UPDATE: I’ve been told that I’ve written sillier things.
I’m meeting with some colleagues over the summer to discuss Thomas Piketty’s new book, Capital in the Twenty-First Century. We’re reading two chapters every week, and I’m going to post my thoughts as we go along. You can think of this as a slow-motion book review.
Chapter 1 provides an overview, and Chapter 2 mainly discusses concepts, so it’s premature to comment on Piketty’s main argument. Indeed, as I go along, I’ll probably need to revise initial impressions in light of later chapters. That said, here are some comments on the first two chapters.
1. Piketty places himself in an intellectual tradition of economists who are concerned with growth and inequality, above all Ricardo, Marx, and Kuznets. Piketty praises each for focusing on these issues, for attending to empirical realities (particularly Kuznets who puts together the first usable data set), and for proposing a theory about the long-term trend of inequality. And he criticizes each for being, well, wrong, and specifically for extrapolating on the basis of inadequate data.
And my first thought is how these comments rebound on Piketty’s own project. One key issue is to what extent can he reliably extrapolate based on his own (much better data). Kuznets was wrong in extrapolating a downward trend in inequality on the basis of his limited data, and so we have to wonder about the basis for extrapolating an upward trend in inequality on the basis of Piketty’s significantly better but still limited data. One notable point is that Piketty extrapolates low long-term economic growth (a key variable in his theory) on the basis of many hundreds of years of history rather than going back to (say) the industrial revolution or later. Is that the right assumption?
2. I suspect, based more on the reviews than on what I have read so far, but partly on that as well, that we will see two Pikettys, or rather two theses:
a. The weak thesis: there is (contra Kuznets) no law that inequality declines with growth; it could go up or down. No one knows.
b. The strong thesis: there is a law that (also contra Kuznets) inequality will increase with growth.
I’m interested to see how this plays out. The weak thesis may contradict some popular right-wing misconceptions but as an academic theory it’s pretty, well, weak. As Piketty acknowledges, Kuznets himself did not consider his own law anything more than speculative. The strong thesis is much more exciting and important, but may end up being just as speculative as Kuznets’.
3. Another issue that I will revisit as I learn more is what exactly we should think about inequality. Piketty is not very clear about this so far. He hints that inequality could lead to very bad outcomes–revolutions, or an oligarchy, or social instability–or maybe it is inherently objectionable. A very old question is whether inequality is worth tolerating for the sake of growth if the bottom rises. Piketty is already clear that we can’t assume that this is the case; nonetheless, we need to know what we think about this before we propose reforms that might improve equality but suppress growth.
4. Finally, I was struck by an interesting observation that Piketty makes in passing about development. He argues that the development success stories–China, South Korea, Japan, and so on–benefited from free trade but not from foreign investment. Meanwhile, foreign investment has actually harmed many countries, especially in Africa and maybe Latin America, because poor members of the public resent foreign wealth in their midst and support populist governments that expropriate foreign investment and destroy the economy. Is there empirical evidence for this argument?, I wonder. If it’s right, it has important implications for how best to help developing countries (namely, don’t invest in them!).
The New York Times reports an upsurge in the number of unaccompanied minors who are entering the United States illegally across the southwest border. I found the data here, and produced the graph above. The data are through May, and so for 2014 I indicate the actual number as well as an estimate through the end of the fiscal year.
What is the source of this wave of illegal migration? The obvious explanation, documented anecdotally in the Times article, is that people living in (mainly) Central America believe that the U.S. government will not deport minors, based on President Obama’s implementation of the Dream Act criteria and other immigration policies.
The administration blames rising crime rates in central America. But while Honduras remains a very dangerous place, homicide rates peaked in El Salvador and Guatemala years ago, and have been declining over the period in which the migration of unaccompanied minors to the United States rose.
A nice illustration of the limits of congressional power. Congress passed the National Defense Authorization Act of 2014, which provided that if the president wants to transfer Guantanamo detainees to foreign countries, he must give Congress 30 days notice, so that it can raise a stink. The president refused to do so before the prisoner exchange since the negotiations took place in secret, and he could not afford the risk that even if he could inform Congress secretly, a leak would occur.
The transfer may have violated the substantive provisions of the statute as well. The statute provides that a detainee can be transferred only if he is no longer a threat to U.S. national security or the transfer is pursuant to a court order. But there was no such court order, and the five Taliban members who were transferred appear to be dangerous people. So much for the statute.
