Past tense please

People that develop new ways of looking at the world are most successful if they can create a new language, a new set of metaphors.  I don’t have the literary skills of a Keynes or a Freud, but then again I also don’t have such grandiose theoretical objectives.

For instance, quite a few people seemed to misinterpret my Mishkin post, which was actually intended to be complimentary.  The point was that if my views are similar to Mishkin’s, how odd can they be?  And yet I am in a very tiny group of people who believe the Fed caused the crash of 2008, so there must be something at least a little bit peculiar about my perspective.

When you are trying to get people to look at the world in a slightly different way, it is easy to be misunderstood.  After my QE post last month, some other bloggers observed that I expected the policy to fail, unless other steps were taken.  That’s not quite right.  I argued the policy had already failed.  That’s past tense.  How could I argue that the policy had failed after just a few hours?  Because in my view the policy objective should not be NGDP, it should be expected future NGDP.  But (you ask) surely the ultimate goal must be to boost actual future NGDP?

Here’s how I look at it.  The Fed should merely worry about targeting expectations.  Until and unless they show any evidence of being smarter than the markets (and right now it doesn’t look too promising) I judge their actions solely based on the impact on expectations.  If someone offers you even odds on a dice role, and lets you choose either 1 through 4, or 5 and 6, the smart bet is 1 through 4.  If you bet 5 and 6, and one of those two turns up, it doesn’t make it a smart bet—it’s a foolish bet that was bailed out by a lucky bounce.  As a citizen and investor I will be thrilled if NGDP grows rapidly over the next 12 months.  And I think there is a possibility that it will.  But right now the odds are stacked against the Fed.  And the Fed is the only institution that has the power to stack the odds in its own favor.  The Fed’s failure to do so is a policy failure, whatever happens subsequently.

This is a new way of thinking about macroeconomics.  I just read the General Theory’s Chapter 12, at Tyler’s suggestion, and will do a post on it in the near future.  Keynes talks about speculative markets as if they are casinos.  Right now we have a central bank that likes to gamble, and likes long odds.  It is no surprise that Wall Street is putting its money on the short odds—policy failure.  Unlike Keynes, I think it’s a big mistake to bet against the market.

[Finally, a short post!]



31 Responses to “Past tense please”

  1. Gravatar of 123 123
    1. May 2009 at 04:05

    Stock market is a good proxy for expected NGDP. Stock market crashed in October, not September. Together with other central banks, fed made an unscheuled cut on October 8, in retrospect we can say that they miscalibrated the cut. The big mistake was in September, but NGDP futures would not have helped.

  2. Gravatar of David Ortmeyer David Ortmeyer
    1. May 2009 at 04:43

    …and what “future expected NGDP” data does the Fed look at to assess if they are on track? Is the stock market the best measure or are there tighter measures (e.g.the TIPS spread), or some index. And do we know that the causality will always run from Fed ==> future expected NGDP (with no feedback loop)?

  3. Gravatar of ssumner ssumner
    1. May 2009 at 07:19

    123, I don’t think you understand a NGDP futures targeting monetary regime. First of all the stock market had already declined quite a bit between May and the end of September, so even if we accept that the stock market is the best proxy for NGDP expectations (and I certainly don’t accept that) the stock market was signaling that money was too tight long before October. More importantly, under an NGDP futures targeting regime the expected growth in NGDP always stays exactly at 5%, so if you believe that sharply falling NGDP growth expectations caused the early October crash (and I believe that is a reasonable assumption), then by your own argument the October crash never would have occurred under NGDP futures targeting. So your comment is logically inconsistent.

    David, I’m not sure what data they use, but if I were the Fed I would mostly rely on market data, but also use some current economic data, as well as the consensus of private forecasters. I think the single best indicator we have is the yield spread, which currently shows about 0.8% inflation over the next 5 years. Real growth is a bit tougher, but stock and commodity prices (when they move together) are one indication. I suppose they could also look at the yield spread, but I don’t think that is as reliable. Obviously I favor the creation of a subsidized NGDP futures market, which would be far better than anything we have today. But even without that market, the Fed has clearly known for some time that there were off-target, and has indeed admitted as much.

    I hope there is some reverse causation, I would hate to think that the Fed doesn’t react at all to changes in the economy. I assume they do react. My criticism is that they don’t react strongly enough to stabilize expectations.

