Nathan Beckmann's Thoughts on Capital in the Twenty-first Century
Capital in the Twenty-first Century by Thomas Piketty. This book is intentionally setting itself in the tradition of 18th century economists that focused on large-scale macroeconomic issues, particularly on the distribution of income and wealth. It has been reviewed well and deemed to be an important contribution to economics, so I'm giving it a fairly in-depth reading. Here's the major points in the book as I've encountered them, and my thoughts:
- He begins with a review of economic thought, with harsh things to say about the modern practice of economics. In his view, economists ignore data or the important questions in the real world to focus on mathematical models in a vain attempt to be "scientific". He wants economics to focus on big questions raised by society with a focus on historical data to tease out the most important factors at play. There are some great quotes in this section about economists knowing nothing about anything but talking as if they understand the world. I'm quite sympathetic to this point.
- The major shift he wants to affect within economics is a renewed focus on inequality and distributional issues. In his view, economists have bought into Kuznets's view that growth is a "rising tide that lifts all boats" without enough data to justify the view. He gives Kuznets a lot of credit for gathering historical data (which subsequent economists haven't done), but thinks that Kuznets's data covered an unusual period of history that isn't representative of long-term trends. His research has produced a much larger dataset covering a larger time period, which he thinks reveals a better picture of the underlying principles.
- The First Law of Capitalism is that share of income from capital (a) equals the rate-of-return on capital (r) times the capital/income ratio (B). This explains the capital/labor income split in society, which is presumably related to inequality since that's what he cares about. (See discussion of American slavery below.)
- His thesis is that the concentration of wealth can be predicted by the relationship between the return on capital (r) and economic growth (g). If r > g, then wealth tends to concentrate and inequality rises. If not, then inequality dissipates. This is more or less a tautology, but it is still important to focus on it. He claims that there is no mechanism that prevents r > g over long periods (see the Second Law of Capitalism below).
- In my view, the real question is whether he has defined his terms meaningfully. He defines capital as assets, tangible or otherwise, that can be traded in some market. This definition excludes human capital like education and skills, because (assuming slavery is illegal) they cannot be traded. But this definition seems to me largely spurious and facile. If a worker is compensated with stock options, then this counts as capital to Piketty even though it is obviously labor income. He also downplays the significance (although he mentions) "entreprenurial labor", ie work done by business owners or other risk takers, which is realized in capital assets. For example, if I work hard for five years to produce an idea which I patent that yields me millions of dollars, why is that capital income instead of labor income? Also, lots of capital income comes not from selling assets but from leasing them. But human capital can be leased: it's called wages. Why, then, is the ability to sell assets a good demarcation criterion? Due to the shifting balance of capital income between different types of capital, this issue raises doubts about many of the historical arguments Piketty makes, which is unfortunate. It really depends on what arguments he makes from this definition of capital. If he argues that capital growth implies inequality, then he is in trouble, because capital ownership is far less concentrated than it was in the eighteenth century.
- He has great review of historical economic and demographic trends. He is appropriately humble about the certainty of the data and points out that the orders of magnitude are what matters. The point of this review is how unusual conditions since 1700, and particularly in the 20th century, have been. Growth before 1700 was very low, probably around 0.1% from 0-1700. Growth since has risen considerably, to around 0.8%. That might not seem like much since we treat growth less than 3% as some kind of disaster, but Piketty points out that even such modest growth rates as 0.8% simply cannot continue over the long haul due to their compounding nature. He also notes that political discourse tends to focus on GDP, which includes population growth, when we really should focus on GDP per capita, aka productivity growth. The former doesn't necessarily imply any increase in the standard of living, and would mislead if demographic growth slows but productivity growth stays constant.
- He next discusses how growth, in productivity or demographics, tends to lower inequality. Because diminishes the importance of inheritance, it evens the playing field and reduces inequality. Inflation can also have a similar effect, especially in the middle part of the twentieth century to erase public debts. This is perfectly sensible, which gives me a few thoughts.
