Michael Woodford is hell bent on creating a monetary economics that revolves around interest rates, with no role for base money. That makes him the opposite of me, as I’d like to get rid of interest rates, and focus entirely on the supply and demand for base money. We both have our thought experiments. He has a “money-less economy” model (which actually has money–reserve balances.) And I have a model with no interest rates (which actually has interest rates–the return to real capital.)
I’m working through Woodford’s paper, and it’s a brilliant piece. But there’s one big flaw. In his zeal to brush aside the monetarist approach, he uses the Japanese QE example to try to discredit monetarism. In doing so he attacks a straw man. He attacks a position that (AFAIK) no monetarist has ever taken. Here’s the proposition that he claims (pp. 55-56) monetarists believe in:
Suppose a central bank injects a lot of cash in the economy. The central bank simultaneously announces that the injection is temporary, and that the cash will be removed the minute prices start to rise. Woodford seems to believe that the quantity theory predicts that the large monetary injection will cause a large increase in prices and NGDP. He doesn’t say so, but presumably he also believes that the quantity theory predicts a big fall in the price level when the money is withdrawn from circulation.
I went to the University of Chicago, and I don’t ever recall learning any theory like that. I’ve read lots of articles by Friedman, Schwartz, Brunner and Melzter, but I don’t recall any of them making that claim. And there’s good reason why no monetarist would make this argument, it would imply completely bizarre fluctuations in the real interest rate. Back in 1993 I published an article that explained this with a simple thought experiment:
The Fed doubles the money supply from $1 trillion to $2 trillion, and simultaneously announces that a year from today the money supply will be cut in half, back to $1 trillion. Suppose the price level doubled, and was expected to fall in half next year. Then the real return from holding cash would be 100% over the next year. Since risk free interest rates are usually around 2%, not 100%, this would mean that money would be wildly non-neutral. But that makes no sense, as the neutrality of money is one of the central tenets of monetarism. Something’s wrong here.
Here’s what monetarists actually believe. A one-time, once-and-for-all, doubling of the money supply will cause a sharp rise in prices. That true whether interest rates are zero or positive. As far as I can tell, Woodford accepts this view. So it’s actually not at all clear where he disagrees with the monetarists.
It seems like Woodford wants to divide the classic monetarist experiment into two parts; the initial injection, and then the permanent increase in the money supply. (Krugman does this too.) The second part (the permanent increase) is then viewed as non-monetarist policy, a sort of signaling issue. The first part by itself has almost no effect, and hence monetarist stimulus ideas are labeled ineffective at the zero bound. But of course by that logic they’d also be almost ineffective at positive rates. If you inject money and pull it out a year later, NGDP won’t change (very much) even if rates are positive at the time of the initial injection.
The problem with this framing technique is that the same could be done with the 2001 interest rate cut I described in this post a couple days ago, which immediately boosted stock prices by 5% and helped prevent a deep post-tech bubble slump. That Fed action could be divided into the initial rate cut, and then expectations of lower future rates (relative to the Wicksellian equilibrium, as longer term interest rate expectations actual rose on the income and inflation effects.) The initial cut did almost nothing. So by that logic the classic Keynesian thought experiment (cutting rates) is nearly as ineffective as the classic monetarist policy. But somehow it worked. To be clear, I’m not contesting Woodford’s claim that the long run trajectory of policy is the key, rather I’m contesting how he interprets that fact.
Woodford then makes some odd assertions, which suggest he thinks Japan’s QE of 2001-06 was a failure.
And as the theoretical models would predict — but contrary to the quantity- theoretic reasoning that had provided the basis for the policy proposal — there was little effect of the policy on [Japanese] aggregate nominal expenditure. As shown in Figure 14, the increased supply of bank reserves raised the total monetary base by 60 percent over the first two years of the policy, and eventually by nearly 75 percent. Yet there was no corresponding increase in aggregate nominal expenditure: nominal GDP was only six percent higher after five years of QE than it had been in the first quarter of 2001, despite a massive increase in the monetary base. And as the figure also shows, deflation (here measured by the GDP deflator) continued unabated.
. . .
The BOJ itself appears no longer to put great stock in pure QE as a policy. Its “comprehensive monetary easing” policy, introduced in October 2010 in response to continuing concerns about the economic outlook, has again resulted in a significant increase in the size of the BOJ’s balance sheet, but does not involve quantitative targets for current account balances.
And yet by his own account the BOJ discontinued QE because it had succeeded in boosting the CPI by a small amount:
As Figure 14 shows, most of the increase in current account balances was reversed, in the space of a few months, after the policy was suspended in March 2006, as a result of CPI inflation that had been measured to be slightly above zero.
So by the BOJ’s criteria it worked. That’s why it was discontinued, that’s why much of the money was pulled out of circulation, and that’s why interest rates were raised in 2006. I think people like Woodford and Krugman misinterpret the Japanese case because by their criteria (and by mine) it failed. Japan shouldn’t have been shooting for zero CPI inflation, but rather a slightly higher number. But QE did what they wanted it to do. Then Woodford suggests the BOJ reacted to this failed experiment by abandoning “QE as a policy.” But in fact they still use it. Yes, it used in conjunction with other tools such as a new and more explicit 1% inflation target, but that’s perfectly consistent with monetarism, broadly defined.
I do think it’s fair to criticize the older monetarists for not putting enough weight on managing expectations (that’s the big contribution of Krugman/Woodford/Eggertsson), but that may be partly due to their advocacy of level targeting monetary aggregates, which avoids some of the pitfalls of purely “forward-looking” policy regimes.
Here’s how I see things:
1. There are no known cases of fiat money central banks trying to inflate and failing.
2. Markets have often responded strongly to slight easing by the Fed, ECB, and BOJ, even though these moves weren’t even 10% of what was needed. Imagine how strongly they’d react to something serious! So the markets certainly don’t think central banks have run out of ammunition. This is important because new Keynesian models suggest that the key to success is raising market expectations of NGDP growth.
3. Everyone agrees that if the Fed bought up all of planet Earth, NGDP would rise sharply. So the “worst case” is that the Fed needs to engage in “unconventional purchases.”
4. Even though it’s extremely unlikely the Fed would ever have to engage in unconventional purchases, even that option is 100 times better than grotesquely inefficient fiscal stimulus.
So if the Keynesians are going to get us market monetarists to abandon our monomaniacal obsession with relying solely on monetary policy to steer NGDP, they’ll have to come up with far better arguments than what I see in Woodford’s paper. It’s a brilliant paper, but nowhere near good enough.