  4. Gravatar of ssumner ssumner
    1. May 2009 at 07:31

    David, I forgot to mention that Intrade has a contract for RGDP. I shows a strong chance of GDP falling in the second quarter, and to a lesser extent in the 3rd quarter. The 4th quarter is 50/50.

  5. Gravatar of 123 123
    1. May 2009 at 10:33

    Stockmarket is an appropriate proxy only for sharp corrections of expected NGDP.

    NGDP targeting regime would have helped in October (by ensuring that the “Greenspan put” is an official policy). But in the first three weeks of September the difference between the actual policy and policy under NGDP targeting would have been too small to prevent the crash.

  6. Gravatar of Rafael Rafael
    1. May 2009 at 14:12

    I´m waiting anxiously for your post on GT´s chapter 12. I´d suggest some comments on Keynes appraisal of the Fisher effect as well.

  7. Gravatar of JimP JimP
    1. May 2009 at 14:47

    Exactly – except that the sentence “The Fed should merely worry about targeting expectations” should be “The Fed should ONLY worry about targeting expectations”

    As I have said before, the Fed keeps telling us their sad and unhappy expectations – thus setting our expectations lower and lower. They just really really don’t understand their job at all. They need to set our expectations, not tell us theirs. It is like some insane slow motion nightmare drift. They could do something about it and they just flat flat refuse to do so. One wonders if they really don’t actually want just what we now have – just as the Bank of Japan claimed to not want deflation when they actually did want it.

  8. Gravatar of JimP JimP
    1. May 2009 at 14:57

    When you think about it – I think the Fed does in fact want exactly what we now have. I get the feeling that they are pleased to be able to say that(their) expected inflation is quite low. This allows them to expand the Fed balance sheet without Volcker and the WSJ going ballistic. They really do believe that the way to fix this is to have the Fed screw around in particular credit markets rather than doing the general (strongly positive NGDP) fix. Just as Obama is not all that unhappy about owing particular car companies and banks.

  9. Gravatar of ssumner ssumner
    1. May 2009 at 15:01

    123, I am puzzled by your response. Most people think the crash of early October was caused by expectations of a deteriorating economy. Obviously if that was the cause (and by deteriorating economy I mean falling AD) then the stock market crash could not have happened under NGDP targeting. Now I would concede that the stock market crash might have been due to some other factor (as in 1987) but you started this discussion with an assumption that the stock market was strongly correlated with NGDP expectations. In any case, if it had been due to some other factor (as in 1987) who cares? That crash didn’t hurt the macroeconomy at all, as long as you have 5% NGDP growth expectations, you don’t have to worry about demand-side recessions. Stock market crashes by themselves aren’t problems at all, they are only a problem if the stock market crashes BECAUSE NGDP expectations plunge. (Obviously a supply shock might still have caused a recession. But it is hard to find a single serious recession in all of American history caused by supply shocks.)

    I also don’t see the point of the September/October distinction. Stocks crash at the precise moment expectations get worse. If expectations aren’t getting worse, stocks won’t crash. There is no “Granger Causality” in financial markets. One doesn’t have events in September causing stock crashes in October.

    Rafael, I forget, does he discuss the Fisher effect in Chapter 12? If so, I’ll discuss it.

  10. Gravatar of ssumner ssumner
    1. May 2009 at 15:07

    JimP, I go back and forth on this. I do think it is probably a mixture of bad judgment and bad priorities. The bad judgment is not being forward-looking enough, the bad priorities is worrying more about inflation than deflation (when the latter problem is worse over moderate ranges (say having inflation 3% too low vs. 3% too high.)

    I recently saw some international Phillips curve data showing the relationship is much stronger in countries that have experienced some deflation. This tells me that it’s not completely symmetrical–the costs of policy errors rise sharply as inflation approaches zero, and then goes below.

  11. Gravatar of JimP JimP
    1. May 2009 at 15:56

    Which is rather odd. Given Bernanke’s background one would think that he of all people would recognize that deflation is much worse than inflation.

