The Obama Ceiling, Again

Shortly after the New Hampshire primary, I offered the following analysis:

Barack Obama got 36% of the vote in NH, which is almost exactly equal to the 35% percent of the vote he got in the entrance poll of Iowa caucus goers. Even with what was perhaps the most positive press coverage for any candidate in recent history, he ended up with zero bounce. Zilch. Nada. This suggests that Obama has a ceiling, at least among white voters, of about 30-40 percent.

And after the Nevada caucuses, I updated the analysis:

The exit polls from Nevada provide more evidence that Obama has been unable to break out of his ceiling of approximately one-third of the white vote. In fact, in each of the three Democratic contests thus far, Obama's support among whites has been remarkably consistent:

Iowa: 35%
New Hampshire: 36%
Nevada: 34%

If this 35% ceiling does, in fact, exist, it's interesting to compare it to Jesse Jackson's performance in 1988. Despite the passage of 20 years and the fact that Jackson and Obama are very different candidates and personalities, Obama hasn't performed significantly better than Jackson. During the 1988 primaries, especially once the race narrowed down to Dukakis and Jackson, Jackson's white support ranged between 20 and 35 percent

.

In a recent LA Times op-ed, Abigail and Stephan Thernstrom argue that this year's Democratic nomination contest has led to "the demolition of the long-held belief that whites simply won't vote for black candidates for higher office. Before the Iowa caucuses on Jan. 3, who could have predicted the remarkable outpouring of white support for Sen. Barack Obama?"

As evidence of this long-held belief, they point out that "after the first three contests, political scientist Philip Klinkner was ready to say that there was a 'ceiling' on white support for Obama of about 35%."

So how wrong was I? To check, I went into the exit polls and vote totals to estimate the percentage of white votes for Obama. Below are the results for every primary since John Edwards pulled out and narrowed the contest to Clinton and Obama. I've excluded caucus states since we don't have exit polls for them and because their smaller turnout makes them somewhat less representative of behavior of rank and file voters. I've also excluded results from New York, Illinois and Arkansas since they are home states or adopted home states of Clinton and Obama (though I would add that including them does not change the numbers much).

State	Obama White %	Total Votes	Obama White Votes
California	0.45	4,416,779	1,987,551
Arizona	0.38	443,324	168,463
Oklahoma	0.29	417,095	120,958
Missouri	0.39	823,376	321,117
Tennessee	0.26	622,586	161,872
Alabama	0.25	536,293	134,073
Georgia	0.43	1,060,767	456,130
Delaware	0.4	94,675	37,870
New Jersey	0.31	1,108,044	343,494
Connecticut	0.49	337,269	165,262
Massachusetts	0.4	1,244,133	497,653
Utah	0.55	124,307	68,369
Louisiana	0.3	384,243	115,273
Virginia	0.52	973,931	506,444
Maryland	0.42	760,314	319,332
Wisconsin	0.54	1,099,661	593,817
Ohio	0.38	2,218,174	842,906
Texas	0.44	2,856,813	1,256,998
Rhode Island	0.43	182,204	78,348
Vermont	0.6	141,849	85,109
Total		19,845,837	8,261,038
Percent			41.63%

As the table shows, Obama's white support has averaged just under 42 percent across each of these primaries. In Utah, Virginia, and Wisconsin, he managed to get a majority of the white vote.  On the other hand, in eight states (Arizona, Oklahoma, Missouri, Tennessee, Alabama, New Jersey, Louisiana, and Ohio), Obama's white support was under forty percent.

What does this all mean? First, Obama has shown in states like Virginia and Wisconsin that he break the ceiling and win the majority of white votes. On the other hand, the overall results, including recent states like Ohio, show the difficulties he faces gaining enough white votes, even among Democrats and in Northern states, to break out of the ceiling I suggested last moth. There might be many reasons for Obama's performance among white voters that have nothing to do with race. As I've suggested, Clinton has strong appeal among working class voters based on the economic performance of Bill Clinton's administration. In addition, white Catholics might be turned off by the Protestant revival style of the Obama campaign. Still, I don't think you can completely rule out race as a factor in Obama's performance among white voters. The exit polls show that race, more than age, income, education, or gender, has been the most important factor in the Democratic primaries. In particular, there's little evidence to suggest, pace the Thernstroms, that Obama's candidacy proves race no longer matters in elections and that the Voting Rights Act is no longer necessary.