Capital Eats the World
Piketty’s Capital in the Twenty-First Century shows that not everything in mainstream economics is worthless.
In their essay last fall on the state of economics, Seth Ackerman and Mike Beggs charged that today’s mainstream is irredeemably captured by conservative ideology. The good news is they’re wrong — Piketty’s work testiﬁes to that.
Contemporary mainstream economics is a politically broad tent, and has a lot to contribute to economic analysis. But it needs to be struggled with, as many have in the debate surrounding Capital in the Twenty-First Century.
Every economics student learns the “Kaldor facts” of economic growth. One of these is that the share of national income going to capital has a long-run tendency to stay constant.
Heterodox economists have been pushing against this stylized fact for a decade, and ﬁnally mainstream economists are recognizing and documenting that far from being constant, the capital share has in fact increased around the world. A noted paper by Lukas Karabarbounis and Brent Neiman documented this in the corporate sector around the world, while Francisco Rodriguez and Arjun Jayadev show it for global manufacturing.
Two explanations for this phenomenon typically thrown around by economists are “trade and technology”: the global supply of labor has increased relative to capital, and technological changes have lowered the price of capital and increased the substitutability of capital and labor. Another explanation, of course, is political, with right-wing ideology and policy ideas diffusing around the world alongside the Great Right Turn and the demise of the Soviet Union.
Piketty suggests that the rise is a long-term structural trend – the outcome of decelerating population and productivity growth coupled with a profit rate (r) that stays steady. But what keeps r high? Piketty never explicitly says. This question is at the heart of the struggle over how to interpret his book.
The neoclassical approach would be to examine three sets of forces in the market for capital that could account for it: supply, demand, and taxes. The supply of capital is given by the savings rate, and one important idea in Piketty is that the taste for savings at the top looks little like the frugal ant saving in order to consume for the future, the conceit of optimal growth theory. Instead, he suggests that a better way to think about savings is through models where accumulation and the building of estates are ends in themselves.
Piketty and others have been exploring these kinds of models in academic papers, where multiplicative shocks to capital accumulation — random fluctuations in tastes, lifespans, fertility and investment opportunities — generate a skewed distribution of wealth. “Accumulate! Accumulate! That is Moses and the prophets,” ran a memorable line from Marx.
If it is an accurate description of the capitalist drive to invest and save, then the forces that drive the wealthy to accumulate might not just be the realization of future consumption, but instead an insatiable drive for security, sociological pressures, psychological fantasies of future empires, or other structural imperatives.
When you start thinking of savings this way, the case for taxing capital becomes much clearer. If the supply of capital is more like immobile real estate and less like footloose cash, basic economics suggests that we can tax it, because it won’t disappear, and you might even be doing some social good.
If people are saving to pass inheritances onto their kids, then the cost of taxing capital is depriving some of trust funds, not a comfortable retirement. Not only does it mean that certain standard theories saying optimal capital taxation is zero don’t hold up anymore. It also means that one-off re-distributions of assets won’t stay equal for long, so some kind of permanent capital tax is needed.
Piketty suggests a fruitful research agenda here. Once freed from the consumption Euler equation, the “frugal ant” view of saving, what theory of private sector savings do we need to best understand inequality and growth? The question about how to tax capital becomes less about the trade-off between savings and consumption, and more about how to implement global taxes to keep capitalists from taking their money offshore.
The other blade of the scissors determining r is the demand for capital. Piketty makes the argument that r is likely to stay higher than g because capital and labor are becoming more substitutable, which could be read as the incoming future of robot capitalism as well as increased trade with labor-intensive countries.
The upshot is that even though capital will keep accumulating, the rate of proﬁt will not fall much because we can keep substituting out workers with it. Peter Frase’s “Four Futures“captured this well; in one of the futures, an abundant, narrowly owned capital stock resulted in relatively low wages for everyone despite high output.
But we have heard this before. Consider the development of the tractor, which mechanized virtually all of agriculture over the 20th century. Somehow new desires and demands sprung up for new kinds of manufactured goods, many of pure entertainment value, and people stayed employed and real wages kept rising.
I do not think there is anything inevitable about how capital-labor substitution could evolve in the future. It is quite possible that future technological and organizational changes are labor-augmenting rather than labor-saving. I’ll return to this below.
Finally, Piketty can combine these supply and demand elements into a complete model of income distribution dynamics. Imagine an economy where capital is accumulated, but there are sudden shocks to savings/bequests as well as wages, and a high elasticity of substitution between capital and labor.
In this model, capital increases as a share of income, the rate of proﬁt doesn’t fall very much (because capital and labor are very easily substituted for each other), and the distribution of capital is very unequal (because persistently high r allows capital shocks to be amplified over time). It can be embellished with increasing rates of return in wealth, reﬂecting the fact that richer people can obtain better ﬁnancial services and diversiﬁcation in order to earn higher rates of return. This model gets most of the way in explaining the stylized facts about capital in Piketty’s book.
