Why Bernanke’s debt-deflation article is wrong

Ben Bernanke published an influential article back in 1983, in which he argued that debt-deflation could worsen a depression by reducing bank intermediation.  He saw the reduction in intermediation as sort of “real shock,” which could not be completely addressed by easier money (otherwise his model would not have differed from Friedman and Schwartz’s.)

In 2008 he was given a chance few academics ever see—he was allowed to try out his theory on the US economy.  The Fed decided to focus on bailing out the banking system in the second half of 2008, rather than adopting an aggressive policy of monetary stimulus.  Indeed the famous interest on reserve program of October 2008 was implemented precisely to prevent the injection of funds into the banking system from ballooning the money supply and raising prices.  The Fed argued that without IOR the fed funds rate would have fallen close to zero.  (The ff target was in the 1.5% to 2.0% range at the time.)

Unfortunately, Bernanke’s theory is based on a misreading of the Great Depression.  The bank panics were problematic, but only because they led to monetary contraction.  The direct effects were trivial.  How do I know this?  Obviously I cannot be sure, but consider the following evidence:

1.  There were more than 600 bank failures each year during the Roaring Twenties, and yet the economy boomed.  Over 950 banks failed in 1926, a relatively prosperous year.

2.  The rate of bank failures did increase in the early 1930s, but they were mostly the same small rural banks that failed in the 1920s, and the share of deposits affected was a small fraction of the total banking system.

3.  There was one exception, during 1933 bank failures rose dramatically.  The deposits of failed banks were 11% of all deposits.  Much of the banking system was shut down for many months.

And what happened to the economy during this “mother of all bank panics?”  Prices and output soared (as I discussed in the previous post.)  This occurred because in 1933 (unlike 1930-32) the bank crisis was not allowed to lead to monetary contraction.

I would never argue that banking problems had zero impact on productivity, but the evidence from the booming 20s, and from 1933, suggests that as long as NGDP is growing, banking difficulties are not a major factor in the business cycle.  And we also know that banking problems don’t prevent NGDP from growing.  So it looks like Bernanke was relying on the wrong model of the business cycle, and fighting the wrong problem.

The Fed was not trying to raise the rate of inflation during late 2008.  Now they are trying to (modestly) raise inflation, up to around 2%.  They should have done that two years ago, and they are still likely to fall short of their goal.  How do I know?  The Fed’s own internal forecasters just issued a new set of macro forecasts, which are essentially telling Bernanke that $600 billion isn’t enough.  More is needed.

Part 2.  The battle of textbook co-authors

It seems to me that this post loosely relates to the recent back and forth between Tyler Cowen and Alex Tabarrok on the question of whether we’d be better off without the Fed.  I don’t have strong views on the question, partly because it’s not clear to me exactly what is being debated.  If the counterfactual to no Fed is that we go back to the gold standard, then I vote for the Fed.  I don’t wish to rehash the issue of whether the macro economy did better before WWI, or after WW2, and in any case I’m not sure that’s the right question (for instance, almost half the population were farmers in the 1800s–so how can one compare unemployment rates?)  Rather I’d like to point out that the Asian boom has just led to a big rise in real commodity prices.  If we returned to gold I doubt other countries would follow.  And I’m not willing to risk our monetary system on the assumption that somehow one country returning to gold would have prevented that commodity price boom from spilling over into higher real gold prices.  Of course a rise in the real value of gold means deflation for any country with a currency pegged to gold.

If the counterfactual is that the Fed is abolished in 2006, and instead the Treasury puts monetary policy on automatic pilot via a NGDP futures targeting scheme, then count me in.  Tyler might argue that we’d have been worse off without someone to rescue the banking system.  I guess you won’t be surprised to learn that given a choice between stable 5% expected NGDP growth (level targeting) combined with a banking crisis of uncertain size, and collapsing NGDP growth combined with the Fed doing its lender of last resort routine, I’ll take the NGDP target.

Part 3.  The paradox of really stupid monetary policy

People have asked me to comment on the new paper by Paul Krugman and Gauti Eggertsson.  I’ve just skimmed the paper, but much of it seems to revive the various “paradoxes” that I have often criticized.  Their innovation is to directly model the debt crisis.

