Can you trust political trends?

If you plot up the presidential results since Reagan came along, you get something interesting: a nice little regression. Using results from 1980-2008 (excluding 1992), you would have predicted that Obama would win in 2012 with about 53 percent of the vote, close to the actual result of about 51 percent, which is pretty surprising considering this method does not take into account any consideration of economics, who the Republican nominee was, or anything else of that nature. Here's the graph:

We should interpret this graph as showing that it has been a little less likely for the Republican to win the presidency in each successive cycle as of late. In 2016, it will be even less likely than it was this year. However, unlikely does not mean impossible, and it would still be consistent with the long-term trend if a Republican wins in 2016.

But how much faith should we put in political trends like this? I went looking for another example to test how well they can predict the future. In this case, it's an increasing Republican trend. I used the trend from the first 20 years of data to predict the election in Year 24, and the results from the first 24 years of data to predict the election in Year 28, and so on. Here's what we get:

The predictions are, for the most part, pretty good'not that great, honestly, but certainly no worse that many of the pundit predictions we saw in 2012.

Until 1932. The trends predict a Republican victory with 60 percent of the vote. Actual Results: The Republican only managed 40 percent of the vote. That's waaaaay down there in the lower right hand corner.

But that's not fair! you might say. These were unusual circumstances! The Great Depression! A popular Democratic governor from a Republican state running for president!

And that's the point. Circumstances change. Parties change. Generic Republican almost never is paired off against Generic Democrat at the national level. Political trends can be interrupted.

Such an interruption would probably be pretty obvious well in advance. Several possibilities would be an extraordinary economic or geopolitical disruption (worse than 2008), a major third party candidate that draws primarily from voters who otherwise would have voted Democratic, a major criminal scandal in the either party, or a major policy shift in either party that alienates a substantial portion of their base. For the interruption to shift votes from Democrats to Republicans permanently, it would have to be accompanied by a major change in the Republican party, namely, the elimination of Republican's openly hostile tone towards so many Americans (is this even possible?).

A few more considerations below.