Saturday, October 30, 2010

QE Has Worked Before: My Reply to Paul Krugman

Paul Krugman appears to be responding to my previous post where I argued it was unfair of him to dismiss Milton Friedman's views based on what happened in Japan.  In his reply he continues to raise doubts about the efficacy of Friedman's monetary prescription for Japan in the 1990s and the United States in the early 1930s.  He then goes on to question those of us who have been calling on the Fed to target the level of aggregate spending through something like a nominal GDP target. Here is Krugman (my bold below):
I think it’s also important to note that Friedman was all wrong about Japan — and that you can argue that he was also wrong about the Great Depression, for the same reason. For what Friedman argued, both for Japan in the 1990s and America in the 1930s, was that all the central bank needed to do was more — push out those reserves into the banking system. This would raise the money supply, and a higher money supply would have the usual effects.

But the Bank of Japan tried that — and found that pushing more reserves into the banks didn’t even lead to rapid growth in the money supply, let alone end the problem of deflation... So, after 2000 the Bank of Japan engineered a huge increase in the monetary base; this was the original quantitative easing. And it didn’t even translate into a surge in the money supply! This is why I’m so skeptical of people who say that all the Fed has to do is target higher nominal GDP growth — in liquidity trap conditions, the Fed doesn’t even control money, so how can you blithely assume that it controls GDP?
Let me begin my response by encouraging Krugman and other monetary skeptics to have a little more faith in the power of monetary policy.  Here is why: QE has been done before in the United States and it worked incredibly well.  It was initiated in early 1934 when FDR and his treasury officials decided to (1) devalue the value of the dollar relative to gold and (2) quit sterilizing gold inflows.  Now this was a radical move at the time, much like QE2 is to many folks today.  The gold standard was viewed then almost as a sacred institution.  FDR was going to weaken it and allow prices to permanently increase.  How dare he! But that is exactly what was needed, a big permanent shock to inflation expectations that  served to stop the deflationary spiral, end the liquidity trap, and allow a recovery in aggregate demand.  Now this policy move was backed up with significant and permanent increases in the monetary base over time: it went from about $8 billion right before the policy change to about $24 billion by the end of the 1930s.  Below are two graphs that shows this remarkable QE program at work.  Here is the monetary base and M2 (Click on figure to enlarge.):

Note how the increase in the monetary base was followed by an similar increase in M2.  Monetary policy--in its unconventional form--was able to influence the money supply.  And note the slowdown in the monetary base in 1937 was followed by a slowdown in M2.  This is when the Fed decided to stall out FDR's QE program and in the process help create the recession of 1937-1938. (See Francois R. Velde for more on this recession.)  

What about total current dollar spending? Here too, FDR's QE program also served to stimulate aggregate spending as seen in nominal GNP:

Again we see that nominal GNP closely tracks the surge in the monetary base.  FDR's QE was a smashing success when it came to shoring up aggregate spending.  So those of us folks who want the Fed to increase and stabilize nominal GDP have a good reason to believe it is possible--it happened before.

So let's recap the highlights here:

(1) QE was tried before in the United States during a liquidity trap and worked incredibly well.

(2) QE not only worked well, but it did so in an economic environment far wore than that found in either Japan in the 1990s or the United States today.

(3) QE worked  because it (i) it reshaped inflation expectations and (ii) was backed up with meaningful increases in the monetary base.  

The Fed is more than able to do the same today.  It certainly can reshape inflation expectations.  Just look at what has happened to them over the last month or so when Fed officials started talking up QE2.  The Fed would be far more effective, though, at shaping inflation expectations by explicitly committing to some nominal target. Thus, the Fed needs to come out this week with more than an announcement that it is going to purchase say $500 billion assets going forward.  It also needs to shake up inflationary expectations and do so in a manner that creates certainty going forward.  The Fed, then, needs to (1) announce an explicit nominal target and (2) say it will do whatever is necessary to hit it (i.e. it will buy/sell as many assets as needed). I just hope it is a level target for total current dollar spending.

Update I: Josh Hendrickson reminds us that QE seems to be working rather well in the U.K.
Update II: The M2 and monetary base data come from Friedman and Schwartz (1963) while the nominal GNP data come from Balke and Gordon (1986).


