yet again, the conventional divorce rate is pure bunk

So Matt Yglesias and Mike Konczal  and presumably others are talking about divorce rates and the economy. Unfortunately, they are taking the conventionally-calculated divorce rate at face value, when in fact the statistic is useless. I bring this up again and again. As it only tracks number of divorces compared to number of new marriages, the conventional divorce rate is both easily susceptible to small numbers of serial marriers, and more importantly, matching the data pool of divorces from all existing marriages against the data pool of just new marriages. A couple that has been together for years and stays together has no positive impact on the divorce rate of any given year, despite the fact that this is precisely what we're interested in. Indeed, a couple that stays together until death is never represented in the divorce rate at all other than in the year that they are married. A more accurate divorce rate, tracking individual marriages until they end in divorce or don't, is a much more difficult statistic to derive, but much more accurate. The best info I've seen is that the true divorce rate peaked in the early 90s, never exceeded 40%, and has declined since.

This isn't some crank view of mine, by the way, but a well-known problem in sociology. Quoth the New York Times: 

But researchers say that this is misleading because the people who are divorcing in any given year are not the same as those who are marrying, and that the statistic is virtually useless in understanding divorce rates. In fact, they say, studies find that the divorce rate in the United States has never reached one in every two marriages, and new research suggests that, with rates now declining, it probably never will.

This particular discussion that Konczal and Yglesias are having is actually the perfect example of when the conventional divorce rate is most misleading. I don't blame them at all for using the conventional divorce rate, as despite its lack of analytic rigor, the popular press never stops using it. Maybe somebody with a better reputation than I have could get the word out.