Those magical, mystical, long and variable lags

The following quotation discusses one of the more perplexing aspects of quantum mechanics:

In 1935, several years after quantum mechanics had been developed, Einstein, Podolsky, and Rosen published a paper which showed that under certain circumstances quantum mechanics predicted a breakdown of locality. Specifically they showed that according to the theory I could put a particle in a measuring device at one location and, simply by doing that, instantly influence another particle arbitrarily far away. They refused to believe that this effect, which Einstein later called “spooky action at a distance,”1 could really happen, and thus viewed it as evidence that quantum mechanics was incomplete.

I don’t plan to explain this phenomenon (and please don’t write in with an “explanation,” as you’ll only convince me that you don’t understand it.)  But regardless of whether there is action at a distance in particles, I am convinced that the concept does not apply to economics.  To be more specific, I don’t believe in “inflationary time bombs” hidden in money supply increases.  And I don’t believe in “long and variable lags” from monetary policy shocks.

Then what do I believe?  These six conjectures:

1.  Monetary shocks can affect sticky wages and prices with a distributed lag, but the lag is not variable.

2.  Monetary shocks can affect output with a distributed lag, but the lag is not variable.

3.  Monetary shocks are very difficult to identify using traditional indicators (interest rates, money supply, etc.)

4.  The “long and variable lag” hypothesis comes mostly from previous misidentification of monetary shocks.

5.  When monetary shocks are correctly identified they have a very powerful and immediate impact on all sorts of asset prices.

6.  If one does not observe the predicted asset price reaction from an apparent monetary shock, then the shock never happened.

I came to these conjectures in several different ways.  One strand of research looked at monetary shocks during the Great Depression.  I decided to investigate a model of money based on its role as the medium of account.  Thus in the 1920s and 1930s it should have been possible to explain price level movements in the U.S. by examining changes in the supply and demand for gold (and in 1933 the price of gold.)

When I applied the model to the US, I was shocked by how well it was able to explain the ups and downs of stock prices, commodity prices, and industrial production.  And there seemed to be almost no lags.  When major central banks started massively hoarding gold after October 1929, stocks, commodities and industrial production plunged.  Private gold hoarding in late 1931, spring 1932, early 1933 and late 1937 had similar powerful and immediate effects.  Gold dishoarding in the summer of 1932, and 1936 had the opposite effect.  And increases in the price of gold had the most powerful and easily identified effect of all (as I discussed in my George Warren post.)

Then I went back and looked at the famous examples of Friedman and Schwartz.  The bank panics had a contractionary effect, but it was hard to find evidence that any of the key Fed policy movies had the effect F&S hypothesized.  The famous discount rate increases of October 1931 (from 1.5% to 3.5%) seemed to have no impact at all.  The same for the famous reserve requirement increases of 1936-37.  And the effect of the 1932 OMOs was ambivalent, and almost certainly not what F&S claimed.

Suppose you were a really smart guy, and you convinced yourself that monetary policy was the key factor driving NGDP growth.  And suppose you also reached the conclusion that fluctuations in NGDP growth explained the business cycle (or at least the demand side cycles.)  But you lacked the theoretical tools necessary to correctly identify monetary shocks.  The base went up in the Depression, so that wouldn’t work.  Interest rates are unreliable for all sorts of reasons.  So you notice that M1 and M2 plunged in the Great Depression.  There’s your indicator.  Even better, M2 tracks NGDP to some extent.  (Of course all sorts of other nominal aggregates also track NGDP, but let’s put aside that annoying complication.)

Now you’re ready for your grand monetary history of the U.S.  Now assume that M2 is not truly the right indicator of monetary shocks, it’s just one you found that works reasonably well for the big shocks.  So someone asks why M2 rose this year, but NGDP did not.  You respond, “just wait, it will eventually increase as well.”  Now you’ve been boxed into a corner.  You say money affects NGDP with a lag (even though in your famous Great Depression example NGDP seemed to fall before M2.)  People are going to ask you how long the lags are, and when can we expect NGDP to rise?  You answer, “it depends, the lags are long and variable.”

But even that isn’t enough.  In the 1980s rapid money growth leads you to warn about a return of high inflation.  The years go by, then decades, but the high inflation never comes.  Now your whole model becomes discredited.  But here is the ultimate irony.  The best parts of your model are discarded (the monetarist transmission mechanism, skepticism about interest rates, the monetary theory of the business cycle), and yet economists of widely varying schools of thought hang on to the weakest part of the model, the slight of hand that pathetically tried to cover up for the model’s failure to identify monetary shocks, the “long and variable lags.”

