model is instructive,lower variance usually implies other differences in firm value under risk aversion.At present,we are not aware of any debt-equity signaling model that considers asymmetric information about firm value driven(at least in part)by asymmetric information about firm risk under risk aversion. Vermaelen (1984),Persons(1994,1997),and McNally (1999)model firms'incentives to repurchase shares,holding investment fixed.In these models,better (potentially undervalued) firms repurchase shares to distinguish themselves from worse firms.The Vermaelen and McNally models use a Leland and Pyle-type managerial incentive structure.The Persons'model uses a shareholder heterogeneity device.Since a share repurchase increases a firm's financial leverage,these models suggest a positive correlation between leverage and expected future cash flows. 3.3. Empirical evidence on leverage signaling There are several empirical predictions that emerge from the capital structure signaling models in which firms'investment is fixed.If better quality is based on higher profitability,and if we assume asymmetric information in firm's profits,more-profitable firms have the incentive (and ability)to maintain higher levels of debt in order to signal their value to the market.Yet, Harris and Raviv (1991)observe that much empirical research,including Titman and Wessels (1988),has found the opposite,a negative cross-sectional relation between financial leverage and profitability.Comprehensive studies by Rajan and Zingales(1995),Frank and Goyal(2000),and Fama and French(2002)confirm this empirical finding. Capital structure signaling models predict that higher-value firms will have more financial leverage,given a firm's book value of assets.Again,cross-sectional findings by Rajan and Zingales and Frank and Goyal are not consistent with this prediction.These studies report a negative empirical relation between firms'financial leverage and the market-to-book ratio. 10
10 model is instructive, lower variance usually implies other differences in firm value under risk aversion. At present, we are not aware of any debt-equity signaling model that considers asymmetric information about firm value driven (at least in part) by asymmetric information about firm risk under risk aversion. Vermaelen (1984), Persons (1994, 1997), and McNally (1999) model firmsí incentives to repurchase shares, holding investment fixed. In these models, better (potentially undervalued) firms repurchase shares to distinguish themselves from worse firms. The Vermaelen and McNally models use a Leland and Pyle-type managerial incentive structure. The Personsí model uses a shareholder heterogeneity device. Since a share repurchase increases a firmís financial leverage, these models suggest a positive correlation between leverage and expected future cash flows. 3.3. Empirical evidence on leverage signaling There are several empirical predictions that emerge from the capital structure signaling models in which firmsí investment is fixed. If better quality is based on higher profitability, and if we assume asymmetric information in firm's profits, more-profitable firms have the incentive (and ability) to maintain higher levels of debt in order to signal their value to the market. Yet, Harris and Raviv (1991) observe that much empirical research, including Titman and Wessels (1988), has found the opposite, a negative cross-sectional relation between financial leverage and profitability. Comprehensive studies by Rajan and Zingales (1995), Frank and Goyal (2000), and Fama and French (2002) confirm this empirical finding. Capital structure signaling models predict that higher-value firms will have more financial leverage, given a firmís book value of assets. Again, cross-sectional findings by Rajan and Zingales and Frank and Goyal are not consistent with this prediction. These studies report a negative empirical relation between firmsí financial leverage and the market-to-book ratio
Although this finding is consistent with the growth-opportunities argument for lower leverage suggested by Myers(1977),Rajan and Zingales raise the possibility that the finding is evidence that firms try to time the market by issuing equity when the market-to-book ratio is high. We note that cross-sectional analyses may not be the best tests for signaling models, since the signal could be lost in the noise of variation in other factors that might drive capital structure.To explain cross-sectional variation,we must also assume that firms are alike on all other dimensions.Put another way,capital structure signaling might be valid,even if other factors explain the capital structure differences across firms.Since it may well be that cross- sectional variation across industries and sectors mask the role of asymmetric information in capital structure choices,stronger evidence(and more in keeping with the spirit of the models) can be found in the event study literature. Another implication of the fixed-investment leverage signaling models is that if a firm's managers believe that future profitability will be higher (lower)than current market expectations, then they should conduct a leverage-increasing (decreasing)capital structure transaction.The empirical implication of this finding is that we would expect positive (negative)stock price reactions in response to capital structure changes that increase(decrease)leverage. This empirical implication may also hold even in the Brick,Frierman,and Kim (1998) model,where better quality firms have the incentive(and the ability)to reduce leverage to signal reduced variance to the market.Increasing (decreasing)leverage is associated with a decrease (increase)firm value,but there will be an additional impact on how firm value is divided between debt and equity value.Using the call option view ofequity,a higher (lower)volatility of firm returns,all else equal,will imply a higher(lower)equity value.Thus,even in this model,it 11