- First, how much of the United States' reputation for meritocracy, growth, etc. is just a consequence of population growth since 1700? Particularly in comparison to Europe, the United States has had much more demographic growth simply because it started from a lower base. He also makes the interesting observation that following WWII, continental Europe (eg, France and Germany) grew extraordinarily quickly until around 1980, at which point they regained footing with Britain and the United States and growth rates equalized. But coincidentally at the same time, a conservative revolution instituted new policies across both regions which were incorrectly (in Piketty's view) interpreted as the cause of the change in relative growth rates. Thus continental Europe still favors leftist policies that predominated before 1980 during their period of strong growth, whereas Britain and the United States have drawn the opposite lesson and favor economic liberalisation. As I mentioned, Piketty thinks the conservative revlution was a simple coincidence and had little impact on the underlying growth trends. A similar story applies to inflation: Britain has avoided inflation even when it was highly indebted, while Germany and France experienced rapid inflation in the middle of the twentieth century. Britain is consequently less concerned about modest inflation while Germany is paranoid about inflation rising above 2%.
- Second, Piketty is somewhat undermining his own arguments against modern economics that, according to him, is too focused on growth. But if growth is what matters to reduce inequality, a growth focus serves his ends as well. His counter-argument would be that growth simply cannot continue at a 20th-century pace going forward, so we need an alternative solution.
- Third, he points out that growth is no excuse for inequality. Just because some talented, productive people can get wealthy in a high-growth regime, it does not follow that the poor are justifiably suffering nor that every talented, productive person will be duely rewarded. In my view this is clearly correct, but it neither follows that interventions to reduce inequality are justified in a high-growth regime. The question is always an empirical one about the costs and benefits of a particular redistribution policy.
- He has a long discussion of the changing nature of capital since the eighteenth century and the shifting role of public wealth. In all wealthy Western countries, capital shifted away from predominately agricultural land towards housing and industrial assets towards the beginning of the twentieth century.
- Europe: At the turn of the twentieth century, foreign assets were a significant net contribution in Britain and France owing to their colonial empires. Throughout the period 1700-1910 capital stayed at roughly 7x national income. Then the World Wars happened, and capital fell drastically throughout Western Europe. (Which he argues was a political effect, not the physical destruction of capital.) Capital fell around 1950 to roughly 3x national income--a major decline. In the decades since, capital has resurged to roughly 5-6x national income. For most of this period foreign assets and public capital are minor contributions. Public capital was a net negative in substantial measure in Britain during the war, and following the war socialist policies made public wealth 30% of national capital in Continental Europe. But these changes were short-lived. Inflation quickly reduced public debts and the liberalisation of public ownership, particular in the 1980s, returned public capital (assets minus debt) to a small amount.
- United States: The USA started from a fairly low capital/income ratio in the eighteenth century, roughly 3x national income. This reflects the abundance of cheap land and the lack of other capital accumulation, owing to a mostly immigrant population. It also explains the USA's reputation as a land where wealth doesn't matter and people can make their own way. Capital gradually increased over the nineteenth century to about 4x national income, and stayed fairly stable at that level through the twentieth century. The USA was much less impacted by the world wars than Europe, as one would expect. However if you break down capital by region, then the picture changes. The Northeast had capital levels similar to Europe by the turn of the twentieth century. Piketty has an interesting analysis here about the role of slavery. If you add slavery as capital to the South, then its capital is roughly 5x the region's income, closer to Europe. In the same period, the North's capital levels are 3x regional income. In this analysis, the South's capital stays fairly constant around 5x income until the 1950s when it drops to 4x income, reflecting the "brutality" of capitalism in the South. To Piketty, this reflects the two sides of American capitalism that have existed throughout its history: egalitarianism on the one hand, and sink-or-swim inequality on the other.
- Public debt has had many different distributive effects over time. On the gold standard where inflation was low, public debts were a boon to the wealthy. It was essentially a claim of the wealthy on the mass of taxpayers. During the post-War period, inflation rapidly eliminated debt and financed public goods, leading to redistribution from the wealthy (who lent the money) to the public. In the low-inflation regime since 1980, public debt is likely once again a boon to the wealthy, although public opinion has not realized this. Piketty also notes that inflation is not a long-term solution to this, since if high inflation is expected then rates of interest will rise to compensate. Inflation can therefore only be used as a "silent default" on public debt when it is unexpected.
