America’s death penalty has been abolished once before, by court decisions in 1972. What did we get for it? Ever since, we have had Charles Manson grinning at us from his prison cell, and now he is networking with his fan base on smuggled cell phones. He has been let off far too easy for his seven horrible murders. While life in prison is adequate punishment for the typical heat-of-passion homicide, it is a travesty of justice for the worst of the worst.

For killers such as Timothy McVeigh, Ted Bundy or the D.C. Sniper, nothing less than death is justice.

There is also good reason to believe an effective, enforced death penalty saves innocent lives through deterrence. It is a basic principle of human behavior that incentives matter. Increase the cost of doing anything, and fewer people will choose to do it.

Opponents say this does not apply to homicide because most homicides are impulsive, but the death penalty is not for most homicides. It is for premeditated murder, and only the worst of that limited subset. In the late 1990s and 2000s, a wave of econometric studies showed a deterrent effect. The best-known study, from economists at Emory University, estimated 18 innocent lives were saved per execution. The Politically Correct forces in academia counterattacked, of course, but the authors responded and showed their results held up even after taking the criticisms into account.

Academic debate goes on, but one thing is clear: The common-sense conclusion that an enforced death penalty will save innocent lives through deterrence has not been disproved. If the death penalty deters and we abolish it, we kill innocent people. If we keep the death penalty and it does not deter, we have simply carried out a sentence that is independently justified as a matter of justice.

Americans’ support for the death penalty as fundamentally just has remained remarkably high and stable over the years. Opponents have therefore turned to other arguments, which have been shot down one by one. The notorious Baldus study on racial discrimination in Georgia’s death penalty went to trial in federal district court. The court found that, to the extent the data proved anything, they proved just the opposite of what Baldus claimed: There was no racial discrimination effect.

The so-called “innocence list,” presently numbering 152, is regularly referred to as a list of people who are actually innocent, but that claim is a discredited fraud. Placement on the list does not require actual innocence, proof of innocence, or even evidence of innocence. North Carolina triple murderer Timothy Hennis was on the list for years before improved DNA technology provided a conclusive match. He had been paraded around as an innocent man, grievously wronged by the system, yet he was guilty as sin the whole time.

Supporters of Troy Davis succeeded in creating a media firestorm with assertions he was innocent. The claim was fully heard in a federal district court before his execution, and the judge found the claim was “smoke and mirrors” and Davis was “not innocent.” Far from showing a dysfunction of the system, this case showed how very careful the system is on claims of actual innocence.

The system presently takes far too long and costs far too much, to be sure. However, most of the additional cost and delay comes from litigating issues that have nothing to do with actual guilt of the crime. To the extent we spend money making sure we have the right person in a capital case, there is no moral justification for spending a penny less if we are going to lock him up and throw away the key instead. We can save large amounts of money, though, by placing reasonable limits on the penalty phase of the trial that follows the guilt phase and especially on the review of the penalty decision.

After one full round of review (one appeal and one habeas corpus petition), we can simply limit all further reviews to questions of actual innocence. Far from increasing the danger of executing an innocent person, this reform would decrease it by requiring the defendant’s lawyers to focus on innocence claims. In the very rare cases of actual innocence, it would reduce the time spent in prison. Overall, because very few cases have even a ghost of doubt of guilt, this reform would save large amounts of money, dramatically reduce delay, and greatly improve the quality of justice.

Mend it; don’t end it.

Kent S. Scheidegger is legal director of the Criminal Justice Legal Foundation.