Jonathan Chait: Confirmation Bias in One Satirical Lesson

Jonathan Chait's article, “Fact Finders,” in the new TNR is one of the most obnoxiously blinkered pieces of self-serving political magazine writing in recent memory. I'm just flabbergasted by the stupidity of this thing. Chait's claim is that liberals by and large are empiricists, willing to go where the evidence takes them, while conservatives (loosely and irresponsibly identified with free-market types) are dogmatists who will unaccountably but doggedly cling to principle even after being brought low by data. The claim is almost self-refuting. It should be impossible for an intelligent and observant person, such as Chait imagines himself to be, to fail to see the ravages of dogmatic narrowness on all sides. To claim the mantle of empiricism exclusively for liberalism (or any -ism) in the teeth of overwhelming evidence that that empiricism is water in ideology's oil is a signal failure of empiricism.

An empiricist about the alignments of empiricists will surely comes to this (my) conclusion:

There are surpassingly few empiricists about politics. And among those wide-minded few there are liberals, conservatives, libertarians, etc. The reason we do not converge on an evidence-based consensus is that we all started with different priors, and all use different methods of updating our beliefs, about which we also have different priors, which causes us to avail ourselves of certain kinds of evidence and not others, to trust certain kinds of studies and not others, to give credence to certain experts and not others, to mistrust certain methods of inquiry, and not others.

Surely Chait has heard of confirmation bias. John Hollander once wrote a book of poems about writing poems. A villanelle about writing villanelles, a sonnet about sonnets. “Fact Finders” reminded me of that sort of thing, but as satire. Assignment: Write an article about how YOUR ideology embodies true blue commitment to just-the-facts empiricism while in fact demonstrating the kind of confirmation bias that makes just-the-facts empiricism all but chimerical in political discourse. Jonathan Chait: A+!

“Fact Finders” also exemplifies a strange lack of coherence about the nature of liberalism, no doubt a symptom of Chait's admiring belief that liberalism is marked by it's lack of coherence. Really.

Chait: “Incoherence is simply a natural byproduct of of a philosophy rooted in experimentation and a rejection of ideological certainty.” I think he thinks he's helping. But, in any case, liberalism is not marked by it's lack of coherence, or by it's promise “to produce certain outcomes: more prosperity and security, especially for the poor and middle classes; a cleaner environment; safer food and drugs; and so on.”

Liberalism is not a list. It's just not. And it is not a list that has incoherence as a natural byproduct of being a list that rejects ideological certainty. Green, Hobhouse, Dewey, Rawls, et al did not see themselves as championing incoherent lists of things people might happen to want. They championed a particular conception of the relationship between the citizen and the state based on what they took to be compelling general normative principles.

Which is why Chait looks dense when he takes it to be a flaw in conservative thinking that it appeals to compelling general normative principles. He thinks it's telling when Milton Friedman says, “Freedom in economic arrangements is itself a component of freedom broadly understood, so economic freedom is an end in itself,” when he should simply be nodding. But this is not a statement of laissez faire though the heavens may fall. Friedman is simply enunciating a fairly banal and unobjectionable liberal principle. Freedom is an end in itself, things that are constitutive of freedom are thus ends in themselves, and economic freedom is constitutive of freedom, so an end in itself. How does Chait think he got his list of socioeconomic desiderata? Handed down from liberal Sinai? It's the list he found on his desk when he started at TNR and TNR's list is definitive of liberalism? Or are the items on the list thought to be instrumental to or constitutive of some liberal good, such as freedom? Perhaps.

Anyway, you might want to disagree with Friedman, for whatever bad reason, and say that economic freedom is not part of freedom broadly understood. But if you replace “economic freedom” in Friedman's sentence and substitute “x” where x is anything that is part of freedom broadly understood, and you still disagree, then there is no intelligible sense in which you are a liberal.

Accepting Friedman's principle, that things that are constitutive of freedom are ends in themselves because freedom is an end in itself, does not disqualify you from good standing as empiricist because empiricism is indifferent between values. If you like freedom, and you can say with adequate specifity what freedom is, then empiricism can help you know how best to maximize freedom. If you like totalitiarian dictatorship, empiricism can tell you how best to sustain a totalitarian dictatorship. If you like safer food and drugs … etc. Free-market types can't fail to be empiricists simply because they are not empiricist liberals, although this at times appears to be Chait's argument.

