What would happen if the Fed set a (positive) interest rate floor?

Caroline Baum has an interesting article on a new book written by DiMartino Booth, who worked at the Fed from 2006 to (I think) 2015:

The Fed regularly publishes a summary of economic conditions in the 12 federal reserve districts, but when real-world information contradicts the Fed’s econometric model, the model wins. DiMartino Booth provided Fisher with real-time information — not seasonally adjusted, to the consternation of the staff — gleaned from an array of market sources and data sets.

Fisher, with his background in business, finance and government, earned a reputation as a maverick inside the Fed. He dissented from the FOMC statement five times in 2008, twice in 2011 and twice again in 2014, in all cases favoring less monetary accommodation or earlier rate hikes than the consensus view.

Fisher retired at the end of 2014; DiMartino Booth followed shortly after, once she realized her real “mission” is to educate the public about the inner workings of the Fed. “Fed Up” succeeds in doing just that. . . .

One of her suggestions made me laugh out loud. The Fed should ship the Ph.D. economists back to academia and use the money saved to hire some crack bank supervisors at competitive salaries for the Fed’s “Sup & Reg departments,” traditionally a second-tier job at the Fed.

A few comments:

1.  In retrospect, it’s clear that all of the Fisher votes cited above were in error. Indeed in my view that’s not even debatable.  The fact that it is debated tells us that we need to reform Fed policy is such a way that clearly mistaken votes are no longer debatable.  Chad Reese and I have a letter just published in The Hill that discusses this issue.

2.  You generally don’t want to rely on non-seasonally adjusted data, which might signal a huge “boom” in December.

3.  Over the past three decades, the Fed has relied far more heavily on academic economists than in the past.  During this period, Fed policy has become vastly more stable than in the prior 70 years, whether you rely on inflation or NGDP growth as your indicator.

The author advocates greater diversity at the Fed: specifically, more staffers with actual business experience and fewer ivory-tower types. She would like to see an increased focus on systemic risk. And she wants Congress to release the Fed from its dual mandate — stable prices and maximum employment — so it can focus solely on price stability. . . .

Where I [Caroline Baum] would challenge DiMartino Booth is on her recommendation that the Fed normalize the overnight rate and pledge never to breach the 2% floor again so as not to punish savers.

I have not read her book and I may be misinterpreting the comment about a 2% floor.  But if the floor refers to interest rates, then I’d say the proposal is either absolutely horrible or mindbogglingly insane.  Let’s start with the best case, absolutely horrible:

1.  One way to make sure interest rates never again fell below 2% is to raise the inflation target to 10%.  That’s a horrible idea, and since it would probably punish savers I don’t think that’s what the quote refers to.

2.  If the inflation target is kept at 2%, then a 2% interest rate floor would be non-credible, because it would be impossible for the Fed to achieve.  But trying to achieve it could easily cause another Great Depression.  A mindbogglingly bad idea. This is why you don’t want non-PhDs making monetary policy.  Indeed, even the brief April to October 2008 2% interest rate floor proved to be disastrous.

PS.  Vaidas Urba sent me another post that comments on the same book.  Looks like a very good blog.

PPS.  The case for the Fed increasing its interest rate target is growing stronger by the day.  Some of the new data looks quite strong for both prices and output. Continued talk of fiscal stimulus is starting to look kind of silly.  For the first time since I started blogging, I see the monetary policy risks as being balanced, instead of skewed to the downside.

PPPS.   Many commenters are unaware of my current views on NGDP targeting.  I currently favor the “guardrails” approach, which is not susceptible to the market manipulation problem.  Nor would lack of trading be a problem.  My attempt to create a small NGDP prediction market is not to be confused with this policy proposal.  I have a new article in the Journal of Macroeconomics that explains my current views on a wide range of monetary policy rule issues.

PPPPS.  A zero percent interest rate floor would be less bad than 2%, as the Fed has the option of QE.

