The Keynesian theory is wrong

So are all the other macro theories, including my own.  (How’s that for a misleading title.)  They all abstract from reality.  But the Keynesian theory is wrong in very important ways that lead to bad policy advice.  Before I discuss a recent example, let me sketch out the basic Keynesian view of monetary policy:

1.  The Fed adopts an expansionary policy.

2.  Interest rates decline.

3.  Aggregate demand increases

4.  Output increases

5.  Inflation rises when the economy nears the natural rate

If you are in a liquidity trap, Keynesians would still adhere to this basic approach, although they would allow for the possibility that monetary policymakers could lower real rates through signals about future intentions to inflate.

I have two problems with this basic approach.  First, they have the transmission mechanism wrong.  It focuses too much on the risk-free interest rate on T-securities.  An expansionary monetary policy increases long run NGDP for reasons that have nothing to do with interest rates.  (Recall that interest rates are only affected in the short run.)  Instead, more money increases expected long run NGDP via the excess cash balance mechanism.  This increase in expected future NGDP raises current asset prices (stocks, commodities, real estate, etc), which then leads to more current AD.  My other problem with the Keynesian approach is that in the real world an increase in aggregate demand will tend to increase prices and output, even if the economy is not at full employment.

In a recent post Paul Krugman uses this faulty approach, and reaches the conclusion that the Fed cannot boost inflation expectations to around 2%; it’s roughly 4% or nothing (where ‘nothing’ might be the current 1% rate.)

But how do you get inflation? Only by having a full-employment economy. So there’s a reserve relationship “” a Phillips curve “” which might look like this:

Now, take these together. If the public believed we would have 2 percent inflation, according to the first figure this would lead to 6 percent unemployment. But according to the second figure, 6 percent unemployment would lead to only 1 percent inflation. So a Fed commitment to achieve 2 percent inflation wouldn’t be credible “” because even if believed, it wouldn’t deliver. And so the whole edifice would collapse.

A commitment to, say, 4 percent inflation, on the other hand, could work: it’s enough to produce full employment, which in turn is enough to validate the public’s expectations.

So there’s a real threshold effect here: you have to have a sufficiently high inflation target even to have a chance of breaking out of a liquidity trap.

In the past I’ve criticized Krugman’s view for being inconsistent with the relatively flat SRAS, but now I can see that he understands that issue–he simply doesn’t think there is any equilibrium point around 2% inflation.

I have two problems here.  First, inflation can occur without full employment, so his premise is incorrect.  As an aside, many conservatives are too quick to dismiss Krugman, and would say something like “He doesn’t even remember the 1970s stagflation!!!!”  Actually, Krugman doesn’t mean inflation, he means increases in inflation.  Of course you might point out that both inflation and unemployment rose in 1974.  But then he’d say “increases in inflation due to demand shocks.”  So the argument isn’t as silly as it looks, but it’s still wrong.

In 1933-34 we had roughly 20% wholesale price inflation and 10% consumer price inflation.  We know this inflation occurred as a result of demand shocks, because the weekly changes in the WPI were closely correlated with the value of the dollar—which was being manipulated by FDR.  And during this period unemployment was near an all-time high, roughly 25%.  You really can’t get a better test of the Keynesian theory that inflation is caused by pushing up against full employment, and it failed miserably.

The bigger problem with his argument is that he assumes lower real interest rates are the factor causing the increase in AD.  That’s why he’s pessimistic; his model tells him that real interest rates must get very low to provide the needed boost to the economy.  In fact, expansionary monetary policy affects all sorts of other variables, including stock prices and exchange rates.  How do I know?  Because I’ve just seen it happen with my own eyes, and it happened at the zero rate bound.  Rumors of Fed easing raised stock prices, lowered the dollar, and lowered 5 year bond yields.  Nick Rowe has a nice new post on how higher stock prices raise the Tobin q (ratio of market to book value of stocks) and thus give corporations an incentive to invest more.  And since human capital is one of the most important corporate “investments,” a higher Tobin q should also lead to more hiring.

In a recent post, Ryan Avent made the following comment:

Another new criticism is that the latest American jobs figures show that more QE isn’t necessary. This complaint is also extremely misguided. For one thing, as my colleague notes, the good October jobs numbers were no doubt boosted by anticipation of the Fed’s new easing programme, which has been lifting share prices and inflation expectations and reducing interest rates and the dollar since late August.