You may recall in an earlier post I discussed how believers in astrology react when I ask if their model could be tested by comparing the average psychological profile of people born in different months.  And the response is always, “well, it’s more complicated than that.”  Yeah, perhaps there are long and variable lags.

I am pragmatist, so it was my Depression research that had the greatest impact on my views of long and variable lags.  But after a while I also realized that the concept was totally at variance with rational expectations and the EMH.  If monetary injections were inflationary time bombs, why didn’t they impact commodity prices, or TIPS spreads, or any of the other asset prices that theory predicts should be impacted.  Then I started reading Svensson’s work on targeting the forecast, and realized the only meaningful indictor of monetary shocks was expected changes in the goal variable (inflation, NGDP, etc.)

Now I understood what had really happened in the Great Depression.  The monetary shocks must have occurred right about at the same time as stocks and commodities crashed.  But they crashed at roughly the same time industrial production plunged.  Not just in 1929, but each time throughout the 1930s.  There were no long and variable lags.

Monetarism came about long before the Ratex and EMH revolutions, and thus lacked the proper tools to identify monetary shocks.  Then there was the long fruitless detour into equilibrium macroeconomics.  By the time sticky-price models came back into vogue, the new Keynesians were driving the agenda.  Models did incorporate rational expectations, but they still had trouble identifying monetary shocks.  Woodford understood that what mattered wasn’t nominal interest rates, but rather the policy rate relative to the underlying (and unobservable) neutral rate of interest.  And not the current interest rate, but the expected future path of interest rates.

But this was all so hopeless vague that real world macroeconomists went back to the old-fashioned Keynesian practice of equating low rates with easy money and high rates with tight money.  When low rates didn’t produce fast growth in 2002, it was attributed to those mysterious “long and variable lags.”  (Very few seemed to notice that money wasn’t easy at all in 2002, it was tight.)  Lags had become an all purpose excuse not just for monetarists, but for macroeconomists of all stripes.  Here are a few recent examples of the inflation “time bomb” approach:

Monetarist Allan Meltzer:

Keynesian Brad DeLong:

Austrian Bob Murphy:

I’m not trying to single out these three guys; this is the standard way that almost all macroeconomists look at things.  And I don’t see how it is consistent with the EMH.

Here’s how I view the monetary transmission mechanism.  The Fed does something to increase the future expected path of the base, or to reduce the future expected demand for the base.  This raises the future expected path of NGDP growth.  This raises all sorts of current asset prices.  It raises the TIPS inflation forecast.  It lowers the Baa/Aaa yield spread.  And all of these things happen right away.  This sets in motion changes in sticky wages and prices that occur over a longer period.  But even those changes start right away.  There is no mysterious “action at a distance” that violates the laws of economics.  No time bombs waiting silently to go off at a future date.  It is those immediate changes in expected NGDP growth and asset prices that cause the delayed changes  If the immediate changes don’t happen, neither will the delayed changes.

How could everyone have gotten it wrong?  Consider the following example.  Suppose we have spent a long time on a certain monetary regime, and people are used to that regime.  When change finally comes it might well be unexpected.  Thus in the 1960s we began to experience faster than normal money growth.  In past decades that would not have lasted long, as the price of gold was fixed.  But now the higher money growth begins to persist, and the public starts to catch on, and AD and prices start rising more rapidly.  It looks like the monetary expansion impacted the economy with a lag.  But what really happened is that the initial expansion was viewed as temporary and thus had little effect (see Krugman 1998).  Only when people began to understand that the policy change was permanent, did they revise their forecasts of expected future money and NGDP growth.  And that change in the forecast of future nominal growth is the real monetary “shock”, the relevant shock not just in my view, but in ‘state of the art’ new Keynesian models.  And there are no lags between that shock and asset prices.



21 Responses to “Those magical, mystical, long and variable lags”

  1. Gravatar of Leigh Caldwell Leigh Caldwell
    14. May 2009 at 14:39

    Interesting. I agree with your basic idea (and Einstein’s) of no-action-at-a-distance.

    I’m not completely convinced that you aren’t simply redefining “monetary shock”, but it’s certainly worthwhile to distinguish between ultimate and proximate causes. In this case the proximate cause of the rise in AD is the public’s realisation that the M growth is permanent; the ultimate cause is the permanent growth in M; and in between the two there are many parallel chains of cause and effect. The lag between proximate cause and effect is indeed zero, but there are still lags in the chain before the proximate cause which you believe is the real shock.

    Two points for future discussion:

    1. Is it possible to predict or model the rate at which the public will catch onto something like this? I feel it is, though I don’t think we have the tools for it yet.

