Tuesday, March 25, 2014

Capital in partial equilibrium

Image source

Review: Thomas Piketty, Capital in the Twenty-First Century

Piketty's primary contribution is to provide an impressive array of data on wealth and income, for several countries, beginning as early as the 1700s in some cases. Note that he does not examine consumption data. The book is an impressive feat and certainly deserves attention, as the facts Piketty provides are crucial to discussions of the evolution of capital and economic inequality in the rich countries. Many reviews have been very positive; there are a lot of positive things I could say about it, but I will leave that to others. The book suffers from some fundamental flaws; in short, while it is heavy on data it is light on serious economics. Readers will find themselves wading through hundreds of pages of opinion and ideological quips, not economic analysis, with interesting charts scattered throughout. The firehose of data can be overwhelming, which may explain why some reviewers internalized his arguments uncritically. Piketty's accomplishments with data collection are admirable. But a book of this size, with the title Capital, should include some economics.

Piketty's data on inheritance are the most interesting and persuasive to me. Inheritance still matters and plays a nontrivial role in the wealth and income distribution. Reducing wealth inequality over time, should we decide to do so, will require serious attention to the issue of inheritance, which more than any other issue lacks a tie to meritocracy (but that does not mean incentives stop mattering!). There may be other arguments, not based solely on inequality, for thinking about inheritance. That said, not all capital is created equal, and the book could have benefited from some focus on distinctions between capital types--particularly in the context of inheritance.

Most of the analysis in the book is more about accounting than economics. Piketty takes nearly everything as exogenous then divides things arithmetically. His ubiquitous r > g heuristic takes both sides of the inequality as given for almost the entire book. Lines like "the richest 10 percent appropriate three-quarters of the growth" (297) enable lazy readers to avoid thinking about what actually determines income. Language about "appropriation" suggests that we live in an endowment economy, as does the claim that post-World War I wealth inequality fell "so low that nearly half the population were able to acquire some measure of wealth" (350). Endogeneity, anyone? Taking income as exogenous leads to other large problems with inference, such as the claim that "meritocratic extremism can thus lead to a race between supermanagers and rentiers, to the detriment of those who are neither" (417). Piketty does not consider the possibility that this race results in more income than otherwise, nor does he consider the notion that an increase in the bargaining power of elite executives could actually come at the expense of capital owners rather than workers. I'm not making an argument for either here; I'm simply suggesting that Piketty's ideological quips don't deserve the certainty with which he delivers them. Models with endowment economies have their purposes, but a 600-page book should be able to relax such strict assumptions. His criticisms of mathematical economics (32, 574) are not surprising given that he relies so heavily on assumptions and mechanisms that would be highly vulnerable to criticism if they were forced into the transparency of a formal model.

This kind of fast-and-loose economic reasoning pervades the book. Piketty attributes the rise of the "patrimonial middle class"--the great home-owning middle class of developed countries--entirely to the rise of capital taxation (373). It's perfectly reasonable to argue that taxation played a role, but it's absurd to give taxation all the credit without further analysis. Piketty relies on this shaky causal claim for his central thesis; presumably unbounded capital accumulation wouldn't be a problem if everyone owned some. Denying that economic forces played any role in bringing wealth to the middle class helps Piketty claim that inequality will spiral out of control and leave us all in poverty unless serious tax reform is effected. This argument also requires him to assure readers that there are no tradeoffs associated with capital taxation: "It is important to note that the effect of the tax on capital income is not to reduce the total accumulation of wealth" (373). We also learn that there is "no doubt that the increase of inequality in the United States contributed to the nation's financial instability" (297). No identification problem here, folks; causal inference is easy! The book is littered with extremely strong claims like these, despite the existence of good reasons to at least be skeptical of some of them. Maybe Piketty is right about these things, but he has not shown it here; and even if his considerable collection of charts and tables is enough to dazzle most reviewers into fawning submission, the data are not sufficient for demonstrating his strong conclusions.

Piketty's data on the rise of middle-class capital ownership raise an important point. A key theme of the book is that poor people don't own productive assets, so they must rely entirely on labor for income. But is taxation and redistribution the only way to address this situation? This poses a difficult question for those who oppose some form of privatization of government retirement programs. One cannot simultaneously claim that owners of capital stand to gain absurd riches in coming decades and that privatization and choice for Social Security is a terrible idea.* This is not the only possible alternative to taxation, but it is a reminder that one way to treat the problem of poor people not owning stuff may be to help poor people, well, own more stuff. But Piketty simply asserts that "only a progressive tax on capital can effectively impede" increasing wealth concentration (439). More generally, Piketty decries the ability of those with large fortunes to access opportunities for higher rates of capital return than those with smaller starting funds, but he makes no mention of the fact that this is due in part to laws banning small investors from participating in alternative investments. By law, if I want to invest in a startup, I can only do it in undiversified ways (like starting my own firm or investing in a friend's). We don't need higher taxes to help lower classes invest better.

Another key weakness of the book is that, like much of the popular debate on inequality, it focuses almost entirely on higher moments of income and wealth distributions while making only minimal effort to provide context about absolute levels of income and wealth. Piketty compares inequality over the centuries, noting that it is returning to pre-World War I levels then claiming that "the poorer half of the population are as poor today as they were in the past" (261). This is only meaningful in relative terms; as Piketty briefly mentions a few times (but does not emphasize), the poor of 2014 are much better off than even average earners from previous centuries: "With 5-10 times the average income in 1800, one would have been in a situation somewhere between the minimum and average wage today" (415, see also 88), and even Piketty admits that this claim relies on dubious adjustments for inflation (how much did a car with air conditioning cost in 1800?). The truth is that most of today's developed-country poor are astronomically, unquantifiably better off than almost anyone from 1800, which raises the question of why Piketty sounds alarms about inequality reaching previous levels. He briefly acknowledges modern international context in chapter 12, but he examines only wealth inequality,** which is tricky, and makes no mention of global income inequality, which has been declining. In any case, Piketty has strong incentives to tell us that the only thing that matters is inequality within rich countries (432)--even if the poorest Americans are better off than most of the world.

This is the great failure of the inequality alarmists generally: a myopia centered around rich countries that have seen massive growth in purchasing power for everyone. For within-country inequality to become a public policy priority, those concerned about it need to make a much stronger case. This isn't just about whether society is totally meritocratic--obviously it's not, as Piketty argues in multiple places. The problem is that Piketty has not performed an analysis of optimal inequality, choosing to rely instead on a lot of straw man arguments and vague references to "democracy." He has provided a lot of data that demonstrate that inequality in countries that produce the top 20 percent of global output is on the rise, and he has suggested policy remedies (that he claims are basically costless), but he has not made a serious attempt to convince fence-sitters that this issue should top the policy agenda.

It is hard to believe that Piketty's predictions for the future--on which his policy prescriptions rely--are much more than undisciplined speculation. He does not have a model. He has shown us historical data from which he has drawn inference using accounting-based counterfactuals and a lot of hand waving. He has made forecasts for the future paths of income and returns to capital--both of which are basically exogenous processes in the text. On these forecasts he hangs his broader predictions. His predictions are worth noting because he is a capable scholar with a tremendous mastery of the data, but his hand-wavy approach to dismissing (or ignoring) weaknesses in his framework limits the ability of his speculations to convince those who don't already agree.

Economists who write books like this have an opportunity to educate the public in economic reasoning. Thinking about macroeconomics is hard. Piketty's methods and rhetoric suggest to readers that macroeconomics is easy. He can dismiss all objections with a wave of his hand. He can ignore policy tradeoffs and potential general equilibrium effects by simply asserting them away. Someone should count how often he uses the terms "no doubt" and "clearly" when drawing huge, highly debatable conclusions about almost everything (e.g., 511). By referencing only charts (if even that) for many of his claims, he is feeding the sloppy and destructive "this one chart proves...!" fad that has spread in the blogosphere; a chart is never sufficient to make causal claims or demonstrate optimal policy. In this sense, Piketty does his readers a disservice. He should have asked them to think harder instead of just gazing at graphs. He should have accompanied his facts and predictions with serious normative arguments instead of assuming his readers automatically share his deep concerns about how the world's top 20% are faring compared to the world's top 0.1%. He should have explained the reasons many economists prefer low taxes on productive capital other than land; even if he finds such arguments unpersuasive, he robs readers of the chance to consider them when he blatantly accuses those economists of intellectual dishonesty (514).*** He should have acknowledged the massive "causal density" problem in macroeconomics and shown readers how economists investigate causal relationships and, more importantly, identify the limits of their knowledge. Instead, he preaches to the converted and to those who are easily overwhelmed by a deluge of charts, knocking down straw men but avoiding the hard questions. The book is many things, including an excellent resource for stylized facts, but serious economic analysis it is not.