Avent’s colleague also noted that:

Similarly, the growth is inconsistent with the contribution of fiscal stimulus which has turned negative lately.

Of course we can’t know for sure what led to the increased hiring in October, but surely the rumors of Fed easing that pushed up stock prices in September and October must have made companies at least slightly more confident that we would not have a double dip recession.

Like Keynes, Krugman loves to mock right wing economists who have their head in the sand about the reality of involuntary unemployment, just because their models say it shouldn’t happen.  But it seems to me that Krugman keeps insisting that QE can’t work (unless it raises inflation expectations to around 4%) even as evidence piles up that it has already had a modest effect on asset prices and growth.  He’s still right that the Fed hasn’t done enough, but the evidence strongly suggests that if they do a bit more, we’ll get a bit better results.  It’s not all or nothing.

PS.  Nick Rowe has a post that explains Krugman’s model.


Tags:

 
 
 

36 Responses to “The Keynesian theory is wrong”

  1. Gravatar of David Pearson David Pearson
    6. November 2010 at 07:38

    Scott,

    August private payrolls were revised upward to 126k. October came in at 150k. Given the margin of error, one cannot conclude that hiring rose from August to September, especially since in September we saw the first Household survey job loss in months.

    So the uptick in hiring was first observable concurrent with the August lows in the stock market. At that time, stocks had declined almost 17% from their May highs. The Bernanke Jackson Hole speech was at the end of the month, so anticipation of QE can hardly have accounted for that uptick.

    In theory, it seems plausible that higher stock prices and anticipated QE had an impact on hiring. The data, however, does not support that thesis.

  2. Gravatar of Shane Shane
    6. November 2010 at 08:43

    What would Bernanke have to have done back in 2008 to avoid the liquidity trap and bloated balance sheet? Other than simply reframing his actions as some sort of level targeting, is there any actual difference in the mechanism the Fed would have had to use to create credibility about being able to hit its target? Or would it simply have been enough to pay negative IOR and then just use ordinary open market operations? Or would a level target require some kind of unconventional QE or even qualitative easing?

  3. Gravatar of Morgan Warstler Morgan Warstler
    6. November 2010 at 09:12

    Stocks Up!
    Gold Up!
    Oil Up!
    Dollar Down!
    Jobs Down!
    Productivity Up!

    One of these things is not like the others…. Productivity.

    Is there any chance we’ve reached a pickle where even if foreign demand increases (weak dollar) companies are bound and determined to not hire, and sock away profits?

    It seems to me that with… with enough Productivity Up! without printing money and we can have:

    Stocks Up!
    Gold Down!
    Oil Up! – we’re running out after all
    Dollar Up!
    Jobs Up!

    —–

    Note: I was kind of excited, because I thought this was Scott lampooning Krugman for Fiscal Crimes Against Nature.

  4. Gravatar of Morgan Warstler Morgan Warstler
    6. November 2010 at 09:12

    Stocks Up!
    Gold Up!
    Oil Up!
    Dollar Down!
    Jobs Down!
    Productivity Up!

    One of these things is not like the others…. Productivity.

    Is there any chance we’ve reached a pickle where even if foreign demand increases (weak dollar) companies are bound and determined to not hire, and sock away profits?

    It seems to me that with… with enough Productivity Up! without printing money and we can have:

    Stocks Up!
    Gold Down!
    Oil Up! – we’re running out after all
    Dollar Up!
    Jobs Up!

    —–

    Note: I was kind of excited, because I thought this was Scott lampooning Krugman for Fiscal Crimes Against Nature.

  5. Gravatar of jkr jkr
    6. November 2010 at 09:27

    Mr. Sumner,

    How would you respond to John Hussman’s recent commentary,

    Bernanke Leaps Into a Liquidity Trap

    http://www.hussmanfunds.com/wmc/wmc101025.htm

  6. Gravatar of marcus nunes marcus nunes
    6. November 2010 at 09:31

    Scott
    What this post is saying in effect is: Let´s stop talking about inflation (and interest rates, Taylor Rules, etc.). Focus on NGDP and try to keep it stable and “on target”!

  7. Gravatar of scott sumner scott sumner
    6. November 2010 at 10:11

    David, It’s not clear whether the private or total jobs number is more meaningful. The total number was the largest in many months, and at a time when fiscal stimulus is slowing. That suggests that monetary stimulus may have played a role. But I agree that one month is not enough, there is too much statistical noise. So you may be right. We’ll need confirmation from several more months to figure out whether the surge is real. I still think the recovery will be weak, but I also think a double dip recession is now far less likely.

