I'm sure there's lots in Frank's book to critique. But Bartels misses the point. He thinks he's helping his argument by pointing out that if you only focus on whites in the poorest third (the poorest quarter of whites) rather than those without a college degree, you get a two-point margin for Bush in 2004 rather than a 23-point margin. But that fact alone is shocking and cries out for explanation. If you just think of how many of those people must rely on food stamps or Medicaid or the EITC for basic survival - or are related to people who do - the question becomes, how could Republicans actually get as many votes as Democrats among that demographic group?
Forty years ago, the GOP was the party of both John Lindsay and Barry Goldwater, George Romney and Strom Thurmond, so it's not surprising that class voting was not heavily polarized. Whereas today the GOP is pretty much 100% Tom Delay -- i.e., politicians who rail against the shiftless poor and cut health benefits at every opportunity. It would be truly shocking if there were zero shift away from the GOP among the poorest quarter of whites during that timespan. In other words, observed shifts in white partisan voting do not represent shifts between fixed points on a continuum; they represent shifts between two points, one of which has moved enormously - and openly - against the interests of poor and working class.
Let me be concrete about what we should expect: In the 1960's the GOP made up its mind to be the party of anti-civil rights. That shift can be vividly observed in the voting data: In the mid-1950's almost half of blacks voted Republican; by the end of the 1970's they voted Democrat 90-10. Likewise, in the 1970's and 1980's the GOP decided to be the party of viciously persecuting the poor. But instead of leading to a massive shift toward the Democrats, Bartels' Figure 3 shows that the Dems' share of bottom-third whites barely changed - it went up by (eyeballing the chart) <5 percentage points.
Imagine if the voting of the white poor had undergone a shift even half as sharp as that of blacks - let's say they went from voting Dem 50% of the time to voting Dem 70% of the time. That would have given the Dems an additional structural 4% or 5% in every election. The Republicans would have lost most of the races they won or - more likely - would have had to moderate their politics. Instead, the GOP took its massive rightward lurch and got away with a poor-white shift against them that was so small as to make virtually no electoral difference. Why?
SA