Is Piketty’s treatment of housing an excuse to ignore him?
French economist Thomas Piketty’s treatment of housing as capital in his blockbuster “Capital in the 21st Century” is not an excuse to ignore his predictions about rising economic inequality. “Capital in the 21st Century” is clear from the beginning that housing—and real estate generally—ought to be included in the definition of “capital” for the book’s purpose, which is to examine the aggregate effect of accumulated wealth that produces an annual return through no effort on the part of its owner.
A whole set of Piketty “rebuttals” attacks that treatment of housing as capital. The critics focus on that aspect of his analysis because a large proportion of the increase in the capital-to-income ratio that he emphasizes is thanks to the accumulation of housing and real estate at market prices. In some countries, the rise in the value of housing accounts for all of the increase in the capital-to-income ratio since 1970. And even beyond housing, Piketty’s approach is not consistent with standard neo-classical economic theory in several important ways—and so the critics have looked to housing as a reason to cling to their theory in the face of his countervailing facts.
But if housing were not counted as part of capital in Piketty’s analysis then the wealth distribution patterns he explores would be even more skewed. He identifies one key historical phenomenon that is unique to the 20th century—the rise of what he calls a “patrimonial middle class,” that is, non-trivial, inheritable wealth holdings by those between the 50th and 90th percentiles of the wealth distribution in advanced economies, rather than the top decile holding nearly all the wealth as was true in the 19th century and previously.
So, even though total capital accumulated (as a percentage of national income) has already reached the level where it was in the late 19th century, inequality has not yet attained its prior height, at least not in Europe, because that capital is partly held by the middle class. And their wealth is largely in housing.
This means that removing housing from consideration would by fiat skew the wealth distribution as much as Piketty predicts it will be skewed in the future if that patrimonial middle class dwindles, which is the dire outcome his whole book warns against. In other words, the arguments put forward as part of an attempt to discount Piketty’s prediction about the rise in capital’s share of income would, if accepted, put his prediction about the evolution of wealth inequality into effect mechanically. So how should economists interpret housing wealth?
Is housing capital?
Claim: Housing is not “productive” capital, like machines or factories or even farmland.
Response: In fact, housing is productive in the sense that it produces a good—shelter, broadly conceived—that economic agents value. Someone who owns his or her home doesn’t have to pay rent, and the owner of an investment property will earn rent exactly as does the owner of some piece of “productive” capital equipment. Second, capital functions as a store of value, and not simply as an input to production—a function that Piketty’s historical data show has been vital over the long run.
To take the argument a bit further, the value in real estate is often a matter of proximity—of “location, location, location”—and proximity to productive economic activity or to agents whom it is valuable to know is a very real economic resource. That proximity is capitalized as real estate. An identicallysized and equipped dwelling in Manhattan, Kansas costs much less than one in the more famous Manhattan because of the productivity and amenity benefits that come from living and working near many other people in New York City.
Arguably, in a world of increasing population density, location has been getting more important, and incumbent owners of real estate have been the main beneficiaries. Klaus Desmet at the University of Charles III and Esteban Rossi-Hansberg at Princeton University have a 2014 paper called “Spatial Development” that emphasizes population densification as the cause of productivity gains, location rent dynamics, and inter-sectoral employment flows in the United States. Piketty doesn’t discuss spatial trends as such, but the dynamics of housing wealth are entirely consistent with the argument in “Capital in the 21st Century.”
Excluding housing and real estate from the capital stock is convenient if the goal is to dismiss Piketty’s data and predictions, but that interpretation is not warranted by economic theory or empirical analysis.
How should the value of housing be calculated?
Claim: Piketty’s use the market price of housing distorts his analysis because housing should be priced according to its discounted rental stream, which is a measure of its “fundamental value.” The market price is subject to bubble dynamics, which according to the standard economic theory of bubbles would occur when the price deviates from some notion of its fundamental value, such as if a fruit tree were to cost more than the appropriately discounted sum of all the fruit it will ever produce.
Response: This critique fails for the same reason the argument that housing isn’t capital fails—because capital, broadly conceived to include wealth, is a store of value, and thus the stream of annual rent from its users isn’t the only relevant aspect of the return to its owner. The price of housing in central metropolitan areas has been on an upward march for the past several decades, in part because economic agents foresee dynamics of the kind described by economists Desmet and Rossi-Hansberg and in part because investing in real estate abroad is an effective way for the wealthy global elite to stash their cash overseas. At times, this has given rise to the property bubbles observed in Japan before 1990 and in the United States before 2006, but the resulting firesales do not erase the long-run trend. The fluctuation in housing market prices probably occurs in part because of misperceptions or misevaluations of the timing of long-run trends, but that does not imply the trend doesn’t exist.