    The fiscal policy advocates are already on record saying fiscal policy is now a drag, so obviously any demand increases that do occur will be monetary, or some third factor.

    Shane, They should have cut interest rates much more aggressively during 2008. We now know that the monthly low for NGDP was around December 2008. That means the Fed had not even cut rates to zero by the time the decline in NGDP was ALREADY OVER. “Liquidity traps had nothing to do with the collapse in NGDP.

    But level targeting was the key. And of course they should not have paid interest on reserves.

    Morgan, No, if NGDP rises fast, so will jobs.

    jkr, People keep sending me that. It’s warmed over 1938 Keynesian ‘pushing on a string’ nonsense. He doesn’t understand that the negative correlation between M and V proves nothing, as the Fed has kept NGDP growth near 5% for most of the past 25 years. M is endogenous.

    Marcus, Yes, and if we insist in inflation targeting I don’t think we need 4% inflation. The SRAS is flat enough that 2% to 3% would promote a brisk recovery.

  8. Gravatar of marcus nunes marcus nunes
    6. November 2010 at 10:38

    Scott
    And the inflation confusion keeps coming up!
    http://online.wsj.com/article/SB10001424052748704366704575598443920302492.html?mod=WSJ_hp_MIDDLETopStories

  9. Gravatar of Yosef Yosef
    6. November 2010 at 10:43

    Professor Sumner,

    You say that you have criticized Krugman’s inconsistencies before, which you have brilliantly, but they are are always with him.

    In an old article Krugman wrote for Slate he said, (http://web.mit.edu/krugman/www/hotdog.html)

    “It is possible for economies to suffer from an overall inadequacy of demand–recessions do happen. However, such slumps are essentially monetary–they come about because people try in the aggregate to hold more cash than there actually is in circulation. (That insight is the essence of Keynesian economics.) And they can usually be cured by issuing more money–full stop, end of story.”

    Full stop, end of story. Not “lowering interest rates” not “fiscal expansion”, but the correct solution of issuing more money. Krugman’s true overall inconsistency is his abandonment of his previously well thought out ideas.

  10. Gravatar of Jim Glass Jim Glass
    6. November 2010 at 10:55

    You are kind. Arnold Kling says that in that column Krugman used The Strangest Macro Model Ever.

  11. Gravatar of JTapp JTapp
    6. November 2010 at 11:33

    “First, inflation can occur without full employment, so his premise is incorrect.”
    Krugman’s Phillips curve shows inflation without full employment. I’ve not seen a textbook that illustrates it otherwise. I think his is drawn rather Friedman-esque, vertical in the long-run. (Why am I wrong?) A Principles AD-SRAS-LRAS would also show inflation without full employment.

    RE: Ryan Avent– he has a new article up on international complaints about QE2 where he makes this statement:
    “There are forms of QE that look more like currency manipulation: unsterilised foreign-exchange intervention, for example…but that is not what the Fed is doing. It is simply trying to do to long-term rates what it has already done with short-term rates.”

    That seems to be a pretty false statement. An increase of the money supply, by definition, is an unsterilized foreign exchange intervention (using the monetary approach to BOP here). On the asset side of the Fed’s balance sheet what they are doing is identical to what it would look like if they were buying foreign exchange reserves instead of Treasury bonds… the change in the monetary base would be the same. Am I wrong?

  12. Gravatar of JTapp JTapp
    6. November 2010 at 11:39

    Jim Glass’ Arnold Kling link cleared it up for me, I understand what Krugman is saying now (so long as Kling understood what he was trying to say correctly). I get your critique now. That is “naive” Keynesianism. I think even Samuelson would disagree with Krugman here.

  13. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. November 2010 at 12:01

    Scott wrote:
    “In 1933-34 we had roughly 20% wholesale price inflation and 10% consumer price inflation. We know this inflation occurred as a result of demand shocks, because the weekly changes in the WPI were closely correlated with the value of the dollar””which was being manipulated by FDR. And during this period unemployment was near an all-time high, roughly 25%. You really can’t get a better test of the Keynesian theory that inflation is caused by pushing up against full employment, and it failed miserably.”

    I agree but I want to clarify this somewhat.

