Capital in the Twenty-First Century
What is more striking when one reads Pareto’s work with the benefit of hindsight is that he clearly had no evidence to support his theory of stability. Pareto was writing in 1900 or thereabouts. He used available tax tables from 1880–1890, based on data from Prussia and Saxony as well as several Swiss and Italian cities. The information was scanty and covered a decade at most. What is more, it showed a slight trend toward higher inequality, which Pareto intentionally sought to hide.30 In any case, it is clear that such data provide no basis whatsoever for any conclusion about the long-term behavior of inequality around the world.
Pareto’s judgment was clearly influenced by his political prejudices: he was above all wary of socialists and what he took to be their redistributive illusions. In this respect he was hardly different from any number of contemporary colleagues, such as the French economist Pierre Leroy-Beaulieu, whom he admired. Pareto’s case is interesting because it illustrates the powerful illusion of eternal stability, to which the uncritical use of mathematics in the social sciences sometimes leads. Seeking to find out how rapidly the number of taxpayers decreases as one climbs higher in the income hierarchy, Pareto discovered that the rate of decrease could be approximated by a mathematical law that subsequently became known as “Pareto’s law” or, alternatively, as an instance of a general class of functions known as “power laws.”31 Indeed, this family of functions is still used today to study distributions of wealth and income. Note, however, that the power law applies only to the upper tail of these distributions and that the relation is only approximate and locally valid. It can nevertheless be used to model processes due to multiplicative shocks, like those described earlier.
Note, moreover, that we are speaking not of a single function or curve but of a family of functions: everything depends on the coefficients and parameters that define each individual curve. The data collected in the WTID as well as the data on wealth presented here show that these Pareto coefficients have varied enormously over time. When we say that a distribution of wealth is a Pareto distribution, we have not really said anything at all. It may be a distribution in which the upper decile receives only slightly more than 20 percent of total income (as in Scandinavia in 1970–1980) or one in which the upper decile receives 50 percent (as in the United States in 2000–2010) or one in which the upper decile owns more than 90 percent of total wealth (as in France and Britain in 1900–1910). In each case we are dealing with a Pareto distribution, but the coefficients are quite different. The corresponding social, economic, and political realities are clearly poles apart.32
Even today, some people imagine, as Pareto did, that the distribution of wealth is rock stable, as if it were somehow a law of nature. In fact, nothing could be further from the truth. When we study inequality in historical perspective, the important thing to explain is not the stability of the distribution but the significant changes that occur from time to time. In the case of the wealth distribution, I have identified a way to explain the very large historical variations that occur (whether described in terms of Pareto coefficients or as shares of the top decile and centile) in terms of the difference r − g between the rate of return on capital and the growth rate of the economy.
Why Inequality of Wealth Has Not Returned to the Levels of the Past
I come now to the essential question: Why has the inequality of wealth not returned to the level achieved in the Belle Époque, and can we be sure that this situation is permanent and irreversible?
Let me state at the outset that I have no definitive and totally satisfactory answer to this question. Several factors have played important roles in the past and will continue to do so in the future, and it is quite simply impossible to achieve mathematical certainty on this point.
The very substantial reduction in inequality of wealth following the shocks of 1914–1945 is the easiest part to explain. Capital suffered a series of extremely violent shocks as a result of the wars and the policies to which they gave rise, and the capital/income ratio therefore collapsed. One might of course think that the reduction of wealth would have affected all fortunes proportionately, regardless of their rank in the hierarchy, leaving the distribution of wealth unchanged. But to believe this one would have to forget the fact that wealth has different origins and fulfills different functions. At the very top of the hierarchy, most wealth was accumulated long ago, and it takes much longer to reconstitute such a large fortune than to accumulate a modest one.
Furthermore, the largest fortunes serve to finance a certain lifestyle. The detailed probate records collected from the archives show unambiguously that many rentiers in the interwar years did not reduce expenses sufficiently rapidly to compensate for the shocks to their fortunes and income during the war and in the decade that followed, so that they eventually had to eat into their capital to finance current expenditures. Hence they bequeathed to the next generation fortunes significantly smaller than those they had inherited, and the previous social equilibrium could no longer be sustained. The Parisian data are particularly eloquent on this point. For example, the wealthiest 1 percent of Parisians in the Belle Époque had capital income roughly 80–100 times as great as the average wage of that time, which enabled them to live very well and still reinvest a small portion of their income and thus increase their inherited wealth.33 From 1872 to 1912, the system appears to have been perfectly balanced: the wealthiest individuals passed on to the next generation enough to finance a lifestyle requiring 80–100 times the average wage or even a bit more, so that wealth became even more concentrated. This equilibrium clearly broke down in the interwar years: the wealthiest 1 percent of Parisians continued to live more or less as they had always done but left the next generation just enough to yield capital income of 30–40 times the average wage; by the late 1930s, this had fallen to just 20 times the average wage. For the rentiers, this was the beginning of the end. This was probably the most important reason for the deconcentration of wealth that we see in all European countries (and to a less extent in the United States) in the wake of the shocks of 1914–1945.
