While I wrote earlier in the week that Europe’s voters have been shifting to the right, it might be thought that the story is different in Latin America, where Peru last weekend elected leftist Ollanta Humala as its new president. Humala defeated the right’s Keiko Fujimori (daughter of disgraced former president Alberto Fujimori) in the second round by the narrow margin of 51.5% to 48.5%.

Humala has since been trying to reassure Peru’s business community that he is basically a moderate, and has promised to maintain economic growth and pursue close ties with the United States. In the past, however, he has been regarded as an ally of Venezuela’s far-left leader Hugo Chavez.

But the perception of a swing to the left needs to be reviewed in light of the full results. In the first round of the election, held two months earlier, there were five serious candidates: Humala had 31.7%, Fujimori 23.5% and three centrist candidates had 44.0% between them. (Adam Carr has the figures.

So if the centrists had been able to pool their votes behind a single candidate — either via a system of preferential voting, or just by agreeing on one of them to be their standard-bearer — they almost certainly would have won: they would have had enough votes to get ahead of either Humala or Fujimori, and supporters of each of them would have preferred a centrist to their ideological opposite.

Instead, however, Peru has a simple two-round system: only the top two finishers from the first round go on to the second round. It’s impossible for a candidate to win from third place, as sometimes happens under preferential voting in Australia. Most of the time the two systems are equivalent, but this is a good example of a case where they were clearly not.

It’s also an important reminder that election results are not always quite what they seem. Results that are remembered by history as marking some decisive shift of opinion may actually be artefacts of a particular electoral system. Public opinion may not fairly translate into political outcomes.

For example, recall the 1998 election in Australia — frequently cited as proof that voters approved John Howard’s GST when he risked an election before its implementation. In reality, of course, the majority voted against re-election of the Howard government: Kim Beazley’s ALP recorded just over 51% of the two-party-preferred vote. But the vagaries of single-member electorates gave the Coalition a narrow majority.

A more significant instance was highlighted last year by Dan Berman at fivethirtyeight.com. The South African elections of 1948 and 1953, which put the National Party in power and began the era of apartheid, did not actually represent a shift in majority support. The more moderate parties still commanded a majority of the vote, but many of their votes were wasted in safe urban constituencies while the Nationals benefited from the weighting given to rural areas.

But this is politics; the perception tends to become the reality. South Africa really did change in 1948, and the fact that the change was in a sense an artificial product of the system ceases to matter much in the long sweep of history.

Similarly, Humala is the legitimate president-elect of Peru, and he’s entitled to make what he can of his opportunity. If he succeeds, this could be remembered as the time his country made a decisive choice. But it’s worth noting that a different and by most measures fairer electoral system would have delivered the election to one of his opponents.