Last week, the National Center for Health Statistics issued the results of a survey revealing that the share of American births that resulted from unwanted pregnancies increased from 9 percent in 1995 to 14 percent in 2002. (Seattle Times reports here.)

That’s bad news. It’s also puzzling.

It’s bad news because babies conceived by accident, when mothers do not want to have a child (or another child), tend to have what social scientists call “adverse outcomes,” as discussed here. They’re more likely to have bad prenatal care, die in infancy, fare poorly in school, and suffer violence at the hands of their caregivers.

It’s puzzling because so many reproductive trends have improved since 1995. Pregnancy rates overall have fallen, as shown in this chart for Oregon . . .


  • . . . especially among teens, as shown in these charts for Oregon . . . 

    . . . and Washington.

    Birth rates overall have fallen. Access to emergency contraceptives has expanded dramatically, as has insurance coverage for prescription contraceptives.

    All of these positive trends would make you expect that the share of births that result from unwanted pregnancies has also declined. But the opposite has happened, at least in the United States overall.

    The trend is less contradictory within Cascadia. Washington has the best data on unwanted births over time, and they show hardly any change—or hardly any change that’s greater than the margin of error. At best, there’s been a tiny decrease in unwanted births.

    Still, you’d expect that drops in pregnancies and births would lead to equally dramatic declines in unwanted pregnancies. You’d think, in fact, that improving pregnancy prevention would show itself first and foremost in a declining share of pregnancies that are unwanted. Instead, everything is shrinking dramatically except the “unwanted” percentage!

    What’s going on here?

    I don’t know.

    The Alan Guttmacher Institute (AGI) in New York is reportedly in the midst of an analysis of this puzzler, and I hope they’ll figure it out. In the meantime, let me underline my ignorance by explaining why the obvious answers are probably wrong or, at least, inadequate.

    Pro-choice advocates argue that the survey results are a sure sign of deteriorating access to abortion services, which is plausible. In Idaho, Oregon, and Washington, the number of abortion providers is lower now than it was two decades ago. In 1981, there were more than 160 abortion providers in these states; by 2001, there were fewer than 100, according to AGI.

    And the next chart, comparing the share of pregnancies ending in abortion (excluding miscarriages) in British Columbia and Washington, lends further plausibility to the theory. In British Columbia, where abortion providers have not decreased in number to the same degree, abortion has grown as a share of all pregnancies. In Washington, it’s shrunk.

    Harassment and intimidation from extremists explains some of the drop in abortion providers, but economic consolidation has also contributed. Abortion services have become a specialized medical subdiscipline, concentrated in the hands of fewer providers who are, in general, very good at what they do. First-trimester surgical abortion, therefore, may now be safer and less expensive, in inflation-adjusted terms, than ever before. Convenient, nearby access to safe abortion services does not extend to small-town residents in the inland parts of Cascadia, but most women who want an abortion can get one, by traveling to a city—the same place they have to go for many other surgical procedures.

    Anti-abortion advocates have a different exlanation for the rise in unwanted births. They suggest that Americans are demonstrating, in the words of an official at the U.S. Conference of Catholic Bishops, a “pro-life shift.” American women may be exercising their freedom to choose by electing to have fewer abortions. This explanation also has some plausibility. Abortion rates are lower in more-conservative states such as Idaho than in more-liberal states such as Oregon and Washington. Maybe a cultural change is making the whole country more like Idaho. (And maybe the Washington-BC divergence in the chart above is explained not by changing access to abortion but by changing social values.)

    Maybe, but I’m dubious. For one thing, British Columbia has the same anti-abortion movements, the same media influences, and the same medical technology (such as ultrasounds that make fetuses seem like babies sooner) as Washington. But abortion trends have diverged.

    For another, if the United States were experiencing such a shift, the recent survey should have found not only that “unwanted” pregnancies but also that “mistimed” pregnancies were being carried to term more often. In fact, if a “pro-life shift” were the cause, one would expect a more dramatic increase in the share of births that resulted from merely mistimed (or “too soon”) pregnancies, rather than from pregnancies that were truly never wanted at all. Women who found themselves pregnant a few years before they intended to be (and were swayed by anti-abortion arguments) would almost certainly be the first to forgo abortions, not the smaller number of women who found themselves pregnant despite their wish never to have a child (again). Wouldn’t the easy cases by “shifted” before the hard ones?

    The Center found no such shift.

    Between the Center’s 1995 survey (careful, enormous pdf) and its 2002 survey (careful, even larger pdf), in subset after subset of American women (young, old, married, cohabiting, Hispanic, white, first-time mothers, third-time mothers, etc.), there’s a marked increased in the share of births that women report as having resulted from unwanted pregnancies. There’s no comparable change for births that result from the far-more-numerous mistimed pregnancies.

    (Oh—and just to muddy the waters futher—the same reasoning also counters the suggestion that a paucity of abortion providers explains the rise in unwanted births. If barriers to getting abortions were the problem, it would presumably afflict the more-ambivalent mothers of "mistimed" births even more than it afflicted women with unwanted pregnancies.)

    This unwanted-mistimed patter is so unusual that I wonder, did some wording change in the survey skew the response? (The survey report claims the wording was identical.) Did 9/11/2001 shift women’s attitude retroactively, making them less sanguine about childbearing, and less positive about their births? (Seems unlikely, given that the survey was taken many months later and that it covered a five year period.) Did the Center’s statisticians just make a mistake?

    And even if there is some statistical fluke explaining this national survey, why isn’t the Washington state “unwanted” number (and it’s similarly static “mistimed” number) dropping with the pregnancy rate?

    My own partly formed suspicion centers on the fewer than 10 percen
    t of sexually active wome
    n (and their partners) who do not regularly use contraception and therefore account for about half of all mistimed and unwanted pregnancies. If everyone but them is getting better at prevention, their pregnancies may loom larger in national and state statistics. (There are flaws with this theory, too, but it’s the best I’ve got at the moment.)

    Anyone else care to theorize?