As Marty Lederman notes, the president paved the way with a signing statement that said that the law conflicts with “separation of powers principles.” A nice echo of the Bush administration, and a reminder that the imperial presidency is bipartisan.
A good case for a class on remedies, it features an unusual harm. “Amy,” a victim of child pornography, is emotionally traumatized when she learns that thousands of people view images of her on the web. Paroline is one of those people, and she seeks restitution from him. It’s not entirely clear to me that he even contributed to the harm, but if he did, determining how much money he should pay her on account of that contribution is a challenge. What is unusual about Amy’s injury is that it requires a large number of contributors but above a threshold additional contributors do not increase the harm. Some thoughts on the case at Slate. There is a good academic paper on the topic, making a similar argument, by Jenna Sheldon-Sherman.
A former student who works for Social Security responds to yesterday’s post:
I read your blog post today on Social Security overpayments and wanted to respond briefly (in my personal capacity, of course, and so my opinions don’t necessarily represent Social Security’s official policy on the matter). I think you’re seriously mistaken about the source of the agency’s claim to recoup alleged overpayments. Rather than a common-law argument of unjust enrichment, the Social Security Act and implementing regulations offer specific authority to collect overpayments using a variety of methods, including tax refund offsets.
The source of authority for collecting overpayments from claimants who received Title II benefits lies in Section 204 of the Social Security Act (codified at 42 USC 404). That statute allows Social Security to prescribe regulations to implement the overpayment process. 20 CFR 404.502 details the specific means by which the agency may collect overpayments. 20 CFR 404.502(b) covers situations when the individual overpaid dies before the adjustment is made, and 404.502(b)(3) specifies that such adjustment can be made by “withholding a lump sum or monthly benefits due any other individual on the basis of the same earnings which were the basis of the overpayment to the deceased overpaid individual”. In the case of the Post story, both Sadie Grice and Mary Grice were entitled to benefits based on the father’s earnings record (some combination of widow’s benefits, mother’s benefits, or child’s benefits, I would guess, though I don’t know for sure and have no personal knowledge of the case other than what’s in the Post article). Therefore, 404.502 would arguably allow collection of the overpayment from Mary because she received benefits on the same earnings record as Sadie did, and Sadie was mistakenly overpaid. 20 CFR 404.520 then permits the overpayment to be referred to the Treasury if the debt is certain in amount, past due, and legally enforceable, for a tax refund offset.
If the benefits in question were child’s benefits that Mary was entitled to solely because she was a minor child of an insured individual (20 CFR 404.350), that’s her entitlement, not her mother’s even though her mother acted as her representative payee because she was 4 years old when she gained the entitlement. I’m a little more unclear on how it would work if Sadie was receiving mother’s benefits (20 CFR 404.339) because of the minor children, but that presumed she wasn’t entitled to widow’s benefits in her own right. In any case, while the propriety of trying to collect such old debts is arguable and the agency should certainly be held to prove the overpayment occurred on the stated earnings record in the specified amount, I feel certain the agency is not relying on a common-law theory of unjust enrichment to do so.
Point taken, but I still believe that this program would not be upheld in court (and perhaps the author of the email agrees in part). I haven’t look at the regulations, but the regulations must be consistent with the statute, and 42 USC 404(b) contains this provision:
In any case in which more than the correct amount of payment has been made, there shall be no adjustment of payments to, or recovery by the United States from, any person who is without fault if such adjustment or recovery would defeat the purpose of this title or would be against equity and good conscience. In making for purposes of this subsection any determination of whether any individual is without fault, the Commissioner of Social Security shall specifically take into account any physical, mental, educational, or linguistic limitation such individual may have (including any lack of facility with the English language).
So the common law (or technically, equity) is relevant after all. And this means, among other things, the doctrine of laches would apply (even if Congress has suspended the statute of limitations); and equity would frown on (as they say) recovery from Mary on account of payments that Sadie may not have no turned over to her, may have spent on herself or the other kids, and so on. Clearly, as a four-year old, Mary was without fault; and the last sentence of the provision would seem to further strengthen her argument. Finally, consider again the policy of going after the oldest child alone for overpayments that benefited all the children if at all; this too is not explicitly authorized by the statute, and would, I am fairly certain, not be considered equitable by a court.
