Fast-forward to 2006. By then, the voters were all 23 years old. They'd witnessed the Twin Towers fall with the same 18-year-old perspective and saw America mired in the same wars in Iraq and Afghanistan. The only difference between the two groups was a higher chance of having registered after 9/11 for the September group.
Yet these essentially identical groups still held different political allegiances in 2006: The fraction of voters born between Sept. 1 and Sept. 15, 1983 who were registered as Republican was two percentage points higher than the fraction of voters born between Aug. 16 and Aug. 31 of that year. (On average, 28 percent of Californians are registered Republicans.) The effect proved to be remarkably persistent: Looking at 2008 registration records, the researchers found that September babies still registered as Republicans 2 percent more often than August ones. The difference in Republican registration rates between the two groups induced by the Sept. 11 attacks has proved to be remarkably persistent, remaining intact at least to 2008.
Might the effect that Kaplan and Mukand have observed merely be the result of inertia—of voters changing political allegiance but not bothering to update their party affiliation? That doesn't seem to be the case. Limiting their analysis only to registrants who changed their address (and hence were forced to re-register) since 2001, they find a gap of more than 6 percentage points between the August and September birthdays—three times the average for the sample overall. Given that the effect doesn't diminish among movers, it's very unlikely that the September birthday effect is purely the result of inertia. (The fact that it's three times larger suggests that movers are somehow different from voters overall. It turns out that they tend to be from zip codes that are richer, whiter, and more Republican than average—movers thus might have their Republican identity reinforced because they're more likely to be surrounded by other Republicans.)
In explaining their results, the authors argue that they are unlikely to be the result of exposure to one-sided propaganda. The August-September Republican registration gap widens for voters living nearby four-year college campuses, which the researchers take as a rough proxy for informed citizenry who are likely to keep up on current affairs and be less swayed by party literature. The authors favor instead the explanation that party registration imparts a self-perpetuating political identity that lives on well past voters' 18th birthdays.
The implications for political operatives from both parties is clear—rocking the vote among college freshman can pay dividends many years down the line, and campaign dollars should be allocated accordingly.
But if political divisions are a matter of social identity, is there any hope of healing the red team–blue team divide in America? The pioneering work of social psychologist Muzafer Sherif may provide some insight. Sherif ran a series of studies with 12-year-old boys in a summer camp-like setting in Robbers Cave State Park in Oklahoma. He divided the "campers" into two teams—Rattlers and Eagles—and despite no prior relationships among the boys, Rattlers befriended other Rattlers and eyed Eagles with suspicion and hostility. It was Red and Blue teams all over again. Yet when Sherif set up challenges that forced all the boys to work together—like pulling a broken-down truck back to camp or solving a water shortage—the animosity between Rattler and Eagle disappeared, never to return. One might have thought that fighting wars abroad and dealing with financial calamity at home would be challenge enough to set aside partisan divides. Then again, perhaps pre-adolescent boys have more open minds than today's electorate.