If God came down and told conservatives that free-markets and smaller government aren't the best way to get the things on the list kept in the offices of the New Republic (“And I know,” God said, “for it is I who made Nature's Laws”) and the conservatives said, “Oh, that's OK God, we've got a different list in the offices of Americans for Tax Reform, but then you knew that,” that's not a failure of empiricism.

Throughout the article Chait blithely sides with government-interventionist economic policy, as if the matter had been decisively resolved through empirical inquiry, that there is no ongoing debate among serious economists, and those on the other side are simply benighted ideologues unmoved by empirical considerations.

Chait cites several conservative punditocracy responses to Michael Kinsley's argument that PRAs can't increase national wealth. Apparently, many of them think personal accounts would be a good idea even if Kinsely was right. To Chait this is evidence of dogmatic commitment to the devolution of power from state to citizen.

This preference for removing power from Washington is simply something that either you accept or you don't. It's neither right nor wrong in an absolute sense. It does, however, make empirical reasoning pointless. Viewed pragmatically, Social Security raises questions about which economics has a lot to say: balancing the tradeoffs between retiree incomes and costs to workers, allocating risk, and so on. Liberal thinking, unlike conservative thinking, actually hinges on the outcome of those questions.

First, many people prefer removing power from Washington on what they take to be empirical grounds about the relative effectiveness of federal vs state vs individual control in achieving some goal they care about. (Again, Chait seems, bizarrely enough, to interpret a commitment to different goals from his as a failure of empiricism.) Second, is Michael Kinsley Jonathan Chait's main source of economic theorizing? I swear that just two weeks ago I heard the 2004 Nobel winner say that a system of social security personal accounts would have a monumental effect on the supply of labor, and thus on growth, and national wealth. So what's an empiricist to do? Throw in one's lot with Michael Kinsley or Edward Prescott?

Well, this question leads us to the point that Chait doesn't understand what empiricism is. Look:

. . . the remedy of smaller government is particularly ill-suited for the problem of health care. The market for medical services does not resemble the market for blue jeans. Among other problems, health insurance firms have every incentive to deny coverage to those most likely to get sick, which makes the health insurance market inefficient and prohibitively expensive. Economists call this phenomenon “adverse selection,” and it is inherent in the private health care market. It cannot be solved without some kind of government intervention.

Let it now go unsaid that the market for medical services does rather resemble the market for blue jeans in the absence of government interference. For now let us consider the example of adverse selection. Chait repeats a a prediction of economic theory, given certain assumptions. In a context of perfect information, there is no adverse selection because of the possibility of price discrimination. If we relax the assumption of perfect information, and allow for informational assymetries, then we get adverse selection. But high theory, based on nosebleed-inducing abstraction, simply generates hypotheses, many of which we know to be false, to have no empirical basis. We could relax or assert different assumptions and get different results. The proclamations of theory have a grip on the empiricist only to the extent that the assumptions have empirical teeth (which is why I don't believe Prescott about the effects of PRAs on labor supply).

Tyler Cowen, at least as good an economist as Michael Kinsley, writes:

Economists miss one of the biggest problems with insurance. We are blinkered by adverse selection models, which imply that the dangerous prospects most want to buy insurance. The opposite is more often true. If you are an irresponsible driver, you are likely to be irresponsible in other spheres as well and not buy auto insurance. On the whole many insurance markets show positive rather than adverse selection. . . this means that the people who most need insurance will be the least likely to buy it.

Now, maybe you think that the fact that people who you think need insurance often don't try to buy it is a problem the state needs to solve, but it's not a problem with the market. The main point is, Chait doesn't know that commitment to economic theory is not tantamount to a commitment to empiricism, and that he just happens to select the set of theoretical assumptions that generates a problem he claims “cannot be solved without government intervention.”

I think that's enough. Let's just grant that if Chait is correct, and liberalism is by nature incoherent, then his article successfully embodies liberal ideals to a spectacularly high degree. And that the guy's no Francis Bacon.