    I do think it is to some degree a matter of personality and basic outlook. Neither Bernanke nor Obama have Roosevelt’s joyful ebullience nor his amazing self-confidence. Look at the deflationisis now – sad grim folks rather glad we are getting exactly what we deserve. (think Buiter or Simon Johnson or Krugman – economic moralists all)

  12. Gravatar of DG DG
    1. May 2009 at 23:36

    Mr Summers – I love reading your blog and your thinking aloud is always entertaining. But sometimes your arguments frustrate me, and this one particularly so.

    On the Fed, who you blame for the crisis, you also argue that “Until and unless they show any evidence of being smarter than the markets (and right now it doesn’t look too promising)” they should only target expectations. But if there is no evidence of them being able to beat the market, why should they be targeting anything at all?

    Take any index – stocks, bonds, commodities, “balanced” – and you’ll find that nearly all investors don’t beat the market. There are very good explanations as to why that is true. Why are the Fed, or any other group of central bankers expected to be so different? Didn’t the Fed cause this crisis in the same way most investors underperform the whatever index they benchmark against – by anchoring around recent experience, overconfidence in their own judgement, and over estimation of the likelihood of “representative” outcomes to name but a few? And this faliure wasn’t just in missing the housing bubble, it was in creating it in the first place, as it was in creating the tech bubble before it etc … Why should anyone at all, let alone (no offence) academics from ivory towers have the foresight required to hit the “optimal monetary policy?”

    If the answer is, “no reason at all” then why should we have an activist central bank at all? I bet you own index tracking stock and bond portfolios and I bet you recommend to your students the same. And if we all purport to have such faith in markets, why can’t rate setting be equally passive?


  13. Gravatar of 123 123
    2. May 2009 at 02:02

    There are two important lags (market information processing lag and monetary policy lag):
    1. In September there was a run on a financial system, but the future consequences of the run were not reflected in the market expectations until October.
    2. In October there was a huge pressure on market expectations, Fed’s reaction under NGDP targeting regime would have helped, but mostly by affecting NGDP in second half of 2009 and later.

  14. Gravatar of Bill Woolsey Bill Woolsey
    2. May 2009 at 02:31


    Sumner believes (plausibly,) that nominal inocme falls now because it is expected to be lower in the near future. If nominal expenditure isn’t expected to be lower in the future, then it will not fall much now. In his view, consumption fell in the fourth quarter of 2008 and remains well below where it was in the third quarter of 2008 now, after the first quarter of 2009 because people expect real income and nominal income to stay low for a protracted period of time. Investmnet fell in the fourth quarter 2008 and the first quarter of 2009 because firms expect nominal and real income to remain low for a protracted period of time. If nominal income was expected to be back to its 1996-2008 growth path by the fourth quarter of 2009, the factors that caused a drop and nominal expenditure and real output in later 2008 and 2009 would largely have not existed.

    I think there is something to this. On the other hand, an important factor is that the changes in the quantity of money now that would be needed to keep nominal income on target in 2009 would offset any shortage of money resulting from changes in the quantity of money now.


    The purpose of Fed policy in Scott’s view (and I agree) is not to “beat” finanical markets, but rather stabilize growth in spending on final goods and services in the income. Financial markets cannot do that because the Federal Reserve has a monopoly on the issue of base money, and the “price” of base money is just he reciprocal of the price level. The way the market system adjusts the price level is through changes in spending on final goods and services (among other things.) Free banking reforms try to reduce the scope of that monopoly. Depending on the nature of the reform, they may leave a base money whose market still must clear through changes in the price level and so, through changes in aggregate expenditure.

  15. Gravatar of ssumner ssumner
    2. May 2009 at 04:53

    Good questions:

    DG, I agree with your comment 100%, which is exactly why I don’t want an activist Fed. I’d rather have the markets set the money supply at a level where the market expects 5% NGDP growth. That’s my NGDP futures targeting proposal. If they can’t do that for some reason, I want then to use whatever market indicators are out there, not their own judgment.


    Suppose the goal is 2% inflation. Suppose the Fed currently expects 3%, but the TIPS market indicates a market expectation of 1%. I want the Fed to go with the TIPS, not their own judgment, and tighten policy.