PPPPPS.  For those who don’t have access to the JM article, the basic idea is as follows.  The Fed sets a 4% NGDP target path, level targeting.  If the economy is currently on target, the Fed commits to take a short position against any NGDP futures trader going long at 5% NGDP growth, and the Fed takes a long position on any NGDP futures traders who go short at 3% NGDP growth.  Thus the Fed might be exposed to loses if the actual NGDP growth rate is outside the 3% to 5% guardrails.  The losses could be large if, ex ante, it’s clear to traders that NGDP growth will be far too high, or too low.

The Fed monitors the trades.  It still has 100% discretion over monetary policy, with the proviso that it be willing to commit to the NGDP positions described above. This is much like Bretton Woods or a gold standard (where they committed to exchange money for gold, instead of NGDP contracts), and no more susceptible to market manipulation than those regimes.  Indeed less so, as the “band” is (effectively) wider than under the gold standard. The Fed can and should ignore a single large trader, who might be engaged in market manipulation.  If it sees lots of traders all going long or short, and if the Fed’s potential losses become too large, it may want to take corrective action.

Even if I were wrong about manipulation, competition among potential manipulators would keep expected NGDP growth in the 3% to 5% range (as the second manipulator could offset the first, and earn larger profits by going the opposite direction).

Here’s another metaphor.  The 3% and 5% guardrails serve the same purpose as the beeping sound when trucks are backing up, and get too close to hitting something.

The basic idea here is to allow me, Scott Sumner, to get filthy rich if the Fed screws up the way they did in 2008.  Since I’m a fatalist who never expects to get filthy rich, I would not expect the Fed to screw up under my proposed guardrails regime. I hope I’m wrong!!




26 Responses to “What would happen if the Fed set a (positive) interest rate floor?”

  1. Gravatar of anon anon
    15. February 2017 at 12:09

    Can you provide a non-paywalled link to your article in the Journal of Macroeconomics?

  2. Gravatar of Keenan Keenan
    15. February 2017 at 12:45


    Thanks for addressing my concern about market manipulation. I’m sure you’ve thought about it. Unfortunately, I’m not going to pay $35.95 for this article. Could you quickly summarize how this NGDP market won’t be able to traded against other instruments? Basically anyone who trades this NGDP futures market shouldn’t legally be allowed to trade anything else. Otherwise there will be a lot of opportunity to use this market to make money trading other markets, not informing useful policy.

    I’m a professional trader, and I am extremely nervous about this idea of NGDP futures market.

  3. Gravatar of Matthew Waters Matthew Waters
    15. February 2017 at 13:08

    Isn’t IOER the interest rate floor? The Fed could create a 10% floor with 10% IOER. It wouldn’t work out too well, with savers probably doing worse overall with the stock crash and bond defaults. But interest rates wouldn’t be below 10%.

  4. Gravatar of ssumner ssumner
    15. February 2017 at 13:19

    anon and Keenan, I’m not sure about copyright issues, so I added a footnote at the end of the post, hopefully that explains the idea.

    Matthew, Yes, IOER is an interest rate floor, indeed a fed funds target could be viewed as a sort of floor. But these are adjusted as the economy evolves over time. I read the proposal as advocating a 2% floor regardless of the macro conditions. That’s what I responded to.

  5. Gravatar of Eliezer Yudkowsky Eliezer Yudkowsky
    15. February 2017 at 14:18

    What I incorrectly deduced this blog post would be, after seeing the title:

    “This of course is a trick question. There’s no such thing as setting an interest rate floor. The Fed can only declare a policy it will implement if interest rates go under a floor. If that action is to destroy money while interest rates stay under the floor, it will potentially cause the next Great Depression. If that policy is to go on printing money while interest rates stay under the floor, and they print enough money, nominal interest rates will rise above the floor as soon as markets realize they’re serious. Unfortunately we have a Fed and a Congress that would almost certainly try to follow the first policy.”

  6. Gravatar of Keenan Keenan
    15. February 2017 at 14:30

    Thanks Scott, sorry to not buy your paper.

    This seems good. I just hope that the Fed keeping the NGDP market between 3% and 5% would spillover to making the population think NGDP growth will be between 3% and 5%.

    So basically, to avoid ’08, if someone with conviction (i.e. Scott Sumner) thinks NGDP growth is going to be 1%, they will sell as many futures as possible, the Fed will compile a big long futures position @ 3% and think “maybe we’re missing something here guys”, and look into it a little more.

