Abstract
We argue that market failure is a major and growing source of income inequality in the United States and in liberal market economies (LMEs) more generally. This market failure takes the form of occupational, educational, managerial, and capital rents that are generated by institutional barriers that restrict the free flow of capital or labor. We suggest that these four forms of rent can partly account for (a) the extreme income inequality at the very top of the LME income distribution as well as (b) the extreme income inequality that is also observed beneath the highest percentiles of the income distribution. The sharp increase in these forms of rent, when coupled with rent destruction at the bottom of the labor market, may well explain much of the takeoff in LME inequality in the past four decades. We conclude by outlining the empirical research agenda and policy prescription implied by a rent-based account.
The dramatic takeoff in income inequality in the United States and many other rich countries has generated one of the most active research literatures in the social sciences. We now have an extensive body of research on why the takeoff in inequality occurred (e.g., Goldin & Katz, 2008), who benefitted most from it (e.g., Atkinson, Piketty, & Saez, 2011; Salverda, Nolan, & Smeeding, 2009), and which countries experienced it in especially extreme form (e.g., Gottschalk & Smeeding, 2000; Salverda et al., 2009). The literature within economics on the causes of the takeoff tends to feature market-based stories about the growing payoff to skill (e.g., Acemoglu, 2003; Autor, Katz, & Kearney, 2008). Although “political economy” analyses of inequality have become increasingly popular, they mainly focus on the political underpinnings of aftermarket interventions (e.g., taxes, welfare payments) rather than on the political underpinnings of the market itself (e.g., Baramendi & Anderson, 2008; Bartels, 2010; Brady, 2009; Frank, 2011; Hacker & Pierson, 2010; Kenworthy, 2011; Manza, 2012). 1
The core problem with this existing literature, in our view, is that it either operates narrowly within a competitive market framework or treats the competitive market as an unanalyzed backdrop against which the politics of aftermarket redistribution occurs. There are, to be sure, some recent political economy accounts that seek to explain the rising wealth and income of the top 1% in term of their power to shape the rules of markets (e.g., Reich, 2012; Stiglitz, 2012). In this article, we offer a programmatic statement on rising income inequality that, like this newer literature, refocuses attention on the existing institutions in so-called competitive market economies. We argue, however, that market failure and the income inequalities it generates are as important to understanding income disparities between the middle and working classes as they are to understanding the disparity between the 1% and the 99%.
The conventional wisdom on the takeoff in inequality embodies both a diagnosis of its sources and a prescription for its reduction. Under the formulation favored by many institutional scholars, market competition is responsible for the high baseline level of inequality in liberal market economies (LMEs), and a more recent wave of deregulation and other competition-enhancing reforms pushed levels of inequality in LMEs even higher. The resulting mantra is that competition breeds inequality and increasing competition breeds ever more inequality. Insofar as the policy goal is to treat inequality, the twin prescriptions emerging from this diagnosis are either to redistribute postmarket income or to increase premarket skill through greater investment in education (see, e.g., Obama, 2009).
These competition-based accounts of inequality and its takeoff are incomplete at best. Even in ostensible LMEs, much inequality is generated by competition-suppressing institutions that deliver income to those at the top of the income distribution. These institutions are often veiled from view and, as a result, have not been targeted by the various “competition-enhancing” reforms occurring at the bottom of the income distribution. The implication is that the rapid growth in inequality in LMEs occurred not only because of competition-increasing change at the bottom (e.g., declining union power, globalization) but also because of competition-reducing processes at the top. If we are correct in this diagnosis, the prescription is clear: We can take on poverty and inequality by repairing noncompetitive labor and capital markets and thereby reducing the inequality that such failed markets generate.
