Mankiw on monetary policy options

Back in 2006 I read an interesting post by Greg Mankiw, which advocated a monetary regime where the Fed would use market expectations to set the monetary policy instrument at the level expected to hit their inflation target.  Given my longstanding interest in futures targeting, I was naturally encouraged to see a highly respected mainstream new Keynesian endorse this type of policy regime.  Before looking at his specific plan, however, it will be useful to review his more recent essay on negative interest on money.

Mankiw got a lot of flack for suggesting (somewhat whimsically) that the liquidity trap could be sprung with a policy of negative interest on cash.  He noted that a Harvard grad student had even suggested a way of doing this—the government could announce that all cash with serial numbers ending with a randomly selected digit would periodically be invalidated.  Mankiw did realize that the plan was politically problematic, but regarded it as a useful thought experiment.

I immediately saw the similarity to my penalty rate on excess reserves idea, and tried to convince Mankiw of the superiority of my approach.  I still don’t know whether he opposes a penalty rate on excess reserves, but I think it is fair to say that he was less than enthused by the idea.  Why?  Because he correctly noted that it would not lower nominal rates significantly below zero, which was the objective of his negative interest on cash idea.  In contrast, I tend to view monetary policy through the lens of changes in the supply and demand for base money, and saw huge advantages flowing from an $800 billion dollar drop in demand for excess reserves (even if it led to increased hoarding of cash.)

Thus the fundamental point of contention was that he saw policy in terms of the new Keynesian transmission mechanism, which focuses on the role played by changes in the Fed’s interest rate target.  In contrast, I took a more monetarist approach to the transmission mechanism.

Over the past few weeks I have occasionally thought back to Mankiw’s 2006 market-oriented policy proposal, and wondered why he didn’t resurrect that idea.  After all, in my previous post I argued that futures targeting offers a foolproof way out of a liquidity trap.  So today I went back and looked at the plan, as you can see it is slightly different from my proposal, and different in a very revealing way.  Under my plan, there would be no single monetary policy indicator.  Each trader could use their preferred policy indicator (the base, M1, M2, fed funds rates, the price of gold, etc.)  My plan would literally turn the investing public into the FOMC.

Under Mankiw’s proposal, the Fed would still determine the appropriate instrument of monetary policy—and he chose (not surprisingly) the fed funds rate.  So he envisions a system where the market would predict the fed funds target most like to hit the Fed’s inflation target.  Now I suddenly understood why Mankiw had not resurrected this idea for the current crisis.  The fed funds instrument seems ineffective in a liquidity trap.  The Fed needs some other, unconventional way of adopting a more expansionary policy stance.  They cannot lower short term nominal rates much further.

So once again, in a completely unrelated policy issue, we find that nominal interest rate targets are the key sticking point.  As long as one sees monetary policy in terms of the control of short term interest rates, one will have difficulty envisioning quick and effective ways out of the liquidity trap.  That’s not to say that Mankiw doesn’t have some other good ideas—for instance he favors a 3% explicit inflation target (level targeting) as a way to boost inflation expectations.  I think that is a great idea, and have also advocated similar explicit targets.  But when you are playing the expectations game, you are always subject to the skeptics who insist central banks’ reputations are too conservative to generate suitable inflation expectations.

In contrast, my idea doesn’t have that problem.  The market becomes the FOMC, so there can never be a disconnect between the target and the forecast.  Mankiw’s market-oriented target idea might also work, but only if we can be sure that we won’t face a zero bound.

BTW, I should probably clarify one point.  In my previous post where I explained the mechanics of futures targeting, I described how futures trades would impact the monetary base.  I did not mean to suggest that the monetary base would become the “policy instrument” in the ordinary sense of the term.  For instance, if most investors were Keynesians, they would presumably buy and sell NGDP futures until the market fed funds rate moved to a level expected to generate on target NGDP growth.  At the zero bound they would have to look at something else, but it might be the base, M1, the price of gold, stocks, etc.  There is no single policy instrument.  Of course changes in the monetary base are a part of the policy proposal, but that is even true when the fed funds rate is the policy instrument—which require OMOs to implement changes in the base.



7 Responses to “Mankiw on monetary policy options”

  1. Gravatar of StatsGuy StatsGuy
    10. May 2009 at 16:45

    How would you suggest making such a mechanism resistant to speculative action and manipulation?

