Capital in the Twenty-First Century

by Thomas Piketty

  49. Contrary to an idea that is often taught but rarely verified, there is no evidence that executives in the period 1950–1980 made up for low pay with compensation in kind, such as private planes, sumptuous offices, etc. On the contrary, all the evidence suggests that such benefits in kind have increased since 1980.

  50. To be precise, 82 percent. See Piketty et al., “Optimal Taxation of Top Labor Incomes,” table 5.

  51. Note that the progressive tax plays two very distinct roles in this theoretical model (as well as in the history of progressive taxation): confiscatory rates (on the order of 80–90 percent on the top 0.5 or 1 percent of the distribution) would end indecent and useless compensation, while high but nonconfiscatory rates (of 50–60 percent on the top 5 or 10 percent) would raise revenues to finance the social state above the revenues coming from the bottom 90 percent of the distribution.

  52. See Jacob Hacker and Paul Pierson, Winner-Take-All Politics: How Washington Made the Rich Richer—And Turned its Back on the Middle Class (New York: Simon and Schuster, 2010); K. Schlozman, Sidney Verba, and H. Brady, The Unheavenly Chorus: Unequal Political Voice and the Broken Promise of American Democracy (Princeton: Princeton University Press, 2012); Timothy Noah, The Great Divergence (New York: Bloomsbury Press, 2012).

  53. See Claudia Goldin and Lawrence F. Katz, The Race between Education and Technology: The Evolution of U.S. Educational Wage Differentials, 1890–2005 (Cambridge, MA: Belknap Press and NBER, 2010), Rebecca M. Blank, Changing Inequality (Berkeley: University of California Press, 2011) and Raghuram G. Rajan, Fault Lines (Princeton: Princeton University Press, 2010).

  54. The pay of academic economists is driven up by the salaries offered in the private sector, especially the financial sector, for similar skills. See Chapter 8.

  55. For example, by using abstruse theoretical models designed to prove that the richest people should pay zero taxes or even receive subsidies. For a brief bibliography of such models, see the online technical appendix.

  15. A Global Tax on Capital

  1. The additional revenue could be used to reduce existing taxes or to pay for additional services (such as foreign aid or debt reduction; I will have more to say about this later).

  2. Every continent has specialized financial institutions that act as central repositories (custodian banks or clearing houses), whose purpose is to record ownership of various types of assets. But the function of these private institutions is to provide a service to the companies issuing the securities in question, not to record all the assets owned by a particular individual. On these institutions, see Gabriel Zucman, “The Missing Wealth of Nations: Are Europe and the U.S. Net Debtors or Net Creditors?” Quarterly Journal of Economics 128, no. 3 (2013): 1321–64.

  3. For instance, the fall of the Roman Empire ended the imperial tax on land and therefore the land titles and cadastre that went with it. According to Peter Temin, this contributed to economic chaos in the early Middle Ages. See Peter Temin, The Roman Market Economy (Princeton: Princeton University Press, 2012), 149–51.

  4. For this reason, it would be useful to institute a low-rate tax on net corporate capital together with a higher-rate tax on private wealth. Governments would then be forced to set accounting standards, a task currently left to associations of private accountants. On this subject, see Nicolas Véron, Matthieu Autrer, and Alfred Galichon, L’information financière en crise: Comptabilité et capitalisme (Paris: Odile Jacob, 2004).

  5. Concretely, the authorities do what is called a “hedonic” regression to calculate the market price as a function of various characteristics of the property. Transactional data are available in all developed countries for this purpose (and are used to calculate real estate price indices).

  6. This temptation is a problem in all systems based on self-reporting by taxpayers, such as the wealth tax system in France, where there is always an abnormally large number of reports of wealth just slightly below the taxable threshold. There is clearly a tendency to slightly understate the value of real estate, typically by 10 or 20 percent. A precomputed statement issued by the government would provide an objective figure based on public data and a clear methodology and would thus put an end to such behavior.

