My views on monetary policy: An update

It’s time to update my current preferences on monetary policy:

Monetary authority structure:

1.  As in the UK, monetary policy decisions should be made by a committee of monetary specialists.  They need not be economists, but they must be experts. Self-taught is fine. Financial regulatory decisions should be made by a separate (Treasury) committee, composed of experts on finance.

2.  The FR Board should continue to be independent.  The regional Fed banks should be abolished.

3.  The Fed balance sheet should be moved over to the Treasury so that the Fed does not incur any balance sheet risk.

4.  Anyone should be able to have a “reserve” account at the Treasury. It would pay no interest.  I.e., electronic cash.

5.  Paper currency should not be abolished (as it provides privacy.)

Policy Tool:

The Fed should use just one tool, open market operations.  Generally these operations should involve Treasury securities.  There’s no need for the Fed to recommend discount lending, reserve requirements and interest on reserves as tools of monetary policy.  Other assets should be purchased only if necessary.  If we don’t get my optimal NGDP level targeting strategy and the zero bound problem occurs frequently, then there is a much stronger argument for making the purchase of other (non-Treasury) assets a routine part of policy.  Bank bailouts should be done by the Treasury, if at all. (Hopefully not at all.)  There should be an iron curtain between the Fed and the banking system, like the separation of church and state.

Policy Target:

Nominal growth targets should be high enough to avoid the zero bound problem.  With level targeting, a 4% NGDP growth rate should be high enough.  We should target expected per capita NGDP growth, level targeting.  (Perhaps growth of one basis point a day is a nice round number, for that future time when “big data” allows us to know daily changes in NGDP.)  A nominal total labor compensation target is better for some countries.  The Fed should completely offset the impact of any fiscal policy action.

Policy Rule:

The Fed should commit to a “guardrails approach”, a willingness to sell unlimited NGDP futures contracts at a implied NGDP growth rate slightly above target and buy unlimited NGDP futures contracts at an implied NGDP growth rate slightly below target.

Policy Accountability and Transparency:

The Fed should report to Congress twice a year on the effectiveness of previous policy decisions.  They should explain whether, in retrospect, policy settings adopted 12 months earlier were too easy or too tight, even if only slightly so.  They should clearly explain the metrics they used to make this determination.  This form of accountability is especially important if NGDP targeting is not adopted, less so if it is.

What am I missing?  I may add a few items based on comments.



17 Responses to “My views on monetary policy: An update”

  1. Gravatar of Brett Brett
    11. January 2019 at 09:36

    4% is too low. I think we could easily sustain 5-6% NGDP as a target without too much of it being lost to inflation.

  2. Gravatar of John Hall John Hall
    11. January 2019 at 10:14

    I’m glad you did this post. It’s a useful thing to do…write down what you believe in a short succinct way. It probably would be a good thing to do something similar on what you think is wrong about the mainstream approach to monetary policy.

    In terms of missing things, I would recommend three things:
    1) I think in the Policy Target section, it might be useful to list a few different cases. For instance, for a large open economy, use NGDP, for a small closed economy, use X, for a small commodity exporters, use Y.
    2) You don’t discuss NGDP futures until the Policy Target section (more generally they should be Policy Target futures?). It seems like you need at least another section discussing the importance of futures and that central banks should subsidize futures markets.
    3) You mention that the central bank should buy unlimited NGDP futures contracts, but you don’t specify which ones. Operationally, I feel like a central bank would have to be very particular about what it does, and that’s probably not necessary for a short summary piece of what you believe. Nevertheless, I think it’s important. Do they only intervene in the one-year ahead contract? Why not intervene in many contracts?

    In terms of criticism, I’ve had an issue with your emphasis on 4% NGDP growth for some time after most of your earlier blog posts favoring 5%. 5% was the pre-Great Recesssion trend and 4% was the trend post-Great Recession. But this was also a very slow recovery from a recession. Is that something you really want to stick to? Or does it not matter what the actual level is so long as we keep it fixed, or change it very slowly. If that’s what you really believe, then put that in instead of the 4%.

