The Lionel Robbins lectures

A few days ago Paul Krugman gave three lectures at the LSE.  I have not heard the final lecture, but I thought I would make a few comments on the first two lectures.  As usual, I will focus on those areas where I disagree with Krugman’s views.  However I should also emphasize that I wouldn’t even waste time analyzing his views if I didn’t regard him as the closest thing we have to Keynes himself.  When Bob Murphy defended me to his Austrian readers with the line “finally, a man worth killing,” I took it as a compliment.  I won’t make the same remark vis-a-vis Krugman because we live in a “brave new world” where one can be arrested for making a joke while waiting to board an airplane.  So I’ll just say Krugman should take my obsession with his every utterance as a compliment.

1.  What happened to the expectations trap?

In 1998 Krugman developed a model of the Japanese liquidity trap.  He correctly noted that one way of getting out of a liquidity trap was to lower the real interest rate through a policy of inflation targeting.  Even at zero interest rates an increase in the money supply that is perceived to be permanent will increase the future expected price level, and hence the current expected inflation rate.  But he also saw a problem with this scenario.  Suppose the central bank had a conservative reputation like the Bank of Japan.  In that case the public may not find a promise to inflate credible, any monetary injections would then be viewed as temporary, and thus would be hoarded.  He called this scenario an “expectations trap.”

I have argued many times that this problem is unlikely to occur in the real world.  I know of no central bank that sincerely wanted modest inflation, but was not believed by the public.  Certainly the Japanese public has no expectation of inflation, but that is because they have observed the BOJ follow deflationary monetary policies for 15 years, even raising rates when necessary to maintain steady deflation.  Nor does the 1930s fit the scenario, because as soon as the US decided to inflate it left he gold standard and prices immediately began rising rapidly, despite high unemployment an a banking crisis.  In contrast, Krugman thinks that it is very hard to get out of a liquidity trap, and given the current US situation, I can see why his views have appeal.

In the LSE lectures, Krugman did not seem to emphasize the time inconsistency part of the expectations trap, the idea that a central bank that sincerely wanted to inflate would not be trusted.  Instead he emphasized two other, slightly different aspects of the problem:

1.  He talked about his discovery that (during a liquidity trap) a doubling of the money supply that was seen as temporary would be hoarded.  And thus prices would not rise at all.  And he presented this insight as if it was his discovery.  Perhaps it was, in the context of liquidity traps, but it has always been known that temporary currency injections have little or no impact on the money supply.

2.  He suggested that monetary injections would only cause a proportionate increase in the price level if they were perceived as permanent.  Again, that is clearly true, but so what?  Almost every year the money supply is larger than the year before, so it’s not like permanent monetary injections are some sort of anomaly.  Rather it is temporary currency injections, like the recent increase in the monetary base, that are rare.

3.  He argued that central banks were still unwilling to target inflation.  I don’t recall the exact words, but it was something to the effect that Volcker had argued that after working so hard to get us back to price stability, he [Volcker] strongly opposed any attempt to create inflation.  This is odd example for a number of reasons.  First, Volcker made the statement in a debate with an actual central banker, and one that did favor inflation targeting.  Second, Volcker didn’t bring us to price stability; he brought us to 4% inflation.  Nor did Greenspan bring us to price stability; he brought us to 2% inflation.  Since when have non-Japanese central bankers believed in price stability?  I thought they all wanted to target inflation?  Nor do I see Congress or the President clamoring for price stability.  Why is inflation targeting then viewed as politically impossible?

Now I suppose you could say Bernanke has brought us to price stability, in the past 12 months the CPI has not risen.  But wasn’t that a politically unpopular policy?  The Phillips curve has returned with a vengeance.  Does anyone really believe that 0% inflation and 9.4% unemployment is more politically acceptable than 2% inflation and 5% unemployment?

