I can help a bit, Kevin

What’s the Deal With Bayesian Statistics?

Nice illustration.  

The key to understand the problem is that a conditional probability means correlation means off the path to equilibrium which requires equally surprising events A and B. 

Thus the conditional probability cannot last as the economy adjusts, and flat earthers carry on the conditional, hedges assume it wall fade.

Hence the problem with Keynes. He always kept the private sector conditional on the public sector when the two groups actually tend to keep their distance so as to avoid the conditional.  Like in California, we all want to avoid having our lives conditional on the public sector unions and tend to avoid that situation.

This is equivalent to removing loops, what hedges do, minimize the deja vu issue.  Nancy has the same problem, she gets the 'Not another time around" because she needs to keep her programs conditional on the republican programs.  Our government thus engages in cycles, the loops never closing.

Glad to see Kevin look into this, it will lead him to closing the system and looking at the complete sequence.  Krugman only wants us looking only one cycle back, he does not want us seeing the constant deja vu moments.

Boy it is rather pleasing to see more and more pundits and economists figure this out.