Recent Posts

Friday, July 22, 2016

NGDP Futures Convertibility -- Unworkable?

Scott Sumner has been championing "nominal GDP futures" ("NGDP futures") as a means of creating "market-based monetary policy." This has provoked all kinds of debate, such as whether the Efficient Markets Hypothesis holds. In my view, this is a distraction from the main point -- the market would not function. Sumner has proposed a few models; this article discusses what he said is his current preference -- futures convertibility.

Michael Sankowski has written a long article about the difficulties with GDP futures implementation. I have not spent much time looking at nominal GDP futures, but I have looked at CPI futures. Those futures failed for reasons very similar to what Sankowski describes. Please note that I see a great many issues with Sumner's idea; this article is only touching on one issue. I am more interested in his comments on the linkage to the monetary base, which is a subject that fits in with my upcoming book -- Abolish Money (From Economics)!

In the paper  "A Market-Driven Nominal GDP Targeting Regime" Sumner writes:
William Woolsey proposed another method of using market expectations to guide monetary policy, which uses a principle similar to the classical gold standard. The basic idea is to make money (currency and bank reserves) convertible into NGDP futures contracts at a fixed price (such as $1.0365) [emphasis mine - BR], in much the same way that currency notes were once convertible into gold bullion at a fixed price. However, unlike the NGDP futures targeting plan discussed earlier, there is no automatic connection between the purchase and sale of NGDP futures and open-market operations by the central bank.
The problem with this idea is that it makes no sense if "NGDP futures" resemble other traded futures.

When you enter into a futures contract, the usual convention is that you pay no money -- whether you are long or short. (You do have to post collateral, which is based upon the price volatility of the contract.)

Let's assume a NGDP future has a notional of $1,000,000, and the price (based on annualised GDP) is $1.06 (corresponding to 6% annualised GDP growth). For numeric simplicity, I assume that the central bank is offering convertibility at $1.05 (5% growth).

If you think GDP growth is going to be 7%, you want to enter into a long position. You have two options:
  1. Put down $0, and enter into a long position at $1.06 (which has a notional value of $1.05 million).
  2. Pay the Fed $1.05 million for the privilege for entering into a long futures position (the Fed would have to be short to balance the open interest) at a price of $1.05. This means that you paid $1.05 million for the privilege of getting a mark-to-market gain of $10,000 on a futures contract. 
The central bank's conversion option is only of value if annualised growth is greater than 105% (or less than -95%). Since it would never be exercised outside of a hyperinflation, it is entirely irrelevant, and the idea is reduced to having a futures on nominal GDP (which would fail in the same way that CPI futures did).

The only way that this can sort-of work is that what Sumner/Woolsey are really talking about is nominal GDP "bonds", where you have to put 100% down to buy them. In this case, the convertibility option is just a one-way price peg (or a two-way price peg, if you can convert the bond to cash) at the central bank's conversion price (he uses $1.0365 for some reason).

The problem with an instrument that you put 100% down is that it is indeed a bond, and then boring bond analysts start asking awkward questions. Who is the borrower? What is the money used for? How will they pay me back? The inability to answer those questions cleanly is one very important reason why futures operate on the "no money down" principle.

(Update: added this discussion.) For those readers who are unfamiliar with the justifiable paranoia of fixed income analysts, let us imagine that crude oil futures operated with 100% money down. Assume that we bought $100 million in crude oil futures. There are two sensible places where the $100 million would go -- either the seller(s) of the crude oil futures, or the futures exchange.

  1. If it goes to the seller, what is to stop fly-by-night operators from pocketing the $100 million, and jetting off to a country without an extradition treaty with my country? Since we cannot research our counterparties in a futures trading environment, this is completely unacceptable. (Investors are obligated to investigate the creditworthiness of each of their over-the-counter (OTC) derivatives counterparties; and smart investors generally do not structure trades to advance large amounts of money through derivatives structures.)
  2. If it goes to the futures exchange, what is it going to do with the money? If the futures exchange is sitting on hundreds of billions in liquid securities, what is to stop it from jetting off with a slice of the stash? (Yes, investors post collateral, but it is a fraction of the notional amounts.) 
Even if we imagined that NGDP futures operated with 100% collateral requirements, it still costs us $0 to buy the contract; we just need to find collateral. That collateral, as the name implies, is still our property. It still makes no sense to "convert" money into the futures contract.

(c) Brian Romanchuk 2016


  1. What's the theoretical mechanism (in their minds) to translate those futures into effective monetary policy which would increase output anyway? This does not make any sense whatsoever to me.

    Why would you even buy NGDP futures if you can buy bonds (I guess indirectly you a remaking the point).