There is only one way I can make sense of Krugman’s argument here.  He must believe that in order to stimulate the economy the Fed would have to create high inflation expectations.  Indeed Krugman has made that claim, but I have never seen him attach a specific number to the phrase “high inflation.”  In his blog he did cite some Taylor rule studies that suggest 5-6% expected inflation is needed to depress real interest rate to the appropriate level.  Perhaps he thinks such high inflation expectations are needed for monetary policy to create a robust recovery.  But if he does, he is wrong.  The SRAS is very flat right now and thus 6% expected NGDP growth would probably break down into roughly 2% expected inflation and 4% expected RGDP growth.

Now let’s return to the issue of temporary currency injections.  He loves to poke fun at the right wing Cassandras who keep warning that high inflation is just around the corner.  Krugman is right that those fears are unfounded.  But what annoys me is the way he misuses history to make his point.  Krugman points to the fact that the monetary base in the US rose sharply in the early 1930s, and again in Japan in the late 1990s and 2000s, and yet prices continued to fall.  So this shows that rapid money growth does not necessarily foreshadow high inflation.  So far so good.  But then he pushes things too far.  He suggested that these real world cases that show not just that monetary injections need not be followed by high inflation, but also that monetary injections at zero interest rates cannot produce high inflation.  He argued that these examples show the futility of trying to escape from a liquidity trap by swapping zero interest cash for zero interest T-bills.  But that is not at all what they show.

Even when the economy is not in a liquidity trap, the monetary base is almost always endogenous in the very short run.  Thus the Fed might set an interest rate target, and the base moves as necessary to meet that target.  Or they might set an inflation target, and again the base adjusts to meet that target.  Very few economists view the monetary base as an end in itself, but rather a means to an end.  Even in a liquidity trap there is nothing stopping central banks from setting an inflation target and using ordinary OMOs in T-bills to hit that target.  Obviously it is possible that they would run out of T-bills at some point, and have to buy T-notes, but this problem is unlikely to occur if they charge a penalty rate on excess reserves.

Of course there is one thing that could prevent this from working, a lack of credibility.  Now we are back to the “expectations trap.”  But why should we worry about expectations traps?  If the Fed announces a policy of targeting inflation at 2% or NGDP growth at 5%, why wouldn’t the public believe them?  After all, that’s pretty much what they have been doing for the past 25 years.

I wish that in the LSE talk Krugman had emphasized this expectations trap aspect of his 1998 paper, at least then I would have known what I was up against.  Instead he drifted over into an entirely different argument, suggesting that the central bank wasn’t willing to set an explicit target.  Well yes, if the central bank tries to fail, then they may well fail.  But if Krugman were to make clear to his audience that that was his argument, they might be a bit more outraged at the world’s major central banks than he is.  Instead he drifted off into little asides about how absurd it was to swap zero interest cash for zero interest T-bills, as if this observation could explain a liquidity trap.  That was Keynes’ 1936 argument, in 1998 Krugman showed that zero rates would not constrain monetary policy without an “expectations trap.”  I wish he would stick to that argument.

To summarize (finally!), if Krugman thinks high inflation expectations are necessary to generate a robust recovery he is wrong.  If he thinks only 2% inflation, or 3% inflation at most is necessary, but that that sort of inflation is politically unacceptable, then I think he is mistaken.  And even if inflation targeting is currently politically unacceptable, if he were to explain to people like Congressman Barney Frank (who has opposed inflation targeting) why it was essential for monetary policy to be effective, it might go a long way to changing the political climate surrounding this issue.  Instead, his LSE lecture simply served to reinforce common prejudices surrounding this issue.  Do you want proof?  Here is how The Economist summarized his views.  Call it the $3 trillion dollar non sequitor:

It also renders conventional monetary policy impotent, as the interest rate that prevents too much saving is below zero.

That creates a role for fiscal policy.

Actually it doesn’t.  It means unconventional monetary policy should be employed until the expected NGDP growth rate equals the policy target.  What makes this passage so bizarre is that fact that they actually included the term “conventional monetary policy.”  Doesn’t the dismissal of that option suggest that we might want to examine “unconventional monetary policy?”  You will not find the latter term mentioned or even alluded to in the article.  Not once.  Why do I call this a $3 trillion dollar non sequitur?  That’s my guesstimate of the budget cost (worldwide) from the assumption that a failure of conventional monetary policy naturally opens the door to fiscal stimulus.  A pretty costly error.

