Winner of the New Statesman SPERI Prize in Political Economy 2016


Thursday 23 May 2013

The Liquidity Trap and Macro Textbooks


Over the last eleven days something unusual has happened – I have not only failed to post a blog of my own, but I have not even read anyone else’s posts. Instead I have taken advantage of a sabbatical term to take a break [1] in Umbria, during a time of year when it is still cool enough to walk, but not too cold in the Piano Grande. Before I left I did write a couple of things that I thought I might quickly post while away, but with the help of the Italians’ penchant for starting dinner late and eating four (or more) courses that idea somehow got lost.  

So I’m spending part of today catching up, and reminding myself why Paul Krugman and Martin Wolf are such great writers. (For example, from the former, a masterful analysis of the decent into worldwide austerity, and from the latter, a perfect short account of why when it comes to government debt the Eurozone really is different.) What I want to pick up on here is this Krugman post, where he questions the description in a Nick Crafts piece of higher inflation as a way out of the liquidity trap as being ‘textbook’. (See also Ryan Avent.)

So is raising inflation expectations to avoid the liquidity trap textbook or not? Let’s take the 2000 edition of the best selling undergraduate macro textbook. Here ‘liquidity trap’ does not appear in the index. There is a page on Japan in the 1990s, and in that there is one paragraph on how expanding the money supply, even if it was not able to lower interest rates, could by raising inflation expectations and therefore reducing real interest rates stimulate demand. One paragraph among 500+ pages is not enough to make something ‘textbook’, so it seems as if Paul Krugman has a point.

Yet how can this be? It is not one of those cases where textbooks struggle to catch up with recent events, because the Great Depression was a clear example of the liquidity trap at work. How can perhaps the major macroeconomic event of the 20th century, which arguably gave rise to the discipline itself, have so little influence on how monetary policy is discussed? Yet it is possible to argue that the discussion is there, in an oblique form. A standard way of analysing the Great Depression within the context of IS-LM, which this popular textbook takes, is to contrast the ‘spending hypothesis’ with the ‘money hypothesis’: was the depression an inevitable result of a negative shock to the IS curve, or as Freidman argued could better monetary policy have prevented this shock hitting output?

A standard objection to the money hypothesis is that nominal interest rates did (after a time) fall to their lower bound. The counterargument – which the textbook also suggests - is that, if the money supply had not contracted, long run neutrality would imply that eventually inflation would have to have been higher, and therefore real interest rates on average would be lower. So in one way the story about how higher inflation could avoid a slump is there.

What is missing is the link with inflation targeting. Because textbooks focus on the fiction of money supply targeting when giving their basic account of how monetary policy works, and then mention inflation targeting as a kind of add-on without relating it to the basic model, they fail to point out how a fixed inflation target cuts off this inflation expectations route to recovery. Quantitative Easing (QE) does not change this, because without higher inflation targets any increase in the money supply will not be allowed to be sustained enough to raise inflation. In this way inflation targeting institutionalises the failure of monetary policy that Friedman complained about in the 1930s. Where most of our textbooks fail is in making this clear.    


[1] Sometimes known as holidays, these are things that we Europeans are forced to take many more of than Americans, leading to great frustration and misery (or maybe not).

8 comments:

  1. Simon,

    "Quantitative Easing (QE) does not change this, because without higher inflation targets any increase in the money supply will not be allowed to be sustained enough to raise inflation."


    That statement contains a strange non-sequitur, i.e. the implication that somehow, if inflation was permitted to rise, QE could cause inflation. I.e. if we *will* inflation to be higher, then QE will cause it? The problem, of course, is that OMOs achieve exactly nothing when the short rate is at zero (or IOR is at the short rate). No matter what we "target."

    And, yes, the expected future money supply could in theory be mapped to future interest rates and therefore have some effect. But such differences in money supply would have to be 10, 20 or maybe 50% above pre ZLB levels. It strains credibility to see how changes to a money supply that's already been increased by two orders of magnitude over the non-ZLB equilibrium level, can have any consequence for post ZLB interest policy. Whatever the post-ZLB interest path, essentially all QE would have to be reversed in order to raise rates at all.

    And all that assumes that IOR is stuck at zero, since otherwise the quantity of reserves can have no consequence *whatsoever* for the level of rates now, or at any point in the future. And why would one assume that? In the case of the Fed, it seems far more likely that they'll just raise rates by raising IOR, whatever the size of balance sheet. And the UK already runs a corridor system.

    What is the theory whereby QE affects the economy when the rate on reserves is perpetually equal to the rate on t-bills?

    Our textbooks don't need to resolve the conflict between the exchange equation, IS/LM and the inflation target. What they need to do is to toss out all the pre-intertemporal-equilibrium neoclassical claptrap, and adopt a framework that recognizes that *contingent* on the term structure of interest rates, the "quantity of money" has zero macroeconomic consequence.

    ReplyDelete
  2. Another high class article ( tautology?).

    I hope my publicising it gets more people reading it down-under.

    Thanks a lot

    ReplyDelete
  3. Mankiw only says that modern currencies are fiat money once in his textbook. (2003 edition.) Is that a basis to say that the view that modern currencies are fiat money isn't the textbook view?

    The view that the liquidity trap is no barrier to monetary stimulus is the textbook view. It's just that it turns out that a lot of macroeconomists (including those advising politicians) don't actually believe what they've been telling us.

