Capital in the Twenty-First Century

by Thomas Piketty

Once both population and output begin to grow steadily, land tends to become increasingly scarce relative to other goods. The law of supply and demand then implies that the price of land will rise continuously, as will the rents paid to landlords. The landlords will therefore claim a growing share of national income, as the share available to the rest of the population decreases, thus upsetting the social equilibrium. For Ricardo,

In fact, all the historical data at our disposal today indicate that it was not until the second half—or even the final third—of the nineteenth century that a significant rise in the purchasing power of wages occurred.

The data we have assembled nevertheless reveal no structural decrease in inequality prior to World War I. What we see in the period 1870–1914 is at best a stabilization of inequality at an extremely high level, and in certain respects an endless inegalitarian spiral, marked in particular by increasing concentration of wealth.

In short, Marx took the Ricardian model of the price of capital and the principle of scarcity as the basis of a more thorough analysis of the dynamics of capitalism in a world where capital was primarily industrial (machinery, plants, etc.) rather than landed property, so that in principle there was no limit to the amount of capital that could be accumulated.

In fact, his principal conclusion was what one might call the “principle of infinite accumulation,” that is, the inexorable tendency for capital to accumulate and become concentrated in ever fewer hands, with no natural limit to the process. This is the basis of Marx’s prediction of an apocalyptic end to capitalism: either the rate of return on capital would steadily diminish (thereby killing the engine of accumulation and leading to violent conflict among capitalists), or capital’s share of national income would increase indefinitely (which sooner or later would unite the workers in revolt). In either case, no stable socioeconomic or political equilibrium was possible.

Marx’s dark prophecy came no closer to being realized than Ricardo’s. In the last third of the nineteenth century, wages finally began to increase: the improvement in the purchasing power of workers spread everywhere, and this changed the situation radically, even if extreme inequalit...

Marx evidently wrote in great political fervor, which at times led him to issue hasty pronouncements from which it was difficult to escape. That is why economic theory needs to be rooted in historical sources that are as complete as possible, and in this respect Marx did not exploit all the possibilities available to him.

the twentieth-century analyses of Simon Kuznets, we might say that economists’ no doubt overly developed taste for apocalyptic predictions gave way to a similarly excessive fondness for fairy tales, or at any rate happy endings. According to Kuznets’s theory, income inequality would automatically decrease in advanced phases of capitalist development, regardless of economic policy choices or other differences between countries, until eventually it stabilized at an acceptable level. Proposed in 1955, this was really a theory of the magical postwar years referred to in France as the “Trente Glorieuses,” the thirty glorious years from 1945 to 1975.9 For Kuznets, it was enough to be patient, and before long growth would benefit everyone. The philosophy of the moment was summed up in a single sentence: “Growth is a rising tide that lifts all boats.”

He noted a sharp reduction in income inequality in the United States between 1913 and 1948.

What are the major conclusions to which these novel historical sources have led me? The first is that one should be wary of any economic determinism in regard to inequalities of wealth and income. The history of the distribution of wealth has always been deeply political, and it cannot be reduced to purely economic mechanisms.

From a strictly theoretical standpoint, other forces pushing toward greater equality might exist. One might, for example, assume that production technologies tend over time to require greater skills on the part of workers, so that labor’s share of income will rise as capital’s share falls: one might call this the “rising human capital hypothesis.”

Another optimistic belief, which is current at the moment, is the idea that “class warfare” will automatically give way, owing to the recent increase in life expectancy, to “generational warfare” (which is less divisive because everyone is first young and then old). Put differently, this inescapable biological fact is supposed to imply that the accumulation and distribution of wealth no longer presage an inevitable clash between dynasties of rentiers and dynasties owning nothing but their labor power. The governing logic is rather one of saving over the life cycle: people accumulate wealth when young in order to provide for their old age. Progress in medicine together with improved living conditions has therefore, it is argued, totally transformed the very essence of capital.

Unfortunately, these two optimistic beliefs (the human capital hypothesis and the substitution of generational conflict for class warfare) are largely illusory.

