This article is a review and critique of the new book by Thomas Piketty, "Capital in the Twenty-First Century." Piketty, a professor at the Paris School of Economics, reviews data on income and wealth inequality in developed countries over the past hundred years or so. He makes bold projections that apparent recent trends of increasing inequality will continue and deepen. Based on his interpretation of the data, Piketty gives strong prescriptions to substantially increase marginal tax rates on income to institute a global tax on capital.
This article is a review and critique of the new book by Thomas Piketty, "Capital in the Twenty-First Century." Piketty, a professor at the Paris School of Economics, reviews data on income and wealth inequality in developed countries over the past hundred years or so. He makes bold projections that apparent recent trends of increasing inequality will continue and deepen. Based on his interpretation of the data, Piketty gives strong prescriptions to substantially increase marginal tax rates on income to institute a global tax on capital.
This article is a review and critique of the new book by Thomas Piketty, "Capital in the Twenty-First Century." Piketty, a professor at the Paris School of Economics, reviews data on income and wealth inequality in developed countries over the past hundred years or so. He makes bold projections that apparent recent trends of increasing inequality will continue and deepen. Based on his interpretation of the data, Piketty gives strong prescriptions to substantially increase marginal tax rates on income to institute a global tax on capital.
By Mark J. Warshawsky This article is a review and critique of the new book by Thomas Piketty, Capital in the Twenty-First Century. Piketty, a professor at the Paris School of Economics, reviews data on income and wealth inequality in developed countries over the past hundred years or so. He makes bold projections that apparent recent trends of increasing inequality will continue and deepen. Based on his interpretation of the data, Piketty gives strong prescriptions to sub- stantially increase marginal tax rates on income and to institute a global tax on capital. The books political significance is high for Tax Notes readers because the Obama administration early on strongly endorsed Pikettys claim with University of California, Berkeley professor Em- manuel Saez that U.S. income and wage inequality has grown significantly over the last 30 years. The administration highlighted Pikettys findings in its first proposed budget, presented in 2009, tying major parts of President Obamas domestic policy agenda to the research. More recently, Obama has stated that inequality is the single most important policy issue in the United States, the defining challenge of our time. The book will also surely resonate in Europe and among international eco- nomic organizations. The books intellectual signifi- cance is high for Tax Notes readers because the statistics reported are based mainly on historical and recent tax records for France, Great Britain, and the United States (and to lesser extents, Germany, Sweden, and other countries). Economic Theory Piketty organizes his analysis around two simple equations that he calls fundamental laws of capital- ism. The first is an accounting definition the share of capital in national income equals the prod- uct of the return on capital and the capital/income ratio. While tautological, the equation is nonethe- less informative because it expresses an important relationship among key variables, each of which can be measured and explained, sometimes inde- pendently and often by various data sources. For example, if the capital/income ratio is 600 percent and the return is 5 percent, the share of capital in national income is 30 percent. Capital is defined and measured as all forms of real property (includ- ing housing) and financial and professional capital (plants, infrastructure, machinery, inventory, pat- ents, and so on) used by companies and govern- ment, all of which can be owned and exchanged, on some market. Thus, capital is largely measured at market prices. The second equation, or fundamental law of capitalism, is that the capital/income ratio is equal in the long run to the savings rate divided by the economic growth rate in inflation-adjusted terms. For example, if the savings rate is 10 percent and the growth rate is 2 percent, in the long run the capital/ income ratio must be 500 percent. While these equations are elementary concepts in the theories of economic growth and development, their relevance to the study of inequality is that the ownership of capital is often concentrated among a relatively small group of the population. Hence, the study of the path of capital is considered essential to the study of inequality. Moreover, labor income can be unequally distributed as well. Finally and these are key points Piketty believes that the return on capital has held fairly steady over time and will continue to do so, while the rate of economic growth declines as the population (that is, labor force) stops increasing and even decreases in many European and Asian countries. Piketty also thinks that the savings rate is fairly steady, regard- less of changes in economic conditions, because it is mainly influenced by the desire of the rich to leave bequests to their children. As we will see, these beliefs lead to a strong prediction of an increasing role for capital in the future, and therefore more inequality arising from bequests, which Piketty views negatively. Capital Ratios and Income Factor Shares Measuring the capital/income ratio over three centuries, Piketty finds that through 1910, the ratio in both Great Britain and France was steady at Mark J. Warshawsky was formerly Treasury assistant secretary of economic policy and is a visiting scholar at the Mercatus Center at George Mason University. Thomas Piketty, Capital in the Twenty-First Cen- tury (Harvard University Press, 2014), ASIN: B00I2WNYJW (Hardback, $40). tax notes
POLICY AND PRACTICE