Capital is tremendously informative, and Piketty's use of literary anecdote is a nice touch; but ultimately this is a chart book, with plenty of economic data but very little economics.

*Piketty devotes a grand total of two paragraphs to this idea (488-9), making two weak, unimaginative arguments against the policy. He cannot imagine a way to transition from PAYGO to private investment accounts, and he believes that trusting retirement to the markets amounts to "a roll of the dice." We know enough about optimal retirement portfolio choices to do better than dice-rolling. For someone in my generation, watching a chunk of my paycheck go toward a program based on the silly assumptions underlying Social Security is the real gamble. In any case, if Piketty truly believes that giving everyone a chance to become an owner of capital is a dice roll, he must explain why he is so certain that capital income is sure to explode in coming decades.

**His discussion of wealth inequality even includes an awkward admission: "The average global fortune is barely 60,000 euros per adult, so that many people in the developed countries, including members of the 'patrimonial middle class,' seem quite wealthy in terms of the global wealth hierarchy" (438). But the middle class doesn't just "seem" wealthy. The developed world's middle class is wealthy by any objective standard, which emphasizes my point that inequalities within developed countries are much less disconcerting that true global poverty.

***Aside from this silly accusation, Piketty never mentions optimal taxation literature aside from a handful of his own papers. His central recommendation of a global wealth tax does not read like a result of optimal taxation analysis. He claims that "it is hard to think of an economic principle that would explain why some assets should be taxed at one-eighth the rate of others" (529), as if "elasticity" and other drivers of optimal tax models are not economic principles.

Monday, March 17, 2014

"A good time to be a water attorney in Texas"


From the NYT:
66 Texas towns and electric utilities [were] exempted from a [water] cutoff for health and safety reasons, even though hundreds of farmers and others who lost their water held more senior rights.

And here's an issue I heard about a lot while growing up:

In Colorado, officials in the largely rural west slope of the Rocky Mountains are imposing stiff restrictions on requests to ship water across the mountains to Denver.

The Western Slope just might win that one, but it may not matter. The Colorado River is crucial to five states. And that's not the only important river that starts in Colorado: ever heard of the Rio Grande? I've seen the words "flush: Texas needs the water" scribbled on bathroom walls in the San Luis Valley, and many in the Colorado basin feel about the same way about Phoenix, Las Vegas, and southern California.

Here's a typical observation about the urban/rural split:

Farmers and others downstream complained that they were surrendering their water while Austin residents continued to wash their cars, groom golf courses and water their lawns.

I alluded to this in my note on landlords: In the long run, does anyone think that the legal water claims of rural interests have a chance against the cities?

I wish I knew more about this topic. It's at least clear that the current legal and market framework will not withstand the coming stress. After hearing about officials in big cities begging people to reduce their toilet flushing but keeping the golf courses running, it's hard not to think that using prices would be an improvement.

Saturday, March 15, 2014

Pecuniary externalities

Bloomberg reports that Blackstone is reducing its purchases of houses, which peaked at $100 million worth per week last year. From the article:

President Barack Obama credited the investors for helping put a floor under the plunging housing market and consumer advocates such as the National Community Reinvestment Coalition later blamed them for soaring prices in some cities.

Saturday, March 8, 2014


Thanks to a happy accident at Amazon, I received my preorder copy of Thomas Piketty's Capital early. I started reading it yesterday. So far my impressions are mostly good, and I think I will learn a lot about the data and about how to think by reading this book.