    123, There is no information processing lag in financial markets (at least not exceeding 24 hours.) But I agree with you in this sense, it took a month for the markets to realize Lehman was going to result in a steep recession. But that was because new information came on the scene in early October. The new information was that the Fed was distracted by bank bailouts and planned no aggressive monetary stimulus. Another new piece of information was that (due to interest on reserves) we were essentially in a liquidity trap and the Fed obviously had no anti-liquidity trap policy (despite Bernanke’s earlier statements that the Fed could handle a liquidity trap.)

    2. The “huge pressure on market expectations” in October was fear of falling NGDP growth for years to come. NGDP would have prevented those fears in the first place. But let’s say your right about the fears still being there. Would NGDP have fallen significantly? No. And it will be helpful to start with an analogy. Suppose the price of gold is $800 and the Fed starts pegging 12 month forward gold at $840 (5% higher.) How likely is it that gold will fall to $700 in the short run? Not impossible, but not likely either because arbitragers could take advantage of the discrepancy between short and long run gold prices.

    You might say gold is nothing like NGDP, and that’s right, but this analogy is more apt than you might think. By far the most powerful factor affecting current movements in NGDP is expected future movements in NGDP. This isn’t just my view, but the view of modern macro in general (both new Classical and new Keynesian.) And they are right. If market are expected to be fine in the second have of 2009, then:

    1. Business investment that takes a year to complete should be strong.

    2. Current housing prices will reflect the expectation of healthy economy in late 2009, not a weak economy.

    3. Bank balance sheets today will be much stronger, as the value of loans will be higher.

    But suppose I am wrong. Your worst case seems to have been a weak economy for a few months, but back to 5% unemployment by the second half of 2009. I’ll take that any day over the estimated 10% unemployment we will have in the second half, and which will stay high for years.

    To summarize, I’d never claim that 12-month NGDP futures targeting can work miracles, obviously it doesn’t completely stop instability in the very short run. But I think that instability would be far less than people might imagine, because expected future NGDP has a powerful effect on current NGDP growth.

    Bill I agree with your comments. The only point I would add is that although changes in the quantity of money are essential, and my plan calls for the optimal change in the quantity of money, the most important thing is that any policy have credibility. Even without NGDP futures targeting the Fed can do much better. In my view even if the quantity of money had not changed appropriately at the September 16 meeting last year, it would not have snowballed into a crash in early October if the market had expected an aggressive “make-up” in October. Exactly such an aggressive make-up occurred in January 2008 (125 basis points) after the blown call in December 2007. The January moves kept nominal and real GDP growth in the first half of 2008 at adequate levels, despite both oil and financial shocks. As long as money is expected soon, velocity adjustments can tide the economy over. What kills an economy (1929, 1937, 2008) is when the market realizes monetary policy is likely to be too tight for years.

  16. Gravatar of Jeremy Goodridge Jeremy Goodridge
    2. May 2009 at 08:01


    A suggestion: it might help if you made a timeline chart showing all the relevant events together (Fed actions, lehman collapses, stock market moves, inflation expectations (as measured by TIPS differentials), consumer spending changes, business investment changes, housing prices, etc, etc). Then it might be much easier to see what affected what — and perhaps could help convince the community of the causal relationships (and that Fed action triggered the crisis).

    I’d be happy to help out in creating such a timeline chart, if you think it would help. Any events/indexes you think should be included?


  17. Gravatar of 123 123
    2. May 2009 at 12:29

    Scott, actually my worst case scenario under NGDP regime is a bit different – weak economy in Q4 2008 (NGDP 0% vs -5.8% actual) and return to NGDP growth of 5% (with RGDP of 1%) in Q3 2009. Why RDGP 1% in Q3 2009? The answer is structural problems in residential and commercial property market and structural problems in financial sector.

  18. Gravatar of Bill Stepp Bill Stepp
    2. May 2009 at 17:45

    JimP writes:

    […] Given Bernanke’s background one would think that he of all people would recognize that deflation is much worse than inflation.