There’s probably some value in focusing on the debt problem, but I see it as mostly reflecting tight money, not as an exogenous shock.  Their counter-intuitive policy advice (savings, wage flexibility are bad) comes from the assumption that the AD curve slopes upward when at the zero bound.  This means that if you have wage cuts (or more saving), both prices and output fall.  Which means NGDP falls.  There are certainly policy regimes where this can occur, and indeed it might have played a role in the 2008 recession.  For instance, if the Fed is pegging nominal interest rates and people suddenly try to save more (or invest less), then the Wicksellian equilibrium interest rate will decline.  If the Fed continues to hold rates fixed, the money supply will decline, as will NGDP.

But note that this assumes monetary policymakers are stupid.  Some might argue that they really are stupid, so the model applies.  I think it’s more reasonable to argue that they are occasionally a bit slow to respond, and they sometimes let AD fall more than they should.  But any serious discussion of macro stabilization policy that assumes the central bank is hopelessly incompetent is not likely to lead to any good policy options.  Who are we supposed to look to for wise policy advice—Congress?

I don’t think money was tight during the German hyperinflation and hence I don’t use interest rates as a benchmark of the stance of monetary policy.  I use expected growth in M*V, or NGDP.  So when I read Krugman and Eggertsson, I interpret them as saying that more saving or wage cuts might be bad at the zero bound, because it would cause the Fed to tighten monetary policy, i.e. it would lead to lower NGDP expectations.  I don’t think that view of the world is capable of providing useful policy advice.

I also don’t think the empirical evidence supports their view of monetary policy, or wage flexibility.  The 1921 recession (with flexible wages) ended quickly.  The 1930 recession (with very sticky wages) . . . not so well.  FDR’s NIRA was a complete disaster.   Industrial production had risen 57% in the 4 months before his high wage policy (due to an easy money policy), and then increased not at all for the next two years (until the NIRA was declared unconstitutional.)

I also disagree with Krugman’s interpretation of Japan, having argued many times that:

1.  The BOJ said they were opposed to inflation.

2.  The BOJ tightened policy to prevent inflation on several occasions during the 2000s.

3.  The BOJ succeeded in preventing inflation.

I’ve never understood how those facts show that a central bank cannot create inflation at the zero bound.  There are people at the BOJ who are currently warning about the danger of inflation, even as deflation has been accelerating.  I know, I’m not sophisticated enough to understand the subtle nuances of Japanese monetary policy.

If the Congress does more saving (i.e. fiscal austerity) the Fed should do more QE, or a lower IOR.  That’s the policy mix recently adopted by the British.  Is our system in America so inept that we must develop special macro models that rely on our central bankers being more incompetent that the Brits?

Don’t answer that question.

Seriously, part 3 of this post does seem inconsistent with part 1.  But don’t we have to work on policy approaches that assume some sort of rationality on the part of policymakers.  Yes, the Fed shouldn’t have let inflation fall to 1%, but at least they moved when it did.  And what’s our alternative?  Does anyone see Fiscal Stimulus II in the near future?  At least the Fed is doing something.

Update:  People complain that I am too tough on Krugman.  But Bob Murphy is even tougher.  Here he finds an amusing contradiction.


Tags:

 
 
 

26 Responses to “Why Bernanke’s debt-deflation article is wrong”

  1. Gravatar of Mark A. Sadowski Mark A. Sadowski
    23. November 2010 at 19:00

    Scott wrote:
    “Yes, the Fed shouldn’t have let inflation fall to 1%, but at least they moved when it did. And what’s our alternative? Does anyone see Fiscal Stimulus II in the near future? At least the Fed is doing something.”

    QE II means nothing as long as IOER is in effect. The Fed is doing nothing and they know it. We are all headed towards DOOM!

  2. Gravatar of ssumner ssumner
    23. November 2010 at 19:15

    Mark, BTW, check out the update I added.

  3. Gravatar of Mark A. Sadowski Mark A. Sadowski
    23. November 2010 at 19:28

    Scott,
    I sort of buy the “income” argument. QE II seems to be increasing long term yields because of increased expectations regarding NGDP.

    But I’m not completely satisfied. If we just eliminated IOER we wouldn’t even be needing to be engaging in QE II.

    We need to get back to good old fashioned monetary policy.

