalance that can only be shifted in one direction. After all, it is possible to move the needle toward laws, practices, and norms more favorable to workers from any given point on the spectrum. And it is possible for that needle to move very far in that direction, theoretically at least, to the point where workers would be able to stop production and employers would have little recourse. Moreover, viewing labor market governance as worker protection can camouflage the extensive employer protection, or pro-employer regulation and practices, that govern most real-world labor markets. In other words, if labor exploitation is the norm, that is not because that is the state of nature but because employers have deployed business and political strategies to make it that way. And if researchers assume that imbalance rather than scrutinize it, they are missing the essential dynamics of real-world labor markets.

Policy Implications

A power-centric model of labor relations bears substantial implications for policy debates. It recognizes that conventional models have a status quo bias in that they compare policy “interventions” to the perfect market rather than to the real world, thoroughly sullied by imbalances of power. These models assume that there is a neutral labor market devoid of power relations, and policies and institutions represent interference in that market that can only be justified by reference to specific market failures. ⁶² Moreover, they assume a trade-off between the efficiency of the perfect market and other goals such as equity that might justify deviations from it. The analysis presented here suggests that while labor unions may create a deadweight loss relative to perfect competition in theory, there is no such thing as perfect competition in practice. In fact, if the balance of power favors employers, then greater worker power would be more likely to improve efficiency than to undermine it. For example, Azar, Berry, and Marinescu estimate that workers produce about 17 percent more value than that reflected in their wage level. ⁶³ If this is accurate, that would mean that increasing labor power would shift labor markets toward a more efficient allocation. So government officials should feel freer to introduce policies to shift that balance to achieve goals such as reducing poverty, moderating inequality, increasing employment, or boosting consumption.

This does not mean that it is impossible to say anything meaningful about the efficiency or the competitiveness of markets but only that we cannot judge these items by their distance from a perfect market. ⁶⁴ The perspective adopted here simply stresses that comparing one point on the line to a perfect market does not tell us whether that point is more or less efficient than other points on that line.

A power-centered model also suggests that we should rethink what competition means in labor markets. Competition does not require limitless jobs or limitless workers, but merely some movement (contestability) based on price. And that necessitates some balance in power. The dynamism of the market comes from the process of competition, not from the perfect transmission of market signals. Standard economic models overestimate the “perfection” of markets, but also their fragility. Hence they exaggerate the danger of market distortion from policy measures. The key is not to strive for perfect signals, because all real-world markets fall short of that ideal, but to preserve basic market incentives for firms to produce better products and workers to do their jobs. ⁶⁵

The standard model can also bear a normative overtone that wages reflect marginal productivity, which simply reflects skill and effort. By assuming away power relations and institutions, these models fail to represent how much wages reflect things other than skill and effort. ⁶⁶ There is nothing inherently fair or just about the status quo, so we should not shy away from trying to create a more just foundation for labor markets.

This perspective also implies that we should place more emphasis on the market governance policy agenda (or “predistribution”) to combat wage inequality. ⁶⁷ I refer to this market governance as “marketcraft” because it constitutes a core government function for economic policy roughly analogous to statecraft for foreign policy. ⁶⁸ Marketcraft strategies focus on the market governance that generates inequalities in the first place, whereas redistribution strategies shift the allocation of income and/or wealth after the fact. One strives to create a more equitable market; the other accepts the market as it is and tries to compensate for its inequities. We could visualize the difference between predistribution policies and redistribution policies in terms of Figure 1. The predistribution (preproduction) policies would be on the left side of the figure, such as labor regulation or education policy, whereas redistribution policies would come in on the far right beyond the scope of the figure (postproduction) after the initial allocation of wages. ⁶⁹ In fact, one could argue that adjusting the market governance model to shift the balance of power in favor of workers would be more market-conforming than imposing price controls, such as minimum wages, because employers and workers would still be responding to incentives at the margins whereas minimum wages would preclude some transactions. ⁷⁰

This suggests that the policy priority should be to alter regulations, practices, and norms that favor employers over workers. The marketcraft policy agenda does not obviate the need for other policies, including macroeconomic management to boost employment, welfare services to support workers, tax policy to redistribute after-tax income, or public investment in education and infrastructure. But it does differ from a more standard welfare state mix in that it prioritizes market governance over redistribution and it advocates empowering workers within markets over protecting them from markets. ⁷¹

