The new trilemma

Ryan Avent has a very interesting new post on monetary policy.  Here’s the punchline:

It is perhaps premature to declare the existence of a new monetary trilemma, that over the medium-term central banks can choose at most two of the following: low inflation, low unemployment, and financial stability. But if Mr Bernanke continues arguing this effectively in favour of higher inflation, we may need to ask why he isn’t pursuing it as an explicit goal.

I’d put it slightly differently.  In a low real interest rate environment you can have at most two of the following three:

1.  Low inflation (actually low NGDP growth)

2.  Stable unemployment near the natural rate

3.  A Keynesian monetary policy regime (interest rate targeting)

Right now we have numbers 1 and 3.  Australia has faster NGDP growth, and hence has 2 and 3.  Or you could go with NGDP futures targeting, and have 1 and 2.

HT:  TravisV


Tags:

 
 
 

19 Responses to “The new trilemma”

  1. Gravatar of marcus nunes marcus nunes
    4. March 2013 at 11:04

    And by pursuing 1&3 he can say with a stright face that:
    “The fact that market yields currently incorporate an expectation of very low short-term real interest rates over the next 10 years suggests that market participants anticipate persistently slow growth and, consequently, low real returns to investment. In other words, the low level of expected real short rates may reflect not only investor expectations for a slow cyclical recovery but also some downgrading of longer-term growth prospects.”

  2. Gravatar of Matt Waters Matt Waters
    4. March 2013 at 11:36

    One thing I have difficulty resolving is this question. Let’s say that 5% NGDP growth means 2% inflation and 3% RGDP growth. What happens if the real interest rate necessary to achieve 5% NGDP growth is less than -2%?

    My only guess is the NGDP targeting regime gets around this question by increasing the monetary base significantly and therefore shifting the demand curve for bonds outward to where the real interest rate is -2%. Enough QE could also theoretically push up the necessary real interest rate for 5% NGDP growth as well.

  3. Gravatar of ssumner ssumner
    4. March 2013 at 11:42

    Marcus, Good point.

    Matt, Then the central bank could buy up the entire global stock of wealth, without triggering inflation of more than 2%. Good for the Fed!!

    The other answer is that NGDP growth depends on more than just real interest rates on Treasuries. The monetary base also matters.

  4. Gravatar of Geoff Geoff
    4. March 2013 at 13:38

    Mr. Smith wants sub-4% NGDP growth, hence he believes the current 4% NGDP growth means money is too loose and inflation is too high.

    Mr. Jones wants 5% NGDP growth, hence he believes the current 4% NGDP growth means money is too tight and inflation is too low.

    Who is wrong about the stance of monetary policy, and why?

  5. Gravatar of Don Geddis Don Geddis
    4. March 2013 at 16:15

    @Geoff: You’ve provided no data in your example about whether money is in fact too tight or too loose, so obviously there is no valid answer to your final question. Depending on the state of the economy, either one could be wrong, or even both at the same time.

  6. Gravatar of Bill Woolsey Bill Woolsey
    4. March 2013 at 16:17

    If “the” real interest rate necessary for 5% nominal GDP were -2%, then nominal GDP targeting at 5% with 3% potential output growth is not consistent with zero interest currency. The heroic open market operations that Market Monetarists would advocate would have to stop once the Fed owns all of the assets it is eligible to buy.

    On the other hand, if the overnight funds rate (or something ilke that) needed to be less than -2 percent, but other interest rates can be higher than that, then heroic open market operations could work. It would involve the Fed purchasing longer and risier assets.

    But also, keep in mind that in the realistic case where the equlilibrium interest rate is less than -2 on short and safe assets solely because people fear that the Fed won’t undertake heroic open market operations, then a change in that attitude by the Fed will result in higher equilibrium interest rates.

    A central bank committed to redeemability in zero interest currency and targeting short term interest rates by open market operations with short term government debt could generate a recession. Breaking some of those self-imposed constraints could promptly end the recession.

