February 24, 2009

red state blue state review-part two

I wish to continue my reviews of this excellent work. I hope my previous review has shown how very valuable and interesting this work is. Here I want to focus on a few issues where I think their conclusions need nuancing-and more than they give it as well as one objection raised I think is invalid.

It strikes me a lot of their comparisons/ the implications of their comparisons are diluted by the importance of the African American Vote in the United States. This is around 10% and more of the American Vote. Indeed in one survey I read African American were more likely to vote democratic in the South than Affluent, Gun owning self described members of the religious right were to Vote Republican! African Americans vote overwhelmingly Democratic with only fairly minor differences on the basis of whether they are born again or not, rich or poor go to church or not etc. They are also of course disproportionately poor and churchgoing. Thus the effect of their vote-will is to increase the class cleavage and reduce the religious cleavage in American life. It will also make poor voters look like they vote less on religious lines than rich voters- even white rich voters do. Gelman and co make some reference to this -but not enough. In other words the apparent (comparative) weakness of churchgoing and income in the US could be a function of race.

And indeed if you look at white voters the big differences on income are for those earning in the bottom 8% of Americans and the top 3%. This makes a lot of sense since the differences between the two parties on income policies (whether the Earned Income Tax credit or high rate income tax) are concentrated at that level. This is discussed by Byron Caplan in this blog post..

I'm quite certain that lies behind Mississippi having the most republican non-churchgoers of any state (Mississippi has the highest % of African Americans. Similarly as they also acknowledge a lot of the reasons why the income gap is higher in red states is that in the South African Americans are a very high % of the poorest voters (in the other republican presidential stronghold of the rocky mountains the gap is much less apparent)-again they do acknowledge this is part of the explanation.

On international compressions it strikes me there is insufficient attention to the role of "dead" cleavages-and this they don’t really address. That is if a factor (such as religion) or class had a large effect on a generation's parents voting it is likely to correlate with their own voting-because both class, religious affiliation . A fictional example (that is perhaps not that far from some countries actual electoral history) should help.

Say Country X has a rightwing and a leftwing party 80% of churchgoers vote for the right and 80% for the left (let's say for the sake of argument it's a new democracy) .Let's say in the Next generation churchgoing stops having any direct effect on voting patterns. However Say churchgoers/ non churchgoers) are 80% likely to have churchgoing /non-churchgoing children-and right-wingers are 80% likely to have rightwing children. A huge correlation will remain between churchgoing and voting for the right even if it has no independent effect (over 60%).

A classic case of this happening was in post war Canada. In the early and mid 20th Century the Conservatives (latter Progressive Conservatives) were above all the party that sought to build up Canadian life according to a pro British, anti American protestant and centralized model. The basis of their support was Protestant Canadians. Catholics voted overwhelmingly for the liberals. A survey in the 1980's (when the issue environment had changed so much it was unrecognisable) found whether one was Catholic or Protestant was still the best predictor of voting Conservative- a better predication than economic class or religiosity. However this was just because the Catholic Liberal voting’s of the early 20th century had had both disproportionally Liberal and Catholic Children. The best proof of this was that converts to Catholicism who were went to church regularly (i.e. those with whom one can see the effect of Catholicism operating independent of historical tendencies) were actually extremely likely to vote conservative.

It may be the weakness of the churchgoing cleavage in America relative to western Europe is very heavily this effect- the lack of pious/ secular religious cleavages in America in the 1950's and the strength of it in so many western European countries.

I also feel I should defend them from some criticism Gelman that they themselves seem weak in answering”How could it be when here religion is out of the political campaigns and discourse and there is no question whatsoever about the faith of candidates?” For a start the latter claim strikes me as dubious and exaggerated (note how every major French rightwing candidate in the last twenty years has been a practising catholic with no leftwing one being so). . In any cases in America talk of faith by candidates tends to be highly generic –at most generically Christian often frankly generically theistic (in a way that could included theists)-crucially it tends to bi-partisan there is not an obvious difference in the way different parties national candidates talk about Faith. (And its difference rather than level of piety that explains how much people vote on religious lines-I imagine people’s religious views had little impact on their voting in the fifteenth century England!)

. AT the same time the differences between left or right or religiously issues can often be sharper. So for example in Italy (whose overwhelmingly catholic nature means such issues are likely to be more a proxy for religiosity than in the united States the right when last in power restricted IVF and embryo research earlier this decade, the left government then sought to bring in "civil unions" (failing partly due to internal divisions) and the right has now sought more restrictive and tight “euthanasia laws” Even in the United Kingdom there weree huge party parliamentary differences for example on the abortion laws shown in very recent votes. it’s no wonder in such circumstances there’s a correlation between churchgoing and voting. It’s not just a legacy of the past.


Vino S said...

Your comments on the effect of the African-American vote in 'diluting' the effect of religiousity and voting is interesting. It is notable that 80%+ of black voters tend to vote for the Democrat (from 1964 onwards).

However, as I said in my comment on your last post, white Southerners are also another anomaly. They are poorer than white voters in the North but more likely to vote Republican. They voted Democrat until the 1960s and then swung dramatically to the Republicans once Johnson signed the Civil Rights Act!

I would be interested to see whether correlations between religiousity and vote and income and vote are stronger in the Northern states or California than they are in the Deep South.

James Higham said...

Different world down that way. Have you been down there?

Sulla said...

Vilno I think on African-Americans we agree.

White Southerners do indeed more Republican than income would suggest- but I think factors like Religion can explain this quite well. What’s certainly the case is that rich and poor white southern vary more than white "coastal" voters in their electoral behaviour (though the class difference is less than among southerners and "costals"ignoring race)

ON your last paragraph I'm pretty certain church going gaps are bigger in the North- however churchgoing is only a partial proxy for religiosity as I've pointed out!

Incidentally it's quite dubious that the Civil Rights Act was that important in switching white Southerners or even the Voting Rights Act. I think it had a very minor effect long term-see for example this excellent work http://books.google.co.uk/books?id=h1P_u6sxY4MC&dq=The+End+of+Southern++Exceptionalism&printsec=frontcover&source=bn&hl=en&ei=oy-zSavsE5DRjAevs4zpBQ&sa=X&oi=book_result&resnum=4&ct=result#PPR7,M1