Why Are Economists Giving Piketty the Cold Shoulder?
[We live according to the tenets of a mythology that props up power and greed, that masquerades as a 'science' for driving rational policy. *RON*]
Marshall Steinbaum, Boston Review, 12 May 2017
|Janitor in Goldman Sachs building. Photo: Andrew Russeth|
The book’s central tenet—that wealth grows faster than economic output, thus concentrating capital (and the income it produces) in ever-fewer hands—confirmed the public’s sneaking suspicion that the workings of the capitalist economy are malfunctioning. Indeed even Goldman Sachs, perhaps the highest-profile symbol of global financial capitalism, put in a note to investors in early 2016, “[I]f we are wrong and high margins manage to endure for the next few years (particularly when global demand growth is below trend), there are broader questions to be asked about the efficacy of capitalism.”
Piketty confirmed the public’s suspicion that the workings of the capitalist economy are malfunctioning.
This is arguably the most Pikettian statement it is possible to make from perhaps the least likely source. But despite Piketty’s resonance with public experience and apparent applicability to the economic environment of global finance, his book was mostly greeted with hostility by the academic economics profession. There was a sense among academic economists that the book was a hostile action from within, and aside from Nobel Prize–winners Robert Solow and Paul Krugman, who both published reasonably favorable reviews in the highbrow popular press, the reaction was, in general, quite harsh.
In the Journal of Economic Perspectives, Daron Acemoglu and James Robinson wrote, “If the history of grand pronouncements of the general laws of capitalism repeats itself—perhaps first as tragedy and then as farce as Marx colorfully put it—then we may expect the same sort of frustration with Piketty’s sweeping predictions as they fail to come true, in the same way that those of Ricardo and Marx similarly failed in the past.” In the Journal of Political Economy, after praising Piketty’s lifelong research agenda assembling inequality statistics for income and wealth (as do all the reviewers named here), Lawrence Blume and Steven Durlauf wrote, “Capital is, nonetheless, unpersuasive when it turns from description to analysis. . . . Both of us are very liberal (in the contemporary as opposed to classical sense), and we regard ourselves as egalitarians. We are therefore disturbed that Piketty has undermined the egalitarian case with weak empirical, analytical, and ethical arguments.”
In a review in Democracy, Lawrence Summers, perhaps the foremost representative of the economics elite at the highest reaches of the recent policymaking apparatus, ushered in what became the academic establishment’s agreed-upon reason for dismissing Capital in the Twenty-First Century, aside from its empirical contribution: that Piketty’s theory that the marginal value of capital diminishes more slowly than capital itself accumulates (hence, the famous formulation, r > g) is at odds with econometric evidence suggesting the opposite conclusion about the “marginal elasticity of substitution” between labor and capital. Hence it cannot be that a rising capital-to-income ratio causes a rising capital share of national income and—since the owners of capital are, to an increasing degree, exclusively among the rich—rising income and wealth inequality.
Matthew Rognlie—then a doctoral student, now an assistant professor at Northwestern—took up that line in even greater detail in an article that eventually appeared in the Brookings Papers on Economic Activity, to which he added that the rising capital-to-income ratio in Piketty’s data is disproportionately the result of the price appreciation of certain scarce stores of wealth, primarily housing and the land it sits on, not the quantity accumulation of productive capital that is the subject of the neoclassical theory of economic growth.
Finally, Jason Furman and Peter Orszag, two senior economic policymakers in Democratic administrations, released “A Firm-Level Perspective on the Role of Rents in the Rise in Inequality,” which, while not a direct response to Capital in the Twenty-First Century, took indirect issue with foregrounding the shift in national income from labor to capital and with Piketty’s story of rising inequality within the distribution of labor income: that it results from the increased power of top executives in a within-firm bargaining context. Instead, they highlight rising dispersion of both firm-level profits and earnings as driving inequality dynamics between an in-group at the highest-profit, highest-earning firms, and everyone else.
But perhaps the greatest rebuke of Piketty to be found among academic economics is not contained in any of these overt or veiled attacks on his scholarship and interpretation, but rather in the deafening silence that greets it, as well as inequality in general, in broad swathes of the field—even to this day. You can search through the websites of several leading economics departments or the official lists of working papers curated by federal agencies and not come across a single publication that has any obvious or even secondary bearing on the themes raised by Capital in the Twenty-First Century, even in order to oppose them. It is as though the central facts, controversies, and policy proposals that have consumed our public debate about the economy for three years are of little-to-no importance to the people who are paid and tenured to conduct a lifetime’s research into how the economy works.
