This is a monumental work about inequality. Despite the title's allusion to Marx's classic (a point emphasized by the dust jacket design), it's neither a primarily theoretical nor a primarily polemical work, though it has elements of both theory and advocacy. Nor is its author (TP) a radical: he taught at MIT, and is thoroughly at home in the concepts and categories of mainstream neoclassical theory. Nonetheless, I think even many who hold less orthodox views about economics will find this book stimulating, valuable and sympathetic in many respects. And all readers ought to find it disturbing.
In the ultra-long comments below, I begin with the book's audience and style (§ 1); then turn to some of the book's main arguments, which are more nuanced than usually reported (§§ 2-6); then to some things that are unclear or missing (§§ 7-8); and I end with some comments about the book's production (§ 9) and some concluding remarks.
1. In the original French edition, TP says that he intended this book to be readable for persons without any particular technical knowledge. In principle, it could be read by a broad, college-educated audience. TP's prose is very clear and direct, with a low density of jargon and a high density of information. (I read the French edition, but Arthur Goldhammer's translation seems to preserve these qualities very well.) The discussion is enlivened by well-chosen references to literature and a sprinkling of sarcastic barbs, both of them techniques that French scholars have developed into art forms (if not as elegant as John Kenneth Galbraith's irony). The allusions here range from Balzac, Jane Austen and Orhan Pamuk to "The Aristocats," "Bones" and "Dirty Sexy Money;" and the sarcasm hits both university economists and The Economist (@636n20), among others.
But: this is a long and demanding book. It talks relatively little about current events or the policies of particular governments, unlike, say, Joseph Stiglitz's "The Price of Inequality" (2012). I wouldn't say Stiglitz's is an easy book, but it was written more with of a popular audience in mind (picking up 270+ Amazon reviews in less than 2 years). TP's presentation is far more methodical and meticulous than Stiglitz's. It helps for the reader to be interested in the fine points of data series and categories, and in the sources of uncertainty in data. Occasionally the discussion will focus on concepts from academic economics, such as Cobb-Douglas production functions, elasticities, and Pareto coefficients; while TP uses words rather than math on these occasions, he generally assumes you pretty much know what he's talking about. Finally, if, as I did, you make it through the whole thing while reading with some attention, I bet dollars to donuts you'll come out of the experience feeling very, very down, on account of TP's message. Actually, that mood will hit you long before the end. Despite its felicities of style, this is an arduous read.
2. The "capital" in the title includes not only farms, factories, equipment and other means of production, but also assets typically owned by individuals, such as real property, stocks and other financial instruments, gold, antiques, etc. -- what's sometimes called "wealth". TP excludes so-called "human capital," because it lacks some features of true capital (ability to be traded in a market, inclusion in national accounts as investment), unless it's in the form of slaves.
The distribution of capital is far more unequal than that of income. Even the Scandinavian countries have a Gini coefficient for capital of 0.58 -- comparable to that for income inequality in Angola and Haiti, among the 10 worst in the world (World Bank figures). For Europe and the US in 2010, the coefficient is at 0.67 and 0.73 respectively, worse than any country on the World Bank income inequality chart. (Of course, the worst countries on that World Bank list have hair-raising capital inequality, too.)
The book's main thesis is that economic growth alone isn't sufficient to overcome three "divergence mechanisms" or "forces" that are in many places returning inequality in income and/or capital to pre-World War I levels. The main mechanisms are:
(A) the historical tendency of capital to earn returns at a higher rate ('r') than the growth rate of national income ('g'), which typically sets a constraint on how workers' salaries grow, symbolized by the mathematical expression, "r > g".
(B) the relatively recent (post-1980) widening spread between salaries, not only between the wealthiest 10% or 1% and the mean, but even within the top 1%.
(C) an even newer inequality in financial returns, which correlates r with the initial size of an investment portfolio -- i.e., different r for different investors.
A point to keep in mind is that g relates to national income, not to GDP. National income = GDP - depreciation of capital + net revenue received from overseas. Among other benefits, this measure corrects for the reconstruction boosts in GDP after wars, hurricanes, earthquakes, etc., since the depreciation term takes the destruction of property into account. Also, an increase in national income usually has two different sources: part of it is truly economic, coming from productivity growth (output per worker), and part is due to population growth. Historically, it's the latter that has dominated.
