75% of Japan's 1992-2007 decline relative to the U.S. is due to demographics

I have heard or read about this in passing but don't recall reading clear numbers to put it in perspective.  Until now.

Paul Krugman:

Using the Total Economy Database — another useful source — I find that from 1992 to 2007 (eve of the crisis), Japanese GDP per capita fell from 88 percent of US GDP per capita to 76 percent. That sounds bad, and it is. But about two-thirds of that decline can be explained by the aging of Japan’s population. According to the OECD factbook, in 1992 working-age adults were 69.7 percent of Japan’s population, compared with 65.5 in the US; by 2007, the Japanese number was down to 64, while the US number was up to 67.

But my math tells me this explains about three-quarters of the relative decline.  Output per working-age person in Japan fell from 82.7 percent of that in the US to 79.6 percent.  So, that's a relative decline of about three percentage points instead of 12.

Yes, I'm knit picking. It helps me to remember things like this.

A more careful analysis may lead to different numbers. One could take into account the whole life cycle of worker productivity in relation to the age distributions in the two countries.  One could get fancier still and account for the joint distribution of age and education. Would it be fair to assume education was exogenous to the relative decline?  Maybe not, but it would still be interesting.