Informal Algorithms, Impossibility Theorems, and New “Learning” Business/Compensation Models for the Credit Rating Agency Industry

Abstract

This chapter adopts agent-based and complexity approaches and introduces new “learning business models” and compensation contracts for CRAs that can solve the conflicts of interest, antitrust, greed, regret, deadweight-losses, complexity, and industrial organization problems inherent in the credit rating agency industry, and each model contravenes Myerson-Satterthwaite impossibility theorem, Arrow's impossibility theorem, Sen's impossibility theorem, Gibbard's theorem, the Gibbard-Satterthwaite impossibility theorem, and the Green-Laffont impossibility theorem. The business models include “quasi multi-sided auctions” wherein at any time t several auctions can simultaneously occur, and the payoff functions of any buyer-seller pair in any auction depends on the bidding done by at least another buyer-seller pair either at the same time, or at a different time. This “long-memory bias” of buyers and sellers in auctions is new in the literature.