Tuesday, January 22, 2008

Unintended Consequences?

In the latest in their occasional Freakonomics column in this past Sunday's NY Times Magazine, journalist Stephen Dubner and economist Steven Levitt tell three stories of unintended consequences. One of their two principal modern examples involves the Americans With Disabilities Act, which, they say, had the perverse effect of lowering the employment rate of Americans with disabilities. Citing a study by economists Daron Acemoglu and Joshua Angrist , Dubner and Levitt explain: "Employers, concerned that they wouldn’t be able to discipline or fire disabled workers who happened to be incompetent, apparently avoided hiring them in the first place."

An interesting story, but is it true? Well, for one thing, it's not clear that Dubner and Levitt understand the argument they're reporting: Fear of back-end lawsuits by dismissed members of a protected class is a frequent argument made against non-discrimination mandates in general, but the particular worry with respect to the ADA has typically been that its requirement of "reasonable accommodations" for disabled workers would impose higher costs on employers for workers they don't fire. The reasonable accommodation requirement of the ADA distinguishes it from Title VII, the general federal anti-discrimination in employment statute. By failing to confine their argument to the disability context, Dubner and Levitt (perhaps unwittingly) suggest that all anti-discrimination law is perverse.

What about the core empirical claim? Here's what Washington University of St. Louis Law Prof. Sam Bagenstoss had to say in a 2004 article in the Berkeley Journal of Employment and Labor Law:
In brief, though I find it hard to disagree with the claim that the statute (at least initially) imposed some negative pressure on employers' decisions to hire some people with disabilities, critics of the statute have argued well beyond their data in urging that the ADA be abandoned. To the contrary, the data suggest that much (though probably not all) of the employment decline for people with disabilities resulted from factors extrinsic to the statute. In particular, I find quite plausible the argument that the 1990-1991 recession pushed an unusually large number of people with disabilities out of the workforce and onto the SSDI rolls--an argument pressed by the economists John Bound, Timothy Waidmann, David Autor, and Mark Duggan--though it is difficult empirically to disentangle that phenomenon from the effects of the ADA. Moreover, whatever the ADA's short-term effects, it seems likely that the statute's net long-term effects on employment for people with disabilities will be positive.
Dubner and Levitt could be forgiven for not reading all of the relevant law review literature; they are after all, not legal academics or lawyers, and disciplinary boundaries are often substantial. But even if they were unaware of the Bagenstoss piece, surely they must know of the whole book on the subject that Bagenstoss was reviewing, as it is very much a work of economics.

In addition to the factors Bagenstoss cites, one might also note---though Dubner and Levitt fail to do so---that discriminating against disabled persons at the hiring phase is also illegal. And as various scholars have noted---including Yale Law Prof Christine Jolls (also an economist) in a 2000 Stanford Law Review article ---there is relatively lax enforcement of the ADA's prohibition of discrimination in hiring. To be sure, stricter enforcement at the front end would not be costless, and thus Dubner and Levitt could rationally oppose it on those grounds. But they don't even mention the front-end enforcement possibility.

At the end of their essay, Dubner and Levitt admit that not all laws have perverse consequences. Yet they leave the clear impression that they think the ADA does and inevitably must. Both claims are highly controversial. Dubner and Levitt present them as unarguable fact. Freakonomics indeed.

Posted by Mike Dorf