I have written a few columns, and spoken with some government IT managers, about the idea of using so-called "prediction markets" as a tool to improve government performance, particularly for managing big IT (or other) projects. Wednesday's New York Times actually ran a long article -- "Betting to Improve the Odds" -- on the growing use of prediction markets inside businesses, by cutting-edge firms such as Cisco, GE, and Google.
In prediction markets, people put cash on predicting various outcomes. The "price" of a prediction will go up and down, depending on how much money people are willing to place on the prediction -- the higher the price, the more likely the outcome is considered to be. Prediction markets thus make use of the "wisdom of crowds" idea.
Prediction markets have gotten a lot of attention in elections, where they have often been more accurate than polls or pundits in predicting election results; there was also an unfortunate incident a few years ago when an innovative effort in the government to use prediction markets to glean information on the likelihood of various events (such as specific terrorist attacks) was squashed amidst an ignorant political uproar.
The most obvious use for prediction markets inside government is to get better estimates of the actual cost, schedule, or performance on big projects, such as IT projects. When this has been tried out in companies, the predictions about schedule coming out of these prediction markets have been superior to those made by subordinate managers to their bosses. The Times article cites predictions about when a company's new stores in China would be open that were far more pessimistic than what senior management was told, and that enabled corrective measures that reduced the delays. One could also imagine prediction markets on the likely success of new organizational change initiatives, as one way to pre-test the probable viability of different ideas.
Who's going to be the pioneer for trying this idea out in government management?
By the way, for new blog readers (and old ones who may have forgotten or never known) -- I try to post on Tuesdays and Thursdays whenever possible.
