Richard L. Hasen (U.C. Irvine Law) has posted The Case for Fraud on SSRN. Here is the abstract:
Laws designed to reduce the level of voter fraud are flawed at multiple levels. There is no good empirical evidence that voter fraud is a serious problem for the election system as a whole--although there are isolated cases in which particular jurisdictions experience serious voter fraud problems. Moreover, the negative effects of laws directed at voter fraud far outweigh the benefits. For example, Voter ID laws suppress turnout and disproportionately effect disadvantage groups. These points are well-established by a consensus of the legal and political science literature.
What is less well understood is that "traditional" voter fraud laws--including the prohibitions on "double voting" and voting on behalf of deceased persons--are deeply antidemocratic and anti-egalitarian. The empirical literature strongly suggests that almost all documented instances of voter fraud are responses to democratic failures. The most common scenario in which significant voter fraud occurs is one in which structural features, including poverty, massive failures in public education, and learned hopelessness result in disproportionately low voter turnout rates among disadvantaged persons. In these scenarios, voter fraud acts as a corrective--approximately that outcomes that would have occurred but for the unjust structural disenfranchisement of minority groups and low income persons. That outcome should be embraced for the reasons identified by Nicholas Stephanopoulos in his recent work on the alignment of policy outcomes with majority policy preferences. Most voter fraud is actually supported by democratic theory.
This analysis does not lead the conclusion that voter fraud should be legalized or decriminalized across the board. Those options would enable powerful groups to utilize voter fraud to enhance their political clout at the expense of the very groups that voter fraud now empowers. Instead, I propose an Equal Protection Clause affirmative defense to voter fraud prosecutions. If the defendant can establish that the voter fraud had a reasonable likelihood of producing election outcomes more representative of the population of the relevant voting district as a whole, then the defendant would be exonerated. As a very rough approximation, this rule would permit most voter fraud on behalf of Democratic candidates but would not be available to those who commit voter fraud on behalf of Republican candidates.