  4. Gravatar of TGGP TGGP
    23. November 2010 at 19:53

    Mancur Olson in “The Rise and Decline of Nations” pointed to evidence that wages/prices had gradually gotten sticker since the 1890s, and that this resulted deflation increasingly being associated with unemployment or drops in real output (in the 19th century there might be strong growth in real GDP alongside significant deflation). He also pointed to evidence that the stickiness of prices varied by economic sector (agriculture had fewer “distributional coalitions”) and that states with higher unionization had more unemployment.

  5. Gravatar of Mark A. Sadowski Mark A. Sadowski
    23. November 2010 at 20:08

    Mancur Olsen didn’t criticize things as they were, he wrote about things as they were.

  6. Gravatar of Doc Merlin Doc Merlin
    23. November 2010 at 21:01

    @Mark:

    “But I’m not completely satisfied. If we just eliminated IOER we wouldn’t even be needing to be engaging in QE II.”

    I don’t think QE II would help regardless. In the long term IOR will result in the fed being the only dollar bank that matters at all.

  7. Gravatar of Rod Everson Rod Everson
    23. November 2010 at 22:12

    Mark wrote: “QE II means nothing as long as IOER is in effect.”

    Mark, I notice you put the emphasis on interest on Excess Reserves, rather than IOR, interest on Required Reserves. I agree that it is the IOER that matters, but could you explain your reasoning as to why paying IOER causes problems, or direct me to a previous post where you explain it?

    I only recently found Scott’s blog and have read a lot of his older posts, but haven’t had time to read the comments.

    Thanks.

  8. Gravatar of Doc Merlin Doc Merlin
    23. November 2010 at 22:53

    @Rod:

    Interest on reserves prevents the money supply from expanding when it needs to, because banks just let their money sit instead of lending it out. This is why the fed has launched such massive QE, and why the yield curve on treasury debt is so fubar right now. You can earn more overnight on reserves than you can even with 6 month treasuries.

  9. Gravatar of Mark A. Sadowski Mark A. Sadowski
    24. November 2010 at 02:54

    Rod,
    Doc Merlin took the words right out of my mouth.

    Well, I’m off to Rowan now (6AM here) to host an 8AM pre-Thanksgiving exam on monetary policy (I’m such a meanie).

  10. Gravatar of Sumner vs. Eggertson & Krugman « Kantoos Economics Sumner vs. Eggertson & Krugman « Kantoos Economics
    24. November 2010 at 06:53

    […] Krugman habe ich kritisiert, dass sie den Fokus zu wenig auf Geldpolitik und Geldnachfrage legen. Scott Sumner ist da noch kritischer (in „Part 3. The paradox of really stupid monetary policy“), […]

  11. Gravatar of Bill Woolsey Bill Woolsey
    24. November 2010 at 07:15

    I don’t agree with you all.

    If the Fed pays interest on reserves, then they need to expand base money more than otherwise, and also pruchase longer term and riskier assets than otherwise.

    If they quit paying interets on reserves, and even charged a bit, this would drive down the yields on other safe and short assets too. While the amount of base money the Fed needs to create, and the amount of longer term assets might be smaller, there is no guarantee that $2 trillion is enough.

    As for the notion that interest rate on reserves makes the Fed the only dollar bank that counts, it all depends on the interets rate they they pay.

    Now, if you are willing to suspend currency payments, then, yes, you can reduce the interest rate “Paid” on reserves (or really the charge for holding them) and so reduce the demand for reserves enough, so that the needed quantity is as low as you like.

  12. Gravatar of Contemplationist Contemplationist
    24. November 2010 at 08:28

    I nominate Scott Sumner and Bill Woolsey to fill the open positions on Fed Committee. Please please

  13. Gravatar of Rod Everson Rod Everson
    24. November 2010 at 09:55

    Doc,

    Yes, that’s how I see it also, which is why I titled my November 16, 2008 blog post: “Paying Banks to Hold Excess Reserves – Huge Mistake, Huge!” link: http://ontrackeconomics.blogspot.com/search/label/Priority

    What I was attempting to get at with my question is whether he makes a distinction between IOR (Interest on Reserves) and IOER (Interest on Excess Reserves.) I think for clarity we should actually talk about IOER and IORR (Interest on Required Reserves) because, in my opinion at least, they have distinctly different implications for policy. When someone uses IOR, it’s difficult to tell exactly what reserves they mean, all of them, including excess, just the excess or just the overall policy.