  7. Oddly enough, the French government once again turned to this archaic method in 2013 to obtain information about the assets of its own ministers, officially for the purpose of restoring confidence after one of them was caught in a lie about evading taxes on his wealth.

  8. For example, the Channel Islands, Liechtenstein, Monaco, etc.

  9. It is difficult to estimate the extent of such losses, but in a country like Luxembourg or Switzerland they might amount to as much as 10–20 percent of national income, which would have a substantial impact on their standard of living. (The same is true of a financial enclave like the City of London.) In the more exotic tax havens and microstates, the loss might be as high as 50 percent or more of national income, indeed as high as 80–90 percent in territories that function solely as domiciles for fictitious corporations.

  10. Social insurance contributions are a type of income tax (and are included in the income tax in some countries; see Chapter 13).

  11. See in particular Table 12.1.

  12. Recall the classic definition of income in the economic sense, given by the British economist John Hicks: “The income of a person or collectivity is the value of the maximum that could be consumed during the period while remaining as wealthy at the end of the period as at the beginning.”

  13. Even with a return on capital of 2 percent (much lower than the actual return on the Bettencourt fortune in the period 1987–2013), the economic income on 30 billion euros would amount to 600 million euros, not 5 million.

  14. In the case of the Bettencourt fortune, the largest in France, there was an additional problem: the family trust was managed by the wife of the minister of the budget, who was also the treasurer of a political party that had received large donations from Bettencourt. Since the same party had reduced the wealth tax by two-thirds during its time in power, the story naturally stirred up a considerable reaction in France. The United States is not the only country where the wealthy wield considerable political influence, as I showed in the previous chapter. Note, too, that the minister of the budget in question was succeeded by another who had to resign when it was revealed that he had a secret bank account in Switzerland. In France, too, the political influence of the wealthy transcends political boundaries.

  15. In practice, the Dutch system is not completely satisfactory: many categories of assets are exempt (particularly those held in family trusts), and the assumed return is 4 percent for all assets, which may be too high for some fortunes and too low for others.

  16. The most logical approach is to measure this insufficiency on the basis of average rates of return observed for fortunes of each category so as to make the income tax schedule consistent with the capital tax schedule. One might also consider minimum and maximum taxes as a function of capital income. See the online technical appendix.

  17. The incentive argument is central to Maurice Allais’s tendentious L’impôt sur le capital et la réforme monétaire (Paris: Editions Hermann, 1977), in which Allais went so far as to advocate complete elimination of the income tax and all other taxes in favor of a tax on capital. This is an extravagant idea and not very sensible, given the amounts of money involved. On Allais’s argument and current extensions of it, see the online technical appendix. Broadly speaking, discussions of a tax on capital often push people into extreme positions (so that they either reject the idea out of hand or embrace it as the one and only tax, destined to replace all others). The same is true of the estate tax (either they shouldn’t be taxed at all or should be taxed at 100 percent). In my view, it is urgent to lower the temperature of the debate and give each argument and each type of tax its due. A capital tax is useful, but it cannot replace all other taxes.

  18. The same is true of an unemployed worker who has to continue paying a hi
gh property tax (especially when mortgage payments are not deductible). The consequences for overindebted households can be dramatic.

  19. This compromise depends on the respective importance of individual incentives and random shocks in determining the return on capital. In some cases it may be preferable to tax capital income less heavily than labor income (and to rely primarily on a tax on the capital stock), while in others it might make sense to tax capital income more heavily (as was the case in Britain and the United States before 1980, no doubt because capital income was seen as particularly arbitrary). See Thomas Piketty and Emmanuel Saez, A Theory of Optimal Capital Taxation, NBER Working Paper 17989 (April 2012); a shorter version is available as “A Theory of Optimal Inheritance Taxation,” Econometrica 81, no. 5 (September 2013): 1851–86.