Despite Piketty’s resonance with the public, his book was mostly greeted with hostility by professional economists.
This dearth of reaction to such a critical work is not healthy. It is as if the rapturous reception by the public increased the resentment among Piketty’s academic economist colleagues. As an appeal to the public to resolve, or at least have a say in, what the experts consider their own domain, Piketty appears to have questioned the very value of having a credentialed economics elite empowered to make policy in the name of the public interest but not answerable to public opinion. The economics elite, it seems, answered by stonewalling Capital in the Twenty-First Century, so it would not have the impact on economics research agendas that it merits.
And yet the purpose of this essay is to catalog the ways that Piketty, and Capital in the Twenty-First Century, in particular, has had influence—great influence, even—on several distinct lines of scholarship to have grown up starting around the time the book was in draft and to have flourished and extended following its publication. In only some of the literatures discussed here is the reference to Capital in the Twenty-First Century overt, but thematically the links are there and it is important that they are acknowledged. Indeed, it is in economists’ interest that they be acknowledged, because without them the public would be justified in thinking that a cadre of well-paid and highly credentialed experts wields great influence over matters of direct relevance to their well-being without having to bother showing, even mildly, that they are remotely concerned with the public’s welfare.
Let’s begin with this question of splitting national income between capital and labor. Piketty’s theoretical apparatus on this question is wholly orthodox, in the sense that he assumes that what determines who gets what is the marginal productivity of each factor of production. Piketty’s crucial empirical finding, only possible with the long-time series data he and his coauthors created for each country, is that the rate of return on capital—the annual income that owning one unit of capital generates—has remained relatively constant over time, approximately 5 percent throughout Piketty’s timescale. That has been the case even as the total amount of capital in the economy has fluctuated a great deal: it declined during the world wars, was low during the middle of the twentieth century, and has gradually accumulated since. Given Piketty’s rather confining theoretical superstructure, this invariance in the rate of return can only be sustained by restrictive assumptions: that capital must remain productive as it is accumulated, and the way in which it remains productive is that it can be used to substitute for labor. Hence a high “marginal elasticity of substitution” in Piketty’s world.
So what of the literature on the capital/labor split in national income? In the book, Piketty relies on a paper by Loukas Karabarbounis and Brent Neiman that estimated a high marginal elasticity of substitution given two sets of facts: capital has gained national income at the expense of labor, and the price of capital inputs to production has been falling. The authors interpret the latter as the cause of the former, and again within the confines of neoclassical production theory, the only way to reconcile those two facts is with a high degree of factor substitutability: when capital becomes cheap, firms switch to using it, displacing workers. The labor displacement exceeds the increase in wages to workers who remain employed, thus reducing the total labor share of national income. That paper does not mention Piketty, but in a follow-up the next year, the same authors foreground the similar model proposed by Piketty.
It is increasingly possible to have a comfortable, rewarding life as a professional economist and never consider the issue of inequality.
The problem is that more plausible estimates of the elasticity of substitution between capital and labor conclude that it is low, even in a model that allows for aggregation across firms into an aggregate elasticity that is greater than the sum of its parts. Changes in factor prices can induce what economists call entry on the extensive margin: firms that wouldn’t have operated at all under one set of prices increase production above zero under another, giving rise to a higher aggregate elasticity. This is what Ezra Oberfield and Devesh Raval do in their 2014 paper “Micro Data and Macro Technology,” in which Capital in the Twenty-First Century is the very first citation, though not a favorable one. Their entire point is that given factual econometric estimates of production parameters, Piketty’s model is not up to the task of explaining his data—the “Piketty Puzzle.” What then?
The major new departure is work that eschews the standard neoclassical assumption that marginal productivity of factors in production set each factor’s share of national income. The aforementioned paper by Furman and Orszag argues that what Piketty (and Gabriel Zucman, in their joint work) identify as a high and invariant rate of return on capital in aggregate in fact reflects uncompetitively high rate of return on capital for a few specific, superstar firms in the economy, and the key task is not to manipulate aggregate macro equations, as Piketty (and neoclassical economists more generally) do, but rather to explain why these superstar firms do so much better than everyone else. The short version, according to Furman and Orszag, is rents—payments that some agents, superstar firms in this case, are able to extract from the rest of the economy either because they have successfully blocked any competitive pressure or because they have bought special treatment through the political system, or some combination of the two, in addition to other mechanisms.