The concept of rent is key to our argument. We adopt the usual definition of rent as returns on an asset (e.g., labor) in excess of what is necessary to keep that asset in production in a fully competitive market (e.g., Congleton, Hillman, & Konrad, 2008, 2010; Tollison, 1982; Tullock, 1967). By this definition, rents exist wherever (a) demand for an asset exceeds supply and (b) the supply of the asset is fixed, whether through “natural” means or through social or political barriers that artificially restrict supply. The “fixed supply” condition implies that labor cannot respond to the price increases that arise when demand exceeds supply. In contemporary labor markets, rent takes on many well-studied forms, including the wage premiums associated with the minimum wage (e.g., Neumark & Wascher, 2010), the union wage and the associated “spillover” effects (e.g., Freeman & Medoff, 1984), between-industry wage differentials (e.g., Cha & Morgan, 2010; Katz & Summers, 1989), and the wage premiums that accrue to licensing and related types of occupational closure (e.g., Kleiner & Krueger, 2008; Weeden, 2002). 2
It is well known that some of the institutions through which workers have historically been able to capture rents have weakened in the past 35 years. The two main developments of interest are the decline in the real value of the minimum wage during the 1980s and the decline in the proportion of the labor force that is unionized. In a now-classic analysis, DiNardo, Fortin, and Lemieux (1996; Lee, 1999) concluded that the minimum wage “held up” the lower tail of the earnings distribution in the late 1970s but that, as the real value of the minimum wage declined, the lower tail quite quickly sank. More recently, there has been some debate about the extent to which the changing value of the minimum wage can indeed explain rising inequality, but even the skeptics suggest that it at least had a nontrivial spillover effect on workers who earned somewhat more than the minimum wage (Autor, Manning, & Smith, 2010). 3
The effects of declining unionization are less controversial. Most obviously, unions are understood to raise wages by providing members with a monopoly over certain jobs, in effect preventing employers from driving down wages by pitting union and nonunion workers against one another. At the same, unions raise the wages of nonunion workers because (a) nonunion employers wish to forestall unionization and (b) the union wage generates widely shared norms about proper pay that are then costly for employers of nonunion workers to ignore. When these effects are taken into account, Western and Rosenfeld (2011) have shown that approximately one third of the rise in inequality between 1973 and 2007 is attributable to the decline in unionization (also see Card, Lemieux, & Riddell, 2003).
The foregoing accounts of rent destruction are, we argue, not so much incorrect as incomplete. By focusing on the decline of institutions that generate rents for workers at the bottom of the wage distribution, the prevailing assumption in much of this literature is that rent destruction, not rent creation, is the main dynamic characterizing advanced industrial societies. In his seminal statement outlining a rent-based theory of social class, Sørensen (2000, p. 1555) predicts that “rents will disappear from structural locations in the labor market,” with the result being a transition toward a “structureless society” without the “nooks and crannies” in which competition is suppressed and rent can be collected. At the same time, Sørensen also locates rising income inequality in the 1980s and 1990s in the newfound incentive of managers to capture a greater share of composite rents, a theme later taken up by Morgan and his colleagues (Cha & Morgan, 2010; Morgan & Cha, 2007; Morgan & Tang, 2007). With these notable exceptions, the bulk of theoretical and empirical efforts to develop a rent-based account of rising income inequality have focused on rent destruction and the diffusion of a market-based logic that exacerbates inequality.
We argue, to the contrary, that rent destruction and rent creation have been deeply asymmetric: Just as rent is gradually being destroyed for workers at the bottom of the income distribution, it is also gradually being created at the top of the distribution. This rent creation has not occurred solely through industrial restructuring and, in particular, the “financialization” that has characterized LMEs (see, e.g., Phillipon & Resehef, 2009; Tomaskovic-Devey & Lin, 2011; Yeldan, 2000). Rather, it has been generated through a much broader set of institutional shifts that have created new rents for those at the top of the income distribution, even as rents for those at the bottom have been destroyed or redistributed.
In this article, we address four such forms of top-end rent: occupation rent, education rent, capitalist rent, and managerial rent. Although most of these forms have been identified in the existing literature on rent, they are largely absent from the literatures on comparative inequality and, especially, on trends in inequality. In the sections that follow, we discuss each of our four types of rent and how it affected inequality trends in LMEs, and we lay out the types of analyses that might be undertaken to test a rent-based account. We begin, however, by rehearsing the conventional wisdom on how inequality is generated in LMEs and why it is increasing. This sets the stage for our more skeptical account. As noted above, we argue that the high and rising inequality of LMEs is not an inevitable consequence of competitive market institutions, but instead is attributable to the particular types of rent that flourish within the United States and, we suspect, within most LMEs.
