Tuesday, October 04, 2005

Larry Bartels on "What's the Matter With Kansas?"

For those looking for a good analysis of Tom Frank's book, What's the Matter With Kansas?, I strongly recommend Larry Bartels's paper from this year's APSA meeting. Bartels points out that there's little if any evidence to support Franks's claim that cultural issues have divided lower income Americans from the Democratic party. Here's the abstract of his paper:

• Has the white working class abandoned the Democratic Party? No. White voters in the bottom third of the income distribution have actually become more reliably Democratic in presidential elections over the past half-century, while middle- and upper-income white voters have trended Republican. Low-income whites have become less Democratic in their partisan
identifications, but at a slower rate than more affluent whites – and that trend is entirely confined to the South, where Democratic identification was artificially inflated by the one-party system of the Jim Crow era.
• Has the white working class become more conservative? No. The average views of low-income whites have remained virtually unchanged over the past 30 years. (A pro-choice shift on abortion in the 1970s and ‘80s has been partially reversed since the early 1990s.) Their positions relative to more affluent white voters – generally less liberal on social issues and less
conservative on economic issues – have also remained virtually unchanged.
• Do working class “moral values” trump economics? No. Social issues (including abortion) are less strongly related to party identification and presidential votes than economic issues are, and that is even more true for whites in the bottom third of the income distribution than for more affluent whites. Moreover, while social issue preferences have become more strongly related to presidential votes among middle- and high-income whites, there is no evidence of a corresponding trend among low-income whites.
• Are religious voters distracted from economic issues? No. The partisan attachments and presidential votes of frequent church-goers and people who say religion provides “a great deal” of guidance in their lives are much more strongly related to their views about economic issues than to their views about social issues. For church-goers as for non-church-goers, partisanship and voting behavior are primarily shaped by economic issues, not cultural issues.


I would add to this that not only is Frank wrong about American politics overall, he's particularly wrong about Kansas. Frank makes it sound as though Kansas has abandoned a long history of supporting Democrats. That's just not true, since it has almost always been one of the most Republican states in the country. The table below gives the Republican % of the presidential vote for both Kansas and the U.S. for each presidential election since 1864 (the first in which Kansans voted). In addition, I've included where Kansas ranked among all states for Republican vote percent. As you can see, in 34 elections out of 36, Kansas has been more Republican than the rest of the nation, and 20 elections it has been among the top 10 Republican states in the country. The only two times that Kansas ran behind the Republican ticket nationally was in 1896 and 1912. The first of these was at the height of the Populist movement, but even then, Kansas was only 3.4 percentage points less Republican than the rest of the country and the Democratic/Populist ticket barely won a majority of the vote. The second time was in 1912 when Teddy Roosevelt bolted from the Republican party. If you look at the combined vote for both Republican candidates (TR and Taft), it was 50.6 nationally, and 53.4 in Kansas.

YearKansasU.S.DifferenceRank
186479.25524.21
186868.852.716.13
187266.555.610.97
187663.14815.13
188060.448.312.15
188461.548.313.22
188855.247.87.46
189248.4435.412
189647.651-3.426
190052.651.6123
190464.856.48.412
190852.551.60.924
191220.523.2-2.733
191644.146.1-227
192064.860.34.514
192461.5547.57
19287258.213.81
193244.139.74.412
19364636.59.54
194056.944.812.13
194460.345.914.41
194853.645.18.54
195268.855.213.65
195665.457.485
196060.549.610.92
196445.0138.56.5112
196854.843.411.46
197267.760.7713
197652.5484.512
198057.950.87.110
198466.358.87.59
198855.853.42.424
199238.937.51.422
199654.340.713.62
20005847.910.111
20046250.711.38

6 comments:

Anonymous said...

Finally somebody else realizes and articulates the digusting inaccuracy and oozing condescension in this appalling book, which seems to encapsulate the view of many Hamilton College students towards the Middle America.

Anonymous said...

This analysis is useful, but it skims over one of the main thrusts of the book: these people are voting for the GOP even as they've abandoned being a party that helps them out. Just like it says, they vote against abortion but abortion's still legal. They don't get abortion illegalized, they get tax cuts for the rich. Etc.

Anonymous said...

What's the point of this table?

Frank doesn't claim that what's happening in Kansas is unique to that state. On the contrary, he sees it as an example of what's happening all over the country.

The test of Frank's thesis against aggregate election results (not -- I'd certainly agree -- a test with much power to judge his thesis against other explanations) is whether Republicans are doing better absolutely, not whether they're doing relatively better in Kansans. In case anyone hasn't noticed, Republicans have won some elections lately.

Philip Klinkner said...

Mr. Ross:

True, Frank is making a claim about the U.S.in general and the Bartels paper I mention does an excellent job of analyzing his national claims. On the other hand, the book is called ""What's the Matter With Kansas?" and in it he makes all sorts of particular claims about Kansas politics, namely that it has abandoned its previous support for progressive/liberal candidates. That's the purpose of the table.

Ambivalent_Maybe said...

Bartels's paper was interesting, but the argument btwn him and Frank seems to hinge on how to define 'working class' in America. For Frank, the key marker is a college education. Bartels briefly considers this, but rejects it w/o giving us any description of what the numbers look like. Bartels too easily equates 'working class' with the lower third of the income distribution.

Anonymous said...

The heart of Frank's argument is that when working-class voters support Republicans they are voting against their own class interests, defined in economic terms, and that this behavior needs to be explained. Frank's explanation for this phenomenon is that conservatives skillfully use a populist rhetoric that focuses on religiosity and traditional morality to appeal to working-class voters, ironically presenting themselves as "populists" despite the relentless faithfulness of Republican policies to corporate interests. Bartels' critique of Frank is thoughtful and makes some good points, such as that lower-income whites (those in the bottom third of the income distribution) tend to mostly favor Democrats and do so on largely economic grounds. But read carefully Bartels' paper does not completely refute Frank. For one thing, he is able to claim that the white working-class has not weakened its allegiance to the Democratic party by defining the working class very narrowly (bottom third), while even he admits that there is evidence that at higher income levels social issues appeals are successful at attracting voters to the Republicans. This procedure of defining the "working-class" as only those in the bottom third is dubious, since Frank would surely argue, correctly, that the majority of voters (not just the bottom third) are hurt by Republican policies that favor the wealthy. So the behavior of a wider category of voters needs to be explained than the bottom third of white voters that Bartels focuses on. Another big flaw in Bartels' argument is that he does have to admit that there has been significant erosion of white working-class support (even using his narrow definition of that class) for Democrats in one part of the country, the South. Though the cause of this erosion is generally considered to be largely racial dynamics (arising from many Southern whites being repulsed by the Democrats' identification with civil rights for blacks and racial equality) it is very likely that the social and religious issues that Frank discusses have played a role as well. The bottom line is that Bartels has a clever argument, but at the end of the day Frank's argument that social issues seem to function to convince many voters to ignore or overlook their own economic interests still sounds valid.