The authors examine the call delay of convertible bonds, comparing the delays of dividend-protected convertibles and non-dividend-protected convertibles. They also investigate the call delays between low-dividend firms and high-dividend firms. Their results reveal the reasons behind the call delay.

Theoretically, firms should exercise the call to force conversion when the conversion value exceeds the call price, which would maximize the shareholder’s wealth. Previous researchers have demonstrated that, on average, firms wait until the conversion value exceeds the call price by 43.9%.

Dividend policy is one explanation for the call delay. Among the dividend-related rationales, one is applicable when forced conversion means the loss of a valuable corporate tax shield. It is argued that the delay can be optimal when the after-tax cost of the coupons is less than the post-conversion dividends to be paid to former bondholders. As a result, high-dividend firms should prefer to delay the call, whereas less optimistic firms should prefer to call and force conversion.

Since 2003, the percentage of callable convertibles that are dividend protected has increased rapidly. When a convertible is dividend protected and a dividend is distributed, the number of shares to be received upon conversion increases so as to leave the bond’s conversion value unchanged.

The authors show that the dividend-related rationales are not applicable to dividend-protected convertibles. The call delay is near zero (2.47 days) for dividend-protected convertible bonds, whereas the average call delay for non-dividend-protected convertibles is substantially longer (67.22 days). The link between dividend protection and a diminished call delay highlights the importance of dividends in understanding convertible call policy.

To establish a stronger result, the authors further examine those high- and low-dividend firms that issue non-dividend-protected convertibles. In their sample, the call delay is significantly longer for high-dividend firms (181 days on average) than for low-dividend firms (18 days on average), whereas for the dividend-protected convertibles, the size of the dividend is not an important determinant of call delay.

The authors attempt to explain the redesign of convertible bonds to incorporate dividend protection. After 2003, when a US tax act reduced the tax penalty on dividends, dividends were expected to increase. Finance personnel at firms viewed dividend protection as a simple method of reducing the difficulty of valuing the convertible at the time of a perceived increase in dividends. Hedge funds may also prefer dividend-protected convertibles because they are easier to price, and hedge fund managers can thus more successfully hedge their positions.

The authors identify the set of convertible bonds issued by US industrial companies over the period from January 2000 to December 2008 by examining the Securities Data Company (SDC) database. To be included in the sample, issuing firms must have an offering prospectus available on the SEC’s EDGAR database and the convertibles must have call features. After some minor exclusions, 471 convertible bonds remain in the sample.

Among those 471 bonds, only 159 were called; the rest were retired as a result of a merger, a bankruptcy, an exchange of the convertible for other securities, or a similar event. In quantifying the call delay, the authors measure the number of trading days relative to the first date it becomes possible to force conversion.

Among those 159 called convertibles, 51 are classified as out of the money at the time of the call announcement and are excluded from the study as a usual practice. The other 108 are classified as called in the money, of which 45 are dividend protected and 63 are not.

The authors show that for dividend-protected convertibles, the cost to the firm’s shareholders of delaying a call is the sum of the fraction of the cost of future after-corporate-tax coupons that is borne by the shareholders plus the value of the insurance the bondholder retains. When a call is delayed, most of the cost is borne by the firm’s shareholders. The estimated call delay cost per proceed would be $2.6 million; thus, there is apparently no delay. The authors also show that high-dividend convertibles experience a call delay that is 10 times longer than that of their low-dividend counterparts.

The results are robust after controlling for other relevant variables.

The findings of the study further support the dividend-related rationales for the call delay because if non-dividend rationales can fully explain the length of the call delay, there should be no difference in call delay between dividend-protected and non-dividend-protected convertibles.