Michael Sankowski, writing at Monetary Realism, critiqued using futures contracts to target nominal GDP. Unfortunately, he misunderstood the proposal. He thought that the Fed would trade the futures contracts at a fixed price right up until the time of settlement. For example, he considered the Fed trading a contract on third quarter 2012 nominal GDP up until the preliminary estimate is announced in late October. He imagined Goldman Sachs putting up $5 billion in margin the day before the announcement, with some huge leverage. And then the result would be a $500 billion payday for Goldman Sachs.
However, the proposal has the Fed cease trading the contract on the third quarter of 2012 before that quarter begins, with recent proposals being three quarters before, at the end of the third quarter of 2011. During the fourth quarter of 2012, the Fed would be trading contracts on the fourth quarter of 2013. Of course, the system could be set up with longer or shorter lags. Sumner often mentions a two year lag. Still, no version suggests less than a one quarter lag. Anyway, the notion that Goldman Sachs, or anyone else, would get wind of tomorrow's announcement of nominal GDP and then trade some huge amount with the Fed the day before is inconsistent with any sensible version of the proposal.
Sankowski's understanding of leverage in futures markets is quite different from mine. In his example, Goldman Sachs supposedly put up $5 billion in a margin account on one day and received a $500 billion return. He understands this as being 100-1 leverage--or perhaps the implication of 100-1 leverage.
The actual proposals for futures targeting involve margin requirements based upon the variance of the targeted macroeconomic variable. If Goldman Sachs takes a position that could generate a $500 billion profit, there is also a chance of a $500 billion loss. The margin requirement for that position would be of the same order of magnitude. For example, suppose a $1000 contract pays $10 for every percent that nominal GDP deviates from target. Suppose there is only a 1 percent chance that nominal GDP will deviate more than 2 percent from target, and that is considered acceptable. The margin requirement would then be 2 percent, or a only $20 a contract. "Leverage" would be 50 to 1. A sale of $250 billion worth of contracts would require a $5 billion margin requirement. If nominal GDP was 3 percent below target, the pay off would be $7.5 billion, providing return on the margin investment of 150%. If, on the other hand, if nominal GDP was only 1 percent below target, then payoff would be $2.5 billion and the rate of return would be 50%.
For Goldman Sachs to get a $500 billion, then the deviation would have to be very large. Let's say it is 5 percent, so that it would require contracts worth $10 trillion. The margin requirement would have been $200 billion. This would be a great return for Goldman Sachs--250%. And, of course, the "cost" wouldn't be the $200 billion, which would be returned with interest. The cost is the risk that nominal GDP would have come in above target. For example, if it was 1 percent above target, the loss on a $10 trillion short position would be $100 billion.
Sankowski also discussed the possibility of using the contracts for hedging. Incredibly, he described a scenario that I understood to be "dynamic hedging." Perhaps Sankowski misspoke, but he seemed to imagine that a company would forecast losses and then sell futures contracts against nominal GDP to hedge the loss. How is that hedging? I had never even imagined such a "strategy" until Arnold Kling suggested that some traders were using it during the financial crisis. I don't remember the details, but it was something like "hedging" the losses from selling default swaps by carefully watching the news, and when risk of default increased, selling the insured bonds short. That makes just about as much sense as "portfolio insurance."
A true hedge would be companies nominal GDP futures contracts on the anticipation that when nominal GDP was below target, sales and profits would be low. And if nominal GDP came in above target, sales and profits would be high. This would reduce the variability of profits on operations plus the net return on the futures contract. But waiting to find out that profits will be low and selling futures--that isn't hedging.
Sankowski's entire approach to the analysis seemed to be about whether futures exchanges would find it advantages to introduce these contracts. Would the member/owners find opportunity to profit from commissions or at the expensive of hedgers? Well, that is hardly important. The purpose of futures targeting is not for a central bank to make money from commissions or at the expense of traders. The primary purpose is to constrain the central bank to seek the targeted goal. The secondary purpose is to provide an opportunity for market participants to reveal their expectations regarding nominal GDP, so that the central bank may adjust monetary conditions according to those expectations.