Cross-Sectional Variation in Stock Price Reaction to Bond Rating Changes: Evidence from India (Methodology-1)

Cross-Sectional Variation in Stock Price Reaction to Bond Rating Changes: Evidence from India (Methodology-1)The study has been conducted in two parts. In the first part, the relationship between company characteristics, bond rating changes and stock returns is examined. The impact of bond rating changes is analysed separately in case of downgrades and upgrades and for each of the company characteristics. We use event study methodology, as developed by Fama, et al. (1969), Brown and Warner (1985) and Campbell, et al. (1997).
The event date ‘T0’ is the date of announcement of the bond rating change. The daily closing stock prices for the period T0 – 120 days to T0 + 20 days are used for analysis in this study.
Daily log returns were also found for the BSE 200 index for the same period.
The market model developed by Sharpe (1963) was used in the analysis. The market model is estimated using ordinary least squares (OLS) regression. The estimation window is a rolling window and consists of 100 days immediately before the day t, on which the return Rk,t is to be computed. This procedure has been used to obtain dynamic values of a k and P k, which change for each day of the event window (-20 to +20 days). This enables us in obtaining more precise values of expected returns for each day of event window. The period consisting of day T0 – 120 to day T0 – 21 was used to obtain the initial estimates of a k and P k and the process is repeated by skipping one day at a time. Autocorrelation was checked using durbin – watson test at 1% significance level. In this study, autocorrelation was detected in 13 cases. GLS estimation procedure is adopted for these cases to ensure efficiency of the estimated parameters.
The daily abnormal returns for the pre-event window, days T0 – 20 to T0 – 1 as well as the post – event window, i.e. days T0 to T0 + 20 were calculated. For drawing inferences about the impact of the event, the abnormal returns so obtained were aggregated. The summation for each company was across time.
The abnormal returns of firms with downgrades and those with upgrades were analysed as separate portfolios. The SCAAR values follow t-distribution and are compared to t-statistic, at 5% significance level, 2-tailed with (T-2) degrees of freedom. This helps to find significant average abnormal returns in the pre-event and post-event windows.
The relationship between the bond rating changes and stock returns is also analysed for portfolios created on basis of different firm chacterstics. The characteristics considered include firm size, price to book value ratio, stock liquidity, leverage, nature of assets (intangibles) and profitability. The first step included arranging the BSE 500 companies in the descending order of their respective firm characteristic value (size, P/B ratio, Leverage, proportion of intangibles and profitability) at end of each year (31 December) from 2002 to 2010.

Tags: , , , ,