Any democracy should aspire to a perfect, fraud-free voting system. But today’s loudest complainers on this issue — mainly Democrats complaining about Republican election victories — should be careful what they wish for, because they might get it.
To see why, let’s apply a little game theory to the problem. Ask yourself under what circumstances vote fraud will be most effective, and have the least risk of being detected.
Other factors being equal, the probability that vote fraud will be detected will scale up with the number of voting machines you tamper with rather than the number of votes. Thus, tampering will be most effective and least risky when relatively many people are being served by relatively few machines — that is, areas of high population density.
Another factor favoring high-density areas as vote-fraud sites is that poll-watchers are less likely to know most of their neighbors by sight, and thus more likely to pass a bogus voter. Also, in conditions of high traffic at the polling station people voting multiple times are less likely to be spotted.
We should expect vote fraud to be more common on behalf of the party controlling the local administration of a ward than on behalf of its opponents, because the controlling party will be better able to get covert access to the machines and better positioned to suppress any evidence of tampering.
Finally, we should expect vote fraud to be generally more prevalent among parties who think it less likely they’ll win an honest vote — that is, parties in an overall minority position. Especially, we should foresee fraud from minority parties who have reason to believe that polls and demographic projections overstate their strength.
What do these predictions tell us to expect, given the facts of American political demography? Well…the most important of these facts is that Democrats tend to carry areas of high population density and Republicans areas of low population density. To an excellent first approximation, the Democrat vote is urban, the Republican vote is rural, and the suburbs swing by local population density. Finally, it is also a well-known fact among political demographers that polls (especially those canvassing the general population rather than ‘likely voters’) tend to oversample Democrats.
(American voting patterns used to be more complicated, mainly because the Democratic party once had a strong rural Southern wing. But, as I have written elsewhere, the Democrats have spent the last forty years reducing themselves to a regional party of the coastal metroplexes.)
This tells us to expect that vote fraud will be primarily be a Democratic crime, especially practiced by Democratic urban party machines and increasing as the now-slender national Republican majority increases.
Both major parties behave as though they believe this to be true. This news story is entirely typical of the resulting skirmishes — a Republican legislature proposes a requirement that voters show photo ID, a Democratic governor vetoes it.
I’m in favor of moves to clean up electoral fraud, but that’s an easy position for me to take because I loathe both major parties. Partisan Democrats should be less sanguine; on the evidence, a perfectly clean voting system would cost them painful losses in swing districts without any corresponding gains in Republican heartland country.