Poverty and Abortion: A New Analysis

There has, for some time now, been an argument floating around out there (mostly among progressive-minded pro-lifers) that perhaps electing pro-choice Democrats is actually more useful in reducing abortion than electing pro-life Republicans, because Democrats reduce poverty and poverty is is tied to abortion. Among Catholic blogs, Vox Nova writer Mornings Minion wrote a statistical analysis of sorts six months back which he continues to point to as proving strong connection between poverty and abortion. In the wider world, Fuller Theological Seminary ethicist Glen Harold Stassen authored a pair of articles back in 2004 arguing (based on what later proved to be drastically incomplete data) that the number of abortions had risen under George W. Bush, and attributing this to Bush's economic policies. Stassen's analysis was then quoted (with rapidly decreasing degrees of accuracy) by Hillary Clinton, John Kerry and finally Howard Dean, who asserted that abortions have gone up 25% under Bush. (see above link to FactCheck.org)

Once it became clear that the data did not even remotely support the claim that abortion had increased under Bush, Stassen and others fell back on asserting that while it was true that the overall drop in abortion rates had continued under Bush, that the rate of its decrease had slowed, and that this was the result of an increase in poverty under Bush.

Now from a moral point of view, I think one must conclude that the point is irrelevant. Individual human beings are moral agents and as such, although they may find themselves under huge temptation to sin based on external pressures, they are free to choose right or wrong action. Thus, while it may be that in a period of relative prosperity people feel less pressure to commit certain crimes and/or sins, we must not see the duty of society to be simply to make sure everyone is too wealthy to want to sin; rather society must retain a strong enough moral sense to encourage right behavior in good times and in bad.

However, after hearing this argument one too many times, I decided to go dig into the data and see if even the strictly factual side of it is true. So far as I can tell, the argument has the following components:

1) Economic well-being (as measured by a low poverty rate and a high median income) has been greater under Democratic administrations than Republican ones in the last 30 years.

2) Abortion rates have either been lower or have decreased faster under Clinton than under Reagan, HW Bush or W Bush.

3) There is a strong correlation between economic well-being and the abortion rate.

I'm going to argue that all of these are partly or wholly false.

First, let's look at how the percentage of individuals below the poverty rate has fluctuated since 1976:

I've marked out the presidential administration on the chart, and I also marked the 1996 welfare reform act, since that appeared to align with an inflection point. It may also be worth noting that according to the National Bureau of Economic Research, there were recession troughs on July 1980, November 1982, March 1991 and November 2001. (Data drawn from US Census data here and here.)

Now let's look at the percentage of individuals below the poverty line, the percentage of families below the poverty line, and the median household income for the last four administrations:

As you can see, what we have is a steadily improving economic trend throughout the period. You can argue that the Clinton administration provided a better economy than Reagan and H. W. Bush did, but only if you concede that W. Bush has done an even better job than Clinton. (It is, of course, possible that the 2007 and 2008 years will put Bush below Clinton, but it seems unlikely that the economy will get bad enough this year to counteract the weight of the early Clinton years. At most, it will make the Bush administration less of an improvement over the Clinton one.) Going back farther, Carter and Nixon/Ford had the lowest poverty rates of any recent president(~11.9%), while JFK/LBJ had by far the highest (~17.5%). The does not seem to be any clear correlation between administration party and economic well being. Instead, the macro trend is that poverty has been steadily declining and median income has been steadily rising throughout the period since 1960.

Let us turn now to abortion rates and any correlation they may show with economic data. The Guttmacher Institute is generally considered the most reliable source for data on abortion in the United States. They are somewhat affiliated with Planned Parenthood, but their data is used and respected by those on both sides of the abortion debate. In their data they provide two statistics which are population-adjusted: Abortion Rate and Abortion Ratio. The former is the number of abortions per 1,000 women aged 15-44. The later is the percentage of pregnancies (which end in either live birth or intentional abortion) which ended in abortion. Thus, the rate gives us a good idea of the population-adjusted frequency of abortion within society, while the ratio allows us to control for lower incidence of pregnancy due to contraception. (Guttmacher data below is drawn from this study.)

The macro trend in abortion rate and ratio is that abortion skyrocketed from 1973 (when Roe v. Wade effectively removed nearly all abortion regulation) through 1980, when both flattened off and began to drop. In 1980, 30% of all pregnancies ended in abortion.

