You can’t please everybody

Several commenters sent me the following by Michael Sankowski, insisting it raised serious questions about my NGDP futures targeting idea:

I never wanted to write this post.

I might not be the world’s biggest fan of monetary policy, but I like NGDP level targets. A NGDP level target would be superior to our current dual mandate structure.

But NGDP level futures contracts are a very bad, even terrible way to moderate the level of NGDP. I thought the fascination with NGDP level futures would have played out by now, without me doing a full on smackdown. But that hasn’t happened.

Instead, we get Noah Smith arguing trading is volatile. It’s not a great reason to avoid NGDP futures. It’s a bit thin.

So I am compelled to write a post I did not want to write.

NGDP level futures would almost certainly hand Goldman Sachs and hedge funds a payday worth over $500bn, while giving almost nothing else to the rest of the economy. Either that, or NGDP level futures would never be traded by anyone.

I wish Michael hadn’t written that post, as I don’t like having to keep replying to criticism by people who have never even studied the proposal.

Noah Smith pointed out that if NGDP futures traders were much dumber than the Fed, they might inject extra volatility into NGDP.  I pointed out that I’d have no objection to the Fed doing its own trading to offset suspicious trades.

Now Michael raises the opposite concern.  The traders might be so smart that they steal $500 billion from the Fed.  I wonder why he didn’t pick $500 quadrillion? Seriously, I’d argue the market need be no larger that $100 million, maybe $10 million.  And I’ve published proposals where the Fed would not take a net long or short position.  So under those versions any $500 billion earned by GS would be offset by losses to other traders.  And I don’t care if no one trades the contracts, as long as the central bank ensures liquidity and a low cost of trades.

More generally, the futures targeting idea has been around for a quarter century, with many articles published by myself and others.  There have also been critiques, the best one by Bernanke and Woodford  (JMCB 1997.)  Garrison and White also did a good critique.  But those critiques only apply to one version, which is not the one currently being pushed by people like me and Woolsey.

I would add that the “futures” market being proposed is unlike any real world futures market.  It is set up not to provide forecasts of future NGDP, but rather to forecast the setting of the policy instrument likely to result in on-target NGDP.  So any knowledge one has of actual futures markets is likely to be very misleading when applied to this policy regime.

There may be a deadly flaw.  But if no one has found it after 25 years and endless debate, I wouldn’t count on it.




30 Responses to “You can’t please everybody”

  1. Gravatar of dwb dwb
    14. July 2012 at 12:14

    I read the “about me” talking about disagreements about operational realities and i think oh boy… and we are handing goldman sachs 3% of gdp based on what exactly? maybe this guy has designed some futures contracts but there sure is some dopey analysis – I just don’t have time to find the flaw(s).

  2. Gravatar of dwb dwb
    14. July 2012 at 12:24

    here is my off the cuff thought:

    they can only make money to the extent they can outguess the market, and if they can make 500 Bn they can also lose it. no risk manager on earth (or regulator) will ever approve that. with ngdp level targeting unexpected ngdp misses will be random, so a bank will be unable to make money consistently. moreover, with the fed ngdplt, the volatility of ngdp will be a lot lower so we have no idea what the true but lower volatility of ngdp will be. even if the volatility of ngdp turned out to be +/-1% (we are talking about 1 std dev) then thats 37.5 Bn *per quarter* (=15000 Bn * 1% /4) probably distributed across many firms, and the stock market can go up that much is one day so in comparison this is pretty paltry.

  3. Gravatar of 123 123
    14. July 2012 at 12:44

    Scott: “Noah Smith pointed out that if NGDP futures traders were much dumber than the Fed, they might inject extra volatility into NGDP.”

    I am writing here to clarify my position that this extra NGDP volatility created by the markets would be much smaller than the volatility of NGDP that the Fed has caused. This was my position all along.