- I have two major problems with Piketty's analysis of slavery in the Southern USA. He uses the capital/income ratio as a proxy for inequality, but his definition of capital does not imply this. When slavery is eliminated, in Piketty's mind this decreases the capital/income ratio because that labor is now free to work for its own benefit rather than the benefit of a select few. That's all fine and good, but only if all forms of capital necessitate ownership by an elite. This is not the case. Small farmers can own land, the public can own corporations, and so on. Piketty himself admits this in his analysis, except when it comes to human capital. In fact, small farmers could own slaves as well. But the moment that the slaves become legally free, we should consider that the wealth of society has declined? He does discuss this issue, but not in any convincing fashion. He says: "Attributing a monetary value to the stock of human capital makes sense only in societies where it is actually possible to own other individuals fully and entirely." This is plainly false. If I want to know whether it is worth it to go to college, I should compare the value of the diploma--human capital--against the costs of attending college. If I'm considering the wages demanded by two competing candidates to hire, then I had better have some way of assigning monetary value to their skills--again, human capital. And if I want to have some idea of the total stock of value in some society, I had better include the skills and knowledge of its people. Piketty's argument is insincere and unconvincing.
- Piketty's Second Law of Capitalism is that the capital/income ratio (B) equals the savings rate (s) divided by the growth rate (g). This is an asymptotic law, simply because B is the sole equilibrium value for B, and all other values converge to it. This discussion introduces some welcome nuance, unlike the discussion of slavery, where Piketty stating that capital accumulation can be a good thing because capital can benefit all of society. His concern is that "the owners of capital--for a given distribution of wealth--potentially control a larger share of total economic resources." This is a fair assessment to my mind, but raises a few other concerns. First, is this book only about ratios? Or are we also concerned about absolutes? One of the most unfortunate features of Marx is his obsession with ratios rather than absolute well-being. Second, does Piketty consider the inter-relationships between these factors? That is, some policy may reduce B but also reduce growth, which makes people worse off over the long run. Another policy may reduce B but concentrate wealth in the hands of a few, increasing inequality. Piketty consciously has not broached inequality at this juncture, but I am concerned.
- He next covers the capital-labor split. Piketty's main point is that the division of income between capital and labor has shifted over the centuries, and it is certainly not fixed by any mechanism at a stable value (contrary to conventional thought). Before WWI, capital varied around 30-45% of income. After WWI, it plumetted to 20% and has gradually increased since. The rate of return on capital has been around 4-5% annually, varying with no clear pattern, maybe decreasing slightly in recent decades to 3-4% but the data isn't conclusive. The main variation in capital's share is therefore the stock of capital, which has changed with the growth. (Growth here is both economic and demographic.)
- Piketty uses his first and second laws of capitalism to compute the share of income going to capital, and claims that the factors "are largely independent of each other". I assume here that he means that the savings rate, rate of return on capital, and productivity growth are independent. This seems like a very bold claim that is casually made and not given any evidence for. Moreover, it contradicts Piketty's own arguments later, but the upshot is that nowhere does he lay out details of how the share of capital income is determined.
- He recognizes that an increase in the capital stock will, all things equal, tend to lower the rate of return. This is simple supply and demand. Logically, a lower rate of return would tend to lower the savings rate as people substituted savings for consumption. (Piketty states that the rich gain little from consumption so will tend to save, but he doesn't explain why they should continue to save at the same rate if the return on capital is falling.) A lower savings rate will obviously tend to decrease the capital stock, but this will lower investment and potentially long-term growth. Its obvious to me that these factors are far from independent and are in fact very highly related, although its not obvious how the relationship will pan out.
- Piketty's analysis does broach the idea of rate of substitution between capital and labor, and he states that the elasticity of substitution appears greater than one historically. That is, it is economically efficient to use more capital than labor, so capital's share is gradually increasing. (This was not the case in agricultural societies, however.)
- He claims that Marx was essentially aware of the dynamics involved with the growth of capital's share of income, but Marx believed that long-term productivity growth was necessarily zero. So he therefore concluded that the capital stock would tend to infinitey, driving the rate of return on capital to zero, and forcing capitalists into non-economic conflict with one another. This is a generous interpretation of Marx. According to my understanding of the meaning of terms Marx used, this is not what he intended. (Particularly economic "crises" meaning depressions, not war, and "contradiction" being a Hegelian philosophical phrase not the common meaning.)
