I’m finally ready to announce the first step in creating an NGDP future market.
[Update: I won't say much now other than that some very generous offers have been received. Tomorrow I speak with an organization that helps fund innovative projects like this, so perhaps I'll be able to report more in the near future. I plan to list the names of the people making offers, but will wait until tomorrow, as I'm checking to see who wants to remain anonymous. The task is slightly more complicated than usual with two different funding targets, but I'll just say at this point that I am very pleased with how things have gone today.]
Future generations of economists will look back on our era with disbelief. Respected economists like Robin Hanson and Justin Wolfers clearly explained how prediction markets could generate all sorts of valuable information for policymakers, at an absurdly low cost. And they were totally ignored by both policymakers and the economics profession. The Fed continued to blunder into errors costing hundreds of billions of dollars (like their decision not to ease policy in the meeting 2 days after Lehman failed, on fear of inflation) partly because they weren’t willing to spend a few thousand dollars setting up and subsidizing prediction markets to gather real time investor forecasts.
The biggest problem by far was the lack of an NGDP futures market. TIPS spreads are fine, but in a world of 2% inflation targeting much of the variation in inflation represents supply shocks, such as unstable commodity prices. Interest rates are almost useless as a measure of the stance of monetary policy, and the money supply is only marginally better. We need real time data on the single most important macroeconomic variable in the universe, expected US NGDP growth, which also happens to be the best indicator of the stance of US monetary policy. (Ben Bernanke once suggested that interest rates and the money supply are not reliable, and that only NGDP growth and inflation can reliably indicate the stance of policy. But inflation is distorted by supply shocks, whereas NGDP is not.)
In recent weeks I have been working with Eric Crampton at the New Zealand Initiative and Robert Quigley-McBride at iPredict. Robert put together the following proposal, with input from Eric and myself:
iPredict is a real-money prediction market run by Victoria University of Wellington. With over 8000 traders, iPredict is operated by students of the University with the purpose of forecasting social, economic and political events. iPredict has designed contracts for trading to allow forecasting of nominal GDP (NGDP). Traders will be able to buy and sell contracts paying $1 if NGDP falls within a particular range in a particular quarter.
An example contract would pay $1 if the US GDP for Q1 2014 were greater than or equal to $17,250 Billion and less than $17,500 Billion, based upon the [initial estimate for quarterly nominal GDP from] BEA Table in Section 1 – Domestic Product and Income, Table 1.1.5, Line 1. We would establish similar contracts spanning the intervals $250 Billion above and below the example contract, with two open ended intervals beyond those for outcomes above or below the ranges.
The set of contracts, for each quarter, will generate probability estimates for the different levels of NGDP, as well as providing a gauge of the traders’ uncertainty, or margin of error, on this estimate.
To support the accuracy of the forecasting, we wish to provide the market with an injection of liquidity. This will help by ensuring there is sufficient incentive for traders with beneficial information to bring it to the market. To support contracts for forecasting NGDP for the next 3 years, we will need to raise $1500. This will be used to provide a market maker on the contracts, which will ensure there is always a reasonable bid-ask spread, and that any trader wanting to bring information to the market can do so even if there is not necessarily another trader to take the other side in the trade.
An additional option to improve the accuracy of these contracts would be to pay interest on the value of the contracts to any traders holding shares over a long period of time. This would reduce trader reluctance to engage in the long term contracts because of the cost of the capital invested. This option would require substantial back-end system development to track the funds that traders have invested in the contracts, and so it would require funding in the region of $30,000.
iPredict is an authorised futures dealer in New Zealand under the Financial Markets Authority. However, they must strongly discourage Americans from trading on iPredict due to US laws. Anyone wishing to trade on iPredict should take note of the Terms of Service 1.5 “Nothing on this Website constitutes an offer or invitation to trade with any person who is under 18 years of age and/or outside New Zealand. In trading on iPredict, traders warrant that they are complying with all laws in the jurisdiction in which they are present. [bracketed portion added by me]
A few comments:
Obviously the proposal can be tweaked with further input from commenters. They think a 5-contract setup is best, and I’d be inclined to go something like:
below 2.5% NGDP growth
2.5% to 3.5%
3.5% to 4.5%
4.5% to 5.5%
5.5% or more.