11 Although this finding is consistent with the growth-opportunities argument for lower leverage suggested by Myers (1977), Rajan and Zingales raise the possibility that the finding is evidence that firms try to time the market by issuing equity when the market-to-book ratio is high. We note that cross-sectional analyses may not be the best tests for signaling models, since the signal could be lost in the noise of variation in other factors that might drive capital structure. To explain cross-sectional variation, we must also assume that firms are alike on all other dimensions. Put another way, capital structure signaling might be valid, even if other factors explain the capital structure differences across firms. Since it may well be that crosssectional variation across industries and sectors mask the role of asymmetric information in capital structure choices, stronger evidence (and more in keeping with the spirit of the models) can be found in the event study literature. Another implication of the fixed-investment leverage signaling models is that if a firm's managers believe that future profitability will be higher (lower) than current market expectations, then they should conduct a leverage-increasing (decreasing) capital structure transaction. The empirical implication of this finding is that we would expect positive (negative) stock price reactions in response to capital structure changes that increase (decrease) leverage. This empirical implication may also hold even in the Brick, Frierman, and Kim (1998) model, where better quality firms have the incentive (and the ability) to reduce leverage to signal reduced variance to the market. Increasing (decreasing) leverage is associated with a decrease (increase) firm value, but there will be an additional impact on how firm value is divided between debt and equity value. Using the call option view of equity, a higher (lower) volatility of firm returns, all else equal, will imply a higher (lower) equity value. Thus, even in this model, it
is possible for leverage-increasing (decreasing)transactions to have a positive (negative)stock price impact. Event studies involving changes in capital structure have provided a significant amount of evidence indicating that information is conveyed.Moreover,the event study findings are consistent with the spirit of the theoretical models of capital structure and signaling.Harris and Raviv summarize some well-known results from the 1980s that,on average,announcements of leverage-increasing (decreasing)transactions have been accompanied by share price increases (decreases),except in the case of public debt issues,which have been accompanied by insignificant share price changes.2 Masulis(1980)and Vermaelen(1981)find that repurchases financed from debt had larger announcement returns than those financed with cash.Since repurchases financed with debt represent larger increases in financial leverage,this finding is consistent with the theory of capital structure signaling. More recent empirical studies have found similar results for four types of leverage- changing transactions:exchange offers (including swaps),forced conversion of bonds to equity, share repurchases,and seasoned equity offers(SEOs).The results in these studies are consistent with the empirical implications of capital structure signaling models that announcements of leverage-increasing(decreasing)capital structure changes result in positive(negative)share price reactions.3 Erwin and Miller(1998)document negative returns for the rivals of firms announcing 2 For example see Masulis(1980,1983),Vermaelen(1981),Dann(1981),Asquith and Mullins(1986),Mikkelson and Partch(1986),Masulis and Korwar (1986),Pinegar and Lease (1986),Shipper and Smith (1986),Kalay and Shimrat(1987),and Israel,Ofer,and Siegel (1989).For surveys of the early event study literature see Smith (1986) and Masulis (1988) 3 For empirical studies of exchange offers and swaps,see Copeland and Lee (1991),Shah (1994),Born and McWilliams(1997),and Lie,Lie,and McConnell(2001).For studies of forced conversions of debt to equity,see Campbell,Ederington,and Vankudre(1991).For studies of share repurchases(leverage-increasing),see Lakonishok and Vermaelen (1990),Comment and Jarrell (1991),Howe,He and Kao (1992),Ikenberry,Lakonishok,and Vermaelen(1995),Ikenberry and Vermaelen (1995),Erwin and Miller(1998),Chan,Ikenberry,and Lee(2001),and Maxwell and Stephens (2001). 12