An additional, and crucial, finding in the Furman and Orszag paper in support of the “rents story” and against the capital-labor substitution story is that the cost of capital has been low or even declining, all while the return on capital has been high. A model of perfect competition would predict these two things would be equal in equilibrium, as profit margins are driven to zero either by expanding output in existing firms or new entrants. But cheap capital has not caused more investment by either incumbents or entrants, just high profits—which, as the Goldman Sachs investors’ note, is a failure of capitalism. Indeed the distinction between high return on capital and low cost of capital could be considered a restatement or reinterpretation of Rognlie’s findings about the dynamics of the capital share: that it is driven by price appreciation of existing capital, like land and housing, rather than quantity accumulation of new productive capital, as the neoclassical theory implies.
That low cost of capital is the crucial point where Simcha Barkai departs from Piketty and gives a formal theoretical treatment to the empirical phenomena documented by Furman and Orszag. His paper measures the value of capital not by its high return, like Piketty does, but rather by its low cost. Given that the cost of capital has declined substantially, replacing all the productive capital that currently exists in the economy would be relatively cheap. The difference between that measure of the capital share and the measure in the national accounts (i.e., what capital is paid, not what it costs) is what Barkai calls the “profit share,” and he concludes that if we allow for it in a macroeconomic model, the profit share has soaked up large declines in both the labor and the capital share in recent decades. Thus, it is not productive capital, but unproductive and uncompetitive ownership, driven by high markups over the cost of production of output, that accounts for shifts in factor shares of national income. By that means do we resolve the “Piketty Puzzle” of the invariant return on capital, alongside a low measured elasticity of substitution between labor and capital.
Barkai’s reliance on a formally uncompetitive model, in which the return on capital remains higher than the cost of capital and the economy operates at less than its potential output on a prolonged basis, fits nicely into the debate over “Secular Stagnation” in the macroeconomy: whether there is a structural absence of aggregate demand that keeps the economy below full employment and drags down both wages (and hence the labor share) and the risk-free interest rate (and hence the capital share, calculated in Barkai’s way). This line of research thus links Piketty’s distributional study with one of the major issues at the forefront of macroeconomics and research into aggregate dynamics: why is aggregate demand low and sluggish (“Secular Stagnation”), and relatedly, why are corporations accumulating retained earnings on their balance sheet (“rising corporate net saving”) rather than investing and expanding operations?
Macroeconomic accounts of secular stagnation have for the most part operated independently of distributional questions, at least on the surface, but they have asked why the economy is spending less than it once did, and implicitly, why income is flowing to economic agents who save at a high rate, relative to where it used to flow—to the rich, rather than the poor. Several theoretical models in this vein have been published, but none of this work actually cites Piketty, and all of it is, at least theoretically, about the consumption and savings decisions of households. Perhaps the closest tie between the literature on macroeconomic dynamics, secular stagnation in particular, and Piketty’s work is “Aggregate Demand and the Top 1%,” by Adrien Auclert and the aforementioned Rognlie, in which Capital in the Twenty-First Century is actually cited.
Moving from the theory of macro-relevant household behavior to the empirics of macro-relevant firm behavior, the literature on the rise of corporate net savings is only just beginning to come to fruition. Karabarbounis and Neiman, along with coauthor Peter Chen, recently released a paper that documents this trend among corporations globally and models why it might be happening: the decline of the labor share and the rise in the profit share, in part thanks to markups, in part thanks to declining net corporate income tax payments. That paper is a more complete and theoretical treatment of the issue of rising corporate net saving than appeared earlier in a paper by Steven Kamin and Joseph Gruber, two senior macroeconomists advising the Federal Reserve Board. Finally, a recent working paper by Thomas Philippon and German Gutierrez compares sluggish corporate investment with what a canonical theory of corporate investment behavior would predict, namely, high investment when profits are high. It too zeroes in on the explanation of high markups and a lack of competition in the output market.