    But, we can imagine a scenario where expectations are not enough and the heroic open market operations or breaking the tie to currency would be necessary.

    And it is even possible that heroic open market operations would not be enough, and the break with currency would be necessary.

    In my view, the Fed should commit to the heroic open market operations and if they don’t do it, break the tie to currency.

  7. Gravatar of Matt Waters Matt Waters
    4. March 2013 at 16:39

    I did not mean my comment to sound like real interest rates are the only thing that matter. I said in my post that the monetary base also matters.

    Maybe my question will make more sense if I start from the EMH/Indexing paradox. Indexing is great for most individual investors (it’s what I do with most of my portfolio). But it could never work if everybody did it. At the most extreme interpretation that the EMH ALWAYS makes prices correct, there needs to be some infinite reservoir of both money and time to correct stocks to their innate value. That’s generally a pretty good approximation, especially for very transparent and fluid markets.

    There is a similar paradox for expectations/Market Monetarism, one that I’ve always had a hard time with. NGDPLT is probably the best goal for a central bank, but getting there requires a mix of concrete steppes and expectations. Volker was in fact able to control NGDP and lower inflation, but it took at least a short period of time of very high interest rates before markets believed inflation was truly ending.

    So, what exactly is the equivalent of the 18% Fed Funds rate when NGDP growth is too low rather than too high? If the Fed only uses the Fed Funds mechanism for policy, there is no restriction to tightening policy but there is the ZLB for loosening. If the Fed merely communicates its intention to NGDPLT without any firepower backing it up, it is at the mercy of the market. The market will increase NGDPLT if that’s what everybody else in the market is doing, but Fed cannot unilaterally punish the market for not getting onboard like it punished those who took out loans at 18% interest and 5% inflation.

    The most obvious mechanism that’s similar to the 18% interest rate is charging interest on excess reserves. If that’s not on the table, then the functional monetary base has to be increased by swapping money with interest-bearing assets. If real interest rates have a floor of -2% due to no penalty interest, then it seems like the Fed may have to buy A LOT of assets to get to remove enough current assets to force new investments or consumption.

    I’m sure this all sounds incredibly sophomoric, but it does seem like the argument should be “the Fed must communicate NGDPLT and then charge penalty interest until the market agrees,” instead of the argument “if the Fed communicates NGDPLT, then magical pixie dust will make the market agree with it.” The concrete steppes such as penalty interest rates have been talked about a good bit, but it seems to take a backseat to expectations in most MM arguments.

  8. Gravatar of Matt Waters Matt Waters
    4. March 2013 at 16:58

    Of course the NGDP futures proposal gets around all of the objections by creating assets de novo that the Fed “buys” when NGDP is too low and “sells” when it’s too high. If NGDP grew at 2% when the Fed paid 4%, then there would be too much free money to the point where the difference is arbitraged. Too bad it just seems to come off as too weird to most people.

  9. Gravatar of Nathan Nathan
    4. March 2013 at 17:26

    Scott, I’m a bit confused about your assertion that Australia has “faster” NGDP growth/higher inflation. I acknowledge that you used relative terms and in those terms your statement is true, but it’s a little hard to see what’s undesirable about a 2.4% inflation rate.

  10. Gravatar of Michael Michael
    4. March 2013 at 17:49

    Matt Waters wrote:

    “I’m sure this all sounds incredibly sophomoric, but it does seem like the argument should be “the Fed must communicate NGDPLT and then charge penalty interest until the market agrees,” instead of the argument “if the Fed communicates NGDPLT, then magical pixie dust will make the market agree with it.” The concrete steppes such as penalty interest rates have been talked about a good bit, but it seems to take a backseat to expectations in most MM arguments.”

    How about the “Benjamin Cole announcement”:

    “Here is the NGDP level target. We are going to MONETIZE US TREASURY DEBT until we reach the target path.”

