Voting and Income in the U.S. and Europe
I wrote over here that “[i]n the US and in Europe, income level is fairly predictive of voting behavior.” That’s broadly true, but as Andrew Gelman pointed out to me in an email, it obscures some major differences. His book Red State, Blue State, Rich State, Poor State includes this graph summarizing international differences in the extent to which high income leads to conservative voting:
I think it’s probably better not to think too much about the Eastern Europe and Latin America cases here and limit our attention to the rich countries. As you can see, it’s generally true that in Western Europe, as in the United States, being rich seems to make you favorable to parties of the right. But there are some exceptions to this rule. And in general, the rich-poor gap is not as sharp as in the US. This is intertwined with the fact that the economic gap between rich and poor is not as large and the fact that the ideological gap between the parties is not as big. In some of these countries, though, I would be interested to know how the parties are being coded. Based on my understanding of the Swiss party system, for example, I’m not sure which party I would label as the “conservative” one—they basically seem to have one left-wing party and then three different flavors of conservative parties.
29 Responses to “Voting and Income in the U.S. and Europe”
Register
RSSLogin
Topic Cloud
ArchivesBlog RollAbout Matt YglesiasContact Me
Donate
March 15th, 2009 at 1:19 pm
Again, in the U.S. it’s crucial to break analyses like this down to a regional level, since often electoral behavior is significantly different in the U.S. South.
March 15th, 2009 at 1:21 pm
Now if we could just find out if there’s a correlation between the Gini coefficient and levels of belief in creationism. Iirc, the US ranks just ahead of Turkey on the latter.
.
March 15th, 2009 at 1:28 pm
I think that there are serious limitations to a graph like this. Most importantly, because, in multiparty systems, it’s not always easy to decide which party is the ‘right-wing’ or ‘left-wing’ one. How would you fit the political parties in, say, Peru, or Russia, or Israel into a convenient left-right framework? In many countries politics has a number of different axes (liberal vs. authoritarian, nationalist vs. cosmopolitan, capitalist vs. socialist, religious vs. secular) which may not all line up the same way.
March 15th, 2009 at 1:29 pm
Complicating matters even futher in the U.S., ethnic and racial minorities tend to be poorer than whites, and I imagine the ethnic/racial divide between the parties outweighs the income divide.
March 15th, 2009 at 1:39 pm
Obviously we need some explanation as to who is classified as conservative.
But even if we assume they performed some rigorous ideological test that avoids the issue of multiparty states this graph is still pretty useless. Basically you are just comparing outliers and putting the data into a scatter plot doesn’t add any value.
March 15th, 2009 at 1:48 pm
Why are there so few liberals in Switzerland? What’s the matter with Switzerland? Don’t the Swiss know that liberals care more about the little people?
March 15th, 2009 at 1:51 pm
You might also want to figure out whether this represents capture.
March 15th, 2009 at 2:02 pm
I thought it was Iceland that believed in all the elves.
March 15th, 2009 at 2:14 pm
and, of course, no graph such as this can accurately represent sentiment in Israel. After all, progressive economic policies in an apartheid state mean something different altogether than those in even a moderately inclusive, non-racialized democracy.
What, for example, do socialized medicine, support for organized labor, or, resource sharing generally, represent in a white-man’s democracy? The answer, simultaneously, is something technically progressive and something else far more right-wing and sinister than ‘conservative.’
March 15th, 2009 at 2:16 pm
I agree with comment number 4, that much of this disparity between rich and poor voters in the US is due to racial and ethnic differences. In particular, it is well known that African Americans vote overwhelmingly Democratic and that they are significantly poorer on average than the rest of the population. I would guess that without African American voters, the US would be in line with European countries in this regard.
March 15th, 2009 at 2:17 pm
Good catch, Cap and Gown, that may well explain the entire difference. Anyone have any idea how poor immigrant communities in Europe tend to vote? My guess is that they are quite socially conservative, but also dependant on state largess and are put off by nativist sentiments among the conservative parties.
March 15th, 2009 at 3:12 pm
Okay that chart is absolute bull puckey. It shows no trends, the vertical axis is meaningless because of countries with more than 2 parties, and on the horizontal axis it gets the facts wrong. Russia has a GDP per capita of $14,800 yet the chart makes it look like they have $1,480. I am sure most of the other countries are wrong too (like Ukraine which is $7,000 but the chart shows having a negative GDP)
March 15th, 2009 at 3:19 pm
@13: I disagree.
The graph illustrates the relative degree to which wealth (and the status generated by wealth) tends towards conservatism in various countries. Nothing more and nothing less.
What is interesting are countries like France and Belgium where conservatism skews towards poverty: suggesting that religion and other social factors play more of a role in political outlook. Which in fact is true. In this light there is nothing surprising about Switzerland.
March 15th, 2009 at 3:20 pm
This graph represents everything I hate about political “science.” It does not appear to correct or account for the relative political power of “conservative” parties in the various countries, or for differences in income distribution that would affect the portion of the population classified as “rich” or “poor.” In fact, it does not even indicate what it means by “rich” or “poor.” And as others have mentioned, it does not (and cannot) account for the vast differences in the definition of “conservative” and “liberal” in the political paradigms of the various countries.