    2. I strongly believe that there’s also no action-at-a-distance in individual motivation. For example, am I saving $1000 now so that I’ll have a better lifestyle in retirement? No way. I don’t believe in this long-distance (and, for that matter, reverse) causation. Instead, I want to identify what payoff I get NOW from saving NOW – psychological, social or whatever. I believe that understanding this will give us a strong insight into behavioural effects such as hyperbolic discounting. And by extension and by your no-lags argument, there may be macroeconomic effects which are analogous to those behavioural patterns.

  2. Gravatar of David Landry David Landry
    14. May 2009 at 15:16


    I’m curious to know what you think of the work Romer and Romer (2004) did to derive a new measure of monetary shocks. They used the fed funds rate as their measure (just bear with me here), but attempted to purge it by accounting for endogenous variation and anticipatory actions by the Fed by using the intended/target (instead of actual) fed funds rate as well as the economic forecasts used by the Fed when those targets were established. Their measure then tries to identify monetary shocks as changes in the intended fed funds rate after accounting for changing expected and future economic variables. As I see it, this sounds similar to your notion of monetary shocks, since it considers how the target changes, not the actual rate, and because they are unrelated to any current information available about the current and expected future path of the economy (exogenous). Just curious to hear your assessment. If you haven’t read it, the abstract probably gives a better synopsis than I did.

    Oh, and the conclusion they found was that monetary policy is far more potent than previous econometric studies indicate, and also solves the annoying “price puzzle”


  3. Gravatar of ssumner ssumner
    14. May 2009 at 15:27

    Leigh, I didn’t explain myself very well. What I meant was not that it took a while for the public to catch on, like they were stupid, but rather that it wasn’t clear whether the policy change was going to be permanent. Only when and if it was seen as permanent does it become a true monetary shock. One point I cannot emphasize enough is that the vast majority of changes in monetary indicators are merely accommodating shifts in the demand for money, and are not shocks at all. So I don’t think I am redefining monetary shocks, as nobody would describe the increase in the monetary base every December as a monetary shock, it is well understood to be accommodating the public’s increased demand for currency. Most monetary base increases are like that, including our current one, albeit for very different reasons.
    BTW, many shocks occurred in the Depression when monetary changes occurred that were (wrongly) thought to be permanent. So it works both ways.

    Regarding your two questions:

    1. No, you cannot predict when the public will catch on because the public is much smarter than you (and I) are. So any attempt to model it is doomed to fail. (This relates to my previous point above.) It would be like predicting the next theory that Steven Hawking will come up with.

    2. That’s a very interesting analogy, that I would have to think about a bit more. My first reaction is that it was the wrong analogy. But with further thought I think you may be on to something. I’m not sure others will see it that way, but I do think you could view the decision to save as a decision that leaves you happier right now than if you didn’t save (and thus worried and fretted about the future.) Is that your point?

  4. Gravatar of ssumner ssumner
    14. May 2009 at 17:35

    David, That’s a great question. (Unfortunately I haven’t read the paper, I probably skimmed it a few years ago, but I am not sure.) Let me break the question into pieces. The question of what is a monetary shock is difficult. First you need to make a sort of “value judgment” about what nominal aggregate promotes macro stability (price level, NGDP, etc.) Say you decide (like me) that NGDP is the key. Then a monetary shock is an unexpected change in M*V. If we had a futures market, a sudden change in the expected NGDP growth path would be a monetary shock, as I define the term. Note that Fed errors of omission and commission are equally important.

    So any attempt to artificially estimate shocks by looking at the Fed behavior is very, very difficult. The Fed simply does not set out to create monetary shocks. (At least it usually doesn’t, in 1920, 1929 and 1981 there were fairly explicit tight money policies. In 1933 there was a very easy money policy.) But a lot of the 1950s-1980 updrift in inflation was Fed mistakes.

    On the plus side, I like the pragmatic way Christy Romer approaches macro issues. She also did a lot of interesting work on the Depression. She is sort of like me except she is a Keynesian and teaches at a much more prestigious university (and now is with Obama.) I also like what you said about the price puzzle. I had totally forgotten about the price puzzle, but it is one more reason to doubt that our traditional approaches to identification are very good. For readers who don’t know, the price puzzle refers to the fact that when traditional interest rate-oriented methods are used to identify monetary shocks, then money seems to cause the price level to go in the wrong direction—more money means lower prices! (I guess thinking about recent history it’s not so hard to understand.) I see this as a huge problem, and if the Romer’s procedure solves that problem, they may be on to something. But as I said in my previous response, what you really want is market expectations—in a strange way they are both the indicator of the shock and the shock itself. Thus if the TIPS spread jumped from 1% to 2% tomorrow, that would INDICATE a positive monetary shock, and also it would CAUSE AD to rise more rapidly. It’s a weird way of looking at things, and I don’t expect many people to agree with me. Of course the true monetary shock would be the thing that caused inflation expectations to jump, and I guess that is what R&R are looking for. But it could be complete passivity on the Fed’s part. I’ll try to take a look at the paper and get back to you. Their approach does sound interesting. Is there a copy on the internet?