    It’s important to distinguish between two types of price expectations. In one case expected inflation rates are assumed to depend primarily on past inflation rates. This model does a good job of describing inflation in the United States for much of the postwar period and is thus a popular model of inflation in the United States.

    The second case is one where price level expectations are more important. In this case, even in a severe recession, people expect that expansionary central bank policy actions can restore a price level, and this expectation can act as a magnet pulling prices (and output) up even in a very serious recession.

    Analysis of the relationship between prices and unemployment during the 1920s and the Depression indicates that the inflation rate was closely linked to the *change* in the unemployment rate, rather than the *level* of the unemployment rate. That is, when unemployment was rising, prices fell, and when unemployment was falling, prices rose.

    The United States was of course following the gold standard at the onset of the Depression, a policy that can produce a relatively stable price level over long periods. Consequently inflation dynamics today are likely to be very different than they were during the 1920s and 1930s. The United States and other countries generally follow policies aimed at maintaining low, positive inflation rates rather than stable price levels.

    On the one hand, this is an argument that, given the high unemployment rate, inflation is likely to remain low and to continue to decline. On the other hand this is also an argument in favor of price (or NGDP) level targeting.

  14. Gravatar of Morgan Warstler Morgan Warstler
    6. November 2010 at 12:09

    Scott do you have some kind of guess at 5% NGDP growth over X years = 5% unemployment

    Have you really looked at it? How long’s it take to get back to 5%, without you cheating – changing fiscal policy, lowering minimum wage, etc.

    No other variables, 5% NGDP get us 5% in how many years?

  15. Gravatar of Shane Shane
    6. November 2010 at 12:32

    Krugman no doubt helped create the obsessive focus on fiscal policy, which helped continued the slump. Why, though, did he change his emphasis, if not his views? Because he became more partisan over time. Why did he become more partisan? Because of the temporary breakdown in rational discourse during the GWB years. Krugman is just one of the various “transmission mechanisms” by which Bush continues his destruction.

  16. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. November 2010 at 13:32

    Scott, wrote:
    “Of course we can’t know for sure what led to the increased hiring in October, but surely the rumors of Fed easing that pushed up stock prices in September and October must have made companies at least slightly more confident that we would not have a double dip recession.”

    With respect to David Pearson and Scott’s exchange let me add the following:

    We all know that according to the major macroeconomic models that as the fiscal stimulus is withdrawn in the fourth quarter of this year that the effect on AD was forecast to be negative.

    We also have the fact that a modest QE2 has been introduced this very quarter.

    In this context will the effect of QE2 be sufficiently positive to officially render a verdict that fiscal stimulus is completely irrelevant given the idea that the central bank is supposedly always at the helm?

    In my opinion no definitive answer is on the way. Right now fiscal and monetary policy is merely in the process of changing who’s leading the dance. (But did MP allow FP to take the lead for a time? It’s so confusing!) (Not.)

  17. Gravatar of W. Peden W. Peden
    6. November 2010 at 13:44

    Mark A. Sadowski,

    One of the problems seems to be that, while there are cases of monetary stimulus without fiscal stimulus (as with this current package) it’s hard, at least for me, to think of cases with fiscal stimulus with no monetary stimulus.

    Even if there were plenty of cases, fiscal stimulus is not neutral with respect to money, i.e. a government borrowing from banks short-term will increase their reserves.

    Still, if QE2 is even a moderate success then it might finally get some people shutting up about liquidity traps, although thanks to Krugman that term is so vague that it will now probably become immortal.

  18. Gravatar of W. Peden W. Peden
    6. November 2010 at 13:47

    Shane,

    That’s a good way of putting it. Krugman was no saint before 2000, but the bad stuff doesn’t really start until he got aggravated about GWB and the Bush era. He was nowhere near so keen to portray everyone who disagreed with him as stupid before then.

  19. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. November 2010 at 14:07

    W. Peden,
    I agree with you at least in this way:

    I don’t believe that there will be a “double-v”. Thus people will start to scratch their heads and wonder if all we really ever needed was simply more money.

  20. Gravatar of Morgan Warstler Morgan Warstler
    6. November 2010 at 14:46

    I think I’ve told you all this before… but Krugma’s saving grace is his performance in Get Him To The Greek. Netflix it.

    It’s as good as Caddy Shack – over and over for years and years.

