This is an early draft of a forthcoming article . . . feel free to send comments here .
“The conventional wisdom holds that financial markets combine the collective wisdom of the millions of investors who participate in them. Designers of the program were hoping to use that market force to help predict political upheaval in volatile places like the Middle East.. . . ”
“By placing thousands of individual bets, the thinking went, market participants would shape forecasts about future events. The value of the contracts would fluctuate based on market participants’ beliefs about the likelihood of specific events taking place – moving higher if they appeared likely and lower if unlikely. Contracts would expire quarterly and extend a year and a half into the future. Each contract would be fully valued at $1 – the payoff for an accurate prediction.”
The recent brouhaha over former Admiral Poindexter’s proposed Policy Analysis Market is much ado about the wrong thing. Sure, it’s in horrifically bad taste to bet on assassinations, government coups and other assorted mayhem. If you get past your natural impulse to recoil over such an unsavory exercise, you are left with a specious idea based upon a faulty premise.
Underlying the basic concept of the Terror Futures Market, as it was more commonly referred to, are the following ideas:
1) Markets accurately predict the future;
2) Human Beings are rational economic players;
3) The information distribution process is highly efficient;
4) Market participants readily capitalize on that information;
5) Markets are free from manipulation.
Each of these ideas are long revered on Wall Street; They are in actuality myths, readily shown to be, at best, misleading, and at worst, deeply flawed. Lets review each in turn:
1) Markets accurately predict the future
There certainly is appeal to the idea that all of the market’s participants collectively know more than any one single investor. But it’s a huge leap of faith to suggest that the market’s actions are predictions which are invariably “right.” Yet its discussed as a given on the Street in hushed, nearly idolatrous tones.
The reality is that markets are actually lousy predictors of future economic activity. It’s a very old joke that the “markets have accurately predicted 11 of the past 4 recessions.” Its funny, because its true.
The Markets forecasting record is no better than that of storefront fortunetellers. One need only review the past 3 years to see the lack of forecasting acumen in predicting an economic rebound . . . Or as its been called since late 2000, the “phantom” second half recovery.
Since the Bear market began some 41 months or so ago, U.S. equities have had five significant rallies. Each rally was greeted by the same refrains from the cheerleaders: “The market is predicting an economic recovery.” The first four of these ‘predictions’ were clearly wrong; The most recent rally has the best shot so far. Assuming this one sticks, that’s a less than glorious success rate of 20%.
On the Nasdaq, for example, there has been five rallies since March 2000. They have ranged from 39.4% to 51.32% in gains, and have lasted between 48 and 118 days. The median move up has been 42.54% over the course of 77 days. (You can download an Excel file of the comparison here)
If the market is a predictor of future economic improvement, than what, pray tell, were these four previous rallies predicting? The rally that began in May of 2000 – which saw the Nasdaq gain 1246.4 – was an oversold bounce from the initial drop from the bubble highs. Not exactly an astute prediction of an economic recovery. So too for the rally which began in March 2001, one year after the initial crash. From the lows, the Dow gained 2243.48 points, while the Nasdaq bounced 708.47 points.
I’ve heard it argued that the markets were predicting the end of the recession, now dated by the NBER to be November 2001. But that’s a way-after-the-fact rationale. If you buy into the theory that the markets are an efficient determiner of why real investors are deploying actual capital, than that excuse does not withstand scrutiny. Investors don’t care about academically determined endpoints of economic contractions; They care about generating real positive returns on capital.
And a positive return was not the result obtained by anyone who believed the market’s prediction that an economic recovery was soon at hand follwoing the March 2001 rally. From their peaks, the S&P, Dow and Nasdaq subsequently dropped 41.59%, 52.39% and 36.59% respectively. If that’s the sort of predicting ability a Terror Futures Market has, do we really want to entrust our well being to this early warning system? We’re probably better off sticking with Tom Ridge’s duct tape alerts.
It’s not just that the first two rallies of the Bear market failed; Neither of the next two rallies predictions of an economic recovery were accurate either. The bounce after the September 11th sell off was the longest lasting move of all five rallies; It was also the strongest for the Dow and Nasdaq (and second strongest for the SPX). All three indexes subsequently gave back about a third of their value in the ensuing sell off (The Dow, Nasdaq and SPX dropped 34.69%, 36.55% and 32.56% respectively). So too with last summer’s rally, the weakest and shortest since the bear market began. That was followed by a move off the October lows, which also failed.
I’ve heard some people argue that investors behave irrationally during Bear Markets, and that these failed bounces are atypical. Does that suggest that Futures investors will behave rationally if the country is on the verge of a biological weapons assault? I’m also forced to ask “what was the market predicting in late 1999?”