    The only way this makes any sense is if you could use them as collateral to leverage, but that makes them no different than any other bonds or collateralised obligations, except that having something material underlying (whether it is mortgages or the full power of a government) you would have a bag of hot air, the value is nothing more and nothing else than the differential with the notional when the derivative is due, as with any other derivatives.

    There is no underlying value underlying them, hence for them to work it would required collateral equivalente to the initial notional (otherwise it would be just a scam), as you explain in your example.

    Do Sumner & co understand how derivatives work? ie. when a derivative is due the market clears and remaining contracts have to be realised? Like if I hold a gold future after the contract expires I have to actually deliver the gold or the equivalent dollars (depending on me being long or short).

    1. Sumner usually says that the Fed will automatically change the monetary base by a fixed amount for every transaction in the futures; the idea is that whenever someone pushes the expected growth above 5%, monetary policy is tightened. How the fact that every single transaction has two counterparties is worked into that is a mystery to me. The fact that he refers to changing the monetary base is the reason why this concept might make my book.

      Since he moved between different contract structures, it is very hard to characterise the idea. Most of his writings assume that the structure will work in a way that he imagines it will, and then he goes off to discuss the implications. I am stuck on trying to figure out the structure.

  2. Ok so is typical monetarist mumbo jumbo ("increasing the monetary base will increase activity because money multiplier effects yadda yadda yadda").

    But the mechanism through which the monetary base would be increased are not well known or thought out.

    The only reasonable counterparty for such transaction would be the central bank itself, but if this is the case I don't know what would make it different than say other central government securities, but worse, as there is no derived interest income for holding such derivatives (considering the futures would be sold by the market maker, in this case the CB, at notional value ofc). There wouldn't be any net addition to the monetary base this way.

    The only way I can see having any sort of effect would be through margin trading and leveraging credit to speculate on the futures. So nothing like we haven't seen before and virtually no different than building any other financial bubble. As soon as the real NGDP growth wouldn't keep up with the market expectations the bubble would crash (not that I think the instruments would have any success at all even if they come with a workable structure).

    All sounds rather desperate and nonsensical to me, but maybe I'm missing something obvious.

  3. I am wondering whether I am missing something as well, but I cannot see how the mechanisms will work, even with a lot of ingenuity on the contract design.

  4. You must be from the "concrete steppes". Apparently, I am also. There is no mechanisms or steps here. There are expectations, and Chuck Norris, and strange monetary bridges that those who believe can perceive.

    I like Scott Sumner and I really enjoy his blog, TheMoneyIllusion. I also think that having the central bank do NGDP level targeting is a good idea, better than the current inflation target. But, like a lot of MMT people, I have doubts about the effectiveness of monetary policy in the first place. And, I can not figure out how the NGDP futures market would work either.

    To me at least, the idea of a NGDP futures market combines a belief in the Powers of monetary policy with a deep faith in the ability of Markets to determine the best outcome. If you are short on faith well then you may never understand it. Which is where I am. Out on the concrete steppes.

    1. Oh, don't get me started on "the concrete steppes."

      How these things tie into monetary policy is another topic (which I will start to touch on later). But if the "futures" contract convertibility will *never* exercised, the instruments cannot have any effect on the economy. There might be an explanation somewhere, but the only explanation that can make sense is that the "futures" contract is actually not a futures contract, but a bond. This then raises a lot of questions how the market would interact with the rest of the economy. (If people can sell short the bonds, who gets the money?)

      Of course, "expectations" solves all of these terminological problems.

  5. Why? I am quite sure that if you got started about the "concrete steppes" that it would be a fascinating read. I would like to know if you think the steppes are the right place to live or a really bad place to be from. I'm not currently thinking of moving, but opinions of the area might be good to know.

    1. The problem with attempting to discuss these topics is that the believers in expectations embed so many assumptions into their world view that it is extremely difficult to explain what they are trying to say in plain English. Since I try to keep my articles at least somewhat comprehensible (and to have conclusions that can be somewhat related to real world data), it is a difficult area to write about.

      It usually takes me at least 600 words to unpack the embedded assumptions within a single paragraph (like the current article, and the next one, which should be out next week).

    2. Hmmn, this explanation is a good explanation, which I often use in my own line of work. Basically it boils down to- I'm not doing all that work for free.:) Nor should you. Very understandable to me. Hopefully I'm not totally misrepresenting you here. And I do very much appreciate your efforts in writing these articles in the first place.

      I have a lot of difficulty with understanding the expectations thing. Its possible that I just can't see the 'big picture' because I'm always looking for the steps. Or maybe I'm just locked into my own ideology, or just plain stupid. Or maybe expectations really aren't all that important. I just don't know. If you figure it out, please write about it. I will read it, even if it isn't for free.