3. The r > g argument has received the most attention. It's to be seen "as an historical reality dependent on a variety of mechanisms not as an absolute logical necessity" (@361). TP finds that this condition has held throughout most of the past 2,000 years. As long as it does, he says, it's the natural tendency of capitalism to make inequality worse -- and the bigger the difference (r - g), the worse that inequality will be. Many commentators about this book make it sound as if this is an obvious mechanism. But if you play with it on Excel, using reasonable values for r and g, it turns out to be slower and more sensitive to initial conditions than you might expect.
Here's a toy example: Let's suppose r = 4%, g = 1.5%, and that salaries rise as fast as g (a very idealistic assumption!); and let's assume these rates are net of taxes or that no taxes apply. I'll compare the situations of three people in Silicon Valley: X, an engineer who made $8.5 million by exercising stock options when the company she used to work for had an IPO; Y, the same company's former HR manager, who made $6.0 million from her options; and Z, a young lawyer at a local law firm, who has a $200,000 salary when we first meet her.
After a year, X has $340K in disposable income, Y has $240K, and Z gets a raise to $203K. Suppose X and Y spend all their income from their capital every year. Eventually, Z can earn more than each of them: it will take her about 37 years to exceed X's annual income, but only 13 years to make more than Y. Now suppose X and Y each save the equivalent of 1.5% of their capital. Then Z will never overtake either one in gross annual revenue, but the situation as to disposable cash is a bit different. After saving, X will always have more cash to play with than Z, but it will take more than 15 years for her to have just 50% more than Z does. As for Y, she'll actually start out with less annual cash than Z, and it will take her 13 or 14 years just to catch up -- even though she's a multi-millionaire.
The true potency of the r > g mechanism comes from its working in conjunction with other circumstances. For example, according to TP's historical data, I've been way too conservative in my assumptions about X's and Y's advantages over Z.
From the 18th through the early 20th Centuries, the people who earned money from capital had proportionally a lot more than they do today: e.g., in 1910, the wealthiest 1% in Europe held > 60% of all European wealth, about triple the share they hold today (see Fig. 10.6). The US was not so extreme, but still very unequal: From 1810 to 1910, the share of the top 1% grew from 25% of American wealth to 45.1% (Fig. 10.5), compared to 33.8% today. So to set our example 100-200 years ago, the endowments of X and Y could plausibly be much bigger relative to Z's wages (especially if we chose, say, Wilhelmine Germany instead of Silicon Valley).
More recently, since the 1980s, most folks with a lot of capital also earn salaries -- and having a lot of capital tends to be correlated with having a salary well above average. So in a more realistic modern example, we should consider that X and Y have moved on to new companies where they receive hefty salaries, which would give each in total a healthy and growing excess of annual spendable cash versus Z. This is the realm of the second divergence mechanism, which is especially formidable in America. In 2010 the richest 1% not only held more than 33% of American wealth, but they earned between 17x and 20x the mean American income (depending on whether capital gains are included). Even the wealthiest 0.1% of Americans work, for average incomes roughly 75x the mean (or 95x, if capital gains are included) (see Table S8.2). At the other end of the spectrum, I was shocked to learn that the purchasing power of the US Federal minimum wage peaked in *1969* -- what was $1.60 an hour back then would be worth $10.10 in 2013 dollars. In those same dollars, the current statutory minimum hourly wage is $7.25 or a bit less (see Fig. 9.1 and nearby text).
On top of these trends, succession to family wealth is becoming important again today, even if not to the full degree it was in 19th Century novels. TP frames this in terms of the dialogue of the worldly Vautrin and the young, ambitious Rastignac in Balzac's "Père Goriot" (1853). Rastignac aspires to wealth by studying law. Vautrin counsels him that unless he can claw his way to become one of the five richest lawyers in Paris, his path will be easier if he simply marries an heiress in lieu of study. Cut to the present: judging by TP's Fig. 11.10, law school might have been the better choice for Baby Boomers, but if you're a Rastignac in your 20s or 30s when you read this, consider marrying up. Maybe you think you'd rather found the next Facebook or Google -- but why work so hard, and against such long odds? TP shows that when Steve Jobs died in 2011, his $8 billion fortune was only 1/3rd that of French heiress Liliane Bettencourt, who has never worked a day in her life.
4. There's another way that "r > g" is inadequate as a summary of TP's argument: TP calculates that during the past century (1913-2012), we've seen r < g, the opposite of its usual polarity (Chapter 10).