Is housing really as substitutable with labor as Piketty assumes?
Claim: Including housing in the stock of capital should reduce Piketty’s assumed value for the marginal rate of substitution between capital and labor, and hence his prediction of an increasing capital share of income. Piketty’s assumption that the marginal rate of substitution between capital and labor is greater than one is important since it implies that the aggregate rental rate of capital will not decline by as much as the stock of capital increases. In other words, workers face no threat of losing their jobs just because more houses exist since there’s “no way to substitute a house for a worker.”
Response: Piketty’s main argument that the marginal rate of substitution is greater than one—that workers are indeed threatened by capital accumulation, including housing—is based on the dual U-shaped historical evolutions of the capital-to-income ratio and capital’s share of income in the very long run. That implies that when capital is accumulated, the resulting decline in the price of capital is not large enough to offset the increase in its quantity. Hence, the total share of income going to capital is higher when there’s more capital. If the marginal rate of substitution were less than one, the capital share would move inversely with the capital-to-income ratio, and if it were equal to one, as neo-classical theorists generally assume, the capital share would not change at all with the capital-to-income ratio.
All the evidence that the marginal rate of substitution between capital and labor is less than one cited by the critics is drawn from relatively short-run studies of the tradeoff between capital and labor at the firm- or industry-level, and there are very good reasons to believe that long-run elasticities are higher. Piketty’s long-run, aggregate evidence already includes the historical value of housing in capital, so this critique doesn’t bring anything more to the table—Piketty already has an excellent empirical case for assuming the marginal rate of substitution is high: the dual U shapes cited above.
Moreover, the aggregate marginal rate of substitution incorporates a much larger range of empirical economic phenomena than simply how easy it is to substitute between two factors in the production of a single good. So interpretations that adhere narrowly to that premise—such as econometric estimates at the firm or industry level—are bound to fail. The neo-classical argument holds that the price of capital is determined by its marginal productivity, and that marginal productivity declines mechanically as the quantity of capital increases. That is the so-called Ricardian scarcity principle, named for the 19th century thinker David Ricardo.
The rate at which it declines depends on how substitutable capital and labor are. The argument that they are not very substitutable implies that additional capital is relatively useless—and hence that its price will get much lower as its quantity increases. Notably, if what is relevant about housing and its value dynamics is that it acts as a store of value, then there’s no reason to believe that diminishing marginal productivity is operative. That concept relates to the additional output produced by increasing the use of one input in production while holding all others constant, but there’s no production going on if what’s being amassed is a store of value.
In this sense, housing wealth accumulation is like hording a precious metal: how useful the metal is in the production of other goods is irrelevant to the value of the horde. Finally, there are strong empirical reasons to believe that the price of housing, and of capital in general, is not only determined by marginal productivity as in the traditional, neo-classical macroeconomic model. That is the subject of my next response.
Are housing price increases due to supply restrictions?
Claim: There’s been a great deal of research into the dynamics of the housing market since the housing bubble burst, starting in 2006, and especially the unsurprising conclusion that local housing supply elasticity is related to price dynamics, and further, that political pressure by homeowners and the mortgage lending industry, especially on the west coast of the United States, has constrained housing supply and led to the enormous price swings. Those supply restrictions have nothing to do with the rise in the capital-to-income ratio and the reasons for it proposed by Piketty.
Response: This explanation for housing price dynamics isn’t actually distinct from Piketty’s narrative. The Economist commentator Ryan Avent wrote about this eloquently: “Over the last few decades technological changes have greatly increased the return to locating in large cities filled with skilled people. Being in such places makes workers more productive and raises the income they are able to earn. But skilled cities have not allowed housing supply to expand to meet rising demand. Housing has therefore been rationed by price, pushing less productive workers toward cities where housing supply growth is higher and housing cost growth is lower.
As a result, fewer people live in the most productive places, and quite a lot of the gain from employment in productive places is captured by landowners earning rents thanks to artificial housing scarcity. This may mean lower overall productivity, more income inequality, and more income flowing to capital rather than labour.” In other words, what we have here is collective political action to make sure the price of housing remains high just as increases in the bargaining power of capital relative to labor have contributed to the decline of the labor share. There is no room in the neoclassical model for these effects—only for the Ricardian scarcity principle and diminishing marginal productivity—but that doesn’t mean they aren’t there.
The answers are clear
Piketty’s framework, including his decision to count housing as capital, does not map directly onto the standard neoclassical economic growth model—but his approach is more consistent with empirical reality in several key ways. The critics who want to cling to their outdated theories have latched onto his interpretation of housing as a way to do so, but given their theory’s many empirical shortcomings, they are seriously misguided.