  20. This is because the capitalized value of the inheritance over the lifetime of the recipient is not known at the moment of transmission. When a Paris apartment worth 100,000 francs in 1972 passed to an heir, no one knew that the property would be worth a million euros in 2013 and afford a saving on rent of more than 40,000 euros a year. Rather than tax the inheritance heavily in 1972, it is more efficient to assess a smaller inheritance tax but to requirement payment of an annual property tax, a tax on rent, and perhaps a wealth tax as the value of the property and its return increase over time.

  21. See Piketty and Saez, “Theory of Optimal Capital Taxation”; see also the online technical appendix.

  22. See Figure 14.2

  23. For example, on real estate worth 500,000 euros, the annual tax would be between 2,500 and 5,000 euros, and the rental value of the property would be about 20,000 euros a year. By construction, a 4–5 percent annual tax on all capital would consume nearly all of capital’s share of national income, which seems neither just nor realistic, particularly since there are already taxes on capital income.

  24. About 2.5 percent of the adult population of Europe possessed fortunes above 1 million euros in 2013, and about 0.2 percent above 5 million. The annual revenue from the proposed tax would be about 300 billion euros on a GDP of nearly 15 trillion. See the online technical appendix and Supplemental Table S5.1, available online, for a detailed estimate and a simple simulator with which one can estimate the number of taxpayers and the amount of revenue associated with other possible tax schedules.

  25. The top centile currently owns about 25 percent of total wealth, or about 125 percent of European GDP. The wealthiest 2.5 percent own nearly 40 percent of total wealth, or about 200 percent of European GDP. Hence it is no surprise that a tax with marginal rates of 1 or 2 percent would bring in about two points of GDP. Revenues would be even higher if these rates applied to all wealth and not just to the fractions over the thresholds.

  26. The French wealth tax, called the “solidarity tax on wealth,” (impôt de solidarité sur la fortune, or ISF), applies today to taxable wealth above 1.3 million euros (after a deduction of 30 percent on the primary residence), with rates ranging from 0.7 to 1.5 percent on the highest bracket (over 10 million euros). Allowing for deductions and exemptions, the tax generates revenues worth less than 0.5 percent of GDP. In theory, an asset is called a business asset if the owner is active in the associated business. In practice, this condition is rather vague and easily circumvented, especially since additional exemptions have been added over the years (such as “stockholder agreements,” which allow for partial or total exemptions if a group of stockholders agrees to maintain its investment for a certain period of time). According to the available data, the wealthiest individuals in France largely avoid paying the wealth tax. The tax authorities publish very few detailed statistics for each tax bracket (much fewer, for example, than in the case of the inheritance tax from the early twentieth century to the 1950s); this makes the whole operation even more opaque. See the online technical appendix.

  27. See esp. Chapter 5, Figures 5.4 and following.

  28. The progressive capital tax would then bring in 3–4 percent of GDP, of which 1 or 2 points would come from the property tax replacement. See the online technical appendix.

  29. For example, to justify the recent decrease of the top wealth tax rate in France from 1.8 to 1.5 percent.

  30. See P. Judet de la Combe, “Le jour où Solon a aboli la dette des Athéniens,” Libération, May 31, 2010.

  31. In fact, as I have shown, capital in the form of land included improvements to the land, increasingly so over the years, so that in the long run landed capital was not very different from other forms of accumulable capital. Still, accumulation of landed capital was subject to certain natural limits, and its predominance implied that the economy could only grow very slowly.

  32. This does not mean that other “stakeholders” (including workers, collectivities, associations, etc.) should be denied the means to influence investment decisions by granting them appropriate voting rights. Here, financial transparency can play a key role. I come back to this in the next chapter.

