I am currently at a Fed conference called “FedListens”, which is evaluating options for improving monetary policy. One focus of the conference is the question of how best to address the zero bound problem. Here I’ll present a few initial impressions, based on the first paper (by Janice Eberly, James Stock and Jonathan Wright) and the following discussion. I should warn you that I may be misinterpreting the paper, but FWIW I’ll give you my impression of where I think it goes off course.

The authors argue that a negative 5% nominal interest rate would have been appropriate during the Great Recession, and then look at various counterfactual strategies for improving monetary policy, given the zero bound constraint. These counterfactuals include asset purchases, forward guidance, and a higher inflation target (3% or 4%), among others. The goal is to get the stance of policy closer to the negative 5% fed funds that would have been appropriate, if not constrained by the zero bound.

In my view, the basic mistake is to (implicitly?) assume that the need for a negative 5% fed funds rate was caused by an exogenous negative shock (something like the financial crisis, for instance), rather than excessively tight money during 2008, which sharply depressed NGDP growth expectations. If I’m right, then they underestimate the benefits of structural changes in monetary policy that lead to faster NGDP growth, and hence a milder recession. One of these benefits is that the real Wicksellian equilibrium rate would have fallen much, much less sharply.

Thus consider the counterfactual of a 6% inflation target. In a simple model where this policy counterfactual had no impact on the exogenous shock that reduced the Wicksellian equilibrium real rate, the Fed would still have had to reduce interest rates to negative 1% to achieve an appropriate real rate during the Great Recession. I find that assumption to be exceedingly implausible. In my view, even a 4% inflation target, and certainly a 6% target, would have prevented the Fed from ever hitting the zero bound. The extra NGDP growth expectations (both inflation and RGDP) caused by a higher inflation target would have prevented the equilibrium rate from falling anywhere near zero (see Australia), and thus none of the unconventional policy options would have been needed.

I believe this critique is related to John Taylor’s criticism of the paper, which was that they had not based their policy counterfactuals on a structural model.

The profession as a whole tends to have a Keynesian approach to these issues. The economy is inherently unstable due to “shocks”, and the Fed is a sort of fireman that comes to the rescue, by trying to depress rates to as close to the equilibrium rate as possible. In the monetarist framework (which is best explained in the work of Robert Hetzel), the Fed is more like an arsonist, creating nominal instability through monetary policy errors. Because wages and prices are sticky, this nominal instability creates labor market instability, which depresses investment and hence the equilibrium interest rate. A more effective monetary policy helps mostly by avoiding causing nominal shocks, not by reacting to instability in the private economy.

I don’t favor a 4% or 6% inflation target, but I believe a 4% or 5% NGDPLT would have had a similar stabilizing effect, which is not picked up in the paper presented.

Another way of putting things is that while most economists want to reduce interest rates to the Wicksellian equilibrium rate during a recession, I want to use asset purchases to raise the equilibrium interest rate.

Alternatively, this conference needs to take NeoFisherism more seriously. Although NeoFisherism as a stand alone theory is wrong, it’s a very useful critique of conventional macro. It reminds us that the ultimate goal should not be to get the policy rate as low as possible, but to create conditions where the equilibrium interest rate is much higher than it was in 2009. We need a regime that reduces the need to cut rates to below zero.



29 Responses to “FedListens”

  1. Gravatar of Brian Donohue Brian Donohue
    4. June 2019 at 08:56

    The evidence of the last decade strongly suggests that QE was effective at increasing (or at least stabilizing) long-term rates, pretty much the opposite of the conventional wisdom.

  2. Gravatar of Christian List Christian List
    4. June 2019 at 08:57

    Here are all the papers, I guess:


  3. Gravatar of Matthew Waters Matthew Waters
    4. June 2019 at 09:18

    The best evidence for an exogenous shock is Gary Gorton’s paper “Panic of 2007.” It gives a soup-to-nuts overview of subprime mortgages up to the ultimate holder.

