Dividend-Protected Convertible Bonds and the Disappearance of Call Delay

Abstract

Firms have not historically forced conversion as soon as possible. Explanations for the delay rely on the size of the dividends that bondholders forgo so long as they do not convert. Constantinides and Grundy (1986) posit that call delay can be optimal when the dividends to be received upon conversion are, or are expected to be, sufficiently large that convertible holders have an incentive to voluntarily convert. Asquith and Mullins (1991) introduce a corporate tax wedge between bondholder losses and shareholder gains and show that call delay can be optimal when the dividends to be received upon conversion exceed the after-tax coupon. We investigate an important change in convertible security design, namely that many recent convertible bond issues are dividend-protected. Dividend-protected convertibles are priced as if no dividends were paid by the firm and dividend-related rationales for call delay become moot. The optimal call policy for a dividend-protected convertible is that of Ingersoll (1977a) and Brennan and Schwartz (1977), namely call as soon as conversion can be forced. We examine the call decisions until 1-1-2011 for 432 callable convertible bonds issued in the period 2000 through 2006. Dividend protection for convertible bonds has increased in popularity over time: there are no dividend-protected convertible bonds in 2000, while 100 percent of the convertible bonds in our sample issued in 2005 are dividend-protected. We document that dividend-protected convertibles are called as soon as conversion can be forced and thus vindicate the original prediction of Ingersoll and of Brennan and Schwartz.