High rates of growth -- or at least, what we're accustomed to thinking of as high rates of growth, 3%-4% or more -- aren't a sufficient explanation. In fact, such rates of growth aren't sustainable in the long term, and were not sustained in most countries; they're mainly a catch-up mechanism lasting a few decades, according to TP. During the period from 1970-2010, the actual per capita growth rate of national income averaged about 1.8% for the US and Germany, 1.9% for the UK, and 1.6%-1.7% for France, Italy, Canada and Australia. The wealthy country with the highest per capita rate was Japan, at 2.0 (Table 5.1). (Think about that, next time you're tempted to swallow what Paul Krugman and other pundits pronounce.) Nonetheless, growth rates in this range appear to be what TP calls "weak" (e.g., @23).
Rather, the main reasons for the flip are the tremendous destruction of capital in Europe due to the two world wars, and the imposition of very substantial taxes on capital, at an average rate of about 30% in recent years. These greatly reduced r.
Despite these trends, inequality has been getting worse during the past few decades. This isn't a paradox, but rather the impact of the other divergence mechanisms, especially the rise of the "working rich" and the spread of inequality in salaries. So we should be quite alarmed by TP's assertion that we'll flip back to r > g during the 21st Century. His explanations for this seem rather more speculative than most of the rest of the book, though it's clear he expects g to remain low. I return to this a bit more in § 7 below.
In any case, it's clear that r > g isn't a necessary condition for inequality to get worse.
5. TP reserves his most anxious prose ("radical divergence," "explosive trajectories and uncontrolled inegalitarian spirals") for the third mechanism, inequality in returns from capital (@431, 439). Those with a great deal of capital are able to earn higher returns on it -- such as 6%-7%, or even 10%-11% in the case of billionaires like Bill Gates and Bettencourt -- compared to those with only a few hundred thousand or millions of dollars, who may earn closer to 2%-4%. This results from two types of economies of scale: the ultra-rich can afford more intermediaries and advisers, and they can afford to take on more risk.
Unfortunately, public records don't provide adequate information on this point, and while TP does look at Forbes's and other magazines' lists of the wealthy, those present many methodological issues. So TP corroborates his findings by looking at the more than 800 US universities who report about their endowments. Most spend less than 1%, or even less than 0.5%, of their endowments on annual management fees. Harvard University spent around $100 million annually (ca. 0.3%) on management of its $30 billion endowment, and earned net returns of 10.2% annually during 1980-2010 (not counting an additional 2% annual growth from new gifts). Yale and Princeton, each with a $20 billion endowment, earned a similar rate. A majority of universities have endowments of less than $100 million, and so obviously can't fork over $100 million to managers; they earned average returns of 6.2% during that period (still better on average than you or me).
TP of course doesn't worry that universities will own most of the world, nor does he find it plausible that sovereign funds from Asia or oil-producing countries will either. The bigger danger, he contends, is private oligarchs, and he believes this process is already underway. Since the officially documented ownership of global assets comes up slightly negative, TP calculates that either the rich are already hiding the equivalent of at least 8% of global GDP in tax havens, or else that our planet is owned by Mars (@465-466).
6. In Part IV of the book, TP considers policy approaches to deal with the three forces of divergence. In short, the answer for all three is a progressive, annual global tax on capital, to be set at an internationally agreed rate and its proceeds apportioned among countries according to a negotiated schedule (@515). This will also need a global real-time reporting system for transactions in capital assets. Many will attack these ideas, but it seems that TP's main intention is to get a serious conversation going. His admits his approach is utopian, but maintains that utopian ideas are useful as points of reference.
What interested me most was that TP doesn't see pumping up g as a viable approach to preventing r > g from returning. For one thing, demographics create some limitations in how far g can be pushed, especially in countries whose populations will soon be declining (or Japan, where that's happening already). For another, the same forces that pump up g can also increase r, at least in theory, so (r - g) wouldn't necessarily change much. The more practical answer then, is to bring down r.
In his final chapter TP turns to the very topical question of public debt, which he sees as an issue of wealth distribution and not of absolute wealth. He reminds us about two of its important aspects: One is that public debt takes money from the pockets of the mass of citizens, who pay taxes, and puts it in the pockets of the smaller group of people who are wealthy enough to make loans to the state. The other point is that nations are rich -- it's only states who are strapped for funds. He calculates that in many countries, a one-time progressive capital tax of up to 20% on property portfolios worth more than 1 million Euro could bring the national debt to zero, or nearly so.
