Here’s a view Krugman attributes to me:

Of course, there’s also an alternative universe in which insurance
companies would never, never treat their clients badly, because that
would hurt their reputations.

Here’s what I explicitly said in my original piece on insurance and reputation:

Am I saying that health insurance companies never play dirty
tricks on their customers?  Of course not.  It’s a big world, lots of
bad stuff happens.  What I’m saying, rather, is that reputation works
well even in industries where firms have big, lumpy liabilities.

If the .5% rescission number that Krugman cites were true and correctly interpreted as fraudulent-in-spirit, I’d admit that reputation in the health insurance industry isn’t working too well.  But I suspect that the real story is usually that (a) insurers are well within the letter and spirit of their contracts, and (b) are fighting genuine customer abuse. 

In any case, no matter how you interpret the .5% number, there’s good reason to be skeptical of it.  A recent non-Krugman editorial in the NYT suggests a figure closer to what I’d expect:

But Congressional investigators found that three big insurers canceled
about 20,000 individual policies over a five-year period — allowing
them to avoid paying more than $300 million in medical claims.

In 2008, U.S. health spending was about $2.4 trillion; 2004-2008 spending would be roughly $10 trillion.  So even if these three big insurers held just 1% of the market each and their behavior were typical, we’re talking we’re about .1% of total spending.  If companies were really targeting big claims, the fraction of affected individuals would be much smaller. 
Question: If Krugman really wanted to fight dishonest health insurance companies, wouldn’t it be more productive to produce a table of companies’ rescission rates, so customers would know where take their business?  Can anyone find a link to such a table?

* Note: Even Krugman hedges his bet with an “if that’s really true.”