  4. Gravatar of curiouseconomist curiouseconomist
    14. July 2012 at 14:00

    Scott, as a step towards a futures market, we could implement pay by performance for central bankers. I expect the FOMC would pay more attention to the dual mandate then! This wouldn’t be as efficient as a futures market, but it would be an improvement.

    I do think your futures market idea will come to fruition at some point (given neoliberal trends). It actually ties in well with Robin Hanson’s “futarchy” idea: “elected officials define measures of national welfare and prediction markets are used to determine which policies will have the most positive effect”. That’s not going to happen any time soon either, but hey, maybe we should pay elected politicians by performance too until then!

  5. Gravatar of Negation of Ideology Negation of Ideology
    14. July 2012 at 14:16

    I’m not an expert on futures, so I may be missing something obvious, but I have a question. Wouldn’t it be easier to have the Treasury sell NDGP linked bonds, like it currently sells inflation linked bonds? Then the Fed could just target the spread between ordinary Treasury bonds and NGDP bonds at an implied rate of 5% NGDPLT growth.

  6. Gravatar of John Hall John Hall
    14. July 2012 at 15:43

    I don’t see how you don’t think it’s a flaw that it’s different from any other futures contract that exists. Practical considerations are as important as theoretical ones.

  7. Gravatar of Steve Steve
    14. July 2012 at 15:53


    I believe the NGDP futures idea is to create futures contracts *CONDITIONAL* on policy instrument settings, so the Fed can choose the setting most likely to produce the desired outcome.

    I’ve advocated the NGDP linked futures before, although Scott has pointed to the circularity problem.

    Personally, I like trying both ideas. I like NGDP linked debt, because there are political economy benefits, i.e., giving creditors a vested interest in economic growth.

    Some worry that even conditional NGDP futures could be manipulated by Goldman. But I don’t see the harm in creating a market and then letting the Fed use it in an advisory role.

    Finally, I do like the idea of making the central banker pay accountable for hitting near targets. There could even be salary and pension clawbacks if the Fed misses NGDP trend by over 10% in a five year window.

  8. Gravatar of curiouseconomist curiouseconomist
    14. July 2012 at 16:20

    Negation of Ideology,

    That would run into the “circularity problem” discussed by Bernanke and Woodford (paper is cited in Scott’s post). Basically, as I understand it, trying to hit an “NGDP Spred” that way (or an inflation target using TIPS spread) would essentially make the spread itself useless information because nobody would bother to trade (as the Fed would offset every move that changed the TIPS spread). That’s how I understand it anyway.

    Scott’s proposal doesn’t fall into that trap because he isn’t trying to hit a measure of NGDP expectations. He wants to use the markets NGDP expectations to set base money (ie. base money would be endogenous) on the correct path for the actual NGDP value targeted. Better explanation of Scott’s position in this old post:

    John Hall,

    I find that argument to be somewhat reaching. After all, the US dollar is the global reserve currency and massively influences the world market. There are trillions of transactions taking place in USD. My point is that with such a large audience, it’s only going to take one explanation of how the system works to get things going.


    I think the NGDP futures proposal you’re talking about is the one that uses interest rates as an instrument (rather than the money base). Here’s Scott’s explanation:

    “You set up a number of NGDP futures contracts auctions, each contingent on a different setting of the fed funds target. You announce that the contracts will only be exercised in the auction that, ex post, most nearly balances the long and short positions of NGDP futures speculators (which means the Fed takes little risk.) And then you set the fed funds rate at that level for the next 6 weeks.”

    I think Scott’s preferred proposal has the Fed standing willing to buy and sell NGDP futures at the NGDP value (or for more liquid trading, a fraction of NGDP) it wants to hit. The private sector then just trades with the central bank using base money, which stabilizes NGDP growth.

    Regarding manipulation of these markets, I don’t think it’s really important because we’re talking about monetary policy here. The Fed has unlimited base money creation power that it can use to push things closer to its own internal forecasts if it thinks there is manipulation. There is no real way to win against an unconstrained central bank.