TAX NOTES, June 30, 2014 1547 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . about 700 percent and then plummeted in the aftermath of World War I to about 300 percent. The capital/income ratio remained at that low level, even declining a bit, until 1950, when it began a climb, reaching 500 percent in Britain and 600 percent in France by 2010. Looking at component parts, one finds that at least part of the plunge in 1920 resulted from the expenditure of significant net foreign capital to pay for the war, and at least part of the recent rise is the result of substantial increases in housing. In Germany, the trajectory is largely the same except that the fall after World War I continued through 1950 as physical capital was destroyed in World War II, so that the capital/ income ratio was only about 200 percent then. The later increase is significant but to a lower level, just more than 400 percent in 2010, again largely be- cause of housing assets. By contrast, for the United States, the capital/ income ratio increased steadily from about 300 percent in 1770 to 500 percent in 1910, falling only slightly after World War I and increasing again through 1930 with the stock market boom. The ratio fell to below 400 percent by 1950 and, thereafter, increased only slightly through 2010 to about 430 percent. Interestingly, the run-up in housing prices, although present in the United States, is much less pronounced than in France, Piketty found, using decade average statistics, although the short-term drop for the United States from 550 percent in 2007 to 430 percent in 2010 surely reflects the volatility of both the housing and stock markets. Income growth, partly the result of rapid population growth from immigration, is also higher in the United States than in Europe, which increases the denominator and therefore lowers the long-run capital/income ratio. Piketty collects similar data for other countries Canada, Japan, Australia, and Italy although generally for shorter periods. He finds a massive run-up of Japans capital/income ratio, from about 360 percent in 1970 to 800 percent in 1990, followed by a rapid decline to 700 percent after the asset bubble popped in the country, and a further de- crease to about 600 percent in 2010. For Italy, he sees a steady and rapid increase from just about 250 percent to 600 percent over that same period. Piketty explains the overall increases in those coun- tries ratios by the second fundamental law of capitalism: low economic growth and high savings rates. From this and other scattered data, Piketty makes some truly bold assumptions and takes two gigan- tic leaps. He creates a world capital/income ratio from 1870 to 2010 and then projects that ratio through 2100. While some would view that exercise as almost a work of fiction, Piketty is serious about the results. He says that the world ratio was 500 percent in 1910, dropped to 260 percent in 1950, and then increased to about 440 percent by 2010. There- after, it is projected to continue to increase to 600 percent by 2060 and to 670 percent by 2100. One must credit Piketty for taking a global view because capital markets have indeed become open and linked in most countries. The simplicity of using the second equation and assuming in the long run an average world savings rate of 10 percent and an economic growth rate of 1.5 percent to project the global capital/income ratio is breathtaking, how- ever. Piketty next moves from the capital/income ratio to the share of capital income in total national income, using the first fundamental law of capital- ism. He shows that there has been an overall downward trend for Britain and France, from about 40 percent in the 19th century to about 20 to 25 percent or even less in most of the 20th century, increasing recently to about 25 percent or a bit higher. Piketty attributes most of this trend to the changes in the capital/income ratio over time, but he allows for some changes in the rate of return as well, with return increases in the mid-20th century and declines most recently. Looking over a shorter, more recent period, Piketty finds that the capital share in the United States increased from 21 percent in 1975 to 29 percent in 2010, with considerable volatility in between, apparently related to the stock market. With the exception of Canada, the other developed countries saw similar share increases for capital over that period. Despite recent lows in interest rates, Piketty says that the total rate of return on capital, averaging across risk types, is still and generally will be about 4 to 5 percent in inflation-adjusted terms. These rates have changed little from the rates of return on agricultural land and government bonds implicit in Jane Austens depiction of Mr. Darcys estate in- come or in Honor de Balzacs description of the dowries of Pre Goriots daughters in the 1810s. So, according to Piketty, while rates of return may fall somewhat as capital increases, most of the projected increase in the capital/income ratio will flow through to the capital share of income. In more technical terms, this indicates that the elasticity of substitution between capital and labor exceeds 1, a controversial claim. Piketty concludes this section of the book by projecting that with a capital/income ratio of 700 to 800 percent and a rate of return of 4 to 5 percent, capitals share in national income will increase to 30 to 40 percent levels close to those of the in- egalitarian inheritance-influenced days of Austen and Balzac. Again, these projection calculations are extremely rough and, indeed, seem exaggerated, COMMENTARY / POLICY AND PRACTICE 1548 TAX NOTES, June 30, 2014 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . but Piketty uses them to advance his central argu- ment that income inequality arising from the in- creasing role of capital in an environment of economic stagnation is growing and will continue to do so, with little in the way of natural checks except for government intervention. Inequality The main way Piketty measures inequality, whether of income, labor earnings, or capital own- ership, is to calculate the share of various top percentiles of the population in the quantity in question (income, etc.) for various countries and periods. For example, he reports that the top per- centile in Scandinavia in the 1970s and 1980s got 5 percent of total labor income, while the next 9 percent got 15 percent. For the same region and period, capital ownership was more concentrated, with the top percentile owning 20 percent and the next 9 percent owning 30 percent. These data are based mostly on annual observations, either from tax records or surveys, and there is no assurance that they represent the same people or families over long periods (even generations), particularly if there is a lot of mobility and volatility in the society and economy. Nonetheless, Piketty calls the top 1 percent the dominant class, the next 9 percent the well-to-do class, the middle 40 percent the middle class, and the bottom 50 percent the lower class. These clearly are arbitrary categories that may or may not correspond to recognizable social and political groupings, such as British nobility in the 1820s. As is usual in Capital, Piketty starts with and concentrates his review of data in France. The upper deciles share of national income decreased from 40 to 50 percent in the 1910s to mid-1930s to 30 to 35 percent today. Almost the entire drop occurred just before and during World War II. By contrast, the wage share has been fairly flat over the entire century, at