This is perhaps a long-winded way of saying that I now think Scott Sumner and David Beckworth have been right all along and we need to ditch talking about “inflation” in favor of something that means “nominal GDP.” I think “total spending” may be the right turn of phrase.
I’m glad Matt Yglesias has declared himself on this point, so I now know what position I’m arguing with when I argue with him.
My reasons for hesitancy about endorsing nominal GDP as the target for monetary policy remain the same:
- It’s a fairly radical departure from established practice (although an interesting one!).
- It perversely “punishes” spurts in productivity (which increase real growth) with higher interest rates, and “rewards” drops in productivity (which reduce real growth) with lower interest rates; similarly, it calls for looser money when there’s a commodity-related inflationary shock (such as an oil spike) because such shocks reduce NGDP growth even though they (temporarily) increase inflation – and for tighter money when the opposite happens.
- It implies either very volatile interest rates (as expectations for NGDP are more volatile than inflation expectations), or unrealistically good forecasting of NGDP, or overly slow responses (we didn’t realize how badly NGDP had dropped in the recent recession until well after the Fed acted). Scott Sumner thinks that a robust NGDP futures market could give the Fed the tool it needs to see changes in expectations swiftly, and respond accurately. I’m more skeptical; the stock market is a pretty good proxy for such a market, and the stock market predicted (as the saying goes) twelve of the last five recessions. People seem comfortable with TIPS as a guide to inflation expectations, but again: inflation expectations have been less volatile than nominal GDP growth expectations.
My hesitancy about “catch-up” spending growth is something I’ve also articulated before: I don’t think there is a symmetry here. I haven’t heard Yglesias say that one problem with the 1980s is that interest rates were consistently too low, and that we needed to keep nominal spending well below 5% for a period to “catch up” with the fact that we had such high nominal spending in the 1970s.
But I don’t feel the need to present a strenuous objection to an NGDP rule. It’s a rule. If I understand right, Scott Sumner thinks it was effectively the rule being followed under the Greenspan Fed in the 1990s, though Greenspan never formally adopted a rule of any kind. Then again, Taylor thinks Greenspan was following his rule. So during the most successful period of central bank management in recent history – we got stable growth, low inflation, and an unemployment rate down to levels not seen since the early 1960s – we don’t know what rule was being followed. Which – maybe – suggests that underlying economic conditions were favorable enough that monetary policy was simply easier in the 1990s than it had been before or would be subsequently.
Which is a big part of why I am resisting the call to focus so much on monetary policy as the solution to our current economic problems.