    Also, I’m not completely sold on guardrails. This idea started appearing on the blog lately, but I’m not sure you’ve really defended it. You’ve used 1% guardrails in the past, but why not 0.5% or 1.5%. AFAICT, you are picking a number out of the air (moreso even than the 4% vs. 5% NGDP target).

    Instead of guardrails, I’d recommend the central bank create an options market on the NGDP futures. It can provide valuable information on the shape of NGDP expectations. They can primarily intervene with futures, but they could also measure tail risk and potentially intervene by buying puts/calls if needed. For instance, if you wanted to implement guardrails with puts and calls, then you buy puts on 5% and buy calls on 3% on an unlimited basis. Market participants can then hedge those by buying and selling futures contracts and the requisite amounts. I’m not sure I’ve fully thought through the implications, but it makes sense to me.

  3. Gravatar of lat lat
    11. January 2019 at 11:39

    Can you point to a post/paper/article wherein you provide more details on the NGDP futures and guardrails idea? I’m familiar with futures/options, but am having difficulty conceptualizing the Fed interacting in a futures market for NGDP. Mahalos in advance.

  4. Gravatar of Ralph Musgrave Ralph Musgrave
    11. January 2019 at 11:43

    Nice clear article. I agreed with 1-5. Re “open market operations” I don’t agree with that because there are no good reasons for government borrowing, as suggested by Milton Friedman, Bill Mitchell and me (links 1 -3 below). No government borrowing means no open market operations.

    Re the “zero bound problem”, what’s the problem? If stimulus cannot be implemented via interest rate cuts, just have government borrow and spend (or as Keynes suggested, print money and spend it, and/or cut taxes). Indeed the latter “implement fiscal stimulus when near zero” is central to the UK Labour Party’s new so called “fiscal rule” (authored by Simon Wren –Lewis, who is an Oxford economics prof). Unfortunately I can’t find the latter “rule” by Googling, but it’s definitely out there somewhere. I may try later an put the link on this Money Illusion site.


  5. Gravatar of Garrett M Garrett M
    11. January 2019 at 12:42

    John Hall’s idea on NGDP futures options is interesting and I’d love to hear Scott’s thoughts on it.

  6. Gravatar of Market Fiscalist Market Fiscalist
    11. January 2019 at 13:16

    I have a question on the policy rule:

    ‘The Fed should commit to a “guardrails approach”, a willingness to sell unlimited NGDP futures contracts at a implied NGDP growth rate slightly above target and buy unlimited NGDP futures contracts at an implied NGDP growth rate slightly below target.’

    While in principal this seems like an elegant way to vary the money supply in order to hit an NGDP target is it actually better in practice than just buying and selling govt bonds (or other assets) to vary the money supply?

    I worry that people will find it gimmicky, or endlessly debate whether it is possible to game an NGDP futures market etc and this will slow down adoption of NGDP targeting.

  7. Gravatar of ssumner ssumner
    11. January 2019 at 14:27

    Brett, I’m not sure I follow your logic. Trend inflation rises one for one with trend NGDP growth.

    John, The link below discusses my guardrails approach. There is no need to subsidize futures trading under a guardrails approach.—Nudging-the-Fed—Scott-Sumner.pdf

    I have no objection to an options market for NGDP futures. I doubt it would provide much useful information if you had a guardrails system in place, but under our current discretionary regime it could provide valuable information about tail risk, as you say.

    I’m indifferent between 4% and 5%, but 4% is a much easier sell to a country that has become used to 2% inflation, and is likely to experience 4% growth. So I usually pick 4%.

    Lat, See my reply to John.

    Ralph, If there is no Treasury debt then buy something else. But of course there’s no chance of the Fed facing that “problem” in my lifetime.

    Garrett, See my reply to John.

    Market, I would continue to use the sale and purchase of government bonds as the tool for adjusting the money supply. The NGDP futures market is like a beeping signal when a truck backs up—the only purpose is to help the Fed avoid getting off course.

    There is no way to game the system, as any attempt to do so would be quite rightly ignored by the Fed.

  8. Gravatar of Doug M Doug M
    11. January 2019 at 16:02

    Can the Fed be independent at the same time as Transferring their balance sheet to the Treasury? It seems that your recommendations are in conflict.

