A Fact in Search of a Theory

A few weeks ago, I read the following fact about U.S. and international stock markets in the Wall Street Journal:

During the two-year period that ended in February, correlation between U.S. and other developed markets was 0.63, according to ING Asset Management. That is a big decline from 2003 to 2005, when they practically moved in lockstep, at 0.93. (The figures are based on monthly movements in the Standard & Poor's 500-stock index and the Morgan Stanley Capital International EAFE indexes.)

This fact raises many questions in my mind, which I have been contemplating without much progress. What causes this correlation to change over time? Does it say something about the nature of the shocks hitting the world's economies? What implications does the changing correlation have? Do we observe more efforts at international diversification when the correlation is low? Does a falling correlation mean that risk premiums should shrink because international diversification reduces risk more? Do risk premiums in fact move together with changes in this correlation?

If anyone knows the answers, please share them in the comments section.