The Limits to Capital
What I’ve described above is the conventional liberal economist’s interpretation of Piketty’s work, the one that Piketty and his critics have coordinated on in public. The question of whether capital will eat the world boils down to the degree of substitutability between labor and a single aggregate capital.
Despite assuming competitive labor and capital markets, this setup explains rising inequality and increasing capital shares, and yields a justification for capital taxation, completely within a neoclassical model. If this was all that was there, it would still be a pretty big advance within economics.
In this conventional interpretation, Piketty stays inside orthodox growth theory. His results arise from modiﬁcations to the savings equation and the marginal-pricing production function, not from alternatives to them. If the problem is just very high substitutability, a variety of labor market reforms are taken off the table, as ﬁrms would just replace workers with machines when you raise the wage. Minimum wages would kill a lot of jobs, and unions would immediately induce ﬁrms to close.
But this is again contradicted by recent evidence on both fronts. More importantly, it misunderstands capital by putting politics outside the production function, rather than inside it.
The increasing elasticity of substitution between “capital” and “labor” may be as much determined by institutions and property rights as by technology. Think of the parallel with slavery. The robot economy and the slave economy may both have higher elasticities of substitution than industrial capitalism. Slaves could do virtually all the tasks of free labor, and were movable assets.
In “The Causes of Slavery and Serfdom,” Evsey Domar famously argued that it was a historical impossibility to have free labor, abundant land, and an aristocracy simultaneously. Free labor and abundant land would make aristocratic claims on labor impossible, abundant land and an aristocracy would require coerced labor, and only scarce land could depress wages enough to allow an aristocracy to coexist with free labor.
Perhaps a similar trilemma exists with abundant robots, dignified employment, and unequal capital ownership.
This sort of institutions-as-primitives thinking is how we should approach the question of capital. Capital is a set of property rights entitling bearers to politically protected rights of control, exclusion, transfer, and derived cash ﬂow. The capital share of income is just the last part of that sequence.
Like all property rights, its delineation and defense require actions of state power, legal standardization, and juridical legitimacy. In the last instance, capital includes the ability to call on the government to evict trespassers, be they burglars, sit-down strikers, or delinquent tenants.
In economics, we capture some of the political dimension of capital with incomplete contracts. Contracts between ﬁnanciers, entrepreneurs, and workers (among others) can never be completely speciﬁed. Instead, large domains of the economic transaction are left to the discretion of one side of the market.
A CEO like Steve Jobs complains about the power exercised by Apple’s shareholders in the late 1980s as surely as Jobs’s workers complain about the tyrannical power wielded by Jobs himself. As Ronald Coase argued, this distribution of power is not outside the market, but part of the transaction. Workers do what they are told because they can be kicked out of the ﬁrm. Capital here is seen as not just a ﬂow of income, but rather a right to exclude and appropriate. Focusing on balance sheets rather than bosses will miss this.
Seeing capital this way also blurs the line between supermanagers and rentiers. Supermanagers happen to have labor market contracts (in the form of bonuses and stocks and options) that entitle them to stupendous income when the ﬁrm is doing well. It is not clear that this is “labor” income as much as it is a form of capital that requires you to run meetings and wear a power suit.
Jointly, the rentiers and the supermanagers have cash flow and control rights inside the ﬁrm, and the institutions of corporate ﬁnance and governance that allocate these powers determine the demand for capital as surely as technology does.
The book is too good to miss this, however. It contains an excellent section on the gap between cash-ﬂow rights and control rights in corporate governance, which suggests a capital demand schedule derived not just from ﬁrm optimization decisions, but from the distribution of power within the firm.
The book points out that German shares are “underpriced” because shareholders there do not have the same level of political power as shareholders in the US and UK, since they have to share power with workers’ councils and other stakeholders. The same thing is true of unions in the US. David Lee and Alexandre Mas shows that strong union victories in NLRB elections once reduced stock prices, yet it is very unlikely they changed the replacement value of the company’s underlying assets.
The everyday encounter most people have with accumulated wealth is not through prices in the market for shoes, or the society pages, but instead the control and threats inﬂicted by their employers, landlords, and bankers. Inequality of income and wealth means that some people live off unjustly earned income, but it also means a lot more people are on the short-end of an asymmetric exchange, toiling away as personal assistants and Mechanical Turks.
This is where Piketty’s Walrasian conventions dampen his contribution: he discusses the ﬁrst, but not the second. It’s like saying slavery is an inequality of assets between slaves and slaveholders without describing the plantation.
