Syndicate content

Do casinos produce anything valuable?

Ryan Hahn's picture

"Are markets simply casinos for betting?" Asli recently asked this question on the All About Finance blog. She argues that financial markets do a lot more than that, and I agree. But there are some markets that come very close to being casinos. Surprisingly, in the wake of the financial crisis, the U.S. Commodity Futures Trading Commision has given permission for one such market—a futures exchange based on the box office revenues of Hollywood movies. Basically, anyone will be able to bet on whether a movie will be a flop or become a runaway success. The NYT has the details.

Many were quick to ridicule the idea. Mother Jones screeched that Subprime Goes Hollywood, and compares a futures exchange for Hollywood films to subprime mortgage-backed securities. A blogger on the Davian Letter also astutely points out that (1) Hollywood insiders have a massive informational advantage over the guy on the street (making it a bad idea for most people to bet), and (2) these contracts create the potential for moral hazard by studios that realize they could do better off by causing a movie to tank (although the NYT's article gives vague reassurances that this will be limited because the size of positions will be too small to create that kind of incentive).

Granted, these are both serious issues. However, one benefit is overlooked in all the commentary I've seen so far. As Friedrich von Hayek pointed out many decades ago, prices are valuable because they convey in a single number an enormous amount of information about the costs of labor, capital, transport costs, etc to produce a good. Even if a futures market for box office returns is nothing more than a casino, it will still produce something of value in the form of a price.

Imagine that a studio wants to know a few years in advance if a particular script might make a good movie. The studio has little informational advantage at this point—all they have is a storyline. The studio could set up a contract on the futures exchange, and within a matter of a few days (assuming, of course, sufficient liquidity) have a decent prediction of whether the movie will be a flop or not. Based on this information, the studio can decide to proceed with investing tons of money into production or to put its money elsewhere. In the long run, the average quality of films produced is increased, as flops are screened out well before money is invested. All this depends on the accuracy of long-term prediction markets—I would say the verdict is still out on this question, but you can have a look at some of the research yourself.

Of course, the world will not be much changed because we have averted a few failed Hollywood films. (Although I would argue we'd have been appreciably better without the likes of The Love Guru...)  But a highly liquid futures exchange for films could eventually provide data on the accuracy of long-term prediction markets. If they turn out to be sufficiently accurate, we might then be able to set up markets on many other questions—for instance, on the likelihood of success of various aid initiatives. In an industry with precious few market mechanisms to sort out the good from the bad, the information produced by a futures exchange could be extremely valuable. 


Submitted by Precious Juno HurtLocker on
While prices convey information they are notoriously bad at determining blockbuster or flop. Bringing the futures market to movies provides one more chance to dumb down the movie selections. Teenage pregnancy movies - flop or not? The Blind Side - flop or not. Sandra Bullock made her bet with her participation and percent of gross. Indian Slum Movie for US market - flop or blockbuster? Anything from Roger Corman? Insiders already have capability in Hollywood to bet FOR movies. A futures exchange just gives a way to short the movie market.

Submitted by anonymous on
Amazing claims. So in your view a company - in this case a studio - that keeps under tight wraps new works knows less about its product than the anonymous public, thus public trading years before the project can give valuable insight?

Submitted by Ryan Hahn on
@Precious Juno HurtLocker I would be interested in seeing any research that backs up the claim that "prices are notoriously bad at determining blockbuster or flop." If the prices are consistently wrong, I will consistently bet against them and make a lot of money. But on a more serious note, I know that Intrade has some markets on movies, but I don't think the number of markets and the available liquidity is enough to make a judgement on how well these markets could potentially work. And HSX (which is the precusor to the real-money market) relies on fake money, which is probably not comparable to a real-money market. @anonymous My claim is that we don't know one way or the other if a futures exchange could produce accurate prices well in advance. I do think it is at least possible that a futures market could deliver accurate prices, but we simply don't have the empirical evidence on this. I think it's worth it to try and find out since the potential upside is big.

One thing that needs to be stressed is that prediction markets will not create any information. They are merely tools for aggregating information about a future outcome. If the outcome is very uncertain, the market will reflect this. HSX attempts to forecast box office receipts after a movie opens (not whether it will be a flop or not). It appears to have *some* usefulness, because it is trying to forecast a continuous variable (revenue). If the market is trying to predict a binary outcome (flop or not), it will generate an expected outcome of, say 60%. This means that an *accurate* market will still be wrong 40% of the time. Furthermore, coming "close", or being "almost right", means that you are completely wrong. It is like a horse race. When you pick a horse to win, but it comes in second (almost right), you lose. The blog links to "research", but this is a very limited sample, on a prediction market vendor blog! For more information see the Toronto Prediction Market Blog:

Submitted by Ryan Hahn on
Hi Paul, First off, you are quite right that a prediction market doesn't create any new information - but only in the same sense that prices for goods and services don't create any new information. They simply aggregate existing information about the cost of labor, etc., but this is still extremely useful. As for the binary issue, I agree that it would be more useful to predict a continuous variable. Having a yes/no threshold leaves a lot of uncertainty about how big of a flop/success a movie might be. As for the research, thanks for pointing that out. The reason I linked to Consensus Point is that they include the most supportive research, but I don't find what they have on their website to be that convincing. I think there is more heat than light in the debate over prediction markets, and I look forward to seeing the results of a highly liquid real-money market like HSX. If it turns out that it is not very predictive, then we will find that out and can drop prediction markets in favor of other tools.

Submitted by Mike on
It seems to me that more than most, this market could create self-fulfilling prophecies. Insiders and experts would have incentive to participate, driving the price up or down, generating hype or derision for a given project in news reports and reviews that influence the decisions of ticket buyers. On the other hand, it might be the thing that helps indie films finance themselves (not that they're having trouble these days) thus giving us a more diverse field of films.

Submitted by Fehmeen on
Movies do not become famous or fail based on a pre-determined set of traits. No mathematical models will accurately predict the outcome of these derivatives. I think this is a highly risky move, more ironical in that the current crisis was caused by huge risk appetites.

Submitted by Ryan Hahn on
@Fehmeen It is true that there is no single formula to predict success or failure - but that is not the point. The argument for prediction markets is that they make better predictions than the next best available tool. It is also true that some things are completely unpredictable (as Keynes explained regarding interest rates or wars 20 years in advance). But it is not clear whether futures contracts for movies fall into this category or not. I am simply arguing that it is worth trying to see if the data are worth anything (and, thus, whether the technology might usefully be applied in other areas).

Add new comment