There's nothing arbitrary about it. A wealthy white woman who violently kills a poor black or hispanic man in the Northeast of the US stands virtually no chance that she will get the death penalty. A poor black or hispanic man who violently kills a wealthy white woman in Texas stands an excellent chance that he'll get the death penalty.
Some people oppose capital punishment because they believe it is randomly assigned. In truth, the extent to which you can predict the assignment of capital punishment is alarming.
The factors which are the main predictors of capital punishment assignment are geography, economic class of the accused, race of the accused, gender of the accused, gender of the victim, race of the victim, and economic class of the victim.
If the death penalty was randomly assigned (say we spin a roulette wheel at the conclusion of all murder convictions and we execute everyone who gets a lucky 13), that would be an improvement in terms of the equal administration of justice.
In addition to the bias with which the death penalty is assigned, please also bear in mind that capital punishment is neither an effective deterrent nor cost efficient (as some mistakenly believe).
The extent to which the death penalty fails as a deterrent is well illustrated by this graph which shows that the states which have the death penalty generally have much higher murder rates:
![](http://www.deathpenaltyinfo.org/MurderRateGraph.gif)
The South accounts for 82% of executions, yet the murder rate rises there; while the Northeast accounts for less than 1% of executions, and the murder rate falls there. If we have a societal need to have murderers (along with some wrongly accused) killed for the sake of our desire for retribution (and when a death penalty case is called "The Defendant vs. State" -- the "State" is "us" so the death penalty kill is explicitly done in our name and on our behalf), we should not kid ourselves with the false idea that capital punishment is a deterrent. That's just a myth some people tell themselves to assuage the guilt, and the studies bear that out. See <
http://www.fbi.gov/ucr/cius_02/02prelimannual.pdf>.
Moreover, those who believe that capital punishment is more cost efficient than life without parole are mistaken.
The pretrial investigation of a capital case is at least 3 times costlier, trial is at least 15 times costlier, appeals are at least 20 times costlier. Contrary to what some people might wrongly presume, the pretrial investigation and the trial are by far the costliest part of the process, not the appeals or the actual incarceration. See “Performance Audit Report: Costs Incurred for Death Penalty Cases: A K-GOAL Audit of the Department of Corrections,” December 2003.
Even when you figure in the extra costs of a longer incarceration in a life without chance of parole conviction, capital cases are still much more expensive (nearly 150% as expensive on average) because of the much greater costs of pretrial investigation and trial associated with such extreme penalties. See Indiana Criminal Law Study Commission, January 10, 2002.
Finally, some might say that the costs of the death penalty are unnecessarily inflated because of the judicial system. The simple answer is that our justice system is already so pressed by the desire to cut costs that the system has already become haphazard to the degree that it has been responsible for the killing of several wrongfully accused on our behalf and in our names. Any debate about further streamlining and cost cutting for the capital punishment assignment system must begin with the question "how many of the wrongfully accused shall we tolerate as acceptable collateral damages to our justice system." Personally, I think our system already tolerates too many collateral kills of the wrongfully accused so I think the level of care (and therefore the cost) of the system should be improved and not cut, but if anyone wants to cut the cost and corresponding level of care we exercise in applying the death penalty, that's a public debate we should embrace, but the first question in that debate requires the identification of the proper threshold for the erroneous killing of innocents.
Perhaps we should start the debate by answering that question.