The Mantra on Inequality in LMEs
The conventional claim within the varieties of capitalism (VoC) literature is that income inequality in LMEs is intrinsically higher than inequality in coordinated market economies (CMEs). This claim appears in Hall and Soskice’s (2001, p. 21) influential VoC statement and is repeated throughout the voluminous literature that followed (e.g., Baramendi & Anderson, 2008; Pontusson, 2005; Soskice, 2010).
The empirical relationship behind this claim is indisputable, but the key question is why inequality is higher in LMEs than in CMEs. With few exceptions, the VoC literature simply takes it for granted that LMEs, resting as they do on market contracts to coordinate exchanges, generate high inequality. The explanatory effort thus presupposes high inequality in LMEs and focuses on the institutions in CMEs that limit the rise of inequality. Within the VoC literature, CMEs have all manner of institutions that limit the rise of inequality (e.g., collective bargaining agreements), whereas LMEs are reduced to ideal-typical competitive markets and thus play the role of institution-free foils (also see Howell, 2003). The implication is that high inequality within LMEs is a consequence of intrinsic inequality-generating features of competitive market economies that have largely gone unchecked. We turn this approach on its head, focusing not on the institutions that suppress inequality in CMEs but on the institutions that generate inequality in LMEs.
What is it about competitive market economies that putatively generate so much inequality? The microeconomic literature identifies six mechanisms of interest. As will be apparent, some of these mechanisms obtain in perfectly competitive economies, while others become operative in the context of minimal departures from that ideal type, in particular departures that allow for imperfect information. We rehearse these six mechanisms below and then identify those featured in the VoC literature as the primary source of high inequality in LMEs.
Conventional Sources of Inequality in LMEs
In all economies, some workers are blessed with abilities that make them more productive in the context of that economy, resulting in higher earnings. The man who is so lucky, for example, as to be seven feet tall, agile, strong, coordinated, and athletic is well suited to be a center in the National Basketball Association (NBA). The extremely high pay of NBA centers may be understood as the rent that accrues to these abilities, abilities that are in fixed and limited supply. 4 If height and other basketball-enhancing attributes were in ready supply or could be easily acquired through surgical or pharmaceutical means, then NBA centers would be exposed to more competition and their pay would be driven down by virtue of that competition. It follows that rents on ability arise when there is a fixed supply of inborn traits and a demand for those traits that exceeds supply.
Similarly, some workers have inborn tastes that make particular activities especially rewarding, such that they are willing to forego some amount of earnings in exchange for the utility that arises from indulging those tastes. It is often suggested, for example, that some people are natural caregivers and are willing to accept low earnings in exchange for the on-the-job rewards that, for them, arise when they tend children. If the distribution of inborn tastes generated too few people willing to accept low pay in these child-tending jobs, the shortage in the supply of child-tending labor would drive wages in child care occupations higher and, in a competitive economy, lead to an influx of child care workers. The implication is that a given distribution of tastes, unless it perfectly matches demand, may disrupt this clearing flow of labor and allow inequalities in pay to persist.
The preceding example pertains to taste for a particular line of work (i.e., child tending), but more general tastes can also have an inequality-generating effect. A taste for “deferred gratification,” for example, can increase downstream productivity and earnings by motivating investments in schooling and other human capital, while a taste for “hard work” leads to differences in the amount of hours spent earning money and the amount of effort that is expended each hour. 5 As long as these tastes are fixed, and workers are willing to trade off pay for them, then the inequality that they generate may be understood as rent-based. Clearly, it is a form of rent that many people find tolerable and even desirable, given that any restrictions on the expression of tastes is viewed as an unwarranted deprivation of liberty. But it is nonetheless a form of rent.
The inequality generated by inborn tastes should be distinguished from that generated by tastes that are acquired through socialization by the family, schools, or peers. It is often claimed, for example, that children in impoverished neighborhoods or families do not develop a taste for deferred gratification and are therefore reluctant to invest in education, even assuming opportunities for these investments are available (see, e.g., Breen & Goldthorpe, 2001). These inculcated tastes, like inborn tastes, can also generate rent by virtue of their scarcity relative to demand. The root source, however, is a particular configuration of social institutions, not genes, and hence the resulting rents should be understood as institutional in origin.