    What is the rationale for reserve accounts at Treasury? It doesn’t seem like the Treasure has the infrastructure in place to take this on. Nor does it seem to be something in their mandate. Do you not like private banks for this function?

    As for the rest of it, I am on board.

  9. Gravatar of dtoh dtoh
    11. January 2019 at 16:16

    1) I think your target is a little low because you’re under-estimating the potential real growth rate (resulting from tax-cuts and the reduction of some government regulation.)

    2) I think there needs be discussion about the trade off between inflation and growth. At what point does the cost of a little more inflation exceed the cost of a little more growth. 2% inflation seems to be a magic number that no one debates.

    3) Bank bailouts are unavoidable, but they should be made so painful to management, highly compensated employees, and shareholders that it rarely happens. Dreaming that TBTF will go away is completely unrealistic.

    4) Once the regime was in place, monetary policy would become super boring, and the rate of OMP would run at roughly the rate of NGDP growth.

    Think you are missing a couple of things.

    5) You don’t sufficiently emphasize the folly of OMP solely within the banking system (e.g. Fed selling Treasuries to Chase and Chase depositing the money in an ER account with the Fed.)

    6) Don’t think you have fully explained the ease of dealing with the ZLB. You mention buying other assets, but there are other solutions such as the Fed imposing a surcharge on net aggregate cash withdrawals (in excess of the NGDP target rate) from the Fed. You can also have anonymous electronic cash that carries a negative interest rate.

    7) It’s silly to involve the Treasury. There is no need, and it just brings in an additional group of incompetent government bureaucrats.

  10. Gravatar of Benjamin Cole Benjamin Cole
    11. January 2019 at 17:06

    It seems to me, in the real world, under the Sumner plan the Fed would end up buying a few trillion dollars in securities during any particular recession, and then transferring those securities to the Treasury. This may work, and quantitative easing may work.

    Ergo, we are willing to print money and buy securities during a recession. In practice, the Treasury may end up owning the securities in perpetuity. See Japan.

    Would it not make more sense to simply print money and finance the federal deficit? This would be the money-financed fiscal programs proposed by Ben Bernanke in Japan in 2003.

    QE as economic stimulus has a bit of the Rube Goldberg in it. You pull this lever in the hopes that that lever will be activated resulting in action C to open door D.

    There is much historical precedent to suggest that money financed fiscal programs are a very effective way to stimulate an economy.

  11. Gravatar of Will Will
    11. January 2019 at 17:11

    Do you have any posts about why you think the regional banks should be abolished?

  12. Gravatar of John hall John hall
    11. January 2019 at 17:31

    I just read the guardrails thing. I still don’t know how I feel about it. What’s to stop it from getting stuck at the lower or upper bound permanently. You also don’t explain why the futures market doesn’t need to be subsidized in that paper (versus your original non-guardrails approach that just targeted 5%). I mean, the whole point of the efficient market hypothesis is that you need the wisdom of the crowd for it to work. If no one pays attention to the market, then you’re not getting the wisdom of the crowd.

    On 4% vs. 5%, let’s start by looking at NGDP and PGDP growth from 1990 through 2005 (because this is a period very close to a 5% pace) and their expectations from the Philly Fed SPF. NGDP grew about 5.2-5.3% over this period annually, while PGDP grew 2.2% on average. When you look at the expectations, NGDP was around 5.3%, but PGDP was around 2.5%.

    Now, let’s update this and look at the same information from 2010 through Q3 2018. We get 4% NGDP growth and 1.7% PGDP growth. However, expectations are 4.5% and 1.8%.

    In other words, we have averaged NGDP of 4% growth, even though forecasters expected 4.5%. In fact, the SPF forecasters NEVER expected a 1 year forecast of NGDP of 4%. They hovered between 4 and 5% almost the entire time. Regardless, we have rarely had 2% PGDP growth or expected growth during this period. We’ve only begun to get 2% expected PGDP growth numbers in the past two years. So, that’s why I think 4% is too low.

    At the end of the day though, it just feels like a number pulled out of a hat.

  13. Gravatar of artifex artifex
    11. January 2019 at 20:33

    I am confused now. I thought there was no zero bound problem, so there should be no need to avoid it?

