Alex Tabarrok, over at Marginal Revolution, makes the fundamental point this morning that what's new about prediction markets is not that they are markets, or that they throw off valuable information, but that they are markets specifically engineered for the sole purpose of throwing off valuable information. Money quote:
"Sundering information markets from trading markets, therefore, is a big advance and one that is likely to lead to better market design for information revelation..."
I'll agree with that, and go one step further. Faster, more accurate information revelation - on its own - is only part of what's needed for prediction markets to take off in corporate strategic planning applications. I've spent years conducting workshops for large organizations using what are, in effect, one-time markets amongst senior executives to assign likelihood values of a range of future events (often as many as 150 in a sitting). These 'conventional wisdom' exercises never fail to produce fascinating insight for the executive team. But in order to make use of the information these markets throw off, decision-makers usually require 1) a shared framework for tracing the links (i.e., dependencies) among events, and 2) a mechanism for identifying the most critical 'fork in the road' events from among a paralyzing constellation of potential ones (regardless of their individual likelihood). Creation of that second-order context may or may not lend itself to further application of the prediction markets themselves.