  21. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. November 2010 at 14:56

    Morgan,
    That reminds me of some of my less unpleasant tutoring sessions.

    http://nymag.com/daily/intel/2010/06/paul_krugman.html

    But on the other hand, at least, they fear me.

  22. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. November 2010 at 15:01

    Morgan,
    I meant “less pleasant”. But I may assume you understood that.

  23. Gravatar of Felix Felix
    6. November 2010 at 15:02

    I’d like to defend Krugman somewhat. The basic insight is that something strange must happen when expected deflation (defined as negative of inflation) is higher than the real rate, because then money would yield a higher return than real investments, thus the money demand will explode, i.e. it will no longer be determined by use of money for transactions needs but by the desire to save. Krugman’s mistake is that he does not take into account that the real interest rate may well be higher in a scenario of higher inflation, and thus, the condition “exp. deflation < real rate" may still be satisfied even if inflation is only moderate. However, the "may" is important. It's a quantitative question.

  24. Gravatar of JimP JimP
    6. November 2010 at 16:02

    Anyone who has not read that Eggertsson paper MUST read it. It confirms what this whole blog is about. Expectations rule all.

    And how are Obama and Bernanke doing in the expectations area? Not as well as they might. But maybe a bit better than before.

    http://www.ny.frb.org/research/staff_reports/sr234.html

  25. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. November 2010 at 16:13

    JimP wrote:
    “Expectations rule all.”

    Enough said.

    Makes a good T-Shirt as well.

  26. Gravatar of MDR MDR
    6. November 2010 at 16:48

    Your blog is always interesting and thought provoking. But you should ease up on the obsession with Krugman.

  27. Gravatar of marcus nunes marcus nunes
    6. November 2010 at 17:31

    Scott
    You have “scared” Arnold Kling. Maybe he hasn´t looked at the MTMs that Mishkin has detailed in his textbook.
    http://econlog.econlib.org/archives/2010/11/scott_sumner_ex.html

  28. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. November 2010 at 17:59

    marcus,
    Terrifying! If only we poor simpletons had paid more close attention to the power of money.

  29. Gravatar of scott sumner scott sumner
    7. November 2010 at 09:12

    Marcus, Yes, just more evidence the Fed made a mistake by describing policy in terms of inflation–it should be NGDP.

    Jim and Yosef, Thanks. I don’t know why people think I am so tough on Krugman. I generally say better things about him than my commenters. Indeed the reason I obsess over him is that I regard him as one of the smartest and most important bloggers on the left. Especially on macro.

    JTapp, Yes, you can have inflation, but not increases in inflation due to a positive demand shock.

    I think you are wrong about unsterilized intervention, but I am not certain (as I regard the issue as supremely unimportant and have never focused on it.) I think others would say it’s different if the Fed bought foreign currencies.

    Mark, Those are good points about inflation, but don’t really rescue Krugman’s argument at all. If the Fed does some sort of unconventional policy, inflation expectations will change with the change in regime. That is the Lucas Critique. For instance, British inflation expectations fell during the week the central bank was made independent in 1997, despite the fact that Britain was not on gold, and despite the fact that past inflation rates hadn’t changed. There is much more to inflation expectations than past rates. Indeed TIPS spreads (and private consensus forecasts) fell sharply in the second half of 2008, even as past inflation changed little.

    Keynesians tend to assume that real growth drives inflation innovations, instead it is nominal shocks that drive real growth.

    Morgan, I want more than 5% now, and 5% after recovery. With 5% right now, it will take a number of years (maybe 4 or 5?)

    Shane, That’s one of the best comments I’ve seen in a long time. Something that would enrage both Republicans and Democrats. Just crazy enough that there may be a bit of truth. 🙂

    Mark, Good point, But I think the verdict is already in on fiscal stimulus. I don’t think Keynesian models would predict such low real growth during a period of fiscal stimulus, even if the level of stimulus was declining. At worst, where should have had trend rate of growth. If Krugman’s explanation is correct (for the slowdown) then 1.3 trillion really wouldn’t have made much difference, as that money would also be withdrawn at some point leading to an even bigger drop in fiscal stimulus, and an even bigger slowdown.

    W Peden. You said;

    “One of the problems seems to be that, while there are cases of monetary stimulus without fiscal stimulus (as with this current package) it’s hard, at least for me, to think of cases with fiscal stimulus with no monetary stimulus.”

