the "it's really hard for a party to win >2 elections in a row" fallacy

[Mr. Morden]

I see people making this claim all the time here: "It's really hard for a party to win a third consecutive presidential election.  And even if they do, winning a fourth term is nearly impossible."

Why do people believe this?  The data don't really bear it out.  It's true that, at least since about 1900, when a party wins one presidential term, they tend to be enormously successful at winning a seond term (1980 being the only case in the 20th century when a party was voted out of the White House after just one term).  But after a party wins *two* terms, and is going for a third or fourth or more, is there really any evidence that the a priori probability for which party will win each election is any different from 50/50?  Not as far as I can tell.

There are two ways to look at these probabilities.  The first is more subjective.  You look at each election individually, and, based on the circumstances of the campaign, and how close the outcome was, you guess at how great of a chance the losing candidate really had.  So let's look at the last 5 presidential elections in which a party was coming off of holding the White House for 8 years: 1960, 1968, 1976, 1988, and 2000.  The incumbent party won only one of those contests (1988), but every single one of the other 4 contests was so close that it was virtually a tie.  The slightest change in circumstances in the incumbent party's favor could have turned each of those elections in the other direction, so from that perspective, I'm not convinced that it's "nearly impossible" for a party to win 3 consecutive times.

The other way to look at it is just by the final outcome: In a given set of circumstances (a party wins 2 consecutive presidential elections, and is going for a 3rd term), what percentage of the time do they successfully get a 3rd term?  In all of American history, there've been 13 presidential elections in which the incumbent party was running for a third terms (not counting 2008, since we don't know the outcome yet).  7 or those 13 times, the incumbent party successfully won a 3rd term.  So it looks like it's not so improbable.  It's true that *recently*, in the last half century, the incumbent party going for a third term hasn't been as successful.  Didn't happen in 1960, '68, '76, or 2000, though it did happen in 1988.  But so what?  5 trials isn't enough to tell you much one way or the other.  In fact, here, I'll flip a coin 20 times, and here are the results:

THTTHHTTTT
HTTHTTHTHT

Wow, look at that.  In those last 10 flips, there isn't a single sequence of three or more flips that give the same result.  If those were presidential elections (with H & T two different parties), then that would correspond to a period of 40 years over which no party held the White House for more than 8 years.  I guess that means that once you win two consecutive terms, your odds of winning a 3rd term are really low, right?  No, of course it doesn't.  Is it clear now how ridiculous that line of thinking is?

[Mr. Morden]

Well, clearly there are individual elections in which each party's chances are significantly different from 50/50, depending on the particular circumstances of that election.  But what I see people do all the time here is try to use historical precedent to predict how hypothetical elections in the future will play out.  For example, people argue that McCain will have a really tough time winning in 2008 because "it's nearly impossible for a party to win 3 consecutive terms".  Or that, even if he does win, the GOP is doomed in 2012 because winning 4 consecutive terms is nearly impossible.  Or that, if Obama wins in 2008, and is reelected in 2012, we'll likely have the GOP win in 2016 because of the difficulty of a party winning 3 terms in a row.

What I'm suggesting is that each of those arguments is bogus.  McCain is in fact the underdog in 2008, but it has nothing to do with historical precedent on parties winning 3 or more terms.  Those historical arguments don't hold up in this case, as I argued in the OP.