Its worth observing that a long term correlation between equities and the economy has never been shown to exist. “Over the past half-century, we cannot demonstrate stable causal links between total return on stocks and GDP, after-tax profitability, capital expenditures or inflation” noted Howard L. Simons, special academic adviser to the Nasdaq Liffe Markets.
The Market does not predict; It discounts, it assimilates, it reflects the collective beliefs of all its participants. Let’s look at the differences.
A. Predictions vs Probalistic Assumptions
What the markets are actually very good at is acting as a “future discounting mechanism.” By doing that, they make probalistic assumptions about the likely course of events. There is a subtle but crucial difference between predicting the future and handicapping likely outcomes.
Consider this analogy: A group of investors and floor brokers decide to try their hand at handicapping the ponies. They arrange a field trip to a famous track, guided tours included. The track management is happy to have this group of high rollers around. They outfit them with programs and wireless handheld devices, which allows them to place bets from anywhere on the grounds, and set them loose on the field.
During the course of their tour, the investors learn a variety of things (separately and as a group) about the various horses. They notice that “Random Walk” is a favorite, paying 3 to 2 odds. They also see that “Upside Surprise” is a longshot, paying off at 100 to 1. One of the bettors speaks to a groom at the paddock. He learns that Upside Surprise comes from esteemed bloodlines. He whispers this to his friend, and soon the story spreads. Upside Surprise drops to a 50-1 bet.
A different bettor speaks with the track physician; He’s impressed with the unique feed and vitamin regimen the Mare enjoys. He tells a few buddies, places his bets, and the odds slide to 30-1.
This scenario plays itself out over and over again. One investor learns how successful Upside Surprise’s training routine has been (20-1). Everyone is pleased when a well regarded jockey is substituted at the last minute for a rookie rider (15-1). As the horses approach the starting gate, it starts to rain. The owner of Upside Surprise is cheered by this, as the Mare is a mudder. “She just loves the slop” he says within earshopt of the group “. . . and I don’t think there’s a horse in the rest of the field that likes to run in the rain.”
Upside Surprise starts the race with odds now at 4-1, while the still favorite Random Walk is running at 2-1.
The bettors at the racetrack are not imbued with any special omniscient abilities. Individually and collectively, they are not prescient. What the analogy illustrates is the natural tendency of the marketplace to make probalistic assumptions about the likelihood of one course of events occurring over another. As each additional piece of information gets assimilated into the marketplace, the odds of one horse or the other winning changes. That’s not the same as claiming the market knew in advance who was going to be the eventual winner (If you really must know, email me here, and I’ll tell you who won).
The future discounting mechanism of the market essentially operates to build into prices a risk/reward ratio. Back in early March, before the Iraq War started, there was very little in the way of positive economic news. At the time, the Fed Funds rate was higher; The Dollar was stronger versus the Yen and Euro, capital gains taxes were higher, and dividends were taxed as ordinary income.
If you bought the Nasdaq at that time, you paid about 1100 for it. The market rewarded those who made a leap of faith by buying when few else would when all else looked grim. Since then, we’ve seen many things change. The War was quick, and relatively speaking as far as wars go, mostly painless. Taxes and interest rates are lower; The dollar is weaker; There are at present a variety signs indicating the economy is getting better.
The market has now worked into its prices the improved probability of a recovery. But that’s not the same as predicting a recovery; If a recovery is perceived as more likely, your reward for buying is lessened. In March of 2003, you were buying a longshot, an unknown horse with a rookie rider. Today, (August 2003), you have more of a frontrunner. Investors today take on less risk because the unknowns from the Spring are now well known in the Summer. But your reward is also much less, as is your perceived risk of further economic meltdown. That is the future discounting mechanism in action . . .
Sources
Federal Government:
Defense Advanced Research Projects Agency (DARPA)
Policy Analysis Market – PAM
Press Prelease: Terror Futures Market Policy Analysis Market Cancelled
Articles:
Pentagon kills ‘terror futures market’
‘Disaster Market’ Brings Up Unresolved Question
Poindexter’s DARPA Casino, redux
Bookmakers for the Bomb-Makers
Finding the Missing Link Between Stocks and GDP
Stock Prices and the Terror Factor
Bush Impeached? Wanna Bet?
Academic Papers:
The Use of Knowledge in Society
How Well Does the Federal Funds Futures Rate Predict the Future Federal Funds Rate? (PDF)
Results from a Dozen Years of Election Futures Markets Research
What do Financial Markets Think of War in Iraq? (PDF)
Noise Trader Risk in Financial Markets
The complexity of price discovery in an efficient market: the stock market reaction to the Challenger crash
Accuracy and Forecast Standard Error of Prediction Markets
Combinatorial Information Market Design
Decision Markets
Idea Futures: Encouraging an Honest Consensus
Logarithmic Market Scoring Rules for Modular Combinatorial Information Aggregation
Shall We Vote on Values, But Bet on Beliefs?
