HP began to explore prediction markets in 1996, but did not even consider applying them to the 2002 HP-Compaq merger. Similarly, Yahoo and Microsoft are two of the companies mentioned most often as being involved in prediction markets (along with their main competitor Google), but I’ll bet
Why doesn't the reaction of the two stock prices to the news tell us enough about whether stockholders benefit?Erich, employee manipulation could easily be a problem if outsiders were unable to trade or unable to evaluate the effects. If outsiders can see that the prices are inaccurate, then the employees will need to spend arbitrarily large amounts of money to maintain the manipulation. Also, the employees face free-rider problems - each employee wants the other employees to devote money to manipulating the market, but doesn't gain from devoting his own money to it unless that money turns out to be the manipulation that stops the merger.
Executives are acting in their own self interest in not allowing prediction markets to elbow in on their sphere power. Part of it is ego, part of it is reaping the rewards of a tournament environment. There will have to be a clear cut, prediction market-driven organization to lead the way to greater acceptance and implementation at these levels.
Furthermore, if a market for mergers did exist, I would expect to see a lot of employees use their raffle points to short positions that may actually threaten their job security. If Yahoo has a rather strong HR talent, the Microsoft HR department should aggressively try to bias the decision as to not become a casualty of synergy.
Doug,Remember, a decision making process that corrects for agent failure is not in the interest of agents!
True, and I'd say thats why successful companies (such as Oracle and Microsoft) are often run by people with large stakes in them. Presumably you'd also have a constant evolution of corporate charters which governed decision-making processes (hopefully helping to weed out bad agents), but I don't know how much of modern corporate structure is set in stone by law. Some of what many boards of directors do seems somewhat incestuous to me.
I don't think its clear that MSFT and YHOO were wrong not to use a prediction market (or at least one publicly known of) to assist in their decision.
Mergers rather consistently produce less value for shareholders than predicted by corporate officials, or so I seem to recall reading. They seem to be motivated by the interests of those who propose the merger. Remember, a decision making process that corrects for agent failure is not in the interest of agents!
Grant, I think I see your point. On the other hand, I think a good prediction about general expert opinion of a policy would perhaps help us get a better sense of the various intangibles that might go into forming an expert consensus.
Most economists might agree that X policy will lead to Y outcome, and this might be taken as an endorsement by policy-makers, whereas if you also asked economists whether they would think Y outcome would be considered good by economists in the future, the economists in the present might generally say no.
Of course, a problem here would be conflating expert values with the values of the general public. Experts probably see things more clearly than the general public in terms of what is, but might have much different values from the public generally, and assuming this, experts generally might not act as good proxies for the public. Perhaps a bet on the opinion of intelligent lay people in 5 or 10 years time would be useful.
Ah, my bad, I thought we were talking about trading a generic futures contract (e.g., a conditional Intrade contract) and not actual stock options. Come to think of it, I didn't know if conditional stock options even exist (or are even legal).
Mike, my first thought is that markets give experts incentives to participate just like anyone else. Experts of course tend to be more sure of their opinions and thus would be more likely to participate in the market and add their private information than laymen might be (so your pool of "experts" would be self-selected from incentives in this case). So I'm not really sure why we'd necessarily care to predict the outcome of "expert" polling.
My first thought on the reasons prediction markets aren't used in cases like this is the degree of private information consolidated in the firms' executives. Their long-term business plans to create value from the merger are not going to be things they are going to want to share with others. Given this, the executives might not feel (perhaps arrogantly) that outsiders have enough knowledge to make accurate predictions. Has anyone written anything on practical uses of prediction markets in cases like this? I am also skeptical I would be able to use them to guide many business decisions for the above reasons.
Or it also could be that most haven't heard of prediction markets yet. Most of the executives I've talked to haven't. The last issue of the HBR talked about the Delphi method, but did not mention prediction markets.
Grant, my thought: it seems better to trust expert consensus over my own opinion, even if experts' incentives aren't completely in my interest. Perhaps I am wrong about this. The polling process would probably need to be fairly explicit at the time betting is started. "We will have X polling company ask Y number of randomly selected Ph.D. Economics professors if they agree or disagree with the following question..."
My impression is a decision market would be the best tool for guessing expert consensus in the future compared to other tools. Perhaps I am wrong about this too!
Robin, would the estimates not be valuable? Or is it they are valuable but it's tough to persuade decision-makers? I'm not quite sure of your question.
Mike, the question is who would be persuaded by such estimates.
James, yes markets can have secret prices, though we expect them to suffer an accuracy penalty in this case.
Grant, this is conditional stock futures.
Psuedo, there are not transparent predictions like this in ordinary stock prices.
these companies are just not serious about finding the highest value applications of prediction markets
Companies are rarely serious about anything. The interesting question is whether the managers are serious. These examples are easily explained as principal-agent conflict and tell us little about the managers' beliefs. It's also possible that they (correctly!) believe that if they let the genie out of the bottle, they'll be forced to use it in examples like these.
Doesn't the stock market make this sort of prediction every day?
I would expect prediction markets to have their greatest effect when they are not (effectively) duplicating an existing market.
Would a prediction market be more prescient than conditional stock futures (i.e., "I'll pay X for YHOO-MSFT if the merger happens", or "I'll pay Y for YHOO and Z for MSFT if the merger does not occur")?
Mike, how do you know you can trust "experts" unless the experts have strong incentives to be as honest and thorough as they possible can be? Further, these experts would have to be chosen in some manner, also without the market incentives that keep prediction market participants honest and diligent.
Would it be possible to set up an effective prediction market in which prices are kept secret? (Buyers and sellers could place bids contingent on final prices.) If so then a third party could take advantage of the fact that a successful merger would probably increase the average price of both stocks. This party could set up a merger prediction market for two companies that might merge and then if the results indicated that a merger would be profitable this party could buy lots of stock in the two companies and then publish the results of his prediction market.
Sorry, let me revise this confusing sentence.
"For example, one bets on whether a poll of experts in a certain field (law or economics or political sciences, et c.) would believe that the merger was good or bad in five or ten years, et c."
Should read: "For example, one bets on whether a poll of experts in a certain field (law or economics or political sciences, et c.) would ***SHOW EXPERTS BELIEVE*** that the merger was good or bad in five or ten years, et c."
The poll would be of experts or non-experts in 5 or 10 years and the results would be what is bet on.
I apologize for any confusion.
Would there be any value in betting on how questions would be answered by experts or non-experts in the future regarding the issue of privacy or other concerns? For example, one bets on whether a poll of experts in a certain field (law or economics or political sciences, et c.) would believe that the merger was good or bad in five or ten years, et c. Perhaps a similar question could be asked of non-experts, regarding whether their online search experiences, et c., are better or worse, and what the answer would turn out to be would be bet on now. I'm not if this approach would be too problematic or not.