    Actually inflation is much worse than deflation. After all, it benefits the government–a chronic debtor and “the biggest mass murderer, armed robber, enslaver, and parasite in all of human history,” as Rothbard put it.
    And as George Selgin has pointed out (_Less Than Zero_), in a a free market (monetary system, but this qualification is reredundant) the natural tendency would be for the rise in productivity to cause a gradual deflation of prices. This would leave factor owners, including workers, better off in real terms, contrary to the infaltion apologists, such as Bernanke and other Fed Heads.
    Selgin notes that the prejudice against deflation exhibited by most professional economists and economic pundits is wrongheaded and exists because of (1) the long memory of the bad sort of deflation of the Great Depression (which was accompanied by falling output and employment), and (2) the failure to understand the difference between the bad and good (free market, productivity inspired) sorts of deflation.
    Bernanke was recently quoted as pining for 2% annual inflation. What he didn’t say was that he was advocating a hidden tax on holders of money balances. At 2% that means the real value of a dollar declines by half in 36 years. Guess who benefits? Uncle Sam, the Mass Murder Inc. and Red Ink Man, that’s who.
    Give me deflation any day.

  19. Gravatar of ssumner ssumner
    3. May 2009 at 04:22

    Jeremy, Several people have mentioned that idea, but for reasons discussed in my “How Did Friedman and Schwartz Persuade Us” post, I don’t think that exercise would help as much as people might imagine. I did several posts showing graphs of all the key variables that I could think of, and put 5 of them them (separately) in the post called “Some Graphs” and then another one a few days later on real estate and industrial production. The bottom line is that all seven of these key indicators were getting a bit worse even before Lehman.

    If you think it would be helpful to put all seven of those graphs on the same chart, along with Lehman and the Sept 16 Fed meeting, and the October 6 decision of interest on reserves (two Fed mistakes in my view) you are welcome to try. I don’t know how to do that stuff, but I’d be glad to post the result if you think it would be helpful.

    Thank you.

    123, Now I think I see your confusion. You seem to assume that 12 month NGDP futures are aimed at getting a rate of 5% NGDP growth 12 months from now. But that is not the idea at all. Rather the idea is to get 5% NGDP between now and 12 months from now. If you look at the numbers you provide, they would result in much less than 5% growth over the next 12 months. If you redo the exercise the way I described it, you’ll see it’s really hard to squeeze a recession into a plausible growth path (unless of course the market is very bad at forecasting, which is always possible.)

    Bill, The benefit of 2% inflation to the U.S. government is too trivial to even talk about. Speaking of mass murderers, the German hyperinflation was followed by democracy, the German deflation of the early 1930s led directly to Hitler. So I don’t see deflation as any sort panacea. The first half of the 20th century the developed countries were far more likely to commit mass murder than during the post-1968 fiat money era, So Rothbard’s reading of history is very different from my own. Government revenue from money creation is tiny compared to tax revenue.

    There is no evidence that inflation could not occur under a free market monetary system. Inflation often occurred under a commodity standard, once for a continuous stretch of nearly 150 years (1500-1650.)

  20. Gravatar of 123 123
    3. May 2009 at 05:05

    “You seem to assume that 12 month NGDP futures are aimed at getting a rate of 5% NGDP growth 12 months from now. ”
    Not really. My assumption was that in 2008 Oct fed targets a series of quarterly NGDP futures (from Q4 2008 to Q3 2010 – each of them separately).

    My worst case scenario assumes big structural adjustment problems (breakdown of monetary transmision mechanism, continuing issues in real estate sector etc.). Note that under my worst case scenario Fed has to pay out lots of money at the expiry of Q4 2008 and Q1 2009 contract.

    My best case scenario is much more optimistic (slowdown in October-Novoember, recovery in December 2008).

  21. Gravatar of Bill Stepp Bill Stepp
    3. May 2009 at 06:58


    If the U.S. government steals one penny from inflation, that’s a crime–a small one to be sure, but theft is theft.
    Government revenue from inflation is small compared to direct tax-theft, but it’s still a crime.
    While the German hyperinflation was followed by democracy, it wasn’t one I’d have liked to have lived under. The German deflation of the ’30s was the bad kind of deflation; the idea that it led directly to Hitler (like Hitler = f (deflation) is very simplistic, even if it encouraged him and his followers.

    I’m not arguing for deflation as a panacea. It’s part and parcel of a free banking system with a gold standard, and that is a lot better than central banking and a fiat standard. Plus Bernanke would have to get a real job.
    Mass murder was more prevalent when criminals like Wilson, Kaiser Wilhelm, Lloyd George, Churchill, Hitler, Stalin, Tojo, and FDR strode the world stage than in the post-1968 era.
    But that is a lucky accident. I don’t see deflation as a causal link to more mass murder.
    Think of the 1815-1914 era–the century of (relative) peace. There was a gold standard, and some countries had free banking, albeit usually in state-truncated form (like the U.S.), and deflation.
    There is a book waiting to be written on the historical relationship between the institutions leading to free market money, and how they acted as a barrier to war.