12 is possible for leverage-increasing (decreasing) transactions to have a positive (negative) stock price impact. Event studies involving changes in capital structure have provided a significant amount of evidence indicating that information is conveyed. Moreover, the event study findings are consistent with the spirit of the theoretical models of capital structure and signaling. Harris and Raviv summarize some well-known results from the 1980s that, on average, announcements of leverage-increasing (decreasing) transactions have been accompanied by share price increases (decreases), except in the case of public debt issues, which have been accompanied by insignificant share price changes.2 Masulis (1980) and Vermaelen (1981) find that repurchases financed from debt had larger announcement returns than those financed with cash. Since repurchases financed with debt represent larger increases in financial leverage, this finding is consistent with the theory of capital structure signaling. More recent empirical studies have found similar results for four types of leveragechanging transactions: exchange offers (including swaps), forced conversion of bonds to equity, share repurchases, and seasoned equity offers (SEOs). The results in these studies are consistent with the empirical implications of capital structure signaling models that announcements of leverage-increasing (decreasing) capital structure changes result in positive (negative) share price reactions.3 Erwin and Miller (1998) document negative returns for the rivals of firms announcing 2 For example see Masulis (1980, 1983), Vermaelen (1981), Dann (1981), Asquith and Mullins (1986), Mikkelson and Partch (1986), Masulis and Korwar (1986), Pinegar and Lease (1986), Shipper and Smith (1986), Kalay and Shimrat (1987), and Israel, Ofer, and Siegel (1989). For surveys of the early event study literature see Smith (1986) and Masulis (1988). 3 For empirical studies of exchange offers and swaps, see Copeland and Lee (1991), Shah (1994), Born and McWilliams (1997), and Lie, Lie, and McConnell (2001). For studies of forced conversions of debt to equity, see Campbell, Ederington, and Vankudre (1991). For studies of share repurchases (leverage-increasing), see Lakonishok and Vermaelen (1990), Comment and Jarrell (1991), Howe, He and Kao (1992), Ikenberry, Lakonishok, and Vermaelen (1995), Ikenberry and Vermaelen (1995), Erwin and Miller (1998), Chan, Ikenberry, and Lee (2001), and Maxwell and Stephens (2001)
repurchases,which supports the notion that the information conveyed is firm level,not industry- level.Maxwell and Stephens (2001)find that repurchase announcements are accompanied,on average,by negative bond returns.This finding suggests that wealth is redistributed from debtholders to stockholders.*Technically,neither share repurchases nor SEOs will change the financial leverage of an all-equity firm,but the transactions appear to convey information anyway,as in Ferris and Sant (1994).The positive share price reaction to repurchase announcements could be interpreted as supporting the signaling model of Miller and Rock (1995),and the negative share price reaction to SEO announcements may be interpreted as supporting the adverse selection model of Myers and Majluf(1984). The event study evidence on announcement of debt issues does not provide much support for signaling theories.Early studies by Dann and Mikkelson (1984),Eckbo (1986),Mikkelson and Partch (1986),and Shyam-Sunder(1991)find insignificant changes in stock prices in response to straight corporate debt issues.Chaplinsky and Hansen (1993)point out that these results might be due to the predictability of the debt offerings,but Manuel,Brooks,and Schadler (1993)report significantly negative stock price reactions to announcements of debt issues that closely precede dividend and earnings announcements.Johnson (1995)finds significantly positive stock price reactions to debt issue announcements for low-dividend payout firms. Howton,Howton,and Perfect (1998)report significantly negative stock price reactions to announcements of straight debt issues without conditioning on dividend or earnings announcements,and that the stock price announcement is inversely related to investment opportunities(Tobin's Q). 4 For studies on seasoned equity offers(leverage-decreasing),see Korajczyk,Lucas,and McDonald(1991),Brous (1992),Jain (1992),Choe,Masulis,and Nanda (1993),Brous and Kini(1994),Eckbo and Masulis(1995),Loughran and Ritter (1995),Speiss and Affleck-Graves(1995),and Clarke,Dunbar,and Kahle (2001). 13
13 repurchases, which supports the notion that the information conveyed is firm level, not industrylevel. Maxwell and Stephens (2001) find that repurchase announcements are accompanied, on average, by negative bond returns. This finding suggests that wealth is redistributed from debtholders to stockholders.4 Technically, neither share repurchases nor SEOs will change the financial leverage of an all-equity firm, but the transactions appear to convey information anyway, as in Ferris and Sant (1994). The positive share price reaction to repurchase announcements could be interpreted as supporting the signaling model of Miller and Rock (1995), and the negative share price reaction to SEO announcements may be interpreted as supporting the adverse selection model of Myers and Majluf (1984). The event study evidence on announcement of debt issues does not provide much support for signaling theories. Early studies by Dann and Mikkelson (1984), Eckbo (1986), Mikkelson