  7. Gravatar of Jerry Brown Jerry Brown
    15. February 2017 at 15:19

    Professor, you mention a 4% NGDP level target in one of the postscripts. I have often wondered what the differences would be with different NGDP level target rates for the Fed. I mean, I think I would favor a higher target rate as you might have noticed from my comments over the years. Because I am mostly labor orientated and think a higher target would tend to favor full employment and therefore workers, at the expense of an increased risk of higher inflation. Is that somewhat correct in your view? And I realize it is a lot to ask, but could you do a post explaining what you see as the effects of different target levels?

  8. Gravatar of Ray Lopez Ray Lopez
    15. February 2017 at 16:36

    Professor Sumner:

    As the trader Keenan says, this NGDPLT ‘crowd source’ policy will force the Fed to bend to the will of the people if it finds it’s taking a position that seems to be against what the market is saying is correct. In that case, if it’s in the interest of the Fed to ‘cut their losses’, how is this any different from today’s Fed and ‘discretionary policy’?

    Let me be concrete (with steps): step 1: The Fed sets a 4% NGDP target path, level targeting.  If the economy is currently on target, the Fed commits to take a short position against any NGDP futures trader going long at 5% NGDP growth, and the Fed takes a long position on any NGDP futures traders who go short at 3% NGDP growth. step 2: the economy is NOT on track, and the Fed finds it is on the wrong side of many bets, that are saying (contrary to the Fed’s 4% target) that NGDP will be at 0%/yr. Hence the Fed sees lots of traders all going short, and the Fed’s potential losses become too large, so the Fed wants to take corrective action. What is the Fed to do? Pls explain in concrete steps what corrective action the Fed is to take. In another post a year ago you implied the Fed will ‘double down’ on its 4% NGDP target, and simply ‘print money’ until it hits that target. Is that still your position today? step 3: supposed the Fed is wrong, and ex ante the losses by the Fed to traders like Keenan are on the order of, say, $4 trillion USD. Who is footing that bill? The US taxpayer?

    I look forward to your answers. So does Capital Hill, believe me they are watching you. Now is the time to put their fears that you are the Pied Piper of ruinous hyperinflation to rest.

  9. Gravatar of Tommy Dorestt Tommy Dorestt
    15. February 2017 at 18:16

    So Trump can fill three open board seats and nominate a new Chair next year. Scott, I fear we are going to be going over the guardrails in one way or another, especially if nitwits like this DiMartini dunce are what the future of a Trumpian FOMC look like.

  10. Gravatar of ssumner ssumner
    15. February 2017 at 18:35

    Eliezer, They could hold rates at 2% for a considerable period, but only at the expense of losing all control over the price level.

    Keenan, You said:

    “This seems good. I just hope that the Fed keeping the NGDP market between 3% and 5% would spillover to making the population think NGDP growth will be between 3% and 5%.”

    Yes, but I’m not really concerned what “the population” thinks, I want to insure that monetary policy is consistent with investors expecting 3% to 5% NGDP growth. In late 2008, growth was obviously going to be much less than that, and any idiot could have become a millionaire. I don’t expect that outcome to ever happen under the guardrails system, because the Fed would not allow it.

    And no worries about the paper, I get zero royalties anyway.

    Jerry, I don’t think a 5% target would help labor any more than a 4% target. But it’s possible that 4% would be better than 1%, due to wages being sticky downwards. The problem with faster NGDP growth is that it leads to higher taxes on capital, which hurts growth.

    Ray, It doesn’t force them to do anything, if they think the market is wrong. If they think the market is right, they should follow its advice.

    Tommy, I don’t think Presidents have much influence on the Fed.

  11. Gravatar of Do I wanna bet Do I wanna bet
    15. February 2017 at 18:46

    I am trying to understand the next step in Keenan’s example after the Fed says:
    “maybe we’re missing something here guys”, and look into it a little more.

    I assume the Fed then eases policy in response to this new information using the same mix of tools they use now, buy bonds reduced or negative IOR, etc. And I assume that would work and NGDP would grow by more than 3%, maybe right around their target at 4%. Which is the whole goal of this system and a great thing for everybody…except Keenan, who bet that this would not happen. But Keenan, being clever, saw this problem ahead of time and never placed a bet in the first place.