  22. Gravatar of JimP JimP
    3. May 2009 at 07:22


    Here is a link to the St. Louis Fed. – they have a real good start on a timeline.

  23. Gravatar of 123 123
    3. May 2009 at 11:57

    A question – could you clarify if you favour NGDP targeting that pegs a multiyear path of NGDP (i.e. if actual NGDP growth is a bit too low one year, NGDP growth target next year is 5% plus last years underperformance), or do your favour a policy regime that pegs NGDP at last years actual NGDP plus 5%?

  24. Gravatar of ssumner ssumner
    4. May 2009 at 04:24

    123, The Fed can’t target more than one futures contract at once (the one tool, one target constraint.) I prefer they target a point estimate of NGDP, computed as a weighted average of two quarterly reports. Each day they target a new one, always exactly 12 months out.
    I would have the Fed try to avoid taking significant net long or short positions, but rather mostly just make the market. This would require them to estimate base demand. Right now that sounds really hard, but remember there is a new contract each day, so they’d learn fast. During normal times it is easy to get a good daily estimate of base demand.

    Bill, A tiny bit of inflation is a crime? Well we can’t even measure inflation very accurately.
    I also like some aspects of 1815-1914. But let’s not be too nostalgic. There was slavery, a Civil War, a Franco-Prussian war, Crimean war. Yes, 1914-45 was even worse, but since fiat money came around in 1968, even a Franco-Prussian war seems unlikely. A lot of libertarian/Austrian types (I don’t know your views) seem too negative about the modern world. There are some bad trends, but some good ones as well.

    JimP, Thanks, that’s a very complete timeline.

    123, I prefer a multiyear path (called “level targeting”), although I’d be fine with either. I need a post on this plan, there are lots of questions on different posts.

  25. Gravatar of Bill Woolsey Bill Woolsey
    5. May 2009 at 03:16

    Scott, the Fed can target more than one futures contract. There is a tool, (buyng or selling the contract) for every target (each contract.)

    What you meant was that it can’t govern its open market operations in bonds in such a way that the quantity of base money will be adjusted so that expected nominal income will be on target.

  26. Gravatar of ssumner ssumner
    5. May 2009 at 05:17

    Bill, OK, but the term “target” suggests “fixes.” Only with the power of the printing press can they be confident of fixing the price. If they use fiscal resources they may be able to peg some other contracts where they want (by taking a strong net short or long position.) Is this what you were assuming?

    I did discuss this idea in an earlier paper (JMCB, 1995) and the more I think about it, the more I think you are right. I argued that the fiscal resources used to peg other contracts would eliminate the “time inconsistency problem” Now the Fed would have an incentive to keep the policy going year after year, so as to not lose money on it’s longer term bets.

    Is that what you had in mind?

  27. Gravatar of Scott Lawton Scott Lawton
    5. May 2009 at 06:23

    @Jeremy Goodridge: I think an effective chart of the appropriate data, marked with events and short explanations is VERY important. I don’t have time to put one together, but I would be happy to offer feedback. If I think the result is compelling, I will try to give it some visibility. (I can also host it on its own domain if that’s useful.) To contact me, just click the link for info … though using my first name for the email address may be a little more reliable.

  28. Gravatar of ssumner ssumner
    6. May 2009 at 05:54

    Thanks Scott.

  29. Gravatar of DG DG
    7. May 2009 at 23:31

    Bill Woolsey and ssumner (been too tied up to respond recently so apologies for the delay) …

    Bill – when I drew parallels with investors “beating the market” I wasn’t trying to suggest central bankers do the same and I hope I didn’t push the parallel too far. What I was trying to argue was that the skills required of a central bank to successfully and consistently and smoothly navigate the economy, and avoid the mistakes which add to that volatility are similar to the skills consistently outperforming investors possess – primarily to see when the consensus is wrong and have the fortitude to act accordingly. Needless to say, such investors are in the minority. For investors this is difficult enough but for central bankers I think it is impossible because there is no monetary incentive to step away from the crowd in the way there is for investors (note that even that monetary incentive doesn’t seem to work too well and that only 15%, not 50% of investors tend to “beat” the market).

