No two liquidity traps are alike

A few years ago it was difficult to find many elite macroeconomists who put much faith in fiscal stimulus.  Now suddenly it is all the rage.  What happened?  The only explanation I can think of is that as soon as nominal interest rates fell close to zero, and aggregate demand seemed to be spiraling downward, much of the profession assumed that we had entered a “liquidity trap.”  I haven’t seen much evidence that very many economists actually know anything about liquidity traps, but perhaps they have a vague memory of the concept being mentioned in their undergraduate macro classes.  Unfortunately, what they learned is almost certainly wrong.

Let’s start with the misconception that Keynes invented the idea.  If one examines the conservative financial press from the early 1930s, one finds frequent references to the “pushing on a string” concept as well as the problem of the zero rate bound.  While perhaps not quite as sophisticated as Keynes’ conception (which itself was pretty primitive), it is clear that the basic idea was almost conventional wisdom well before the General Theory was published.  In the early 1930s it was generally conservatives who used this argument against progressives who favored monetary stimulus.

In the General Theory itself Keynes was famously ambiguous about the idea, but outside the General Theory it was clear that he thought monetary policy was impotent in a severely depressed economy facing near zero interest rates.  That’s why he advocated fiscal expansion in the 1930s.  The only liquidity trap mentioned in the General Theory was the famous 1932 open market purchases by the Fed, which were widely viewed as having failed.  Friedman and Schwartz argued that they did not fail, as did Christina Romer.  But I think that Keynes was basically correct (as did stock and commodity market participants, who drove prices sharply lower during these operations.)

But Keynes was right for the wrong reason.  The problem wasn’t that monetary expansion failed to boost the economy, but that open market purchases failed to boost the money supply (as M1 and M2 fell.)  The monetary base did rise slightly during the bond purchases, but most of the expansionary impact on the base was offset by an outflow of gold, as the operation contributed to devaluation fears.  This example of monetary policy ineffectiveness did not represent a liquidity trap, but rather the well known constraints imposed by the international gold standard.  How do we know this?  Because as soon as FDR abandoned the gold standard in 1933 prices and output began rising rapidly–and the increase was clearly correlated with dollar devaluation.

By 1938 we were back on the gold standard and back in a liquidity trap.  But again, it was not at all what it seemed.  There was no evidence that the Fed was “trapped” in a deflationary cycle, as policy had become entirely passive, with changes in the monetary base merely reflecting gold flows.  Again, this was exactly what one would expect under a monetary regime where the currency was pegged to gold

By the late 1990s Japan faced near zero rates, falling prices, and this time there was no gold standard constraint.  So what happened?  The standard view is that Japan was enmeshed in a liquidity trap, i.e. that they wanted to escape the cycle of deflation but the zero rate bound prevented monetary stimulus from being effective.  There are two big problems with this perception.  First, if the Bank of Japan really wanted to end their deflation, why did they not adopt unconventional monetary policies like currency depreciation, which offers what Svensson calls a “foolproof” escape from the liquidity trap?

Let’s give the BOJ the benefit of the doubt and assume they were afraid of foreign reaction to currency depreciation.  (In that case their problem would be very similar to what faced the Fed in 1932, an exchange rate constraint on policy.)  But even if that were the case, how could they have experienced an almost steadily falling GDP deflator from 1994 until today?  Why did they let the yen strongly appreciate?  And why did they raise rates in 2000, despite the fact that the GDP deflator continued to fall?

Let’s again give the BOJ the benefit of the doubt, and assume they made a mistake in raising rates in 2000.  They did quickly reduce rates to zero again in 2001.  But why did they again raise rates in 2006, despite the continual fall in the GDP deflator?  And why was the BOJ unable to even agree on a 1/4 point cut in late 2008, as their economy was slipping back into severe recession.  At some point one has to stop giving the BOJ the benefit of the doubt, and assume that the steady 1% to 2% percent deflation experienced almost continually from 1994 to today reflects the BOJ policy preferences, and is not a “trap” at all.

Krugman argued that liquidity traps are actually “expectations traps” and that the BOJ was a prisoner of its own conservative reputation.  I certainly wouldn’t argue with Krugman’s assertion that money supply increases that are expected to be temporary will have little or no impact on AD (as I made this argument (in the JEH) 5 years before he did.)  But to call this a “trap” one would have to assume that the BOJ sincerely wanted to end the deflation.  Unfortunately, there is no evidence (in their behavior) that they did.