    To my mind, IORR is harmless, but a concern for taxpayers as it’s just the government deciding to give away more money. IOER is, I think, a completely different animal and, as you say, keeps banks from lending the funds.

    Thirty years ago I decided that the Fed had all the tools they needed to manage the money supply. That decision was based on the relatively stable policy tools in place from WWII to 1980. Then NOW accounts were approved, followed later by changes in reserve requirements, changes in the period of reserve calculation, etc., and now with IOER. Many of those changes diminished their tools, but more important, obscured their previous effectiveness.

    As Scott mentioned in the P.S. to an earlier post, in 2001, I authored a monograph warning that the Fed would lose control of monetary policy (as previously practiced) if they made the mistake of allowing the fed funds rate to fall to zero. What I foresaw as an issue has indeed become an issue. What I’m hoping is that people reading the monograph come to realize that there is, quite possibly, an easy way out of all this. It’s linked to from the post I referenced above.

    Rod

  14. Gravatar of Mark A. Sadowski Mark A. Sadowski
    24. November 2010 at 10:07

    Rod,
    You wrote:
    “What I was attempting to get at with my question is whether he makes a distinction between IOR (Interest on Reserves) and IOER (Interest on Excess Reserves.) I think for clarity we should actually talk about IOER and IORR (Interest on Required Reserves) because, in my opinion at least, they have distinctly different implications for policy.”

    I absolutely agree. That’s why I always exert the extra key stroke to put in the “E”. It’s an important distinction.

  15. Gravatar of Rod Everson Rod Everson
    24. November 2010 at 11:18

    Mark: “I absolutely agree. That’s why I always exert the extra key stroke to put in the “E”. It’s an important distinction.”

    Thanks for clearing that up Mark. In your discussions, have you come across many others who feel similarly? I only recently decided to rejoin this sort of discussion when I found a reference to some of Scott’s thinking, so I’m really out of the loop on the financial side of things.

    I sort of did my career in reverse order, starting out as the sole bond portfolio manager at a state retirement system in the South in 1975 (the pay was so low they couldn’t get anyone with experience.) Then, I ended up as a government bond salesman in Chicago for several years and left the business at a relatively young age. At one point, I had job offers to work at two money center banks and for a major Wall Street firm’s chief economist, and was pretty well connected to what was going on in the financial world. My interest was primarily in what the Fed was doing, since that’s what drove bond prices the most at the time.

    I wrote the monograph almost 10 years ago, put it on the web a couple of years ago now, and haven’t paid much attention since, until now. Scott’s blog has been very interesting so far and I intend to check out a few of the others that he links to.

    Rod

  16. Gravatar of Mark A. Sadowski Mark A. Sadowski
    24. November 2010 at 11:41

    Rod,
    “In your discussions, have you come across many others who feel similarly?”

    Not many outside of The Money Illusion (and I read several econblogs). And actually I think that I may have picked up the habit of inserting the “E” from Scott although his use of it is less consistent than mine. I think that it’s safe to say that thanks to Scott there’s more recognition of the problem of IOER here than anywhere else.

  17. Gravatar of scott sumner scott sumner
    24. November 2010 at 19:09

    Mark, I agree about IOR.

    TGGP, Those observations sound reasonable.

    Rod, We are trying to reduce the demand for base money. Since required reserves are required, bank demand for RRs is less sensitive to IOR than the demand for ERs.

    Bill, You said;

    “I don’t agree with you all.
    If the Fed pays interest on reserves, then they need to expand base money more than otherwise, and also purchase longer term and riskier assets than otherwise.”

    I’m not sure why you think we disagree. I agree that IOR forces the Fed to be more expansionary than otherwise.

    Thanks Contemplationist.

  18. Gravatar of Rod Everson Rod Everson
    24. November 2010 at 21:14

    Scott,

    Originally the IOER was set 65 basis points lower than the IORR. Do you know the current situation? Also do you know of a link to a regular Fed release that reports the levels? Thanks.

    Rod

  19. Gravatar of Bogdan Bogdan
    24. November 2010 at 21:32

    There is this note by the St. Louis Fed :

    http://research.stlouisfed.org/publications/es/08/ES0830.pdf

    It seems that initially the interest on excess reserves was 75 bps lower than on required reserves and then after November 21 it was set at the lowest federal funds rate of the maintenance period.