    The entire model of activist central bankers (and therefore a large chunk of macroeconomic theory?!) falls down because it places an expectation on central bankers to know more than everyone else and to act on that knowledge which is as unrealistic as the expectation for fund managers to beat the index.

    As for the inevitability of central banks setting price because they are monopoly issuers … such a practice is actually an abuse of monopoly power. The free banking idea is one I find very interesting and I have a lot of sympathy for. But it’s a different argument and in my experience seems a bit of a step too far for some people. Mention it and everyone thinks you’re a militant libertarian. So I try to leave it out of my arguments. Maybe you’re right though and it’s actually the same argument. I guess I’m trying to figure out in my own mind if that’s true and if free banking necessarily follows from restrictions on central bank rate setting (and I’d love some help here). Other monopoly abuses aren’t tolerated, why should this one be?

    ssumner – I think your argument really is very neat and intellectually elegant. But I have two problems with it. The first is conceptual – why target a rate of growth if you don’t know what the underlying trend rate of growth is? If you target a rate which is incorrect or inappropriate you will cause distortions in one way or another. You can’t make an economy grow consistently above (or below) its trend rate, for example, any more than you can hold an exchange rate above (or below) its equilibrium level, prop up or hold down a stock market, a housing market etc. Any attempt will be ultimately futile and will create distortions elsewhere. Also, I know we can all run clever techniques, filters, and other econometric WMDs but let’s be honest here – we have no idea what trend growth is at any point in time. The Japanese rate of trend growth has fallen spectacularly in the last few decades. Who saw that coming? Who knows where it’s going?

    The second is more technical. How liquid are the financial markets you need? I know there are CPI and GDP futures, but have you ever tried to trade them? They are close to meaningless and can be swung around by all sorts of things which have little to do with forecasts of the economy. Trying to extract market expectations of the future is worthwhile, but you are inevitably subject to a real-time reading on crowd madness, and all the volatility that entails. Remember the Kansas Fed’s clever model on inflation expectations which looked great until you needed it because the markets went nuts last autumn, and all sorts of previously dormant risk premia kicked in rendering the reading a nonsense? How would you conduct policy in those circumstances?


  30. Gravatar of ssumner ssumner
    8. May 2009 at 03:55

    DG, I think you misunderstood my proposal. I did not argue the Fed should target real GDP, which would be madness, as you indicate. Rather I favor they target NGDP, for which there is no “underlying rate of growth.” Central banks set the rate of growth for NGDP, whether the are “active” or “passive” in their policies, not the market.

    I don’t know about the KC Fed problem you mentioned, please send me a link if you can find one. I do know that last fall the TIPS market was screaming deflation—if only the Fed had paid attention we wouldn’t be in this severe recession. The type of market I envision is a stand alone futures market, so it would be much less distorted by risk premia than markets like TIPS, which are also impacted by liquidity differences between TIPS and conventional bonds. The market would be subsidized to insure adequate liquidity, so that is not a problem. If you think the markets are irrational—let the Fed forecasting unit trade contracts as well, that would make the market more efficient, and the taxpayers could make a nice profit from fleecing the stupid speculators who trade in the market.

  31. Gravatar of スコット・サムナーのマネー観、マクロ観 by Scott Sumner – 道草 スコット・サムナーのマネー観、マクロ観 by Scott Sumner – 道草
    9. October 2011 at 00:10

    […] マクロの標準的な見方では金融政策および財政政策ツールの組み合わせは、「長く可変のラグ」を伴ってマクロな経済に幅広く作用するとされる。その伝達メカニズムは「ブラックボックス」のようなものの中にあってほとんど分かっていない。つまりまず政策行動がこの謎の箱の中に入り、一年後にマクロ経済的な結果となって飛び出してくる。VARモデルを使ってその効果を見積もることはできる。私はこれは資産市場が非効率であることを暗黙に仮定していると考えるので、そのような物の見方は受け入れられない。インフレーションに織り込まれる長期の効果は直ちにTIPSスプレッドとCPI先物価格に反映されるはずだ。それが正しい政策評価の方法である。   […]

Leave a Reply