    An alternate universe of only Treasury purchases in 2008 would have moved the crisis up to March with Bear’s failure. Lehman and AIG both had to rely on overnight funding and would have quickly also failed. In the alternate universe, the Treasury would not have backed money market funds and the Fed would not have purchased private ABS.

    The trillions in run-prone shadow banking liabilities differentiated the 2007-08 dollar from the Australian dollar. Even with a flood of new reserves for Treasuries, the shadow banks did not have adequate Treasuries for their immense outflow. The outflows wired money to regular banks, such as JP Morgan, but the regular banks would take months to set up and approve irregular financing of the shadow banks.

    Now, the Fed should have *also* done massive Treasury purchases. The Fed wrongly sterilized new reserves in 2008 and had little “one-way” asset purchases. However, Treasury or Agency MBS purchases alone would not have been sufficient given a Bear/Lehman/AIG failure. Financial regulation is utterly important to reduce runnable liabilities without government backing, as Morgan Ricks as proposed. The financial regulation part is a big piece of both Australia and 50s/60s US monetary stability.

  4. Gravatar of Alex S. Alex S.
    4. June 2019 at 09:34

    The Eberly, Stock, and Wright paper distinguishes “slope” and “level” policies. I think the terms “credit policy” and “monetary policy already basically describe this, and should not be so easily forgotten.

    I recall a 2009 Bernanke speech stating that QE is credit policy not monetary policy.

    Does anyone know if there are studies that look at QE’s effect on market-based measures of inflation expectations?

  5. Gravatar of rayward rayward
    4. June 2019 at 09:50

    In the financial crisis, with asset prices plummeting, the Fed pumped liquidity into the financial system to stop the fall. And it worked. But try though the Fed might, including with below zero interest rates, owners of capital wouldn’t invest in productive capital. And wouldn’t. And wouldn’t. And wouldn’t. And even despite a trillion dollar fiscal stimulus (the Trump corporate tax cut), still won’t – much. How does the Fed stimulate investment if the owners of capital prefer safe investments or take their chances on rising asset prices in the financial markets? Sumner sees the world as a good market monetarist would see the world, by creating higher expectations of economic growth by targeting NGDP. Would owners of capital expect higher economic growth and therefore invest in productive capital just because the Fed targets it? Maybe. Roger Farmer, sort of a market economist, would go all in for rising asset prices as the key to prosperity by committing the Fed to buy equities (equities!) if the NGDP target isn’t achieved. Implicit in Farmer’s approach is that owner’s of capital must not only expect higher rates of return, but be guaranteed it with rising asset prices. I’d buy equities if the Fed was committed to keeping their prices rising. Is that the world of today?

  6. Gravatar of ssumner ssumner
    4. June 2019 at 10:30

    Everyone, It seems to me that people are taking the size of the crisis as a given, and not contemplating how big it would have been in a counterfactual where the Fed did enough to keep NGDP expectations growing at 5%/year. In that world, pressure on the banks is far smaller, although still elevated. Real growth is far higher, although still lower than trend. Inflation is somewhat elevated, there is a bit of stagflation.

    Don’t make the mistake of taking “the crisis” as a given. It was an ongoing problem that got continually worse as NGDP expectations fell throughout 2008, indeed BECAUSE NGDP expectations fell throughout 2008.

    The Great Recession was partly a real shock that might have pushed unemployment up to 6% or 6.5%, and also a secondary nominal shock caused by tighter than desired money which pushed it up to 10%.

  7. Gravatar of James Alexander James Alexander
    4. June 2019 at 10:33

    Do you get a chance to point out the obvious? FedFirelights not FedFirefights. Or are you too small a minority to get a chance to speak?

  8. Gravatar of Matthew Waters Matthew Waters
    4. June 2019 at 11:51

    Sorry, I know this is long and I have tried to avoid long comments. But I believe the institutional details are important. It’s sad that the true underlying details of the crisis are not more widely known.

    Basically, a Fed with helicopter money could have certainly offset a run, but I am certain a run would have still happened with constant NGDP growth.