One final point on liquidity traps.  I do understand that a penalty rate on excess reserves may not eliminate a liquidity trap.  Krugman said he was told that when Japanese rates hit zero the only consumer goods that seemed to be selling were safes.  But when you think about it that way, isn’t it surprising that we have not seen something similar here?  We are also supposedly in a liquidity trap—why aren’t Americans hoarding large quantities of currency?  I find it ironic to think about all the criticism the BOJ got from elite salt water economists for not trying more aggressively to stimulate the economy.  But at least they pushed enough cash into the banking system where a lot spilled out into cash held by the public.  The Fed hasn’t even gone that far.  They are so frightened that the QE might cause inflation, that they “confessed” (Robert Hall’s term) that they adopted the interest on reserves program so that banks would hold on to the reserves.

2.  Paul Krugman vs. Amity Shlaes

I don’t recall his exact words, but Krugman seemed to have a very low opinion of the work of Amity Shlaes.  He specifically ridiculed her claim that FDR’s high wage policy slowed the recovery and extended the Depression.  I confess that I haven’t read her book yet (it is on my desk and will be read in the next few weeks.)  But I will defend her anyway.  My hunch is that her appraisal of FDR’s wage policy will be less harsh than mine.

Krugman used a peculiar AS/AD model that he claimed applied to a liquidity trap.  In his model both the AS and AD curves were upward-sloping, with the AD curve being somewhat steeper.  If you draw it on a piece of paper, you will see that in that case a decrease in AS would actually cause both prices and output to increase.  So on a sheet of paper FDR’s high wage policy might have boosted output, but what about in real life?  No evidence was provided.  As we will see, it’s easy to see why.

Let’s have a contest.  I claim there is no theoretical claim by any major economist that is so clearly refuted by the evidence, as is Krugman’s claim that FDR’s high wage policy didn’t hurt the economy, and even might have helped the recovery.  For my evidence, check out this data from a very early post, which some of my more recent readers might not have noticed:

There were actually five wage shocks, four of which are easily dated.  As part of the National Industrial Recovery Act, FDR ordered an across the board 20% hourly wage increase in late July 1933, and then further increases in the spring of 1934.  At the same time the workweek was reduced about 20%.  The NIRA was declared unconstitutional in 1935, but a minimum wage was instituted in November 1938, and raised a year later. To say the IP data is bad for the Krugman interpretation would be an understatement.  These numbers are horrendous:

Table 12.2: Four month (nonannualized!) growth rates for industrial production

Before        After
July   1933 wage shock   +57.4%     -18.8%
May   1934 wage shock   +11.9%    -15.0%
Nov.  1938 wage shock   +15.8%     +2.5%
Nov.  1939 wage shock   +16.0%     -6.5%

You’ll notice that I left out the fifth wage shock, but it’s no better for Krugman’s view, just messier.  Historians argue that the huge union drives of late 1936 and 1937 were due to both the Wagner Act and FDR’s massive election victory.  Whatever the cause of the union gains, they led to rapid wage increases in late 1936 and much of 1937.  This time, monthly industrial production did not fall immediately, as prices were also rising fast in late 1936 and early 1937.  But when prices stopped rising industrial production began falling sharply under the burden of high wages.

In fact, there is even more evidence against the view that the high wage policy helped.  In the mid-1990s Stephen Silver and I did a study of wage cyclicality in the 1920s and the 1930s.  In the 1920s a regression of IP on the WPI and nominal wages shows the WPI is strongly (and positively) correlated with industrial production, whereas nominal wages are essentially uncorrelated.  In the 1930s the WPI is again strongly correlated with IP.  Only this time nominal wages also show up highly significant—with a negative sign.  Then we looked for a “break point,” where nominal wages suddenly became strongly and negative correlated with industrial production.  The break point turned out to be 1933.

Everyone is inspired by stories like “Rocky” where the underdog beats the seasoned pro.  Usually Hollywood drags it out for 15 rounds to build up suspense.  If there was a debate between Krugman and Shlaes on the impact of high wages on the recovery it would be over in one round.  A TKO for Ms. Shlaes.

