Scott Sumner  

Is the Fed allowed to create GDP prediction markets?

Did Edwin Cannan Think a 100 P... Is Ciudadanos a hope for Spain...

In the comment section of a recent blog post, David Andolfatto asked the following question:


Does it fall within the Fed's present Congressional mandate to create markets in the instruments you want? (I honestly don't know. If we do, I'll do everything I can to help.)


David is a distinguished blogger and also works at the St. Louis Fed. We were discussing my proposal that the Fed set up and subsidize trading in NGDP and RGDP prediction markets (sometimes called futures markets.)

First a bit of background information:

1. Private companies sometimes create prediction markets to forecast important variables such as sales or revenue. This concept is based on the "wisdom of crowds."

2. There is a lot of academic research by people like Justin Wolfers and Robin Hanson that suggests that prediction markets are useful.

3. We know that the Fed cares about market forecasts, as the minutes of Fed meetings show that variables such as TIPS spreads (implicit inflation forecasts) get discussed.

4. In theory, NGDP is a better indicator of demand expectations than TIPS spreads (which are affected by both supply and demand shocks.)

5. The Fed spends a lot of money each year on economic research. Prediction markets are relatively inexpensive.

6. The US government has previously approved prediction markets such as Iowa Electronic Markets for academic research purposes.

7. These markets would be extremely useful for two reasons:

a. They provide market indicators of expected growth in demand.
b. They would show the impact of unexpected monetary shocks on both NGDP and RGDP expectations.

This latter point is especially important. Some Keynesian theories deny that monetary policy can impact expected NGDP growth, when at the zero bound. Real Business Cycle theory suggests that a monetary policy shock that impacted NGDP growth expectations would not impact RGDP growth expectations. Market monetarism suggests that monetary shocks can impact both NGDP and RGDP growth expectations at the zero bound.

I believe that it is clearly within the Fed's mandate to spend a modest sum of money setting up NGDP and RGDP prediction markets. But it doesn't matter what I think it matters what Congress thinks, and what the Fed thinks that Congress thinks.

Sometimes when I travel to DC I meet Congressional staffers, who ask me how they could help. Here's one good area. It would be great if we could get some important Congressional figures to go on record as supporting the concept of the Fed setting up prediction markets to ascertain useful market forecasts, which could help make monetary policy more scientific. The cost is trivial and the potential benefits are huge.

PS. In my view it would be better if Congress said it was OK with them, but up to the Fed. Why not have Congress mandate these markets? I think as soon as you go down that road things get very politicized, and people become much more worried about a loss of independence. I find it hard to believe that the Fed wouldn't want to do at least a pilot study, if they had a clear go-ahead from Congress.

PPS. This is also something that Treasury could consider doing. I would think that at some point the Fed and Treasury would want to sit down and discuss where it fits best, once they accept the basic concept. I tend to prefer the Fed, as (rightly or wrongly) it's viewed as being more independent.

PPPS. The St Louis Fed where David works has spent money setting up a vast database ("FRED"), which is incredibly useful in economic research, teaching, blogging, journalism, etc.

PPPPS. It would help if other bloggers and journalists would encourage the Fed to set up some prediction markets.

Comments and Sharing

COMMENTS (14 to date)
Dan writes:

The operation of and participation in prediction markets is generally regulated by the CFTC. Adam Ozimek wrote a paper on the topic and might be a helpful resource to you:

Vaidas Urba writes:

The Fed could offer term deposits with NGDP-linked interest via auction process.

ThomasH writes:

"Some Keynesian theories deny that monetary policy can impact expected NGDP growth, when at the zero bound."

No doubt "some" do, but isn't the more typical "Keynesian" view that conventional monetary policy -- manipulation of short term interest rates -- cannot impact NGDP growth when those rates are close to zero AND (implicitly) that there is a lot of political resistance to non-conventional monetary policy such as QE, exchange rate "debasement," announceing a higher inflation target, etc.?

bill writes:

What should we make of a Fed that doesn't seem concerned with a sub 2% interest rate on the 10 year Treasury? That's a huge market and the market price is almost certainly predicting sub 4% NGDP growth for the next decade.