Important people don't rely on Amazon glitches to get books early. Ryan Avent has already written some comments about the introduction. I'll comment on two points. First, as Avent notes, Piketty does do some obligatory economist-bashing in the intro. This sort of thing is everywhere in the popular press (and the blogs). My hunch is that it stems from academic habits of selling one's work by claiming that it is a nontrivial improvement on existing literature*. Unfortunately, nonspecialist readers are not likely to understand that. I'm not convinced that telling lay people that the bulk of economists are stubbornly clinging to deliberate ignorance is a good thing on net. Whatever internal fights exist in the profession, probably everyone is better off if the public doesn't think economists are idiots than if it does. Tone it down, people.

Second, despite his complaints about theory, Piketty notes that he will be using some simple toy models to illustrate points. It's good to give nonspecialists some idea of how models can leverage our intuition and help make sense of data. As most people probably know, Piketty's primary framework is the relationship between r and g, where r is returns to capital and g is aggregate income growth. When r>g, owners of capital receive a growing share of income and accumulate wealth rapidly. This is simple but profound--but it is partial equilibrium. Returns to capital are endogenous, which is the point of Karl Smith's comments that prompted this post from me. For some forms of capital, there is an adjustment mechanism that can push returns down. Output growth is endogenous, too, and it's not hard to image a single force that both pushes r down and pushes g up. Hopefully Piketty will shed some light on this.

UPDATE: My full review of Capital is here.

* A less charitable reading of this behavior is that each specialist wants to secure reader loyalty by convincing nonspecialists that all other specialists are quacks. I suspect this happens a lot in the blogosphere, but I have no reason to think it matters for Piketty.

Sunday, March 2, 2014

Are landlords the future global plutocracy?

Karl Smith has two posts about land that I have been reading over and over for weeks. I'll start with this one. In response to recent work by Thomas Piketty, Smith argues that we should carefully distinguish between land and other forms of capital when talking about wealth. He starts with this chart from Piketty:

Smith's observation is that the surge in wealth of capital holders after the 1950s is driven primarily by housing (and, therefore, land). Using the example of Paris, he notes that legal restrictions on architecture and zoning render land extremely expensive, so "only those willing to pay the highest rents or mortgages can live in Paris and the rest are pushed to the suburbs." The value of Parisian land "does not stem from the creation of new structures, plant or equipment." It stems from severe supply constraints. It's hard to create more land. It's impossible to create more land in downtown Paris. 

Growing up in western Colorado, the richest people I knew were landowners. They came from old agricultural families that had been steadily selling off little chunks of land to movie stars, fashion moguls, and subdivision developers. Maybe other people think of rich guys as the ones who work in banking and drive a new Porsche, but the rich guys I knew drove 15-year-old pickups and wore cowboy hats. They were rich because the supply of private land near decent airports and skiing is fairly limited, and their parents had given them a chunk of that supply. What they haven't sold to Ralph Lauren will be passed on to their own children, who can sell a few more acres and pass on the rest. So Smith says, "irreproducibility and inheritance are the essential features of land in classical economics." And those who didn't inherit face ruthless pecuniary externalities.

Ryan Avent took issue with Smith's post, noting that US data look a bit different from French data. I think he makes a few good points to which I'll return, but Smith effectively rebuts him by noting that the data hint at fairly strong responses of capital wealth to capital returns, while land wealth just steadily grows. He uses the dot-com bust to illustrate that capital wealth is destroyed when it grows too big relative to national income--"this automatic correction mechanism is not an anomaly but a fundamental feature of capital." Smith ends with:

In the wake of the subprime crisis, I understand the temptation to rally against big banks and global finance. However, Lehman Brothers is dead. Sam Zell, founder and CEO of Equity Residential, is still alive. This is not an accident. The future does not belong to high flying titans. It belongs to dogged men and women who squirrel away rent checks when times are good, and buy [their] home when times are tight. This is the tyranny of land. Ignore it at your peril.

Now back to Avent. His argument is that land is more like productive capital than Smith allows, as the reason for its high demand in key locations is that it facilitates transactions, both within firms and between firms and workers. It "reduces the cost of exchange" because it allows workers to locate in productive cities. I like this heuristic, and to the extent that it's accurate Avent's point is reasonable. A good example that illustrates the point is Goldman Sachs, which has been moving an increasing share of its operations to Salt Lake City in a strategy that amounts to land price arbitrage. The cost of fragmenting production across regions is now low enough that the price of land in New York kills the advantages of locating workers near the main headquarters, at least for some tasks. You can work out the marginal consequences for New York landlords. More generally, even though land is a production factor with fixed supply, production functions have substitution elasticities. Skype can accomplish a lot of things that used to require physical presence; in that sense it is a substitute for city land, and substitution works against the notion that fixed factors must command a growing share of income over time (this is literally textbook macro; see Romer p. 42).

