How not to rig an election

Nicolás Maduro’s brief flirtation with democratic accountability appears to be over. As no doubt you’ve heard, the Venezuelan president’s hand-picked electoral commission announced on Monday that he was the winner of Sunday’s presidential election, allegedly with 51.2% of the vote on the basis of 80% counted, as against 44.2% for the opposition’s Edmundo González and 4.6% in aggregate for the other eight candidates. (See Friday’s preview here.)

That’s a modest enough number – it doesn’t intrinsically ring alarm bells in the way of, say, the 80.1% claimed by Belarus’s Alexander Lukashenko in 2020. More damaging is the fact that there was an unexplained delay before any figures were released, that no subsequent details have been forthcoming (despite promises, and despite the fact that the system is electronic, so tabulating data should be easy), and that the figures are radically inconsistent with what the opinion polls had predicted.

But the clincher is in the numbers themselves. As observers have pointed out, they show an uncanny degree of precision, with zeros in the next four decimal places for all three: Maduro has 51.199997%, González 44.199999% and others 4.600004%. A Spanish expert calculates that the chance of producing that result randomly is one in a hundred million.

Why, you might ask, only four decimals – shouldn’t a fake result be exact, to any order of precision you want? No, because votes only come in whole numbers. If you start with a (possibly real) number of votes and apply those nice round percentages to them, you’ll get a number with some fractional component: Maduro’s vote, for example, comes out at 5,150,092.288. Rounding that to the reported 5,150,092 produces the small discrepancy in the sixth decimal place.

So this is a smoking gun, pointing to an electronic version of what I referred to on Friday as “direct old-fashioned ballot-stuffing.” Of course it’s not all there is; the opposition also has copies of actual tallies obtained from a large number of polling places, which point to the expected landslide victory for González. But if anyone tries to tell you (as a few unreconstructed leftists are still saying) that the opposition has produced no evidence of fraud, the government’s own numbers are enough.

And when the difference between claimed and real results is so large, lots of people can figure that out from their own lived experience. Hence the widespread nature of the popular anger, which looks to be at a different level from previous anti-Maduro agitation. That doesn’t mean it will be successful; the government is well-armed and ruthless, and is still portraying all its opponents as “fascists, criminals and imperialists.” But survival from here is going to be difficult.

Maduro also looks increasingly isolated at an international level, with only a small band of fellow-Putinists still in his corner. But I want to take a proper look at that tomorrow, by which time there might be more news.

And speaking of numbers, just a quick update on the betting odds that we discussed on Monday. The gap between Trump and Harris has closed quite sharply: according to Lott and Stossel’s aggregator, the implied chance of a Republican victory has fallen since then from 57.2% to 53.7%, and the Democrats’ has risen from 42.7% to 46.0%. Sportsbet now quotes Harris close to even money, at 11-10 against.

One obvious explanation is an Economist/YouGov poll released this week that gives Harris a two-point lead, a gain of five points for her in the space of a week.