  9. Gravatar of Benjamin Cole Benjamin Cole
    14. July 2012 at 19:01

    The futures market is an interesting idea, but it seems to trigger fears.

    PR-wise, I prefer Market Monetarists stick to a QE program that targets NGDP growth. That is, instead of announcing QE of $100 billion a month for six months, try announcing $100 billion in QE until an NGDP growth target of 6 percent is hit for four or six consecutive quarters.

    If the QE does not work, then we will be wiping out of national debt (I advocate only buying Treasuries).

    Egads, if a shrewd businessman was running the federal government, would he not seize at a chance to eliminate debt while spurring growth? We may be passing up once-in-a-lifetime opportunity to wipe out debt by the trillions.

  10. Gravatar of ssumner ssumner
    14. July 2012 at 19:04

    dwb, Good points.

    123, Ok, I’m glad to hear that. But you write that as if you are Noah Smith. Surely you aren’t speaking for him?

    Curiouseconomist. First they need to simply set up the market and subsidize trading. Later we can worry about attaching it to policy. It’s a scandal that they still haven’t even created the market.

    Negation, They could do that, but there are some problems:

    1. The liquidity differences in the two market might distort the spread.
    2. There’s a “circularity problem” discussed by Bernanke and Woodford in 1997.’

    John, I’ve considered all sort of practical issues in the various papers that I’ve published on this topic. I don’t see any serious problems that can’t be dealt with pretty easily.

  11. Gravatar of ssumner ssumner
    14. July 2012 at 19:05

    Ben, Yes, we are missing some easy opportunities here.

  12. Gravatar of Jon Jon
    14. July 2012 at 19:38

    Noah Smith also doesn’t understand that stock prices are future looking. Stock prices show more volitility than dividends because stock-prices are about the expectations and are being compared against realized dividends. If Noah’s mental model was right, past Performance would be an Indication of Future Results.

    Ironically, his attempt to take the market down a notch demonstrates once more that the market knows more than the technocrats (as embodied by Noah) do.

  13. Gravatar of Morgan Warstler Morgan Warstler
    14. July 2012 at 20:34

    Note: my idea, of course, actually works…

  14. Gravatar of DKS DKS
    14. July 2012 at 22:05

    I see that “NGDP futures” is one name that gets used for two different tools.

    Sometimes, “NGDP futures” is used to mean purely “indicative” futures, which would trade like the TIPS spread today. These futures provide information for the central bank, but don’t feature any central bank commitment.

    But then there are “guaranteed” futures, which the central bank could use to create an NGDP standard that’s a modernized equivalent of the gold standard.

    “Guaranteed” futures trade very differently from “indicative” futures.

    “Indicative” futures would trade just like TIPS today.

    “Guaranteed” futures would trade like a gold standard, where the central bank commits to accept the worse side of the trade, precisely in order to launch an ultimately self-fulfilling trust in the exchange rate.

    Think about a gold standard, where the central bank pledges to redeem $1500 at 1 ounce of gold. If gold is worth more than that, dollars flow into the bank and gold flows out. If gold is worth less than that, gold flows into the bank and dollars flow out.

    Of course the central bank is “losing” on every redemption! The willingness and ability to persistently lose money is precisely what makes the gold standard commitment credible.

    An NGDP futures standard would work like a gold standard, run by a bank with unlimited supplies of gold.

    If the central bank has committed that 2018 NGDP will be $20 trillion, and it’s looking like NGDP is going to be just $19 trillion, then lots of people will sell 2018 NGDP futures to the bank at the guaranteed rate for cash. Which puts money into the economy, until NGDP hits the desired level.

    Likewise, if it looks like 2018 NGDP is going to come in at $21 trillion, lots of people will buy NGDP futures from the bank at the posted $20 trillion rate for their cash, pulling cash out of the economy, until NGDP contracts to the desired level.