and Partch (1986), and Shyam-Sunder (1991) find insignificant changes in stock prices in response to straight corporate debt issues. Chaplinsky and Hansen (1993) point out that these results might be due to the predictability of the debt offerings, but Manuel, Brooks, and Schadler (1993) report significantly negative stock price reactions to announcements of debt issues that closely precede dividend and earnings announcements. Johnson (1995) finds significantly positive stock price reactions to debt issue announcements for low-dividend payout firms. Howton, Howton, and Perfect (1998) report significantly negative stock price reactions to announcements of straight debt issues without conditioning on dividend or earnings announcements, and that the stock price announcement is inversely related to investment opportunities (Tobinís Q). 4 For studies on seasoned equity offers (leverage-decreasing), see Korajczyk, Lucas, and McDonald (1991), Brous (1992), Jain (1992), Choe, Masulis, and Nanda (1993), Brous and Kini (1994), Eckbo and Masulis (1995), Loughran and Ritter (1995), Speiss and Affleck-Graves (1995), and Clarke, Dunbar, and Kahle (2001)
These event studies of capital structure signaling with fixed investment generally indicate that capital structure transactions do convey information.At present,it remains unclear whether the information pertains to managers'private information about future profitability,changes in risk,and/or managements'belief that shares are simply misvalued by the market (even if the managers have no private information about future profitability or risk.) In addition to the short-term impact of announcements on share prices,several studies examine long-term firm performance subsequent to capital structure changes.Harris and Raviv summarize the early attempts at analyzing long-term performance.Cornett and Travlos (1989) show that firms'earnings tended to increase (decrease)after leverage-increasing (decreasing) events.Dann,Masulis,and Mayers (1991)report evidence of earnings increases after share repurchases.Israel,Ofer,and Siegel(1989)report than analysts revise firms'earnings estimates downward in response to firms'offers to exchange equity for debt. There is now an extensive empirical literature on the issue of firm performance after capital structure transactions.Copeland and Lee (1991),Shah (1994),Born and McWilliams (1997),and Lie,Lie,and McConnell (2001)examine exchange offers,the cleanest leverage- changing transaction,holding investment constant.Copeland and Lee report that systematic risk dropped (rose)after leverage-increasing(decreasing)exchange offers.This finding is consistent with managers increasing (decreasing)leverage when they foresee a change in firm value due to a decrease (increase)in their firm's business risk.However,this finding does not explicitly correspond to existing capital structure signaling theory.Shah's work shows that leverage- decreasing transactions are followed by abnormal reductions in operating cash flows (not earnings),but that leverage-increasing transactions are not followed by increases.On the other hand,on average,leverage-increasing offers precede declines in the systematic risk of the firm's 14
14 These event studies of capital structure signaling with fixed investment generally indicate that capital structure transactions do convey information. At present, it remains unclear whether the information pertains to managersí private information about future profitability, changes in risk, and/or managements' belief that shares are simply misvalued by the market (even if the managers have no private information about future profitability or risk.) In addition to the short-term impact of announcements on share prices, several studies examine long-term firm performance subsequent to capital structure changes. Harris and Raviv summarize the early attempts at analyzing long-term performance. Cornett and Travlos (1989) show that firmsí earnings tended to increase (decrease) after leverage-increasing (decreasing) events. Dann, Masulis, and Mayers (1991) report evidence of earnings increases after share repurchases. Israel, Ofer, and Siegel (1989) report than analysts revise firmsí earnings estimates downward in response to firmsí offers to exchange equity for debt. There is now an extensive empirical literature on the issue of firm performance after capital structure transactions. Copeland and Lee (1991), Shah (1994), Born and McWilliams (1997), and Lie, Lie, and McConnell (2001) examine exchange offers, the cleanest leveragechanging transaction, holding investment constant. Copeland and Lee report that systematic risk dropped (rose) after leverage-increasing (decreasing) exchange offers. This finding is consistent with managers increasing (decreasing) leverage when they foresee a change in firm value due to a decrease (increase) in their firmís business risk. However, this finding does not explicitly correspond to existing capital structure signaling theory. Shahís work shows that leveragedecreasing transactions are followed by abnormal reductions in operating cash flows (not earnings), but that leverage-increasing transactions are not followed by increases. On the other hand, on average, leverage-increasing offers precede declines in the systematic risk of the firmís