Now I understand that all these explanations are a bit ad hoc, and I would be the first to admit that they do not apply to current Fed policy.  I don’t believe the Fed is as conservative as the BOJ  Unlike the BOJ, they probably do sincerely wish to avoid mild deflation.  And so I can see how most people might assume that we are in some sort of liquidity trap.  If so, it might be the first true liquidity “trap” to occur under a fiat money regime.

What’s most frustrating about recent Fed policy is that they sometimes act as if they believe that they have done all they could, and at other times act as if they have more ammunition that could be used in an emergency.  For instance, their explanation for the policy of paying interest on reserves only makes sense if they were not in a liquidity trap, but rather were anxious to avoid an overly expansionary policy stance.  On the other hand, Bernanke has advocated fiscal expansion, which only makes sense if he thought monetary policy could do no more.  And yet Bernanke has indicated that the Fed could adopt unconventional methods (such as buying long term bonds or foreign government bonds) if things got very bad.

With nominal GDP now estimated to be falling at a 10% rate, and as a result, the entire world banking system tottering on the edge of collapse, I wonder how the Fed defines “very bad”?

Richard Timberlake once argued that “A proper central bank does not fail because it loses all its gold in a banking crisis.  It only fails if it does not.”  He was referring to the fact that when the U.S. left the gold standard in 1933, it still had the world’s largest reserves of gold.  Those reserves were supposed to be used in an emergency, that’s the whole point of reserves.  To leave a gold standard without having come close to exhausting one’s reserves is shocking evidence of excessive timidity.  Now the Fed has essentially abandoned monetary stimulus, without having even tried to see what would happen if it aggressively expanded the base without paying interest on reserves (or better yet what would happen if it charged interest on reserves.)  The policy of paying interest was avowedly contractionary–the stated aim was to support the Fed funds target above zero.  Even worse, in the past few weeks they have begun rapidly contracting the monetary base, even as interest is still being paid on reserves.



17 Responses to “No two liquidity traps are alike”

  1. Gravatar of Bill Woolsey Bill Woolsey
    24. February 2009 at 07:04

    We know that Bernake believes that problems with credit markets can create a serious “supply-side” impact on real output. If total spending continues growing on target, then there would be staglation. Slower growth or shrinking real output combined with higher inflation.

    This “theory” has “value,” in that it shows how problems in the banking sector can result in reduced output even if prices and wages always adjust so that real spending equals capacity. During the “heyday” of new classical macroeconomics, this would be important.

    So, to prevent problems in the shadow banking system from causing reduction in real output, the Fed begins to do all sorts of things to help particants in the shadow banking system. These firms (investment banks and money center commecial banks) all have solvency problems. And the Fed is backing them up in ways that puts the “solvency” of the Fed at risk. The Fed could end up having guarateed a lot of bad assets, become insolvent, and require a Treasury bailout.

    I am not sure about the exact legal standing of this sort of behavior, but certainly, the Federal Reserve act emphasises a need for the Fed to be financially sound.

    Paying interest on reserves only makes sense if the Fed isn’t trying to maintain expenditure, but rather trying to funnel funds to the shadow banking system. Fewer people want to be ultimate lenders in that sector and would rather hold insured deposits. The Fed is taking that money and trying to make sure it is used to finance securitized loans.

    Of course, now, nominal expenditure is dropping.

    Why not a more agressive monetary policy rather than fiscal policy? Perhaps it is the greater risk of loss if the Fed accumulates even more unconventional assets?

  2. Gravatar of ssumner ssumner
    24. February 2009 at 14:21

    Thanks Bill. I agree that financial turmoil can create a supply-side problem, but I don’t think it is as serious as some others have argued. The banking system was partially shutdown throughout much of 1933, when the dollar was devalued, and yet industrial production initially rose much faster than prices. More importantly, Bernanke clearly favors higher nominal spending, so the Fed’s passivity cannot be explained by the assumption that they are worried about stagflation. How do I know?

    1. Bernanke has called for fiscal stimulus–which makes no sense if he was worried about stagflation.

    2. In today’s testimony Bernanke suggested that for the next several years both inflation and real growth would come in below the Fed’s long run target.