    (I’ve left out interest rate effects, but that’s a standard futures issue, not a special NGDP futures issue.)

    “Guaranteed” NGDP futures do not trade like stocks. They do not trade like the TIPS spread, either.

    “Guaranteed” NGDP futures trade like an exchange rate. In fact they *are* an exchange rate, backed by a central bank, and the exchange is between two very similar countries – the United States of today, and the United States of the date of contract settlement.

  15. Gravatar of DKS DKS
    14. July 2012 at 22:34

    Incidentally, the “free money” will go disproportionately to the *first* traders to realize that NGDP is coming in low (or high). So contrary to the Sankowski post, every macroeconomics-minded investor will want to exchange futures for cash as early as they see the economy moving off course.

    Suppose we had guaranteed (exchangeable) NGDP futures right now, and the guaranteed exchange rate called for a 10% nominal growth in the economy over the next year. The economy’s not growing at 10%, so I sell the Fed NGDP futures for next year, in return for cash now. I use the cash to buy Treasuries. What happens next?

    (1) Nobody else sells futures to the Fed, the economy grows at much less than 10%, and I realize a huge profit when the contract settles.

    (2) Lots of people sell futures to the Fed, lots of cash goes into the market, the price of assets soars, and I realize a huge profit right now when the price of my Treasuries goes up. (Meanwhile, my contract settles at about zero, since all that cash put the economy on the track the Fed was aiming for.)

    So if I have information about the direction of the economy, it’s in my interest to act on it whether or not other people do. What’s more, since other people almost certainly will act on that information, it’s in my interest to act on it right away, because otherwise the asset price increase will have already taken place and it’ll be too late.

    So Sankowski’s argument that Goldman Sachs would buy or sell the futures on the last day before settlement completely fails to understand the effect on asset prices (and the broader economy) of guaranteed NGDP futures.

    But beyond Sankowski, we do need to keep straight whether we’re talking about indicative or guaranteed NGDP futures, because they’re both useful but they’re very different, as different as TIPS and a pegged exchange rate.

  16. Gravatar of 123 123
    15. July 2012 at 02:01

    Scott, this has nothing to do with Smith. I wrote this because I was surprised about the extent of our disagreement in the “Taxing doghouses” thread.

    My position is that many market monerists are giving too little attention tomthe sources of macro volatility against the central bank target.

  17. Gravatar of Rien Huizer Rien Huizer
    15. July 2012 at 02:12


    If I remember it correctly, NDGP futures were to offer the Fed a better tool for estimating NDGP in real enough time. I have never gon to the lengths Michael Sankowski went here (and not really tried to pick apart his piece either) but intuitively it always looked like the prototype that deserves rigorous testing before any pactical use involving real resources.

    Sankowski has a very good point: a generically new futures contract, with highly unconventional market influencing features (the Fed, a gvt agency, insolvency proof) would make a market) may either attract no market interest whatsoever, or have such flaws that it can be manipulated in such a spectacular fashion that public opinion would force termination.

    As I understand it, an Sumerian NDGP futures market would discover (continuous) pricing for standard fractions of NDGP on a strip of future dates (and assuming cash settlement).
    That pricing would provide information to the Fed for its use in monetary policy.
    Q 1 What instruments would the Fed be using (apart from intervening in the NDGP futures market and thus directly influencing the price level for those futures (and with it society’s expectations about Fed views on that price level?)? and why would they work better than with an instrumentarium lacking those futures?
    Q2 Why would anyone trade something about which there can be high level confidence: the NDGP prices the Fed wants? (apart from trying to benefit from flaws and operational errors) or worse, would the Fed allow prices to deviate temporarily to later turn around an hit the markets where it hurts (as in standard currency intervention practice) in which case the Fed would be in a position to consistently earn a positive return, at the expense of other market players. If the target variable is an exchange rate, intervention by credible parties tends to be unnecessary; threats sufice. Non-credible parties typically fail to achieve their objectives.
    Q3 To what extent would such a contract, with a strong anti-competitive role on the part of the Fed built in (who else would want to make a market in competition with an ominpotent player) be immune to legal challenges (after the launch?)