    The expectations argument paints over details with such a broad brush that it’s nearly unfalsifiable.

    The institutional details matter, which is why I highly recommend Gary Gorton’s “Panic of 2007” as well as “Sizing Up Repo” by Krishnamurthy.

    The Reserve Primary Fund lawsuits, when State Street suspended the Funds’ overdraw capability the morning after Lehman bankruptcy, are also instructive. The Fund had met $1.00 NAV requests every thirty minutes. The nature of the panic was that new reserves from OMOs would have to flow to buying the Reserve Primary Fund’s assets *in a matter of hours*.

    Meanwhile, hedge funds had prime brokerage accounts at Lehman’s subsidiary in London which were essentially bank accounts. Withdrawals were forced on the hedge funds and hedge funds sold prime brokerage claims for as little as 10 cents on the dollars. After 10 years, the claims ultimately paid over 100 cents on the dollars since British bankruptcy laws pay interest.

    The crisis traces all the way up to fiduciaries and a failure of their regulation:

    1. *All* mutual funds can lend up to 33% of their assets. Very few people realized that their index fund in their 401k lends out up to 33%. The cash received as collateral is then reinvested in MMFs, foreign bank accounts and CP/Repos of broker-dealers or foreign banks. Of course the ultimate mutual fund owner has no idea about these collateral reinvestment arrangements. I don’t feel like an idiot with regard to financial markets and I didn’t realize their breadth until a year or so ago.
    2. Pension and insurance companies also relied on the opaqueness of security lending. AIG’s downfall was directly tied with its security lending rather than merely AIGFP. Without the Fed’s bailout, the Texas and New York insurance subsidiaries would have gone into receivership. Ultimately the life insurance and annuities of these subsidiaries had a backstop of guaranty funds, which are then funded by assessments on healthy insurers.
    3. As fiduciaries for non-accredited investors, publicly traded companies only disclosed assets in the shadow banking system under “cash equivalents.” The security laws for registered securities are set up to disclose material issues and the security laws clearly failed on disclosing material issues.

    Ultimately, relatively little of very low-capital and low-liquidity shadow banking system had truly accredited investors investing their own money. IMO, a run on this system was absolutely inevitable. If we limit monetary policy to:

    1. Purchases of Treasuries and Agency MBS.
    2. Traditional, relatively conservative discount window lending.
    3. Helicopter money in the form of tax cuts and Treasuries issued directly to the Fed.

    Then IMO, monetary policy would have been successful at hitting *nominal* GDP, but there would have been significant costs to *real* GDP through the shadow banking system declaring bankruptcy and then lawyers sorting it all out. The whole system was far too opaque for even helicopter money to sustain the shadow banks’ outflows.

    To conclude, it’s certainly not wrong to wish for “hell or high water” NGDP targeting. But the real-world institutional and information constraints would not have stopped the run. The inevitable run would have merely been offset.

  9. Gravatar of Christian List Christian List
    4. June 2019 at 13:26

    I’m positively surprised by this mainstream (?) paper. At least they get two things: “Stronger sooner“, as well as a higher inflation target. I think you can’t expect much more at the moment.

    Your firefighters vs. arsonist metaphors sound a bit nitpicky (or pedantic) to me. Of course it would be a very big deal if you were 100% right, but I now understand better after this paper, why the mainstream does not want to buy into it right now. Who is 100% right, ever?

    Let them do “stronger sooner” and a higher inflation target – and the end result should be the same, for the time being.

  10. Gravatar of Benjamin Cole Benjamin Cole
    4. June 2019 at 15:59

    Excellent blogging, and Scott Sumner appears to take the stance that the Fed is not trustworthy, in terms of macroeconomic policy making. Often too slow, often faint-hearted and sometimes wrong.

    I will posit that in the current economies of developed nations the trend is towards disinflation or deflation, not inflation.

    A thought experiment given the above two paragraphs: should the US have an automatic holiday on Social Security taxes whenever unemployment goes above 4%?