    I recall that Keynesians predicted high inflation from the Reagan deficits. (Did that include Krugman?) Yet NGDP growth plunged from about 10% to near zero in the 1981-82 period, when we saw very expansionary fiscal stimulus and tight money.

    In 1968 it was reversed, and NGDP kept rising briskly.

    So when there is a test, monetary policy generally wins out.

    Morgan, Are you saying Krugman was in the film? Oops, I guess he is.

    Felix, That’s a good observation, but I’d make a few points.

    1. Real currency demand has not been rising sharply, as it’s supposed to if the real return from currency dominates other assets.

    2. I don’t see it as a quantitative question if you mean “Does the Fed have the capability to produce 2% inflation expectations?” If you mean can they do this under restrictive assumptions about what tools are available, then I do agree. It’s really a question of does the Fed feel comfortable getting very aggressive. There are so many transmission mechanism (10 in Mishkin’s textbook) that they have enough ways to overcome the zero rate trap.

    JimP, Yes, Eggertsson is very good on expectations. We once had a long email back and forth.

    MDR, I probably need therapy. But check my response to Jim and Yosef above.

    Marcus and Mark, I left a long comment over there. Perhaps I should do a post, as I suppose people don’t read comments.

  30. Gravatar of marcus nunes marcus nunes
    7. November 2010 at 10:10

    Scott: “Post required”. Here are 2 comments on your comments back a A. Kling´s blog:
    1. Sumner should really pay more attention to pros in investing. They pay a great deal of attention to what the Fed intends to do. The stock market tanked earlier in the year when it became clear that the Fed would end monetary pumping. It languished all summer until it became clear again that the Fed would resume monetary pumping. Machlup explained in the 1930’s that most new money goes into assets, especially the stock market and most working (not academic) financial experts know it.

    And Sumner ignores Uncle Miltie’s warning that the lags between policy and effect are long and varied. Today, monetarists think that expectations overcome those lags. Their faith in expectations takes on theological fervor at times. They seem to believe expectations trump every other economic law.

    2. To say you don’t believe in bubbles is to say you don’t believe in wide-scale misallocation of capital. Given the number of unused homes in the US I don’t see how you could possibly come to that conclusion.

    If it is just about NGDP why doesn’t the Fed put out an RFP for $1T worth of teddy bears, or why dont we just invade another country at a cost of $1T. Higher NGDP = more jobs; problem solved.

    To me it is all about return on capital, not capital invested. I do think the market is the best mechanism for ensuring good returns on capital, BUT only if is getting the right interest rate, FX signals, etc. Clearly this is not happening.

    I’m sure QE will result in some more NGDP, but I am also sure a good chunk of it will be wasted. They key for me as an investor is to figure out where that waste will occur, get in early, and get out before a crash. I hate that my job has been reduced to that, but that is the signal the Fed is forcing on me. Unlike Scott (and most other economists) I am actually allocating large chunks of capital in the economy..and I can tell you it is far from being optimally invested.

  31. Gravatar of Andy Harless Andy Harless
    7. November 2010 at 16:31

    The bigger problem with his argument is that he assumes lower real interest rates are the factor causing the increase in AD….In fact, expansionary monetary policy affects all sorts of other variables, including stock prices and exchange rates.

    As a minor point, I don’t think someone with a Nobel prize for international economics will disagree about the exchange rate mechanism, but he has pointed out that the slump is an international phenomenon, so beggaring thy neighbor by affecting exchange rates really only transfers the problem rather than solving it, and other countries (assuming they are reasonable, although I realize that may not be a valid assumption) will react with policies that largely negate the effect of QE on exchange rates.

    As to the other things affected by monetary policy, real interest rates are just a shorthand for talking about those things. Asset prices are just the reciprocal of expected asset returns. When the asset involved is a T-bill, it’s easier to talk in terms of the interest rate. We aren’t used to talking about expected stock returns as an interest rate, but the principle is the same. The problem is that there is one asset — cash — that has a guaranteed nominal return of zero. In order to get people to buy other assets (stocks, etc., or more importantly, houses, software, factories, refrigerators, etc.), we need to convince people that the nominal return on those assets will be high enough to justify the risk. The only fool-proof way to do that is by making the inflation rate high. For any given low inflation rate (say, 2%), there is no guarantee that the real full-employment expected return on real assets exceeds the required risk premium by that amount. Thus there is no guarantee that such inflation rate is consistent with full employment.