But I suspect it's not enough. There are reasons to believe that cities aren't going away and that where workers locate will continue to matter. Some jobs that used to be done only in New York can now be done elsewhere, but they must still be done in a city. And land is more than Avent suggests. It provides its users with collateral services and enters utility directly. It depreciates very slowly, more like human capital than physical capital, and it's easy to bequest. And Americans' taste for living space seems to be growing, so there's no reason to assume that demand for land is going to abate.

Avent also notes that breaking the policy grip of landlords will not necessarily result in giving workers a larger share of income. Rather, the gains would accrue to owners of intellectual property. Again, though, intellectual property is much less conducive to intergenerational wealth accumulation than is land.

Izabella Kaminska got in on the discussion as well, emphasizing the following:

In more established economies, however, land value is being increasingly determined by exclusive and finite access rights, and/or improvement restriction factors. Limits on how high you can build, how much of the green-belt can be urbanized, and so forth. 
This is what makes access-based land values so pernicious. There is no correction mechanism to offset the ever incremental returns.
What's really going on here is a complicated political economy tradeoff that no American politician will touch; but Eric Crampton points to some leaders Down Under who are being frank with voters. Australia's Matt Cowgill:

There's a trade-off at play here, one that can't be wished away or ignored. With a growing population, you can't restrict rising density in established suburbs, prevent sprawl on the urban fringes, and prevent housing from being unaffordable. Pick two of the three.

And New Zealand's Bill English:

So if Auckland wants to grow now, it has to grow out because you don't want it to grow up. Now that's a fair choice, but please don't stop it from growing out as well, otherwise we'll get another few years of 15% house price growth.

After you've recovered from the shock of seeing politicians tell voters that tradeoffs exist, solve for the equilibrium here. Apparently nobody likes sprawl and nobody likes density. Is there enough desire for low real estate prices to overcome these preferences? Cries for better zoning aren't going to come from incumbent land owners, are they?

East Coast dwellers may be unaware of another important factor in land values that will increasingly plague a few key US cities in the future: water. This was the dominant political issue in western Colorado. There is a perception--probably accurate--that cities have the political power to secure water, while rural areas don't. Phoenix, Las Vegas, and Los Angeles will always have water, a fact that will only supplement the forces that are drawing people into cities. Some of those places may not face the steep cost of living faced by the mid-Atlantic and Northeast, but give it a few years and they just might.

One last point is that real estate prices have asymmetric consequences. High prices are good for owners, but low prices may not provide a lot of benefits to non-owners. Remember that pecuniary externalities can bite in the presence of incomplete markets. Homeownership didn't surge when house prices collapsed. A lot of non-owners are credit constrained. Breaking into the landlord class isn't easy if prices only fall during credit crises.

I think the implication of Smith's argument is that we're marching the wrong way if we focus only on returns to owners of capital but ignore land. Capital accumulation has a way of self regulating, but land doesn't. If wealth concentration is pernicious, land wealth is more pernicious than capital because it is more likely to be durable. This supplements the standard list of reasons that Ramsey models typically imply high and low optimal taxes on land and capital, respectively, as Smith notes.

It is worth noting, though, that it has become increasingly easy to own small quantities of both land and capital, through REITs and stocks, and these holdings can be expanded gradually through tiny marginal investments. In my view, before we start trying to design a tax system to preempt Piketty's hunch about the future of capital's income share, we should see if there are ways to get more of the working class into land and capital ownership. For example, it seems to me that you can't believe both that returns to all forms of capital will continually outpace returns to labor in the future and that giving people the option to invest their Social Security accounts in equity securities is bad for them. This is not a call for policy, but if there are policy nudges that can make more regular people into owners of stuff, that may be a better way to go than to complicate the tax-and-transfer system further.