22

Admati, Demarzo, Hellwig, and Pfleiderer (ADHP, 2018) note that static models of optimal leverage have assumed firms have no prior debt. In this case, the leverage that maximizes firm value also maximizes value to the initial equity owners. However, using a simple two-period model with zero coupon debt and default possible only at maturity, ADHP prove two startling results: (i) when prior debt is extant, it will never benefit equity holders to retire debt, no matter how high the current leverage; and (ii) it will be in the equity owners’ interest to issue sequential rounds of additional debt, until all the tax advantages of debt are exhausted: the “Leverage Ratcheting Effect” (LRE). An immediate conclusion is that one-round (static) models of optimal debt issuance with no prior debt provide poor guidance as to a firm’s optimal leverage. We examine these contentions using an alternative model of debt, with rollover at a proportional rate m and average maturity = 1/ m, introduced in Leland (1994a). We show that when the average maturity of debt is substantially longer than 5 years, considerable further debt will indeed be issued, although issuance ceases well before tax benefits are exhausted. With 5-year average maturity, very little additional debt is issued under reasonable calibrations. With 3-year average maturity, no additional debt is issued and it may actually be optimal for the firm to buy back debt, in contradiction to the LRE. We explain why our model gives differing results.