Compensating Differentials
In all economies, some jobs are so unpleasant, risky, or undesirable that the wage rate must be increased to motivate enough workers to take such jobs (e.g., Rosen, 1986). Although the simple bivariate relationship between desirability and wages is positive, the standard presumption is that, were all other sources of inequality netted out, the relationship would be negative. If one focuses on income inequality (as opposed to “utility inequality”), it follows that compensating wage differentials can explain some portion of the inequality observed in LMEs.
Human Capital Investments
Because human capital investments, such as schooling, bring about higher productivity, employers will pay more for workers who have made these investments. This higher payment may also be understood as compensation for training costs and, as such, will yield inequality only in cross-sectional earnings. In the absence of barriers to the flow of labor, lifetime earnings will be equalized across workers, given that they will invest in human capital up to the point where investments equal expected returns.
This highly stylized human capital story is complicated by the distribution of tastes and abilities, insofar as the decision to invest in human capital may itself require learned or inherited tastes or abilities that are in short supply. It may well be that only those young people with a well-developed taste for deferred gratification are willing to make a costly investment in medical training. If there are naturally occurring (i.e., genetic) or institutionally generated (e.g., class-specific socialization) shortfalls in this particular taste, the wage returns of doctors will exceed what would prevail if labor freely flowed to the high-return investment. The resulting wage returns in excess of the costs of training are properly understood as rent.
Efficiency Wages
In some markets, managers may pay their workers in excess of what the market demands, as doing so makes them more efficient or productive (e.g., Akerlof & Yellen, 1986; Shapiro & Stiglitz, 1984). When direct measurement of work effort is costly, overpayment in the form of efficiency wages becomes the cheapest solution to the problems of shirking or malfeasance. The simple logic here is that, because the likely pay cut associated with getting fired is high, workers will devote maximum effort to their jobs to avoid getting fired. Alternatively, firms may “overpay” to induce high-productivity workers to apply for an open position, thereby reducing worker selection and monitoring costs. The fundamental problem in both cases is that acquiring information about likely productivity or on-the-job performance (i.e., monitoring) is so costly that it is cheaper to overpay. It follows that inequality arises in LMEs because some firms pay their workers efficiency wages and others do not.
Statistical Discrimination
Because it is often too costly for firms to measure anticipated productivity at the individual level, they may resort to group-level measurements as a shortcut assessment of individual-level capacity. Members of some groups (e.g., Whites, males) may be paid especially high wages without such wages reflecting true individual-level differences in ability or productivity (Arrow, 1973; Phelps, 1972). This mechanism creates inequality that cannot be attributed to ability or tastes, although it is again related to the information shortfalls that generate efficiency wages.
The foregoing list of the sources of inequality in competitive economies is standard fare, but we have tried to be more rigorous than usual in distinguishing between sources of inequality that imply rent and those that do not. The distinctive feature of this list, for our purposes, is that it acknowledges only two types of rent: (a) the “endowment rent” that arises because ability or tastes are fixed in a distribution that does not match demand and (b) the “information rent” that arises because firms cannot directly measure future or current productivity, thereby generating efficiency wages and statistical discrimination.
The VoC literature, with its rational-functionalist orientation (see Streeck, 2010), accepts this account of the sources of LME inequality. It therefore implicitly recognizes some forms of rent within LMEs but ignores others. It is simply assumed that LMEs are so close to the perfect-competition ideal type that only the “inevitable” rents of modern economies, those attributable to the natural distribution of endowments or to high information costs, will be in evidence. It follows that, if only LMEs would adopt the inequality-suppressing institutions of CMEs (e.g., collective bargaining, strong unions), inequality in those countries would be greatly reduced.
This line of reasoning rests on a misdiagnosis of the sources of inequality in LMEs. There is good reason to suspect that LMEs are rife with rent, not just endowment and information rent, but rent that has deeply institutional sources. These institutional rents generate levels of inequality in LMEs that far exceed what would be observed if the mechanisms reviewed above were the only sources of inequality. As we discuss below, such rents tend to develop at the top end of the labor market, meaning that they generate the peculiarly top-heavy form of inequality found in LMEs (see Figure 1). The upshot is that the much-vaunted inequality-suppressing institutions of CMEs would have only limited utility in combating inequality in LMEs.