Articles Supporting the idea:
Best Guess: Economists explore betting markets as prediction toolsScience News On Line, Erica Klarreich, October.18,2003; Vol. 164, No. 16
Futures Trading and the Internet, NYT, By STEVEN WEBER and PHILIP TETLOCK, August 11, 2003
Toward A Bettor Future, IEEE Spectrum, Steven M. Cherry, August 8, 2003
The Disaster Market, Slate, Daniel Gross, August 8, 2003
Deep-sixing a bright idea, U.S. News, Lou Dobbs, August 11, 2003
Tell It Slant, Economic Principals, David Warsh, August 3, 2003
Betting on Terror: What Markets Can Reveal, NYTimes, Floyd Norris, August 3, 2003
A Safe Bet, WSJ, REUVEN BRENNER, August 1, 2003
A Good Idea With Bad Press, NY Times, HAL R. VARIAN, July 31, 2003
Terror ‘Market’ Not Totally Idiotic, Dan Gillmor, San Jose Mercury News, July 30, 2003
Betting On Terror, Reason on line, Ronald Bailey, July 30, 2003
What Weird Futures Can You Buy?, Slate, Brendan I. Koerner, July 29, 2003
DECISIONS, DECISIONS, New Yorker, James Surowieckim 2003-03-17
The Story of the Idea Futures Web Site, Ars Electronica, Robin Hanson, 1995
Idea Futures, Wired, Robin Hanson, Sep 1995
Articles Opposing the idea:
Terrorism: There’s No Futures in It
LifeLog, ‘Terror Futures’ Top List of Bad Ideas
A callous scheme? Bet on it
Economics Can’t Solve Everything, Can It?
Why Terrorists Trade Stocks
Other Futures Markets:
TradeSports
Weather Futures
American Action Market
The Foresight Exchange Prediction Market
Iowa Electronic Markets
Fatdog Exchange
Long Bets
Aggregators & Others
BlogShares Futures Market
MetaCritic
Neotek Markets Project
MarTek
Update (October 10, 2003)/:
Policy Analysis Market Announcement
double click for larger image
Further Sources:
References:
Chen, K.-Y., and C.R. Plott. Preprint. Information aggregation mechanisms: Concept, design and implementation for a sales forecasting problem. Available at http://www.hpl.hp.com/personal/Kay-Yut_Chen/paper/ms020408.pdf.
Hanson, R. 2003. Combinatorial information market design. Information Systems Frontiers 5(No. 1):110-119. Available at http://hanson.gmu.edu/combobet.pdf.
1999. Decision markets. IEEE Intelligent Systems 14(May/June):16-19. Available at http://hanson.gmu.edu/decisionmarkets.pdf.
Further Readings:
2002. The Bank of Sweden Prize in Economic Sciences in Memory of Alfred Nobel 2002. Nobel Foundation press release. Oct. 9. Available at http://www.nobel.se/economics/laureates/2002/press.html.
Additional Sources:
Kay-Yut Chen
Web site: http://www.hpl.hp.com/personal/
Kay-Yut_Chen/research.htm
Robert Forsythe
Department of Economics
C120 PBB
University of Iowa
Iowa City, IA 52242-1000
Timothy J. Groseclose
Stanford University
Graduate School of Business
Stanford University
518 Memorial Way
Stanford, CA 94305
Robin Hanson
James M. Buchanan Center
MSN 1D3, Carow Hall
George Mason University
Fairfax, VA 22030-4444
John O. Ledyard
Division of Humanities and Social Sciences
Mail Code 228-77
1200 East California Boulevard
California Institute of Technology
Pasadena, CA 91125
Charles R. Plott
Division of Humanities and Social Sciences
Mailstop Code 228-77
California Institute of Technology
Pasadena, CA 91125
Thomas Rietz
University of Iowa
Department of Finance
College of Business Administration
S252 Pappajohn Business Administration Building
Iowa City, IA 52242
Vernon L. Smith
Law and Economics
George Mason University
Fairfax, VA 22030
Koleman Strumpf
University of North Carolina, Chapel Hill
Gardner Hall, CB#3305
Chapel Hill, NC 27599-3305
Shyam Sunder
Accounting, Economics and Finance
School of Management
Yale University
New Haven, CT 06520
Justin Wolfers
Political Economy Group
Graduate School of Business
Stanford University
518 Memorial Way
Stanford, CA 94305