The actual contracts would have dollar amounts specified, however, as the percentages change with revision to NGDP at the starting point.
We would use initial estimates of NGDP, as the later revisions are unforecastable before the initial announcement, so we’ll get an unbiased forecast. The advantage is that investors don’t have to wait as long for the payoff.
We have to exclude Americans because . . . well because America is no longer a free country. More specifically there are two problems. First, iPredict does much more than the Iowa Electronics Market, and hence doesn’t have their waver from the Feds. And second, there are now incredibly burdensome “know thy customer” rules that America imposes on foreign financial firms dealing with US citizens. A small entity like iPredict (run by the Victoria University of Wellington) obviously doesn’t have the funds to meet all these burdensome regulations. Now when I travel around the world almost everyone I talk to in the financial area regards US citizens in roughly the same way they view the Ebola virus. They see our government has being a sort of madhouse.
You will notice two sums mentioned. The $1500 figure is the cost of setting up the market. If that’s all we can raise the market will still go ahead. However it would be nice if we could raise an additional $30,000, so that the market can be subsidized with interest-bearing margin accounts, which will help insure more trading, more liquidity.
The market may well fail. There is currently no NGDP market in the private sector, presumably because there is little market demand. That’s the whole point of prediction markets. Corporations setting up internal prediction markets to forecast next year’s sales don’t do so to meet “market demand;” they want information, the wisdom of crowds.
The market is only a first step. The goal is to shame the economics profession and policymakers into to setting up a better funded market in the US, with government funds supporting the market. I’m not normally a fan of big government, but consider:
1. Market NGDP expectations are a pure public good.
2. The cost of a superb market would be trivial, a few millions.
3. The benefit would be in the hundreds of billions, if not trillions.
4. The trivial cost could easily be more than recouped by cutting the Fed budget for economic forecasting.
It’s a no-brainer for libertarians, but also for thoughtful liberals who agree with people like Christy Romer and Michael Woodford on the importance of NGDP, and the need for wise policymaking. I see no good arguments against creating an even better market than we will be able to establish.
My initial inclination was to try something like Kickstarter. But I’m no good with computers, and very busy. So first I’m going to try to raise all the funds from one or two rich people. This really shouldn’t be hard. It should appeal to everyone from liberals like George Soros, to moderates like Michael Bloomberg, to libertarians like the Koch brothers and Peter Thiel. I recall a few years back that someone sent me a youtube of a west coast billionaire (Valve Corp?) giving a talk, and suddenly casually mentioning the need for a NGDP futures market. The startup cost ($1500) is pocket change for a billionaire, and even the $30,000 is very affordable. If that doesn’t work, we’ll try Kickstarter.
As a reward I’m willing to name the market after the big donor when I write about it in my blog. So if Mr. Jones donates $1500, I’ll call it the “Jones NGDP market,” unless they prefer anonymity. If Mr. Smith later kicks in $30,000, I’ll rename it the Smith/Jones NGDP market. Extra typing for me. And I still type with two fingers!
This could be a history making initiative. I implore potential donors to think about what this might lead to, and to consider whether they want to be seen as the visionary who began the long process of modernizing economic policymaking.
Don’t send me money; contact me and I’ll set you up to talk directly to the people in New Zealand. If we do get the market set up, I hope my non-American readers will participate, at least in a small way. It would help to make the market a success. But again, the real goal is to shame the people in power into stopping all this Nostradamus nonsense, and move economics and economic policymaking into the 21st century. It’s long past due.
I can respond to comments, or you can email me directly at firstname.lastname@example.org.