When the National Center for Health Statistics released figures in October showing that illegitimate births had dropped by 3 percent in 1995, President Clinton was quick to claim credit. It was the first decline in out-of-wedlock births in a generation, and it showed, he said, that "the American family is getting stronger and we are making responsibility a way of life."
Unfortunately, there are far better explanations for most of the drop, all having to do with how a few states collected their numbers. In 1995, California, the most fecund state, changed its hopelessly outdated practice of considering a Hispanic child illegitimate if its parents have different surnames. Now, if the child's surname combines the parents'—in other words, if Pedro Escobar and Maria Lopez produce Jesus Escobar-Lopez—the state considers the child legitimate, whereas it didn't before. This change, by itself, lowered California's tally by some 20,000—half the nationwide decline.
Michigan presented another anomaly. It failed to file complete information on out-of-wedlock births for 1995, so the federal government just as-signed it the average for the other 49 states. But illegitimacy in Michigan ran well above the national average in 1994, and there is no reason to think this didn't continue in 1995.
The seemingly good news on illegitimacy loses still more of its luster when you put California and Michigan together with New York and Connecticut. These four states account for a quarter of all U.S. births, and yet none of them, for reasons of privacy, records the marital status of mothers on birth certificates. Instead, they infer legitimacy from such things as paternity affidavits and the appearance of a father's name on birth certificates.
No one knows for sure what would happen if these states went about this task more directly, as the 46 other states do. But when Nevada switched from inference to actual head counts in 1995, illegitimate births leapt by 20 percent. How would the president explain this jump?