    Of course they would need a toned down version of that.

  11. Gravatar of Don Geddis Don Geddis
    4. March 2013 at 17:53

    Matt Waters: Your analysis sounds fine to me, so I don’t see why you’re so concerned about it. As we know from the concrete steppes, all the Fed requires is a sufficient credible threat. But it doesn’t have to actually carry out the threat; expectations do most of the work.

    So what is the specific threat? Unlimited QE. Continue to purchase Treasuries (or MBS) until NGDP expectations rise sufficiently.

    The only part of your comment that I’d disagree with is: “the Fed may have to buy A LOT of assets to get to remove enough current assets to force new investments or consumption.” No, the Fed only needs to threaten to “buy a LOT of assets”. Once the threat becomes obviously credible, the actual work of raising NGDP takes place through the monetary base, and changes in velocity. It is not required that the Fed actually buy enough assets to “force” economic behavior change directly.

  12. Gravatar of ssumner ssumner
    4. March 2013 at 18:15

    Matt, You are mixing up two very separate issues:

    1. What’s possible in theory.
    2. What’s possible in reality.

    Not just you, but many others start by talking about the theoretical possibility of a liquidity trap, and then seemlessly shift over to the real world and continue assuming that the theoretical possibility is relevant.

    I’ve addressed both 1 and 2 numerous times. In the real world the Fed needs to merely set an appropriate NGDP growth target, and peg the price of NGDP futures contracts via OMOs.

    In theory this might not work, as you say, or it would require the Fed to buy assets that they are currently not allowed to buy. In that case they have many other options, which I have discussed numerous times. But that possibility is not very interesting as it doesn’t apply to the world we actually live in.

    I would add that market monetarism doesn’t stop working in the world you describe, we simply use other tools. And no, the “other tools” don’t include fiscal policy.

    And regarding “pixie dust,” tell that to investors in the Japanese stock and forex markets.

  13. Gravatar of Steve Steve
    4. March 2013 at 19:09

    I agree, except that I worry that even 1 and 2 are incompatible if the ‘low’ in low inflation is too low. I worry that the welfare maximizing NGDP rate is between 5 and 7 percent rather than 3 to 5 percent. But I would accept any NGDPLT > 4 percent.

  14. Gravatar of Matt Waters Matt Waters
    5. March 2013 at 07:28

    I agree with all that. Like I said here:

    “If the Fed merely communicates its intention to NGDPLT without any firepower backing it up, it is at the mercy of the market. The market will increase NGDPLT if that’s what everybody else in the market is doing”

    Enhanced communications could be enough if it moves the market’s thoughts on what the rest of the market is doing. For NGDP growth in particular, markets are a Keynesian beauty contest where an individual actor will increase their spending if everybody else is increasing their spending.

    I’m not sure what the other loosening tools are other than IOR penalties and QE, but certainly penalty interest is the most straight-forward way (except for NGDP futures) to make sure NGDP goes up if the market does not find the Fed credible.

  15. Gravatar of Rademaker Rademaker
    5. March 2013 at 07:30

    The real trade off, in my opinion, is between consumer price inflation and asset price inflation. Central banking policy has been very successful at pushing the former form of inflation into the domain of the latter, but the excesses in the latter category have built up to such extremes that they can’t be upheld easily. My guess is we will see a reversal of that transfer: back from asset inflation to CPI inflation as the extreme overhang of debt gets put on the central bank’s balance sheet where it will be monetized away. Of course this process will dispell many illusions people hold about the effect and usefulness of monetary policy.

  16. Gravatar of ssumner ssumner
    5. March 2013 at 08:58

    Steve, 3% would fine for Japan, but I see your point.

    Matt, They could set a higher NGDP growth target. I’ve discussed the tradeoffs extensively in other posts.

    Rademaker, There is no tradeoff. Expansionary monetary policy raises both prices.

