Bringing some empirical input from a little different presidential electoral system.
Here in Brazil the system for presidential elections has two turns. All candidates run for the first turn. If someone has >50% votes, that candidate wins. If none has it, the two with most votes run for the second turn, when the most voted wins.
In the last three elections (2002, 2006, 2010), there were 6, 8 and 9 presidential candidates. There was a second turn in all of them, with candidates from the same parties (Worker’s Party and Social-Democratic—the labels don’t mean much, though) at all three. The third-place candidates, though, were from different parties in each election, with 17,9%, 6,85% and 19,3% of the votes. The 2006 election was Lula’s reelection, and the votes for the first and second place in the first turn were the closest of all three years (48,6%/41,6%).
I’m not sure about what the data means and how Eliezer’s line of thought would have to change to apply here. I think it changes the picture a little because even getting to the second turn is seen as some kind of victory. So people can vote in the candidate they really prefer because, hey, maybe they can get to the second turn and have a chance! And the chance that there will be no second turn is small, so the part about “keeping the wrong lizard out” is postponed.
In France we have a run-off system too, and I think some complaints about American politics I hear would disappear if they adopted that—though others might appear, for example regional parties like there are in Canada.
Unlike what you describe in Brazil, it’s not that rare for candidates that are not part of the two “main” parties to get to the second turn (Le Pen did in 2002), or even to get elected (The centrist Giscard in 1974). The “main” parties are also more likely to split up / reorganize themselves than the US ones.
The Czech Senate (upper chamber of the parliament) uses top-two runoff too and a lizard-lizard competition in the second round happened in 14 out of 27 districts, which is only slightly more than half (lizard defined as a candidate of one of the two strongest parties). Remarkably, there was no lizard-free second round, at least one lizard succeeded everywhere. In the second round, 20 lizards won. It means that out of 13 lizard vs. non-lizard competitions, 6 were won by lizards. It seems to indicate that lizard victories aren’t largely due to “strategic” voting: if it were so, non-lizards would massively outcompete lizards in a direct confrontation.
though others might appear, for example regional parties like there are in Canada
What reason for appearance of regional parties is present in a top-two runoff system and absent in the plurality system?
Actually the current polarization is recent, I think. In 1989 some guy from an obscure party got elected. But brazilian democracy itself is young: we’ve only had 6 direct presidential elections since the demilitarization in 1985, and in the last 5 the second turn, when it happened, was between the Worker’s and Social Democratic parties. So the polarization is recent, but there’s not much data before it.
Bringing some empirical input from a little different presidential electoral system.
Here in Brazil the system for presidential elections has two turns. All candidates run for the first turn. If someone has >50% votes, that candidate wins. If none has it, the two with most votes run for the second turn, when the most voted wins.
In the last three elections (2002, 2006, 2010), there were 6, 8 and 9 presidential candidates. There was a second turn in all of them, with candidates from the same parties (Worker’s Party and Social-Democratic—the labels don’t mean much, though) at all three. The third-place candidates, though, were from different parties in each election, with 17,9%, 6,85% and 19,3% of the votes. The 2006 election was Lula’s reelection, and the votes for the first and second place in the first turn were the closest of all three years (48,6%/41,6%).
I’m not sure about what the data means and how Eliezer’s line of thought would have to change to apply here. I think it changes the picture a little because even getting to the second turn is seen as some kind of victory. So people can vote in the candidate they really prefer because, hey, maybe they can get to the second turn and have a chance! And the chance that there will be no second turn is small, so the part about “keeping the wrong lizard out” is postponed.
And more, voting for the right lizard doesn’t change the probability of the wrong lizard’s first round victory.
In France we have a run-off system too, and I think some complaints about American politics I hear would disappear if they adopted that—though others might appear, for example regional parties like there are in Canada.
Unlike what you describe in Brazil, it’s not that rare for candidates that are not part of the two “main” parties to get to the second turn (Le Pen did in 2002), or even to get elected (The centrist Giscard in 1974). The “main” parties are also more likely to split up / reorganize themselves than the US ones.
The Czech Senate (upper chamber of the parliament) uses top-two runoff too and a lizard-lizard competition in the second round happened in 14 out of 27 districts, which is only slightly more than half (lizard defined as a candidate of one of the two strongest parties). Remarkably, there was no lizard-free second round, at least one lizard succeeded everywhere. In the second round, 20 lizards won. It means that out of 13 lizard vs. non-lizard competitions, 6 were won by lizards. It seems to indicate that lizard victories aren’t largely due to “strategic” voting: if it were so, non-lizards would massively outcompete lizards in a direct confrontation.
What reason for appearance of regional parties is present in a top-two runoff system and absent in the plurality system?
Actually the current polarization is recent, I think. In 1989 some guy from an obscure party got elected. But brazilian democracy itself is young: we’ve only had 6 direct presidential elections since the demilitarization in 1985, and in the last 5 the second turn, when it happened, was between the Worker’s and Social Democratic parties. So the polarization is recent, but there’s not much data before it.