- Regardless, his essential point is that in a low-growth regime there is no reason that the rate of return on capital and savings rate can't stay fairly stable. This implies that the capital stock will increase and capital's share of income will steadily increase.
- We next turn to inequality, the real thrust of the book for which the first two parts were preliminary. Piketty begins once again by laying out the playing field---what is the typical division of capital ownership, capital income, and labor income? This discussion reveals what seems to me the real motivation for separating human capital and other capital: human capital is far more equally distributed than other capital. In other words, labor income is much more egalitarian than capital ownership (as defined by Piketty). This part of the book is very convincing and useful: "Ultimately, inequalities of wealth in the countries that are most egalitarian...appear to be considerably greater than wage inequalities in the countries that are most inegalitarian."
The main criticism I have is that Piketty is, like Marx before him, apparently obsessed with ratios rather than absolutes. Socialist economists tend to focus exclusively on the relative share of income rather than whether people are better off, which logically and empirically is what people tend to care about.
- The big picture of wealth and income distribution is that, before WWI, inherited wealth was extremely important. One could get far, far wealthier by marrying an heiress than through one's own labor and talent. Following WWI, this seemed to change as inflation and growth whittled away at past fortunes. Piketty worries that the world is returning to a capital-dominated world once again. The distribution of labor income follows what most would consider a "reasonably unequal" distribution: the top 10% earn ~25%, the bottom 50% earn about the same. In the wealth distribution, on the other hand, the top 10% take much more and the bottom 50% has almost nothing. Before WWI, the top 10% owned 90% of wealth. Even at the height of socialism in the West, Scandanavian countries still had 50% of wealth concentrated in the top 10%. In the USA presently, the top 10% owns roughly 70% of wealth. So unlike I had apparently been led to believe, wealth ownership of stocks etc is not equally distributed at all through society. The effects are even more extreme when comparing the top 1% to the top 10%.
- Piketty points out that understanding the differences between these two distributions is very important, because inequality is not inherently unjustified. One could make a meritocratic argument in favor of unequal labor income on the basis of merit, and similarly for the accumulation of capital within a lifetime. It is far harder to make an argument in favor of extreme inequality in inherited wealth.
- Piketty also uses wages instead of total compensation (post tax & transfer). I don't understand why this is justified. We have established a tax and social safety net for the explicit purpose of mitigating inequality. This can have knock-on effects that exacerbate wage inequality, despite total compensation's inequality decreasing. It seems methodologically unfounded to use labor income pre-tax, pre-transfer as a measure of social inequality. To give him credit, however, Piketty does seem to use individual income rather than household income (although this isn't perfectly clear) that should give a better picture.
- Similarly, what is the justification for focusing on intra-national rather than international inequality? If the poor and middle class of the rich world are stagnating while the poor of the developing world prosper, that is certainly relevant to the discussion. It may not justify the accumulated inequality of the upper centiles, but it completely changes the picture of capitalism's evolution. Piketty might respond that global data isn't readily available, but that is exactly the type of lazy work that he criticizes in his introduction. He should focus on the relevant question, not the one that is easy to answer. It seems to me that a focus on inequality within the rich world, where even the poor are objectively prosperous by every historical standard, is unjustified when so many truly poor people exist in the world.
- On income inequality, Piketty considers the traditional explanation of wages measuring marginal productivity. He is skeptical of this explanation because of the nature of income inequality within Anglo-Saxon countries (all of which are showing extreme concentration of labor income, unlike Continental Europe). In Anglo-Saxon countries, the "winners" are the top 1% or even 0.1%--inequality is highly concentrated. But there is very little to differentiate the education or skills of the top 0.1% or 1% from the top 5%. So empirically it seems unlikely that this income concentration is purely meritocratic. Moreover, Piketty points out that a "super-manager" inequality implies that the major of income growth is going to managers, individuals whose marginal productivity is inherently difficult to measure. (What is a CFO's individual productivity?) Instead we rely on board members--other executives--to determine pay, who are hardly disinterested parties.
- Piketty begins a discussion of capital inequality by noting that Kuznet's observations of the "collapse of income inequality" were not in fact due to more equal distribution of labor income, but the destruction of wealthy fortunes in 1910-1950. It was not a natural force of capitalism realizing our democratic aspirations. Prior to 1910, the top 1% of families owned 50% of capital in Europe (roughly stable) and 45% in America (trending upward). There was no downward trend apparent anywhere.
