For instance, perhaps I propose (and the speculation market agrees) that some form of regulation of some industry will increase some social welfare statistic. But at the same time someone else proposes (and the speculation market agrees) that some form of industry taxation will increase some social welfare statistic. If the statistic goes down, it might be very difficult to tell which policy was wrong, or if both were wrong. It could also be that the statistic goes up, even though one policy was wrong, because the benefits of the other were so great.

I know you argue that, insofar as these speculators are wrong, they lose money and can no longer bet. But I think in practice this process would act too slowly. For instance, by the time Bill Gates lost a significant fraction of his fortune he could put enough “noise” into the markets to override thousands of academic experts. If Bill Gates were a completely rational person and consistently lost money, maybe he would give up early. But in practice he would win some and lose some, if only by chance, and it’s unclear to what extent his psychology would take this as motivation.

On the other hand, I guess the system doesn’t have to be perfect — it just has to be better than what we have. In our present system, I think, there is an element of speculation: a firm can lobby and gain political influence in return, and in a sense it determines how much money to spend on lobbying based on its perception of the probability of success and the return. However, in the present system, the firm makes these bets vis-a-vis its individual benefit, regardless of what is good for other firms are people at large. In your system, at least the purpose of the bets is to improve some social welfare statistic.