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):
READER COMMENTS
John Hall
Dec 15 2024 at 8:36pm
“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%. ”
In that last sentence, the Fed would only be making profits (or losses) at all if they are actually participating in the market, which implies that the futures price has moved to pricing in 3% (or less) or 5% (or more) growth and then subsequent realized growth comes in within the guardrails.
“In this respect, NGDP futures targeting is a lot like level targeting without NGDP futures contracts.”
I see these as more separate, but oh well.
Anyway, my recollection of the original NGDP futures idea was that the purchases and sales have a direct impact on the monetary base. The rationale was to avoid the circularity problem. It seems like the way you describe the guardrails approach here you don’t have that same direct connection and thus leave yourself open to that circularity criticism.
Scott Sumner
Dec 15 2024 at 10:14pm
No, the circularity problem occurs when the central bank looks at a futures price for direction. In my proposal, the price of NGDP contracts never changes. Instead, the central bank reacts to trading volume (if they choose to do so.) You can think of this type of market as predicting the instrument setting that leads to on-target NGDP growth.
Ahmed Fares
Dec 15 2024 at 10:20pm
re: the nonsense that is EMH
I’m a self-trained securities analyst who made a living investing for thirty years following the methods used by Buffett and Graham. I once measured my CAGR for a 20-year period, and it was 18%, this is a Canadian market which had a CAGR of 6% over the same time period. That’s an alpha of 12%.
Now this is not supposed to happen according to EMH except by pure luck, something about that given enough monkeys flipping coins, some monkeys will get a long run of heads. First, this from a previous article by Scott Sumner:
Warren Buffett, out of sample
Brad DeLong weighs in.
Weekend Reading: Warren Buffett and Michael Jensen: the 1984 Columbia University Orangutang Debate
Vaidas Urba
Dec 16 2024 at 3:38am
ChatGPT 4o on the debate:
Tyler Cowen and Scott Sumner are clearly talking past each other in their debate over o1’s analysis of inflation dynamics. The core issue lies in their differing goals and frameworks. Cowen defends o1 as a reflection of the professional consensus, while Sumner critiques the consensus itself. This means they are debating separate issues: Cowen on how well o1 captures mainstream views, and Sumner on whether those views are valid in the first place.
Their monetarist perspectives also differ sharply. Cowen references classical monetarism, rooted in Friedman’s idea of stable relationships between monetary aggregates and nominal income. Sumner, by contrast, advocates market monetarism, emphasizing nominal GDP targeting and recognizing the instability of the money multiplier in modern monetary regimes. These differing frameworks make their arguments fundamentally incompatible.
However, Cowen’s prompt engineering is problematic. By framing o1’s task as defending the consensus, he biases its output to align with mainstream views, sidelining legitimate critiques like Sumner’s. This stacks the deck in favor of Cowen and avoids addressing whether the consensus itself might be flawed. In doing so, Cowen conflates representing professional opinion with advancing genuine understanding.
Todd Ramsey
Dec 16 2024 at 9:38am
Scott, if you have time to respond I would appreciate it.
1. If the NGDP futures market is not large enough, traders will willingly lose money in the NGDP futures market to promote a Fed response that they can profit from in other, larger markets. That eliminates the information-collating benefit of the NGDP futures market. The NGDP futures market needs to be huge–similar in size to the Treasuries market, which eliminates that possibility; adds to the wisdom of crowds by increasing the size of the crowd; and allows Fed operations in the market to directly alter the path of NGDP.
2. I don’t think it’s desirable or feasible to change from a 1-year to a 2-year contract at the Fed’s discretion. What happens to the money of people who depended on the duration of the 1-year contract, and need the money at that time? Also, we’re trying to avoid discretion at the Fed, right?
Agreed that a one-year contract is too short during an event like COVID. So make it always a two-year contract, and offer a new contract every quarter. The Fed would only participate in one contract at any given time, but the other contracts would continue to trade until expiration.
Scott Sumner
Dec 16 2024 at 3:56pm
Your second question is based on a misconception. I’m not suggesting that the Fed invalidate existing contracts, or change their maturity. I’m suggesting they extend to 2 years on brand new contracts, after it’s clear that something like Covid will lead to lower NGDP in the very short run.