about 25 percent of total labor income. The collapse in income for the top percentile started earlier, after World War I, and was more dramatic from 20 percent in 1910 to 8 percent by 1945 and increasing slightly to 9 percent by 2010 while top wage shares are remarkably stable, at 6 percent over 100 years. Piketty also shows that the share of income of the top 0.5 percent coming from capital and labor inverted between 1932 and 2005. He calls these trends the fall of the rentier and the rise of a society of managers. Note, however, that those data exclude capital gains. Piketty then presents comparable data for the United States. The share of the top decile in total income was about 40 percent in the 1910s, it in- creased to 45 percent in the 1920s-1930s, plummeted to just above 30 percent during the years of World War II, remained at that level through 1980, and climbed back to 45 percent by 2010. If capital gains are included, the levels are somewhat higher on average and far more volatile. Piketty attributes much of the plunge in income inequality during World War II to the activity of the federal govern- ment in restricting wage increases. Given his ulti- mate focus on tax policy, it is surprising that he ignores the potentially more significant changes in tax law and administration during the war. Piketty provides data showing that most of the increase in income inequality in the United States from the mid-1980s forward is attributable to the top percentile, some of it to the top 1 to 5 percent and almost nothing to the top 5 to 10 percent. Indeed, in 1986 the share of the top percentile in total income was 9 percent. It jumped to 13 percent in 1988 after passage of the Tax Reform Act of 1986, followed the path of asset markets up and down thereafter, and by 2010 increased to 17 percent. Further, Piketty finds that regarding labor income data, about two-thirds of the increase in income inequality is attributable to the rise of wage inequal- ity, which he attributes to the advent of superman- agers. Note that wages include all bonuses and stock options. Looking at other countries, Piketty finds that increases in income inequality are generally similar to those in the United States but smaller. In the United Kingdom, the top percentile share increased from 7 percent in 1981 to 15 percent in 2010. Over the same period, the share increased from 5 percent to 9 percent in Australia, 8 percent to 12 percent in Canada, 4 percent to 7 percent in Sweden, 7 percent to 10 percent in Japan, and 9 percent to 11 percent in Germany. Note that despite a large increase in the capital/ income ratio in France, the increase in income inequality there was the smallest among the devel- oped countries surveyed. By contrast, the increases in reported income inequality were large in the United States and the United Kingdom, while the capital/income ratio remained flat in the United States and increased in the United Kingdom, but to a lesser extent than in France. Those observations are inconsistent with Pikettys central point that we should be concerned about the growth of capital and tax it heavily because of the dire implications of income inequality. Piketty then turns to the inequality of capital ownership. He finds that in France the top decile owned 90 percent of capital in 1910 but that that share dropped steadily to 60 percent in 1970 and thereafter remained constant. According to Piketty, a similar pattern and levels may be found for the United Kingdom and Sweden, although there were small increases after 1980. For the United States, the height in the ownership of capital by the top decile COMMENTARY / POLICY AND PRACTICE TAX NOTES, June 30, 2014 1549 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . was reached at only 80 percent in 1910, fell to 65 percent by 1970, and increased slightly to 70 percent by 2010. At face value, these statistics connection to Pikettys central hypothesis is unclear the con- centration of capital ownership has remained flat or increased slightly, while changes in capital/income ratios and income inequality have diverged across countries. Piketty cites progressive capital taxation in explaining why the concentration of capital own- ership has not increased more, but he does not analyze this view deeply, for example, by examin- ing inter-county differences in tax policy. Chris Giles, an editor at the Financial Times, recently found some errors and questionable judg- ments in Pikettys calculations of the concentration of capital ownership. Giles makes a case that the concentration of capital in the United Kingdom has actually declined in the last three decades and that it has remained flat (at a lower level than Piketty reports) in the United States. Piketty has given a response. In my opinion, Giless view is a more reasonable read of the available data. The difference of opinion about these key results further under- mines the support for Pikettys central hypothesis of a positive and causal relationship between capi- tals prominence and income inequality recently and into the future. Finally, relying mainly on French data, Piketty looks at the role of inheritance versus savings in the accumulation of private wealth. For France, he finds that the annual inheritance flow was about 20 to 25 percent of national income during the 19th century and until 1914; it then fell to less than 5 percent in the 1950s and returned to about 15 percent in 2010. (A similar trend is apparent in German data.) As- suming that the rate of return on capital exceeds the rate of economic growth, which is asserted repeat- edly in the book, Piketty projects that inheritance flows will continue to increase, to perhaps as high as 23 percent of income in 2100. This would imply that more than 90 percent of wealth in France will be inherited, up from 70 percent currently. Because, according to Piketty, rentiers are the enemies of democracy, this would a bad outcome for society. It may seem strange to American eyes and ears that bequests and not retirement savings, which are generally used up in a workers lifetime, are cred- ited with the creation of most wealth. Historically, of course, retirement from work is a fairly recent social creation, coincident with the increase in life expectancies and income in the years before and following World War II. Still, it is surprising that Piketty does not give retirement saving a greater role in explaining recent trends and projecting the future. Recall though that in France and Germany, pay as you go public retirement income and health programs are generous and provide the resources for most workers to have long and com- fortable retirements. So retirement asset savings will be crowded out in those countries and account for much less than in the United States, the United Kingdom, Canada, and