  1. please correct me if I'm wrong, but what you're suggesting is quite unlike what the Fed is proposing to do, is it not? One actually sees money going into the economy, the other does not.

  2. This comment has been removed by the author.

  3. Dr. Beckworth, the conservative think tank Economics21 (Keith Hennessey, Ed Lazear, etc.) has a post up critiquing QE, price level targeting and NGDP targeting given current U.S. national income data. I was wondering if you could address it.

    They write:
    Unmooring inflation expectations is a dangerous game and using the “price level” as a target seems especially dangerous now given that most market participants think in terms of changes to prices – inflation and returns on assets. The targeting of nominal GDP could be even more dangerous since it would deemphasize real changes in output in favor of “catch-up” based purely on increases in the price level.

  4. Anonymous,

    I am not sure what you getting at here, but if you mean that QE2 will mean the Fed buying long term treasuries and letting them sit in banks as excess reserves then you have a point. However, based on Fed officials talks and the last FOMC minutes, it is possible they may announce an explicit nominal target. Doing so would go a long ways in shoring up the velocity of money and reducing the excess money demand/liquidity problem. This would get money flowing and into the economy again.

  5. JDTapp:

    Let me take a closer look at the piece, but the excerpt below doesn't make a very coherent statement about NGDP targeting.

    By the way, there is a new article this week the National Review that comes out in favor of NGDP targeting. So there is a prominent conservative magazine that does support the idea. (Sorry, no link except for subscribers)

  6. People are proposing fiscal Austerity which should be massively deflationary. Wouldn't this overwhelm any positive effect that QE2 would have on inflation? How can monetary policy raise inflation "expectations" when fiscal policy is creating deflation in reality?

    Can monetary policy be effective if fiscal policy is pulling in the wrong direction?

    -jonny bakho

  7. Didn't QE just work in the UK?

  8. Let's look at now. The gold standard shock of 75 years ago isn't especially relevant.

    Money supply (billions) increased from 800 to over 2000.

    M2 (billions)went from about 7800 to 8800. Less than dollar for dollar.

    And the M1 multiplier is DOA at 0.88.[1][id]=MULT&s[1][range]=1yr

    This is awfully weak tea.


  9. Question for all: Does the House Subcommittee overseeing The Fed have the right to stop the printing presses ? I've read that the Repubs plan on installing none other than Ron Paul as head of this subcommittee, and he will without a doubt attempt to stop the printing..

  10. Why wasn't the BoJ successful? Lack of credibility? The Fed may have plenty of credibility to halt inflation but none to increase it.

  11. Mr. K -

    I doesn't look as if Congress has any hands-on kind of control. They can amend the the Federal Reserve Act at any time.

    Mr JzB

  12. From what I understand about the gold exchange standard, exiting it reshaped inflation expectations BECAUSE it enabled "meaningful increases in the monetary base."

    In other words, markets in 1930 saw the monetary base as "fettered" by gold and, as a result of expected gold hoarding, prices collapsed. In 1933, from one day to another, FDR broke this perception.

    What do markets expect today? There is no perceived exogenous limit on the monetary base. Prices have been rising mildly, not collapsing.

    If markets expected the zero bound to constrain the Fed, I could see how unconventional policy would have a large impact on expectations. I don't think markets have had that perception since 2008. If anything, markets may have run ahead of the Fed in terms of expectations of QE2 and a level target, and there is a risk of disappointment. The move in the 5yr TIPS spread to 1.7% is welcome, but it still falls short. How does the Fed close the gap to an implied level target? I would argue that if it doesn't close immediately after an announcement, the Fed's credibility would be in question, and the mechanism for rebuilding it (yet more QE?) is not clear.

  13. Jazzbumpa:

    Comparing the increase in the monetary base to M2 during QE1 is misleading. The Fed was not directly trying to stabilize inflation expectations then. Rather it was first and foremost trying to save the financial system. Bernanke himself said we should not call it QE but CE (credit easing). The large increase in the monetary base was a byproduct of the Fed's efforts to save the financial system, not the ultimate objective itself. In fact, the Fed chose to manage this byproduct increase in the monetary base by paying interest on excess reserves, a sure way to stall big changes in the money supply.

