Abortion, Unwed Births, and Crime

Back in 2001, law professor John Donahue and economist Steven Levitt made headlines with a study in theQuarterly Journal of Economics suggesting that legalized abortion may account for almost half the drop in the U.S. crime rates during the 1990s. Now a study by John R. Lott Jr. of the American Enterprise Institute and his colleague John Whitley showcase substantial evidence that shows just the opposite, particularly that the legalization of abortion has increased murders in the U.S. by more than 7 percent.

Lott and Whitley reach this finding by improving upon the methodology of Donahue and Levitt. First they document how widespread abortion was in 45 states that restricted abortion prior to 1973, allowing them to include data since 1970 in their analysis where Donahue and Leavitt had presumed no abortions had taken place. The researchers also use the Supplement Homicide Reports instead of the arrest reports in the Uniform Crime Reports, allowing them to more accurately disaggregate the number of murders committed by different age groups in each state.

The researchers then arranged their data to examine the number of murders committed each year, allowing them to track each cohort over time and account for the number of abortions in their respective states in what would have been their year of birth. If Donahue and Leavitt’s theory is right—that abortion eliminates those in the population most likely to commit murder—then the panel data would reveal significantly lower murder rates among age cohorts that were born immediately after abortion was legalized. Yet the regressions conducted by Lott and Whitley reveal that when abortion rates increase, so does the murder rate when the remaining offspring born during that time come of age. In multiple specification tests, which included more than 5,000 regressions, their coefficients are remarkably robust, revealing what they call “a strong consistent positive relationship between abortions and murder.”

Their analysis found that the average state had 25,443 abortions in 1980 and an average rate of 24.5 abortions per 1,000 females, ages 15 to 44. Every additional abortion per 1,000 women was associated with about a 0.9 percent increase in murders in any given year. Because Roe v. Wade resulted in a dramatic jump in abortion rates (which they track over time), they estimate that the legalization of abortion is associated with a 7.2 percent increase in murders, or about 700 more homicides per year in the country by 1998.

Seeking an explanation for their findings, the researchers use state data to confirm that legalized abortion is also associated with increased rates of out-of-wedlock childbearing, a theory set forth more than a decade ago by economist George Akerlof. In their tests, Lott and Whitley found that each 1,000 more abortions is associated with a 0.6 percent increase in non-marital births, or nearly 10,000 additional children born to single mothers every year since 1980, when 1.6 million babies were aborted. The two believe this helps explain their documented increase in murder, as violent crime is disproportionately committed by young males who were born out of wedlock.

Lott and Whitley concede that the legalization of abortion, like many laws, has yielded both winners and losers. “But here,” they lament, “the net effect appears to be a reduction in human capital and an increase in crime.”

(John R. Lott Jr. and John Whitley, “Abortion and Crime: Unwanted Children and Out-of-Wedlock Births,” Economic Inquiry45 [April 2007]: 304–24.)