other countries where pri- vately (and sometimes publicly) funded retirement plans and accounts are widespread. Figure 1 shows retirement-related assets (such as employer-provided pension entitlements, annuities, and IRAs) as a share of household net worth in the United States over the 1981-2013 period. Retirement-related assets are about a third of house- hold net worth. Clearly, retirement saving is signifi- cant even by this narrow measure. Moreover, even housing functions as a type of retirement saving in these Anglo-Saxon countries without complete social insurance programs for care needs at the end of life. Many households in the middle class and above use housing equity to pay for home healthcare and nursing home care, which are expensive, particularly for those with Alzheimers and other chronic diseases that can incapacitate an individual for many years. More broadly, households use other types of nonspecial- ized assets mutual funds, deposits, and so on to finance retirement spending. A rough estimate would point to the retirement savings motivation as explaining at least half of capital accumulated in the United States, much more weight than Piketty gives to the life cycle hypothesis of saving, which detracts from the bequest motive. To an American audience, Piketty also severely understates the role of entrepreneurs and overstates that of inheritance in the creation of wealth. For example, in the United States, megabillionaires like Bill Gates and Warren Buffett did not inherit their assets and are setting up charitable foundations to receive them for the benefit of future generations. Piketty may include the entrepreneurial effect in a high general return on capital, but that approach ignores the truly unique and personal catalyzing and organizing contribution of the individuals in- volved in establishing new businesses, raising capi- tal, hiring workers, and creating and marketing new technologies. Pikettys Prescriptions for Tax Policy Without taxes, society has no common destiny, and collective action is impossible (Capital, at 493). With these stirring, but somewhat debatable, words, Piketty gives his recommendations for tax policy. He begins with a historical review of the ratio of tax revenues to national income. For the countries examined (France, Sweden, Britain, and the United States), the tax ratio was flat at about 10 percent from 1870 through 1910 and was used to fund regalian functions that is, the police, courts, roads, army, and so on. From 1910 through COMMENTARY / POLICY AND PRACTICE 1550 TAX NOTES, June 30, 2014 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . 1950, the tax ratio increased to 25 to 35 percent as social spending rose rapidly, for example, on public pensions and education. It continued to increase through 1980 from 30 to 55 percent as healthcare was added to social insurance but flattened out thereafter to 30 percent in the United States, 40 percent in Britain, 50 percent in France, and 55 percent in Sweden. In Pikettys view, European levels of revenue probably represent close to the upper limit because of a public reluctance to pay higher taxes when income growth is slowing and attributable to the productive inefficiency of the public sector. Still, to counteract the increase in inequality of income and capital ownership that Piketty sees and foresees, he recommends an in- crease in the progressivity of taxation on income, on inheritances, and on capital directly. Piketty reminds us that the income tax generally applies to both labor income and at least some capital income (dividends, interest, rents, and so on). Taxes on capital also include any levy on the flow of income from capital, such as the corporate income tax, as well as any tax on the capital stock itself such as property, inheritance, and direct wealth taxes. In most countries, social insurance taxes are placed just on labor income and are often dedicated only to the social insurance purpose, sometimes through an earned benefit formula for the individual. Taxes can be proportional, regres- sive, or progressive. Piketty claims that the progressivity of income and estate taxation has a direct positive effect, reducing the inequality of income and capital own- ership. He says that the spectacular decrease in the progressivity of the income tax in the U.S. and Britain since 1980 . . . probably explains much of the increase in the very highest earned incomes (Capi- tal, at 495). 1 Piketty is also concerned that tax 1 Piketty has to be arguing loosely here because the top inheritance tax rate in France has been consistently and signifi- cantly below that in the United States from 1980 through 2010, even as the concentration of capital ownership increased in the United States and remained flat in France, according to his calculations. Also, the U.S. top income tax rate increased from 28 percent in 1988 to nearly 40 percent through 2000, even as the share of reported income (including capital gains) of the top percentile rose from 16 percent to 22 percent over that period. Figure 1. Employer-Provided Retirement Resources, Annuities, and IRAs As a Share of Household Net Worth, United States, 1982-2013 Source: http://www.federalreserve.gov/datadownload/Choose.aspx?rel=z1. Years 1982 2013 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 COMMENTARY / POLICY AND PRACTICE TAX NOTES, June 30, 2014 1551 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . competition between countries in Europe has led to cuts in corporate tax rates and to the exemption of capital income from the progressive income tax. Piketty reviews the history of top income and inheritance tax rates. He shows that there were substantial increases for all countries during and following World War I and that in the United States and Britain in particular, top rates were again increased dramatically during the Depression and World War II, remaining at high rates through 1980 when they were cut. Piketty says that Germany and France used means other than tax policy to control capitalism after World War II, including public ownership of companies and the direct setting of executive salaries. For the United States and Britain since 1980, he is critical of the practice of corporate governance that has led to an explosion of executive salaries, which he attributes to the cutting of the top income tax rate. Piketty says that lower marginal tax rates encourage executives to negotiate harder for higher pay. He does not explain, however, why the owners of corporate capital, who are also facing lower marginal tax rates, would not also put up a tougher fight against executive demands in re- sponse. Based on these interpretations, arguments, and considerations, Piketty recommends the following tax policy, in particular for the United States. The top income tax rate should be 80 percent, levied on incomes exceeding $500,000 or $1 million. Also, to develop the meager U.S. social state and invest more in health and education (Capital, at 513), tax rates of 50 percent or 60 percent should be imposed on incomes above $200,000. 2 Piketty acknowledges the difficulty of making these changes, which he attributes to the political process being captured by the 1 percent and to a drift to oligarchy. He says there is little reason for optimism about where the U.S. is headed (Capital, at 514). Pikettys boldest recommendation is a progres- sive global tax on capital coupled with a high level of international financial transparency to enable collection of the tax. This annual tax would be in addition to current income tax, social insurance tax, inheritance tax, and other capital taxes. As men- tioned above, Piketty is motivated by what he sees as harmful tax competition among European coun- tries in what he views as an endless inegalitarian spiral and a lack of transparency, leading to the widespread use of illegal tax shelters. He also believes that for the wealthy, capital rather than measured income is the best way to assess contribu- tive capacity. His proposal would establish a tax schedule applicable to all wealth around the world. The resulting revenues would somehow be apportioned among and within countries. Piketty suggests the following annual rate schedule: 0.1 percent for net assets below 200,000; 0.5 percent for between 200,000 and 1 million; 1 percent for between 1 million and 5 million; 2 percent for between 5 million and 1 billion; and 5 to 10 percent for above 1 billion. All types of assets would be included, at market value. Part of the motivation for the higher tax rate on the wealthy is his view that the wealthy get a much higher real return on capital: 6 to 7 percent. That estimate is rough, based on data from the Forbes survey of the worlds billionaires, and does not control for exposure to risk. Piketty places a high premium on transparency to increase our knowledge of capital ownership inequality; to improve financial regulation, espe- cially in handling banking crises; and to force governments to broaden international agreements on the automatic sharing of financial data. With broader agreements, the national tax authorities would receive all the information needed to accu- rately compute the net worth of every citizen including information from countries like the Cay- man Islands and Switzerland. In this regard, Piketty is complimentary of the Foreign Account Tax Com- pliance Act and critical of various weaker EU directives. Public transparency would be required of corporations and individuals in situations where there is no other way to establish trust (Capital, at 570). How that approach would be consistent with privacy and even the safety of the individuals and their families is unclear. To reduce public debt, Piketty suggests a pro- gressive, one-time tax on private capital or inflation. Economic Critiques In summarizing and reviewing Capital above, I have already begun to critique some aspects of the book. There are, however, two major areas that warrant further critical comment: the rate of return on capital (and the claimed inference of a rapid growth of capital) and U.S. income (wage) inequal- ity. Piketty underplays the fact that in the United States, short-term interest rates have been zero for several years, long-term interest rates on nominal government bonds are less than 3 percent, and rates on inflation-indexed securities have been negative. These are the rates available to all income and wealth groups, not just the lower ones. According to universally accepted finance theory, rates on those low-risk securities serve as the base for the expected 2 It is hard to imagine why the United States would want to spend more on health, when it already devotes a far larger share of its income to it than all other developed countries do. COMMENTARY / POLICY AND PRACTICE 1552 TAX NOTES, June 30, 2014 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . rates of return on other, riskier types of capital so that as the rates decline, so do all other rates of return. As further evidence of current low rates of re- turn, see the data on yields on long-term British government bonds in figures 2 and 3. Figure 2 shows yields on consols and inflation rates over a long historical period. We indeed see the 5 percent yield in the 1810s experienced by Mr. Darcy, al- though contrary to Pikettys assertion, inflation was relevant even then, at least during war times. But in any event, the yield is almost always lower than 5 percent historically and certainly is lower in real terms, and especially in 2013, when due consider- ation is given to trends in inflation. Figure 3 is a direct measure of the real (that is, inflation- adjusted) low-risk rate because it shows the yield on inflation-indexed, long-maturity U.K. govern- ment bonds. In recent periods, the yield is negative and has generally been around 1 percent. Therefore, Pikettys assertion of a fairly constant 4 to 5 percent real rate of return on capital flies in the face of past, recent, and current experience and would certainly be unlikely to hold true if the capital/income ratio increased further. As a related matter, most scholarly evidence shows that the elasticity of substitution between capital and labor is less than 1, particularly if housing is included in the measure of capital. In the United States, the savings rate has declined in recent decades at the same time stock and housing prices have increased, leading to greater household wealth. So there are indeed automatic mechanisms in market systems applying to the two fundamental laws of capitalism to prevent the continual expan- sion in the capital/income ratio and any income inequality that would arise therefrom. Much has been written questioning the data produced by Piketty and his colleague Saez on increasing U.S. income and wage inequality. Har- vard economist Martin Feldstein made the follow- ing three points in a May 15, 2014, editorial in The Wall Street Journal: Figure 2. U.K. Consol Yields and Inflation Rates From 1728 to 2013 Long-TermGovernment BondYields 10-Year MovingAverageforAnnual InflationRate 20 15 10 5 0 -5 -10 D o c u m e n t a t i o n Years Sources: http://www.dmo.gov.uk/ceLogon.aspx?page=about&rptCode=D41and http://www.measuringworth.com/calculators/inflation/result.php. 1 7 3 3 1 7 4 1 1 7 4 9 1 7 5 7 1 7 6 5 1 7 7 3 1 7 8 1 1 7 8 9 1 7 9 7 1 8 0 5 1 8 1 3 1 8 2 1 1 8 2 9 1 8 3 7 1 8 4 5 1 8 5 3 1 8 6 1 1 8 6 9 1 8 7 7 1 8 8 5 1 8 9 3 1 9 0 1 1 9 0 9 1 9 1 7 1 9 2 5 1 9 3 3 1 9 4 1 1 9 4 9 1 9 5 7 1 9 6 5 1 9 7 3 1 9 8 1 1 9 8 9 1 9 9 7 2 0 0 5 2 0 1 3 COMMENTARY / POLICY AND PRACTICE TAX NOTES, June 30, 2014 1553 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . 1. TRA1986 lowered the top rate on all income from 50 percent to 28 percent and considerably expanded the tax base by reducing the use of deductions and exclusions, importantly limit- ing the use of top-hat pension plans and other forms of deferred compensation. Because the top income earners and the corporations that pay them are sensitive and quick to react to changes in tax rules, more income was paid as taxable salaries but did not increase the total resources available to those earners. 2. The income tax data used by Piketty does not include the corporate income of profes- sionals and small businesses before TRA 1986s repeal of the General Utilities doctrine and the decline of the top personal rate to less than the corporate rate. Those changes prompted high-income taxpayers to shift their business income out of taxable corporations and onto their individual returns. That was achieved in part by having the companies pay the individuals interest, rent, or salaries, or by converting the businesses to subchapter S cor- porations. These changes in taxpayer behav- ior substantially increased the amount of income included on the returns of high- income individuals. This creates the false im- pression of a sharp rise in the incomes of high-income taxpayers even though there was only a change in the legal form of that in- come. (Feldstein claims that this change hap- pened gradually, but again, because of the rapidity by which the well-to-do react to the advice of their tax advisers and the long lead time of TRA 1986, I think it is more likely to have occurred quickly. Indeed, the data used by Piketty show a particularly rapid rise in inequality in the immediate years after the passage of TRA 1986.) 3. Social Security, other retirement benefits, and health benefits (both from the government and from employers) are a large and growing part of the personal incomes of low- and middle-income households but, while consid- ered earned, are largely not taxed. When com- paring the incomes of the top 10 percent of the population with the total personal incomes of the rest of the population, including these benefits would show a much smaller rise in the relative size of incomes at the top. Regarding the point about employer-provided health benefits, my own empirical research for a recent period has shown that the rapidly increasing cost of healthcare largely explains the increase in Figure 3. Yields (Percentage) on U.K. Inflation Linked Treasury Bonds (2035 Treasury Stock), November 25, 2002, to June 3, 2014 Close of Business Day Source: http://www.dmo.gov.uk/ceLogon.aspx?page=about&rptCode=D3B. -1 -0.5 0 0.5 1 1.5 2 2.5 S e p t . 3 , 2 0 0 2 J a n . 9 , 2 0 0 3 M a y 1 9 , 2 0 0 3 S e p t . 2 3 , 2 0 0 3 J a n . 2 9 , 2 0 0 4 J u n e 8 , 2 0 0 4 O c t . 1 2 , 2 0 0 4 F e b . 1 7 , 2 0 0 5 J u n e 2 8 , 2 0 0 5 N o v . 1 , 2 0 0 5 M a r . 9 , 2 0 0 6 J u l y 1 8 , 2 0 0 6 N o v . 2 1 , 2 0 0 6 M a r . 2 9 , 2 0 0 7 A u g . 7 , 2 0 0 7 D e c . 1 1 , 2 0 0 7 A p r . 2 1 , 2 0 0 8 A u g . 2 7 , 2 0 0 8 J a n . 2 , 2 0 0 9 M a y 1 2 , 2 0 0 9 S e p t . 1 6 , 2 0 0 9 J a n . 2 2 , 2 0 1 0 J u n e 2 , 2 0 1 0 O c t . 6 , 2 0 1 0 F e b . 1 1 , 2 0 1 1 J u n e 2 3 , 2 0 1 1 O c t . 2 7 , 2 0 1 1 M a r . 5 , 2 0 1 2 J u l y 1 3 , 2 0 1 2 N o v . 1 6 , 2 0 1 2 M a r . 2 6 , 2 0 1 3 A u g . 6 , 2 0 1 3 D e c . 6 , 2 0 1 3 A p r . 1 5 , 2 0 1 4 COMMENTARY / POLICY AND PRACTICE 1554 TAX NOTES, June 30, 2014 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . reported wage inequality in the United States. 3 Over the seven-year period for which I obtained compensation data from the Bureau of Labor Statis- tics (BLS), health cost increases fully accounted for changes in the distribution of income. In other words, without rising health costs, there would have been virtually no change in income inequality over the period of 1999-2006. As background, most employers pay workers a combination of both wages and fringe benefits, which include paid time off, retirement plans, health coverage, and other benefits. The role of fringe benefits has grown significantly over time: In 1950 benefits made up only 7 percent of total compensation, according to the Bureau of Economic Analysis; today, benefits make up 20 percent of compensation. Most of these benefits are not in- cluded in taxable income. Also note that these fringe benefits have become more widely distrib- uted across income groups over time, owing in part to the tightening of nondiscrimination rules in the tax code that applies to employers providing ben- efits to their full-time workers. In particular, when pension plans were first widely established to get around the effect of wage controls and higher tax rates in World War II, they primarily benefited higher-paid workers, what we now measure as the top decile. This is now untrue a significant provision of benefit plans extends much lower down in the earnings categories. Both economic theory and empirical findings indicate a trade-off between salaries and benefits: If benefits become more expensive, wage growth will suffer. Indeed, the fact that employer costs for family health coverage, according to the Kaiser Family Foundation, rose from around $4,200 in 1999 to nearly $11,800 in 2013 helps explain why average wages have stagnated in recent years. Total com- pensation continued to increase, but rapidly grow- ing health costs ate away at wages and non-health benefits. But not every employee is affected in the same way by rising health costs. The dollar cost of health coverage is similar for high- and low-income work- ers, which means that healthcare makes up a far larger share of total compensation for low-income earners than for the top 1 percent. I obtained unpublished data from the BLSs National Compen- sation Survey that show that for the lowest-paid full-time workers in 1999, health coverage made up around 6.2 percent of total compensation. Those workers were in the 30th percentile of the overall wage distribution. For middle-income workers, em- ployer health contributions made up 7.2 percent of compensation. This was not because their health coverage was more generous in dollar terms so much as because health coverage is more predomi- nant in middle-income jobs. But for the top 1 percent of earners, health coverage made up just 4 percent of compensation. Now, consider what happens when health costs increase rapidly. Although rising healthcare costs eat away at wage growth for everyone, the effects will be largest for the working and middle classes because their health costs are so large relative to the rest of their compensation package. The BLS data show that from 1999 through 2006, health coverage rose from 6.2 percent to 12.2 per- cent of compensation for a lower-wage worker, a massive increase for a seven-year