    Remember, Sumner has more or less said the Fed is often not an effective policy instrument.

  11. Gravatar of Benjamin Cole Benjamin Cole
    4. June 2019 at 16:19

    OT but in the ballpark:


    There is an interesting chart in the above post. Large fractions of adults who are already employed want to work more, that is they want to work more hours.

    This is a remarkable in and of itself, as Americans already work several hundred hours a year more than, say, Germans or French.

    But this is another indicator of how squishy and malleable the US labor force is. Eyeballing the stats, it looks like there’s 40 to 50 million Americans who would like to work more hours.

    My guess is the right solution is to keep the labor force increasingly employed so that job switching and realignment of workers to positions can be activated.

  12. Gravatar of ssumner ssumner
    4. June 2019 at 17:30

    James, It would be as if I said the Earth was flat.

    Matthew, It’s unlikely that Lehman would have gone bankrupt if the Fed had been doing inflation targeting.

    Christian, The Fed is not going to adopt a higher inflation target.

  13. Gravatar of Christian List Christian List
    4. June 2019 at 17:57

    Scott, Well then let’s hope they learnt anything from 2008. They could at least hit their actual target when it’s most needed, right? But Bernanke was in charge back then, and yet still those mistakes. Now Powell is in charge, so I assume “stronger sooner” will actually mean “lesser later”. I’m so curious how this ride will eventually end and when.

  14. Gravatar of Benjamin Cole Benjamin Cole
    5. June 2019 at 02:48

    Idle pondering: Is the global slump in equities the last few months (until yesterday) a result of the much headlined “trade wars”—or really from too-tight money?

    The math for the “trade war” just isn’t there…Sino imports and exports are not much, compared to US GDP. Same on the other side of the fence, in China.

    The US is still sourcing and selling from and to the rest of the world.

    The ECB seems unwilling to do more, and the Bank of Japan is similarly stuck, despite inflation in both regions well below target. The People’s Bank of China was worried about credit growth, but seems to have backed off (the commies showed up at HQ).

    The central banks are doing their favorite thing, which is to suffocate economic growth while delivering sanctimonious sermonettes.

  15. Gravatar of Mark Mark
    5. June 2019 at 03:54

    Ben, the markets were doing quite well from January to April, then abruptly reversed and went into a downturn in May when the US-China trade war heated up. That seems pretty strong evidence that the trade war is causing the downturn.

    There’s a lot more to the trade war than tariffs. There is a lot of non-tariff economic warfare going on, such as the US sanctions on Huawei. This stuff, as well as the tariffs, will have a lot of hard-to-predict indirect effects such as slowing global innovation and causing companies to make decisions that are less optimal from an economic point of view to reduce political risk.

  16. Gravatar of Matthew Waters Matthew Waters
    5. June 2019 at 08:45

    NGDP growth up to March 2008 was still above the 2001 and 1991 recessions. Bear needed the ad hoc Maiden Lane vehicle to not declare bankruptcy. Maiden Lane made the Fed’s money back, but it was out of line with the discount window collateral/margin previously published.

    When Lehman declared bankruptcy, NGDP YoY growth was still at levels equal to trough of 2001 and 1991 recessions. The plummet in NGDP happened after Lehman’s bankruptcy. If Lehman was not robust enough to handle 2001 NGDP growth rate, I can’t feel especially sorry for them.

    Finally, unlike 1991 and 2001, the Lehman bankruptcy brought a generalized run on money market funds and prime brokerage accounts. The run made the natural rate heavily negative. The run-prone market structure was exogenous to the Fed.

  17. Gravatar of Michael Sandifer Michael Sandifer
    5. June 2019 at 08:50

    Matthew Waters,

    If you’re correct, is it necessarily bad to let the firms fail that fail even when NGDP growth stays on trend?

    Also, it’s hard to believe that allowing nominal growth to fall as much as it did didn’t greatly contribute to the financial crisis. Surely, it made it worse.