Top 10% income shares in coordinated market economies and liberal market economies.
The Mantra on Trend in Inequality
If the conventional wisdom about the LME-CME gap in income inequality fails to appreciate the generative role of institutional rents, so too does the conventional wisdom about trend in LME inequality. The long-preferred economic explanation for the takeoff in wage inequality, skill-biased technological change (SBTC), is by now quite familiar: Exogenous technological changes increased the demand for, and marginal productivity of, highly skilled workers, while it reduced the demand for low-skill workers (for a review, see Acemoglu, 2003; also see Autor et al., 2008; Card & DiNardo, 2002). As a result, the wages of skilled workers rose dramatically, while the wages of low-skilled workers fell.
In the SBTC account, rising inequality is a byproduct of smoothly functioning labor markets that are responding to an exogenous shock, namely technological change. The proximate source of growing inequality is rising returns to skill and hence human capital investments. In a competitive economy, workers should respond to price signals by investing in the training needed for high-premium jobs, thereby increasing the supply of labor for those jobs and driving down their wages. The temporary, “quasi-rents” generated by a short-term shortage of skilled labor will, through the usual equilibrating flows, be eliminated over time. We should see wage premiums in the long term only insofar as they rest on (a) abilities that cannot be acquired through education or training or (b) tastes for deferred gratification (and other prerequisites of investments) that are in short supply and are so entrenched that they do not respond to wage signals. The latter wage premiums can, as noted above, be understood as rent on endowments. 6 All manner of complexities could prevent markets from equilibrating in the ever-present short run. For example, Acemoglu (2002) argued that SBTC is endogenous, given that a ready supply of high-skilled workers creates the conditions for firms to implement ever higher-skill production regimes. This does not resolve the more fundamental problem of explaining persistent market disequilibria. After all, even if the “goalposts” keep moving, the forward-looking worker will still overinvest in skills for the present economy in anticipation of ever-increasing skill demands in the future. The answer to this puzzle rests in appreciating the role of institutional barriers to the free flow of labor. We suggest, in other words, that high returns to skill persist because rent-generating institutions prevent labor from responding to price signals and create permanent market disequilibria.
Although SBTC remains the dominant explanation of rising income inequality, it is certainly not the only explanation on offer. As we noted above, the institutionalist account has it that LMEs once had at least some CME-like institutions protecting low-skill workers (e.g., unions, the minimum wage), but as these institutions gradually dissipated LMEs became purer in form and inequality took off. There is still debate about the relative contribution of these shifts to rising wage inequality (e.g., Autor et al., 2008; Autor et al., 2010). The critical feature of these arguments, for our purposes, is that they focus exclusively on the demise of the institutions that allow workers at the bottom of the wage distribution to capture rent. These arguments are thus well suited for the purpose of explaining the decline in income at the bottom of the LME income distribution.
Can these institutionalist accounts also explain the rise in income at the top of LME distributions? If one sought to do so, it would presumably require a story about how the rents stripped away at the bottom are then transferred to the top. Although the rent formerly accruing to workers could be destroyed altogether, as is the case when barriers to competition are eliminated (e.g., via globalization or antitrust enforcement), it is also possible that such rent could be redistributed to the top. In research inspired by Sørensen’s efforts, Morgan and his colleagues show that, in the United States, low-skilled workers in formerly rent-rich industries lost wages relative to their compatriots in rent-poor industries (e.g., Morgan & Tang, 2007), while managers and professionals amassed greater wealth as stock prices grew (Morgan & Cha, 2007).
In our view, rent redistribution of this sort, while important and intriguing, does not fully capture the range of institutionally generated rents in contemporary LMEs. The latter rents have also emerged from new constraints on the supply of skilled labor (e.g., new credential requirements, more restrictive licensing requirements) and from barriers that prevent the rising demand for skilled labor in already-closed positions from being met by growing supply. As suggested below these rents have disproportionately benefited workers at the top of the income distribution, not only the much-vilified 1% but workers further down the income distribution as well (see Figure 2).