    3. I don't think expectations are very important in non-financial business and institutions. Is the last thing you fall on when basing decisions on, for very obvious (and good) reasons. Economists have completely overblown the importance of expectations. Is a giant exercise of curve-fitting and a wrong foundation which explains away many problems when modelling. Is a convenience.

      Most successful enterprises are driven by as much data as possible, and the less you relay on expectations the best, and the more solid your data is, the best. Voodoo is not a good business strategy. Finance I don't think is very different either.

      Hence expectations playing a big role on daily operations of either the public or private sector are completely overblown, a very small percentage of investment decisions are based on subjective expectations. You can have the problem of poor data analysis, or poor data gathering, but those are very different problems altogether.

    4. Jerry Brown:
      "Hmmn, this explanation is a good explanation, which I often use in my own line of work. Basically it boils down to- I'm not doing all that work for free.:)"

      Not exactly, although I do operate this site as a means of publicising my books. Long-time readers should have noticed continuous site layout changes highlighting the fact that my books are now available at large online retailers.

      I am currently stuck on trying to get my next book in shape, so I find I cannot divide my attention too much. I can produce articles on random topics that I already understand well without difficulty, but I find it hard to pursue more than one line of research at a time. (This is complicated by the fact that I am also doing consulting work on other topics.)

      Otherwise, if you have questions, I can put them on my "to do" list. I cannot always get back to them, but it gives me an idea for new material. My preference is for material that I can reuse in a book that will have a long shelf life, so I do not spend a lot of time discussing what the next Payrolls number (or whatever) will be.

      The problem I have with Market Monetarism is that it makes little sense to me, and so it is hard to characterise. I would have to plough through the research to make sure that I am describing their views correctly. So I am now taking out small bits of what they write, and then describe what how I interpret it. Not as exciting as the big picture arguments that other writers have, but you need to understand their assumptions in order to follow the big picture discussion.

      As for expectations in particular, I write a lot about them in with regards to interest rates. In that area I am an expectations fundamentalist; in my view, that is how you explain yield changes. (There are some qualifications to my views, but that is my starting point.) In fact, it's often the "free market" conservatives who worry about expectations breaking down in the bond market, so it's an interesting subject.

      Outside financial markets - in a low inflation environment - expectations are probably overblown in modern macro (as per Ignacio's comments). In a high inflation environment - which we have not seen for decades in the developed world - inflation expectations matter a lot. How we get from the "low inflation" to the "high inflation" environment is a topic I am currently agnostic on.

    5. Dr. Romanchuck, just so you know for sure- my comments were not intended to be critical of your article, nor do I expect you to spend your time replying to my questions. (Unless you want to). I am flattered that you have and enjoy the discussion. Thank you. Often times what I intend to write in a comment is not the same as what others perception of that comment is. Am working on that.

      Now as far as expectations being relevant to interest rates in the bond market, I do want to question you especially if you are talking about "sovereign" government bonds in the MMT parlance. Why should I care what expectations are in regards to that market, even if they were important in other markets? I ask because I have come to be a "full-bore" believer in the Bill Mitchell school of MMT but at the same time, constantly question it.

    6. No offense taken; I have to remind myself that I am supposed to be making money doing this.

      I have no problem taking questions; I do not get a lot of them. If I am answering one person's question, I may be covering the concerns of a lot of people who are not asking directly. Although I have topics I want to cover, I do need suggestions. However, I cannot guarantee rapid turnaround if the answer requires research.

      Yes, I am talking about sovereign bonds (or the swap curve). If you are a bond investor, you make money off of yield changes, so you are worried about expectations.

      If you are not a bond investor, bond market expectations offer a read on the future path of the economy. (At least in the current institutional framework; if central banks locked the short rate at zero, which is a MMT view, then expectations about short rates are meaningless.) One can debate whether those expectations are accurate, but they have been better than the economic consensus.

  6. Brian you are completely right about expectations importance being context-dependent. It was unfair of me to characterise them as unimportant, I'm biased by the current environment when it comes to expectations and my previous post comes from that biased view. From a macro perspective I cannot see the current trend changing, even if we assume supply shortages and the downfall of our current industrial model it would probably translate through deflation or stagflation than inflation (except maybe the odd currency crisis and hyperinflation) because the natural outcome would be a freezing of activity and transactions.

    The topic of how you get from low to high inflation is a topic on it's own, and also context dependant. It can't be properly modelled because IMO is a multivariate problem where we lack sufficient data to extrapolate to a model, and lack mature operative constructs in the first place (ie. how we define political mismanagement and compute values to include in a multivariate statistical model).

    Demographic factors, supply chain problems, political corruption, social trends, technological shocks, all contribute to modify inflation environment, and all have intra-factorial effects and are context dependent. This is why I think a complete (and mathematically sound) "Theory of Inflation" is a chimera which won't be solved any time soon.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.