period. As a result, while total compensation for this group rose by 41 percent from 1999 through 2006, wages grew by only 28 percent. For a middle-income worker, health coverage increased from 7.2 percent to 10.4 percent of compensation. And while compensation grew by 34 percent from 1999 through 2006, wages 3 See Mark J. Warshawsky, Can the Rapid Growth in the Cost of Employer-Provided Health Benefits Explain the Observed Increase in Earnings Inequality? Pension & Benefits Daily (Feb. 3, 2012). Growth of Earnings and Total Compensation, 1999-2006 Health Coverage as Percent of Compensation Growth, 1999-2006 Earnings Percentile 1999 2006 Earnings Compensation 30th 6.2% 12.2% 28% 41% 40th 8.0% 9.9% 26% 28% 50th 7.2% 10.4% 27% 34% 60th 6.8% 11.1% 27% 36% 70th 7.3% 9.6% 28% 34% 80th 6.8% 8.5% 30% 36% 90th 6.5% 7.3% 31% 33% 95th 5.5% 7.1% 34% 38% 99th 4.0% 4.3% 35% 36% Source: Mark J. Warshawsky, BNA Pensions & Benefits Daily, Feb. 3, 2012. Unpublished BLS National Compensation Survey data. Bottom 30 percent of workers omitted to exclude part-time employees. COMMENTARY / POLICY AND PRACTICE TAX NOTES, June 30, 2014 1555 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . grew by only 27 percent. For the top 1 percent of earners, health costs increased from 4 percent to 4.3 percent of total compensation. Because of the smaller role healthcare plays for top earners, their earnings grew nearly as quickly as their compensa- tion 35 percent for earnings and 36 percent for total compensation. Total compensation that is, the total wages and benefits paid to employees did not grow more unequal from 1999 through 2006. In fact, total compensation for the lowest-paid full-time workers increased more quickly than for the top 1 percent. But rising health costs suppressed wage growth much more for lower- and middle-income workers than high-income earners, with the result that re- ported wage and income inequality rose signifi- cantly. These data show that in the absence of rapidly rising health costs, income inequality wouldnt have budged from 1999 through 2006. 4 If healthcare were producing additional value commensurate with its rising costs, these changes in the makeup of workers compensation wouldnt matter, just as we wouldnt care much if employers paid workers somewhat lower salaries but contrib- uted more to their retirement plans. But many studies suggest that the extra spending in the U.S. healthcare system is often of marginal benefit to patients. We pay more, but we often dont get more. Workers would have been better off paying less for healthcare and seeing higher wages on their pay stubs. And policy to reduce inequality would be better directed at reducing the rapid growth in the cost of healthcare than at increasing tax rates in the higher brackets. Another criticism of Pikettys interpretation of his findings has been made by Scott Winship in the May 19, 2014, issue of the National Review. Because tax returns count all gains when they are realized and members of the top 1 percent strategically time the sale of their assets after holding them for years, all of the gains accruing over time are counted on a single tax return in years close to asset-market peaks. This increases the share of capital income accrued by the top of the income strata, since its concentrated in one year. Indeed, there is a strong element of stock market performance in both the income (including capital gains) and wage (includ- ing stock options) data reported by Piketty, spiking with the booms and falling with the busts. Winship also attributes some of the trend of the reported increase in income inequality to a declining house- hold size over recent decades as marriage rates dropped, divorce rates rose, and elderly widow- hood increased and to a conflation of tax returns with households and persons. Finally, getting into the methodological weeds, one finds there may be another artificiality intro- duced into Pikettys results by swings in the per- centage of the adult population (tax units) filing income tax returns. Since the full-scale implemen- tation of income tax withholding during World War II, the percentage of units filing income tax returns has averaged about 90 percent. But that percentage fluctuates significantly. For example, from 2000 to 2012, it declined from 96 percent to 90 percent. Still, Piketty and Saez assume that all nonfilers get a constant 20 percent of average income, which is somewhat arbitrary, whereas one might have ex- pected the income of nonfilers to increase some- what as the percentage of filing declines, particularly if the decline is attributable to changes in tax administration and law intended to remove low-income taxpayers from filing status. Hence, the rise in inequality is probably overstated. Note that this criticism holds true in the opposite direction as well: When the percentage of filers exploded during World War II from 14 percent in 1939 to 85 percent in 1945 it is likely that the average income represented by new filers declined, so the sharp decline in inequality during the war reported by Piketty is probably overstated. Capitals Style and Tone Now, allow me to make some personal com- ments about the books style, approach, and tone. Although the book is jammed with extensive data, written by a professor of economics and published by Harvard University Press, Capital is not a work of science. The discussion of empirical method in the text is exceedingly brief, and although the summary of theory is in parts quite good and intuitive, it is neither nuanced nor complete. Most importantly, we are given no information or think- ing by which to judge what levels and types of inequality are bad immoral or harmful. Nor is there any inclusion of the empirical effects of cur- rent policies and practices tax, social insurance, welfare, charity, and so on that effectively reduce inequality in all countries in modern times. Those are notable omissions for the discussion of policy, given that there is significant relevant philosophical and economic literature to use. By Pikettys admission, the data reported are sometimes smoothed, interpolated, combined, or adjusted. There seems to be no logical consistency 4 A Congressional Budget Office study also incorporated healthcare costs into its analysis of income inequality. The CBO found a smaller role for health costs in driving income inequal- ity. But that study indirectly estimates health benefits for employees using household survey data, which are subject to error, and healthcare data sets from the 1970s. By contrast, the BLSs National Compensation Survey is more current, and that data is collected directly from employers, providing more timely and reliable data. COMMENTARY / POLICY AND PRACTICE 1556 TAX NOTES, June 30, 2014 ( C )
T a x
A n a l y s t s
2 0 1 4 .