- The fundamental reason why wealth reaches such high concentrations is that historically the rate of return on capital is greater than the growth rate, or r > g. This is an empirical fact, not a theoretical necessity, but one that makes sense upon reflection. The simplest example is a near-zero growth society, with land as capital. Any income on this capital--eg renting it out--will lead to a concentration of capital. Historically, capital had a return of 5%, much greater than the historical growth rate.
- A very simplistic (too simplistic for Piketty) model of time preference shows that the rate of return on capital will equal people's time-rate preference.
- Wealth does not concentrate indefinitely because of (i) lower returns the more capital there is available (supply and demand), and (ii) random shocks that destroy fortunes, even very large ones. The former can take a very long time to operate, however. Simple models show the latter leads to a power law distribution of wealth parameterized by the difference r - g (higher returns on capital vis a vis growth lead to greater concentration).
- Wealth has not returned to nineteenth century levels due to a lack of time, progressive taxation, and unusual growth of the post-war years.
- He is not optimistic about inequality in the twenty-first century, however, because of global competition for capital that may reduce progressivity of taxation, declining demographic growth, flatlining productivity growth, and (most interestingly) more efficient capital markets which will increase the return to capital and limit its negative shocks. This last point also argues against a meritocratic justification for inequality.
- He has a long and very thorough argument that, much like capital, inheritance was a significant factor in wealth in 1910, collapsed during the wars, but is now making a comeback. He is highly critical of the life-cycle theory of saving (that people save during their productive years to fund retirement), pointing out that there is no backing for it in the data and that people have plenty of other reasons to save (security, limited consumption needs, passing along wealth, status, etc.). Inheritance always accounts for ~12% of national income annually (compared to ~20% in 1900) and is projected to stabilize at ~15-25%. Most shocking, inherited wealth as a percentage of wealth was 90% in 1900, 45% in 1970, is already back to 70% and projected to rise to 80-90%. So much for the end of inheritance. However the good news is that this inheritance is much less concentrated--currently and going forward, the top inheritances earn roughly the same income as the top laborers. This is a major difference resulting from the establishment of a middle class in the twentieth century: "A society structured by the hierarchy of wealth has been replaced by a society whose structure depends almost entirely on the hierarchy of labor and human capital."
- Piketty cautions that we can't be sure wealth won't return to the concentration seen in earlier times. He believes that if this were to happen, then it would lead to political upheaval because democratic societies would not accept that state of affairs. He goes on to discuss the source of capital income--rents. "Rent" has turned into a dirty word in modern society and economics, but this is misleading as to the role of capital income. "Rent" is used by economists to denote a market failure that leads to inefficient profit by one party over another. But capital income is not a market inefficiency--it is the just conpensation of the marginal productivity of capital in the economy. There is therefore no economic principle that supports the idea that more perfect markets would reduce the return on capital and naturally diminish the concentration of wealth. "This logical contradiction cannot be reoslved by a dose of additional competition. Rent is not an imperfection in the market: it is rather the consequence of a 'pure and perfect' market for capital."
- Piketty briefly turns to global inequality and admits "it is by no means certain that inequalities of wealth are actually increasing at the global level: as the poorer countries catch up with the richer ones, catch-up effects may for the moment outweigh the forces of divergence." It is good that he admits this, since the core argument of his book is essentially that empirical measurement shows that inequality is increasing. But he only has sufficient data for intra-national analysis of rich countries. He is able to build a coherent story about why inequality should increase, but perhaps he has missed some important aspect of competition that prevents inequality. If so, then it is certainly decreasing global inequality that causes him to miss it. It therefore seems like a very important point, and one that he doesn't dwell on enough.
- Data suggest that capital shows significant economies of scale, and larger fortunes can realize a larger return. This is yet another force of concentration. The strongest evidence Piketty presents in support of this is on university endowments, where it is clear that the larger one's fortune, the more (and better) financial managers one can afford. This leads to returns of over 10% for Harvard and only 6% on average.
