At the Cato Monetary Conference this week, Scott Sumner said he had a "modest" proposal, that there should be a highly liquid futures market in Nominal Gross Domestic Product (NGDP). This caught my attention, as the futures market is a topic near and dear to my heart (I write about it every week).
Sumner is known for his view that the Fed should target NGDP as the basis for monetary policy. So a futures market that predicts it would be convenient. Let's look at his idea more closely.
What, precisely, is to be traded? If I buy an NGDP future, what is delivered at contract maturity? It's clear what I get when I buy a wheat future or crude oil future. With Nominal GDP futures, there's nothing to receive because NGDP is not a commodity or even a security.
A futures contract is a derivative, but an NGDP future would not be derived from anything.
So how is an NGDP contract supposed to settle? The only thing I can think of (Sumner doesn't say) is that the NGDP contract will pay based on the published NGDP number. For example, if NGDP comes in at $19T, then the contract might pay out $19,000 (depending on how many NGDP units are represented by the contract).
GDP numbers are revised a few times after they are published. Is the NGDP contract to pay when the initial number comes out, even though it might be wrong? Or will NGDP be the only contract where both parties' capital must be locked up for months until the Bureau of Economic Analysis publishes the final number?
Moving on to the next problem, let's look at a real futures contract like wheat. Suppose the bid price on a March wheat future is $7.00 and the ask price on spot wheat is $6.50. There's 50 cents of profit to anyone who can buy wheat in the spot market, sell it in the futures market, and store it for the duration. This is called carrying, which is an act of arbitrage.
Arbitrage anchors the price of wheat futures to the price of wheat in the spot market. Both prices can move up and down in response to changes in supply (demand being pretty stable in wheat) or speculation that supply will change.
By contrast, the NGDP market is for speculators only. It has no connection to a real price, and no arbitrageurs. I am all for making gambling legal, if people want to bet on NGDP, the weather, or the horse race. But that's not Sumner's point. He believes that this market will predict NGDP accurately enough to manage the economy without causing recessions and depressions.
Sumner declares this market won't merely be liquid. It will be highly liquid. Let's consider that.
Each market has a different degree of liquidity. The liquidity comes from the character of the thing being traded, not from a government proclamation. For example, copper is more liquid than lumber. Silver is more liquid than copper and gold is more liquid than silver.
A liquid market has a continuous bid and ask. That does not mean someone is always buying or selling. On the contrary, it means someone stands ready to buy or sell at any time. Who and why?
In my wheat example, I mention the carry arbitrage. Suppose the cost to carry -- interest and storage -- is 30 cents. Each trader has to decide his minimum profit threshold. Suppose Joe is willing to do it for 20 cents. He needs to get at least 50 cents on the deal. He can sell a March for $7.00, which is why he bids $6.50 on spot. Joe will not pay one penny more, unless the price of March wheat goes up.
Joe and his competitors are why there's always a bid (and ask) in wheat.
Obviously, there is no carry arbitrage with NGDP futures. There won't be much liquidity either.
An NGDP market may convenient for monetary planners, but without a reason to exist it won't work the way Sumner hopes.