Empiricism

Jonah Goldberg and Jonathan Chait have been debating over whether liberals are more "empirical" than conservatives. It all started with Chait's article in the New Republic:

Imagine that God were to appear on Earth for the unlikely purpose of settling, once and for all, our disputes over economic policy. And suppose that, to my enormous surprise, he announced that every empirical claim advanced by conservatives was correct. Cutting taxes produces such great economic growth that even the poor benefit. Privatizing or eliminating social programs like Medicare and Social Security will cause the elderly to save more money and enjoy higher living standards. Slashing regulations, by eliminating unintended side effects, actually does a better job helping those whom the regulations were intended to help than the regulations themselves. Suppose that God presented these conclusions so convincingly--if his stature alone did not suffice--that everybody immediately accepted them as truth.

How would liberals respond? No doubt by rethinking and abandoning nearly all their long-held positions. Liberalism, after all, claims to produce certain outcomes: more prosperity and security, especially for the poor and middle classes; a cleaner environment; safer foods and drugs; and so on. If it were proved beyond a doubt that liberal policies fail to produce those outcomes--or even, as conservatives often claim, that such policies hurt their intended beneficiaries--then their rationale would disappear. It may be hard to imagine liberals advocating capital gains tax cuts as a way to lift up the working stiff. But that's just because there's no evidence to show they do. If the evidence were to change, so would the liberal mindset. The point is that liberalism has no justification other than the belief that liberal policies produce beneficial outcomes.

Now imagine the opposite were to happen. God appears in order to affirm liberal precepts: Current tax levels barely affect economic incentives, social programs provide tremendous economic security at modest cost to growth, and most regulations achieve their intended effects without producing undue distortions. Would economic conservatives likewise abandon their views? Some certainly would, but a great many would not. Economic conservatism, unlike liberalism, would survive having all its empirical underpinnings knocked out from beneath it.

. . . [C]onservatism, unlike liberalism, overlays a deeper set of philosophical principles. Conservatives believe that big government impinges upon freedom. They may also believe that big government imposes large costs on the economy. But, for a true conservative, whatever ends they think smaller government may bring about--greater prosperity, economic mobility for the non-rich--are almost beside the point. As Milton Friedman wrote, "[F]reedom in economic arrangements is itself a component of freedom broadly understood, so economic freedom is an end in itself."

* * *

The contrast between economic liberalism and economic conservatism, then, ultimately lies not only in different values or preferences but in different epistemologies. Liberalism is a more deeply pragmatic governing philosophy -- more open to change, more receptive to empiricism, and ultimately better at producing policies that improve the human condition -- than conservatism.

Chait then cherry-picks several examples that show specific liberals being more "empirical" than specific conservatives.

Two responses come to mind:

1. You can't say that one philosophy is more "empirical" than another without knowing what the goals of the policy are. You just can't. An analogy: Who is in the best "health": A marathon runner, a champion weightlifter, a top sprinter, or a woman who has lived to 105? You have to decide whether "health" consists of endurance, power, speed, or pure longevity. Until you make that choice, there isn't any "empirical" question that is capable of being answered.

Same for economic policy. Which is "better": (1) a policy that maximizes freedom and autonomy, or (2) one that maximizes personal security and minimizes risk? There isn't any "empirical" way to say which goal is "better." But Chait repeatedly pretends that if some conservatives prefer (1), it is only because they are unempirical as to (2). Chait is fundamentally confused here. He doesn't understand that not everyone is even trying to maximize number (2) in the first place.

Nor does he understand that if conservatives prefer goal (1) while liberals prefer goal (2), both sides have made ideological choices that have nothing to do with "empiricism." A conservative could with equal confusion say that "liberals are unempirical" because they prefer their own set of policies even if it is conclusively demonstrated that conservative policies are empirically the best at maximizing freedom.

2. Who says that liberals consistently value empiricism, come what may? My impression is that many liberals, like many conservatives, are highly resistant -- to the point of being immune -- to any empirical evidence that they dislike. Let someone come out with a study showing that the death penalty deters murder (e.g., Cass Sunstein and Adrian Vermeule), or that charter schools are an improvement (Caroline Hoxby), or that school busing doesn't improve student performance (be sure to read the story of James Coleman), or that lower taxes improve economic growth (Gwartney and Lawson), or that school vouchers have any advantage at all (Paul Peterson et al.), or that the welfare state encourages broken families, and liberals hop out of the woodwork to take potshots at the study.

After all, no study is beyond all criticism. Maybe the sample size wasn't large enough, or there was selection bias, or there was omitted variable bias, or there were too many cross-correlated variables, or there's another data set that is somehow better than what the researcher used, etc., etc., ad nauseam. Do liberals mention such criticisms because they are all hard-headed empiricists who demand the utmost rigor out of every study that comes their way? Maybe a few of them are. But most liberals want the world to agree with their ideological conclusions, and are determined to ignore or discount any counter-evidence.

And to be fair, the same goes for many conservatives' attitudes toward studies involving global warming, racial profiling, a few studies showing the opposite conclusion on the minimum wage (Card and Krueger, for example) and the relationship between taxes and economic growth, or articles about the French health-care system. I'm not trying to argue that conservatives are generally better at being empirical. (To do so, I would have to cherry-pick examples as Chait did.)