  18. Gravatar of ssumner ssumner
    5. June 2019 at 19:04

    Matthew, You said:

    “If Lehman was not robust enough to handle 2001 NGDP growth rate, I can’t feel especially sorry for them.”

    I completely agree, but that has no bearing on what we’ve been discussing.

  19. Gravatar of Benjamin Cole Benjamin Cole
    6. June 2019 at 02:55

    Well, with oil prices falling, the Fed can again say the low inflations rates (that have persisted for more than 10 years) are again “transitory.”

    Note to Mark:

    I do not contest the timing of the Wall Street pullback. I contest the analysis of why.

    A lot of people blamed 2008 on something other that the Fed being too tight. It sure seemed tied to a property collapse.

    The math on the trade tariffs is…well, not serious. So we see a delay in the 5G rollout in the US—-that matters?

    People keep saying the Fed and the Bank of Japan and the ECB are “easy.” Yeah, that is why there is no inflation in Japan, almost none in Europe and the US is well below target.

    Why are global inflation rates so low for so long? Why do people pay for the right to own a German bund?

    Come on Mark, money is tighter than the Sphinx’ sphincter.

  20. Gravatar of Bob Bob
    6. June 2019 at 06:48

    First of all, your summary of the situation at the time is fantastic. Definitely not too long. And I agree with you that a run was probably inevitable. My preference has always been receivership, wipe out the shareholders, recapitalize, and IPO it. The bailouts were toxic for so many reasons.
    That said, I think Scott is right that most of the general public’s suffering (10% unemployment) would have been avoided by maintaining NGDP.
    And it’s hard not to think our politics would be better now if we avoided some of the toxic bailouts, and limited the general public’s suffering in the form of high unemployment and distressed local governments.

  21. Gravatar of Matthew Waters Matthew Waters
    6. June 2019 at 10:57

    I mean, of course there are exogenous swings in the natural rate. The Fed funds rate decreased already starting from 2007. The start of the panic, in August 2007, was a primary motivator. The Fed also started extraordinary measures in August 2007.

    Bear’s near bankruptcy happened very suddenly. As Lehman depended on run-prone funding (repo and prime brokerage), I strongly argue that Lehman would have also faced too many outflows. Remember that Lehman’s financial position was worse than they publicly disclosed.

    The Reserve Primary Fund also had Bear assets and Bear’s failures would have provoked a similar run, followed by a suspension of withdrawals.

    IMO, Bear’s failure would have necessitated zero rates and significant Treasury purchases to meet NGDP target. The purchases should be done through an open-ended tender offer at zero rates up to some maturity.

    For institutional or information asymmetry reasons, the new money may not have moved to funding outflows from shadow banks. The Gary Gorton Panic of 2007 paper gives the extreme information asymmetry involved with “toxic assets.”

    Even though the bankruptcy process itself destroys a lot of value, perhaps the RGDP costs of numerous bankruptcies would have been warranted for long-term RGDP.

  22. Gravatar of Benjamin Cole Benjamin Cole
    7. June 2019 at 01:54

    Wages remain a drag on the Fed’s putative 2% average inflation target.

    “WASHINGTON (Reuters) – U.S. unit labor costs were weaker than initially thought in the first quarter and costs declined in the prior period, suggesting inflation could remain moderate for a while.

    The Labor Department said on Thursday unit labor costs, the price of labor per single unit of output, dropped at a 1.6% annualized rate, instead of falling at a 0.9% as reported last month. Data for the fourth quarter was revised to show unit labor costs falling at a 0.4% rate in the October-December period, rather than increasing at a 2.5% pace as previously reported.”



    Back-to-back quarters of declining unit labor costs, along with a 10-year-run of unit labor costs increasing around 1% a year.

    There is a source of inflation in the US economy—housing costs along the West Coast, NYC and Boston. Not due to monetary policy and not due to wages—-due to property zoning.

    But…in reading Fed literature, there is abundant squeamish hysteria regarding “low” unemployment rates.