One possible explanation of this is that the skills and productivity of these top managers rose suddenly in relation to those of other workers. Another explanation, which to me seems more plausible and turns out to be much more consistent with the evidence, is that these top managers by and large have the power to set their own remuneration, in some cases without limit and in many cases without any clear relation to their individual productivity, which in any case is very difficult to estimate in a large organization.

the return of high capital/income ratios over the past few decades can be explained in large part by the return to a regime of relatively slow growth. In slowly growing economies, past wealth naturally takes on disproportionate importance, because it takes only a small flow of new savings to increase the stock of wealth steadily and substantially. If, moreover, the rate of return on capital remains significantly above the growth rate for an extended period of time (which is more likely when the growth rate is low, though not automatic), then the risk of divergence in the distribution of wealth is very high.

he notes it was the high growth rate of the intra war period that was responsible for relative ewuality not the destruction from the war

When the rate of return on capital significantly exceeds the growth rate of the economy (as it did through much of history until the nineteenth century and as is likely to be the case again in the twenty-first century), then it logically follows that inherited wealth grows faster than output and income.

It is possible to imagine public institutions and policies that would counter the effects of this implacable logic: for instance, a progressive global tax on capital.

Note, for example, that it was not until the coming of the twenty-first century that the wealthy countries regained the same level of stock-market capitalization relative to GDP that Paris and London achieved in the early 1900s. This comparison is quite instructive for understanding today’s world.

In the late nineteenth century, conservative French economists such as Paul Leroy-Beaulieu often used this argument to explain why republican France, a nation of “small property owners” made egalitarian by the Revolution, had no need of a progressive or confiscatory income tax or estate tax, in contrast to aristocratic and monarchical Britain. The data show, however, that the concentration of wealth was as large at that time in France as in Britain, which clearly demonstrates that equality of rights in the marketplace cannot ensure equality of rights tout court.

the question of what share of output should go to wages and what share to profits—in other words, how should the income from production be divided between labor and capital?—has always been at the heart of distributional conflict.

In traditional societies, the basis of social inequality and most common cause of rebellion was the conflict of interest between landlord and peasant, between those who owned land and those who cultivated it with their labor, those who received land rents and those who paid them. The Industrial Revolution exacerbated the conflict between capital and labor, perhaps because production became more capital intensive than in the past (making use of machinery and exploiting natural resources more than ever before) and perhaps, too, because hopes for a more equitable distribution of income and a more democratic social order were dashed.

Taking all these elements into account, what is the “right” split between capital and labor? Can we be sure that an economy based on the “free market” and private property always and everywhere leads to an optimal division, as if by magic? In an ideal society, how would one arrange the division between capital and labor? How should one think about the problem?

The growth of capital’s share accelerated with the victories of Margaret Thatcher in England in 1979 and Ronald Reagan in the United States in 1980, marking the beginning of a conservative revolution.

The most natural and useful way to measure the capital stock in a particular country is to divide that stock by the annual flow of income. This gives us the capital/income ratio, which I denote by the Greek letter β. For example, if a country’s total capital stock is the equivalent of six years of national income, we write β = 6 (or β = 600%). In the developed countries today, the capital/income ratio generally varies between 5 and 6, and the capital stock consists almost entirely of private capital.

The capital/income ratio for the country as a whole tells us nothing about inequalities within the country. But β does measure the overall importance of capital in a society, so analyzing this ratio is a necessary first step in the study of inequality.

The First Fundamental Law of Capitalism: α = r × β I can now present the first fundamental law of capitalism, which links the capital stock to the flow of income from capital. The capital/income ratio β is related in a simple way to the share of income from capital in national income, denoted α. The formula is α = r × β where r is the rate of return on capital.

From 1900 to 1980, 70–80 percent of the global production of goods and services was concentrated in Europe and America, which incontestably dominated the rest of the world. By 2010, the European–American share had declined to roughly 50 percent, or approximately the same level as in 1860. In all probability, it will continue to fall and may go as low as 20–30 percent at some point in the twenty-first century. This was the level maintained up to the turn of the nineteenth century and would be consistent with the European–American share of the world’s population (see Figures 1.1 and 1.2).

the lead that Europe and America achieved during the Industrial Revolution allowed these two regions to claim a share of global output that was two to three times greater than their share of the world’s population simply because their output per capita was two to three times greater than the global average.21 All signs are that this phase of divergence in per capita output is over and that we have embarked on a period of convergence.

other words, the rich countries are doubly wealthy: they both produce more at home and invest more abroad, so that their national income per head is greater than their output per head. The opposite is true for poor countries. More specifically,

But they have benefited far more from open markets for goods and services and advantageous terms of trade than from free capital flows.