Actually, TP doesn't believe that such a drastic reduction in debt levels is urgent, any more than he believes that such a gigantic tax is politically feasible. But his observation puts the lie to the notion that one must raise consumption taxes or income taxes (or, for that matter, experience economic growth) to reduce debt levels.
7. There were a couple of rare instances where I didn't feel the text was sufficiently clear. TP very graciously replied to my emailed inquiries about these matters, but without that input, I'd have remained quite confused by them.
(a) The first arose in Chapter 1, where α (alpha) is defined as designating the "share of income from capital in national income." According to the perhaps intemperately named "first law of capitalism," α = rβ, where β is the ratio of the stock of capital to the flow of national income (and r is as above, the rate of return on capital).
But an important category of income from capital is capital gains, the profits you make when you buy assets cheap and sell them dear. Unrealized capital gains make up a substantial part of the fortunes of Bill Gates, Steve Jobs and other billionaires mentioned in the book. And capital gains are *not* included in national income, according to the algorithm for computing that quantity. (Nor are they included in GDP.) This makes the use of the preposition "in" confusing -- does it mean that capital gains aren't considered as income from capital?
This issue seems to have its root in academic economics, where α appears as a parameter in the neoclassical growth model developed by Robert Solow. The model represents an economy that produces one type of good -- i.e., it's all about making and selling stuff that gets consumed, so capital gains aren't considered. (In a sense, this model supplies a lot of the motivation for Part II of the book: the academic debate over the relative shares of capital and labor in the national income, i.e., the size of α and whether it changes with time, is a long and at times contentious one. But you can still benefit from reading Part II without knowing that.)
The answer I got from TP is that because capital gains don't seem to be very important in the long term (>100 years), netting out to roughly zero over such periods, he didn't consider them when discussing α. The subject of capital gains does come up later in other contexts, though, and TP does consider them important in the short-term (which in some contexts can mean a timescale of several decades).
(b) The second issue relates to TP's prediction that our current condition of r < g will flip back to r > g later this century. TP mentions that for the past 100 years, wartime destruction and, later, an average 30% tax rate on capital have brought r below g, despite currently weak growth rates in many countries. The data in the book, though is rather opaque about the relative contributions of these factors. Also, the book's clearest explanation of why matters might reverse rests on the possibility that countries will compete to attract capital by a race to the bottom in capital tax rates, allowing r to edge back up. This sounded rather too speculative to warrant such definite conviction about the return of r > g.
I checked the online material, and found the Excel file (not the pdf file) of supplementary Table S10.3, which mentions some of TP's assumptions. Among other things, this makes it clear that TP factors in destruction of capital from WWII in calculating r even for the most recent 50 years. It seems plausible that this will be less important going forward, so that even a 30% average tax rate on capital might not be sufficient in and of itself to prevent r from popping above g again ... maybe. I'm still not entirely convinced that TP's argument about the future of r is among the strongest in the book; but I'd be even less so if I hadn't consulted the online information.
8. No book can talk about everything pertinent to its theme, so it's all too easy to think of things one wishes had been included. Still, I was disappointed that the book was conventional both in its thinking about economic growth, and in its thinking less about growth's environmental consequences.
TP tells us that part of "the reality of growth" is that "the material conditions of life have clearly improved dramatically since the Industrial Revolution" (@89). Its main benefits include its roles as a social equalizer, and as a "diversifi[er] of lifestyles" (@ 83, 90). "[A] society that grows at 1 percent a year ... is a society that undergoes deep and permanent change" (@96).
Growth's equalizing effect, though, comes largely from population-based growth, whereas "a stagnant, or worse, decreasing population increases the influence of capital in previous generations" (@84). So is a country already in that condition, such as Japan, supposed to open its doors to immigrants? As an immigrant to Japan myself, I can appreciate that there are many social, cultural and political reasons why this could be a bad path both for country and for many of the immigrants as well. How about focusing on productivity-led growth instead? Maybe, because "in a society where output per capita grows tenfold in a generation, it is better to count on what one can earn and save from one's own labor" (@84), instead of relying on an inheritance. The problem is, this takes for granted that gains from productivity improvements will be shared with labor, rather than shareholders. Yet Part II shows that labor's share has been flat or declining. In Japan, productivity improvements nowadays tend to come from using temporary employees instead of higher-paid permanent ones, and from using robots in lieu of employees at all. These have worked out to be more methods for enhancing inequality, than for abating it.