    In sports betting if the Las Vegas “forcast” is wrong people place bets on one side and the spread adjusts, but my bets against them do not make the Lakers a better team. If the Lakers a -2 for a game and a fair line is -10 I rush to place a bet before other people realize and the spread adjusts. If Fed policy was too tight under the “guardrails” policy and I was the first to notice I would not bet at all figuring if I do i would help save the country from an NGDP shortfall, but I would lose out personally on the transaction.

    I guess you could argue that at equilibrium this makes the system less effective but not completely ineffective, but it is not clear to me.

  12. Gravatar of ssumner ssumner
    15. February 2017 at 19:36

    Do, That’s true to some extent in many other markets. If you buy one year forward copper futures, that encourages copper companies to produce more, which drives down the price of copper.

    Fortunately, any single trader has such a tiny impact on the aggregate market that this would not prevent trading from occurring.

  13. Gravatar of Jerry Brown Jerry Brown
    15. February 2017 at 19:41

    Thanks Scott. Isn’t it a bit true that wages are sticky upwards also, but less so than downwards? I mean for a company with a lot of employees in a particular type job? And I am sure I am missing something, but why does NGDP growth lead to higher taxes on capital any more than leading to higher taxes on labor and consumption? Capital is mostly property employed in production in my understanding. Higher NGDP growth would tend to make that property more valuable? Given the previous investment required to purchase (produce) it? I don’t consider idle money to be capital, although that could be wrong.

  14. Gravatar of Benjamin Cole Benjamin Cole
    15. February 2017 at 20:43

    Excellent blogging.

    The “low interest rates hurts savers” argument is the hallmark of a weak mind devoted to tight money.

    Savers do not buy bonds? Buying and developing property is not saving (investing)?

    Beyond that, see this:


    Total savings deposits in US depositary institutions were about $4 trillion in 2008. Now about $8.6 trillion.

    Savers are flooding savings accounts with money. They have a funny way of showing they are being “hurt.”

    Side note: I hope Market Monetarists can be convinced to be a little more aggressive than a 4% NGDPLT. How about 5% to 6%.

    Look, we want voters to choose free enterprise and sane, civil candidates.

    So, “tight” labor markets and cheap housing are good policies.

    Another percent or two of inflation is well worth it to obtain a voting population that embraces a free market system.

    On the other hand, if the way to “fight inflation” is to periodically unemploy millions of Americans and never less than one in 20 who want a job, then the voting population may well choose a Trump. Maybe they should.

    Indeed, I think Bernie Sanders could have beaten Trump.

    Why not? Housing is way too expensive, jobs are scarce, health insurance through the roof, war-spending chronic, and the establishment says illegal immigration is not something that can be fixed.

    Well, lah-de-dah.

    So vote for Hillary or Romney?

    BTW, Pudzer out as DOL secy. He had illegal maids.

    I guess the entire Cabinets of the last few administrations, and the entire present Trump entourage, has illegal butlers, maids.

    Zoe Baird got bounced for having an illegal maid…and an illegal chauffeur.

    Do you suppose a wall will get built?

  15. Gravatar of Julius Probst Julius Probst
    16. February 2017 at 02:17

    Scott, this is a fascinating new book that describes all monetary policy actions undertaken by the BOJ since the middle of the 1990s, including the most recent policies with Quantitative and Qualitative Easing and Yield Curve Control. It’s written by a former board member of the BOJ.
    I started to read the first two chapters and it’s a good read.
    He describes how the “lack of boldness” largely contributed to the economy being stuck in a low-inflation environment for more than 2 decades.


  16. Gravatar of Randomize Randomize
    16. February 2017 at 07:58

    Dr. Sumner,

    As an alternative, what do you think of the treasury issuing NGDP-adjusted bonds and the Fed targeting the “spreads” with regular bonds? I’d like to hear your thoughts.