I don’t like upward sloping AD curves.  I don’t even like the standard textbook AD curve.  The AD curve should be a rectangular hyperbola.  That is, it should be drawn as a constant level of  . . .  you guessed it, NGDP.



21 Responses to “The Lionel Robbins lectures”

  1. Gravatar of Lord Lord
    16. June 2009 at 22:03

    Industrial production isn’t gdp. All this shows is flows were redirected. Try again.

    An alternative proposal to trying to increase credit would be direct flat cash payments to citizens allowing them to reduce debt or spend as they see fit. This could counteract the direct impact of credit only being available for those who don’t need or want it and allow deleveraging, printing not borrowing it until growth returns, attempting to maintain ngdp. How does that sound?

  2. Gravatar of Bill Woolsey Bill Woolsey
    17. June 2009 at 03:01


    From Krugman’s point of view, elite university free market economists are the enemy. They are the enemy because they favor a society with low taxes and limited government.

    Understand that Krugman’s expectations model was just a intellectual ammunition in the battle. It was a clever weapon becaues it used the enemies tools against them. Intertemporal optimization with market clearing. It could not be dismissed by the enemy as unscientific.

    And, _your_ key problem is that you continue to fail to see the killer to the free market “ideologues.” Using monetary policy to get out of the recession requires inflation. And, inflation is _BAD_.

    So, you will never convince Krguman that we should not rapidly move the U.S. to Swedish levels of government spending. He supporta that. Coming up with reasons why monetary policy won’t work so that we need more government spending if we are not to spend a decade in recession–that is the point.

    Why should Krugman want inflation? It was hatred of inflation by the American people that destroyed the aftermath of Camelot and the Great Society. Here we were, making progress towards the holy grail of Sweden, and then, inflation began rising. We had Carter… and then, worse, Reagan.

    And, we don’t need inflation. Just massive govenment spending. And, yes, really high taxes in the future. Of course.

    Now, are _you_ going to convince those who support small government and low taxes (like yourself) that the Fed should increase base money by more? You keep on forgetting that free market economists _don’t like inflation_. Every time you argue that the solution is to raise inflation expectations, you loose support. Your reductio arguments that suggest really high inflation are just suicide.

    Yes, I did read above, about what you said about the flat SRAS curve. But don’t you see the contradiction? Aggregate demand is supposed to rise because of lower real interest rates because of higher expected inflation. But, aggregate demand will rise and there won’t be higher inflation becasue the short run aggregate supply curve is flat. People will expect inflation sufficiently high so that the real interest rate will be sufficiently low so that they will spend enough to raise nominal income back to target, but this won’t cause the inflation that they expect?

    Now, there _is_ an argument here. And you sometimes make it. And that is something like that if people expect expenditure to rise, and that makes real income rise, then this will raise the natural interest rate, and so the current low nominal interest rates will now be sufficiently low that spending will rise. Nothing there about higher expected inflation cuasing lower real interest rates today.

    Also, if people actually expected deflation (which they don’t now, though maybe they did last fall), then breaking the deflationary expecations would lower real interest rates. (nominal interest rates are low, but real interest rates were high if poeple expected delfation.)

    Sorry, but words matter. And less deflation, more inflation.. no, it isn’t the same.

    While your expertise on the Depression is great, you need to be careful. I think reversing the deflation of the first have of 30 would have been a good thing. After three years, I guess it was a good thing. But… we are not in that situation.

    But, you alwaya are arguing like we are. So, the rapid inflation caused by the Roosevelt administration is taken as a good thing. Well maybe back then it was a necessary evil. (It was reversing deflation.) Yet we don’t have any significant deflation to reverse now.

    So, the intellectual tools of the interest rate targeters, that use inflation expectations are just not appropropriate. Aggergate demand depends on real interst rates. Usually, we manipulate that by changing nominal interest rates. Expected inflation can impact real interest rates, and so, generating inflation expectations can impact aggregate demand.

