Date of Award

12-12-2006

Degree Type

Closed Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Risk Management and Insurance

First Advisor

Shaun Wang - Chair

Second Advisor

Sam Cox

Third Advisor

Omesh Kini

Fourth Advisor

Eric Ulm

Abstract

Committee Chair: Dr. Shaun Wang Major Department: Risk Management and Insurance In this thesis I employ regime switching models on a unique dataset of bond downgrades to examine the information value of timely downgrades. I use ratings from a Nationally Recognized Statistical Rating Organization (NRSRO) and a non-NRSRO as proxies for the arrival of public and private information. Regime switching models allow us to identify the time at which a discrete shift in the underlying stock return process takes place, estimate the distribution of returns in each regime and also observe the duration of each regime associated with the day of the downgrade. The first contribution is proposing an alternative way to perform an event study. First I define a regime switching model with two regimes: one of low and high volatility. The probabilistic nature of regime switching models allows us to identify the exact day on which stock returns switch to a high volatility regime. This is directly observed through the estimated daily conditional probability of being in one of the two regimes. In summary, I find that stocks switch from a low-volatility regime (1.92%) to a high-volatility regime (6.10%) on the day of the downgrade. The high-volatility regime lasts for about three days and it is mainly driven by downgrades of the smaller bond rating company (non-NRSRO). The second contribution is to propose a method to quantify stock return distress costs associated with downgrades. This measure is based on the capital asset pricing model, uses the parameters of the regime switching model and the estimated daily conditional probabilities of being in each regime. I find that distress costs on stock returns range from 9.49% to 12.91% for the 10 days prior to the day of the downgrade when assuming unity for the market price of risk. The magnitude and direction (sign) of my estimates are consistent with prior literature on the information value of bond ratings. The third contribution is to propose an extension to regime switching models to the bivariate case with a common shock. I show through a state-contingent model how shocks to the economy may cause a one time loss that affects a portfolio of stocks. I derive the frequency and severity implications of such exogenous shocks on regime switching models.

Share

COinS