Free Lunch: Piketty meets his match

Sandbu, Martin, "Free Lunch: Piketty meets his match"Financial Times, March 20, 2015. Available at:http://www.ft.com/cms/s/0/bbf09ba8-cd51-11e4-9144-00144feab7de.html"...Rognlie does two things. In terms of theory, he notes a difference between the concept of substitutability used by Piketty and that commonly measured, depending on whether you measure capital (and output) before or after adjusting for depreciation (the investment needed just to offset the wear of machinery over time). Properly measured, he shows, the empirical evidence overwhelmingly shows that capital cannot substitute for labour to anywhere near the degree needed to fulfil Piketty's prophecies. Empirically, Rognlie looks more closely at capital's share of income, and finds that with the exception of one sector, net shares have remained largely constant aside from short-term fluctuations. (They fell slightly until the 1990s and have caught up since. Gross shares have increased more steadily, but that's because of greater use of capital that depreciates faster - software, for example.) The exception is housing, which accounts for a big part of capital accumulation and for all of the long-term rise in the capital income share. Something has clearly been going on in the housing sector - but as Rognlie points out, since the housing ownership is much more evenly spread than productive capital, it may be less worrying for inequality....."
Free Lunch: Piketty meets his match

Martin Sandbu

How could economists not fawn over Matthew Rognlie? A 26-year-old graduate student has delivered a technical critique of Thomas Piketty's theory of capital-driven inequality that is gaining status as the most devastating yet put forward. What started with a comment on the Marginal Revolution blog has grown into a paper that Rognlie is presenting at the Brookings Panel on Economic Activity today - the prime US conference on applied policy economics.

Recall that Piketty, in addition to adding significantly to our empirical knowledge about how incomes and wealth are distributed, theorises that capital accumulation plays a big role in exacerbating inequality and will continue to do so in the future. The idea is that the return on capital will remain sizeable while GDP growth slows (the famous r > g). This means the stock of capital will grow bigger relative to national income and so the share of capital returns in total income will grow at the expense of labour income.

The objection has always been that as capital accumulates, the return on it will fall. Piketty has shown that the rate of return has been remarkably stable historically, and theorised that it can remain so. Technically, how fast returns to capital fall depends on the "elasticity of substitution", ie the ease with which capital and labour can be substituted for one another in the production process. The more easily workers can be replaced by machines, the more capital the economy can absorb without reducing the reward paid to capital owners. If substitutability is moderate, the incremental capital becomes harder to employ productively, and returns fall.

Rognlie does two things. In terms of theory, he notes a difference between the concept of substitutability used by Piketty and that commonly measured, depending on whether you measure capital (and output) before or after adjusting for depreciation (the investment needed just to offset the wear of machinery over time). Properly measured, he shows, the empirical evidence overwhelmingly shows that capital cannot substitute for labour to anywhere near the degree needed to fulfil Piketty's prophecies.

Empirically, Rognlie looks more closely at capital's share of income, and finds that with the exception of one sector, net shares have remained largely constant aside from short-term fluctuations. (They fell slightly until the 1990s and have caught up since. Gross shares have increased more steadily, but that's because of greater use of capital that depreciates faster - software, for example.) The exception is housing, which accounts for a big part of capital accumulation and for all of the long-term rise in the capital income share. Something has clearly been going on in the housing sector - but as Rognlie points out, since the housing ownership is much more evenly spread than productive capital, it may be less worrying for inequality.

Rognlie's work is impressive, and he is receiving well-deserved star treatment today. His discussants are Robert Solow - the nonagenarian Nobel-winning economist known for his theory of economic growth - and Bradford DeLong. DeLong has posted his comments on Rognlie on his blog; they are entertaining and illuminating, though more for raising questions than giving answers.

The upshot is not that inequality doesn't matter. It is that it's hard to explain rising inequality in terms of capital accumulation. Instead, Rognlie concludes, we should focus on changing (and increasingly unequal) distribution within labour income.