James Taranto explains the Roe effect...

...or, How many future Al Frankens have to be butchered like mad cow infected animals before the Democrasses get it?

(Thanks to OpinionJournal.)

Compounding the GOP advantage is what I call the Roe effect. It is a statement of fact, not a moral judgment, to observe that every pregnancy aborted today results in one fewer eligible voter 18 years from now. More than 40 million legal abortions have occurred in the United States since 1973, and these are not randomly distributed across the population. Black women, for example, have a higher abortion ratio (percentage of pregnancies aborted) than Hispanic women, whose abortion ratio in turn is higher than that of non-Hispanic whites. Since blacks vote Democratic in far greater proportions than Hispanics, and whites are more Republican than Hispanics or blacks, ethnic disparities in abortion ratios would be sufficient to give the GOP a significant boost--surely enough to account for George W. Bush's razor-thin Florida victory in 2000.

The Roe effect, however, refers specifically to the nexus between the practice of abortion and the politics of abortion. It seems self-evident that pro-choice women are more likely to have abortions than pro-life ones, and common sense suggests that children tend to gravitate toward their parents' values. This would seem to ensure that Americans born after Roe v. Wade have a greater propensity to vote for the pro-life party--that is, Republican--than they otherwise would have.

The Roe effect would have made itself felt before post-Roe children even reached voting age. Children, after all, are counted in the population figures that determine states' representation in Congress and the Electoral College. Thus, if the greater prevalence of abortion post-Roe affected statewide fertility patterns, the results would have begun showing up after the 1980 reapportionment--in the 1982 election for Congress, and the 1984 election for president.

The first post-Roe babies reached voting age in 1991, in time for the 1992 election. In 1992 the Roe effect would have been minimal, since it was limited to a small segment of the electorate (18- and 19-year-olds), who tend not to vote. The affected segment of the population grows with each election, ranging up to 23-year-olds in 1996, 27-year-olds in 2000, and 31-year-olds in 2004. The Roe effect is compounded over generations. Children who are never born do not have children or grandchildren.

Critics of the Roe effect hypothesis point out that abortion does not necessarily diminish a woman's lifetime fertility. A woman may, for example, have an abortion while in college, but later marry and bear children--children she might not have had, had she been forced to carry her collegiate pregnancy to term. Yet it is not clear how much this might mitigate the Roe effect. Some women do abort their final pregnancy, and delayed childbearing is one manifestation of the Roe effect. If a woman has a child at, say, age 30 rather than 20, one additional census passes before the child counts toward his state's congressional and electoral college apportionment, and two or three presidential elections pass before he reaches voting age. The compounding element applies here as well; if a woman has a daughter at 30 rather than 20, the daughter reaches childbearing age a decade later than she otherwise would have. Moreover, attitudes about abortion and politics are subject to change with age and experience, and usually in a conservative direction. Thus, some women who delay childbearing contribute to the Roe effect on both ends: by having abortions when they are young, single, and pro-choice, and by bearing children when they are older, married, and pro-life.

Has the Roe effect borne itself out in practice? The results are mixed. In terms of reapportionment, the trend is decidedly in favor of Republican states. The 30 states George W. Bush carried in 2000 had 271 electoral votes, a bare majority. Reapportionment after the 2000 census increased that number to 278. In the 1980s, they were worth only 267 electoral votes, not enough for a majority; in the 1970s, 260. The trend continues: Of the 10 fastest-growing states in 2003-04, Bush carried nine in 2004. (One of them, New Mexico, went for Al Gore four years earlier.)

But Roe effect doubters can point to 2004 exit-poll results that found 18- to 29-year-old voters--i.e., those born after 1975, who correspond closely with the post-Roe generation--were the only age cohort that supported John Kerry over Mr. Bush, by 54% to 45%. Yet caution is in order in interpreting these results. The Roe effect does not predict that younger voters will be more apt to vote Republican than older ones, only than they otherwise would be. Putting the Roe effect to a real test will require a longitudinal look at these voters. How will their voting pattern change, as they grow older and more settled? In any given year, the youngest age cohort will include a high proportion of lower-income and never-married voters, both traits that are highly correlated with voting Democratic. Marriage, in particular, tends to correspond with conservative attitudes on abortion and other social issues, and therefore with voting Republican. According to 2004 exit polls, Mr. Bush outpolled Mr. Kerry among married voters, 57% to 42%, while Mr. Kerry beat Mr. Bush among singletons, 58% to 40%.