Regarding your first question, I would not be at all concerned about a lack of trading. It would merely indicate that the Fed is on target. That’s good!
If you put a guardrail along the highway, the project is not a failure if no car brushes up against it.
Todd Ramsey
Dec 16 2024 at 5:04pm
I’m not concerned about a lack of trading.
I’m concerned if the NGDP security market is not large enough, a trading firm could push the contract up against the guardrail, keep pushing so the Fed keeps resisting, knowing they will lose money in the “too small” NGDP market. They do so because they have bets in the opposite direction in larger markets like Treasuries, Corporate Bonds, or Equities. They are willing to lose in the small market in order to gain in the large market.
If this were to happen the information value of the NGDP market is corrupted.
Matthias
Dec 16 2024 at 8:20pm
Even if the market is supposedly small, someone else could still bet against the would be market manipulator and make a lot of money by correcting the prices.
Any such attempt at manipulation will automatically enlarge the market in a sense, by attracting other participants to a profit opportunity.
Todd Ramsey
Dec 17 2024 at 1:28pm
If the NGDP market is too small: Market Manipulator short sells NGDP contracts, pushing the price against the bottom guardrail. They simultaneously bet against Treasury and Corporate bonds, in anticipation of future inflation driving fixed income prices lower.
The Fed reads the NGDP market signal that is (unknown to them) polluted by the Market Manipulator, and takes action to increase future NGDP.
The anti-Market Manipulator wins as Matthias says; and the Market Manipulator loses IN THAT SMALL MARKET. However, the Market Manipulator wins much bigger in the larger markets where they placed large bets.
Summary: If the NGDP security market is too small, the quality of information available to the Fed — which is the purpose of the market in the first place — is poor.
HOWEVER, if the NGDP security market is sufficiently large, two good things happen: first, the quality of the information improves, and second, the Fed’s participation in the market is the mechanism by which the Fed conducts the equivalent of open market operations, altering MV sufficiently to materially affect future NGDP.
Scott Sumner
Dec 17 2024 at 11:46pm
If a single manipulator tried to move the market, there would be no reason for the Fed to respond by changing its monetary policy. I’m not at all concerned about that sort of risk. Competition among speculators has a way of preventing excess returns from manipulation.
The Fed should respond if thousands of speculators are all lined up on one side of the market.
Brent Buckner
Dec 17 2024 at 8:19am
I suspect that there should always be two year contracts: the measurement lag in housing goes out beyond 12 months.
bill
Dec 16 2024 at 7:57pm
I fear that you are underestimating the odds that the Fed makes a mistake and loses many trillions. If the Fed were you and 11 like minded people, your idea would work. But our actual existing Fed has already lost trillions for the economy in 2008/2009 and allowed large inflation in 2022 even though in both cases the markets were screaming to change course.
Scott Sumner
Dec 17 2024 at 11:48pm
That’s exactly why I favor this policy. They refused to change course in 2008 and 2021 because they had nothing to lose. I want them to have something to lose.
I’m confident that they won’t want to go in front of Congressional committees having lost trillions of dollars in easily avoidable trades.
Market Fiscalist
Dec 16 2024 at 7:59pm
Can you clarify:
“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.”
I see how creating new money to pay out traders who bet on low NGDP growth will increase NGDP growth but I’m not seeing how the mechanism works when the payouts are to people betting on high NGDP growth. I assume that in this case the fed sells assets to fund the payouts but as this money is then transferred to the traders I don’t see how the net effect will be to reduce NGDP growth.
Scott Sumner
Dec 17 2024 at 11:50pm
The monetary policy response has nothing to do with payouts in either direction. Monetary policy would operate through ordinary open market operations, as it did prior to 2008. The payouts would not be a part of monetary policy (and would probably be relatively small in size.)
Market Fiscalist
Dec 18 2024 at 12:10am
OK, thanks for the clarification.
Is the aim of the NGDP futures market just to incentify the fed not to lose money by having to pay traders when NGDP growth moves out of range or is there a more fundamental purpose for it?
I
Market Fiscalist
Dec 18 2024 at 9:55am
I suppose it also provides a “wisdom of the crowds” type mechanism for the fed to track NGDP expectations.
KG
Dec 18 2024 at 1:06pm
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.
