More Misbehavioral Economics
I say, again and again, that it is an embarrassing non-sequitur to argue that people are “irrational” and then leap to the conclusion that they need benevolent paternal guidance from the state. After all, if people are irrational then voters are irrational, politicians are irrational, bureaucrats are irrational, etc. To this, Ezra Klein responds:
I’m not sure what exactly it is that Will finds so inexplicable here. Behavioral research often finds that consumers act irrationally in certain situations. So given a specific set of constraints, they may underestimate future risk, prove oversensitive to loss, exhibit significant status quo bias, and so on and so forth. All problems.
Now, the government may be made up of people, but it is not made up of people carrying out transactions under these conditions.
Perhaps Ezra is right, but only because people acting inside government institutions are much less likely to themselves bear the cost of their mistakes, and will therefore likely make more of them. There is no way to wriggle out of the fact that people who win elections are just like the rest of us.
I really wish people would pay more attention to Vernon Smith, who invented experimental economics, won the Nobel Prize for it, and remains by far the most philosophically rigorous theorist of the relationship between individual rationality and institutional performance. (Ted Bergstrom’s paper here [pdf] is a good overview.) What Smith’s work shows is that, yes, individuals in isolation don’t act according to canonical postulates of rationality, but that well-structured market institutions will nevertheless tend to converge on the efficient outcome, as if the agents were behaving with full “rationality”, even though they are in fact limited, confused, and ignorant. The “rationality” of the outcome is more a function of the structure of the institution than of the “rationality” of those acting inside it.
Responsible social science therefore compares the way real people perform when embedded in different real-world institutional settings. What you surely don’t do is perform selective empirical work to discover an “anomaly” in decision making, and then deploy a priori high theory to infer that one set of institutions (markets) won’t work, because, in fact, the performance market institution might turn out to be indifferent to the anomaly or limitation. That’s what Smith has proved. If you’re going to be an empiricist, then be an empiricist, and actually test the effect of the anomaly in the performance of the relevant institutions. Until you do this, it’s either arbitrary, naive, or willfully ideological to posit another set of institutions (government) as a fix. Because there may be nothing to fix. And, even if there is, government may be the wrong kind of institution to fix it. You’ve got to run the experiment.
There is a great deal of carelessness in generalizing the results of anomaly-focused behavioral economics. As Steven Levitt and John List write in their short article on behavioral economics in Science ($$$) this month:
Perhaps the greatest challenge facing behavioral economics is demonstrating its applicability in the real world. In nearly every instance, the strongest empirical evidence in favor of behavioral anomalies emerges from the lab. Yet, there are many reasons to suspect that these laboratory findings might fail to generalize to real markets. We have recently discussed several factors, ranging from properties of the situation — such as the nature and extent of scrutiny — to individual expectations and the type of actor involved. For example, the competitive nature of markets encourages individualistic behavior and selects for participants with those tendencies. Compared to lab behavior, therefore, the combination of market forces and experience might lessen the importance of these qualities in everyday markets.
List has run a number of field experiments that show that this is the case. Smith has run a number of lab experiments that show that the frequency of a “mistake” goes down as the cost of making it goes up.
Ezra continues:
An easy example is the research on opt-out 401(k)s. We know, from the economists, that investing in 401(k)s is generally a wise idea. We know, from the statisticians, that far fewer people do it than should. We know, from the behavioralists, that far more people would do it if the default setting put you in the 401(k), rather than forced you to wander down to HR and specifically ask for it. And so folks in the government, acting with more information and in a different context than folks in an office, think up a policy to “recognize the power of inertia in human behavior and enlist it to promote, rather than hinder, saving.”
At exactly which point in this process does Will fear that the same irrationality that keeps someone from creating a retirement account will foul up a regulator’s efforts to ease their way into a retirement account?
As I said to Dan Ariely in our chat, I think behavioral work is really valuable, especially when it suggests to us how people might better structure their affairs to get more of what they want. I think the evidence shows that 401(k) opt-out defaults are often a good idea, and that businesses ought to make that part of their standard labor contract, if that is something that they think would be appealing to their prospective employees.
I also think that this minor fact about the general distaste for filling out complicated forms can hardly be used to justify further encroachments on the right of individuals to negotiate the terms of their contracts with employers. I think Ezra’s argument here is both strangely narrow and ungenerously extreme. I don’t doubt that non-terrible policies are sometimes successfully enacted. To doubt that would be a bit like a market skeptic doubting that anyone ever succeeds in buying a candy bar. That would be terrifically dense. What I doubt, very strongly, is that the discovery of “irrationalities” undermines the authority of market institutions more than it undermines the authority of government institutions. Are people more or likely to behave irrationally when voting for their congressman or when buying a sandwich? Do buyers for private organizations sign contracts for $76 screws? Etc.
So, no. I don’t fear the mandatory opt out 401(k) plans in particular will be a giant debacle. But I do fear that half-baked behavioral economics is being used to undermine support for market institutions in general, way ahead of the evidence. And I fear that a fundamentally confused assault on “rationality” is being used successfully to promote paternalistic control by elites and, necessarily, to encourage the docility of those who are to be controlled.
[Added: If you have not read Ed Glaeser's "Paternalism and Psychology" [pdf], then you probably should.]