    The thinking behind QE2 is far different. Now the Fed wants to directly arrest the decline in inflation expectations and core inflation. This is already happening with inflation expectations. Money supply measures like M2 and MZM are also responding. Stating their nominal goal clearly would make the Fed even more effective. In short, QE1 is a bad comparison because it had a very different objective than QE2.

  14. Lord:

    I think the BOJ was very successful because it was credible. Its goal, however, was a flat price level (i.e. 0% inflation). I am hoping Scott Sumner does a post on this again as a reply to Krugman.

  15. Josh,

    Good point. I hope Krugman takes notice.

  16. David -

    Thanks for the clarification.


  17. Aren't the the Fed QE in the 1930s and the UK QE today really a case of stimulus by devaluation?

    I imagine that is what Prof Krugman would say in reply, and that is why he is so concerned about China's currency intervention.

    What am I missing?

  18. David, re your response to JzB re QE1/M2 etc....I don't understand what the difference is. QE1 added reserves, just like QE2 is going to do. It doesn't really matter how they added. In fact, if I were to nitpick, I'd say credit easing was better because it actually took dodgy securities off the banks. Saving the system? That's great, but the end result was an increase in reserves by the boatload (no matter what new name the Fed assigns to it). So, if they wanted to free up lending space, then what a terrific way to do it hey?. But we know that reserves are not required by banks in order to lend, so the ultimate response is the same. Pretty much nothing but an indeterminate fall in bond yields of little consequence to borrowers and end users who would rather not borrow it appears. We also know that the link from Base to M2 is not exactly stable, and therefore neither is the link from M2 to GDP, nominal or otherwise.

  19. Anonymous:

    Fair points on the unstable link between base, M2, and nominal income. But during much of QE1 inflation expectations were falling. With QE2 inflation expectations are already taking off (money supply measures M2 and MZM are growing already too). Also the fact that the Fed started paying interest on excess reserves indicates it was not interested in seeing big expansion of the money supply.

    Again, the Fed this time has been talking up the important of inflation (and thus by implication spending)where as before it was all credit stabilization. They are purposefully communicating a change in inflation expectations this time. That is a very important change. The questions is whether they will follow through.

  20. Interesting article. Cant say I know never mind understand all details of US QE1 structure/impact etc but from my understanding it certainly worked better than UK QE1. The Fed directly purchased MBS, which helped banks balance sheets, and even stimulated some lending (!). Also biought US Treasuries of course, which brought down borrowing costs for the government and many corporations. (UKQE1 by contrast did virtually nothing for the real economy, small business lending and at level of switched on economist explains here )

    Having said that, not sure US QE2 will have much impact, yields on government debt being so low. Banks want to consolidate, and households do not want to take on more debt at the moment – interesting to note US personal savings rate is still quite high on historically basis, fallen a little bit recently i know...see chart here

  21. Excuse my ignorance, but in all my reading I still am confused about where the Fed gets what it needs to buy up securities. Does it sell assets from its portfolio, as some are saying it's just an asset swap of short term securities for long, or does it just credit the accounts of sellers? The latter would be the "creating ex nihilo" or printing money meme used by many. Or does it depend on the circumstances?

  22. This argument is based on unsound logic. The entire purpose behind QE is to increase lending. Mr. Bernanke has incorrectly diagnosed the current malaise has a banking crisis when it is in fact a household crisis. Mr. Bernanke believes he can make money so cheap that anyone will borrow. The problem is that we have a debt overhang and households are more eager to pay down debt than they are to take on more debt. We see this every quarter in the Z1 data.

    In saying that QE "worked" in the GD you assert that the US economy experienced a lending boom that resulted in recovery. But no such thing occurred. In fact, lending remained tepid throughout the the 30's and into the 40's. Lending did not boom again until after the war started.

    This has been evident in Japan and the UK during their QE programs as well.

    As for the dollar, the argument is the same. QE did not reduce the pound or Yen in Japan and the UK. In fact, both currencies climbed during the program (as did interest rates). Therefore, any economic benefit was not the result of trade benefits. Evidence of QE1 here in the USA tells the same exact story.

    This may have "worked" when we were on the gold standard, but modern evidence shows that QE most certainly does not generate inflation when you're not on a gold standard.