    IMHO the Feb should use a suite of indicators that (a) cannot be systematically abused or manipulated) and (b) closely reflect the information that could be gained if NDGP figures were available in real time and publicly and irrevocably commit (upon failure to hit the target for more than six monthsthe Fed mandate would revert to the Treasury and all Fed personnel would go to jail for the rest of their lives, forfeiting their extended family’s entire wealth, etc) .

    Then the politicians should declare that from now on they will take no responsibility for demand-side cyclical policy, but concentrate on supply side (efficient taxation, institutional design, abolishing the legal profession, asking the PRC to run US infrastructure etc) and develop a concept of national accounts and budgeting more closely aligned with business management . accounting. They could for instance send a few people to Singapore to study public sector management there.

    The Fed would then be solely responsible for cyclical smoothing and everyone would be happy.

    Looks like a world that would make neither the Fed nor the politicians entirely happy..I guess a more modest ambition for monetary policy: ie to avoid a s much as possible cyclical disturbances from the financial sector (from poor money supply management to dealing with Lehman-like situations). The politicians would still be allowed to buy votes pretending to support the economy and the Fed would have a real job. If that would involve a consensus among economic technicians and self interested politicians that level NDGP targeting was to replace the existing dual mandate (which makes the Fed the fall guy as we will see after this election), that would not be a bad result.

  18. Gravatar of Mike Sax Mike Sax
    15. July 2012 at 02:36

    Why Bain a four letter word

  19. Gravatar of dtoh dtoh
    15. July 2012 at 03:43

    It’s pretty easy to set up a futures market. (I’ve done it before). You really only need to do 2 or 3 things.

    1. Find buyers and sellers.

    2. Get a license from whoever owns the index. In the case of NGDP, not a problem since government published statistics are in the public domain.

    3. If you want the futures to be exchange traded convince one of the exchanges to trade the future.

    I actually think it would be pretty easy to find buyers and sellers. As others have noted however, if the Fed is stabilizing the price, there would not be a market.

  20. Gravatar of von Pepe von Pepe
    15. July 2012 at 09:37

    “I pointed out that I’d have no objection to the Fed doing its own trading to offset suspicious trades.”

    Since the NGDP futures model is your ideal, I thought under your ideal you were against discretion on the Fed’s part?

    Or, does your “market” solution allow for fine-tuning by the Fed? This seems rather ad-hoc to prempt certain objections.

  21. Gravatar of ssumner ssumner
    15. July 2012 at 09:50

    DKS, Yes, it’s often compared to the gold standard.

    Rien, Almost everything you say is completely wrong. You’ve totally misunderstood what I am proposing. And you don’t seem to understand how futures markets work. Instead of me wasting lots of time pointing out the many errors in your post, maybe we could save time if you actually read my proposal, and then you’d be in a position to critique it thoughtfully.

    To cite just one example, you speculate that no one would trade the contracts. I would, and if no one else did then the Fed would write me a check for $1,000,000, or whatever amount of money the government uses to subsidize the market.

    You might start by googling my post “spot the flaw in nominal GDP futures targeting”

    dtoh, You said;

    As others have noted however, if the Fed is stabilizing the price, there would not be a market.”

    No, the fact that the government stabilizes the price has no bearing on the question of whether there is a market. These aren’t artificial price controls. The price would reflect market forces 100% of the time. The government is able to stabilize the price because it is a major player in the money market. It’s like saying there’d be no wheat market if the government owned a big wheat farm.

