|
Future Enterprise- Prediction Markets |
|
|
Prediction markets are an increasingly important forecasting tool based on evolutionary information aggregation and selection. Participants trade in contracts whose payoff depends on unknown future events, as in any financial or betting market, with the price of contracts directly interpreted as a market-generated forecast of some unknown quantity.
In a truly efficient market, prediction market contracts and pricing reflects all available information and therefore holds out the promise of yielding the most unbiased forecasts of available information.
Such tools are extremely useful for estimating the market’s expectation of a particular event occurring and because of their success in predicting public events and corporate outcomes, have generated substantial interest among the business community as well as social scientists and policymakers.
Prediction markets cover a wide spectrum of applications including election,technology, economic and financial futures. For election-betting markets the price of a contract reflects the probability of a candidate winning an election, canvassing all available information, including polls, the state of the economy and recent policy pronouncements. They also focus on economic statistics such as industrial production, retail sales and business confidence levels.
There is also a growing number of web-based prediction markets run by companies that provide a range of trading and betting services, including some pseudo-markets in which participants trade in virtual currencies. In addition companies such as HP, Microsoft and Google are using this method to test the likely success of future web-based technologies.
Importantly, predictive markets have consistently outperformed alternative forecasting methods, including opinion polls, Delphi methods, quantitative models and expert forums. Quantitative models for example try to turn historical trends into predictive outcomes, but even these complex mathematical tools fail to predict human behaviour and psychology.
In the traditional efficient market hypothesis- EMH, prices are best understood over the longer term by standard economic analysis. However there is now a comprehensive challenge to this core assumption, with fundamental differences as to how prices are set in non-utilitarian markets. For example it is realised that these are often governed by human behaviour, such as a tendency towards herding or swarming within a context of high levels of uncertainty. In environments of high uncertainty, individuals instead look for patterns, then form beliefs and act on these, modifying their beliefs by continuing to learn and adapt from experience in Bayesian mode.
Simulations such as Internet–based games are also applied. Market modelers run simulations of financial markets populated by virtual agents or traders that continually make buy or sell decisions, for example in stocks or foreign currencies. The agents try to identify patterns in price movements and use them to predict future trends.
The models generate realistic markets with winners and losers, rallies and crashes, generating an emotional mood and psychology of its own. Such mathematical models predict market behaviour with a precision that traditional economic theory cannot match. Agent-based models can also explain how the buying patterns of a small number of investors can quickly spread and influence those of many others, resulting in mass panic or other collective movements, as happened recently in the global financial meltdown, reproducing the statistical properties of real markets and predicting real prices.
Future trends
In the future it will be possible to create and experiment with entire simulated mini-economies such as those now emerging in virtual social and business network environments such as Second Life. Predicting markets and the probability of future scenarios and outcomes will therefore become an essential tool for strategic planners in the future enterprise.
Only registered users can write comments.
Please login or register. |
