Je ne sais pas

Preliminary Draft: Please do not Quote

Moral Hazard with Rating Agency: An Incentive Contract Approach

July, 2003

Bappaditya Mukhopadhyay
Management Development Institute, P.O.Box 60, Sukhrali, Gurgaon 122001, India [email protected]

Abstract: In this paper, we address the issue of possible moral hazard that rating agencies might have. We discuss the feasibility of possible incentivecontracts that can ameliorate this problem. We ?nd, that incentive payments to the rating agency based on expected returns on debt will do away with the moral hazard problem. JEL Classi?cation No. G140, G200, G290 Key Words: Credit Rating, Information Production, Moral Hazard.

1

Introduction

Rating agencies are unique because, the users of their services do not pay directly for theinformation. Credit rating agencies evaluate and rate debt instruments. They publish ratings on the riskiness of such instruments. The ?rm, whose instruments are rated, pay for these services, while the ratings are available as public information, and hence, the users of these ratings, the investors, do not directly pay. Another, interesting fact associated with the ratings industry is that it is aregulated oligopolistic industry all over. In India, bulk of the ratings are done by the two leading rating agencies in India, CRISIL (Credit Rating Information Services of India Limited) and ICRA (Investment Information and Credit Rating Agency of India Limited). Often the arguments in favor of having such a few rating agencies are that, they promote ‘unhealthy competition’.1 These two facts lead toan obvious problem of moral hazard. In this paper, we discuss incentive contracts that will tackle such moral hazard issues. We model the credit rating agencies as information producing ?nancial intermediaries. We consider a simple model where the economy consists of four risk neutral agents – an investor, a ?rm, a credit rating agency and the regulator (the government). The ?rm has a project withuncertain returns requiring ?xed initiation costs. The ?rms have private information regarding the default probability of their projects. The ?rm does not have capital to initiate the project and hence, has to approach the investor. This investment is raised by o?ering debt. The investor is rational and has Bayesian beliefs. The rating agency charges a fee to the ?rms that come to it to be rated.It then evaluates the ?rms using its screening technology. Information production by a rating agency depends upon the screening function it has and the evaluation standard it sets. This function is such that the accuracy with which the rating agency can infer a ?rm’s type, increases with stringent evaluation standard. The fee is charged ex ante and is same across all the types. The screeningfunction used by the rating agency distinguishes across types with less than perfect precision. The precision level can be improved by the rating agency if it sets higher evaluation standards. There has been a substantial empirical study, on the e?ectiveness of rating agencies. Partnoy (2001), provides a brief survey of empirical work till date. Empirical studies by Ederington, Goh and Nelson (1996)compares the information e?ectiveness of rating agencies and stock market analysts on market movement and e?ciency. White (2001) also study the e?ective1

Whether competition is desirable or not, is dealt in Mukhopadhyay (2002).

1

ness of bond ratings and proposes welfare implications. Partnoy (2001), tests the hypothesis that ratings are e?ective, although the market may actuallyanticipate it in advance and hence, ratings are published with a lag. Recent attempts in theoretical modelling of rating agencies, start with Nayar (1993). In his paper, Nayar establishes the case for voluntary ratings as against compulsory ratings for the Malaysian ?rms. Kuhner (2001) develops a model that identi?es the likely scenarios where the investor may completely disregard or base their…

This article was written by admin