Wednesday, March 25, 2009

A political puzzle...

Following a hunch, I pulled together historical data on US political participation. Here's the result:


The turnout data is from data is from wikipedia. Non-presidential election years are averages for the previous and following election. Polarization data comes from the difference between House party medians in the first two DW-NOMINATE dimensions. Basically, polarization measures how much Representatives' votes broke along party lines.

Here's the puzzle: up until around 1900, the two lines track quite closely. The raw values are about the same, and they tend to move up and down together. But after 1920, the polarization and participation stop tracking each other. Why?

7 comments:

Chickpea said...

Hi Abe,

I just happened upon your blog via tpmod. I'm not a political science major, but I have a thought about why the lines diverge. Do you know when unions started forming in the US? When did 'big organization' lobbies start to affect voting? Just a thought...

Dave said...

I spent some time looking into this.

My first thought was that as America became more prosperous citizen’s worried less about politics and thus started a downward trend in turnout independent of partisan politics. I ran a regression using voter turnout data from IDEA international and GDP per capita from the Penn State World Table for a cross section of nations and election years (from 1972-2004).

I regressed voter turnout on log of GDP per capita and both Freedom House’s Political Rights and Civil Liberties index. It seems I was completely wrong (at least at first blush) with a 1% increase in GDP per capita correlating with a 5% increase in voter turnout (t=7.37). Of course there might be a third factor moving both together (i.e. historical longevity, many rich nations have been democracies for a long time, robust free markets, or any other number of confounding variables.)

The Freedom house indices are very odd. As expected, a poor rating on civil liberties is negatively correlated with voter turnout (t=-2.53) however a poor rating on political rights is positively correlated with voter turnout (t=2.53). Of course the coefficients are small enough (1.5% and -1.8% for each increase on the 7 point scale) that I’m willing to write them off as a type 2 errors, if you don’t have any objections, and head back to the drawing board.

Dave said...

I showed the chart at a get together with the Cowles (Rob and Erin) and Matt Gee last night and they also pointed out that Universal suffrage might have something to do with it. Can you break out the turnout numbers by gender?

Margaret said...

It seems like a lot was changing in that "Fin de Siecle" period. Dave mentioned Rob's very good point about Universal Suffrage. There was also a war on pretty close in there, and I'd say there's plenty of support for the idea that WWI was the end of an era. You've also got the big bosses of Tammany Hall, which started being fought in the early 1900's. (Can you tell I'm just pulling on all my half-remembered pieces of American History?). I guess what I'm saying is, I don't have a clear thing to point to, but is it possible that events combined to make the spirit of democracy change?

That's a crappy answer, isn't it. Hmmm...I guess I'd just have to keep thinking. But I feel that the answer would be non-obvious without a more thorough understanding of history than I have.

Abe said...

All -
Whoa a lot of good comments in a very short time. Let me recap the ideas you've put forward.

*trade unions
*prosperity
*universal suffrage
*machine politics


Did I miss anything?

Most of these relate to urbanization and industrialization. It seems like these ought to affect patterns of voting and polarization, but I have a hard time saying exactly how.

One observation: the polarization data is based on the behavior of Congressmen and the turnout data is based on the the public at large. So any story that explains these two facts together must account for a change in the relationship between political elites and the electorate at large.

Universal suffrage seems like a good bet for explaining that kind of trend. The simplest story is that female, minority, and unpropertied citizens were less likely to turn out to vote. Therefore, turnout fell as voting rights were extended to these groups.

This might be true, but it can't be the whole story. If the change in trends is just due to "watering down" of turnout due to expanded suffrage, we would expect the relationship between polarization and turnout to resume once the influx of new voters was over, say by the 1940's. I tested this with a simple regression and it doesn't hold up -- there is no relationship between the two trend lines after about 1900. So the simple suffrage story can't account for the pattern we see. either.

I'm still stumped on this one.
- Abe

Dave said...
This comment has been removed by the author.
Dave said...

Hey Abe could you send us a link to this data. Margaret and I want to throw a few regressions at it and see if anything sticks.