Recent work on the topic of reputation systems from U. of Michigan researcher Paul Resnick in collaboration with staff at Harvard's Kennedy School of Government.

The full paper can be downloaded in PDF.

Abstract: Recommender and reputation systems seek to inform potential customers by securing current consumers' feedback about products and sellers. This paper proposes a payment-based system to induce honest reporting of feedback.

The system applies proper scoring rules to each buyer's report, looking to how well it predicts the report of a later buyer. Honest reporting proves to be a Nash Equilibrium.

To balance the budget, the incentive payment to each buyer is charged to someone other than the one whose report that buyer is asked to predict. In addition, payment schemes can be scaled to induce appropriate effort by raters.