In the previous post I explained why the commonly held notion of long and variable lags is a myth. I didn’t mention the reason for this myth. It results from economists trying to cover up the fact that the monetary policy indicators they have fallen in love with (interest rates, the money supply, etc) don’t actually identify changes in the stance of monetary policy. Recent work by Woodford and others shows that it is changes in future expected monetary policy that drive current AD. Krugman also did important work in this area, and used this basic idea a few weeks back when he argued hawkish statements by Fed officials were already slowing the economy.
And in fairness to Mr. Bernanke, discord among senior officials also makes it difficult for policy to change expectations: it would be hard to credibly commit to higher inflation if this commitment were constantly being undercut by speeches out of the Richmond or Dallas Feds. In fact, I’d argue that loose talk by some Fed officials is already having a negative economic impact.
Suppose Fisher, Plosser, and Hoenig make a bunch of scary sounding statements about exit strategies, and that causes people to expect tighter money next year. Why would that cause the economy to slow right now? There are many reasons, but the easiest explanation involves a simple example using a perfectly competitive industry. Suppose tighter money is expected next year. Economic theory predicts this will reduce AD next year, and also depress commodity prices. But if future expected commodity prices decline, then inter-temporal arbitrage also causes current prices to fall. If wages are sticky, then lower current commodity prices will cause lower output of commodities like zinc.
Those more comfortable with Keynesian explanations might look at the process through interest rates. Tighter expected money next year will raise expected future real short term interest rates. This will raise current real long term interest rates. And this might slow housing. I personally don’t like using this mechanism, because if tight money leads to slower expected growth, it might actually depress longer term rates. Even so, it would contract current output by lowering the current price of houses.
Here’s what I find interesting about these expectations stories. Suppose you think of monetary policy in the way most people do—as open market operations that change the money supply. In that case, there may actually be LEADS in monetary policy, i.e. the effect may occur before the cause. A future money supply change may cause a current change in demand and output.
Those who are philosophically inclined may be a bit uncomfortable with the thought of effects occurring before causes. “Wait” (I can just hear you saying) “it isn’t the future money supply change that causes a current change in spending, it is the CURRENT EXPECTATION of a change in future monetary policy, and future AD.” OK, I’m willing to go along with that. We should not describe changes in monetary policy in terms of current changes in the money supply, but rather expected future changes in the money supply. And of course we also know that it isn’t just the money supply that matters, velocity is equally important. So what really matters is changes in the expected future level of M*V, or NGDP. It is changes in expected future NGDP that best characterizes changes in monetary policy, at least if we don’t want effects to occur before causes. Of course that’s exactly what I have been arguing in my blog for 18 months.
And by the way, if we define monetary policy changes in terms of changes in expected future M*V (as we should), and if the fiscal multiplier is estimated under the assumption that monetary policy is held constant, then it is a truism that the expected fiscal multiplier is always precisely zero.
Don’t like that result? Then give me your definition of “monetary policy changes” where:
1. Cause doesn’t follow effect
2. The definition does not imply money was “easy” in the Great Depression (when rates were low and the monetary base soared.)