Since 1789, our Constitution has enshrined disproportional representation, in both the United States Senate and in the Electoral College. For all of our complaining about the Electoral College on grounds of disproportional representation, we rarely if ever hear the same complaint about the Senate (although such was not the case in the early days of our government). Indeed by comparison with the Senate, the College is a paragon of proportionality.
However, another feature of the College is perhaps even more concerning: with winner-take-all allocation of state votes, the Electoral College increases the sensitivity of election results to error, whether administrative or fraudulent in nature.
Winner-take-all allocation is not inherent in the College itself: how each state casts its College votes is determined by the state itself, and this interesting article explains how the winner-take-all system took hold. (Essentially, states wanted to maximize the power of their electoral vote for their preferred presidential candidate.)
Why is the Electoral College system less error-tolerant than a staight popular vote? In a close election one or two closely-contested states can be tipping points in the Electoral College vote (e.g. Texas and Illinois in 1960, Florida in 2000). The fact there is no widespread voter fraud is irrelevant. Through winner-take-all allocation, a handful of votes could be worth 3.8% of the College total (Illinois), 4.8% (Florida) or 7% (Texas) – easily enough to swing the overall contest. No widespread error is necessary – in the case of fraud, just a willingness to manipulate a few locales with close votes and dense populations. If notions of fraud seem far-fetched, there is still discussion about the November 1960 presidential vote in Illinois, although fraud was never conclusively demonstrated. As for election hacking, I’ve mentioned previously that this isn’t conceptually diffucult, and security experts have demonstrated its in-principle feasibility.
Of course a straight popular vote can also be very close (and several have been, particularly the popular vote in 1880) but then the argument we heard before the election holds more sway: the dispersion and variety of state voting processes makes fraud a greater challenge.
It’s a legitimate concern when a small number of erroneous votes, well within the number found in one metropolitan area, might even conceivably alter a national election outcome. The Obama administration is quite right to look into the matter of election security. I personally don’t care that these scenarios are not probable – we need full confidence in electoral outcomes, particularly close outcomes, and we’ve had at-risk scenarios as recently as 2000.
A lack of confidence in voting results also erodes confidence in the underlying government. Witness the dispiriting spectacle of a president-elect questioning an election count after he won, an episode that hopefully will never be repeated. But to get to indisputable election confidence, we need a system that is certiably reliable. An extensive debate about electronic voting is now ongoing, and while it’s probably an over-simplification, my reading is this: certifiable security is still a problem, but in principle electronic balloting is more accurate and certifiable than hand vote-counting. In the US worries about privacy and states’ rights will compound the security discussion, but it would be great to see a serious commitment in this direction.