Top 10% income share in the United States, 1917-2010.
Top-End Institutional Rents in LMEs
We next introduce the four main forms of institutional rent and how they are implicated in the spectacular run-up of LME inequality. Although we focus on the U.S. case, the growth in top-end institutional rent is likely also implicated in the takeoff in inequality in other LMEs.
Occupational Closure
We begin with the twin claim that occupational closure is an important rent-generating institution and that such closure is also an important source of rising income inequality. By occupational closure, we are referring to the practice of establishing barriers that protect occupational incumbents against competition from others who might want (a) to enter the occupation or (b) to provide the same or similar products or services from a “close” or competing occupation (see Weeden, 2002). These barriers raise the wages of incumbents by increasing an occupation’s control over the supply of labor and by reducing competition from other potential providers of the product or service (Kleiner, 2006; Kleiner & Krueger, 2008; Weeden, 2002).
Have the rents stemming from occupational closure increased over time? Although this remains an open empirical question, the available evidence shows a rapid increase in the proportion of workers in licensed occupations. In the United States, this proportion now exceeds the proportion of unionized workers (Kleiner, 2006), an increase partly due to the expansion of licensed occupations and partly due to the diffusion of licenses to new occupations. The result in either case is that ever more rent is being collected at the top of the income distribution (where most licensure is found). In and of itself, this development does not challenge the SBTC story that demand for skill is growing, but it does suggest that the quasi-rents in skill-demanding occupations will persist over the long run only when occupational closure prevents labor from flowing in. That is, rising demand for skill is alone insufficient to account for the takeoff in income inequality, as workers in a truly competitive market will naturally opt to acquire the requisite skill and thereby drive the premium back down. If inequality is to be locked in, the beneficiaries of the rising demand must have tools at their disposal (e.g., licensure) that allow them to control competition and prevent other workers from cashing in.
Education Rent
The mechanisms behind education rent are similar, but they focus less on barriers to practicing certain skills and more on the acquisition of the skills themselves (also see Bol & Weeden, 2012). It is by now widely appreciated that returns to education have increased in the past four decades, initially driven by rising returns to a college degree and, more recently, by rising returns to postbaccalaureate degrees. 7 In accounting for these rising returns, we suspect that workers are not opting against education because they are lazy, irrational, or lack the required proinvestment tastes, but rather because various constraints prevent them from acting on the market signal. The supply of potential college students is kept artificially low because children born into poor families are trained in poor schools that do not prepare them for college, while the number of available university degrees (i.e., the “demand” for students) is kept artificially low because administrative or institutional rules, rather than price signals, determine that number.
As a result of these institutionalized bottlenecks, the supply of college-educated labor has not increased as rapidly as one would anticipate in an era of rising returns (also see Goldin & Katz, 2008). Indeed, only 30% of each birth cohort now earns a college degree, which is not much higher than in the 1970s (Hout, 2009, 2012). These bottlenecks in the supply of college-educated labor imply that those fortunate enough to secure a college degree from a traditional nonprofit institution are artificially protected from competition and thereby reap rent. 8 This is an underappreciated form of market failure and likely a major source of rising income inequality.
Capitalist Rent
In the case of capitalist rent, our focus shifts from rising inequality in wages and salaries to the increasing share of national income that goes to capital rather than labor. By some estimates, this share rose over the past 40 years from 34% to 40% of national income, a sizable increase if not as dramatic as the increase in income inequality (Kristal, 2010, 2012). The conventional explanation for this shift is that “capital-biased technological change” increased the productivity of capital relative to labor (see, e.g., Kristal, 2010) and therefore the share of income going to capital. This account, like the SBTC story, ignores sociopolitical factors in favor of narrowly construed technological ones.
