A Moderate and Pragmatic Proposal for Monetary Stimulus

In March of 2009 I presented a proposal for monetary stimulus, in the form of a petition.  I think it’s fair to say that it didn’t attract much attention.  The commenter Benjamin suggested it’s time for a new and specific proposal.  After all, now that fiscal stimulus has failed to generate an adequate recovery, there is a renewed focus on the need for monetary stimulus.

There is no point in proposing my dream monetary policy–NGDP futures targeting and all.  The Fed would never contemplate anything so radical at this time.  Instead I am going to suggest something that just might be acceptable, should the Fed decide the economy needs more demand.  The term ‘moderate’ refers to the fact that I won’t ask the Fed to deviate from their 2% implicit inflation target, and the term ‘pragmatic’ refers to the fact that I won’t ask for risky and untested ideas such as negative interest rates on excess reserves and/or NGDP futures contracts.

Any monetary stimulus proposal has at most three primary components:

  1. A bigger supply of base money
  2. Less demand for base money
  3. A commitment for greater monetary stimulus in the future

I will try to use all three approaches, and do so in a synergistic plan that consists of more than merely the sum of the parts.

1.  Quantitative easing

The Fed should commit to doing as much quantitative easing as necessary to hit its macroeconomic objectives.  They might want to initially commit to a specific figure like $100 billion a month, but over time I believe they should adjust the amount of QE to reflect the demand for base money.  What makes this so tricky is that the demand for base money is itself very sensitive to expectations of NGDP growth, i.e. expectations about whether the policy will fail.  Thus QE will work much better if combined with other policy tools that increase the likelihood of success.  If the Fed does QE, and QE alone, it is very possible that it will be no more effective than the previous $1 trillion in reserves that were injected into the banking system in late 2008.

People like Andy Harless have suggested the Fed buy long term bonds.  I’d prefer just the opposite—the purchase of low risk T-bills and short term T-notes.  By doing so they avoid the risk of significant capital losses if (as I expect) the operations had to be reversed as a result of a robust economic recovery.  I understand that most people envision the transmission of monetary policy through the Keynesian lens of changes in interest rates.  And that it looks like the only interest rates that can now be significantly lowered are the longer term rates.  But I just don’t believe that we can get a robust recovery in NGDP growth without substantially higher long term rates.  Yes, it’s possible that long rates would fall sharply immediately after the Fed purchased lots of T-bonds, and then rise sharply a few months later as economic recovery picked up.  But I have trouble reconciling that scenario with rational expectations.  And I am reluctant to recommend a mechanism that seems to rely on sophisticated bond traders being too dense to understand what is going on.  In Andy’s favor, something like that quick reversal did seem to occur in the spring of 2009, but I think we’d be pushing our luck to rely on it happening again.  For me the transmission mechanism is higher asset prices (stocks, commodities, commercial RE, etc) as expectations of future NGDP growth rise.  That avoids the paradox that we seem to need lower long term interest rates, even as higher rates are associated with prosperity.

Before proceeding to the other two policies, it might be helpful to use an automobile analogy.  Consider QE to be like the car’s engine.  If the transmission is not engaged, the car will not move.  In addition, the car needs to be steered, so that it doesn’t shoot off into a ditch.  The other two proposals are intended to do just that.  Lower IOR can serve as a sort of transmission mechanism, making sure the added QE actually drives the car forward.  And price level targeting will serve as both a transmission mechanism and the steering mechanism letting the Fed know if they have done too much or too little QE.

2.  Eliminate interest on excess reserves

At first glance this seems like a no-brainer.  Many people seem concerned about all the money the Fed has been printing.  Eliminating IOR would allow for the same amount of monetary stimulus with a far smaller monetary base.  Apparently the Fed is concerned about the effect on MMMFs, which might see their rates of return fall to near zero, and thus threaten again to “break the buck,” as occurred to one fund in late 2008.  In that case, I presume that my idea of negative rates on ERs (recently endorsed by Blinder!) would be even more problematic for the MMMF industry, so I won’t advocate that now.

This is not my area of expertise, but the need for monetary stimulus is so great that I think the Fed should eliminate IOR on excess reserves and hope for the best.  Indeed the need is so great that I think they should proceed even if it necessitates the Fed engage in some sort of microeconomic intervention in the MMMF industry that is otherwise undesirable.

I would add that the proposal only applies to ERs, not required reserves.  Thus banks may still be able to profitably offer checking accounts, as each additional account creates a derived demand for more required reserves, which still earn interest.  If the Fed is worried about how removing IOR would impact the banking industry, they have several options.  They could actually raise the rate on required reserves.  This would still encourage banks to reduce their holdings of ERs, which would no longer earn any interest at the margin.  Or, they could “grandfather in” existing reserve holdings for each bank, and only eliminate IOR for the new money that is to be injected in my first proposal; the roughly $100 billion a month in QE.

By eliminating IOR, any additional QE is more likely to find its way out of ERs and into circulation.  But even that may not be enough.  The last proposal is by far the most important.  It provides additional impetus to getting the new money into circulation, and also calibrates how much is needed:

3.  A 2% price level growth path from September 2008, level targeting

Cutting edge monetary models by people like Woodford emphasize that in a liquidity trap you really need price level targeting, not inflation targeting.  The problem with inflation targeting is that it is “memoryless.”  That means if you miss your inflation target for last year, you “let bygones be bygones” and continue to shoot for the same inflation target next year.  Despite the name, “level targeting” doesn’t really mean keeping the price level constant (unless zero inflation is the target) rather it means trying to return to the planned price level trajectory anytime you temporarily diverge from the desired inflation rate.  The intuition is so appealing that none other than Ben Bernanke recommended that the Japanese do exactly that when their CPI had undershot their zero inflation target in the early 2000s.  Bernanke suggested they should temporarily aim for 3% or 4% inflation to catch-up to their original target path.