    Like you, I am very skeptical about the Fed taking on all sorts of risk in an attempt to prop up the banking system. Hamilton argues that the Fed has not succeeded (in all its unconventional actions) in reducing risk spreads. He and I think they should focus on AD, which is what would help the banking sector the most. I would also emphasize that I am advocating that they shoot for the same goals as I would of in the absence of a financial crisis. I don’t favor a bank bailout through printing money, but rather the 5% nominal growth that I think is best under ordinary conditions. If that number is too high, we should get out of the recession first, and then gradually scale back to a lower nominal growth figure. But that has been the pattern in recent decades, and a deflationary crisis is not a good time to set a lower target.

    I think the Fed can do all it needs to without buying risky bonds. There are $10s of trillions worth of T-bonds, foreign government bonds, and agency debt. The base would never need to get anywhere near that large. The problem now isn’t the amount of OMOs, it’s the lack of credibility.

  3. Gravatar of dieselmcfadden dieselmcfadden
    25. February 2009 at 05:26

    well, what would be possible explanations for the fed’s passivity?

    you’d think they’d certainly be motivated enough.

  4. Gravatar of David Pearson David Pearson
    25. February 2009 at 09:16

    I have the same question. Why is Bernanke so manifestly reluctant to target expectations? He’s essentially telling us this is all a misunderstanding: a mis-pricing of risk that can be remedied by targeting credit spreads. Perhaps he thinks monetary policy is better targeted “up the food chain”. If de-levering leads to deflation, and rising risk perception leads to de-levering, then target risk perception.

    You argue that its not risk perception but deflationary expectations that are the cause of our problems. Where is the evidence in TIPs spreads, the gold price, or the behavior of monetary aggregates? All are pointing to rising inflationary expectations (from their deflationary lows in the case of TIPs). This is simply diametrically opposite from the experience of 1930-1933, as is the rapid, consistent growth in the monetary base.

    I would argue Bernanke is wrong because de-levering is a consequence of credit mis-allocation by firms and individuals, and not a mis-understanding. Further, I would argue that inflationary expectations arise from concerns that: 1) the government will nationalize a large portion of the 280% of private debt-to-gdp; and that, further, the government will face crowding out and have to resort to monetizing the resulting deficits.

  5. Gravatar of David Pearson David Pearson
    25. February 2009 at 09:25

    BTW, I would respectfully suggest that a material rise in Treasury bond yields should make you re-think your model. A yield back-up would be consistent with the thesis that inflation expectations, following a debt crisis, are primarily driven by fears over government debt growth. I should point out that Credit Default Swap spreads are rising for most G7 countries, and they reached a high for the U.S. today.

  6. Gravatar of KJR KJR
    25. February 2009 at 11:21

    Certainly, further use of monetary policy should play a role in the current stimulus efforts as the policies outlined thus far are likely to be ineffective in terms of closing the output gap and generating substantial economic activity. Its evident that the latest iteration of the stimulus package is relying on the traditional Keynesian framework, based on substantial spending multipliers. Given the current economic condition, the effects of the proposed fiscal spending policies will likely fall short of expectations (too small, likely to be allocated extremely inefficiently, Ricardian equivalence as a real factor in consumer and business behavior, etc). Furthermore, the tax cuts that have passed are also unlikely to generate the desired effects. We would have been better served with payroll tax cuts as suggested by Mankiw.
    As a result of the aforementioned shortfalls of the current stimulus policy it seems necessary that an effective use of non-traditional monetary policy will have to be employed alongside the current stimulus in order to generate economic growth. While the fiscal spending and tax cuts are likely to have some stimulus effect, albeit multipliers are likely to fall short of expectations, it seems likely that monetary policy can indeed be effective in the current environment despite the fact we are facing a zero bound.
    Take as an example, the very entity that most economists are relying on to lead us out of this economic slump, the consumer. Over the past decade a significant increase in household debt has fueled the type of PCE that has sustained economic growth. As a matter of fact, household debt as a percentage of PCE reached an all time high of 140% in 2008 and as you would expect there is a significant correlation between the increase in household debt outstanding and the percentage change in PCE, in real terms and on and annual basis (p value <.001). It is reasonable to suspect that the absence of a proper stimulus will lead to a significant de-leveraging in household debt outstanding and in turn further decreases in consumer demand (why should the financial institutions have all the fun).
    As previously mentioned, it seems unlikely that the proposed fiscal policy will be effective in containing the de-leveraging in the consumer sector, and a significant increase in the money supply coupled with aggressive inflation rate targeting rhetoric from the Fed could better serve to stabilize current household debt levels, increase consumer demand, and in turn support PCE.

