Wednesday, March 14, 2007

The blind spot of "Freakonomics"

The other day, former senator Zell Miller offered a practical argument against legal abortion, which the Atlanta Journal-Constitution reported:

Miller . . . declared that abortion has contributed to the military's manpower shortage, the Social Security crisis, and the flow of illegal immigrants into the United States.

"How could this great land of plenty produce too few people in the last 30 years? Here is the brutal truth that no one dares to mention: We're too few because too many of our babies have been killed," Miller said.

"Over 45 million since Roe v. Wade in 1973. If those 45 million children had lived, today they would be defending our country, they would be filling our jobs, they would be paying into Social Security," the former Georgia governor said.

Nows when it comes to illegal aliens, that actually makes a lot of sense. Illegal aliens usually fill entry level jobs, jobs that used to be taken by teenagers and younger people, who are now fewer.

This is related to The Wall Street Journal's Roe effect hypothesis (abortion makes the population more Republican because Democrats have a greater propensity to abort their babies) and to the claim by economist Steven Levitt, author of "Freakonomics," that abortion reduces crime (because mothers in crime-prone demographics are more apt to abort).

Levitt, however, says on his blog that Miller is wrong:

Miller . . . makes a key mistake in his logic. While it is true there have been many millions of abortions (although according to the official statistics more like 35 million than 45 million), even if those abortions had not occurred, there would not be that many more Americans today. The reason is that the primary impact of an abortion is not to reduce a woman's lifetime
number of children born, but rather, to simply shift the timing of a woman's fertility from early in life to later in life.

Based on a paper by John Donohue, Jeff Grogger, and I which will be out in a few weeks, I would estimate that each teenage abortion reduces lifetime babies born to the mother by maybe one-tenth of a child, or possibly even less. (For a woman who gets an abortion in her forties, the impact is obviously larger, but there are very few of those type of abortions.) The key to our abortion argument is that women shift their births to a time when they can better care for the children. So even though there is not a big change in the size of the cohort born, the kids still turn out less criminal. Miller's statement, however, is all about the cohort size, not about the unwantedness.

But Leavitt himself misses the point: In demographics, timing is also important, not just number of births per woman per lifetime. This point was addressed in the sociological journal Society:

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.

Either son or daughter also enters the labor force that much later.

Let's apply this to an area Miller discusses: Social Security. Suppose Woman X has a daughter at age 20, and her daughter has a daughter at age 20. When Woman X turns 65, her 45-year-old daughter and 25-year-old granddaughter will both be of working age, probably contributing to Social Security.

Now suppose Woman Y waits till age 30 to bear a daughter, having aborted at 20, and her daughter does the same thing. When Woman Y turns 65, her daughter will be 35 and her granddaughter will be 5. The working-age population among Woman Y's descendants is 50% lower than among Woman X's.

Delayed childbearing slows population growth, which means that the population at any given time will be lower, even if every woman eventually has exactly the same number of children. It's surprising that Levitt would miss this point.

No comments: