Tyler Cowen recent asked o1 Pro this question:
What is the biggest problem with the idea of the Fed subsidizing an NGDP futures market, as economist Scott Sumner has proposed? Feel free to give an answer relying on theoretical macroeconomics, agency problems, finance theory, and the like.
Here’s part of the response:
One of the most serious challenges with the proposal to have the Federal Reserve subsidize an NGDP (Nominal Gross Domestic Product) futures market is the inherent vulnerability to market manipulation and distorted incentives. While the idea is elegant in theory—using a market for NGDP futures as a direct feedback loop into monetary policy—several interrelated issues arise:
Market Manipulation and Gaming the System:
If the Fed stands ready to adjust monetary policy based on the NGDP futures price, then traders have strong incentives to move that price in directions favorable to their own portfolios. Because the central bank would be compelled to respond to the signal emanating from the futures market, well-capitalized or strategically coordinated groups of traders might manipulate the quotes to influence policy actions. For instance, if participants believe the Fed will ease policy in response to lower NGDP expectations, they could short the NGDP futures contract to force the Fed’s hand, benefiting from subsequent asset price responses. The subsidization of the market, meant to improve liquidity, ironically lowers the cost of such manipulation by making the trading environment more attractive to those with non-informational motives.
Information Asymmetry and Noise Trading:
The existence of a subsidy may attract traders who are motivated less by fundamental views on future NGDP and more by the possibility of collecting the subsidy or exploiting the Fed’s reaction function. Under normal circumstances, futures markets aggregate dispersed information and reflect genuine expectations about future economic conditions. However, if the market is dominated by participants responding to the Fed’s predictable policy feedback loop, the price signal becomes less about true economic fundamentals and more about speculators front-running the central bank. Consequently, the quality of the informational content in the NGDP futures price can be severely degraded.
Due to the Efficient Markets Hypothesis, I don’t think that this would be a serious problem. If market manipulation were possible, then competition among market manipulators would drive the rate of excess return down close to zero. If someone manipulating monetary policy the “wrong way” could make $X in profit, someone reverse manipulating in the opposite direction could make even larger profits.
Long ago, I realized that this was a losing battle. Thus for the past decade I have been advocating an NGDP futures targeting approach that is clearly not subject to the risk of market manipulation. I call it the “guardrails approach”, and it is explained in chapter 5 on my free online book. Once again, I feel I need to explain this idea so that critics of my policy respond to what I am actually proposing.
Suppose the Fed has a 4% target for NGDP growth over the next year. (During Covid, I would have recommended temporarily switching to 2-year forward contracts.) The Fed could simply announce a willingness to take unlimited long positions on NGDP futures contracts based on 3% growth, and unlimited short positions on NGDP futures contracts based on 5% growth. In that case, the Fed would profit any time the actual growth rate was within these two guardrails, between 3% and 5%.
That’s it. That’s the entire proposal. Aren’t I also suggesting the Fed must respond to the futures market? No, they can completely ignore it when setting monetary policy, if they wish. No subsidy is required.
So let’s consider some objections:
1. What if no one trades the contracts? I don’t care. To quote Bob Dylan, “There’s no success like failure”. If no one trades the contract, it’s likely because traders expect NGDP growth to fall within the 3% to 5% range.
2. What if no one trades the contracts at a time when policy is far off course, like 2008, or 2022? In that case, I would trade the contracts, and I’d get really, really rich. To anyone with half a brain it was obvious in late 2008 that NGDP would soon undershoot the Fed’s desired NGDP growth rate. It was obvious in 2022 that NGDP growth would overshoot the Fed’s desired growth rate. So I’d get rich.
But didn’t I previously argue that the EMH suggests that it’s actually very hard to get rich? Yes I did. But think about the implication of that fact. If it’s obvious that the Fed is far off course, and no one is trading NGDP contracts, that would mean that it’s easy to get rich. And that implies that in reality lots and lots of people would be trading those contracts if the Fed were obviously far off course.
And we’re not done yet. The Fed clearly doesn’t wish to lose vast amount of money on dumb NGDP futures trades. Imagine having to explain to Congress that they lost a fortune on trades that anyone could see were foolish. To prevent that from happening, the Fed would likely adjust policy until the market expectation of NGDP growth was within the 3% to 5% range.
So why isn’t this guardrails approach subject to market manipulation? It will be helpful to consider two scenarios:
Case A. Trading is fairly quiet for a time, and then a large number of traders begin to line up on one side of the futures market.
Case B. Trading is fairly quiet, and then George Soros suddenly makes a billion dollar bet in the market.
If you were a Fed official, which of these two scenarios would make you more likely to shift your policy stance? The answer is obvious—the whole point of NGDP targeting is to leverage the “wisdom of crowds”, the idea is to expand the FOMC voting members from 12 to as many as 8.2 billion members. George Soros is rich, but he’s still just one opinion.
There are two ways of thinking about this proposal:
1. It gives the Fed useful information.
2. It adds constraints on policy that make policy more credible.
It actually does both, but I believe that the credibility aspect is far more important than the information aspect. Even the Fed knew that NGDP in 2009 was going to be too low. Even the Fed knew that NGDP in 2022 was going to be too high. Guardrails force the Fed to “do the right thing”.
In this respect, NGDP futures targeting is a lot like level targeting without NGDP futures contracts. Under level targeting, markets also strongly pressure the Fed to “do the right thing”. If we had had a level targeting regime in place in late 2021, then once markets saw that NGDP growth was going to be too high, they would have pushed market interest rates higher, in anticipation of the tighter future monetary policy required to bring NGDP back down. But those higher interest rates would have quickly reduced nominal spending. Under level targeting, whenever the Fed is asleep at the wheel, the markets move the wheel (market interest rates) for the Fed.
Guardrails and level targeting—two ways for markets to force the Fed to do the right thing.
Perhaps some day I’ll have people respond to my actual proposal. Until then, I feel like this guy (painted by Titian):
