Corporate call policy of convertible bonds

Bob Yau-ching Chan, Purdue University

Abstract

This dissertation provides a comprehensive study of corporate call policy of convertible bonds. It extends previous work by examining an exhaustive set of bonds issued from 1970 through 1989. Empirical results presented in this study provide insights on why firms delay calling their bonds and the market response to convertible call announcements. The effects of call protection on observed call behavior are analyzed. Call protection features, especially price protection, have become increasingly popular since 1980. However, the adoption of call protection features is not associated with a significantly longer observed call delay. This result does not provide support for the conjecture that call protection features might have induced a bias in call delay measurements (Asquith and Mullins (1991) and Asquith (1992)). The dissertation also studies determinants of corporate call policy of convertible bonds. Recently, Asquith and Mullins (1991) propose that firms may delay calling their bonds when the dividend exceeds after-tax coupon payments to bondholders. Consistent with this cash flow advantage hypothesis, a positive and significant relation between cash flow advantage and call delay is found. The presence of an after-tax cash flow advantage is associated with a 24.5% higher call premium ratio. However, only 18.5% of the called bonds shows a positive cash flow advantage, while call delay is significant for all bonds. Call delay in the sample with no cash flow advantage is attributed to the tax benefits of debt and stock volatility. This result provides support for the tax hypothesis (Mikkelson (1981)) and the financial distress hypothesis (Jaffee and Shleifer (1990)) when cash flow incentives of call delay are absent. Finally, the stock market response to call announcements is analyzed. Overall, bond calls are associated with an average after-tax cash flow saving of $1.2 million per year. Regression results of call announcement period abnormal returns shows that the free cash flow generated by forcing conversion is negatively related to the size of abnormal returns. When the free cash flow variable is not included in the regression, both stock volatility and the pre-announcement stock price runup are negatively related to abnormal returns. This result provides partial support for the adverse selection hypothesis. Furthermore, there is some support for the price pressure hypothesis, evidenced by the negative and significant regression coefficients of the effective dilution factor and the post-announcement stock rebound. Lastly, there is no support for the tax hypothesis.

Degree

Ph.D.

Advisors

Phillips, Purdue University.

Subject Area

Management

COinS