    Well, if we have any kind of true inflation target (even 2%) then that means you cannot manipulate inflation expecations to lower real intersest rates too much. It is obvious when you favor price level stability, I think. But even if you go with the central bank consensus of 2%, you are left with claiming we can get out of the liquidity trap by creating inflationary expectations of however high necessary.. oh, but it can’t be above 2%.. that is the target.

    The short run as curve bandage on that argument won’t work. You have to just give up the inflationary expecations/real interest rate transmission mechanism.

    In other words, sotck to nominal income targeting. And forget the inflationary expectations.

  3. Gravatar of Alan Alan
    17. June 2009 at 04:13

    Hi Scott

    Great post once again.

    The Fed seems (at least by their actions, even if not by their words) to have this future inflation ‘time bomb’ view in their minds whenever they talk about expanding QE. It’s probably why they’ve focused most of their ‘easing’ on credit easing (i.e. intermediary of last resort) rather than conventional QE.

    There was a recent NYT article about the problems the Fed’s having in this role as (e.g. the Fed’s getting a lot of pressure to add leisure boat firms and the like to their list of eligible repo collateral). I find it hard to believe that the Fed finds this job easier than engaging in conventional QE by, for example, penalising reserves. Also, the added political scrutiny created by credit easing (compared to QE) also seems at odds with the Fed’s desire to regain their pre-crisis level of independence.

    More and more, I’m getting the impression that the desire, by both the Fed or the Treasury, to inflate their way out of this crisis is not there. Recent announcements by Geithner on the need to bring the budget deficit down through a combination of higher taxes (e.g. a possible VAT) and lower expenditures, seems to mesh with the Fed’s lack of interest in unconventional policy. We can expect Ricardian equivalence to do the rest!

  4. Gravatar of ssumner ssumner
    17. June 2009 at 04:18

    Lord, That’s right. In the midst of the depression when the steel and auto factories went back to work they didn’t hire unemployment industrial workers, rather they took EMPLOYED doctors, lawyers, accountants, teachers, barbers, farmers, etc. Thus non-industrial production fell to offset the growth of industrial production. Why didn’t I think of that?

    Bill, I am saying that the Wicksellian equilibrium real interest rate at the optimal NGDP growth rate is more than negative 5 or 6%. And I do think I can show that is true without any reference to interest rate theory. All I need is the SRAS. Here’s my idea. Ask Krugman what sort of NGDP growth he’d like to see over the next 12 months. Say he answers 6%.
    We all agree that we are likely to fall short on current policy. We all agree that 5 or 6% expected inflation would cause us to overshoot that expected NGDP target. So there obviously must be some in between monetary policy that will produce 6% expected NGDP growth over 12 months. And given the current flatness of the SRAS, that expected NGDP growth rate means something like 2% expected inflation. Where is my logic wrong?

    You may be right in your cynical appraisal of Krugman, but that doesn’t affect me at all. I am not trying to change Krugman’s mind, I am trying to change the minds of those who read Krugman, and think he is saying that monetary policy has run out of ammunition. People like the reporters for the Economist. (BTW, they have linked to me three times, so I know some of them look at this blog.)

    I think Krugman really does believe his 1998 model. When you read those papers it seems very clear (to me) that he really believes what he is writing. I just think that he is confused about the political implications of his model. He is too defeatist about monetary policy. Even in his recent blog he says things like ‘we aren’t ready for inflation targeting yet.’ But that implies he thinks if we get desperate enough we will become ready. My point is that we are already desperate enough, it just that not enough people know that it can work.

    Regarding inflation, by 1933 the WPI had fallen in half. I have no idea what was optimal during that period. But here is what happened in the four months before the wage shock:

    WPI rose 14%, broader indices less
    IP rose 57%. broader RGDP less.

    That seems like nice NGDP growth, and a nice P/Y split. We both favor level targeting, but when things have been bad so long it is hard to know where to start from, as you correctly point out. I keep bringing this up because it shows that monetary policy can work in a depression/banking crisis, which you know, but almost no one else does know.