The logic of the argument presented thus far suggests that it is impossible to identify a single “fair” balance of power or rewards between firms and their workers, so judgments hinge on values and policy priorities. If we can identify specific laws and regulations that favor employers, then that begets a set of policy targets and offers some guidance in setting priorities among them. This might include prohibiting noncompete clauses, restricting mandatory arbitration, permitting multiemployer bargaining, revising labor law to facilitate union formation, and reclassifying some gig workers as employees. ⁷² And while we cannot pinpoint an ideal balance of power, we have strong evidence that the current balance tilts too far in the direction of employers based on a variety of measures of public welfare. For example, we know that some countries have labor regimes that have provided more gains to workers and achieved greater equality than other countries without compromising growth. ⁷³ We know that labor power has waned over the past forty years in the United States; workers have not gained much of the productivity gains during that time; and this inequality has contributed to secular stagnation. ⁷⁴

Toward a New Political Economy of Wages

So what might an interdisciplinary political economy of wages look like? As suggested throughout this essay, economists could integrate power in the broad sense—including political power, social networks, and social status—into the heart of their models; and noneconomists could be more explicit about what their research findings tell us about market outcomes, especially wages. ⁷⁵

What does this mean for theoretical models and empirical research in concrete terms? On the theory side, the analysis presented here begets one clear conclusion: Economic models of wage formation should begin from the assumption of employer power, not the perfect market. This provides a more accurate depiction of actual labor markets, a more useful model, and a more fruitful basis for an interdisciplinary political economy of wages. ⁷⁶ The perfect market model may be more parsimonious, but it is fundamentally misleading for both analysis and policy because it excludes some of the most important factors affecting wage levels.

Manning bases his case for a monopsony model on two things: There are important frictions in the labor market, and employers set wages. ⁷⁷ The frictions give employers some market power over workers, and employers can exercise this power because they set wages. He also notes that workers typically need their jobs more than employers need any individual worker. I would add that employers tend to enjoy a structural advantage in that they have more political power, more social status, and stronger personal networks.

On the empirical side, we could then test these models, looking at how political and social factors influence the balance of power in labor markets and how that balance in turn shapes wages. We might begin with a meta-analysis of existing studies on how specific elements of power affect wages. For example, Gould and Kimball find that US state right-to-work laws (which prohibit requirements for employees to join a union or to pay union dues as a condition of employment) reduce wages by 3.1 percent, and the Bureau of Labor Statistics estimates that unions increase wages by 20 percent. ⁷⁸ This meta-analysis would take an initial stab at estimating the relative weight of different elements of power (such as political power, social status, or union organization) on wages, and thereby provide some guidelines for further research.

The monopsony model offers a particularly promising framework for economists and other social scientists to engage each other's work directly. Political scientists and sociologists might begin with their existing studies of power asymmetries in labor politics and labor relations, and simply extend the causal analysis to wage levels, as in Figure 1. For example, scholars of comparative political economy could apply their analysis of cross-national variations in levels of corporatism (patterns of the intermediation of business and labor interests) or varieties of capitalism (including labor relations systems, financial systems, and interfirm networks) to develop and test hypotheses regarding the impact of these institutions on wage levels. ⁷⁹ Or sociologists might build on their scholarship on variations in corporate governance or business networks across industrial sectors or firms to quantify the impact on wages. ⁸⁰ And scholars could explore more dynamic models, as proposed above, in which firms not only adopt business strategies to boost profits, but they also deploy political strategies to facilitate their favored business strategies. ⁸¹

A few political economists and economic sociologists have situated wage levels as outcomes in their studies. In addition to the sociologists introduced above (Rosenfeld and Wilmers), political scientists have studied how corporatist institutions (such as centralized wage bargaining) restrain wages or how the varieties of capitalism (including labor market institutions) affect wage inequality in industrialized countries. ⁸² Yet most scholars of labor relations in other disciplines leave it to economists to explain variations in wage levels, and they do not engage with theoretical models or empirical research on monopsony.