In addition, the composition of the largest fortunes left them (on average) more exposed to losses due to the two world wars. In particular, the probate records show that foreign assets made up as a much as a quarter of the largest fortunes on the eve of World War I, nearly half of which consisted of the sovereign debt of foreign governments (especially Russia, which was on the verge of default). Unfortunately, we do not have comparable data for Britain, but there is no doubt that foreign assets played at least as important a role in the largest British fortunes. In both France and Britain, foreign assets virtually disappeared after the two world wars.
The importance of this factor should not be overstated, however, since the wealthiest individuals were often in a good position to reallocate their portfolios at the most profitable moment. It is also striking to discover that many individuals, and not just the wealthiest, owned significant amounts of foreign assets on the eve of World War I. When we examine the structure of Parisian portfolios in the late nineteenth century and Belle Époque, we find that they were highly diversified and quite “modern” in their composition. On the eve of the war, about a third of assets were in real estate (of which approximately two-thirds was in Paris and one-third in the provinces, including a small amount of agricultural land), while financial assets made up almost two-thirds. The latter consisted of both French and foreign stocks and (public as well as private) bonds, fairly well balanced at all levels of wealth (see Table 10.1).34 The society of rentiers that flourished in the Belle Époque was not a society of the past based on static landed capital: it embodied a modern attitude toward wealth and investment. But the cumulative inegalitarian logic of r > g made it prodigiously and persistently inegalitarian. In such a society, there is not much chance that freer, more competitive markets or more secure property rights can reduce inequality, since markets were already highly competitive and property rights firmly secured. In fact, the only thing that undermined this
equilibrium was the series of shocks to capital and its income that began with World War I.
Finally, the period 1914–1945 ended in a number of European countries, and especially in France, with a redistribution of wealth that affected the largest fortunes disproportionately, especially those consisting largely of stock in large industrial firms. Recall, in particular, the nationalization of certain companies as a sanction after Liberation (the Renault automobile company is the emblematic example), as well as the national solidarity tax, which was also imposed in 1945. This progressive tax was a one-time levy on both capital and acquisitions made during the Occupation, but the rates were extremely high and imposed an additional burden on the individuals affected.35
Some Partial Explanations: Time, Taxes, and Growth
In the end, then, it is hardly surprising that the concentration of wealth decreased sharply everywhere between 1910 and 1950. In other words, the descending portion of Figures 10.1–5 is not the most difficult part to explain. The more surprising part at first glance, and in a way the more interesting part, is that the concentration of wealth never recovered from the shocks I have been discussing.
To be sure, it is important to recognize that capital accumulation is a long-term process extending over several generations. The concentration of wealth in Europe during the Belle Époque was the result of a cumulative process over many decades or even centuries. It was not until 2000–2010 that total private wealth (in both real estate and financial assets), expressed in years of national income, regained roughly the level it had attained on the eve of World War I. This restoration of the capital/income ratio in the rich countries is in all probability a process that is still ongoing.
It is not very realistic to think that the violent shocks of 1914–1945 could have been erased in ten or twenty years, thereby restoring by 1950–1960 a concentration of wealth equal to that seen in 1900–1910. Note, too, that inequality of wealth began to rise again in 1970–1980. It is therefore possible that a catch-up process is still under way today, a process even slower than the revival of the capital/income ratio, and that the concentration of wealth will soon return to past heights.
In other words, the reason why wealth today is not as unequally distributed as in the past is simply that not enough time has passed since 1945. This is no doubt part of the explanation, but by itself it is not enough. When we look at the top decile’s share of wealth and even more at the top centile’s (which was 60–70 percent across Europe in 1910 and only 20–30 percent in 2010), it seems clear that the shocks of 1914–1945 caused a structural change that is preventing wealth from becoming quite as concentrated as it was previously. The point is not simply quantitative—far from it. In the next chapter, we will see that when we look again at the question raised by Vautrin’s lecture on the different standards of living that can be attained by inheritance and labor, the difference between a 60–70 percent share for the top centile and a 20–30 percent share is relatively simple. In the first case, the top centile of the income hierarchy is very clearly dominated by top capital incomes: this is the society of rentiers familiar to nineteenth-century novelists. In the second case, top earned incomes (for a given distribution) roughly balance top capital incomes (we are now in a society of managers, or at any rate a more balanced society). Similarly, the emergence of a “patrimonial middle class” owning between a quarter and a third of national wealth rather than a tenth or a twentieth (scarcely more than the poorest half of society) represents a major social transformation.