I am asking the Fed to do the exact same thing that Bernanke recommended for the Japanese.  The only difference is that the Japanese inflation target is 0%, whereas the Fed’s implicit target (they don’t have an explicit target) is believed to be about 2%.

[As an aside, some have argued that the implicit target is actually 1.5% to 2%.  On the other hand over the past two decades they have behaved as if their implicit target is a tad over 2%.  And you could argue that a period of 9.5% unemployment is not the best time to try to bring inflation down to 1.75%.  So you could just as well argue for 2.25% inflation, given actual Fed policy over recent decades.  My 2% proposal is a very reasonable and moderate compromise between their actual policy of slightly over 2%, and the oft-mentioned range of 1.5% to 2.0%, which centers on 1.75%.]

If we track core inflation since the liquidity trap started around September 2008, we find that the core CPI has fallen about 1.4% below the Fed’s implicit 2% target.  In that case, the Fed should commit to trying to raise prices by 2.7% a year over the next two years, and 2% thereafter.  Markets currently seem to expect about 1% annual inflation over the next two years (based on TIPS spreads.)  Why would an extra 1.7% inflation have such a big effect?  Because the SRAS curve is fairly flat when unemployment is high.

After the 1982 recession, Paul Volcker engineered 11% NGDP growth over the first 6 quarters of recovery.  Real growth was 7.7%.  I am not claiming we could achieve the same.  After all, the supply-side fundamentals are not quite as strong as in 1983, as many right-wing economists have pointed out.  But only the most extreme RBC economist could claim the SRAS is vertical, and that moving inflation from 1.0% to 2.7% for 2 years would not substantially boost growth.  FWIW, I’d expect 2.7% core inflation to result from 7% to 10% NGDP growth, implying a real recovery of 4.3% to 7.3%.  That’s much better than we are currently getting.

Now for the hardest part, how do we make it all happen?  A target is just an aspiration, isn’t it?  Not quite.  That is true of inflation targets, but not price level targets.  Importantly, price level targets are both goals, and commitments to do something later to catch up if you fail to meet your goal.  And that future commitment is also very important.  We won’t be at the zero rate bound forever, thus it’s important for the Fed to give markets some sense of the price level trend line they plan to return to when the recession is over.  If the trend line is about what TIPS markets currently expect, we’ll get a weak recovery.  If it is the one I suggest, we’ll get a much more robust recovery.  Would Congress object?  I doubt Barney Frank (a Congressional inflation target critic) would complain, if it was explained that prices had fallen short and that the proposal was aimed at boosting growth.

So what about my first proposal to do QE?  How do we know when we have done enough?  I believe we should follow Svensson’s maxim that we “target the forecast.”  The Fed should do QE until its internal forecast of core CPI growth two years out is on target.  Bernanke and Woodford once pointed out that sole reliance on market CPI forecasts (such as the TIPS market) would create a circularity problem.  There are ways to rely completely on market forecasts w/o a circularity problem, but they are too radical for the Fed, as the market would essentially become the FOMC.  But Svensson’s idea for targeting the Fed’s own internal forecast is eminently reasonable.  Set your steering wheel at a position where you expect to reach your destination.  Do the amount of QE that leads you to expect on-target core CPI growth.

I’m certainly not suggesting the Fed ignore market signals.  They should look at a wide range of market indicators.  Bernanke and Woodford acknowledged those could be helpful, as a supplement to the Fed’s internal structural models.  And let’s face it; QE is going into uncharted waters, so we can’t rely exclusively on structural models.  The Fed should not let market forecasts diverge too far from their policy goal.

4.  Summary

There is nothing radical in my proposal:

  1. QE with T-bills is a plain vanilla open market purchase.
  2. A zero IOR rate is the way the Fed operated for 98% of its history.
  3. Two percent core inflation is widely seen as the Fed’s implicit target.
  4. Level targeting has an impeccable pedigree, with strong support from people like Bernanke and Woodford

And let’s be clear about one thing.  I am not proposing any sort of dramatic change in Fed policy.  Their policy has generally been roughly 2% inflation (I’d prefer 5% NGDP growth.)  All I am saying is let’s stick to that policy.  It is the hawks who suggest policy settings likely to reduce inflation below the Fed’s traditional 2% target (despite 9.5% unemployment) that are the true radicals.

In his public speeches Bernanke really has no choice but to advocate current policy.  And he is reluctant to change policy if it leads to a badly split FOMC.  That gives 3 or 4 hawks an effective veto on change, and insures the status quo hawkish policies continue.  But what does Bernanke believe in his heart?  Does he support the views of people like Charles Plosser, who seem to claim faster NGDP growth would not boost RGDP growth?  Does Bernanke now hold views that are completely inconsistent with his entire academic career, and his recent advocacy of fiscal stimulus?  Or would he prefer something closer to the plan that I have outlined.  Someday we’ll find out the truth.

HT:  Marcus, Liberal Roman, Benjamin Cole

About Scott Sumner 490 Articles

Affiliation: Bentley University

Scott Sumner has taught economics at Bentley University for the past 27 years.

He earned a BA in economics at Wisconsin and a PhD at University of Chicago.

Professor Sumner's current research topics include monetary policy targets and the Great Depression. His areas of interest are macroeconomics, monetary theory and policy, and history of economic thought.

Professor Sumner has published articles in the Journal of Political Economy, the Journal of Money, Credit and Banking, and the Bulletin of Economic Research.

Visit: TheMoneyIllusion

Be the first to comment

Leave a Reply

Your email address will not be published.