  5. Gravatar of David Landry David Landry
    14. May 2009 at 22:47

    Thanks for your thoughts Scott. I used my university’s online database of academic journals to get it, but it’s also available through David Romer’s website:

    If the link doesn’t work for whatever reason, it’s called “A New Measure of Monetary Shocks: Derivation and Implications” and appears in the September 2004 edition of the American Economic Review.

  6. Gravatar of Leigh Caldwell Leigh Caldwell
    15. May 2009 at 00:16

    Funny, that same issue of AER contains a Robert Sugden paper that I have indepedently come across twice this week (once in a presentation by Sugden in a psychology seminar and once in someone else’s Powerpoint about framing and loss aversion). Coincidences will happen…

    Thanks for your clarifying comments Scott. On point 1, I agree that the public is not stupid, but in some cases you can predict the aggregate cognitive behaviour of a population without assuming they are sheep. In marketing, for example, firms often use methods which take advantage of the public’s departure from strict rationality (framing, signalling effects and so on) and I think that the effects of these can be systematically predicted.

    But certainly at present we don’t have the tools to do it at anything like the sophistication that would be required in this case, and you may well be right that we never will.

    On point 2 you have understood me correctly: that feeling of being happier is exactly what I mean. And I wonder what it is that leads us to be happy in this way. I suspect that some aspect of it is traditionally rational, in that people do calculate their lifetime expected utility and get a good feeling from equalising it.

    But even that kind of rationality is a learned behaviour, and I would imagine that the learning differs between individuals, and that it is subject to strong cultural influences. Similarly, a component of this happiness must come simply from a general unquantifiable feeling of safety. This lack of quantification could easily lead to under- or (less likely) over-saving.

    I guess the conventional answer would be that the market will sort it out, and set a rationally correct average saving level, but I’m not sure if that’s always true. In particular, there are equivalent effects in any other area of future planning – for instance planning to carry out a project over a six month period – which are much less open to market forces than the savings example. Maybe we need to set up miniature markets in the hitting of project deadlines…how about a Microsoft Project futures index alongside that NGDP market?

  7. Gravatar of Bill Woolsey Bill Woolsey
    15. May 2009 at 02:11


    I don’t aggree that “the public” is smarter than “us.” I think the opposite is true. I am sure that there are other people smarter than me. More importantly, I think it is likely that people who devote full time to trying to figure out if nominal income is going to be on target will be able to do better at it than assigning team of government selected economists to develop the best macroeconomic model they can–even if all the members of that team are smarter and better trained than I am.

    To me, a monetary shock is an excess supply or demand for money. The quantity of money can be observed reasonably well–better than nominal income, not as well as particular prices or interest rates. The demand for money is not observed. It is how much I want to hold, you want to hold, and on and on. I don’t have a good idea about what I want to hold, much less have confidence in my ability to figure it out for everyone else.

    Assuming that some particular measure of the money supply is “the” money supply and that the demand for “the” money supply is proportional to nominal income aways, so that measures of changes in that measure of the money supply are monetary shocks–yes, I can see there are some problems with that approach. I suppose I should read the Romer paper, but looking at some short term interest rate to figure out monetary shocks is worse. (Excess supplies or demands for money will cause changes in lending or borrowing behavior and so changes in some interest rate. Maybe.)

    Anyway, here is how the lags work. There is an excess supply of money. People take time to adjust their their spending plans. I think Darby described this as a shock absorber approach. I suppose you could argue that it really isn’t an excess supply of money until people start to spend it, but I don’t think that reflects the human experience. Something changes. You know it is changes. You are going to respond. And this process takes time.

    However, in the discussion regarding the Austrian Trade Cycle theory, and so, in a slightly different context, it has become more and more implausible to me that this sort of disequilibrium can be maintained for long. In that context, banks can lend money into existence. Those holding the money accept it because they intend to spend it, but it isn’t filling a demand to increase money balances. The borrower can capture goods and services, but no one is “saving” to free them up. I think this is possible–but for how long?

