Scott Sumner recently came up with a new proposal for the Fed:
I’m going to propose a compromise between the current policy of $40 billion bond purchases each month, and a radical policy of immediately targeting the forecast. Have the Fed start QE3 at $40 billion per month, and then increase their purchases at a rate of 20% each month, until they have achieved their policy goal (of equating predicted nominal growth with desired nominal growth.)
This proposal would most likely pack a punch and raise expected nominal income growth. It would also once and for all settle the debate on the efficacy of monetary policy at the zero bound. There is no doubt in my mind what the outcome would be.
A big problem, though, with implementing so much monetary firepower is the absence of a well defined policy goal. Currently, the Fed has a vague policy goal of improved labor market conditions in a context of price stability. No one really knows what that means. Scott would like to change that by having the Fed target the NGDP forecast using NGDP futures contracts. This is a great idea as it would put monetary policy on autopilot, increase transparency, and respond to money demand shocks. But the Fed is along way from adopting such a goal.
Here is a suggestion for a less ambitious way for the Fed to target the forecast. First, the Fed would set a target for average nominal household income growth over the next year. Second, the Fed would contract with multiple polling organizations to do a weekly poll where they ask households how much they expect their dollar incomes to grow over the next year. Third, the Fed would take some average of these polls and compare it to the Fed's targeted growth rate for nominal household income. Fourth, the FOMC wold then conduct open market operations to bring household's expected dollar income growth in line with the Fed's target growth rate.
If more immediate feedback was needed, the Fed could contract to have the polls done twice a weekly. This still wouldn't give the instant feedback a NGDP futures contract would, but it would be straightforward to implement. The Thompson Reuters/University of Michigan Survey of Consumer Sentiment currently asks that question every month. So it should not be that hard for them and other polling firms like Gallup to ask the question weekly.
Question: would this proposal be susceptible to the circularity critique? If so, what could be done to fix it?
Very interesting ideas.
ReplyDeleteIt seems to me, with so much data online, that tracking economic growth in real time should be easier.
Perhaps very large retailers--Wal-Mart, Target etc. would agree to provide data, if confidential. They track sales daily.
Certainly other data must be out there in real time, like toll bridge crossings in NYC, or container loads at ports etc.
I worry about polling, although certainly it could be useful. More and more people avoid any pollsters, btw. In the old days, you could reach people by telephone. Cell phones have killed that. Samples are getting skewed. Even exit polling at US elections---once considered the gold standard of polling---have proved misleading (or proved our ballots are not properly counted).
It is odd that our data collection ability has radically improved in the last 20 years, and the amount of data that is collected (privately) has exploded, but we still respond to monthly and already out of date figures liked trained donkeys.
I do think it is susceptible to the circularity critique, although given the imperfect nature of the forecast, that wouldn't necessarily be a deal breaker. The other potential problem is that people might interpret "household income" as wage income, excluding capital income. Thus a boom might initially boost capital income, and be missed by this survey
ReplyDeleteScott,
ReplyDeleteI was afraid it might be. In response to your second point, maybe the question be precisely phrased so that it mentions capital income. In any event, my first choice would be to go straight to NGDP futures.
Does anyone know why I can post comments from my office computer, but not my home (iMac) computer?
ReplyDelete