    Can orthodox macroeconomics adjust to the new reality? Will textbooks accommodate the facts on the ground?

  23. Gravatar of Ram Ram
    7. June 2019 at 06:10

    This observation seems particularly relevant at the moment. It seems the market thinks it more likely than not that we will have at least 1 25bp cut by July, 2 by September, and 3 by December. At the same time, the 5-year TIPS spread suggests expected inflation is ~50bp shy of the 2% target. That seems to suggest that the market believes either a 75bp cut should happen well before December, more than 75bp should be cut by December, or both, in order to hit the 2% target. It’s entirely possible the Fed will reduce the FFR target before the end of the year, but fall short of what the market has priced in, and thereby cause a recession. But since it will co-occur with a rate cut, the world may view what’s happening as an “exogenous” shock, rather than the Fed predictably failing to deliver the appropriate level of AD.

  24. Gravatar of Michael Sandifer Michael Sandifer
    7. June 2019 at 11:19


    Off-topic, in case you didn’t see it:


    The Trump Treasury has accused Germany and Italy of currency manipulation, apparently unaware they aren’t sovereign currency issuers.

  25. Gravatar of ssumner ssumner
    7. June 2019 at 15:51

    Ram, Good comment.

    Michael, I should do a post on that.

  26. Gravatar of Benjamin Cole Benjamin Cole
    7. June 2019 at 16:14

    The latest jobs report was hardly inspiring.

  27. Gravatar of Tilting at M3 Tilting at M3
    16. June 2019 at 09:56

    They’re ignoring the Lucas critique. This is literally a replay of a conference Stanley Fischer held in 1980. Keynesians don’t learn. From the best paper Lucas ever wrote:

    I take the purpose of this session to be to elicit views on economic policy from economists of different points of view. The particular title of the session, “Macroeconomic Policy, 1974/75 : What Should Have Been Done?” does not seem to me useful for this purpose, as I will explain below, so I will adopt a somewhat different approach…

    I have developed the reasoning underlying this point elsewhere (Lucas 1975). (Indeed, it follows from modern control-theoretic views of policy evaluation almost independently of one’s views on expectations formation.) I have been impressed both with how noncontroversial it seems to be at a general level and with how widely ignored it continues to be at what some view as a “practical” level. One could ask for no better illustration of this than the question motivating this session: “Macroeconomic Policy, 1974/75 : What Should Have Been Done?”…

    From the point of view of those involved in economic management, the position that policy should be dictated by a set of fixed rules seems at best a partial response to the question: What should be done, now? To one with some responsibility for monetary policy in 1974, say, it is not very helpful to observe that monetary growth “should have” proceeded at a constant 4% rate for the 25 years preceding. Moreover, even if a move toward a policy of fixed rules were desired, it could be done in innumerable ways, presumably with different consequences, and a criterion based on long-run average performance offers no help in choosing among them. What advice, then, do advocates of rules have to offer with respect to the policy decisions before us right now?


  28. Gravatar of ssumner ssumner
    17. June 2019 at 09:17

    Tilting, I agree that the Lucas Critique is important, but I don’t see what it has to do with FedListens. I see no evidence that they are ignoring this issue.

  29. Gravatar of Tilting at M3 Tilting at M3
    17. June 2019 at 13:48

    Scott, if they took the critique seriously, and thought that policy was endogenous, then they couldn’t take the 2008 decline as exogenous and ask “what needs to be done, now?” As opposed to your question, which is “what should the policy regime have been leading up to 2008 that would have prevented the decline in NGDP expectations”

    >”In my view, the basic mistake is to (implicitly?) assume that the need for a negative 5% fed funds rate was caused by an exogenous negative shock (something like the financial crisis, for instance), rather than excessively tight money during 2008, which sharply depressed NGDP growth expectations. If I’m right, then they underestimate the benefits of structural changes in monetary policy that lead to faster NGDP growth, and hence a milder recession. One of these benefits is that the real Wicksellian equilibrium rate would have fallen much, much less sharply.”

Leave a Reply