Even Adam Smith suggested measuring a person’s income by the “quantity of that labor which he can command.” This has normally been taken to mean income of the rich relative to the wage. But it also means looking at “command”: what privileges and obligations can one demand from the soul purchased (or rented)?
An economy that allows indentured labor means that wealth can purchase more power over people; an economy with robust union contracts means that capital is trammeled in its control over the shop ﬂoor. From sexual harassment on the job to the indignities of gentriﬁcation and nonproﬁt funding, a world of massive inequality is a world where rich people get to shape environments that everybody else has to accept.
Piketty repeatedly announces that politics plays a large role in the distribution of income. But he neglects that the distribution of income and wealth also generates inequalities of larger privileges and prerogatives; wealth inequality together with a thoroughly commodiﬁed society enables a million mini-dictatorships, wherein the political power of the rich is exercised through the market itself.
In a thoroughly marketized world, the wealthy can purchase educational reform, the charity of their choice, think-tanks, legislative language, and faceless TaskRabbiters willing to work for a pittance. While feudal lords were wealthy, the absence of certain types of markets made their social power somewhat independent of wealth; the regalia and mounted vassals were an independent basis of status and were not simply purchasable.
But there is an important and nasty complementarity between massive inequality in income and wealth and a commodiﬁed, “fully-incentivized” world. When every action can have pecuniary rewards attached to it, and every source of well-being can be priced at exactly a person’s willingness to pay, the social power commanded by the rich is magniﬁed in a way that is difficult to see when comparing a dollar in 1920 with a dollar today.
Piketty’s big policy idea is taxing wealth directly, progressively, and globally. This is certainly important to put on the table, up there with other global problems like climate change, intellectual property, open borders, and, dare I say, reparations.
But the focus on taxes is again a straightjacket imposed by the equality-versus-eﬃciency lens through which too many public ﬁnance economists see policy issues. The preferred policy instruments are always taxes and transfers, when it is not at all clear that these alone are the best tools for reducing inequality (although they are surely useful for increasing it). This is the same technocratic spirit that makes American liberals love the Earned Income Tax Credit as the only redistributive arrow in the state’s quiver.
The structure and limitations of Piketty’s argument also explains the love the liberal American policy wonk has for it. It comes with a Zip ﬁle full of spreadsheets, a clear argument reasoned from data and common sense, the charisma of the economics profession, and a policy prescription that is technically feasible and politically hopeless.
Like the policy expert, it has neither utopian demand-it-all energy nor the concrete backing of a political actor aiming to win. The book reminds the American wonk community that if only their people could run the show, they have the expertise (and the data!) to produce ﬁnely-calibrated optimal policies without politics.
But the collapse in the capital and top income shares after World War II (and other wars) came along with radical transformations of all kinds of economic institutions, with millions of dead, sui generis geopolitics, and a host of newly mobilized popular forces. The obligations enshrined in balance sheets were destroyed by financial collapse and war, and kept in check by social democracy and postwar growth. Little in the way of clever policy advice mattered for any of this.
Where Do We Go from Here?
Piketty’s book reﬂects the promise and current limits of economics as a discipline. The ideas, which are powerful, could not have originated anywhere but mainstream economics. They require a command of the mathematical models of growth and taxation, and only economists would appreciate the painstaking reconstruction of the balance sheet data.
But Piketty oscillates between paying homage to fundamental forces of technology, tastes, and supply and demand, and then backtracking to say that politics and institutions are important.
So how to do better?
A first step could be a multisector model with both a productive sector and an extractive, rent-seeking outlet for investment, so that the rate of return on capital has the potential to be unanchored from the growth of the economy. This model could potentially do a better job of explaining r > g in a world where capital has highly profitable opportunities in rent-seeking rather than production, and it would generally disassociate the growth of the productive economy from the growth of abstract wealth. When people say neoliberalism was good for growth, they tend to be looking at the stock market, not GDP or wages.
More fundamentally, a model that started with the financial and firm-level institutions underneath the supply and demand curves for capital, rather than blackboxing them in production and utility functions, could illuminate complementarities among the host of other political demands that would claw back the share taken by capital and lower the amount paid out as proﬁts before the ﬁscal system gets its take.
This is putting meat on what Brad Delong calls the “wedge” between the actual and warranted rate of profit. Commentators have listed their pet policy proposals under this umbrella, from strengthening labor and tenants movements to weakening intellectual property rights and ﬁnancial regulation. And yes, maybe even selective inflation of nominal claims, as with the repudiation of the gold indexation clause in 1933.
We need even more and even better economics to ﬁgure out which of these may get undone via market responses and which won’t, and to think about them jointly with the politics that make each feasible or not. While Piketty’s book diagnoses the problem of capital’s voracious appetite, it would require a different kind of model to take our focus off the nominal quantities registered by state ﬁscal systems, and instead onto the broader distribution of political power in the world economy.