    I agree with your last point criticizing inflation targeting. As you point out it may not get us the real interest rate we need. But my point was slightly different. The current inflation forecast is well below two percent for the next couple of years. In my view to get the expected inflation rate up to 2%, the Fed would have to get expected NGDP growth much, much higher. I admit that’s a judgment call. But I think even with 2% inflation targeting, we’d be much better off right now. But yes, I agree NGDP is the right target.

  5. Gravatar of ssumner ssumner
    17. June 2009 at 04:25

    Lord, I forgot the second part of your comment. I am not talking about credit. I am talking about monetary policy. If we give money to citizens where does it come from? If the government borrows it then the economy as a whole does not see it’s balance sheets improve. Most assets in one place are offset by more debts elsewhere. The key is to get more money into the economy, so that nominal values can rise.

    I apologize if my first answer was too sarcastic, but your dismissal of the IP data just seemed incredible to me. BTW, this is exactly the sort of data people like Krugman also rely on. Indeed in his blog he used the even less representative “monthly investment data” from 1933.

  6. Gravatar of Joe Joe
    17. June 2009 at 04:32


    I think you agree with Krugman more than disagree. The disagreement really come from the fact that Krugman feels that fiscal policy is the only relevant option, while you feel that monetary policy is the only relevant option. My hunch is you are both right. Monetary policy should create the expectation of future growth and fiscal policy should be used by governments to make long term investments that have no possible known payoff..i.e. the government is the only player than can have a very long term view of its investments; the interstate system always come to mind..I am sure someone did a cost benefit in the 40’s, but who could have predicted that 70 years later it is still an integral part of our economy. The same goes for R and D; the government can give money purely for research with no currently known application; then 30 years later it is the backbone of treating major diseases…protease inhibitors discovered in and odd bacteria; restriction enzymes that are now used in the production of many biological therapeutics.)
    Also, I feel you are too open minded to be completely sold on Shlaes book; she really butchers history(although she is right about the wage act). When NYU Stern took her on staff they lost my Alumni contributions.


  7. Gravatar of Alex Alex
    17. June 2009 at 04:51

    “I don’t like upward sloping AD curves. I don’t even like the standard textbook AD curve. The AD curve should be a rectangular hyperbola. That is, it should be drawn as a constant level of . . . you guessed it, NGDP.”

    Let me guess… something like MV=PY?

  8. Gravatar of Giedrius Giedrius
    17. June 2009 at 05:00

    I think you meant “non sequitur”. Could you also tell how you arrived at $3 trillion figure?

  9. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    17. June 2009 at 08:50

    ‘…the interstate system always come to mind..I am sure someone did a cost benefit in the 40’s, but who could have predicted that 70 years later it is still an integral part of our economy.’


  10. Gravatar of azmyth azmyth
    17. June 2009 at 09:04

    A world with a real liquidity trap would be absolutely wonderful. The government could monetize the entire debt and there would be no inflation. If you really believe in liquidity traps, you should agree with debt monetization when we’re in one. Can the concept of liquidity trap survive the reductio ad absurdam critique? How do you think Krugman would respond to the suggestion to start monetizing the debt?
    Krugman does not believe in scarcity. When you don’t believe in scarcity, you come to really bizarre conclusions, like that 9/11 was good for the economy and decreasing AS causing prices and output to increase.

  11. Gravatar of Lord Lord
    17. June 2009 at 10:23

    The vague assertion was wage policy prolonged the depression. I assume this means gdp was lower than otherwise. It could also mean employment was lower than otherwise but that is not typically the measure of depression. It is quite reasonable it kept employment lower than otherwise. It just does not imply gdp was lower than otherwise. Industrial production grew less rapidly. Consumption likely grew more rapidly. The effect on gdp is unclear. If workers were less likely to save it and pay down debt, it could well have boosted gdp even while leaving employment low. (Not that I believe this but it is possible.)

    That is why I said print it, and announce to everyone it is being printed. Have it come from out of thin air. My point here is credit based money creation doesn’t work well when would be borrowers are uncreditworthy and undeserving of credit, say due to diminished collateral values or unemployment, and those that could borrow remain unwilling to do so because of concern over their own debt and future. I have my doubts about the possibilities of current money production processes, though asset sales are certainly one way.