  8. Gravatar of ssumner ssumner
    15. May 2009 at 05:56

    David, Thanks for the link. A couple comments. I don’t understand the last full paragraph in column one of page 1067. If bad data can lead to one type of error, why isn’t it symmetrical? Why couldn’t you also have the other type of error? I don’t get that point at all.

    There is something kind of odd about their procedure. They are essentially looking for “inexplicable” movements in the fed funds rate. Is that right? That is, movements not explained by some sort of Taylor Rule type procedure. Something that cannot be explained by macro data, including macro forecasts. Is that right? If so they are in a sense equating “mistakes” and “monetary shocks.” Now of course there are actually reasons for the Fed’s actions. And so in a sense you have a missing variable problem. The Fed doesn’t simply change rates for no reason, they must have something in mind. Thus the R&R model is calling “monetary shocks” changes in the ff-rate that their model is not able to explain, changes due to more subtle or more complex reasons.

    I can think of all sorts of those other factors. One of the best (for their model) would be changes in central bank preferences. Thus the Philips curve stuff in the 1960s may have led to a greater preference for inflation, and the Volcker changeover may have led to a greater preference for disinflation. Bu I also think a big part of their monetary shocks is merely misidentified responses to data not in their model, or data in the model, but not interpreted the right way (levels vs. changes, etc.)

    I find the CPI estimates (with the 22 month lag) implausible. In the clearly identified monetary shocks of the interwar period (1921, 1930, 1933, 1938) you see the CPI respond very quickly away.

    But I could easily believe that their model is better than previous VAR studies, as it does get rid of the price puzzle, so it is a nice paper.

    By the way, they mentioned between meeting ff rate changes. I wonder how often they occur, and if they are an indicator of dramatic changes in the economy.

    Leigh, I think public behavior can be predicted where it is not easy to profit from those predictions. For instance, the relative popularity of various childrens’ names can be predicted. Rich kids names gradually filter down to the middle class.

    Your Microsoft comment reminded me that many firms now have internal prediction markets to forecast variables like firm sales. Wolfers and Hanson have done research on this.

    I was also reminded of evolutionary psychology. We presumably evolved to get instant gratification from doing things that are in our long term interests. (I.e. activity to produce children gives pleasure 9 months before the children arrive.) There is an obvious evolutionary benefit to this as well. I don’t know about others, but I often derive more happiness form anticipating some future thing that I am looking forward to, than the event itself.

    Bill, I mostly agree. I do think the idea of excess cash balances (relative to demand) must be at the heart of any model of monetary stimulus. So we agree on that key point. I also think that the key is the time path of future monetary policy. Thus if an easy money policy is expected for one month, and then is expected to be followed by an extremely tight money policy for 3 years, stocks and commodities will fall immediately. The current easy money will not cause a one month blip up, followed by a sharp downturn in the economy–the downturn will occur right now. But even this is not necessarily at variance with what you are saying. If people really did expect that sort of weird reversal after one month, they might start hoarding cash now (in anticipation of the deep recession) and thus even right now policy would be tight in the sense that money demand might rise more than money supply. So perhaps there is no disagreement at all.

    Another area I would slightly differ, is the talk about the time it takes people to spend the excess cash balances. I agree that the attempt to get rid of excess cash balances is what drives NGDP higher in the long run. But I also think that in many cases monetary policy changes in a way that leads to higher future expected NGDP growth, what Keynes called “confidence.” In that case the AD stimulus can occur even before the new cash actually enters the economy. How? But rising NGDP expectations pushing up asset prices. So for me, the only lag of importance isn’t how long it takes people to spend the new money, but rather how long it takes rising asset prices (commodities, stocks, commercial real estate, home prices, etc.) to trigger higher industrial production. And I think that happens pretty fast. People like to say the stock market turns up 6 months before industrial production. But that’s not the real issue. The real issue is how long it takes before stocks influence IP. And in the first few months of the recovery stock prices might not be high enough to justify new investment. It’s not just the rate of change in stock prices that matters, it is also the level (note that stock prices are still pretty low now, despite the recent run-up.) And of course sticky wages have to adjust too–that’s where the longest lag comes in.

