Has Thomas Piketty's famous book Capital in the 21st Century been refuted by a 26-year-old graduate student? Not really.
The student in question, Matt Rognlie of MIT, has published a paper titled "Deciphering the Fall and Rise in the Net Capital Share" that adds valuable depth and context to the discussion on wealth that Piketty inaugurated. The paper's genesis is also a heck of a good story, ably related by Jim Tankersley; it will, Tankersley notes, "almost certainly be the first paper at the prestigious Brookings Papers on Economic Activity that was commissioned based on a blog comment."
But the nexus of the good story, the heavy politicization of Piketty, and the fact that almost nobody actually read Piketty's best-selling book have also opened the doors to confusion, and helped create a misimpression that Rognlie debunked the book's core findings. He finds that Piketty got some things wrong, sure, but hardly discredits the entire argument.
Here's the quick version of what you need to know:
- Piketty's empirical finding that wealth inequality has risen drastically is untouched.
- Piketty's empirical finding that the capital share of national income has risen (and the share going to workers has fallen) is also untouched.
- Rognlie finds that when you account for depreciation properly, the capital share's increase has been less dramatic than Piketty says.
- Rognlie finds that when you account for depreciation properly, the capital share's increase has been entirely about housing.
- Many economists are much less persuaded by Piketty's r > g model and predictions of the future than they are by his backward-looking empirical work.
- Whether you endorse Piketty's policy prescriptions continues to hinge overwhelmingly on issues that are outside the scope of either his or Rognlie's work.
Housing is central to the growing capital share of income
For the past several years, there has been a lot of discussion of the fact that wages and salaries paid to workers are a declining share of overall GDP. Dylan Matthews reviewed several leading theories on why this is happening recently, including the growing use of robots, the decline of labor unions, the rise of global trade, and other big concepts people care about. Theory five on his list was that the declining labor share is simply a question of bad statistics.
The issue here is depreciation. When a firm buys a computer, the value of that purchase erodes with time. A $2,000 computer bought in 2012 is not worth $2,000 in 2014. When you take that into account, business profits look less impressive. Rognlie says this theory explains most but not all of the apparent rise in the capital share:
Once you factor in depreciation, Rognlie finds that across a range of developed countries the rise in the capital share of income has been modest. What's more, it is entirely accounted for by the housing sector rather than by the kind of business investments — factories, office buildings, equipment, patents, etc. — that put the capital in capitalism.
Does this mean Piketty has been refuted?
That really depends on what it is you mean by "Piketty." Capital in the 21st Century is a long book, and it says a lot of stuff. If your main takeaway from Piketty's book was, "Wealth inequality has risen a lot — back to pre–World War I levels — and I am alarmed about it," then there is really nothing in Rognlie's research that would change your mind about it. This is what the vast majority of people who've talked about "Piketty" over the past decade or so have meant. A line of empirical research that Piketty pioneered with Emmanuel Saez, and that has been continued by various collaborators, shows that wealth inequality levels are higher than previously realized, and generally rising.
In his book, however, Piketty adds to that research by proposing that the future of inequality is likely to be dominated by inherited wealth. This is tied to his view that the rate of return on capital (r) is going to be higher than the overall economic growth rate (g). Piketty shows that this has been the case in the past, and then argues that it is likely to be the case in the future. Rognlie's paper argues that r > g is false for the non-housing sector, and thus challenges a particularly striking claim of Piketty's, though likely not the one that got a mass audience interested in his book.
Does housing count as capital?
Rognlie shows that capital income is on the rise "only if you count housing," which is in some ways less profound than it sounds. Capital is a term of art in certain kinds of macroeconomic theory, and in that context it makes a lot of sense to wonder about which kinds of things are really capital.
But in his book, Piketty very clearly and expressly uses the term capital ("le capital" in French) as an exact synonym for wealth ("le patrimoine" in French).
This was perhaps not a great idea, as it has tended to get discussions of his work mired in the Cambridge Capital Controversy and other obscure and annoying corners of economics. But if you just say wealth instead of capital, then obviously you have to count real estate wealth as a form of wealth. If you were to inherit a giant condo on Fifth Avenue and a beach house in the Hamptons and a ski lodge in Colorado, you would become a wealthy person.
What's more, as Guillaume Allègre and Xavier Timbeau show, a general increase in real estate prices leads to an increase in wealth inequality. This is true even though (as Rognlie says) housing wealth is more broadly owned than wealth in stocks and other financial assets. Wealthy people own more expensive homes (and just more homes, as in both vacation properties and rental properties), so if the price of all homes doubles, the absolute wealth gap between rich people and middle-class people grows.
Haven't lots of people besides Piketty talked about the shifting balance of power between labor and capital?
Yes. Rather than thinking of Rognlie as narrowly refuting Piketty, which is misleading, it's better to say he's offering a profound challenge to a much wider set of writings about labor's declining clout in the modern economy.
Rognlie's findings rather directly call into question the relevance of a lot of punditry about the future of work in an era of growing automation. After all, nobody thinks that there have been huge recent technological advances in the realm of landlording. Yet if labor's falling share of national income is entirely accounted for by the increased returns to housing capital, then it seems we should be looking at housing-specific trends to explain the problem. Rather than robots, the problem is almost certainly snob zoning rules that prevent the construction of new affordable housing in expensive areas.
Could upzoning reduce wealth inequality?
Rognlie cites work by economists Ed Glaeser, Joseph Ryorko, and Raven Saks to argue that exclusionary zoning practices have contributed greatly to lack of housing affordability and that this should be more central to the wealth inequality debate. Lawrence Summers, likewise, argued in a review of Capital that "an easing of land-use restrictions that cause the real estate of the rich in major metropolitan areas to keep rising in value" should be an important element of the policy agenda to address Piketty's concerns.
When I interviewed him last year, Piketty indicated that he is open to a wide range of complementary policies relating to housing and intellectual property and other things, but doesn't regard any of that as an ultimate substitute for progressive taxation.
Where does this leave Piketty's wealth tax idea?
It's doubtful that the technical issues around the return to capital actually have much bearing on the merits of Piketty's core policy proposal: the imposition of a small progressive tax on net wealth.
The main criticisms of this idea are that it would be difficult to implement and damaging to economic growth. The question of whether taxes on capital income hurt investment and economic growth is, as you would expect, subject to controversy, and really nothing about it hinges on the housing capital versus other kinds of capital issue. And of course, the very large practical implementation issues around a comprehensive wealth tax are either solvable or not solvable, regardless of Rognlie's points.
WATCH: 'President Obama on why income inequality has skyrocketed