    ReplyDelete
  4. The ZLB in Mankiw's text book (2011): http://gregmankiw.blogspot.co.uk/2013/05/the-zlb-in-my-favorite-textbook.html

    ReplyDelete
    Replies
    1. Of course any textbook published in the last few years is going to contain more discussion of the ZLB or liquidity trap, which is why I looked at the 2000 edition. But what is really interesting from the exert you link to is that it still does not note that raising inflation expectations would require raising/breaking an inflation target for a time. So my discussion still applies even to textbooks published in 2011! What a strange subject macroeconomics is.

      Delete
  5. Got the idea that the FED is more looking at the 'BS position'. Rising and/or less dropping prices of several assets.
    It makes people wealthier (or less poor) and they will spend more (at least that is the theory). In that way it works very similar to higher inflation. One pushes into spending the other is supposed to create extra spending power.
    By not creating excess inflation assets like pensions will not in real terms drop in value. With all consequences like increased saving. Made even worse by lower real interests (and subsequently lower bondprices most of the back up of eg pensions).

    A choice by the CB. 2 opposing forces and a lot of politcs. And it is a bit different now, first dip I have seen in which people worry about their savings and pensions in this form.

    ReplyDelete
  6. An issue that has been ignored by the macro-bunch, is the way people (be it consumers or investors) look at inflation.
    Please have a look at the Forex market for that.
    In a nutshell and very much simplified: inflation can roughly be distinguished in good and bad inflation (for FX purposes that is).
    'Good' inflation is inflation caused by peaks in economic activity. 'Bad' inflation has other causes (basically the thing not being under control). Basically what we see in Japan now.
    While 'good' inflation is seen as a positive for the FX price as it is covered by growth, 'bad' inflation is seen otherwise. More as a symptom of an unhealthy situation. And works the other way around.

    The point I want to make is that markets (and consumers are a market in the way they are buying and selling stuff) are able to look in a much less simplified way at inflation, as basically your theory does. And probably at other important economic phenomena as well.

    My calculated guess derived from the FX is that when things look artificial or strange this might happen. Or too far of the mark (sort natural equilibrium) or last too long.

    Applying those things to this issue, you're 'inflation push' (unfortunately) ticks all these boxes:
    Artificial: CB policies
    (Simply) Strange: first time most people experience a thing like this
    Too far of the mark: effectively lower zero bound has that already in it.
    Too long: we are now >5 year in a crisis (or something very close to that) and the end is realistically speaking not in sight. 5 year is already longer than the 'normal' business cycles we are used to.

    Especially the too far of the mark interests me in particular. I personally doubt if you go that far of the more or less standard situation things will still work in a similar way as in the standard situation. At least that should be apoint to consider. Several general laws of nature say they probably donot. There is very little or no material on this as well as simply the situation (fortunately) doesnot occur regularly. And a lot of those situations were caused by structural problems that are every time different as well anyway.

    The emotions that caused a lot of people to reject your views are probably (subconscience) based on similar stuff. If you want to bring the message through you probably have to adress it there. With as extra complication that you will be dealing with mainly emotional and not logical/rational people.
    It simply looks highly unlikely that in most countries a decision like this will be taken when either there is a political/electoral platform for it (if it goes wrong the politicians that did it will hang high) or people donot know what to do next (desperate people). And when things look like desperate measures they already have a big minus to start with.

    ReplyDelete
  7. ON QE and inflation
    1. Fully agree that in practice you almost certainly will need real inflation to go up before markets/people buy higher inflation expectations.
    2. However you miss one important point. Which plays for massive stuff like QE. Or for the oil price.
    Basically it comes to the fact that the word inflation doesnot mean the same thing to all people. And measuring it in a currency has its limitations as well.

    Start with oil. Trimming your beard costs now say 20 GBP, and next week 25 GBP. Nobody will see it different than a 25% pricerise (beardinflation if you like). Works perfectly well to use the currency as the measurement tool. Only issue that some people will have relatively more beardtrims than other (but that is not relevant for this point).
    With oil it is different. Oil is priced in USD (to make it properly tradeble), but it has another relation with oil than beardtrimming has with GBP. With oil oil is not only priced in USD, but USD is at the same time also priced in oil.
    And not only in the way that the general value of a currency is reflected in (all) prices, like with beardtrimming or an espresso or a bread.

    Same sort of thing looks applicable with QE in its consequences. It leads to higher other asset prices. So GBP or USD becomes cheaper in those other asset prices.
    It is assumed that there is a currency and there are things goods/assets. Reality is much more complex however. For beardtrimming that model works well. It doesnot necessarily do that for say oil. However as we work in general with a price in USD we interpret it in a similar way. Which doesnot have to be correct.
    Or seen from the other angle a currency is also an assetclass (which makes things even more difficult to work with as you will not have an easy measurement tool).

    From there QE via pricerises (or better relative value rises) of (other) assetclasses increases inflation. In the way that the value of the USD goes down in comparison with the rest of the assetclasses.
    This is however no inflation as officially measured or you refer to. However it is inflation as people at least some of them experience it. And that experience is the reason your mechanism works. In other words it is a pretty relevant point. QE might however push people as it mainly addresses top incomes (plus effects mainly some buys the asset part not the consumer products so much) into a certain direction with their spending.
    Which is probably what we see now nearly all goes into assets and very little in consumption.
    So at the end of the day your endconclusion is probably correct only in the way to it you miss an important step. Which eg obscures that this aspect of QE in its present form doesnot seem very effective.

    With inflation we also have issues on topics of perceived and real, or as you mentioned VAT increases; or in the EZ per country; or in the EZ as well why not leave countries like the PIIGS out all together they need deflation (as least in EZ ideas) to rebalance their competitive position.

    ReplyDelete

Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.