The Review of Financial Studies/v6n 4 1993 market index and is calculated with daily data.[Our results are invari- ant to including a second industry portfolio in the market model,to using a Scholes-Williams(1977)adjustment,and to freezing or float ing the beta-calculation window with each month.The precise method of adjusting equity returns is described in the Appendix.Additional data,such as firms'capital structure or bond ratings,were either hand- collected or read from the COMPUSTAT or the Lehman Brothers database.1s 2.Bondholder Wealth Losses 2.1 Changes in capital structure Table 3 describes the capital structure changes that our LBO firms undergo,as obtained from the COMPUSTAT data base.We find that these firms typically at least triple both their long-term liabilities and their debt in current liabilities.16 Except for Colt Industries,17 the leverage increases are large by any measure. An increase in a firm's leverage need not imply that outstanding bondholders necessarily lose wealth.According to Kim,McConnell, and Greenwood (1977),bonds can be protected against future issues with priority covenants (which prevent issuance of debt of equal or higher priority or shorter duration/maturity).However,Franks and Torous (1989)document that bond covenants are not likely to protect bondholders in case of financial distress.18 In their sample,bankruptcy proceedings stretch over three and one-half years,indicating that bankruptcies can be quite costly.They also show that deviations from absolute priority are the rule rather than the exception.Thus,since seniority rules for bonds that continue to be outstanding are not ex ante enforceable,it is likely that leverage increases produce wealth losses.Moreover,none of the bonds in our sample had explicit put or conversion features. 2.2 Changes in bond characteristics The changes in capital structure are accompanied by the expected downgradings in debt ratings for the bonds.When ranked on a scale We neither excluded any bonds from our sample because of missing information on CRSP and/ or COMPUSTAT,nor were any of our bonds retired during or immediately after the LBO.Only one Leaseway bond was defeased several months after our event window. 16 Because of missing data,the number of entries is different in the ex ante and ex post columns. Thus,the difference in means should not be used to take the ratio of ex ante and ex post means. Colt experienced a leveraged recapitalization a year before the LBO announcement.(Nevertheless, Table 4 indicates that Colt's bondholders suffered substantial losses.) See also Eberhart,Moore,and Roenfeldt (1990)and Weiss (1991). 970