    Others are concerned it could spook investors and lead to a dollar crash and inflation. It is tough managing everyones expectations.

  12. Gravatar of Joe Joe
    17. June 2009 at 11:28

    Maybe a printing press financed money gift? That should raise NGDP expectations no?

  13. Gravatar of Current Current
    18. June 2009 at 00:55

    azmyth: Modgiliani proved in the 40s that the “liquidity trap” cannot survive this problem.

    Once contracts have ended and forward planning is involved the situation becomes very different. As Scott has often mentioned.

  14. Gravatar of ssumner ssumner
    18. June 2009 at 04:37

    Joe, You are partly right, but much of the stimulus is not about worthy infrastructure/research projects. I.e. it would be much smaller if monetary policy had made the recession milder.
    I will review her book in a future post.

    Alex, Yes, AD=MV=PY

    Giedrius My $3 trillion guesstimate was based on:

    1. 787 billion US stimulus plus enough automatic stabilizers to push it well over a trillion. In other words the 2 trillion US deficit would be much less than a trillion w/o the severe recession. And didn’t even include automatic stabilizers for the out years. So its a low estimate.

    2. Other countries also have fiscal stimulus, and automatic stablizers. Guesstimated their total excess deficits at 1.5 to 2.0 trillion.

    It’s just a ballpark estimate. Obviously I am also including the ECB as part of the policy error. The rest of the world were mostly victims.

    Patrick, Good point.

    Azmyth, I actually made that point in a paper I published a few years back, and also used the term “reductio ad absurdum.”

    Krugman admits that the liquidity trap would eventually end. He worries that to prevent hyperinflation central banks would then have to sell off most of the government debt, and with inflation now higher, it would be sold at a big loss. I have argued that this isn’t a big problem in our current situation, for various reasons.

    Lord, You don’t want to divide it up between consumption and IP, the relevant categories are IP and services.
    I just don’t find that hypothetical at all plausible, but it is theoretically possible. But I don’t think that would be Krugman’s defense, it’s a real longshot.

    On money creation we agree, but you also need a target.

    Joe, Regarding the helicopter drop of cash, see my response to azmyth. It could be done, but might overshoot in the long run.

    Current, Thanks, I didn’t know that about Modigliani.

  15. Gravatar of Current Current
    18. June 2009 at 05:00

    See page 132 of the book “The Critics of Keynesian Economics”

    Although that book contains lots of heterodox criticisms the one by Modigliani is really quite orthodox.

    I haven’t read all of this paper and I don’t understand it all.

  16. Gravatar of ssumner ssumner
    19. June 2009 at 04:23

    Current, Yes, that is his famous IS-LM model.

  17. Gravatar of Current Current
    19. June 2009 at 04:59

    Yes, I’ve read it now. It’s interesting to see Keynesian economics in it’s original form not mediated by a textbook author.

    It does contain this argument you are making here that any liquidity trap would have to be a very special case.

  18. Gravatar of ssumner ssumner
    20. June 2009 at 08:09

    Current, I didn’t read it all. Do you remember whether the special case was because you needed sticky prices forever? Or did you need the zero interest rates to be expected to last forever (which was Krugman’s point.)

  19. Gravatar of Current Current
    22. June 2009 at 06:04

    Modigilani’s models indicate to him that it is sticky wages that allow underemployment equilibria, not the liquidity preference. He gives a Keynesian model of demand for money and a “Quantity theory” one. (I think his idea of Quantity theory is a bit unusual). He shows that both give the same results if sticky wages are assumed.