To sum up, historical experience suggests that the principal mechanism for convergence at the international as well as the domestic level is the diffusion of knowledge. In other words, the poor catch up with the rich to the extent that they achieve the same level of technological know-how, skill, and education, not by becoming the property of the wealthy.

Beyond the central issue of convergence, however, the point I now want to stress is that the twenty-first century may see a return to a low-growth regime.

is important to decompose the growth of output into two terms: population growth and per capita output growth.

In fact, however, growth on the order of 1 percent a year in both population and per capita output, if continued over a very long period of time, as was the case after 1700, is extremely rapid, especially when compared with the virtually zero growth rate that we observe in the centuries prior to the Industrial Revolution.

The central thesis of this book is precisely that an apparently small gap between the return on capital and the rate of growth can in the long run have powerful and destabilizing effects on the structure and dynamics of social inequality. In a sense, everything follows from the laws of cumulative growth and cumulative returns, and that is why the reader will find it useful at this point to become familiar with these notions.

Conversely, a stagnant or, worse, decreasing population increases the influence of capital accumulated in previous generations. The same is true of economic stagnation. With low growth, moreover, it is fairly plausible that the rate of return on capital will be substantially higher than the growth rate, a situation that, as I noted in the introduction, is the main factor leading toward very substantial inequality in the distribution of wealth over the long run. Capital-dominated societies in the past, with hierarchies largely determined by inherited wealth (a category that includes both traditional rural societies and the countries of nineteenth-century Europe) can arise and subsist only in low-growth regimes. I will consider the extent to which the probable return to a low-growth regime, if it occurs, will affect the dynamics of capital accumulation and the structure of inequality. In particular, inherited wealth will make a comeback—a long-term phenomenon whose effects are already being felt in Europe and that could extend to other parts of the world as well. That is why it is important for present purposes to become familiar with the history of demographic and economic growth.

growth can thus increase social mobility for individuals whose parents did not belong to the elite of the previous generation. This increased social mobility need not imply decreased income inequality, but in theory it does limit the reproduction and amplification of inequalities of wealth and therefore over the long run also limits income inequality to a certain extent.

One should be wary, however, of the conventional wisdom that modern economic growth is a marvelous instrument for revealing individual talents and aptitudes. There is some truth in this view, but since the early nineteenth century it has all too often been used to justify inequalities of all sorts, no matter how great their magnitude and no matter what their real causes may be, while at the same time gracing the winners in the new industrial economy with every imaginable virtue. For instance, the liberal economist Charles Dunoyer, who served as a prefect under the July Monarchy, had this to say in his 1845 book De la liberté du travail (in which he of course expressed his opposition to any form of labor law or social legislation): “one consequence of the industrial regime is to destroy artificial inequalities, but this only highlights natural inequalities all the more clearly.” For Dunoyer, natural inequalities included differences in physical, intellectual, and moral capabilities, differences that were crucial to the new economy of growth and innovation that he saw wherever he looked. This was his reason for rejecting state intervention of any kind: “superior abilities … are the source of everything that is great and useful.… Reduce everything to equality and you will bring everything to a standstill.”8 One sometimes hears the same thought expressed today in the idea that the new information economy will allow the most talented individuals to increase their productivity m...

Basically, the eighteenth century suffered from the same economic stagnation as previous centuries. The nineteenth century witnessed the first sustained growth in per capita output, although large segments of the population derived little benefit from this, at least until the last three decades of the century. It was not until the twentieth century that economic growth became a tangible, unmistakable reality for everyone.

Finally, productivity growth in the service sector has generally been low (or even zero in some cases, which explains why this sector has tended to employ a steadily increasing share of the workforce), so that the price of services has increased more rapidly than the average of all prices.