A l l
r i g h t s
r e s e r v e d .
T a x
A n a l y s t s
d o e s
n o t
c l a i m
c o p y r i g h t
i n
a n y
p u b l i c
d o m a i n
o r
t h i r d
p a r t y
c o n t e n t . in the picking and choosing among methods and different types of data sets. This broad approach makes the data look more convincing than they really are; other scholars would instead show series breaks and several alternatives and list caveats. Pikettys straight use of informal journalistic data sources with an inadequately disclosed method, such as Fortune magazines tracking of the worlds billionaires, would never appear in a top profes- sional economics journal (I think/hope). Although generally well written and well orga- nized, the book would have benefited from tighter editing, since it is sometimes repetitive. The re- peated invocations of Balzacs Pre Goriot, while at first clever and truly illustrative of the dynamics of inheritance and income in historical France, get annoying and tiring by the fourth and fifth and sixth mentions. And despite the books global am- bitions and American popularity, it is mainly and consistently about France and Britain in terms of formal data and depth of historical explanations, and indeed for many of its projections and conclu- sions. The new work in the book, and its main contri- bution the models of future capital/income ratios and capital shares is informal and impres- sionistic, almost no better than back-of-the- envelope calculations and armchair philosophy. The discussions throughout the book are not bal- anced or comprehensive, failing to consider past criticisms, additional data, or alternative explana- tions or points of view. For example, Piketty does not consider that reported developments might come as narrow behavioral reactions to changes in tax law and administration, and indeed might arise from the inherent nature of the reporting and collection of the data. And while he initially allows for good (that is, justified) as well as bad inequality, Piketty abandons that pretense by the middle of the book and, as his ideal, suggests capital ownership shares substantially more evenly distributed than observed even in Scandinavia of the 1970s-1980s. And then there is the lack of respect for past work by the pioneers and leaders of the economics profession. For example, Piketty dismisses and mis- represents the views of the 1971 Nobel Prize winner and early pioneer in the empirical study of income inequality, professor Simon Kuznets, on the evolu- tion of income inequality across time and countries as magic, a fairy tale, motivated by professional hubris, and a product of the Cold War. Are economics and economists really that small? Concluding Observations Lets not end this review on a critical note. Piketty is to be credited with the hard work of collecting the massive data sets and writing a long but accessible book with a sustained argument based on strong and generally openly stated opin- ions. Moreover, income inequality is a universal and timeless issue for those concerned with a fair and free society, as we should all be. Fairness is also a major consideration in tax policy. That is particularly true today as policymak- ers for the federal and state governments recognize that their structural levels of spending, now and especially for the future, are severely underfunded and overpromised. Either spending must be cut (but how to do so fairly?), or taxes must be in- creased (again, how to do so?). It is no surprise that those who support raising taxes and revenues have championed Pikettys work because it justifies advocates a more progressive approach to tax policy, generally thought to be more popular with the electorate than across-the-board tax increases. On the other side of the debate, Pikettys focus on raising the taxation of capital and entrepreneurial effort is criticized by those who believe it will harm the engine of economic growth, the revving of which is the best remedy for revenue shortfalls. Moreover, they believe his approach will impinge on the individuals freedom to pursue economic opportunity. Although Pikettys strong conclusions about capital, inequality, and the implications for tax policy are not supported by his data and projec- tions, other insights are possible. For example, the clear increase in capital in France and Britain attrib- utable to a rise in housing prices and assets, likely concentrated in increasingly expensive and finite Paris and London, argues for more enlightened policy to spread infrastructure, housing develop- ment, and cultural activities across all parts of those countries. In the United States, the reported increase in wage inequality clearly provides support for new strategies and policies to address one of the key underlying causes the rapid rate of growth in the cost of healthcare. Also, for the United States and Britain, some of the reported perhaps some actual increase in income inequality is attribut- able to stock market performance. While this is mostly a reflection of the greater risk taken by the higher-income groups, it makes sense to more ag- gressively offer these higher expected returns to lower-income groups on an opt-out, low-cost basis, say through add-on personal accounts centrally managed as part of Social Security reform. COMMENTARY / POLICY AND PRACTICE TAX NOTES, June 30, 2014 1557 ( C )