Cass R. Sunstein and Thomas J. Miles have an op-ed in today’s L.A. Times. It’s another one of those pointless exercises that tries to measure “activism” and “partisanship” with studies that purport to be “objective” while ignoring the real issue: are judges deciding cases correctly?
With regard to partisanship, Sunstein and Miles say:
[W]e examined all cases in which members of the court, using settled principles, evaluated the legality of important decisions by federal agencies, such as the Environmental Protection Agency, the National Labor Relations Board, the Occupational Safety and Health Administration and the Food and Drug Administration.
We used clear and simple tests to code the decisions of these agencies as either “liberal” or “conservative.” For example, we counted an environmental regulation as “liberal” if it was challenged by industry as too aggressive, or as “conservative” if it was challenged by an environmental group as too lax.
We used equally simple tests to code the decisions of the justices. If a member of the court voted to uphold conservative and liberal agency decisions at the same rate, we deemed him “neutral,” in the sense that his voting patterns showed no political tilt. If a justice showed such a tilt, we deemed him “partisan.”
The problem is: what if all the decisions they are looking at are really liberal?
I can’t express the problem any better than this commenter did at the University of Chicago Law Blog:
Perhaps I am missing an element of the analysis in the LA Times article (which was really more of a synopsis), but it doesn’t seem like the partisanship ranking really tracks partisanship in terms of potential bias (the conventional connotation of the term). Imagine a world in which agencies have a “liberal” bias and that bias manifests itself in the agency’s willingness to overstep the bounds of its proper authority in defense of liberal positions. In that world, wouldn’t a justice who voted against upholding decisions that exceeded agency authority be described as “partisan conservatives” in this analysis even though the justice’s decisions were not influenced at all by the liberal/conservative result being reached by the agency?
Sunstein and Miles make a similar error when they examine “activism” vs. “restraint”:
If a justice regularly voted in favor of agencies, we deemed him “restrained,” because he proved willing to accept the decisions of another branch of government. If a justice was unusually willing to vote against agencies, we deemed him “activist,” in the literal sense that he frequently used judicial power to strike down decisions of another branch.
But here, they at least acknowledge that their artificial labels elide the question of whether the decisions are correct:
Note that the terms “restrained” and “activist” are purely descriptive, and so permit an objective test of judicial behavior. It is possible that a justice who is restrained, in our sense, is wrong, and that an activist justice is right.
Indeed. But, like studies that examine activism by reference to the rate at which justices vote to overturn laws, the caveat swallows up the utility of the analysis. As I said about a similar study in September 2006:
If California passed the “Imprison All Arabs Law of 2006″ tomorrow, and a court were to strike it down, this study would label such a ruling an “activist” decision.
Let’s take a couple of similarly absurd examples in the regulatory context. Let’s pretend that the FDA, without any good reason, imposes a 20-year certification process for a new cure for cancer — one that has saved millions of lives throughout the rest of the world with no side effects. If a court struck that down, Sunstein and Miles would label that decision “activist” — even though it would be obviously correct.
If the court upheld a bribery-induced FDA approval of thalidomide as a cure for morning sickness, that would be “restrained” — even though the regulation the court upheld would be insane and would lead to tragic birth defects.
Why do we keep paying attention to studies like this?
UPDATE: More here.