(James Q. Wilson is the author of "Moral Judgment" and "The

Moral Sense." ----New York Times)

USA:

July 10, 2000----

For those who support capital punishment, as I do, the possibility

that innocent people could be executed is profoundly disturbing.

No human arrangement can guarantee perfection, but if perfection is not

possible, then the number of errors ought to kept as low as possible. For

that reason, it is worth studying "Broken System: Error Rates in Capital

Cases," the recent report by Professor James Liebman and others at the

Columbia University Law School, especially since that document has

stimulated an outpouring of media coverage.

Its essential finding is that, for the last two decades or so, courts have

found "serious, reversible error" in a large fraction of the cases they

reviewed. These errors, the report claimed, often involved weak or

incompetent defense attorneys and the withholding of important evidence

from the juries.

But notice what the report did not say. Its authors did not attempt to

discover whether any innocent person had been executed, and they made no

claim that this has happened. Instead, they said that the large number of

appeals leaves "grave doubt whether we do catch" all of the errors. The

clear implication is that, were the truth known, we might well be

killing many innocent people.

But that truth is not known. The Death Penalty Information Center, a

rallying point for opponents of execution, reports that since 1973, when

the Supreme Court reinstated the death penalty, 69 people have been

released from death row after they were found to be innocent. But the

center does not say that any innocent person has been put to death,

though if it had found such a case it surely would have proclaimed it.

The Columbia University report shows that death sentences are intensively

reviewed by appeals courts. Some critics of these reviews think they

take too long and involve too many unnecessary bites at the apple, and

that may be true. But if we are to err, it is best that we err on the

side of safety.

Nine or 10 years usually pass between the imposition of the death

penalty and its being carried out. It took 19 years and appeals heard by

more than 30 judges before Gary Graham was executed last month in Texas.

It is hard to imagine that this much time is necessary for an adequate

appeal, but offsetting the cost and delay is the assurance of only a

small chance that an innocent person will be killed. The 5,760 death

sentences handed out since 1973 had, by 1995, led to only 313 executions.

Mr. Liebman suggests that the high rate of appeals means that serious

errors are often made by the trial courts. But before we can accept that

conclusion, we must first know whether the errors were serious enough to

affect the outcomes of the cases when they were sent back for new

trials. Did an "error" cause a new trial that set aside the death

penalty? Unfortunately, Mr. Liebman was able to learn this for only a

small number of the reversals.

Because of Supreme Court decisions, every death-penalty conviction

leads to an appeal to the state's highest court. About two-fifths of

these cases were reversed. As I read the report, we have no information

about what happened in the new trials.

Then there are state appeals after convictions. These also led to many

reversals, but we don't know what happened to the great majority of these

cases when they were retried because trial courts ordinarily do not

publish their findings. Mr. Liebman and his colleagues managed to find

301 cases that had been retried, but we have no idea whether these were

representative of all of those appealed or were only a few dramatic ones

that somehow came to the attention of outsiders.

Of these 301 new actions by trial courts, 22 found that the defendant

was not guilty of a capital crime, 54 reimposed the death sentence and

247 imposed prison sentences.

Then there were appeals to the federal courts that also led to

reversals in about 2/5 of the cases, but again we are not certain

what happened in all the new trials.

The report also lumps together cases going back to 1973 with those

decided more recently, even though the Supreme Court in 1976 created new

procedural guarantees that automatically overturned many of the

death-penalty decisions made between 1973 and 1976. It is not clear from

the Columbia report what fraction of its reversals date back to these

big changes in the rules.

In short, in the vast majority of death-penalty cases we have no idea

whether the finding of error that led to a reversal was based on a legal

technicality, a changing high-court standard about how a capital crime

ought to be tried or a judgment that the defendants might be innocent.

All we know for certain is that a lot of death-penalty cases are

reviewed over a long period of time -- a fact that dramatically reduces

the chances of innocent people having been executed.

More procedural reforms may be coming. Congress is now considering a bill

that would require federal courts to order DNA testing, at government

expense if the defendant is indigent, whenever DNA evidence from the

crime is available. It also would require states seeking federal

crime-control funds to certify that they have effective systems for

providing competent legal services to indigent defendants in

death-penalty cases.

But more might be done at the state level. States ought to have laws that

create imprisonment without possibility of parole for first-degree murder

convictions, and the judge in every such case should instruct the jurors in

the sentencing phase that they can choose that or the death penalty. This

allows jurors who may have some doubts about the strength of the evidence

or some other plausible worry to hedge their bets if they are so inclined.

Not every state now has such laws. In Texas, the alternative to the death

sentence is life in prison, but without an absolute guarantee that the

offender will actually spend his life there. Jurors rightly suspect that

the perpetrator will find some way to get back on the street, and so

they often vote for death.

The American Law Institute, a group of legal scholars that designs uniform

state legal codes, has recommended that even when a jury decides that

capital punishment is appropriate, the judge should be allowed to bar

the death penalty if the evidence "does not foreclose all doubt

respecting the defendant's guilt." The states have not adopted this

rule, but perhaps they should, especially if this change could be

coupled with procedures designed to reduce the seemingly endless number

of post-trial appeals.

In the meantime, we ought to calm down. No one has shown that innocent

people are being executed. The argument against the death penalty

cannot, on the evidence we now have, rest on the likelihood of serious

error. It can only rest, I think, on moral grounds. Is death an

excessive penalty for any offense? I think not, but those who disagree

should make their views on the morality of execution clear and not rely

on arguments about appeals, costs and the tiny chance that someday

somebody innocent will be killed.

(