    Another confusion: bank bailouts should be done by the Treasury? Is this because the efficiency loss from suboptimal monetary policy (the Fed not meeting its target, assuming the target is good) is greater than the deadweight loss from past, current and future taxes used to fund the bailout (I doubt that), or is it because having a bailout done by the Fed sets a precedent, has slippery slope tendencies, or some such concern?

    I like John Hall’s suggestion of creating an options market, but I like idea futures with a scalar range better. The guardrails are the most reliable way to hit the target. I see no reason the Fed should have any other instrument. (Other than political feasibility, but if feasibility is a concern then there’s no point in discussing bailouts or universal accounts at the Fed.)

  14. Gravatar of ssumner ssumner
    12. January 2019 at 10:07

    Doug, Yes, they can still be independent. As it is, all their profits go to the Treasury.

    I have a recent Econlog post that explains why I want to abolish the Fed balance sheet and transfer it to the Treasury.

    In our system, cash and reserves are provided by the government, not private banks. I’m assuming that system stays in place. My only change is that reserve accounts be open to all, not just banks. Why should banks get special treatment?

    Banks would still provide interest bearing accounts.

    dtoh, I do think that a reduction in government regulation could boost growth, but the total amount of regulation is still increasing. The slight slowdown in the rate of increase is not enough to boost growth significantly. Recent tax cuts will only slightly boost trend growth. In any case, it will be years before my system would be accepted, and by that time will know if trend growth is higher. So far the evidence suggests that it is not much different.

    I agree that the zero bound can be dealt with, but I’d rather not have to deal with it. I’d prefer a small Fed balance sheet.

    Will, They are a waste of money.

    John, No subsidy is needed because the system doesn’t require any actual trading to occur. It’s just a warning device.

    Your question is like asking “What if people don’t use the guardrails along the highway? What if they drive in the center of the road?”

    So what? That’s what we want. No trading of NGDP futures would be great, as the wisdom of crowds would be saying that policy is optimal.

    Artifex, The zero bound does not prevent monetary stimulus, but it does lead to a large Fed balance sheet. Look at Japan and Switzerland. I’d rather avoid that.

    I’d like to create a wall between banking and the Fed, because the Fed does a better job with monetary policy when they are not distracted by banking issues, which are unrelated to monetary policy. I’d also like to have the Fed avoid doing environmental or medical regulation.

  15. Gravatar of Ralph Musgrave Ralph Musgrave
    13. January 2019 at 03:00


    I’m sure stimulus COULD BE implemented by having the Fed buy up all sorts of assets other than government debt, but that involves the Fed buying assets which are regarded as being better owned by the private sector. I.e. under the latter arrangement, the Fed is entering COMMERCE, and central banks do not normally do that. I.e. that involves distortion.

    In contrast, if stimulus is implemented simply by having the central bank create more base money, which is used by government to raise public spending and/or cut taxes, that is not distortionary. Public spending is widely seen as desirable, and tax increases and cuts are perfectly normal. Plus the latter ploy does not necessarily involve an chance to public spending as a % of GDP.

  16. Gravatar of Matthias Goergens Matthias Goergens
    13. January 2019 at 23:26

    ‘What else?’ depends on how you want to balance feasibility of your suggestions vs quality.

    Eg ‘Just abolish the fed, and almost all finance specific regulation to herald free banking.

    But politically, that’s never gonna fly. (Though you could keep a rump Fed around, who’s job of stabilsing ngdp will presumably become rather easy.)

    For something more politically realistic, your suggestions weren’t bad.

  17. Gravatar of ssumner ssumner
    14. January 2019 at 10:43

    Ralph, You said:

    “In contrast, if stimulus is implemented simply by having the central bank create more base money, which is used by government to raise public spending and/or cut taxes, that is not distortionary.”

    That’s crazy. Of course it’s distortionary to do tax or spending changes in order to stimulate AD, not because they pass any cost/benefit test.

    In any case, I prefer a monetary regime where the central bank would only need to buy Treasuries. Fiscal policy is like burning piles of money–utterly wasteful.

Leave a Reply