In the rent-based account, the emphasis is instead on the newfound leverage of capitalists to appropriate rent that workers were once able to secure. The growing power of capitalists arises because (a) unions have weakened and no longer protect workers against competition from other domestic workers and (b) international barriers to the flow of capital have likewise weakened and no longer protect domestic workers against competition with foreign workers (e.g., Morgan & Tang, 2007; also Katz & Summers, 1989). The bargaining power of capital increases insofar as workers must now negotiate without the protection once afforded by unions or by “territorial closure” in the form of an implicit national monopoly on the provision of labor (Brubaker, 1993). In addition, capitalists in some industries have achieved higher market concentration of late and firm sizes have grown steadily (White, 2002), developments that not only increase firm rents but also, in conjunction with weak unions and strong threats of “offshoring,” provide additional leverage for capitalists to capture a greater share of the firm rents. This process is especially pronounced in the financial sector and in the financial service arms of nonfinancial firms (Tomaskovic-Devey & Lin, 2011), but it is not limited to those sectors (Kristal, 2010).
Managerial Rent
The literature on managerial rent is relatively well developed. In the managerial sector, rents are understood as stemming from three main sources: pay-setting institutions that provide managers with performance-based bonuses, pay-setting institutions that allow the interests of board members to intrude, and new forms of managerial closure. The first form of managerial rent emerges as a solution to a classic principal–agent problem: To prevent managers from making business decisions that advance their own interests over those of the firm (e.g., Stiglitz & Edlin, 1995), owners have tied managerial pay to firm performance (e.g., through stock options) and, as a result, have raised managerial pay and increased inequality (Morgan & Tang, 2007; Sørensen, 2000). The second source of managerial rent refers, by contrast, to pay-setting institutions that divorce managerial pay from firm performance. The classic example here is the practice of allowing boards of directors, who have a private interest in acquiescing to the CEO, to effectively set the CEO’s pay (for a review, see Bebchuk & Wiesbach, 2009). Likewise, there is a growing literature on benchmarking norms that set CEO pay at the “going rate,” an institutional practice that provides a market-pay veneer because it is presumed that the going rate is not artificially raised by rent (see DiPrete, Eirich, & Pittinsky, 2010). Much debate persists over whether the high pay secured through these pay-setting practices reflects marginal productivity or rents (e.g., Bebchuk & Wiesbach, 2009; Gabaix & Landier, 2008; Murphy & Zabojnik, 2007).
In our view, these debates over the principal–agent problem and CEO compensation have distracted us from appreciating the growth in rents enjoyed by the vast cadre of middle managers, not just top executives. These rents stem from a special form of occupational closure that creates a disjuncture between the supply of managerial labor and the demand for it and, in turn, generates a broad-based increase in managerial pay. The demand story behind closure-based rents in the managerial sector is well known: (a) globalization introduces a new need to supervise and oversee far-flung operations, (b) product and service markets have been further expanded through new technologies, and (c) the intrafirm division of labor has become increasingly complicated and has thereby required new layers and types of control and supervision (e.g., Frank & Cook, 1995).
This increase in demand yields rent because the supply of labor is prevented from straightforwardly meeting the increase. Just as demand for managerial capacity was growing, managers embarked on a classic professionalization project, primarily through managerial credentialism. In the context of this project, formal education for managers went from “ornamental to essential” (Mayo, Nohria, & Singleton, 2006, p. 119), with the MBA in particular emerging as an important screening device and source of social capital. To be sure, business schools have also expanded rapidly, and nonacademic purveyors of business training have emerged as alternatives (Pfeffer & Fong, 2002). Even so, the number of students seeking entry into business schools still far exceeds the number of MBAs produced, as indicated, for example, by Graduate Management Admissions Test registrants in the United States. This newfound credentialism, while likely a less effective restriction on supply than occupational licensure, is nonetheless a source of rent that may be as important as principal–agent rent and CEO overcompensation.
Conclusion
We have argued that LMEs are rent-laced economies and that asymmetries in institutional rents account for the high levels, and rising levels, of inequality in LMEs. This formulation belies the oft-repeated mantra that competition begets inequality and that growing competition begets growing inequality. We have suggested, to the contrary, that the high levels of inequality in LMEs are in part attributable to institutional rents, and that rising income inequality in LMEs is caused by deeply asymmetric institutional changes that have destroyed rent at the bottom of the income distribution and allowed ever-increasing rents to be extracted at the top of the income distribution. Although this account of rising income inequality is still little more than a hypothesis, we think it is at least worth subjecting to as much scrutiny as the SBTC hypothesis.