  9. Gravatar of David Landry David Landry
    15. May 2009 at 09:57

    I think your understanding of their approach is correct. If I’m in the Romers’ shoes, I think to myself “what would a shock to monetary policy look like?” Well a shock would have to be exogenous, meaning that it is not related to current information about the economy, and thus unpredictable based on what we know about the economy. But the fed funds rate, which is presumably the policy instrument, moves for a variety of reasons, as you have pointed out many times in your blog – market forces, reaction functions by the Fed, and anticipatory actions by the Fed. Using the Fed’s targets, instead of the actual manifested rate, helps us purge the effects of market forces. To control for the reaction function and anticipatory actions of the Fed, the Romers simply use the forecasts and estimates the Fed had available when setting the targets. In so doing, all endogenous effects of the economy, whether it be the nudges in the fed funds market or the manner in which the Fed reacts to changing economic variables, are accounted for and so whatever movements are left should not be based on relevant economic information and therefore exogenous. You’re right that deviations from normal Fed reaction to economic variables happen for a reason – the Romers posit changes in procedures (e.g. putting more emphasis on quantity measures), non-systematic “beliefs” of the central bankers (theoretic or otherwise), a shift in “tastes” (e.g. aversion to inflation), political motivations, atypical economic motivations, and “personalities, moods, and idiosyncratic views about the state of the economy that the participants bring to the meeting, the persuasiveness of their rhetoric, and so on.” The theme is that these decision-influencing factors are not in any way directly related to the economic information at hand.

    Though I’m not sure they intended it to be read this way, I see this from the perspective of a market trying to predict the Fed. The Romers note that the Fed has been relatively consistent in their reactions to economic conditions for roughly the past half century. Markets then get a good idea of what to expect from the Fed as economic variables change. As long as the Fed has good economic information, or at least is misinformed in a predictable way, markets can expect the Fed to react in a predictable way. If they deviate from these typical reactions at any given time, then markets are caught off guard – it is considered a monetary shock. Say for simplicity that inflation was the only relevant economic information the Fed used. Say that they normally lowered the fed funds rate 1% for every 1% point fall in inflation. If all of a sudden inflation falls by 1%, but the Fed only lowers the fed funds rate by .5%, as far as the market is concerned, this is a contractionary monetary policy shock. Maybe it happened because of a change in chairman’s preferences – but the market can’t really predict that, so it is (I think rightfully) considered a shock. When I went through these thought experiments when first reading this paper, I immediately thought of your claim that the Fed was not expansionary enough this past fall – I believe you called it a “contractionary surprise”(?) I’d be really interested to know if the Romer measure of shocks would concur.

    As to your concern about page 1067, I’m not sure I completely understand it either, but here’s my take. It seems that they take the Greenbook forecasts to be just as good if not better than private forecasts, at least in the modern era. I’m not sure if this is true, but they make a forceful argument for it early in the paper. However, they seem to be less certain of its prowess compared to private/market predictions in the early years. So I think their point is that the Fed will still react systematically to bad information, but if markets have more accurate forecasts and think the Fed is using these more accurate forecasts, systematic reactions by the Fed to incorrect data will be interpreted by markets as shocks. Thus their point that some monetary shocks will be misclassified as systematic responses. However, systematic responses will not be misclassified as monetary shocks since we only care how the Fed uses its own information, not the correct information that it may not have. Does that make sense?

    If the Fed were to adopt your futures targeting idea, then monetary shocks as defined in this study would cease to exist, since the market controls monetary policy in this scenario. Therefore, if I understand your scheme correctly, “shocks” would only occur when there were changes in expectations. This is a curious phenomenon because normally changes in expectations manifest themselves in demand. Now shocks would always directly affect both the demand side and the monetary side, and never just one.

    I can’t really comment with any authority on the Depression CPI stats, except that maybe this study doesn’t really apply then since Fed thinking so drastically changed. I know that’s not very helpful…

  10. Gravatar of ssumner ssumner
    15. May 2009 at 16:56

    David, You’ve gotten deeper into this than me, so I’ll take your word on some of the stuff that I haven’t fully thought out. But here are a few brief comments:

    1. Even if R&R include variables that the Fed may care about (inflation and real growth), they may not include them in the right way. Consider the very steady and predictable quarter point increases from 2004-07, there must have been nearly 20 of them. Were they all the reaction to specific economic data? If so, I don’t think they would have been so smooth and predictable. It seems to me that the Fed must look at inflation and real GDP data in a very complicated way, which is not necessarily captured in the specific assumptions used by R&R. So I believe that some of the so-called shocks may have been responses to inflation and output data (or forecasts) that were viewed differently by the FOMC than by the R&R model.

    2. To me, monetary shocks are partly the unexpected part of the announcement (as in September and December 2007, when markets responded strongly to announcements) but also partly market responses to economic data that the market doesn’t think the Fed will handle well when it does make a decision. An example would be data indicating we might be sliding into a liquidity trap. This is why stocks may crash between meetings–the market may believe the Fed won’t handle the situation well. The stock market crash at 2:15pm, right after their December 2007 announcement was one kind of shock, and the crash in the first 10 days of October 2008 was the other type of shock. Errors of commission and omission, if you will.