March 15th, 2009 at 3:56 pm
We have a two party system which has become extremely polarized. More parties might not only make us look different on this chart, but it might actually change the disparity in wealth.
March 15th, 2009 at 4:33 pm
Re: What is interesting are countries like France and Belgium where conservatism skews towards poverty:
In France, as i understand it, the poor gravitate towards the extreme left (communists and Trotskyists) and to the extreme right (Front National). Currently the FN is bigger than the communists, so overall it looks like a skew towards conservatism. But it isn’t, really- its a skew towards the extremes. I don’t think that the French poor are particularly more supportive of moderate conservatives than they are of the moderate-left Socialists.
For a long time, Catholics in France tended (by a heavy margin) not to vote for the traditionally anticlerical Socialists, but this was no longer the case by, I think, about 1980.
March 15th, 2009 at 5:15 pm
The chart tells you nothing, there are equal # of countries above and below and all around that line. Also there is no quantification of what “rich” is in a country, and how “rich” can vary from country to country. Is it a strict % of people? Is it a $ value? Is it a certain multiple of media income, what is “rich”? The same also goes for what “poor” is.
And the values on the only axis which can be checked (GDP / capita) is WRONG.
It looks to me like the author of the book wanted to prove a theory he had so he slapped together a crappy graph. When the graph didn’t clearly show what he was expecting he decided to just say it did in the hopes no one would bother trying to understand his crappy graph.
March 15th, 2009 at 5:49 pm
I gotta agree with JimboSlice. These graphs tell you nothing. They are a conclusion looking for data.
March 15th, 2009 at 7:06 pm
The chart tells you nothing, there are equal # of countries above and below and all around that line.
I agree the graph is pathetic and I assume it was probably misused to justify some preconceived argument.
But it does weakly show we can’t reject the null hypothesis that the ratio of rich to poor conservatives is uniformly distributed by GDP per capita. I am not sure the value of this or whether it was the authors intent but it is something.
March 15th, 2009 at 7:28 pm
Gordon, that is correct, and I do not know what the author intended the graph to show. All I know is that young Matthew has taken the graph to mean:
As you can see, it’s generally true that in Western Europe, as in the United States, being rich seems to make you favorable to parties of the right
Which is not shown at all.
March 15th, 2009 at 8:02 pm
Hi all. Thanks for the comments. For some discussion and (I hope) clarification, see here.
March 16th, 2009 at 12:41 am
Politics in Taiwan depend very little on economic policies; they’re almost entirely based on the reunification/independence issue.
March 16th, 2009 at 1:30 am
An additional confounding variable has to do with enfranchisement. A huge difference between the U.S. and many Western European countries is that naturalization and citizenship is a much more difficult process in Europe. This means that the racial/ethnic divides that you may see more prominently in the U.S. (and also present on the economic axis) would disappear in many European democracies since these same people would not be part of the vote…
March 16th, 2009 at 1:56 am
In particular, it looks like we miscoded some of the GDP per capita numbers, which doesn’t affect our conclusions but is a bit embarrassing.
It might not affect your conclusions, but it does affect the credibility I give to those conclusions. My skeptic meter goes humming anytime I see a graph with obviously wrong numbers. If you miss obvious errors like those it shows the figure was not checked, and leaves people wondering what else could be wrong. Are any of the vertical axis variables wrong? What about the countries, did you switch any of those around?
I based my comments about the graph being crappy because the obvious errors make it look slapped together, with no real care taken.
March 16th, 2009 at 10:27 am
Jimbo: Fair enough. As you’ll see in our latest post, it turns out that it’s not so clear what the right numbers are. In particular, if we had given Russia a GDP per capita of $14,000, others could’ve pointed out that, no, we were wrong, it’s actually $7,000. Which means that at the very least we should’ve had some explanation in our Notes section of what numbers we used. This is the sort of discussion we would always have with election or poll data where there is ambiguity. We didn’t do this with the GDP numbers because we didn’t realize the extent of the problem, which indeed represented a gap in our expertise in this area. Hence your skepticism was surely warranted. For our general conclusions, it’s less of a concern than it might have been, because (a) it only affected the x-axis, which we were only using for display purposes anyway, and (b) our substantive conclusions are taken from Huber and Stanig’s more thorough analysis (a point we did discuss in the book). But, yes, had I seen a graph with numbers that upon first glance seemed completely wrong to me, I’d be dismissive too.
March 17th, 2009 at 10:13 am
Did you see this in all age groups? If not, is it possible that this just means that old people are more conservative and richer?
March 19th, 2009 at 7:15 pm
I’m not big on commenting, but nice post.
April 8th, 2009 at 9:51 pm
After reading through the article, I feel that I really need more information on the topic. Can you share some more resources ?
April 15th, 2009 at 6:30 am
The style of writing is quite familiar . Did you write guest posts for other bloggers?