What structural changes occurred between 1914 and 1945, and more generally during the twentieth century, that are preventing the concentration of wealth from regaining its previous heights, even though private wealth overall is prospering almost as handsomely today as in the past? The most natural and important explanation is that governments in the twentieth century began taxing capital and its income at significant rates. It is important to notice that the very high concentration of wealth observed in 1900–1910 was the result of a long period without a major war or catastrophe (at least when compared to the extreme violence of twentieth-century conflicts) as well as without, or almost without, taxes. Until World War I there was no tax on capital income or corporate profits. In the rare cases in which such taxes did exist, they were assessed at very low rates. Hence conditions were ideal for the accumulation and transmission of considerable fortunes and for living on the income of those fortunes. In the twentieth century, taxes of various kinds were imposed on dividends, interest, profits, and rents, and this changed things radically.
To simplify matters: assume initially that capital income was taxed at an average rate close to 0 percent (and in any case less than 5 percent) before 1900–1910 and at about 30 percent in the rich countries in 1950–1980 (and to some extent until 2000–2010, although the recent trend has been clearly downward as governments engage in fiscal competition spearheaded by smaller countries). An average tax rate of 30 percent reduces a pretax return of 5 percent to a net return of 3.5 percent after taxes. This in itself is enough to have significant long-term effects, given the multiplicative and cumulative logic of capital accumulation and concentration. Using the theoretical models described above, one can show that an effective tax rate of 30 percent, if applied to all forms of capital, can by itself account for a very significant deconcentration of wealth (roughly equal to the decrease in the top centile’s share that we see in the historical data).36
In this context, it is important to note that the effect of the tax on capital income is not to reduce the total accumulation of wealth but to modify the structure of the wealth distribution over the long run. In terms of the theoretical model, as well as in the historical data, an increase in the tax on capital income from 0 to 30 percent (reducing the net return on capital from 5 to 3.5 percent) may well leave the total stock of capital unchanged over the long run for the simple reason that the decrease in the upper centile’s share of wealth is compensated by the rise of the middle class. This is precisely what happened in the twentieth century—although the lesson is sometimes forgotten today.
It is also important to note the rise of progressive taxes in the twentieth century, that is, of taxes that imposed higher rates on top incomes and especially top capital incomes (at least until 1970–1980), along with estate taxes on the largest estates. In the nineteenth century, estate tax rates were extremely low, no more than 1–2 percent on bequests from parents to children. A tax of this sort obviously has no discernible effect on the process of capital accumulation. It is not so much a tax as a registration fee intended to protect property rights. The estate tax became progressive in France in 1901, but the highest rate on direct-line bequests was no more than 5 percent (and applied to at most a few dozen bequests a year). A rate of this magnitude, assessed once a generation, cannot have much effect on the concentration of wealth, no matter what wealthy individuals thought at the time. Quite different in their effect were the rates of 20–30 percent or higher that were imposed in most wealthy countries in the wake of the military, economic, and political shocks of 1914–1945. The upshot of such taxes was that each successive generation had to reduce its expenditures and save more (or else make particularly profitable investments) if the family fortune was to grow as rapidly as average income. Hence it became more and more difficult to maintain one’s rank. Conversely, it became easier for those who started at the bottom to make their way, for instance by buying businesses or shares sold when estates went to probate. Simple simulations show that a progressive estate tax can greatly reduce the top centile’s share of wealth over the long run.37 The differences between estate tax regimes in different countries can also help to explain international differences. For example, why have top capital incomes in Germany been more concentrated than in France since World War II, suggesting a higher concentration of wealth? Perhaps because the highest estate tax rate in Germany is no more than 15–20 percent, compared with 30–40 percent in France.38
Both theoretical arguments a
nd numerical simulations suggest that taxes suffice to explain most of the observed evolutions, even without invoking structural transformations. It is worth reiterating that the concentration of wealth today, though markedly lower than in 1900–1910, remains extremely high. It does not require a perfect, ideal tax system to achieve such a result or to explain a transformation whose magnitude should not be exaggerated.
The Twenty-First Century: Even More Inegalitarian Than the Nineteenth?
Given the many mechanisms in play and the multiple uncertainties involved in tax simulations, it would nevertheless be going too far to conclude that no other factors played a significant role. My analysis thus far has shown that two factors probably did play an important part, independent of changes in the tax system, and will continue to do so in the future. The first is the probable slight decrease in capital’s share of income and in the rate of return on capital over the long run, and the second is that the rate of growth, despite a likely slowing in the twenty-first century, will be greater than the extremely low rate observed throughout most of human history up to the eighteenth century. (Here I am speaking of the purely economic component of growth, that is, growth of productivity, which reflects the growth of knowledge and technological innovation.) Concretely, as Figure 10.11 shows, it is likely that the difference r >