Public Opinion Watch: The White Working Class

Revisiting a controversy. Plus: the growing Democratic issue advantage.


Article created by The Century Foundation.

Several months ago, Larry Bartels, a political scientist at Princeton, wrote a very interesting paper, “What’s the Matter with What’s the Matter with Kansas?” critiquing Thomas Frank’s book and its view of white working class politics. David Gopoian and Ralph Whitehead, in turn, published a short critique of Bartels’ analysis, questioning his definition of the white working class and the political conclusions he drew from his analysis.

Since then, Frank himself has replied to Bartels at length, Bartels has replied to Gopoian and Whitehead, and Gopoian and Whitehead have replied to Bartels’ reply. Now, Matt Yglesias has helpfully summarized this controversy in an article in the latest American Prospect.

This is a very complicated and important debate and I don’t intend to rehearse its many aspects here. (For those who are gluttons for punishment, I am currently writing a paper with Alan Abramowitz that will go into the entire controversy in detail.) But I did want to comment a bit on some basic definitional issues that are raised by Yglesias’ article. Yglesias fairly points out some of the problems with the education-based definition of the white working class (those lacking a four year college degree) used by myself and others. He seems particularly perturbed that this definition makes the white working class pretty large and that many of its members have okay incomes, rather than being poor.

I don’t see either of these things as being real liabilities. Looked at politically, the Democrats’ problems with white voters are also pretty large and have never been concentrated among the poorest whites. So the education-based definition matches up pretty well. For that matter, so would an occupation-based definition. Given adequate data (most political surveys, including the exit polls, don’t have enough detail), if you looked at whites with non-professional, non-managerial occupations, that too would produce a pretty large group (similar in size to the non-college-educated group), many of whom would have okay incomes.

I don’t see why we should worry about this. After all, it was one of the great achievements of postwar U.S. capitalism that it became possible to have a middle-class life with a working-class job and/or credentials. Why define down the white working class to be the white poor? Especially when it produces anomalies that, in my view, far outweigh those you get with an education or occupation based definition. They include at least the following:

1. Bartels defines the white working class as those whites who fall in the bottom third of the income distribution. Note that this is in the bottom third of the overall income distribution, so we are not talking about the bottom third of white voters, but rather a substantially smaller group. According to Bartels’ own data, over the 1984–2004 time period, whites in the lower third of the income distribution amounted to about 23 percent of white voters.

2. As Gopoian and Whitehead point out, most of the white working class (almost two-thirds), under the Bartels’ definition, isn’t working (perhaps we should call them the “white non-working class”?). There is some dispute about the exact magnitude of the non-working figure, but, no matter whose figures you use, actual workers are a minority (in contrast to the middle and upper thirds of the income distribution for whites, who Bartels tosses out of the working class, where around three-quarters are actually working).

3. Most damaging of all, in my view, defining the white working class as, in essence, the white poor throws out of the white working class the very kind of workers who traditionally are most associated with that group. Using Bartels’ definition, for example—while one must make inferences from inadequate historical data—it appears highly unlikely that the typical autoworker, steelworker, construction worker, mechanic, etc. back in the late 1940s and 1950s could have qualified for Bartels’ white working class. They just weren’t poor enough.

And today? Not too different. Consider these data from the Economic Policy Institute—the average unionized blue collar job in the United States in 2003 paid $22.74 an hour (presumably the average wage of whites in these jobs was somewhat higher). That’s way too high to qualify for the Bartels white working class—and that’s leaving out any possible income from a spouse.

To me, this just doesn’t make any sense. So let’s think of the white poor as, well, the white poor and not confuse them with the white working class, a genuine and vexing political problem for the Democrats. I’ll close with these hair-raising data from the 2004 exit poll:

Among non-college-educated whites with $30,000–$50,000 in household income, Bush beat Kerry by twenty-four points (62 percent to 38 percent); among college-educated whites at the same income level, Kerry actually managed a 49 percent to 49 percent tie. And among non-college-educated whites with $50,000–$75,000 in household income, Bush beat Kerry by a shocking forty-one points (70 percent to 29 percent), while leading by only five points (52 percent to 47 percent) among college-educated whites at the same income level.

In other words, the more these non-poor voters looked like members of the white working class (by my definition), the less likely they were to vote for Kerry in the 2004 election. That’s a problem—a big problem—that no defining down of the white working class is going to take away.

Growing Democratic Issue Advantage

Last week, I noted how an intense anti-incumbent mood seems to be brewing, unlike anything we’ve seen since 1994. This week, I draw your attention to some very interesting data released by the Pew Research Center which suggest that the Democrats are gaining an issue advantage of unusual magnitude that might feed that anti-incumbent mood.

In the Pew survey, they first asked respondents what they thought the most important national problem was, followed by a query about which party could best handle that problem. The result was a 41 percent to 27 percent advantage for the Democrats on handling the most important national problem. That 14-point gap is the largest measured by either Pew or Gallup since this question was first asked in 1987. By way of comparison, the GOP had an 11-point advantage on this measure in March, 2002 and a 3-point advantage in July, 1994.

These data can be broken down by the type of national problem cited by respondents. Democrats had strong advantages in every area but one (security/terrorism): the economy (+21); social/domestic (+22); Iraq (+19); and foreign policy (+30). This compares to last January, when Republicans were actually favored overall and on social/domestic issues (by a point) and Democrats only led by 5 points on Iraq and 17 points on foreign policy. Even on security/terrorism, while the GOP still leads by 18 points today, that’s down from an overwhelming 39 point margin at the beginning of last year.

More evidence that we are headed for a very interesting election that could see some big changes.