    Krugman is no doubt aware that the expected real return on assets will be higher (compared to what it is now) when the growth rate is high enough to converge to full employment. For Krugman’s argument, the relevant real interest rate is the risk-adjusted interest rate from that full-employment path. You may disagree about whether that interest rate is less than negative 2%, but that’s an empirical question. There’s nothing wrong with the model he’s using.

  32. Gravatar of JTapp JTapp
    8. November 2010 at 16:21

    Scott, Greg Ip actually answered my objection to him saying QE wasn’t unsterilized forex intervention in this interview with Felix Salmon.
    “Buying either bonds or foreign currency with newly printed money are both QE. A monetarist would argue their impact is the same: by increasing the domestic money supply, they ultimately raise the price level and lower the purchasing power of the currency, which should drive down its foreign exchange value…in practice, however, the two policies have very different effects, especially by altering expectations. Domestic QE influences what investors think the government wants bond yields to be, while FX intervention does the same for the currency’s value. Not surprisingly, the latter is much more divisive globally.

    So, isn’t your position then a non-monetarist one according to Ip? 😉

  33. Gravatar of ssumner ssumner
    8. November 2010 at 18:01

    Thanks Marcus, A few quick reactions.

    1. I don’t believe money goes “into” markets.
    2. I think Friedman was wrong about lags (because he misidentified monetary shocks.)
    3. Even without bubbles, misallocation of capital can occur–which should be obvious if you think about it.

    Andy, I don’t agree on Krugman’s view of the exchange rate issue. He was highly critical of the Germans a few weeks back when they accused the US of doing beggar-thy-neighbor policies. He says that’s a phony issue, and I agree.

    You said:

    “As to the other things affected by monetary policy, real interest rates are just a shorthand for talking about those things.”

    But then how can anyone talk about zero rate traps? There’s no upper limit on the price of houses, or corporate assets, or commodities, or foreign exchange. I’m not saying you are wrong, but I am saying that people who look at monetary policy through the interest rate mechanism only, reach radically different conclusions from if they did as you suggest, and look at other assets.

    You said;

    “The only fool-proof way to do that is by making the inflation rate high. For any given low inflation rate (say, 2%), there is no guarantee that the real full-employment expected return on real assets exceeds the required risk premium by that amount. Thus there is no guarantee that such inflation rate is consistent with full employment.”

    I strongly disagree. Suppose you depreciate the dollar enough to create 2% expected inflation. And suppose the SRAS is very flat. Then this argument would be wrong. Krugman responds by arguing that the central bank can’t depreciate the dollar, and cites the Swiss predicament as an example. I think that is a weak argument. A central bank can peg a currency for decades, by supplying an infinite amount of money at the pegged rate. It’s simply not a problem in the real world. Indeed I think the Fed has just disproved it once again, by lowering the value of the dollar, something Krugman insists can’t occur unless you have something like 4% inflation. I see no reason why the BOJ could not have raised Japanese inflation to 2%, it was not a choice between 0% and 4%; that ignores the many ways in which money affects AD, and focuses solely on interest rates.

    You said;

    “You may disagree about whether that interest rate is less than negative 2%, but that’s an empirical question. There’s nothing wrong with the model he’s using.”

    Again, there is something wrong with a model that says a central bank can’t depreciate their currency in the foreign exchange market. I simply don’t find that credible. And if you use a closed economy model, then you still have the flexible prices of all sorts of real assets like stocks and commodities. Again, the model treats the economy as if there is simply a single all-important interest rate. But that’s not how the economy works, there are lots of asset prices that matter, and most aren’t up against the zero bound. So he is using a model with no relevance to reality. It’s a world with interest rates on Treasury bonds, and goods prices that are sticky. And where cash and T-bonds are perfect substitutes at the zero bound. But what about goods prices that are not sticky? If you have them, then the model breaks down.

    We weren’t “converging on full employment” in 1933, but all sorts of asset prices were soaring. His model can’t handle that.

    JTapp, I wasn’t trying to give you my view so much as what I thought the conventional view is. I agree with the monetarists to some extent. The most important impact is on the money supply, which asset you buy is much less important.

  34. Gravatar of Andy Harless Andy Harless
    9. November 2010 at 15:04

    He was highly critical of the Germans a few weeks back when they accused the US of doing beggar-thy-neighbor policies. He says that’s a phony issue, and I agree.