In addition Barkai’s work places the question of how factor shares are determined, if not by their marginal product, at the center of a post-Piketty economics research agenda. This is precisely what Suresh Naidu begins to explore in “A Political Economy Take on W / Y,” (the ratio of wealth to output), which Piketty himself approves as the way forward for research post–Capital in the Twenty-First Century. Such a research agenda would necessarily foreground politics and power operating at a deep level in the macroeconomy, the one phenomenon that neoclassical economics was never able to take on seriously or honestly—a major reason for the absence of inequality-relevant research from broad swathes of the field. But the accumulation of evidence and theory since Piketty’s publication makes it obvious that such a research agenda is vitally necessary, and thankfully it is starting to take shape.
It is in the best interest of economists to take Piketty seriously, lest they give the impression of caring little for the public’s welfare.
In particular, the literature on the common ownership of firms in an industry by a small set of asset managers highlights the issue of both the rising power of shareholders and their increasingly active management of companies in their portfolio, even when they are not considered “activist shareholders.” Asset managers have consolidated to the point that even where competition appears to exist among firms, in fact, the same small set of shareholders runs all the firms in an industry, in such a way that the shareholders benefit from weak competition. Several papers document this for airlines and banking and also show that common shareholders tend to reward the executives at the firm they own on the basis of industry-wide profits, not firm-specific ones. Not surprisingly then, competing on price to snag market share from the competition is not a popular management strategy.
Others have shown that common shareholders, who tend to be large, “passive” investment funds, in fact operate in concert with smaller activist shareholders. The activists, who explicitly campaign among the other shareholders to intervene in the management of firms in order to increase profits, can more easily gain control of a firm when there are fewer other shareholders in control of larger blocks of shares. In effect, common ownership serves to solve the activists’ coordination problem of winning a shareholder election: the electorate is smaller and thus a larger share of that electorate is pivotal. That, in turn, disciplines the behavior of management: above all, reward the owners. (Increasingly, shareholders and management are one and the same, as the rise of Private Equity and the decline in the number of publicly traded corporations shows.)
Finally, there is preliminary support for the idea that common ownership contributes to low corporate aggregate investment and high net savings. Firms that exist and collude for the benefit of shareholders have no incentive to plow their profits back into expanding their output by long-term investment—instead, they are retained or paid out for the shareholders’ use.
An interesting observation about the common ownership literature is that all or nearly all of its contributors work in business schools, not the more traditional loci of original research in academic economics departments. In part this reflects the rapid growth of research budgets in business schools as demand for their professional credentials has skyrocketed. But there may be something else going on: as the “inequality debate” increasingly requires the setting aside of received wisdom—almost all of it theoretical and developed over the last seventy years by academic economics—the researchers with the weakest ties to the departments where that theory was developed might be the ones who are most willing to question its empirical relevance.
A central economic critique of Piketty is that his explanation for rising inequality foregrounds the capital-labor split, whereas in the data, it is inequality within the distribution of labor income that accounts for the lion’s share of the rising overall income inequality. It is worth noting that this critique is itself slightly out of date: in 2016, Piketty, Saez, and Zucman published “Distributional National Accounts” and showed that, while rising inequality is indeed a labor income story up through about 2000, after that it becomes a story predominantly about capital income. Thus the issue of factor shares really is the main event and subject for research interest, contra many Piketty critics.
But Piketty does indeed have a story of rising labor income inequality: that the senior managers of large corporations are able to take home an increasingly disproportionate share of total corporate revenues, at the expense of the other workers they employ. This is a very different story of inequality within the labor income distribution than economists are used to telling. Their standard account is that earnings inequality is caused by differences in “human capital,” and the central observable counterpart to the theoretical notion of human capital is educational attainment.
The sorting of high-wage workers into high-paying jobs is really the story of low-wage workers being pushed out.
Piketty’s book and the inequality statistics that preceded it already called that theory into question as a viable explanation for the rise of the 1 percent—after all, the 1 percent are not educated in a substantially different way than the top 10 or even the top 25 percent. And within the top 1 or 10 percent, it is not the most educated whose earnings have increased the most. Hence there was already reason to doubt that a human capital model could do the job.
Where Piketty’s critics have pounced, however, is on the alternative he offered: intra-firm bargaining. In fact, there is good evidence that the rise of inequality in labor income in the population as a whole is driven substantially by inequality between firms—that high-paying firms (and their workers) have prospered and low-paying firms and their workers have stagnated or declined. This is where Furman and Orszag get their “superstar firms” story, which is based mainly on the increased sorting of highly-credentialed workers into high-paying firms.