The graph above selects data from 1980 through 2006. I'm assuming that the rapid increase in abortions from 1973 is the result of increasing social acceptance and availability, and thus can not be expected to correlate to any factor under examination other than availability and lack of regulation.

As you can see, the downward trend in abortion ratio and rate is very clear. I've plotted the rate against a simple linear trendline, and the relation between these shows an R-Squared of 0.96. The trend for the poverty rate is also downwards from 1980 to 2006, but the slope is much more gentle and the correlation to a simple linear trendline is much lower (0.5). It would be most accurate to say that the poverty rate is oscillating with a slight downward trend. Yes, there is something of a correlation between the abortion rate and ratio and the poverty rate, but only to the extent that both are generally trending downward.

Let's try to get a good visual feel for this. On the top chart, I've plotted the abortion rate against the poverty rate. The RSQ is .47. On the bottom chart, I've charted the abortion rate against the number of years since 1980. The RSQ is .96.


So, based on the trends of the last 25 years, if you wanted to successfully predict what the abortion rate will be in a given year, you're going to be far more accurate if you base your projection on the relation between the number of years since 1980 than if you base it on the poverty rate.

But let's try something else. Say that we look at all the periods during which the poverty rate was going up and see if the abortion rate went up or went down less in those years. The following chart selects the three periods since 1979 when the poverty rate went up year over year.

The trend is actually very interesting. As the poverty rate was rising sharply from 1979 through 1982, the abortion rate dropped. The same thing happened during the poverty rate rises of 1990-1993 and 2001-2004. Even more interesting, however, is that during these first two downturns, although the abortion rate levelled in the first case and continued dropping in the second, the abortion ratio increased during both of those periods. What that means is that although women were not getting abortions at a higher rate during these periods, they were conceiving at a higher rate. People avoided pregnancy at a higher rate (thus decreasing the number of planned pregnancies) but did not abort at a higher rate.

If people were aborting more because of the rise in poverty, one would expect to see the actual abortion rate go up during these periods. Instead, we see that people avoided pregnancy (thus decreasing the percentage of total pregnancies which were "planned pregnancies") but actually reduced the rate at which they aborted.

Perhaps because the increase in poverty under the current administration has not been as severe, as those in the early 80s and early 90s, there has been no rise in either the abortion rate or ratio during that increase in poverty rate.

But has the rate of decline in the abortion rate and ratio slowed under Bush as compared to Clinton?

At first it might appear so. During the Clinton administration the abortion rate declined an average of 0.55% per year, versus 0.38% per year from 2001 to 2005 (more current data is not yet available from Guttmacher.) However, something interesting appears when we look at the abortion ratio, which has declined more rapidly under Bush than under Clinton. So the apparent Clinton advantage is a result of an overall decrease in pregnancy under Clinton. If we control that by looking at the ratio (the percentage of pregnant women who abort) we find that it has actually declined faster under Bush than under Clinton.

This got me interested in looking at how the pregnancy rate and the abortion ratio are related. I used the abortion ratio and the total number of abortions from the Guttmacher data and combined them with the annual total population figures from the Census Bureau. To get a roughly accurate female population figure, I divided the US population in half. Then I used the abortion ratio and the total number of abortions to calculate the total number of pregnancies, and I divided those by the female population to get a pregnancy rate. (One of the reasons this rate runs so low is that this would be total female population, not just the female population of childbearing age.) I charted these two rates below. (Note that this chart flows in the opposite direction from the others in regards to date.)

This perhaps begins to hint at some of the real causes of abortion trends over the last thirty years. Note that right after Roe the pregnancy rate increases by 25% over eight years. During the same period, the abortion ratio increases 50%. In effect, by decreasing the obstacles to ending an unwanted pregnancy, Roe significantly increased the number of unwanted pregnancies, resulting, in fact, in a net increase in births. As the abortion ratio has consistently fallen over the last twenty years, the total number of pregnancies has fallen back into a more normal trendline.

Why has abortion really been falling? I think it's significant that the abortion rate is falling in such a tight correlation to the number of years since the peak. This indicates, it seems, some sort of self-correcting mechanism going on. Perhaps it's partly a re-introduction of restrictions on abortion, both cultural and legal. Perhaps it's partly a build-up of painful experience, which has overcome the initial impression that the costs of getting pregnant (and getting out of getting pregnant) are not as high as they were before 1973. Either way, it seems that some force that is building with time is continuing to drive the abortion rate down without any current signs of slowing.