    Within the next week or two I plan a couple posts on monetary policy in those key months in 2008. So that might allow us to revisit this issue from a slightly different perspective

  11. Gravatar of Larry Larry
    15. May 2009 at 20:37

    My head exploded again when I read this post.

    Can this mean that you can increase money as much as you like for as long as you like as long as expectations don’t change? Is that what the Fed is unknowingly relying on as it tries to get credit flowing? Or is money the real variable, and expectations just a faulty image of that variable’s future path?

    And what drives expectations? Are there actually expectations of future NGDP growth out there anywhere?

  12. Gravatar of ssumner ssumner
    16. May 2009 at 04:48

    Larry, Not only are there expectations of NGDP grwoth “out there”, we have a pretty good idea of what they are. We have inflation forecasts from the TIPS market showing low expected inflation. We have consensus forecasts that average out private sector economists’ forecasts–and they also show low inflation. Real growth is a bit tougher but we having the Intrade betting market showing expectations of falling real GDP in the last three quarters (as a whole) of 2009. We cannot get inside everyone’s head, but all the data we have points to low inflation.

    Ask yourself why this has happened, why has the money supply gone up without triggering inflation? Because money demand has gone up even faster. So money is gaining value. Analogy; oil supply went up in 2006-08, but the value of oil went up, not down. Why, because oil demand rose faster than oil supply.

  13. Gravatar of Bill Woolsey Bill Woolsey
    16. May 2009 at 05:37


    I know you realize this, but it might be helpful to mention that the expected future monetary policy has an impact now by impacting money demand now. It is impacting velocity now.

    So, a drop in the quantity of money this month, combined with an inflationary policy for the next few years, causes “flexible” prices to rise now. This raises industrial production now. OK. How? How can prices rises now? The flow of income hasn’t changed yet? How can anyone finance additional expenditure? They borrow? But who lends? What would the lenders have done with that income? Ah… the demand for money falls. If we assume the supply of money fell, it must be that the demand fell by more than the supply, creating an excess supply of money consistent with increasing prices now.

    As I have said before, we _know_ that this cannot always be possible. Suppose the money supply drops to zero today but next year it is supposed to be really, really high? My view is that the economy will grind to a halt now.

    I also think that some kind of speculative demand for money, regarding real inerest rate changes, is going to have to play a key role in these stories, where the demand for money changes, in perhaps heroic proportion, to offset changes in the quantity of money.

  14. Gravatar of Joe Calhoun Joe Calhoun
    16. May 2009 at 05:57


    A few questions:

    1. In this framework would the innovation of the internet in the 90s qualify as a monetary shock? In other words, productivity expectations changed, NGDP expectations changed, money demand increased but the Fed didn’t meet that demand and therefore there was a monetary shock?

    2. Does the recent stock market rally indicate a recent monetary shock? Did NGDP expectations change recently for the better?

    3. Is there any way to predict what quantitative change in policy will change NGDP expectations?

    4. The Treasury started using on the run TIPS in the real yield curve composition in December. This action caused a big dip in official real interest rates in December because older bonds that would have been adjusted down in price during deflation were excluded from the curve. Does this make the TIPS, real yield curve less valuable as an inflation expectations tool?

  15. Gravatar of ssumner ssumner
    17. May 2009 at 06:07

    Bill, I think the best way to answer your comment is with an example. I think we mostly agree–tell me if you think my interpretation of the following makes sense:

    At the beginning of 1932 the monetary base was about 7.75 billion. During 1932 it gradually rose to 8 billion. In February 1933 it spiked to 8.8 billion, then quickly fell back to 8 billion in the spring. At the end of 1933 it was still about 8 billion. By the end of 1934 it was around 10 billion. So 1933 was the most “contractionary” of the 3 years, even ignoring the February spike.

    Here is my take:

    In 1932 the Fed injected about a billion in OMOs but the policy didn’t lead to higher expected prices because of the constraints of the gold standard. The base rose only slightly because the OMOs were mostly offset by the outflow of gold. And even the base increase was mostly offset by higher demand for base money (bank failures and low nominal interest rates.)