What might this empirical agenda entail? Consider, first, our claim that occupational closure, and not skill upgrading alone, underlies persistently high and growing pay for skilled occupations. At minimum, the closure argument predicts that the takeoff in inequality in LMEs occurred between occupations, not just within them (see Mouw & Kalleberg, 2010; Weeden, Kim, Di Carlo, & Grusky, 2007; Williams, 2013; cf. Kim & Sakamoto, 2008). A more direct empirical test would show that wage growth was most pronounced among (a) occupations with well-developed closure at the start of the takeoff and (b) occupations that ramped up their closure before or near the beginning of the takeoff. The closure effect should be especially apparent among occupations that experienced a pronounced increase in demand (often via computerization and related technological change). Conversely, occupations (e.g., clerical workers) that upgraded but that lack closure should not show a persisting upward wage trajectory, precisely because they lack the capacity to close off against the burgeoning supply of laborers who can secure the requisite training.
In assessing education rents, we might begin by asking why the growing payoff for schooling has not been met by an equilibrating increase in schooling (also see Hout, 2012), and, in particular, whether the main bottlenecks are the supply and demand forces we have identified. The first question to take on is whether trends in the returns to schooling continue to reveal a takeoff even after trends in occupation rent are parsed out. Because highly educated workers tend to be found in rent-collecting occupations (e.g., managers, doctors, lawyers), the standard Mincer-style regressions conflate changing returns to education with changing occupational rent, with the resulting possibility that the seeming increase in returns in education may be nothing more than rising occupation rent in disguise. If, however, the substantial upward trend in returns to schooling persists with detailed occupation controls in place, the next question to ask is how this trend is generated and, in particular, whether it can be attributed to (a) a declining supply of youth who are well prepared for college (perhaps because of a growing preparatory disadvantage among poor children) and (b) a declining number of slots among high-payoff colleges (attributable either to declining state funding for public schools or a slot-rationing strategy among elite private institutions). The market failure hypothesis receives support insofar as the supply of college-educated labor does not well respond to changes in the payoff to college (under appropriate assumptions about the lag structure).
The capital rent hypothesis can be assessed by examining, at the industry level, the effects of various types of closure or market failure (e.g., unionization, concentration) on labor’s share of income within industries (see, e.g., Kristal, 2012). Because the trend in labor’s share varies considerably by industry, there is an opportunity to assess whether that variability reflects differences in the rate of computerization or, as we suspect, differences in trends in foreign competition and other sources of rent. Finally, a managerial rent hypothesis requires us, at minimum, to show that managerial occupations have indeed experienced a relatively steep takeoff in wages, even with controls in place for rising education returns. We also need to show that this changing payoff is associated with the emergence of institutions (e.g., the MBA) that worked to constrict the supply of high-level managers. The underlying demand for managers is unobservable, but the rent hypothesis becomes more plausible insofar as the constriction of supply is linked both to the rise of closure institutions and the takeoff in managerial pay.
If our market failure account is indeed on the mark, what are the implications for efforts to mitigate poverty and inequality? The United States and other LMEs have long fashioned their inequality-reducing interventions under the assumption that inequality is an unfortunate by-product of highly efficient competitive markets. This model leads to a focus on premarket skill enhancement (e.g., education) and aftermarket redistribution (e.g., through taxation and redistribution) rather than market repair. We have argued, by contrast, that LMEs would end up with less inequality, not more, if various forms of closure and supply bottlenecks were purged from our labor market institutions. Although redistributive tax policy is justifiable on a host of grounds, it is just as important to fashion inequality-remediating policy that is guided by the simple principle of repairing rent-ridden markets. Because public opinion in LMEs is, on average, quite sympathetic to the market principle, the prospects for reform that focuses on eliminating corruption, closure, and bottlenecks may be greater than the prospects for significant and lasting aftermarket redistribution.
Footnotes
Acknowledgements
The authors wish to thank Youngjoo Cha and Jennifer Todd for their skilled research assistance and Thijs Bol and Stephen L. Morgan for useful comments on a prior draft.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.