February 28th, 2008 15:45
It’s worth noting that Thaler and Sunstein have used very clearly defined examples with strict perimeters. How many of those are out there? Once the policy becomes “the government gets to try to nudge people into making the ‘right’ decision”, does the content of “the right decision” become a fight between insurance companies that want to sell additional psychiatric care insurance and drug companies who think that you’d you really benefit from more Prozac if only you’d give it a chance? Or more St John’s Wort?
Dismiss this as a violation of the original scientific motivations behind libertarian paternalism all you want. Do you really think that clever attorneys and marketeers won’t be able to find a way to fit whatever they’re selling into the “ignorance” and “inability to rationally process future effects” piece of consumer irrationality? Do you really think that the populace is going to be able to monitor the lobbyists pushing their own product? By definition the populace is already incapable of doing that even if they knew exactly what was going on, right?
February 28th, 2008 17:38
The company I work for defaults new employees into the 401(K) program, even though they match part of this contribution. No coercion needed.
It’s funny to see how surprised some people are that companies actually want to make their compensation packages attractive over time.
If there’s a sensible policy that’s good for employees, companies that are competing for labor will tend to use it when it seems to make sense, and reject it when it doesn’t (e.g. industry specifics differ, alternatives change, etc.).
This isn’t perfect, but it’s much more reliable than government policies adapting appropriately.
February 28th, 2008 19:36
Yes, it’s all too easy to apply the 3-step “enlightened” theory of patronomics:
1) identify institutional failure
2) identify policy to correct failure
3) declare victory
Rodrik was calling this Second Best Economics a while back. I have been calling it vulgar second best theory for a while longer. Why? Because there is no rigor; all they have to do is identify any failure in step 1, including the asymmetric information or imperfect information, the tied-for-second-to-last refuges of scoundrels. Okay, door opened, move on to 2, where the same level of scrutiny is not performed on the policy. Institutional failure and bias in government? Move on, nothing to see here. Or perhaps it’s time that Mr. Klein pick up a copy of _Narcissistic Process and Corporate Decay_ or any of Niskanen’s work on bureaucracy.
What exactly is it that makes 401(k) plans so awesome that everyone must be forced into them? Might it not be the case that some people would derive greater benefit from higher take-home pay and more liquidity? But no, we must treat everyone as a child who ***must*** place the maximum amount in the private mutual fund offered to them.
By the way, isn’t this prescription (privately managed mutual funds) being offered by the same people who vehemently denied that similar levels of choice in the Social Security program would have results no less dire than the opening of the 5th or 6th Seal just prior to the Apocalypse? I may be exaggerating; they certainly were. And I’m not even in favor of that policy change.
February 29th, 2008 14:07
>There is no way to wriggle out of the fact that people who win elections are just like the rest of us.
I tend to agree with this statement, but there’s no a priori reason for it to be true. To play devil’s advocate, why isn’t it reasonable that the people that win elections might be, on balance, slightly more rational than the average American? Especially if the average voter is more rational than the average American? To put it another way, what if only people deemed sufficiently rational were allowed to vote? Or what if only people deemed sufficiently rational were allowed to serve? In such a situation, surely it’s possible (probable?) that the dicta of a so-called “rational class” may serve the average American better than their own ill-informed decisions.
February 29th, 2008 17:45
What Ezra’s example conceals is that even allowing opt-outs is quite a concession from the government on most issues! Sure, for 401k’s default-and-optout is a move in his direction, but for quite a few other things its a move in our direction. Think how many other things we are forced not to do at all, for our own good — no opt-outs allowed. I’d be fine with Ezra’s policy if we could apply it to drugs, prostitution, safety-belts, helmets and everything else that comes under the category of “most likely bad for us” but “some people want it anyway.”
So all I have to do to participate in these things is carry a form around with me that says “I’ve opted out of the default legal restrictions.”
The cops are gonna love this.
March 2nd, 2008 15:49
I wrote a long-winded reply to this, although Scott Wood captures a good deal of what I went for. Behavioral economics has very foreboding implications for the desirability of regulation by democratically accountable actors. Establishing that voters are rubes doesn’t exactly commend the prospect of letting their representatives run wild.
March 4th, 2008 01:56
I think Will is pretty much right here. But I also think libertarians should be pretty happy about this new “soft” paternalism. It seems to me that the more time we spend arguing about exactly how and which way we should “nudge” behavior, the less time will be spent writing the typical types of regulations that DON’T have an opt-out feature.
March 5th, 2008 15:57
Let’s just allow the irrationality to exist even if it harms the market. Just accept it.
March 6th, 2008 14:39
[...] Will Wilkinson (via overcoming bias) [...]
March 7th, 2008 14:54
Perhaps the greatest challenge facing behavioral economics is demonstrating its applicability in the real world.
I suspect the major application is generating justifications for additional government control of something or other.
I have hardly ever seen it used for any other reason.
March 10th, 2008 01:22
[...] Will Wilkinson / The Fly Bottle » Blog Archive » More Misbehavioral Economics Questioning the assumption that if people are “irrational” the Government is right/entitled/should intervene and make people do something different. i.e opt-in/opt-out of pensions savings. (tags: paternalism economics psychology rationality) [...]
March 17th, 2008 08:25
[...] article reminds me of something I read on Will Wilkinson’s blog. He discusses irrationality in economics and has this gem: I say, again and again, that it is an [...]