  7. Gravatar of ssumner ssumner
    25. February 2009 at 17:59

    Dieselmcfadden, I am also puzzled by the Fed’s passivity–it goes against Bernanke’s academic writings (where he argues that liquidity traps don’t prevent expansionary monetary policy.) Maybe he thinks he’s done enough and is waiting for results. But then why does he advocate fiscal expansion?

    David, One problem with this blog is that I am 4 months too late. You are right that deflationary expectations have moved to roughly 0% inflation expectations. But I still think that 0% inflation in this monetary regime is just as painful as negative three percent under the gold standard (where 0% was the norm.) I also pay more attention to nominal GDP than inflation, and nominal GDP is still falling fast.

    Lot’s of people think the big increase in the monetary base means inflation in the future, but that didn’t happen in Japan (which also had big fiscal deficits in the 1990s.) And I think our MB data is even less meaningful than Japan’s as the payment of interest has artificially ballooned bank reserve holdings. The Fed will quickly reduce reserves if inflation picks up–indeed they have already begun doing so (prematurely in my view.)

    BTW, the monetary base rose rapidly between 1930 and 1933.

    Yes, T-bond yields have backed up somewhat (and that may reflect long term monetization of the debt worries) but they’re still very low. Gold prices may reflect extreme uncertainty, not inflation expectations. Gold was hoarded during the deflationary early 1930s.

    For what it’s worth, yesterday Bernanke said he expects inflation and real growth to both come in below normal for years to come. Unfortunately, I think he’s right.

    KJR, I agree that monetary stimulus is preferable to fiscal stimulus. I am fighting a battle on two fronts; some say we don’t need more stimulus, as we will soon face high inflation. Others say we are in a liquidity trap. If I was a conservative who was a bit worried about inflation being somewhat higher than desirable, I still might support monetary stimulus because:

    1. Some stimulus is inevitable
    2. A small monetary policy error is better than throwing trillions into fiscal stimulus, bank bailouts and nationalization. We won’t have hyperinflation–the error might be 5% inflation instead of 2%–in other words back to 1988-89. That doesn’t look so horrible right now. And I don’t even think that will occur.

    I think we actually agree here, I am just fleshing out my reasons a bit more fully. I also agree that if we must do fiscal stimulus, then Mankiw’s payroll tax cut is a much better way of reducing unemployment than pork barrel projects.

  8. Gravatar of David Pearson David Pearson
    26. February 2009 at 13:28

    The monetary base was around $6b at YE 1929, and four years later it was $7.2b. Yes, it did grow, but its certainly a lower growth rate than we have seen the past nine months.

    Was Japan the exception or the rule? The Fed in 1934-1937 managed to create inflation, as have the vast majority of stimulative (monetizing) central banks facing output gaps following debt crises. Against this weight of evidence we have one example of stubborn deflation in the face of large monetary stimulus: Japan.

    Gold is hoarded for a reason. Why hoard gold over paper currency? The latter is easier to store and vastly easier to exchange for goods. You would only do so if you expect paper currency to decline in value.

    I think its helpful to find a set of indicators that can tell you when your view of the world is wrong. If yields continue falling along with gold, then I would admit mine is. On the other hand, if gold and yields continue to rise in the face of declining output…

  9. Gravatar of ssumner ssumner
    26. February 2009 at 17:50

    David Pearson, Japan had a large increase in the base, and very mild deflation. The US. in 1929-32 had a mild increase in the base, and sharp deflation. So in both cases the deflation was considerably worse than the monetary base data would lead you to expect IF YOU ARE A MONETARIST. Which I am not.

    You are right that gold was hoarded due to devaluation fears. But it was not fear of devaluation against goods and services, it was fear of devaluation against gold. The 4 episodes of gold hoarding in the Depression; late 1931, spring 1932, early 1933, and fall 1937 were all motivated by devaluation fears, and all were deflationary.