- Inflation is not the tax on wealth that many claim. "Some people think, wrongly, that inflation reduces the average return on capital. This is false, because the average asset price (that is, the average price of real estate and financial securities) tends to rise at the same pace as consumer prices." Inflation is only a tax on wealth for those who keep their money in cash--it is not hard to develop a low-risk, value-preserving portfolio hedged against inflation. Even worse, those with more wealth are often those best situated to accept a small amount of risk and with the resources to properly diversify against inflation. So the net effect of inflation, although complex, can actually act against the less wealthy rather than in their favor.
- Piketty next turns to what society should do about inequality in the final part of his book. He begins simply by analyzing the evolution of the state in the 20th century: consumes ~40% of national income, most of which is spent on pensions and other social services (only a small part on defense and outright redistribution). Both those who want a return to a 19th-century small government and those who want an expansion of government similar in scope to the 20th century are going to be disappointed. The "modern social state" is here to stay, but it will not double in size again.
- This section then takes a turn for the worse. Much of the remainder of the book is not founded on empirical work that is the bedrock of the first three parts. Instead, Piketty merely states his opinion in strong terms. This is unfortunate in two respects. First, for the people who are disinclined to like his book for ideological reasons, it gives ammunition to throw out the first three parts without rebutting his research. Second, for the people who are sympathetic to his view, it may give a wholly unjustified veneer of respect simply by association. In my opinion, the book would have been improved if the final part were simply deleted. Regardless, here are my thoughts:
- He begins with a superficial and unconvincing account of public services.
His treatment of education is terrible. He starts by decrying the inequality of access to higher education, especially in the USA. But he has no intelligent analysis of the proper role of higher education or the source of inequality. Should everyone go to college? He treats the problem entirely as one of public financing rather than of the quality of public education. To be sure there are financing issues at play, but tuition is already heavily subsidized and grants and scholarships reduce the sticker price drastically. He ignores this fact. He also simply dismisses equity financing without any stated reason.
But the more serious issue is the blind focus on financing to begin with. The simple fact is that college is hard for many people and most don't finish. Many of the people not attending college now are those who would struggle to complete their degree and would therefore get little benefit. This has two obvious causes: inequality of realizable potential, limited by either nature or circumstance, and the terrible state of much of lower education. Piketty does not discuss this, but rather seems to assume that if only the public would pay for it, everyone would see the same benefits from a college degree. This is just wrong.
This gets to the last point, which is how valuable a college degree would be going forward if everyone had one. Piketty's own data shows that education does not decrease income inequality, yet he somehow implicitly believes that universal access to higher education is a democratic imperative. He ignores the human capital vs signalling debate completely. He also doesn't discuss the evolving labor market and what jobs will benefit from a degree--stated differently, does a college degree inherently increase productivity and therefore wages on its own, so that income inequality is a product of unequal education? Or is the availability of jobs in the labor market inherently unequal (ie, with technological progress and globalization), and these jobs just happen to go to those with college degrees? In the former world there is an argument for greater access to higher education, although still not necessarily one for public financing. But in the latter world increased access to higher education would make no difference and is simply a waste of money.
This is a very interesting debate and Piketty doesn't even acknowledge it. This is a very low point in the book. Ultimately it makes me skeptical that he's pulling a fast one in other places that I missed and really weakens his argument.
- His tax proposals are not based on empirical data. He suggests that there should be a global, progressive tax on wealth. The first point he makes is fair--we should have a proper, transparent accounting of wealth so that democratic debate can proceed from facts not speculation. I don't see a real argument against this position. His specific tax proposals are less well reasoned and, frankly, not worth commenting too much upon. He nowhere considers the incentive effects that his proposals would cause. He simply asserts as a matter of faith that taxing capital progressively would not lead to reduced output or other perverse incentives that reduce efficiency. This seems wrong on its face--a progressive tax will encourage people to exploit absurd corporate structures to hide their wealth in smaller units, where it will be less productive for the economy but more lucrative for the owner. Piketty does not address such concerns.
I listened to a radio interview with him where came across as far more reasonable. His position was simply that he doesn't know the future nor the optimal policy, but he is convinced that the current wealth inequality in the USA could be improved upon (i.e., "is 2% of wealth owned by the bottom 50% the social optimum?"), and his proposal is one way forward. Rather than beat up on his suggestions in the book, I'll simply leave it there.