Does debt market timing increase firm value?

Abstract

Survey evidence indicates that firm managers try to time debt markets when choosing the maturity of new debt issues, but we do not know whether these strategies increase firm value. This research examines differences in value across nontimers and timers, where timers are defined as firms that follow either a naïve strategy of choosing long-term debt when the term premium is low or a strategy from Baker et al. (2003) based on the predictability of future excess bond returns. After controlling for various determinants of firm value, the research finds no differences in value across timers and nontimers. It also documents that the timing strategies do not increase firm value and do not affect announcement effects of long-term debt offerings. The results suggest that corporate debt markets are efficient and well integrated with equity markets.

Acknowledgements

I thank Shane Johnson, Ji-Chai Lin, Chip Ryan, Alex Butler and seminar participants at Louisiana State University, Wilfrid Laurier University, University of Arkansas, Southern Illinois University, Sungkyunkwan University and 2004 MFA Annual Meeting for helpful comments.

Notes

¹SeeLucas and McDonald (1990), Korajczyk et al. (1991) and Choe et al. (1993).

²The methods that classify debt issuers into timers and nontimers are explained in detail in Section III.

³This research includes firms implementing timing strategies over the period 1983 to 1997. Because it investigates the impact of timing strategies on change in firm value 3 to 5 years after sample firms implement the strategies, the data extends to 2002 even though the sample period ends in 1997.

⁴SDC New Issues Database covers only investment bank underwritten debt issues, which is limitation of this research.

⁵As Guedes and Opler (1996) document, LYONs were issued infrequently

⁶Traditionally, the term, ‘medium term note’, was used to describe debt issues with maturity >1 year but <15 years. Yet, this is not a true characteristic anymore. The maturities of medium term notes are sometimes <1 year or >30 years. For instance, Walt Disney Corporation issued a security with a 100-year maturity off its medium-term shelf registration in July 1993.

⁷All of the untabulated results are available on request.

⁸Inside ownership data is mainly collected from ‘Compact Disclosure’. Yet, ‘Value Line Investment Survey’ is also used to collect the data for the issuers during 1983 to 1988 since Compact Disclosure is available after 1988.

⁹The item of notes payable from ‘Research Insight’ includes bank acceptances, bank overdrafts, loans payable to officers of the company, loans payable to parents and consolidated and unconsolidated subsidiaries, loans payable to stockholders, notes payable to banks and others, brokerage houses’ drafts payable, telephone companies’ interim notes payable and advances from parent company and commercial paper.

¹⁰Baker et al. (2003) also use expected inflation rates when they estimate the excess bond return equation. Because the results are not qualitatively different, actual inflation rates are used.

¹¹A probit model is used to forecast the maturity of debt. The result of the probit model is available upon request.

¹²To construct the sample for event study, the research examines 2289 straight long-term debt offerings over the sample period, 1983 through 1997. To find out announcement dates of each issue, three sources, ‘Lexis-Nexis Business News’, ‘Wall Street Journal Index (WSJI)’ and a filing date from ‘SDC’ are used. The dates reported about the offering for the first time in ‘Lexis-Nexis’ and ‘WSJI’ are primarily used as announcement dates. For issues that are not reported in the two sources, filing dates from ‘SDC’ are used as announcement dates. If long-term debt issues come with equity or convertible debt simultaneously, those dual issues of debt and equity are removed from the sample because Masulis and Korwar (1986) and Billingsley et al. (1994) show that those dual issues are accompanied by significantly negative announcement effects, consistent with general findings about equity issues. To confirm whether the sample still includes dual issues, filing dates of common stock, preferred stock and convertible debt issues from ‘SDC’ are double-checked. If the filing dates are the same for debt and equity issues, the debt issues are removed from the sample. In this way, 1716 long-term debt issues with announcement dates are obtained. Then, those debt issues are matched with daily stock return data from ‘CRSP’. To be included in final sample, stock returns for those issuers should be available on the event date and estimation period. Finally, a sample of 1423 straight long-term debt offerings is obtained for the event study.

¹³The regression results on change in q from year–1 to year +5 are unreported because those are qualitatively the same as those in Table 5.