The 2002 elections might seem like a long time ago. But in vital ways they still provide the best guide to the rapidly approaching 2004 elections.
Unfortunately, no national demographic results were available in 2002, because the Voter News Service exit poll's computer systems crashed on Election Day. But pollsters eventually mailed in 17,872 completed interviews. The Roper Center for Public Opinion Research inspected them, pronounced the data as reliable as ever, and recently put the raw results on sale. I bought them and laboriously crunched through it for a multi-part series for United Press International.
This is a debate of profound significance for the future of the Republican Party—and of the Republic itself.
The opening round came in 1997, when Peter Brimelow and Edwin S. Rubenstein published a National Review cover story entitled "Electing a New People." Brimelow and Rubenstein demonstrated statistically that the long-term impact on Republican candidates of continuing to import millions of low-skilled people would be dire. (Click here for their 2000 update.)
Needless to say, there was little overt response. Barone simply ignored the argument in succeeding editions of his allegedly definitive Almanac of American Politics.
But, by early in this decade, the new mainstream of Republican consultants and publicists aligned with George W. Bush had accepted a caricature of the Brimelow-Rubenstein thesis: that immigration has already changed the electorate so radically that it would be political suicide to try to limit it. Instead, like a man with a hangover having more of "the hair of the dog that bit him," they argued that the GOP's only hope was to speed up the transformation of the electorate by passing an amnesty for illegal aliens in the hope that this would curry favor.
In late November of 2000, while the Bush vs. Gore recount in Florida was still going on, I wrote an article for VDARE.com—"GOP Future Depends on Winning Larger Share of the White Vote"—that added a new level of refinement to the demographic discussion by distinguishing between the short and long terms. I said:
"Here at VDARE.COM, we've discussed repeatedly how dire will be the long-term impact of immigration on the Republican Party. It's crucial to understand, however, that the long-term has not quite arrived. The GOP is not yet held hostage. It still has a window of opportunity—definitely stretching through the next recession but maybe not to the recession after that—to save itself by changing the immigration laws. This can be seen by examining the 2000 election results closely. The reason George W. Bush struggled so much to eke out a 271-267 win in the Electoral College (assuming that he can hold on to it) is not that he got crushed in the minority vote 77% to 21%. No, it's that he commanded only a measly 54% of the white vote…
Indeed, the basic logic of my position—increasing your party's share of the majority brings in more votes than increasing your share of the minority—is so obvious that I couldn't believe that Rove actually believed his spin to the ex-English majors and other innumerates who constitute the vast majority of reporters.
As I pointed out last year, when the crunch time arrived in the next election, Rove dumped minority outreach and went "hunting where the ducks are." He launched a massive get-out-the-vote drive among the Republican base. (VDARE.COM house style, to which I must modestly bow, is to call this appeal to the white majority "The Sailer Strategy.")
Some sharper liberals noticed Rove's ploy. Democratic pollster Ruy Teixeira came to the same conclusion. He told me, "The demographic theme of the 2002 election for the Republicans was 'Round up the usual suspects,' and they did a good job at it." The results for the Republicans were excellent.
But the lack of national exit poll data has allowed many commentators to go on making up fairy tales about the GOP winning by broadening its tent.
In analyzing the Roper data, I'll concentrate on the races in 2002 for the House of Representatives, since those are more comparable from year to year than the Senate or Governor races.
As I wrote for UPI, the actual Rove strategy (as opposed to the one that he talked about so much) brought these changes:
"[The GOP's] share of the non-white vote dropped from 25 percent in 2000 to 23 percent. That mattered little, however, because its share of the white-vote segment grew from 55 percent to 59 percent. Further benefiting the Republicans, the white portion of the electorate increased from 81 percent to 82 percent [because of improved turnout], even though the total population is becoming less white each year. The result was that the GOP became more dependent upon white voters, with whites casting 92 percent of all votes for Republicans, up from 90 percent in 2000."
Here are some further useful details from this trove of Election 2002 numbers:
Conventional Republican commentators like Barone assumed that Asians, being prosperous and law-abiding—the "new Jews," as he thinks of them—would automatically assimilate into the GOP. Unfortunately for Republicans, they are now voting like the "old Jews"!
But the Hispanic vote always fluctuates in parallel with the white vote, just many points further to the left. The white-Hispanic spread was 20 points in 2000 (55 v. 35) and actually rose to 21 points (59 v. 38) in 2002. In the last dozen House races going back through 1980, this white-Latino "ethnic gap" has held relatively steady at 19 to 28 points.
Moreover, the GOP's share of Hispanic ballots in 2002's Senate and gubernatorial races was significantly worse than in the House: only 33 percent in each.
In 2000, white women gave only 50 percent of their votes to the Republican House candidates, but in 2002 that figure reached 57 percent. Still, only eight percent of black women voted Republican (a typical result) and 37 percent of Hispanic women (there's never been much of a gender gap among Hispanics). Result: overall, the gender gap narrowed—the GOP won 50 percent of the total female vote for the House for the first time in decades.
Click here for all the details on voting by sex—and, more importantly, marital status. Married women vote significantly more Republican. In 2002, 56 percent of married women voted for the GOP (similar to their husbands' 58 percent) compared to 39 percent of unmarried women (and 44 percent of unmarried men). There's an especially large partisan difference between married women with children (58 percent Republican) and unmarried women with children (32 percent).
What about the much-touted Muslim vote? The best evidence against my assumption that Rove is a sharp cookie is the ludicrous and possibly catastrophic effort he cooked up with immigration-booster Grover Norquist to win the supposedly crucial Muslim masses in the 2000 Presidential election. Incredibly, as part of Rove's outreach, President Bush was supposed to meet with Muslim and Arab spokesmen to announce progress in eliminating airport security profiling of Muslim and Arab passengers on…the afternoon of Sept. 11, 2001. Something or other came up.
You would think that if Rove were going to expose Americans to a greater risk of terrorism, he'd at least sell out to a sizable voting bloc.
Yet it turns out that the fraction of voters who declared themselves to be Muslims in 2002 was miniscule: no more than 0.3 percent. The sample size was much too small to be reliable—but, for whatever it's worth, the interviewees voted 90 percent Democratic.
Hmmm. Maybe Karl's not such a "Boy Genius" after all.
In fact, not for the first time, the whole Stupid Party Establishment look, well, stupid.