  17. Gravatar of Geoff Geoff
    6. March 2013 at 21:39

    Don Geddis:

    “@Geoff: You’ve provided no data in your example about whether money is in fact too tight or too loose, so obviously there is no valid answer to your final question.”

    Suppose a man was tied to a pole his whole life. Suppose his captor barely kept him alive. Suppose I said “This man would have a better life if he were let loose. His weak strength and poor health are due to the fact that his mobility is “too tight“.

    Now suppose someone who thought like you, who was focused on “data” only, challenged my argument and said:

    “You claim that his mobility is “too tight”, because relative to some ideal of him not being tied to the pole, he’d otherwise have greater mobility? OK, show me “the data” for this. Show me this “greater mobility” data. For all I see are his movements as they are with him tied to the pole. That’s the only data we have. What “looser mobility” are you talking about? There’s no data for such looser mobility, so how do you know his strength and health would be improved with this non-observed “looser mobility”?

    Now, going back to the topic at hand, the same principle applies. I cannot show you “the data” for a world that does not exist, but I say should not be ignored, namely, the dataset that would otherwise exist had there been a private competitive free market in money economy. That is the “ideal” standard by which I judge the “tightness” and “looseness” of money, the same way the standard by which I would judge the mobility of a man tied to a pole is the “ideal” of him not tied to the pole. We can’t observe this data of course, but that doesn’t mean we should ignore it.

    Asking me to show you historical data associated with the standard to which I define “tight” and “loose” money, when that standard is unobservable, would be like asking the man tied to the pole for data that shows what his life is like untied from the pole, so as to justify the argument that his health and strength would be improved if he were not tied to the pole.

    In short, the burden of proof is not on me to prove to you that a free market in money is the ideal standard. The burden of proof is on you to prove to me why introducing violence in what would otherwise be peaceful market competition, is justified.

    You also cannot consider “the data” either, and I won’t demand that you show me, because you would be just as unable to provide me with the counter-factual data that doesn’t exist, as I am with you.

    Lesson: Not all knowledge is, nor even in principle could be, grounded on observations only. Observations are actually not the ground for what we know about economics.

    “Depending on the state of the economy, either one could be wrong, or even both at the same time.”

    Only one can be right. The problem is what standard you are using to judge it. The standard I use doesn’t presuppose the validity of that which I am judging! The same cannot be said for you.

  18. Gravatar of Geoff Geoff
    6. March 2013 at 21:49

    Don Geddis:

    Or, I’ll approach this in another equally effective way:

    Recall, I asked:

    “Mr. Smith wants sub-4% NGDP growth, hence he believes the current 4% NGDP growth means money is too loose and inflation is too high.”

    “Mr. Jones wants 5% NGDP growth, hence he believes the current 4% NGDP growth means money is too tight and inflation is too low.”

    “Who is wrong about the stance of monetary policy, and why?”

    You said there is not enough data to give an answer.

    OK, suppose the “data” that you’re probably thinking of is the following: Unemployment has been 5% and annualized price inflation has been 10%.

    Who is wrong now?

    Using your approach, you’re going to have to insert an arbitrary standard for what rate of unemployment and price inflation “should be”, according to yet another arbitrary standard that would likely fall into some “natural” or “normal” trend assumption.

    But then suppose

    Mr. Smith wants 5% unemployment, hence he believes the current 5% unemployment means money is too loose and inflation is too high.

    Mr. Jones wants sub-5% unemployment, hence he believes the current 5% unemployment means money is too tight and inflation is too low.

    Who is wrong and why?

    I hope you will eventually realize that you’re going to be in a realm of subjective preference, and not traditional “scientific method, data gathering, theory, falsification, etc”.

  19. Gravatar of Geoff Geoff
    6. March 2013 at 21:50

    Sorry, meant to say

    Mr. Smith wants 6% unemployment, hence he believes the current 5% unemployment means money is too loose and inflation is too high.

Leave a Reply