From his own blog

I have a very unconventional way of thinking about the monetary transmission mechanism. A monetary shock is essentially a change in the future path of the money supply relative to velocity, in other words, a change in the expected NGDP growth. If the NGDP target is 4%, for instance, than an expected NGDP growth rate of 3% is tight money, and an expected NGDP growth rate of 6% is easy money.

The Taylor rule says that you adjust interest rates by a specific amount in response to recent past performance of inflation and unemployment. The Sumner rule says you do whatever it takes to make market indicators of expected future nominal GDP growth hit your target.

A question that I discussed at lunch and afterward with some George Mason folks and John Haltiwanger of Maryland is: why didn’t Ben Bernanke expand the money supply more last fall? Possible answers.

1. He did not really foresee how badly things would turn out in terms of unemployment. Sumner would say that Bernanke should have looked at the market indicators that suggested very low nominal GDP growth going forward.

2 (or perhaps 1a). He thought that as long as there were no major bank failures, the economy would not do too badly.

3. With interest rates at zero, he needed to do quantitative easing, for which he lacked the authority, and Henry Paulson would not co-operate (Tyler Cowen seems to like that explanation).

I lean toward (2). He was more worried about the banking system than the rest of the economy.