Political scientists and sociologists could incorporate monopsony models more directly into their research via three steps. First, they would generate baseline hypotheses about how specific types of employer power, such as political influence or social status, affect wages and employment. Second, they would develop hypotheses about how political or social factors might affect these levels of employer power. And third, they would supplement that analysis with hypotheses about how and why those political or social factors themselves might vary. And then they could test these various hypotheses empirically. To put this in terms of Figure 1, the first set of hypotheses would focus on the relationship from the power balance to wages. The second set would turn to how other factors—such as government regulation and business practices—affect that balance of power. And the third would shift backward one more step in the causal chain, exploring how politics shapes government policy or how social context conditions business practices.

For the first step, scholars might begin with the presumption that different types of employer power would have similar effects on wages and employment. Economic models suggest, as noted above, that monopsony produces lower wages and lower employment, resulting in the exploitation of workers and a deadweight loss. ⁸³ The logic here would be that a balance of power in favor of employers relative to workers would have the same effect on wages and employment regardless of its source. Yet it is also possible that the type of power could affect the impact of that power. For example, employers with greater political power might be inclined to reduce wages but to maintain employment, rather than to reduce both, because they might have more sensitivity to political backlash against layoffs.

The next step would delve further into the diverse forms of power, which should play to the comparative advantage of scholars outside of economics. For example, scholars might compare and contrast how political influence or social ties affect employer power. Nathan offers an example of this type of analysis, as outlined above, by showing how the features of particular industrial sectors affect the power of multinational buyers relative to the workers at their suppliers. ⁸⁴

Perspectives from outside economics might also enrich analysis of the worker side of the balance of power. As we move from a narrow definition of monopsony as a dominant employer in a particular geographical location to a broader one as employer power more generally, the concept becomes more relational. That is, it reflects the resources of the employer relative to the countervailing power of the workers.

Finally, scholars might investigate how the sources of employer power themselves might vary across time, space, industry, and worker identity. For example, monopsony power in the narrow sense varies primarily by geographic region, whereas political power might vary by the party in power; network power might vary by industrial sector; and status power might vary by worker identity. This research might also generate more insights into the durability of monopsony power and the dynamics of change.

As stressed in earlier sections of this article, some economists have indeed incorporated broader elements of power into their models of wage formation. This article simply suggests that they could benefit from further forays into related scholarship in other fields, especially regarding the multiplicity of forms of power. Recent work by economists suggests some promising ways to accomplish this. Bivens, Mishel, and Schmitt provide a review of the literature on how market power affects wages and stress that economists should go beyond conceptions of market power as market concentration. ⁸⁵ To do so, they favor models of “dynamic monopsony” that posit that labor market frictions can make markets behave as if there were a concentration in the labor market. Those frictions—such as incomplete information, transportation costs, and family care responsibilities—simply reflect the normal functioning of real-world markets as opposed to the perfectly competitive models of textbooks. And they argue that the US government has boosted the relative power of employers over time through policies such as tight monetary policy, a low federal minimum wage, a lack of support for union power, and the de facto acceptance of employer practices designed to limit workers’ bargaining position.

Wilson and Darity suggest that addressing how race affects wages requires going beyond standard individual-centered models to consider structural and institutional factors. ⁸⁶ Stratification economics adopts an interdisciplinary approach that combines economics, sociology, and social psychology and centers on group formation, group identity, and group action. ⁸⁷ Persistent racial inequality arises when a dominant group seeks to maintain hierarchy. Under this framework, it can pay (literally) for a group to invest in or associate with a group identity.

This article has challenged the conventional wisdom on labor markets, advancing the following propositions: (1) There is no such thing as a labor market that is not socially and politically constructed. (2) All real-world labor markets reflect specific balances of power. (3) The balances of power reflect not only the abundance or scarcity of market (exit) opportunities but a wide range of political and social factors. (4) Therefore, laws and regulations to shift those balances do not constitute “interventions” into free markets. (5) Such laws and regulations are not necessarily inefficient or undesirable, and they do not require a particular justification based on market failures. In sum, this article posits that power is endogenous to labor markets. There is a spectrum of power balances between employers and workers but no state of nature devoid of power relations. And the balance of power is a major determinant of wages. Hence, we should begin our analysis from the assumption of employer power rather than the perfect market. We should expand the model of wage formation to include a broader array of political and social factors, especially the many facets of power. And we should specify just how much those facets of power cost workers.