Both population growth and productivity growth have other costs, too. The rapid growth of output TP alludes to could only be of the transitory, catch-up sort, such as China has been experiencing since the 1980s. The environmental consequences of that haven't exactly been benign. Nor does the book give any consideration to the environmental consequences of population growth, when the population in question aspires to a wealthy country's per capita environmental footprint.
So are countries with declining populations doomed to oligarchy until all the other countries in the world can agree on a global capital tax? Obviously there are better ways to proceed. Such as examining whether growth truly is necessary for further improving health and other material conditions of life, even in an already-wealthy country. And inquiring whether deep and permanent change is a virtue in itself, or whether good sorts of changes can be achieved without following policies obsessed with growth. Exploring such questions thoroughly would certainly have been outside the scope of this book, but failing even to hint at their existence was either a missed opportunity or a lapse of imagination.
9. In addition to the good translation, some other aspects of the book's transition to English succeed. The US hardcover has sewn signatures; my closely-read and much-shlepped French copy, which comes in at nearly 1,000 pages in a perfect binding, is already showing signs of loose leaves. The US edition has a pretty good index, whereas the French lacked one entirely. It's not quite complete, though: e.g., you won't find the above-mentioned references to Mars, "Bones" or The Economist in it, and I noticed a few references to Japan that were missing, too. On the other hand, the notes didn't fare as well. The notes in this book are long, discursive and informative; you really should read them. The French original used footnotes, but Harvard opted for endnotes, which means you'll either be doing a lot of flipping back and forth, or else ignoring a lot of good material.
A mixed blessing in both editions is that the technical appendix has been punted online. The package is generous, and includes files for the book's tables and figures in both pdf and Excel formats. The expository appendix (evidently translated by someone other than Dr. Goldhammer) includes hyperlinks to pertinent scholarly articles. Downloading the 2013 paper TP wrote with Gabriel Zucman, "Capital is Back," along with its own humongous technical appendix, might be a good choice: the present book's technical appendix refers to this often. If you want all relevant Excel files (including, e.g., some UN data and TP's comments to the Angus Maddison historical data), be sure to scroll through the pdf of the appendix and click on appropriate links, since several such items are absent from the website's "Piketty 2014 Excel files" folder.
Unfortunately, no one can know if this website will be maintained a few decades from now, or how easy it will be to read .pdf and .xls files by then. Just as is the case today with books by leading mid-20th Century economists, this is the sort of book that scholars will still want to read in future, even after it's out of print. They'll be very frustrated by the many cross-references to the technical appendix (at least 100-200 times by my eyeball count) if the information has vanished. I hope that in the not-too-distant future TP will freeze and publish a hard copy of this supplemental material for archival purposes.
It's also surprising that not even the website provides a comprehensive bibliography. The technical appendix includes a number of references, but these are spread out over a list at the beginning and more references embedded into a chapter-by-chapter commentary. Even this fragmented resource doesn't pick up many of the books and articles mentioned in the printed book's endnotes/footnotes. Again, I hope TP or the publisher will remedy this soon.
===
Among its other accomplishments, the book demolishes a couple of abstractions from the 1950s that economists have cherished for decades. One is the "Kuznets curve," according to which income inequality first rises, then peaks and thereafter declines as per capita GDP (or earlier, GNP) continues to rise. Another is the Modigliani "life-cycle" saving theory, which posits that the people save for their retirement and then spend pretty much everything by the time they die. TP's long runs of data show that both of these theories were plausible, if ever, at best only during a brief era around the time they were formulated, when both capital and income were distributed in a more egalitarian way.
How will the economists of today react to this book? Paul Krugman didn't provide an encouraging sign in his blog a few days after the US edition appeared. First thing he did was to try to "understand" it by plugging TP's data into another abstract 1950s-era mathematical model. The vast majority of mainstream economists didn't see the 2008 crash coming, but after it happened they insisted that their models weren't defective. If an historical event of that magnitude couldn't make a dent in their worldview, one has to be a great optimist to believe that this book will. More realistic may be to hope that this book's impact can be political. Luckily, that isn't up just to economists, but to readers like us.
Capital in the Twenty-First Century
4.5 | 5,481 ratings
Price: 34.96
Last update: 07-31-2024
Top reviews from the United States
A. J. Sutter
An important book (in both the English and French senses of the word)
Reviewed in the United States on March 17, 2014
Anne Mills
A Milestone in Economic Thought
Reviewed in the United States on June 14, 2014