Most striking, however, is that when you hold educational attainment and other observable worker characteristics constant, pay is starkly different depending on the firm where you work, even within narrowly-defined education categories, industries and occupations. This is prima-facie evidence that the human capital model in a competitive labor market is an increasingly poor way to explain earnings inequality. Wages for similar workers do not, in fact, equilibrate across firms.
The evolution of a single paper, “Firming Up Inequality,” is instructive in this regard. The first version, released publicly in May 2015, explicitly placed itself in opposition to Piketty’s story of intra-firm bargaining and CEO domination. Instead, it focused on firm heterogeneity and the mystery of the lucky and unlucky firms and their workers—the idea that informed Furman and Orszag. But the paper underwent a serious revision in the space of a year (Furman and Orszag’s paper was released in the interim), and the version dated June 2016, takes a very different line: it is not superstar firms, but rather wage segregation and, implicitly, monopsony.
The sorting of high-wage workers into high-paying firms, it argued, is really the story of low-wage workers being pushed out of high-wage firms. Whereas once the corporation, broadly defined, was a relatively egalitarian social construct employing people at every percentile of the earnings distribution, it is increasingly segregated, existing only to the benefit of its top earners and relegating the rest to “subordinate” firms, in what the economist and former Labor Department official David Weil has called the “fissured workplace.” By this means, “lead firms” evade responsibility under labor law (as well as custom and social pressure) to provide benefits and compress the earnings distribution at the firm level. If workers are outside the firm’s walls, so the story goes, then it is easier to squeeze them for concessions on benefits and working conditions and harder for them to make claims on the firm’s profits, which can therefore be retained for insiders.
The newer version of “Firming Up Inequality” makes it more consistent with a long-developing strand of labor economics research associated with the senior scholar David Card, who has long focused on departures from the competitive model of the labor market and the notion that human capital and observable worker characteristics are what set wages. He and his coauthors released a review paper in 2016 that, in addition to rationalizing firm-specific wages with a model of imperfect competition in the labor market, also links the issue of firm wage heterogeneity with the matter of firm productivity heterogeneity. This is a source of great controversy in the subfield of industrial organization, which in turn links back to the question of “superstar” firms and how they come to exist.
Once again, we have now ventured from the realm of papers that explicitly cite Piketty to those that do not. But that does not mean we have ventured outside Piketty’s sphere of influence—once it becomes clear that shifting power dynamics within firms can manifest as interfirm inequality given the strategic behavior involved in deciding who is in and who is out, the story Piketty tells in Capital in the Twenty-First Century looks well able to assimilate the new data on interfirm inequality.
So where does that leave us, and specifically, where does it leave Capital in the Twenty-First Century, three years after its publication? It seems strange, perverse even, to say that its influence has been “quiet” when it has had great influence on public debate. But what this tour of the landscape of academic economics tells us is that, despite its hostile reception, Piketty’s influence, and that of this book in particular, continues to grow in the academic realm and is not likely to wither and die anytime soon—much as that might pain the harshest critics or the many more who have kept their distance.
For the latter, unfortunately, it is all too easy to keep looking the other way. It is increasingly possible to have a comfortable and rewarding life as a professional economist and never even consider the broad issue of inequality or the controversial explanations for and consequences of it that Piketty offers. Social norms used to require economists to at least take on broad public sentiment and to consider the issues of the day when setting their agendas, but the amount of money available for economics research and teaching has never been higher, no matter the esteem (or lack thereof) in which economists are held by the public. High officials in government, in corporate boardrooms, in courtrooms, and in university administrations, alumni bodies, and boards of trustees still want to hear what economists have to say (or at least to make a point of ostentatiously seeking out their advice and approval), and to have that approval validated in public.
All of which avoids the crucial question: are we actually doing or saying anything to make the economy serve the people who inhabit it? Economists could very easily spend their individual and collective lives avoiding that question as the economy crumbles around them, with Piketty’s book serving as little more than a cry in the wilderness. Right now, there is no assurance it won’t end that way, but by reading between the lines, my suspicion—and hope—is that Piketty is not one in a series of pop–social science fads. Rather, his work on inequality is an agenda-setting and generation-marking intellectual achievement, potentially as explosive (albeit with a longer fuse) in academia as it has been outside of it.