Nick writes:

I don't understand ... Keynesians believed in 2010-2013 that the fed could ease money with unconventional policy, but it would be more practical politically to do massive fiscal stimulus???

Hazel Meade writes:

Aren't government bonds basically NGDP prediction measures?

Todd Ramsey writes:

Do you foresee these prediction markets as an end in themselves? Or as a stepping stone to a true NGDP futures market, where Treasury issues securities linked not to interest rates, but to NGDP, in volumes large enough that the Fed can effect monetary policy through purchases and sales of theses securities?

And if the latter, would legislation be required to create these securities, or could Treasury do so through administrative action?

Scott Sumner writes:

Dan, Thanks for that info. It's hard to imagine them opposing a Fed-created market.

Vaidas, Yes, but I think you'd get a cleaner estimate with a simple prediction market.

Thomas, I saw a lot of Keynesians claiming QE is ineffective at the zero bound because banks would just hoard the cash. I have no doubt some Keynesians agreed with you, but I'd guess it was a distinct minority.

BTW, that's a really bad argument because effective unconventional monetary stimulus would be far LESS controversial than what the Fed actually did. The zero bound would have lasted for a much briefer period, and less QE would have been involved.

Bill, Good point. Even worse, they are concerned, but believe it is telling them that money is too easy, not too tight.

Nick, Good point.

Hazel, Long term rates are correlated with expected NGDP growth, but not perfectly. Right now expected NGDP growth is probably higher than the 10 year bond yield.

bill writes:

OMG! You are so right. They really do think that a low 10 year rate means money is too loose. Crazy. BTW, I think one reason the Fed is not interested in an NGDP futures market is that it would make their jobs less fun. They love going to their meetings and pontificating and saying wise things. If the meeting was just "oh, 3.9%?, that's a little low, so let's loosen a little until we're back to 4.0%", it would be boring. They want to feel important.

Jeff writes:

Does this plan really need Fed involvement though? Couldn't private entrepreneurs set up something like the Iowa electronic market Come to think of it, why not just pitch this idea to Iowa? They're quite nice people down there in Tippie.

Michael Byrnes writes:


I know of at least one person who is doing exactly that!

jamesxinxlondon writes:

There is clearly more innovation in the regional Feds than the central Fed.

The Atlanta Fed RGDP Nowcast is a good innovation too, and certainly influences the Hypermind quarterly NGDP competition. Perhaps they could stretch it to an NGDP Nowcast.? And then take both a bit further ...

I wonder whether the central Fed would get huffy if this innovation went too far. Almost certainly, but could they stop it?

Ricardo writes:

If the costs really are trivial, why not have Mercatus do it?

ThomasH writes:


[Keynesians believed in 2010-2013 that the fed could ease money with unconventional policy, but it would be more practical politically to do massive fiscal stimulus???]

What I mean is that "Keynesians" think that monetary authorities when faced with a demand shock are politically constrained from using non-conventional policies vigorously enough to stabilize the real growth trend. This implies that there will be a substantial period when increased, deficit financed, investment ("massive fiscal stimulus") would be be appropriate. Investing in activities with present costs and future benefits which have NPV>0 at the (near ZLB borrowing rate) ought not be politically controversial. Of course using unconventional monetary policy ought not be controversial, either, so what combination of fiscal and monetary policies represents the politically constrained optimum is debatable.

I think it is correct to fault many such "Keynesians" for not helping overcome the political opposition to more vigorous monetary policy. I think it is not correct to count them among opponents of monetary policy and even less to suggest that they oppose monetary policy because the favor fiscal policy over monetary policy

Comments for this entry have been closed
Return to top