In order to find our way through this vast aggregate of activities, whose growth accounts for much of the improvement in living conditions since the nineteenth century, it will be useful to distinguish several subsectors. Consider first services in health and education, which by themselves account for more than 20 percent of total employment in the most advanced countries (or as much as all industrial sectors combined). There is every reason to think that this fraction will continue to increase, given the pace of medical progress and the steady growth of higher education. The number of jobs in retail; hotels, cafés, and restaurants; and culture and leisure activities also increased rapidly, typically accounting for 20 percent of total employment. Services to firms (consulting, accounting, design, data processing, etc.) combined with real estate and financial services (real estate agencies, banks, insurance, etc.) and transportation add another 20 percent of the job total. If you then add government and security services (general administration, courts, police, armed forces, etc.), which account for nearly 10 percent of total employment in most countries, you reach the 70–80 percent figure given in official statistics.17

The consequence of this is that a country that privatized its health and education services would see its GDP rise artificially, even if the services produced and the wages paid to employees remained exactly the same.19 It may be that this method of accounting by costs underestimates the fundamental “value” of education and health and therefore the growth achieved during periods of rapid expansion of services in these areas.20

Before trying to answer this question, it is important to recall that past growth, as spectacular as it was, almost always occurred at relatively slow annual rates, generally no more than 1–1.5 percent per year. The only historical examples of noticeably more rapid growth—3–4 percent or more—occurred in countries that were experiencing accelerated catch-up with other countries. This is a process that by definition ends when catch-up is achieved and therefore can only be transitional and time limited. Clearly, moreover, such a catch-up process cannot take place globally.

At the global level, the average rate of growth of per capita output was 0.8 percent per year from 1700 to 2012, or 0.1 percent in the period 1700–1820, 0.9 percent in 1820–1913, and 1.6 percent in 1913–2012. As indicated in Table 2.1, we find the same average growth rate—0.8 percent—when we look at world population 1700–2012.

there is no historical example of a country at the world technological frontier whose growth in per capita output exceeded 1.5 percent over a lengthy period of time. If we look at the last few decades, we find even lower growth rates in the wealthiest countries: between 1990 and 2012, per capita output grew at a rate of 1.6 percent in Western Europe, 1.4 percent in North America, and 0.7 percent in Japan.21 It is important to bear this reality in mind as I proceed, because many people think that growth ought to be at least 3 or 4 percent per year. As noted, both history and logic show this to be illusory.

predict whether the actual rate will be 0.5 percent, 1 percent, or 1.5 percent. The median scenario I will present here is based on a long-term per capita output growth rate of 1.2 percent in the wealthy countries, which is relatively optimistic compared with Robert Gordon’s predictions

A society in which growth is 0.1–0.2 percent per year reproduces itself with little or no change from one generation to the next: the occupational structure is the same, as is the property structure. A society that grows at 1 percent per year, as the most advanced societies have done since the turn of the nineteenth century, is a society that undergoes deep and permanent change.

In North America, there is no nostalgia for the postwar period, quite simply because the Trente Glorieuses never existed there: per capita output grew at roughly the same rate of 1.5–2 percent per year throughout the period 1820–2012. To be sure, growth slowed a bit between 1930 and 1950 to just over 1.5 percent, then increased again to just over 2 percent between 1950 and 1970, and then slowed to less than 1.5 percent between 1990 and 2012. In Western Europe, which suffered much more from the two world wars, the variations are considerably greater: per capita output stagnated between 1913 and 1950 (with a growth rate of just over 0.5 percent) and then leapt ahead to more than 4 percent from 1950 to 1970, before falling sharply to just slightly above US levels (a little more than 2 percent) in the period 1970–1990 and to barely 1.5 percent between 1990 and 2012.

Great Britain and the United States, postwar history is interpreted quite differently. Between 1950 and 1980, the gap between the English-speaking countries and the countries that had lost the war closed rapidly. By the late 1970s, US magazine covers often denounced the decline of the United States and the success of German and Japanese industry. In Britain, GDP per capita fell below the level of Germany, France, Japan, and even Italy. It may even be the case that this sense of being rivaled (or even overtaken in the case of Britain) played an important part in the “conservative revolution.” Margaret Thatcher in Britain and Ronald Reagan in the United States promised to “roll back the welfare state” that had allegedly sapped the animal spirits of Anglo-Saxon entrepreneurs and thus to return to pure nineteenth-century capitalism, which would allow the United States and Britain to regain the upper hand. Even today, many people in both countries believe that the conservative revolution was remarkably successful, because their growth rates once again matched continental European and Japanese levels.

To back up a bit: the first crucial fact to bear in mind is that inflation is largely a twentieth-century phenomenon. Before that, up to World War I, inflation was zero or close to it. Prices sometimes rose or fell sharply for a period of several years or even decades, but these price movements generally balanced out in the end. This was the case in all countries for which we possess long-run price series.