Market-based systems such as stock exchanges and commodities exchanges align financial incentives with good judgement about risk and return. Such systems provide self-reinforcing rewards for good judgement, and combine the opinions of many people into sets of statistically weighted numbers representing the expected values of the available options.
Producing estimates in this manner can often be more accurate than traditional methods, particularly when traditional methods contain inherent biases or conflicts of interest. For instance, a market-based system produced by HP and the California Institute of Technology (CIT) outperformed the traditional sales forecasting process - which awarded bigger bonuses to account managers who underestimated expected sales. In light of the bias in the traditional system, it is not surprising that the relative "share price" in this internal private market system was a better predictor of actual sales distribution in 15 out of 16 trials.
When considering a market-based approach to decision-making it is worth taking into account that the market may produce incorrect decisions in the following circumstances:
- When a small number of traders exercising bad judgement, "bets it all" on a wrong decision, and there is insufficient liquidity for the rest of the market to effectively disagree.
- When inaccurate or incomplete information is available, and the market does not provide sufficient incentives for traders to do their own research to "fill in the gaps".
- When misleading information is available, causing the majority of traders to make a wrong decision, while only a small number, through luck or analysis, make the right decision.
The great advantage of a market system is that over time it impartially and proportionally gives more influence to those who exercise good judgement, and takes influence from those who exercise poor (incorrect) judgement. Thus, over repeated trading sessions, those exercising better judgement will tend to exert more influence over the system.
While it may be predictive of actual sales volume within a product portfolio, a market-based system combined with the traditional performance incentive of comparing expected sales to actual sales in order to calculate bonuses, may be the cause of inaccuracies. For instance, the account managers may have the best information and insight, yet might be excluded from the market because of the conflict between their performance incentives and the market incentives for good judgement.
(An example of how this conflict would come into play is that if the account manager bets against his own products, he has a financial dis-incentive to perform well, creating a "self-fulfilling prophecy".)
Despite such challenges, the market mechanism is an effective way to measure and reward good judgement for anticipating future options. Like any other market, however, there must be safeguards against abuse, which requires careful monitoring. In other words, once the market has been established, it is not reasonable to expect it to operate effectively without supervision.
Given that modern stock markets already possess the infrastructure for detecting and dealing with abuse, and are constantly investing in and improving this infrastructure, would an organization considering market mechanisms for predicting sales volume be better served by existing public markets and exchanges (i.e. by floating an actual stock on the TSX.V, for example)? Perhaps, but this would still need to be weighed against the value of the strategic operational insight that internal investors would have compared to arms-length investors with potentially less information. And if a hybrid solution were employed (by providing the results of the internal market mechanism to the public), will competitors benefit from this extra information?
A second issue is how to select an optimal performance incentive for account managers. Is it fair to base the performance incentive on the difference between expected sales and actual sales? Perhaps not. The problem with this approach is easy to see, by comparing the following situations:
- If the market accurately predicts, or over-predicts the performance of an over-performing account manager, he or she may receive no bonus.
- On the other hand, if the market under-predicts the performance of an under-performing account manager, he or she may receive a large bonus.