  22. Gravatar of ssumner ssumner
    15. July 2012 at 09:54

    von Pepe, As long as the Fed wasn’t constantly losing lots of money, they would not be significantly moving the market price. If they did continually lose lots of money, then I’d think they’d have the good sense to stop doing so.

    Both Woolsey and I have frequently discussed this option in the past. The key is that you want people (like me!) to be able to get rich any time the Fed screws up, as they did in late 2008. Allowing the Fed to trade doesn’t prevent people from getting rich anytime the Fed screws up. They key is to allow free entry into the market–whether the government wants to speculate a bit is a trivial concern in comparison.

  23. Gravatar of Manny C Manny C
    15. July 2012 at 14:12

    Hi Scott
    Can you please provide references to the papers you wrote on this subject?

  24. Gravatar of Saturos Saturos
    15. July 2012 at 23:26

    Critics, why would NGDP futures work any less well than the gold standard?

  25. Gravatar of Major_Freedom Major_Freedom
    16. July 2012 at 00:25

    Critics, why would NGDP futures work any less well than the gold standard?

    Since when does one have to be an advocate of a gold standard if one is a critic of NGDP?

    To answer your question though, a gold standard would be superior to NGDP targeting fiat money due to the fact that gold money production would be subject to the market test of profit and loss. If investors invest too much capital into gold production, too much meaning relative to other goods, then they will incur losses. It would be like investors doubling investment in, and doubling the production of, table salt. There is only so much goods that people are willing to sacrifice in order to make available spending for salt (BTW, one cannot increase the real demand of everything by inflation).

    Investment in salt production undergoes the market test. The market process balances each industry in accordance with marginal utility.

    This would be the case for money production as well if it was precious metals based. Only if people are in such desperate need of more money will the relative profitability of further gold production be justified. But we cannot know the marginal utility of money vis a vis other goods unless we observe the profitability of gold production (or rather, more accurately, entrepreneurs will make judgments in a state of partial ignorance on the future desires people have for money relative to other goods).

    Central banks do not have this information. They do not act in accordance with the market process. They cannot test their money production against profit and loss. They can create $100 or they can create $100 billion, and they cannot know if the former is profitable while the latter is not. This is why all central banks operate according to non-market based “rules” of money printing: “Target” this or that statistic cooked up by social engineer wannabe “economists” (read: political strategists) with an attitude.

    We can know if there was relatively too much investment in the production of houses, because we can observe the losses that occur. But we cannot know, Sumner cannot know, the Fed cannot know how much fiat money production there should be, because the market in money production is monopolized by the state (and the state privileged banks with Fed backstop).

  26. Gravatar of Negation of Ideology Negation of Ideology
    16. July 2012 at 08:32

    Thanks Scott, Steve and curiouseconomist – I think I understand it better now after your responses and re-reading the “Spot the flaw … ” post.

  27. Gravatar of ssumner ssumner
    16. July 2012 at 09:55

    Manny, The first was in the Bulletin of Economic Research, 1989. The best is probably in Contributions to Macroeconomics, 2006. Also see the paper in Economic Inquiry, 2006, with Aaron Jackson.

  28. Gravatar of Potpourri Potpourri
    17. July 2012 at 04:06

    […] Scott Sumner and von Pepe go toe-to-toe. Von Pepe tells Sumner that the word “market” doesn’t mean what he thinks it […]

  29. Gravatar of Why $500bn and not $500 quadrillion? I bothered to do the Math. | Monetary Realism Why $500bn and not $500 quadrillion? I bothered to do the Math. | Monetary Realism
    18. July 2012 at 13:27

    […] Sumner responded to my first post about NGDP futures. He seems to think I pulled the $500bn number out of thin air. He is […]

  30. Gravatar of Now It's Scott Sumner vs. Mike Sanowski Time | Last Men and OverMen Now It's Scott Sumner vs. Mike Sanowski Time | Last Men and OverMen
    25. February 2017 at 11:42

    […]    […]

Leave a Reply