    Your last point is a very good question (I guess it is one Karl Popper would ask.) It’s hard for me to answer, but I’ll try:

    1. If my futures targeting idea didn’t make nominal GDP growth more stable.
    2. Assuming idea 1 is a nonstarter, then if ordinary NGDP targeting didn’t make real GDP less volatile than under the current discretionary regime.
    3. If the most sensible predictions one could make about nominal GDP growth from various market indicators (stocks, commodities, bond spreads, etc) did worse than the Fed’s prediction of nominal GDP growth. (Although this is less essential.)

    Regarding the specific issue of liquidity traps; it’s hard to think of evidence that allows one to distinguish between my view and Krugman’s view. I’d guess that he’d find my interpretation of the Japanese case less than convincing. The problem is that it is very hard to distinguish between a central bank that is sincerely trying to create a bit of inflation, and one that is merely accommodating higher money demand from its previous overly contractionary policy. But I did the best I could in the post.

  10. Gravatar of Jerry Jordan Jerry Jordan
    27. February 2009 at 07:44

    Re: Japan, they raised rates in 2000 because the US Tr told them to; throughout they were confused about what was going on with yen prices and the exchange rate. Hayek, and especially Mises, warned about the confusion associated with “inflation” and “deflation.” In the land of yen prices that translated to US$600 scotch, US$50 melons, palace grounds worth more than the State of California, the “law of one price” as they began to open up meant that either the exchange rate had to depreciate dramatically (which foreign trading partners would not permit) or yen prices on many things had to fall (which a central bank tried to prevent). It was as Meltzer wrote about the UK and US trying to reset pre-WW gold prices even tho they had had different inflation experiences.
    Letting yen prices fall to world levels was not deflation (increased purchasing power of money) any more than letting post-communist prices in E. rise to world levels was “inflation.” Yet, fallable central banks make mistakes that convert necessary relative price adjustments into aggregative problems.

  11. Gravatar of ssumner ssumner
    28. February 2009 at 12:18

    Jerry, You might be right that they raised rates in 2000 because we told them to. However:

    1. I am not sure if that is so.
    2. If it is so they were foolish to follow our orders–any mild trade sanctions would have hurt them less than deflation
    3. We were stupid to tell them to do so, as it doesn’t really help our trade balance
    4. Even if you’re right, it would still support my point that their was no Keynesian liquidity trap–it would be like the constraints of the gold standard that I discussed earlier. No non-Keynesian would argue with the proposition that if a central bank is browbeaten into adopting a tight money policy, it may not be able to escape deflation.

    These arguments aren’t necessarily directed against your point (as I’m not sure your view of the liquidity trap), I just wanted to make sure other readers didn’t get the wrong idea.

  12. Gravatar of TheMoneyIllusion » An open letter to Mr. Krugman TheMoneyIllusion » An open letter to Mr. Krugman
    1. March 2009 at 07:40

    […] that this problem may not limit the Fed’s options as much as one might imagine.  The post is here, but the basic idea is that the two famous liquidity traps (the U.S. in the 1930s and Japan more […]

  13. Gravatar of finance guru finance guru
    2. March 2009 at 17:23

    i dont usually comment, but after reading through so much info i had to say thanks

  14. Gravatar of Settlement Settlement
    19. March 2009 at 18:10

    I just wanted to say that I love this site

  15. Gravatar of TheMoneyIllusion » Links to my views on money/macro TheMoneyIllusion » Links to my views on money/macro
    23. September 2011 at 04:50

    […] early critique of the “liquidity trap.”  A longer and more recent […]

  16. Gravatar of どの流動性の罠も似ていない by Scott Sumner – 道草 どの流動性の罠も似ていない by Scott Sumner – 道草
    23. September 2011 at 07:08

    […] by Scott Sumner // サムナーのブログから “No two liquidity traps are alike“(24. February […]

  17. Gravatar of Jimmy Jimmy
    16. October 2011 at 09:04

    My friend taught me the the rinks which show that the BOJ aimed at the “neither inflation nor deflation (without CPI bias) “, but I am mute. Are there the person understanding Japanese near you?

Leave a Reply