    “This last result deserves closer consideration. It is usually considered as one of the most important achievements of the Keynesian theory that it explains the consistency of economic equilibrium with the presence of involuntary unemployment. It is, however, not sufficiently recognized that, except in a limiting case to be considered later, this result is due entirely to the assumption of “rigid wages” and not to the Keynesian liquidity preference. Systems with rigid wages share the common property that the equilibrium value of the “real” variables is determined essentially by monetary conditions rather than by “real” factors (e.g., quantity and efficiency of existing equipment, relative preference for earning and leisure, etc.). The monetary conditions are sufficient to determine money income and, under fixed wages and given technical conditions, to each money income there corresponds a definite equilibrium level of employment. This equilibrium level does not tend to coincide with full employment except by mere chance, since there is no economic mechanism that insures this coincidence. There may be unemployment in the sense that more people would be willing to work at the current real wage rate than are actually employed; but in a free capitalistic economy production is guided by prices and not by desires and since the money wage rate is rigid, this desire fails to be translated into an economic stimulus.”

    He deals later with the limiting case. His criticism is a bit different from yours. He points out that positing a liquidity trap with flexible employment is analytically troublesome for his model:
    There is one case in which the Keynesian theory of liquidity preference is sufficient by itself to explain the existence of underemployment equilibrium without starting out with the assumption of rigid wages. We have seen (Section 5) that, since securities are inferior to money as a form of holding assets, there must be some positive level of the rate of interest (previously denoted by r”) at which the demand for money becomes infinitely elastic or practically so. We have the Keynesian case when the “full-employment equilibrium rate of interest” is less than r”. Whenever this situation materializes, the very mechanism that tends to bring about full-employment equilibrium in a system with “flexible” wages breaks down, since there is no possible level of the money wage rate and price level that can establish full-employment equilibrium. From the analytical point of view the situation is characterized by the fact that we must add to our system a new equation, namely r””r”. The system is therefore overdetermined since we have 9 equations to determine only 8 unknowns.”

    He goes on to say that if wages are flexible then they will go down until they are no longer flexible. Then the earlier criticism will come back into play.

    Elsewhere in the paper he points out that “the IS curve, unlike the LM curve, describes not instantaneous relationships but only possible positions of long-run equilibrium.” And that it is the long-run equilibrium rate of interest that must be at the level which provokes a liquidity trap. AFAIK he doesn’t though notice that this isn’t really plausible.

    It’s all very old fashioned and uses very simple models of expectations. As you might expect I don’t like it very much.

  20. Gravatar of azmyth azmyth
    22. June 2009 at 18:22

    Scott: “He worries that to prevent hyperinflation central banks would then have to sell off most of the government debt, and with inflation now higher, it would be sold at a big loss.”
    That’s an interesting point and it took me some thinking to figure out a good answer. There are 2 issues here: What is the best way to end a liquidity trap, and once you have done so, what is the best way to prevent hyperinflation. These questions can be separated and don’t need to be answered by the same agency. The Fed’s bonds never need to be resold on the open market if interest rates are unfavorable. Congress might decide to tax and burn to get rid of the excess liquidity, or required reserves could be dramatically increased (even above 100%). Krugman favors fiscal stimulus, but eventually that will produce inflation if it is not matched by increased taxes. Keep in mind that according to his model, bonds are perfect substitutes for currency in a liquidity trap, so printing more bonds increases the money supply just as quickly as printing cash. I assume that he would argue to increase taxes once the liquidity trap was gone, but the Fed could just as easily soak up the liquidity instead of the IRS. The Fed could also buy assets that are unlikely to decline in value during an inflationary period, such as inflation indexed bonds or real estate. That would send a clear signal to the market that they intend to inflate, which would help the credibility issue.

  21. Gravatar of ssumner ssumner
    23. June 2009 at 06:21

    Current, Thanks, It sounds like he has some valid insights, but still doesn’t fully understand how fragile the Keynesian model really is.

    Azmyth, I agree there are many options for the Fed, but we need to be careful as they all are costly:

    1. Sell off bonds at a loss
    2. Pay interest to encourage banks to hold reserves
    3. Raise reserve requirements, which is a tax on banks

    I agree that the fiscal option he proposes is most likely far more costly than any of the three items above, so I think we are on the same page.

    Your last idea is a good one. I proposed something similar way back around late February or early March. The post has “insider trading ” in the title. The idea is that the Fed can profit if it knows that it intends to carry through with a low inflation policy, (or a high inflation policy.)
    James Hamilton also discussed the advantages of the Fed buying indexed bonds in his blog.

Leave a Reply