There’s a lot of economic news release style swaps on various CFTC regulated exchanges that basically have no trading volume. In fact, the vast majority of newly launched futures/swaps end up delisted within 5 years due to no (open) interest. In this example, the lack of trades in the NGDP contract would mean that your trades would quickly push the market towards equilibrium, but the actual amount of money you could make would be tiny. Don’t believe me: go and trade the natural gas EIA swap on ICE. There’s no bid, no offer, no daily volume, and the settlement is conducted by broker assesment.
If you believe in the EMH why does no one launch a highly liquid NGDP contract? Supposedly the interest is so huge, this represents massive amounts of free money on the sidewalk. In fact, the market for these contracts is already complete since there are (r)GDP contracts that already trade on kalshi, and there’s a super liquid market for inflation swaps that trade OTC with banks. Why is there a deep liquid market on inflation but nothing on rGDP or the unemployment rate? There’s really not a good consistent explanation for why some contracts succeed and others fail to launch.
Todd Ramsey
Dec 19 2024 at 9:50am
Can anyone identify if something is wrong with this analysis?
If the NGDP guardrails market is too small, a Market Manipulator could push prices against a guardrail, incurring small losses in the futures market, to trigger policy changes that create larger profits in broader financial markets. This could undermine the information value of the NGDP futures market and its utility for monetary policy. Addressing this issue requires ensuring sufficient market depth and liquidity to make such manipulation costly, alongside monitoring for external incentives to distort signals.
Detailed Analysis:
The Hypothetical Mechanism of Manipulation:
The manipulator bets against the NGDP futures guardrail market, knowing it is relatively illiquid and therefore easier to influence.
This pushes the market price toward a guardrail, signaling to the Fed that NGDP expectations are outside the target range.
Anticipating the Fed’s policy response (e.g., monetary tightening or loosening), the manipulator places larger bets in more liquid markets (e.g., US Treasuries or corporate bonds), where the Fed’s actions have broader ripple effects.
The profits from these larger markets outweigh the losses incurred in the small NGDP guardrails market, making the strategy attractive.
Implications for Information Value:
The NGDP futures market becomes a distorted signal, driven not by true aggregate expectations of NGDP growth but by the manipulator’s strategic behavior.
The Fed’s reliance on this market as an indicator would lead to suboptimal policy responses, undermining the credibility and purpose of the guardrails system.
Preventing Manipulation:
A. Liquidity and Market Depth:
A small market is inherently vulnerable to manipulation because the cost of moving prices is low. To reduce this risk:
The Fed must ensure sufficient liquidity in the NGDP futures market by encouraging participation from a wide range of traders, including hedgers, speculators, and institutional investors.
Market-making mechanisms (such as subsidies for liquidity provision) could be introduced carefully, though they must avoid distorting incentives.
B. Monitoring and Surveillance:
The Fed could monitor the NGDP futures market for unusual trading patterns, such as a single participant consistently placing large trades near the guardrails without aligning with broader economic indicators.
Such patterns could trigger reviews and potentially lead the Fed to discount the signal from the futures market temporarily.
C. Policy Discretion and Multiple Signals:
The Fed could treat the NGDP guardrails market as one of several inputs for monetary policy, rather than as a single dominant signal.
Cross-referencing guardrails market data with other indicators (e.g., labor market data, inflation expectations, credit spreads) would reduce the risk of overreacting to manipulated futures prices.
D. Deterring Manipulative Arbitrage:
The Fed could make it explicit that it monitors market activity and reserves the right to delay or condition its policy response to futures price signals. This would deter manipulators who depend on predictable and immediate Fed reactions to profit in larger markets.
Is This Problem Inherent or Fixable?
While the risk of manipulation is valid, it is not inherent to the guardrails approach itself but rather a feature of any small and illiquid market with significant external consequences.
Building a robust and liquid NGDP futures market and integrating it into a broader framework of policy inputs can mitigate this issue. The guardrails system remains viable if designed and implemented carefully to resist manipulation.
Conclusion:The hypothesis highlights a legitimate risk, especially in the early stages of an NGDP futures market’s development. Addressing this requires attention to liquidity, monitoring for manipulative behavior, and using multiple data points to guide monetary policy. By doing so, the Fed can preserve the information value of the NGDP guardrails market and leverage its potential to improve policy outcomes.
Comments are closed.