  17. Gravatar of Ray Lopez Ray Lopez
    16. February 2017 at 08:34

    @Sumner: “Ray, It doesn’t force them [the Fed] to do anything, if they think the market is wrong. If they think the market is right, they should follow its advice.” – please explain more. If the Fed thinks the market is wrong, and they continue to hold rates steady at the 4% NGDP target path, level targeting (or try to), isn’t the Fed taking a loss, when the rest of the market has the opposite bet to the Fed? When these futures contracts settle, doesn’t the Fed have to pay money to traders who took the opposite side of the bet to the Fed? Please explain. I really want your scheme to work (well, honestly I think it’s irrelevant but it would be cool to see your scheme in practice) but I don’t understand the mechanics of it. I think others feel the same way. Question to anybody reading this: please answer the above, is the Fed taking a loss or not? Dtoh?

  18. Gravatar of Jerry Brown Jerry Brown
    16. February 2017 at 09:15

    Ray, it doesn’t matter to the Fed if they take a loss. It is hard for me to comprehend how an agency that can create money at will can really be understood to take a ‘loss’ in terms of that money. But I agree that I would like to see Scott Sumner write more about NGDPLT. I always find that educational.

  19. Gravatar of rayward rayward
    16. February 2017 at 09:27

    How I learned to stop worrying and love deflation. That seems to be the WSJ’s new motto. Sometimes I wonder if it’s meant as satire. Maybe that’s why Sumner questions Dr. Strangelove’s sanity.

  20. Gravatar of Doug M Doug M
    16. February 2017 at 13:57

    punishing savers.

    I am aware that there is a camp out there that says that very low T-bill rates hurts old people. People who have diligently saved their pennies over the years, and parked their savings in low-risk investments with the hope that interest off of this nest egg will last them through retirement. Any cut in Fed funds which most directly influences short-term, low risk investment yields, is pure hatred for the elderly.

    This is a view of saving and investment that is at lest 2 years out of date. A saver who wants to minimize risk should match the duration of their investments with the time over which they expect to draw on that cash flow. Yes longer term bonds have more price risk, but paradoxically more price risk may be less total risk.

    If you were truly wedded to this minimum interest rate to protect savers, I would say, that you would have to put a cap of $ 100,000 maximum assets with this guaranteed floor.

  21. Gravatar of Ray Lopez Ray Lopez
    16. February 2017 at 16:41

    @Jerry Brown – “Ray, it doesn’t matter to the Fed if they take a loss” – so you have no problem with the Fed’s balance sheet going from $1T before 2008 to $4T today? I hear you, and I agree Sumner should blog on this, but the Fed itself says it is concerned and will someday unwind this balance sheet. I think there’s a disconnect between what Sumner believes and what the Fed (and most other people) believe.

  22. Gravatar of Jerry Brown Jerry Brown
    16. February 2017 at 21:06

    Ray, I have some concerns about the distributional aspects involved with the Fed balance sheet increasing so much. Because the reason it has increased so much is that it was buying financial assets from those who already had those assets. And that purchasing by the Fed would tend to inflate the value of those assets, as far as I can tell. And I have no confidence that those who benefited from that asset price increase deserved it. But other than that- yes, I have no concerns about the size of the Fed balance sheet.

  23. Gravatar of ssumner ssumner
    17. February 2017 at 17:06

    Jerry, Yes, a little bit sticky upwards as well.

    Thanks Julius.

    Randomize, Not a fan of that idea. The same problem TIPS spreads have—time varying risk premia.

  24. Gravatar of Petja Ylitalo Petja Ylitalo
    18. February 2017 at 10:49

    So you are suggesting a system where traders bet against FED on NGDP growth, doesn´t it seem a bit unprofitable to bet against FED on FED policy?
    Though i guess it could be fixed with traders betting on “NGDP will be under 3% or FED will lower interest rates” or something similar.

  25. Gravatar of ssumner ssumner
    19. February 2017 at 14:53

    Petja, If it’s unprofitable that’s good, isn’t it.

  26. Gravatar of Petja Ylitalo Petja Ylitalo
    23. February 2017 at 02:40

    If it doesn´t stop people from betting, since if people stop betting you don´t get data anymore. Though i guess that there will be some willing people as long as the odds are good enough, so there´s a good chance that this won´t be that bad of a problem, just something to take into account.

Leave a Reply