Of course it is often difficult (but not impossible in all situations) to prove a negative. I.e., it is impossible to prove ghosts don't exist, hence all those popular ghost hunting shows (people seem to like the idea of possible sensationalistic findings no matter how unlikely, which I think is what also, partly, drives the voter fraud stories). Moreover, it is difficult to prove a negative about the degree of magnitude of certain phenomena if one keeps expanding the theory ad hoc. I.e., impossible to prove that I don't have an huge Oedipal complex  when you only have evidence for a small one IF you modify the theory ad hoc, in light of the evidence, to include repressed complexes. In the case at hand, it is impossible (in a strict sense) to prove that there aren't larger amounts of voter fraud than what has been uncovered. Nonetheless, people have TRIED hard, very hard, to come up with significant numbers of this kind of fraud....and it just has never been secured.

 

This raises the question: what amount of fraud is significant, i.e., worthy of additional action? Related, and equally important, is the question of what actions are likely to be effective AND not produce perverse consequences? 

Putting these two elements together (how much fraud must there be before we agree there is something to be fixed and what effective fix doesn't create perverse consequences) I think the decision rule is somewhat like the rhetorical point some voter fraud activists have already identified: every fraudulent vote denies somebody their proper vote. Likewise, everybody denied their proper vote with a new policy (this can be studied, in theory, through ex post evaluations or ex ante reasoning) intending to reduce fraud is a count against that law. I.e., a sort of cost benefit analysis of how many legit voters you are frustrating vs. fraudsters you are frustrating. It may be that a one for one trade off is not important here, but the point is to put the numbers out on the table: Are we preventing thousands of legit voters from voting for each fraudulent vote we prevent? Tens of thousands to one? Once we view it this way, we begin to ask other questions: hmmm....maybe there's another solution, or maybe the solution is worse than the problem, or maybe we don't understand the problem well enough to know what solution could possibly be effective (that is, if the rare cases of fraud tend to be produced by felons that were  not aware of the eligibility standard, than IDs and aggressive volunteer poll watchers may not be the solution likely to stop this problem).

In short, given how harmful some reforms being proposed can be, I would argue that proponents need to demonstrate that fraud is occurring on the order of SEVERAL magnitudes higher than the available evidence to outweigh the costs of some proposed reforms (and only if those reforms (a) truly address the fraud and (b) address the fraud that is demonstrated, not unrelated kinds of fraud).

An example of laws that prevent fraud but likely would not be found to reduce the ability of legit voters from voting are fines. If the fines are high enough nobody in their right mind would commit fraud, of the sort under discussion, if they think it is even remotely possible that they would be caught since the possibility that the fraud that you commit is likely to affect any election is very low (and note that the more fraud you commit the more likely you are to get caught). Another possible reform (or continuation of existing policy) would be to simply continue investigations into fraud in a manner that does not intimidate legit voters. Even if there is little fraud to prosecute and the amount that exists is very unlikely to swing an election, it seems reasonable to assure the public that possible cases are tracked. In short, if people persist in believing that, somehow, legitimate cases of fraud are dropped (and that, somehow, only the poorer cases are chosen to be debunked by prosecutors), then I am all for spending some more money on hunting these ghosts, only if to keep the word out that fraud will be prosecuted and to continue to demonstrate the nature of the problem (i.e., demonstrate that its scale is far, far smaller than many believe). Finally, if we find that felons are the ones that a mistakenly voting, education or clarification/simplification in the process of restoring voting rights seems a reasonable solution (this might also argue for simply returning their rights more rapidly).

 

In other words, until the scale is shown to be larger, why fix what is not broken? Let's just reassure the public and stick with policies that do not have perverse consequences. In short, it seems as if the problem of voters committing fraud is a mosquito and some are recommending that we go at it with a shot gun.

 

-Doug Hess