We note the following recent references to prediction markets:
In a post on intelligence reform, Professor Stephen Bainbridge reminds us of the shortsightedness of killing the DARPA-PAM initiative:
...government bureaucracies are the supertankers of the political world. They have so much institutional momentum that it's very hard to get them to change direction, even with the right captain at the helm. Certainly, however, current risk averse leadership compounds the problem. The most innovative thing I've heard from the government about how to wage war on terror was DARPA's proposed prediction market. The media and the left went nuts over it, even though prediction markets have a very good track record in an astonishingly wide range of areas. Instead of rewarding and defending those who came up with this innovative idea, however, the intelligence leadership and the administration threw them to the media wolves. Dumb.
Another antidote to bureaucratic ineptitude he notes, is increasing the use of outside experts. Unfortunately, whether in government or a large corporate environment, the effectiveness of the smartest, most outspoken outsiders is often diminished by three things:
1) The tendency of bureaucracies to search for and vet outside experts based on a preconceived view of the problem at hand and the solution they'd like. Past relationships with experts whose opinions are well regarded internally also play a role.
2) The gradual, corrosive effect (for even the most clear-headed, ethical individuals) of knowing who they are working for and what their benefactor wants to hear.
3) The tendency of any bureaucracy to close ranks and marginalize those viewpoints they are unprepared to accept, (also part of the challenge of getting prediction markets adopted.)
Prediction markets are not a panacea, but they do offer the important advantage of shifting the initiative for bringing in and selecting outside experts from those inside a bureaucracy to the experts themselves (problem #1). They also avoid problem #2 by reversing the incentive to participate in mass delusion. The incentive instead, is to seek a position that will be borne out in the future by objective facts at the expense of plausible but incorrect theorizing.
Craig Cline opines on Yahoo's Tech Buzz Game:
This “Game” has a serious purpose underlying it: it’s the theory of Yahoo Researchers the “price” of each “stock” represents the aggreated opinionon [sic] which alternative future will dominate searches. By inference, the results will reflect when people believe interest will shift from one product to the next. O’Reilly worked with yahoo to help formulate the wuestions [sic] asked in the “game” in such a way that answering them could be predictive of [where] a given technology is going. The resultant “trends” therefore will be very real data that can be used by pundits and analysts to determine when adoption of, say, OS X Tiger, will switch from Panther.
Unfortunately, there's a gaping hole in this logic. What 'everyone' is searching for and talking about is often different from whether a new tech product will be successful. Having been a technology analyst in the past, I've seen more cases where 'buzz' represented wishful thinking, hysteria and/or corporate sponsorship of 'buzz' itself, and far fewer where it foretold actual product adoption or market share - much less financial success for investors.
As I've said all along, those more concrete types of outcomes make for far more useful and interesting prediction markets in a corporate setting, even as they are less fast-paced. For those disinclined to believe this, see here - a fascinating archive of flops that at one time dominated the media 'buzz', including my personal favorite any-day-now-really-just-keep-sending-us-money innovation: pen computing.
Finally, David Beisel weighs in on exactly this point from a VC perspective:
While I fully realize the intention of [the Yahoo Tech Buzz] game is purely for fun, it is perhaps perpetuating the notion that all that matters is the hype and buzz. Just because this market predicts Friendster to be the premier social network in the future, it isn’t necessarily so. A good business is not made on hype alone.
Amen, David. If, as the saying goes, "you can't manage what you can't measure", we should be careful not to over-measure one element of business success (particularly one that's been notoriously misleading throughout history), just because the measuring itself is easy and likely to generate a lot of buzz for the measurer.
Comments