According to a report in the Washington Post, the government does just that.
A few weeks ago, with no notice, the U.S. government intercepted Mary Grice’s tax refunds from both the IRS and the state of Maryland. Grice had no idea that Uncle Sam had seized her money until some days later, when she got a letter saying that her refund had gone to satisfy an old debt to the government — a very old debt.
When Grice was 4, back in 1960, her father died, leaving her mother with five children to raise. Until the kids turned 18, Sadie Grice got survivor benefits from Social Security to help feed and clothe them.
Now, Social Security claims it overpaid someone in the Grice family — it’s not sure who — in 1977. After 37 years of silence, four years after Sadie Grice died, the government is coming after her daughter. Why the feds chose to take Mary’s money, rather than her surviving siblings’, is a mystery.
If, as I think, the government’s claim is based on common law principles (as opposed to specific statutory authorizations), it’s claim is probably invalid. The issue is not the age of the claim–as the article notes, Congress eliminated the statute of limitations for government claims of this sort. Nor is this a case of a child inheriting the debts of a parent. In a way, this is a run-of-the-mill unjust enrichment case. Consider this simplified version.
The government overpays Sadie on account of her Social Security benefits. Sadie, unaware of the error, cashes the check and gives the proceeds to Mary. Neither knows of the error. The government is entitled to obtain restitution from Sadie or Mary.
Putting aside the time lag, this is our case. But there is a problem for the government. It has no evidence that Sadie spent the money on Mary, as opposed to her other children, or for that matter on herself. (Maybe the law required Sadie to spend the money on her children, but if she didn’t, the government has a (defunct) claim against Sadie, not against the children.) According to the Washington Post:
The government doesn’t look into exactly who got the overpayment; the policy is to seek compensation from the oldest sibling and work down through the family until the debt is paid.
I can almost imagine why this crazy policy was (presumably) approved by government lawyers. If you have a valid restitution claim against multiple people, you can go after whoever you want; you don’t have to go after all of them. The purpose of the payments was to benefit the children, and so overpayment would have benefited all of them, if it was spent on them. And so if it’s easiest to start with the eldest sibling, that’s a policy judgment that is consistent with the law.
But under the principles of unjust enrichment, the government needs to prove that Mary actually received money from Sadie, or that the money was spent on her in a way that made her better off. It can’t.
This is a very good book, which milks insights out of two dead-ish fields–torts and contracts–and one that has never come to life–restitution. I’m impressed by the creativity of the authors. My favorite of their many ideas is “anti-insurance”–where two contract parties agree that if the promisor breaches, he must pay damages to a third party (who pays for the privilege) rather than to the promisee. This mechanism prevents the promisee from relying excessively on performance while preserving the promisor’s incentive not to breach unless efficient to do so. It’s called anti-insurance because it eliminates the promisee’s right to damages (a kind of insurance) in case of breach.
A meditation on our political culture.
I argue for such a statute here.
In The New Republic, Noam Scheiber advocates price regulation and subsidization of legal services, so as to counter the advantages that the rich enjoy in our legal system. I criticize his argument here. My argument begins:
In days of old, litigants would hire champions to assert their claims in trial by battle. The rich could afford more skilled warriors, and so were more likely to win their lawsuits (and less likely to lose their heads). One could imagine proto-liberals at the time proposing, quite sensibly, that everyone receive a champion of equal quality. Whether this would have improved justice is another matter.
As Andrew Rudalevige notes, in response to fellow Monkey Cage inmate, Erik Voeten, presidents do not exercise power only through executive orders. Moreover, many executive orders are trivial while others are important, so one can learn only so much from their absolute numbers. For that reason, it is important to look at other measures of executive power. The graph above shows the number of pages in the Federal Register each year to provide a rough sense of regulatory activity of the executive branch. For many reasons, this measure is extremely crude, but it reinforces two important points: that executive power has increased dramatically since World War II, and that in recent years any particular president such as Obama or Bush does not act much differently from his predecessors.
