Weekly Philo of Economics: Analytical Egalitarianism makes the New York Times

In this week-end's New York Times, the economist, Tyler Cowen, reminded us of the noblest features of the history of economics: the "egalitarian and civil libertarian core." (On his his own blog, Cowen also calls it "cosmopolitan" for good reasons because the piece attacks the nationalist bias in much public policy economics.) In particular, he argues that these are linked: "If you treat all individuals as fundamentally the same in your theoretical constructs, it would be odd to insist that the law should suddenly start treating them differently." Cowen's additional claim, the economist's mantra that "economic analysis is itself value-free," is more problematic. Let's leave aside the non-trivial, problematic jump from positive analysis to normative advocacy (recall the Summers-Khan exchange.) As I have argued elsewhere, non-trivial values are inevitably smuggled into what counts as an "individual" (and "same") as well as any attempts to aggregate these. 

I was especially pleased  that Cowen linked his treatment to the scholarship of my friends and mentors, Sandra Peart and David Levy, on what they have called “analytical egalitarianism.” (Here and here [the latter contains a contribution by (ahum) me].) On Cowen's account analytic egalitarianism rejects the competing theoretical approaches that:

postulate natural hierarchies of religion, ethnicity, caste and gender, often enforced by law and strict custom. Economists too often forget that we are part of this broader battle of ideas, and that we are winning some enduring victories.

Quite right! But even so Cowen does not seem to have fully grasped the crucial insight of analytical egalitarianism. In effect, Cowen treats analytical egalitarianism as a theory that [a) postulates a homogeneity of human nature (when it comes to motives to action). This is indeed a necessary condition in analytical egalitarianism. But it is not sufficient: what needs to be added is that (b) the theorist has to be included in his or her model with the same motivational setup as the subjects studied. (To be clear (a) and (b) are jointly necessary for analytical egalitarianism, but they are also not sufficient.) Too often economists tacitly treat their own public utterances and policy-suggestions as themselves disinterested or un-incentivized (by considerations other than  truth) and the product of a value-neutral method. Why does this matter?

Public policy economists compete for scarce resources: status, grants, government jobs, and the privilege to advice policy-makers. As policy-advisers they can influence policy outcomes or offer the articulation that is supposed to justify it. (They have other roles, too, of course.) For an analytic egalitarian these activities need to be put inside the models that economists use and deploy in these policy-sensitive contexts. (This is especially so because the basic (welfare/partial equilibrium) framework that economists use make it too easy to get the policy result one wants out of the model used.) When Larry Summers goes to Pakistan to give policy advice (recall my post), his utterances systematically leave out the incentives governing the World Bank and his own personal career as servant to the rich and powerful. 

Okay, this all sounds very abstract, so let me put this in terms of Cowen's own example. In reporting Michael Clemens's 2011 paper, Cowen summarizes:

[U]nrestricted immigration could create tens of trillions of dollars in economic value, as captured by the migrants themselves in the form of higher wages in their new countries and by those who hire the migrants or consume the products of their labor.

Fair enough. Given that we are (nearly) all (also) consumers, we all gain? Well, that's not obvious. Mass immigration probably depresses wages in some job-markets; that's good for consumers and those "who hire the migrants," but not obviously so for those who get displaced or whose salaries get lowered. We would need to know if the possible falls in incomes in this latter group is fully compensated by the lower prices of the goods they buy. (Cowen recognizes a related problem: considerable immigration also puts non-trivial strains on welfare states, but that's not the same issue.) 

Now applied welfare economists do not like to treat such distributional effects. As the theoretician behind neo-Liberalism, Arnold Harberger (the Godfather of the Chilean Chicago Boys), put it in a classic 1971 article: "when evaluating the net benefits or costs of a given action (project, program, or policy), the costs and benefits accruing to each member of the relevant group (e.g., a nation) should normally be added without regard to the individual(s) to whom they accrue." (Harberger 1971, 785; cf. this critical treatment of Harberger by Khan 1992, which treats other, even more important issues.) Of course, economists are not blind to the fact that their policy proposals might have distributional consequences; so they often pay lip-service to compensating the losers of any policy (sometimes with a nod to so-called Harberger triangles.)

But, more often than not (as in Cowen's editorial) the distributional effects are buried in the small print. Of course, given that the folk who are negatively affected barely register among the incentives governing the contemporary, ambitious public policy economist it should be no surprise that their interests are not given the attention they deserve. (The concerns of these often show up as 'friction' in the language of economics.) I do not mean to suggest that distribution trumps other concerns or that this issue is the most important problem with applied welfare economics. But the major advantage of adopting analytical egalitarianism is that the market for policy economists ought to be included in one's analysis of the policy prescriptions recommended by these.

So, to recap from earlier blog posts (here, here, and here) below is the minimal, conceptual core of analytic egalitarianism.

[1] Naturally all people are (motivationally) equal, and most salient differences are the product of institutions (law, education, culture, etc).

• [A] This includes the philosophical/economist-expert
• [B] Even if [1] is not quite literally true, (nearly) all people consider themselves to be (at least as) equal when it matters to them.

[2] From [1A] it follows that the the expert should be put inside the model or conceptual framework

[3] From [1B] we learn that human affairs are (at least) in part intentional systems. (That is, we can't have a fully extensional model.)

Second, in practice analytic egalitarianism implies (at least) three important norms for how the experts and public policy ought to interact:

1. Experts can't keep themselves (their incentives/their roles, etc) out of the model/proposal. In practice this means that we can't simply assume that philosophers are disinterested truth-seekers in the context of policy.
2. Experts shouldn't promote policies where the down-side risks of implementation are (primarily) shifted onto less fortunate others.
3. Experts should make an effort to educate policy and opinion-makers to counter-arguments to the policies they advocate.

In practice, of course, considerable and controversial contextual judgment is required in evaluating the suitability and content of these norms (which may involve a weighing of conflicting interests, etc.).