In his NYT mag piece, Paul Krugman blames macroeconomists for believing the world behaved as well as the math behind their models. His solution? To re-embrace Keynes.

But in a provocative counterpoint on VoxEU, Scott Sumner says looking back to Keynes won't solve the problem. Instead he claims that macroeconomists didn't trust their models enough, and that the latest developments in the field should have helped prevent the crisis.

Sumner's main argument is that the Fed didn't loosen monetary policy before and after Lehman's bankruptcy, as most of us believe, but in fact tightened it--this despite signs the economy was slowing rapidly. But how can that be, especially since Allan Meltzer and the like have been crying about the hyperinflationary risks of easy money?

It boils down to the Fed's ability to pay interest on excess reserves. Although the Fed has created a vast amount of extra cash, banks are happy to just hold on to it because the rate being paid on reserves (that is, the money they park at the Fed) is just as good as or better than what banks could get elsewhere.

The Fed says this is a good thing because it will allow the central bank to raise rates without causing inflation once the economy starts looking normal again. The Fed can do this by keeping the rate it pays on reserves roughly equal to, or above, the fed funds rate--that is, the key short-term interest rate. (Right now, the reserve rate is 0.25% while the fed funds rate is between 0-0.25%).

The big problem with this, says Sumner, is that it prevents the economic activity from getting back to normal sooner:

...most economists ignored the Fed’s 6 October 2008 decision to start paying interest on reserves – which meant the Fed bribed banks to hoard all the extra liquidity they were injecting into the system. If this sounds unfair, consider that the Fed itself indicated that these payments were necessary to prevent market interest rates from falling; an explanation that Woodward and Hall (2008) correctly described as “a confession of the contractionary effect.”

The Fed’s decision to double reserve requirements in 1936-37 is often cited as a contractionary mistake that prolonged the Depression. The 2008 decision to pay interest on reserves had the same effect, increasing the demand for reserves. Surprisingly, few economists seemed to notice the parallels with 1937, despite the fact that in 2009 nominal GDP is expected to fall at the fastest rate since 1938. I have argued that the Fed should instead charge a penalty rate on excess reserves, and the Swedish Riksbank recently adopted this strategy.

Sumner also points to stock, bond, real estate, and commodity markets -- which all crashed between July and November 2008 -- as further evidence that monetary policy was too tight "for the needs of the economy." He goes on to point out some research over the last decade which may have helped prevent the post-Lehman crisis. The entire thing is well worth reading.