As I note in a comment on NYT’s Room for Debate, the “executive order” imbroglio coming out of the State of the Union speech is strange. The White House told newspapers before the speech that the president planned to sling about executive orders like Zeus with his thunderbolts, and they duly reported it on their front pages. Republicans duly exploded with outrage. The speech itself has a single mention of executive orders (“I will issue an executive order requiring federal contractors to pay their federally-funded employees a fair wage of at least $10.10 an hour”). The president continues in this vein, saying that he is going to do a bunch of other extremely minor things using his existing statutory authority, though it would be better if Congress would chip in with some legislation. The resulting controversy about presidential power is entirely manufactured–by both sides. Maybe the president’s strategy was to look fierce to his supporters while not actually doing anything that might get him in trouble with Congress.
In my response to their thought-provoking paper, I argued that the supposed fallacy that Eric and Adrian identify depends on empirical claims about judicial behavior in a way that they denied. My point was that although the targets of their critique may make different assumptions about what motivates judges and what motivates political actors in the other branches, those assumptions are not necessarily “inconsistent” if the different treatment is justified by the different institutional norms and constraints that operate on judges, as compared to other political actors (which I consider to be at least in part an empirical question). Neither this point – nor any other one I made – depends on controversial claims about the nature of truth or logical consistency, postmodern or otherwise.
In their brief rejoinders, Eric and Adrian continue to insist that their argument does not depend on any empirical claims about what motivates judges. But in so arguing, each of them contradicts himself and concedes my original point in the process.
Adrian first says of the kind of argument they were examining that “it is caught in a dilemma — it can survive filter (1) only by taking a form that causes it to be weeded out by filter (2).” I take Adrian to mean here that the argument can avoid the charge of inconsistency (filter (1)) but only at the cost of making implausible empirical assumptions about how judges act (filter (2)). But then he goes on to say that by the time we are considering the empirical question (filter (2)), “the fallacy has already dropped out by that point; it is not affected at all by whatever happens in the debate at the second stage.” But how can it be that the fallacy is “not affected” by what happens at the second stage if, as he has just said, it can “survive filter (1)” by making empirical claims that filter (2) then “weeds out”?
Eric makes the same error in even more efficient fashion. He says, “we sometimes argue that they escape the problem only by making implausible arguments. But the inside-outside problem does not depend on our skepticism about these specific arguments being correct.” Eric’s second sentence contradicts his first. He acknowledges that the targets of their critique can “escape the problem” (of inconsistency) by making what he considers to be implausible empirical arguments. But then he insists that their charge of inconsistency does not depend on those empirical arguments about judicial behavior being implausible. But how can that be the case if, as he has just said, the scholars can avoid inconsistency if those empirical arguments are correct?
I don’t think I’m the postmodernist in this debate.
I generally follow Johnson’s advice never to respond to critics, but this is the season for breaking resolutions. So let me offer a brief rejoinder to Charles Barzun’s response to the Posner/Vermeule paper on the Inside/Outside Fallacy; both are recently published by the University of Chicago Law Review.
Eric and I suppose that successful arguments (in constitutional theory, inter alia) must pass through two separate, independent and cumulative filters: (1) a requirement of logical consistency (the inside/outside fallacy is one way of violating this requirement); (2) a requirement of substantive plausibility (not ultimate correctness).
With respect to some of the particular arguments we discuss in the paper, we say that the argument is caught in a dilemma — it can survive filter (1) only by taking a form that causes it to be weeded out by filter (2). Now in some of those cases, I take it, Charles disagrees with us that the argument fails the second filter. He is of course entitled to his views about that. But the inside/outside fallacy — which is the first filter — is strictly about the logical consistency of assumptions, not their plausibility. Thus the fallacy has already dropped out by that point; it is not affected at all by whatever happens in the debate at the second stage. It’s just a muddle to say that because Eric and I do happen to have substantive views about what counts as plausible for purposes of the second filter, we are therefore smuggling substantive content into the first filter. Not so — unless one subscribes to the postmodern view that logical consistency is itself a substantive requirement, thereby jettisoning the distinction between validity and truth. (In some passages, Charles seems willing to abandon himself utterly to that hideous error, but for charity’s sake we ought not read him so, if we can help it).
So when Charles says that the inside/outside fallacy smuggles in substantive assumptions, I think that’s a confusion that arises from failing to understand the distinction between the two filters. The reader of Charles’s piece should be alert for skipping to and fro between these distinct questions of logical consistency and plausibility.
(And see Eric’s earlier reply.)
I examine the grudging case at Slate.
Source: Cingranelli-Richards Human Rights Dataset (0 (bad) to 2 (good)).