    It’s a phony issue because the ECB can (and should) counteract the beggar-thy-neighbor effect by easing its own monetary policy. But if the ECB (along with other central banks) were to do so, then the exchange rate mechanism wouldn’t work, which was my point. It only works if our neighbors are willing to be beggared, and Krugman’s position (with which I agree) is that they shouldn’t complain, since it’s their own fault if they fail to react.

    But then how can anyone talk about zero rate traps?

    Even without bonds, there is a zero rate trap, because money has a zero interest rate. (That is, currency does, inherently, and the banking system is committed to offering unlimited quantities of currency on demand in exchange for other forms of money.)

    there is something wrong with a model that says a central bank can’t depreciate their currency in the foreign exchange market

    There is something wrong with a model that assumes foreign central banks will passively tolerate that depreciation. Perhaps it is realistic, perhaps the ECB and BoJ are foolish enough not to loosen as the dollar (and RMB) strengthens, but it strikes me as a strange assumption.

    what about goods prices that are not sticky?

    Most flexibly priced goods are inputs rather than outputs. If you bid up the prices without raising the expected inflation rate for outputs (i.e., rigidly priced goods), that’s an adverse supply shock. And its a supply shock with demand-side feedback that would make it very hard to raise aggregate demand sufficiently. (By “hard” I mean you’d need a very big shock, which would have severe supply-side effects, so you would get large price increases with little if any increase in real output.) Be careful what you wish for when it comes to realistic models.

    We weren’t “converging on full employment” in 1933, but all sorts of asset prices were soaring.

    I’m not sure how this contradicts Krugman’s model. The average actual inflation rate between 1933 and 1948 was over 4%, and most of that inflation came after we reached full employment. That experience seems consistent with a view that 2% inflation is not an equilibrium.

  35. Gravatar of anon anon
    10. November 2010 at 14:37

    The “risk free” part is also completely contradictory to reality. T-bills carry the risk of inflationary devaluation. And by inflation I mean the reality based inflation rate, not the fake one.

  36. Gravatar of ssumner ssumner
    11. November 2010 at 10:06

    Andy, I misread your argument because Krugman has also argued that in a liquidity trap a central bank might find it impossible to depreciate its currency. Indeed a few months back he cited Switzerland as an example of a country that was trying to depreciate it’s currency, and failed.

    He also cited Japan in previous years, although no one would argue that other countries would have completely negated a Japanese devaluation in the 1990s, as the rest of the world wasn’t even in a liquidity trap.

    But I accept your point, it is possible that the US would not be able to depreciate its currency against foreign currencies. But I also argue that it can be depreciated against other assets like commodities, stocks, and real estate–which do not suffer from a zero bound. So I’m still not worried about monetary policy being impotent at the zero bound. It’s an empirical question, and I think the empirical evidence strongly supports my argument.

    You said;

    “I’m not sure how this contradicts Krugman’s model. The average actual inflation rate between 1933 and 1948 was over 4%, and most of that inflation came after we reached full employment. That experience seems consistent with a view that 2% inflation is not an equilibrium.”

    It’s not clear to me what we are debating here. I thought the Keyneisans argued that inflation occurred as one approached full employment. In 1933-34 we were as far from full employment as we have ever been, and we had really fast inflation. I was not using 1933 as an example of 2% inflation, although I don’t see why it’s not possible.

    If Krugman wants to tell some story about commodity prices and supply shocks, that’s fine. But I’ve never once seen him make that claim. He’s always insisted the SRAS curve is really flat, and that we wouldn’t even get 2% inflation w/o much faster growth. In my view the effects on commodity prices occur precisely because AD rises. In any model I know this is associated with growth. And I’m not just talking about commodities, there are also assets like commerical real estate, and stocks. There is more corporate investment and new construction when prices rise.

    One other problem with your commodity comment. You’d already ruled out the exchange rate mechanism, so it must apply in the closed economy case. But it doesn’t–here’s why:

    In a closed economy model where commodity producers are price takers (as most are) you can only get higher prices if demand rises or supply declines. Monetary stimulus will not cause supply to decline. So if prices rise (and wages are sticky) we can infer that commodity producers will produce more output in equilibrium. But that implies FINAL GOODS output will rise as well–otherwise where would all those commodities go?

    anon, There is not much inflation risk in T-bills, they are very short term. But technically your point is valid.

Leave a Reply