    I would expect liquidity of a direct market on the inflation rate to be thin relative to the liquidity on – say – the entire bond market? This creates incentives to manipulate the inflation market (at a loss) to earn money in the bond market as a result of Fed action. Presumably, other investors could counter, but this puts us in a massive game – with extreme uncertainty surrounding outcomes, and huge incentives for collusion among large (invisible and hard to regulate) private funds.

    Or, were you suggesting that the “inflation target” be derived from some much larger market (say, the bond market)?

    Or, do you simply mean identifying a derived inflation projection from an existing market, and using this to create a “guiding principle” for the Fed to use in determining their rate setting actions?

  2. Gravatar of Nick Rowe Nick Rowe
    10. May 2009 at 18:25

    Scott: I would describe your proposal as saying there is one policy instrument: changing the base via unsterilised OMOs. But people are free to look at whatever indicators they think are important. It’s multiple indicators, not multiple instruments.

    If, for example your proposal were changed to read: “The Fed will raise the price at which it buys and sells gold by $0.01 for every $100 spent on the NGDP futures contarct, I would say that’s a gold price instrument, but people are free to look at any indicators they want.

  3. Gravatar of septizoniom2 septizoniom2
    11. May 2009 at 01:17

    another excellent post. safe to say all of your posts are of the highest quality.
    have you ever run a thought experiment of what would be the end result of doing nothing? i wonder about the nature of the demand for money in times of panic.

  4. Gravatar of ssumner ssumner
    11. May 2009 at 03:08

    Statsguy, All your questions are answered in my previous post, where I discuss the mechanics of the policy. Here I’ll just add that I seem to recall studies of prediction markets showing manipulation is not a big problem. In addition, the Fed would be free to offset manipulation (suspiciously large trades by one person or bank) with its own OMOs. Plus the market will be subsidized so it will be large.

    Nick, I think we both might be wrong, but am not sure. If the price of gold is the instrument in your example, then the price of NGDP futures is the instrument under my plan.

    I find this very confusing, so I might well change my mind again. But here was my original thought–if I am not mistaken, the monetary base is not currently viewed as the instrument of policy, (at least until last year) the fed funds rate was the instrument. But under fed funds rate targeting the base changes to move the fed funds rate where you want it to go. So one can hardly argue that something is an instrument, merely because OMOs cause it to change. And my plan clearly does not call for targeting any particular base level, rather it calls for targeting NGDP futures prices. But then maybe I don’t really know the correct definition of “instrument.”


    One could write a book on doing nothing. There are many “nothings” each with vastly different consequences:

    1. No change in interest rates—which will eventually make the price level explode to zero of hyperinflation

    2. No change in the monetary base–a problem if hoarding or dishoarding sets in.

    3. No change in M3–still susceptible to hoarding in an FDIC world, less so in the 1920s.

    3. No change in the price of gold or foreign exchange–here it depends on how stable other currencies are. This can work reasonable well Bretton Woods, or not so well Argentina late 1990s.

    4. No change in the price of an index futures contract–best option.

  5. Gravatar of Dann Ryan Dann Ryan
    11. May 2009 at 04:37

    This past week I was at a conference where the keynote speaker was a certain member of the Fed Board of Governors. He was speaking completely off the record and answering questions.

    I really wanted to ask him his thoughts on your negative reserve rate policy, but lacked the nerve. I couldn’t formulate my question in a manner that would not make me sound completely daft.

  6. Gravatar of Aaron Jackson Aaron Jackson
    11. May 2009 at 06:14

    Scott, in reference to Statsguy’s issue of market manipulation and your response: Here are some papers on why market manipulation may not be much of an issue. See Hanson, Oprea and Porter (2006), Wolfers and Zitzewitz (2004), and Hanson “Foul play in information markets” (2006).

    Dann, you should have went for it. We had the former president of the Boston Fed speak at Bentley University a couple months ago. A student of mine asked her afterwards about the interest payment on reserves issue, and she kind of danced around the question, before finally implying that she thought there were vested interests (I think in reference to the people who designed the policy) who were determined to see the program continue, and had some political clout. This is all secondhand, of course, from my student, but that is the impression I got from the explanation from my student.

  7. Gravatar of ssumner ssumner
    11. May 2009 at 16:23

    Thanks Aaron, I recall that there were studies showing that there was little risk of market manipulation, but I didn’t recall which studies.

Leave a Reply