  20. Gravatar of Lorenzo from Oz Lorenzo from Oz
    24. November 2010 at 23:34

    So, Irving Fisher in 1933 was right about the importance of price level (that, in the great booms and depressions, each of the above-named factors has played a subordinate role as compared with two dominant factors, namely over-indebtedness to start with and deflation following soon after; also that where any of the other factors do become conspicuous, they are often merely effects or symptoms of these two. In short, the big bad actors are debt disturbances and price-level disturbances … It is the combination of both””the debt disease coming first, then precipitating the dollar disease””which works the greatest havoc. … if the foregoing analysis is correct, it is always economically possible to stop or prevent such a depression simply by reflating the price level up to the average level at which outstanding debts were contracted by existing debtors and assumed by
    existing creditors, and then maintaining that level unchanged.
    And he uses the example of FDR and Sweden to show it can be done) and Ben Bernanke in 1983 was wrong. 50 years of progress in macro-economic understanding!

  21. Gravatar of Bill Woolsey Bill Woolsey
    25. November 2010 at 06:09

    Scott:

    My posts are generally addressed to you, but in this case, I was refering to other comments about how QE does no good if the Fed pays interest on reserves.

    Mark and Rod:

    I oppose the existence of reserve requirements, and that is probably why I have difficulty with distinguishing between interest on required and excess reserves. While it is true that with required reserves, charging banks to hold them won’t cause an increase in the quantity of deposits, it will reduce the interest rate banks pay on deposits that require reserves. This reduces the demand to hold the deposits, which is expansionary.

    In practice, banks can use sweep accounts to reduce required reserves. And so, charging banks interest on required reserves would result in more sweep accounts, more excess reserves, and then, by purchase of securities, an expansion in the quantity of money.

    If it weren’t for the scheme/scam of sweep accounts, then banks might start clamoring to be freed from reserve requirements if they had to pay to hold required reserves. Good! Reserve requirements should go.

  22. Gravatar of Rod Everson Rod Everson
    25. November 2010 at 08:33

    Bill:

    To my knowledge no one has ever proposed charging interest on required reserves, i.e., a negative IORR. While I don’t advocate taking this route, to get excess reserves circulating it would work to impose a negative IOER, i.e, charge interest to hold excess reserves.

    As for doing away with reserve requirements, I assume that you realize that you’re proposing doing away with the fractional banking system and that we might as well do away with the Fed then as well, since its primary purpose is to manage a fractional banking system. So, are you advocating a return to 1800’s style banking? A gold standard system? What?

    Regarding sweep accounts, this is one of those “innovations” that make managing the money supply within a fractional banking system more difficult. They probably should have been discouraged at the outset, but that’s unlikely to happen within a regulatory system that is so bank-friendly that it advocated using taxpayer dollars to subsidize banks via IORR, something never done before and completely unnecessary.

  23. Gravatar of ssumner ssumner
    25. November 2010 at 08:42

    Rod and Bogdan, Yes, They were unified in November.

    Lorenzo, yes, Fisher had it right.

    Bill, Thanks for clarifying that.

  24. Gravatar of Thomas Barton, JD Thomas Barton, JD
    27. November 2010 at 15:23

    Scott, how do you think the Irish should resolve their extreme situation ? Can a nation that small with that much exposure to banks of other nations do anything other than wholesale default and a liquidation of their sovereign banking system ? Or can they not do anything other than muddle through while the Eurozone comes to grip with its larger problems ?

  25. Gravatar of ssumner ssumner
    28. November 2010 at 06:48

    Thomas, It’s seems to me that default would have been better. Let foreign banks provide banking services to Ireland–other small European countries do it that way. There may be an argument that I am missing, but I don’t know what it is.

  26. Gravatar of Keynes General Theory didn´t make Christy Romer´s Book List on “Learning from the Great Depression” | Historinhas Keynes General Theory didn´t make Christy Romer´s Book List on “Learning from the Great Depression” | Historinhas
    18. February 2012 at 14:55

    […] quite likely Bernanke´s “Credit View” had very “negative” implications. According to Scott Sumner: Ben Bernanke published an influential article back in 1983, in which he argued that […]

Leave a Reply