    In February 1933 the base spiked as the Fed partially accommodated the demand for cash and reserves during a severe banking panic. After March 1933 the price level soared (as did output) with the WPI up about 20% by January 1934. Why was the money supply flat at about 8 billion, even after the expected falloff when the bank panic ended? Because prices rose in anticipation of future base increases. These expectations came from dollar devaluation, which led to a higher expected money supply in 1934 and later. Why did real base demand fall in 1933 (even compared to 1932?) Because during the period when the dollar was depreciating, people didn’t want to hold an asset that was losing value. So the policy worked this way:

    1. No change in the base during much of 1933
    2. A fall in the real demand for base money
    3. The lower base demand was caused by expectations of a rising price of gold (dollar depreciation.)
    4. The rising price of gold led to expectations of higher future supply of base money.

    In 1934 the base rose about 25%, but WPI inflation slowed sharply from 1933. Why? Because we refixed the dollar to gold at the end of January 1934. At that point, people no longer expected the dollar to depreciate further, and demand for dollars went up sharply. In a weird sort of reverse Granger causality, the rapid money growth of 1934 caused the rapid inflation of 1933. Of course the actual causality was expectations of higher money growth.

    You are right that I didn’t explain this process well. To the extent that future expected money growth matters, it works through changes in the current real demand for money, as you say. I keep coming back to the 1930s because when you get extreme shocks it is easier to see the principles that underlie this process.

    I really don’t know what would happen if the Fed had such a tight money policy that it cut into the bare amount needed for transactions balances, but where this contractionary policy was associated with a big increase in the future expected growth of the money supply. Nominal interest rates would probably rise, but beyond that I think it would depend on the magnitudes involved, as you say.

    1. I think the internet boom is a monetary shock only if Fed policy is slow to adjust, which it might be in reality. The internet boom does cause RGDP to rise, but increases in RGDP have no direct effect on NGDP. The possible indirect effect is as follows: The boom raises the return on capital, forcing up real interest rates. That could reduce the demand for base money, and raise velocity. But if the Fed is NGDP targeting, the productivity boom should have no impact on NGDP. The rise in the real interest rate would be offset by a fall in expected inflation, and nominal rates need not rise.

    2. The stock rally probably did signal greater NGDP growth expectations. The 5 year TIPs spread rose from 0.6 to 1.0%. Real growth expectations probably also rose, as indicated by rising commodity prices. Unfortunately, we don’t have a precise measure of NGDP growth expectations.

    3. No.

    4. Mankiw discussed the changeover in his blog around early December. It did distort the data somewhat at that time. Right now the TIPS spread is 1.0% over 5 years. If there is some risk of 5 year deflation, the actual amount might be lower. This could be checked by looking at 10 year TIPS with five years remaining until maturity. My hunch is that they would no longer differ greatly. If they did, it would indicate residual fear of deflation.

  16. Gravatar of Joe Calhoun Joe Calhoun
    17. May 2009 at 06:35


    Thanks for taking the time to answer. You’ve been great about interacting on this blog, especially with those of us who aren’t economists. I worry about asking the proverbial dumb question sometimes, but I hope I’m adding to the conversation. Thanks to some of the other commenters too such as Bill Woolsey. I’m just trying to continue learning and the opportunity to interact with professionals is greatly appreciated.

  17. Gravatar of TVHE » Hand waving, model making, and variable lags TVHE » Hand waving, model making, and variable lags
    18. May 2009 at 12:08

    […] would suggest that anyone who is interested in economic models reads this post, namely because I agree with it […]

  18. Gravatar of Steve Steve
    18. May 2010 at 07:59


    a question on the lags: I have tried to find some more evidence on lags of monetary policy. Aner whereever I look, it is usually taken as a fact. Books and articles just use it, not explore it or quesiton it.

    Mishkin’s book also just takes lags as given.

    So my question is this: who else (except you) thinks that lags are neither long nor variable, but actually a myth? And if it is only you, why do other people (good economists I suppose) keep that view, even though it makes no sense with rational expectations? Are there any reasons for why there might in fact be lags?


  19. Gravatar of ssumner ssumner
    19. May 2010 at 07:34

    Steve, Sorry, but I just don’t know what readings to suggest. My sense is that there are some lags for broader aggregates, but less than most think (as monetary shocks are misidentified.) But I also think macroeconomists should focus much more on asset prices–for which there are no lags at all.

  20. Gravatar of Steve Steve
    25. May 2010 at 05:30

    Thanks, Scott. One other question: I remeber you linked to a Nick Rowe post on some transition that the Bank of Canada made, and it worked (temporary higher inflation, or transition to lower inflation I don’t remeber exactly). I cannot find that post, could you help me?


  21. Gravatar of ssumner ssumner
    25. May 2010 at 08:27

    Steve, Sorry, I don’t recall. Try googling themoneyillusion and key words like Nick Rowe.

Leave a Reply