Timothy McVeigh is the ideal poster boy for the death penalty, it is often said. He is an unmistakably guilty, unrepentant, rational, calculating, confessed mass murderer who can complain neither of racism (he’s white) nor of an unfair trial (he had good lawyers). If anyone ever deserved execution, he does. Even leading anti-death-penalty scholar Hugo Adam Bedau has said: "I’ll let the criminal justice system execute all the McVeighs they can capture, provided they’d sentence to prison all the people who are not like McVeigh."

There are two untidy little problems here. The first, of course, is the disclosure just six days before McVeigh’s now-postponed May 16 appointment with death that the FBI had failed-most likely through incompetence, not design-to turn over to his attorneys more than 3,000 pages of possibly relevant documents. The second, less noticed, problem is that executing McVeigh, with the attendant media circus, seems more likely to provoke future carnage than to deter it: It risks lighting some other potential terrorist’s psycho-pathological fuse, just as the 1993 Waco disaster lit the fuse that led to the bombing in Oklahoma City two years later.

With luck, McVeigh’s execution-if and when it comes-will not be avenged. But can we be confident of that, after watching him glory in the fame that mass murder has brought, to the point of suggesting that his execution be televised? And after seeing the Middle East convulsed by suicide bombers craving martyrdom? Terrorists "are not deterred by the threat of severe penalties," in the words of a 1976 book by Frederick J. Hacker. "In fact, they are attracted by dangerous risks and the expectation of a heroic death."

This is no small consideration. President Bush and former Vice President Al Gore both justified the death penalty in their October 18 debate exclusively on the ground that it saves innocent lives by deterring killings. "It’s the only reason to be for it," Bush asserted. "I don’t think you should support the death penalty to seek revenge. I don’t think that’s right. I think the reason to support the death penalty is because it saves other people’s lives."

If executing McVeigh could have the opposite effect, and could endanger innocent lives, why take a chance? Why not lock him up forever instead, throw away the key, and keep him out of sight and out of mind?

Most capital murderers are not terrorists, of course. But most Americans now believe that the death penalty does not deter murders. So do more than 80 percent of leading criminologists, according to a 1996 survey. Indeed, some claim-not all that convincingly-that routine executions provoke more homicides than they prevent, through what academicians call the "brutalization effect" of teaching that it is correct and appropriate to kill those who have gravely offended us.

Exemplifying the journalistic conventional wisdom on deterrence is Washington Post columnist Richard Cohen, who wrote in September: "Over the years, crime figures have been analyzed and re-analyzed, and always the conclusion is the same: Capital punishment fails to deter." Come to think of it, my own column seems to have asserted last year that "there is no convincing evidence" that the death penalty deters murderers-at least, not "at the current pace of executions."

But now come three Emory University economists with a major study that contradicts the findings of most other academicians (not to mention Cohen and me) by concluding: "Our results suggest that capital punishment has a strong deterrent effect…. In particular, each execution results, on average, in 18 fewer murders-with a margin of error of plus and minus 10."

Uh-oh. The multiple-regression analyses underlying this conclusion are way over this columnist’s head. But the authors of the study (economics department Chairman Hashem Dezhbakhsh and professors Paul H. Rubin and Joanna M. Shepherd) have respectable professional credentials. And their analysis, published on the Internet in February, is based on more-recent and detailed data and (it appears) more- sophisticated statistical techniques than any previous study.

Here’s a more accessible statistical tidbit noted by Dudley Sharp, resource director of Justice for All, a Texas-based, pro-death-penalty criminal justice reform group: "The major U.S. jurisdiction with the most executions is Harris County [Houston, Texas], which has seen a 73 percent decrease in murder rates since resuming executions in 1982-possibly the largest reduction for a major metropolitan area since that time."

What would folks such as Richard Cohen say if they believed that each execution saves 18 lives-or, as the authors claim in acknowledging their margin of error, at least eight lives, and as many as 28? Cohen has told us already: "It would be one thing if the death penalty really was a deterrent. Then opponents like me would be in a fix. I’d still have the same moral qualms, but I’d be hard-pressed to argue that we ought to suffer a high murder rate just to make a point about the value of human life."

My sentiments exactly. (But should such logic lead us to the ironic extreme of sparing McVeigh to avoid provoking imitators, while executing lifers who kill fellow prisoners?)

It’s true, of course, that most homicides are crimes of passion or committed by mentally unstable people, impulsive teenagers, and others who are probably not deterrable. But the question is not whether most potential killers can be deterred. The question is whether any can be. Common sense says, yes.

All criminal penalties are based on the incontestable theory that most (or at least many) criminals are somewhat rational actors who try so hard not to get caught because they would prefer not to be imprisoned. And most are even keener about staying alive than about avoiding incarceration. Who could argue with a straight face that robbers pondering whether to kill the witnesses would be indifferent-in every case-to the expectation that, if caught, they would be tried, convicted, sentenced, and executed within (say) 18 months?

We do not have such a regime, of course. Nor should we, given how dramatically it would increase the risk of executing innocent defendants. These days, only about one in every 150 killers-and perhaps one in every 30 capital murderers-is executed, often after more than 10 years of judicial appeals. The long odds and delays attenuate the deterrent effect, but seem unlikely to eliminate it, especially given the push by states such as Texas to make execution for capital murder more swift and certain.

Statistical studies and common sense aside, it’s undeniable that the death penalty saves some lives: those of the prison guards and other inmates who would otherwise be killed by murderers serving life sentences without parole, and of people who might otherwise encounter murderous escapees.

So those of us who lean against the death penalty must confront the very real possibility that abolishing it could lead to the violent deaths of unknown numbers of innocent men, women, and children. And those who are still skeptical that the death penalty deters any killings must also confront the risk-benefit calculus suggested by political scientist John McAdams of Marquette University: "If we execute murderers, and there is, in fact, no deterrent effect, we have killed a bunch of murderers. If we fail to execute murderers, and doing so would in fact have deterred other murders, we have allowed the killing of a bunch of innocent victims. I would much rather risk the former. This, to me, is not a tough call."

It’s still a tough call to me-certainly for as long as so many states use such a grotesquely unfair and error-riddled process to select defendants for death. And even if the process were more fair, the death penalty would still involve heavy costs: the unavoidable risk of executing the innocent; the racial divisiveness of a system regarded (correctly or incorrectly) by so many African-Americans as riddled with racism; the affront to many Europeans and others who see America’s apparent enthusiasm for executions as a barbaric anachronism; even the evidence suggesting that executing capital murderers costs more than locking them up for life.

How to weigh such costs against the possibility of saving innocent lives? The answer comes easily to believers in retribution, such as the late Mike Royko: "Anything less than the death penalty [for murder] is an insult to the victim and society. It says … that we don’t value the victim’s life enough to punish the killer fully." And to believers in the immorality of the death penalty, such as the late Justice William J. Brennan Jr.: "The deliberate extinguishment of human life by the state is uniquely degrading to human dignity."

But for me, such moral absolutes tend to collapse into empirical questions such as how many innocent lives might be saved. I guess it’s time to learn multiple-regression analysis, or to find an unbiased econometrics translator. Any volunteers?