“Just when I thought I was out… they pull me back in” I tried to resist asking “what does “Nominal GDP targeting even mean.” I managed, but now Krugman is burying the hatchet and I am digging it up.
So what is the proposal ? That the Fed have a target for 2012 nominal GDP or first quarter of 2012 nominal GDP ? Even if it isn’t measured, the Fed could try to get the November 2011 nominal GDP it wants. As far as I know, advocates of nominal GDP targetting don’t even acknowledge this question.
I have some longer and more substantive thoughts below.
update: spelling of title corrected. I thank Brad DeLong
I think Krugman understates his case when he claims that the Fed can’t target nominal GDP when we are in a liquidity trap. I would define targeting X as making the conditional expected value of X equal to the target. There will be a disturbance, but if the expected value is different from the target no matter what one does, then on can’t target X. The concept of daily GDP is meaningful (although it would be crazy to try to measure it and correct accounting for inventories would be key). Do quasi-monetarists really think that the Fed can make the expected value of tomorrow’s nominal GDP whatever it wants ?
I admit I am being fairly twitty, but I think this question isn’t totally stupid, because I think it shows that they just don’t think about what monetary policy can and can’t do. The Fed can move the Fed funds rate very fast. The Fed can change the money supply quickly at least if it wants to reduce it or we are not in a liquidity trap. Nominal GDP can only jump if prices are flexible. Monetary policy is effective because they are sticky. We have a problem.
OK a more serious issue. Can the Fed get the 2012 annual nominal GDP it wants by buying Treasuries. Jah Hatsius (and Brad DeLong) argue that the Fed should declare its intention of buying whatever quantity it takes of long term Treasuries to achieve a nominal GDP target. But what if there is no such quantity ? Then the announcement would be a false claim.
Is there any such quantity ? I think not. Certainly not if one wants 2011-2012 growth to be well over the trend growth rate say 10% (Brad DeLong seemed to call for this when he said to target the level). I admit that I will go rational expectations and argue that if it can’t happen, then people won’t believe it can happen. So I assume model consistent expectations.
First the Fed can drive the risk premium on Treasuries to zero. The public will accept returns equal to the risk free rate if the Fed buys all the treasuries on the market. So far, I think this would be a small effect. There just isn’t much room for a risk premium even in 30 year rates. Importantly, a zero short term rate and a zero risk premium does not imply a zero long term rate. Even without a risk premium the long term rate would be equal to an average of current and future short term rates. We are discussing about a mega QE policy not a credible commitment to causing high inflation in the distant future when unemployment is normal (nor are we discussing a unicorn or a flying pig).
I would expect a small effect on interest rates — about on the order of the effect of operation twist (of course I predicted a smaller than observed effect of operation twist). Note this is for about $10 trillion of QE. The effect of the volume on the market on risk premia is not linear.
The risk is mostly inflation risk (not default risk — that is best hedged with canned food, bottled water and guns). But wait aside from the illusory wealth of Treasuries, total inflation risk adds up to around zero as there are nominal creditors as well as nominal debtors.
I have admitted I use rational expectations. And I have assumed that Fed policy now doesn’t have a big effect on expected Fed policy in 2020.
OK so mega QE if needed to target nominal GDP levels might work if it massively affects expectations somehow, even though there is a rational expectations equilibrium with a small change in expectations. But that sure sounds a lot like the confidence fairy to me.