  33. The optimal rate of the capital tax will of course depend on the gap between the return on capital, r, and the growth rate, g, with an eye to limiting the effect of r > g. For example, under certain hypotheses, the optimal inheritance tax rate is given by the formula t = 1 − G/R, where G is the generational growth rate and R the generational return on capital (so that the tax approaches 100 percent when growth is extremely small relative to return on capital, and approaches 0 percent when the growth rate is close to the return on capital). In general, however, things are more complex, because the ideal system requires a progressive annual tax on capital. The principal optimal tax formulas are presented and explained in the online technical appendix (but only in order to clarify the terms of debate, not to provide ready-made solutions, since many forces are at work and it is difficult to evaluate the effect of each with any precision).

  34. Thomas Paine, in his pamphlet Agrarian Justice (1795), proposed a 10 percent inheritance tax (which in his view corresponded to the “unaccumulated” portion of the estate, whereas the “accumulated” portion was not to be taxed at all, even if it dated back several generations). Certain “national heredity tax” proposals during the French Revolution were more radical. After much debate, however, the tax on direct line transmissions was set at no more than 2 percent. On these debates and proposals, see the online technical appendix.

  35. Despite much discussion and numerous proposals in the United States and Britain, especially in the 1960s and again in the early 2000s. See the online technical appendix.

  36. This design flaw stemmed from the fact that these capital taxes originated in the nineteenth century, when inflation was insignificant or nonexistent and it was deemed sufficient to reassess asset values every ten or fifteen years (for real estate) or to base values on actual transactions (which was often done for financial assets). This system of assessment was profoundly disrupted by the inflation of 1914–1945 and was never made to work properly in a world of substantial permanent inflation.

  37. On the history of the German capital tax, from its creation in Prussia to its suspension in 1997 (the law was not formally repealed), see Fabien Dell, L’Allemagne inégale, PhD diss., Paris School of Economics, 2008. On the Swedish capital tax, created in 1947 (but which actually existed as a supplementary tax on capital income since the 1910s) and abolished in 2007, see the previously cited work of Ohlsson and Waldenström and the references given in the appendix. The rates of these taxes generally remained under 1.5–2 percent on the largest fortunes, with a peak in Sweden of 4 percent in 1983 (which applied only to assessed values largely unrelated to market values). Apart from the degeneration of the tax base, which also affected the estate tax in both countries, the perception of fiscal competition also played a role in Sweden, where the estate tax was abolished in 2005. This episode, at odds with Sweden’s egalitarian values, is a good example of the growing inability of smaller countries to maintain an independent fiscal policy.

  38. The wealth tax (on large fortunes) was introduced in Fran
ce in 1981, abolished in 1986, and then reintroduced in 1988 as the “solidarity tax on wealth.” Market values can change abruptly, and this can seem to introduce an element of arbitrariness into the wealth tax, but they are the only objective and universally acceptable basis for such a tax. Nevertheless, rates and tax brackets must be adjusted regularly, and care must be taken not to allow receipts to rise automatically with real estate prices, for this can provoke tax revolts, as illustrated by the famous Proposition 13 adopted in California in 1978 to limit rising property taxes.

  39. The Spanish tax is assessed on fortunes greater than 700,000 euros in taxable assets (with a deduction of 300,000 euros for the principal residence), and the highest rate is 2.5 percent (2.75 percent in Catalonia). There is also an annual capital tax in Switzerland, with relatively low rates (less than 1 percent) due to competition among cantons.

  40. Or to prevent a foreign competitor from developing (the destruction of the nascent Indian textile industry by the British colonizer in the early nineteenth century is etched into the memory of Indians). This can have lasting consequences.

  41. This is all the more astonishing given that the rare estimates of the economic gains due to financial integration suggest a rather modest global gain (without even allowing for the negative effects on inequality and instability, which these studies ignore). See Pierre-Olivier Gourinchas and Olivier Jeanne, “The Elusive Gains from International Financial Integration,” Review of Economic Studies 73, no. 3 (2006): 715–41. Note that the IMF’s position on automatic transmission of information has been vague and variable: the principle is approved, the better to torpedo its concrete application on the basis of rather unconvincing technical arguments.