In the mid 1950s, Simon Kuznets published historical data on the distribution of income in the United States, covering a 35-year period beginning in 1913. This was the first time any such statistics had been compiled for analysis, and they showed a sharp reduction in inequality over that period. In subsequent work, Kuznets argued that income inequality is likely to increase when countries begin to industrialize and experience economic growth, but that eventually the trend in income inequality begins to reverse. This bell-shaped relationship between economic growth and inequality became known as the Kuznets curve, and it broadly represented evidence that economic growth would ultimately and automatically solve problems of income inequality.
In his new book, Capital in the Twenty-First Century, Thomas Piketty compiles data on income inequality across countries and across much longer time horizons than Kuznets did. For some countries, he presents centuries of data. What Piketty finds is that while Kuznets was correct about income inequality for the period he examined, the implications of the Kuznets curve do not hold. Using the share of income earned by the top 10 percent of income earners as his measure of inequality, Piketty shows that income inequality declined beginning in the late 1930s and early 1940s, but began rising again in the late 1970s and early 1980s. This pattern of inequality is not confined to one particular country, but rather is broadly apparent across developed countries over the course of the 20th century.
Piketty notes that the rise in income inequality is also coincident with a rise in the capital-to-income ratio (i.e., the ratio of the stock of capital to total income) over the same period, and argues that the two phenomena are related in a precise way: When the rate of return on capital rises faster than the rate of economic growth, income from capital rises more than total income. As a result, the share of income that goes to owners of capital rises at the expense of the share of income that is going to wage earners. Piketty supports this claim by presenting evidence that the capital share of income has been rising for the past few decades. As the capital share of income rises, wealth and inheritance become a substantially more important source of income relative to wage income.
Overall, this presents a bleak vision of the future. According to Piketty’s view, if the capital-to-income ratio continues to rise, we are likely to continue to see rising income inequality and an ever-greater importance of inheritance and wealth. Nevertheless, lost in the midst of the volume of historical evidence is any notion of what is optimal. Certainly the presence of perpetually increasing income inequality is undesirable, but just how much inequality is too much inequality? What will be the result, and what does it imply for policy?
While there are places in the book where Piketty gives hypothetical benchmarks of potentially desirable allocations, these benchmarks are entirely arbitrary. His theoretical framework is ill-equipped to determine the optimal level of inequality, and one could argue that it is similarly ill-equipped to make predictions about the future. The reason for this is that what Piketty presents as a theory is really an accounting identity: His conclusions simply follow from the definition of the capital share of income. When the rate of return on capital rises faster than economic growth, the numerator (the capital stock) rises more than the denominator (total income). This implies that capital income is rising faster than total income.
Nonetheless, Piketty uses this theory to make predictions, arguing that growth is likely to be lower in the near future than it has been in the recent past and that this is likely to exacerbate the increase in the capital share of income. However, one is left to wonder what happens to the rate of return on capital when the growth rate of the economy changes. Perhaps, as Piketty assumes, the rate of return on capital can be held constant. Nonetheless, it would be preferable to understand the dynamics by which the growth rate of the economy and the rate of return on capital are determined in equilibrium, given how important they are to Piketty’s framework and predictions.
Piketty believes that the increasing importance of inheritance is problematic. He suggests that a greater reliance on inheritance is inconsistent with free, democratic, and meritocratic societies. But does this mean that all inheritance should be eliminated? If not, what types of policies should be enacted to curb an excess dependence of income on inheritance? Piketty says that the best tool to “avoid an endless inegalitarian spiral and to control the worrisome dynamics of global capital concentration” is a global tax on capital. This would be a progressive tax, in that it would tax larger fortunes more heavily than smaller ones.
Piketty’s policy solution is logically consistent with his concern regarding the growing importance of inheritance, but it is inadequate. A better conceptual framework for devising tax policy — and one that is consistent with the theoretical literature in this field — would be the following: Some people are born to wealthy parents, and others are born to poor parents. This is what is known as an “idiosyncratic risk.” The government might want a tax policy that insures individuals against this risk; this would tend to be a policy that has high taxes on inheritance. On the other hand, high taxes would discourage the accumulation of wealth. The government therefore must balance the desire to insure individuals against the lottery of birth with the desire to encourage the most productive members of society to be as productive as possible and, in the process, accumulate wealth.
The implications of this framework are at odds with the policy solutions that Piketty suggests. Specifically, while this framework suggests that the optimal tax on inheritance should be progressive, it also implies that the marginal tax rate should be negative. In other words, the optimal tax policy for inheritance is to subsidize inheritance and to reduce the size of the subsidy with the size of the inheritance. It is easy to understand the intuition behind this conclusion: Subsidizing inheritance prevents the deterrent effect of taxation, and the greater subsidization of inheritance for children of poor parents reduces the risks associated with birth.
This conclusion is important not only because it differs from the policy outlined by Piketty, but also because it is derived from a framework in which the idiosyncratic risks that one worries about when discussing inheritance and inequality are explicitly taken into account.
A separate but related issue is the persistence of wealth. Economists Gregory Clark and Neil Cummins have tracked individuals with the same surnames in England over the period from 1858 to 2012, and found that wealth is indeed highly persistent across generations. This would seem to provide support for Piketty’s view that the growing importance of inheritance is likely to exacerbate inequality in the income distribution; but in fact, these findings cast doubt on the effectiveness of his preferred policy solution. Specifically, as Clark and Cummins note, the persistence of wealth across generations is roughly constant over time, and robust in the face of very significant changes in institutions and tax policies. In addition, with various other co-authors, Clark has found that these results hold across several countries. As the authors note, the high degree of persistence in wealth across generations and tax regimes is probably owing to the fact that the inheritance of wealth is correlated with a great deal of other family-dependent traits, such as family structure, work ethic, and genetics. Given that widely different tax policies across generations have had little effect on the persistence of wealth, it is unclear how a global tax on wealth would produce substantially different results.
The preceding arguments cast doubts on Piketty’s policy prescriptions. A separate question is whether the empirical claims he makes hold up under close scrutiny. For example, Piketty’s main argument is that inheritance is becoming a more important determinant of income, that this is evident from the rising capital share of income, and that this trend is likely to continue. But what if he’s wrong about the capital share of income? Using data compiled by Gregory Clark from every decade in England from 1200 to 1860, I calculate that the average capital share of income has been roughly 21.6 percent — a percentage not substantially lower than the most recently observed figures. Even after we update these data using decade averages since 1860, the average does not change.
The fact that today’s capital share of income might be of a different magnitude than average is not sufficient to argue that the capital share has actually changed. In reality, one must demonstrate that today’s capital share of income differs from the historical average by more than what can be explained by randomness. The capital income share observed today in England is not different in a statistically significant way from the average over this period, so what Piketty depicts as a rising trend might in fact be simply the result of randomness. Given the central role that the capital share of income plays in Piketty’s explanations and predictions about income inequality, caution about his theses appears to be in order.
Despite these criticisms of Piketty’s theoretical framework and policy recommendations, one should not dismiss the contribution he has made with this book. Capital in the Twenty-First Century is a significant contribution to our understanding of the distribution of wealth and income — in particular, in his evidence suggesting that there is no inherent tendency for the distribution of income and wealth to become less unequal as countries grow. In addition, the sheer volume of data Piketty has compiled and summarized in this text is a significant contribution to economic science. Nonetheless, there is much work to be done in understanding the mechanisms that lead to inequality, and the policy implications thereof.
— Joshua Hendrickson is an assistant professor of economics at the University of Mississippi.