Corporate risk management and dividend signaling theory. Georges Dionne and Karima Ouederni. HEC Montréal. (May 2010)

Transcription

1 Corporae risk managemen and dividend signaling heory Georges Dionne and Karima Ouederni HEC Monréal (May 010) Absrac This paper invesigaes he effec of corporae risk managemen on dividend policy. We exend he signaling framework of Bhaacharya (1979) by including he possibiliy of hedging he fuure cash flow. We find ha he higher he hedging level, he lower he incremenal dividend. This resul is in line wih he purpored posiive relaion beween informaion asymmery and dividend policy (e.g., Miller and Rock, 1985) and he asserion ha risk managemen alleviaes he informaion asymmery problem (e.g., DaDal e al., 00). Our heoreical model has esable implicaions. Keywords: Signaling heory; Dividend policy; Risk managemen policy; Corporae hedging; Informaion asymmery. JEL classificaion codes: G35; G3; D8. 1 Inroducion Signaling heory saes ha changes in dividend policy convey informaion abou changes in fuure cash flows (e.g., Bhaacharya, 1979, Miller and Rock, 1985, John and Williams, 1985). Dividend signaling suggess a posiive relaion beween informaion asymmery and dividend policy. 1 In oher words, he higher he asymmeric informaion level, he higher is he sensiiviy of he dividend o fuure prospecs of he firm. Several empirical sudies aemp o es he informaional conen of dividend changes, ye hey disagree abou he sign and he significance of he effec of informaion asymmery on dividend policy. Anoher srand of lieraure suggess ha corporae risk managemen alleviaes informaion asymmery problems and hence posiively affecs he firm value. Informaion asymmery beween managers and ouside invesors is one of he key marke imperfecions ha make hedging poenially beneficial. DeMarzo and Duffie (1995) and Breeden and Viswanahan 1 Evidence ha informaion asymmery posiively affecs dividend policy has also been documened by he free cash flow heory (e.g., Lang and Lizenberger, 1989). 1

2 (1989) argue ha hedging reduces noise around earnings sreams and hus decreases he level of asymmeric informaion regarding he firm value. DaDal e al. (00) provide empirical evidence supporing hese heoreical sudies. In his paper we exploi he documened ineracion beween he level of informaion asymmery and he dividend policy, along wih is ineracion wih corporae risk managemen. We argue ha risk managemen alleviaes he asymmeric informaion problem, which is a main deerminan of dividend policy. Though many sudies ha examine dividend policy deerminans include several measures of informaion asymmery, none, o our knowledge, consider hedging among hese measures. Exending he signaling framework of Bhaacharya (1979), we provide heoreical suppor for he effec of corporae risk managemen on dividend payou policy. We find a negaive relaion beween he hedge raio and he incremenal dividend payou. The remainder of he aricle is organized as follows. In he second secion we summarize he main lieraure relaed o he relaionships beween informaion asymmery, corporae risk managemen and dividend policy and highligh he poenial conribuion of he paper. In he hird secion we presen he heoreical model and is implicaions. The fourh secion concludes. Lieraure review The implicaions of asymmeric informaion beween he manager and ouside invesors have been he subjec of exensive research in he corporae finance lieraure. We focus on wo major proposiions in his lieraure: (i) corporae risk managemen alleviaes asymmeric informaion problem and (ii) dividend policy is relaed o asymmeric informaion. Taken ogeher, hese proposiions imply ha he dividend policy of a firm and is risk managemen policy may be inerrelaed because boh are relaed o he level of informaion asymmery..1 Informaion asymmery and dividend policy In heir seminal paper, Miller and Modigliani (1961) suppor and demonsrae he irrelevance of he dividend policy on he firm value. However, corporaions used and coninue o payou considerable amouns of dividends o heir shareholders. Moreover, empirical sudies idenify a significan effec of dividend changes on sock reurns (e.g. Aharony and Swary 1980). They confirm ha he announcemen of dividend increases or iniiaions (decreases or cus) are usually associaed wih posiive (negaive) excess sock reurns. Consequenly, some

3 heoreical papers relax he hypohesis of perfec capial marke and provide raionales for dividend paymen and is posiive effec on he marke value of he firm. The dividend lieraure rely essenially on he exisence of wo marke imperfecions o explain why firms pay ou as much of heir earnings o shareholders: informaion asymmery and agency conflics. Based on hese marke imperfecions wo respecive dominan heories aemp o explain he corporae dividend policy. The signaling heory suggess ha, in he presence of informaion asymmery beween insiders and ouside invesors, changes in dividend policy convey informaion abou changes in fuure performance of he firm. Thus he higher he level of informaion asymmery he more informaive are dividend changes and he higher is heir effec on he firm s marke value. The free cash flow heory suggess ha in he presence of agency problems, managers wih subsanial cash flows and poor invesmen opporuniies have incenives o wase money in perquisies or negaive NPV projecs, raher han pay i ou o shareholders. Thus hey inerpre dividend changes as an indicaor of he misuse of cash flows by he manager of firms wih high balance shee liquidiy and low invesmen opporuniies (e.g. Lang and Lizenberger, 1989). Accordingly, hey sugges a posiive relaionship beween he dividend policy and he level of informaion asymmery. Several empirical sudies aemp o es hese heories and o provide argumens, in favor or agains hem. In his paper we focus on he main heoreical papers and empirical ess of he dividend signaling heory (see Allen and Michaely, 003, for an exensive lieraure review). Among he firs and he mos known signaling models are hose of Bhaacharya (1979), Miller and Rock (1985) and John and Williams (1985). A crucial assumpion of hese models is ha dividend is a cosly signal ha i is no worhwhile for firms wih bad fuure prospecs o imiae. In oher words hese models are based on he exisence of a dissipaive cos ha allows signaling o occur and makes a separaing equilibrium feasible (e.g. ransacion coss, underinvesmen, axes). For insance, in Bhaacharya (1979), he dividend increases he probabiliy ha he firms will need cosly exernal financing o finance is fuure invesmens. Unless he firm has really good fuure earnings prospecs i will be relucan o increase is dividend. Acually, in he case of insufficien payoff he firm will resor o exernal financing and incur he relaive ransacion coss. In Miller and Rock (1985), he deviaion from he opimal invesmen level is he cos of he dividend. When more profiable firms pay a sufficienly high dividend i becomes unaracive for less profiable firms o cu heir invesmen o imiae his signal. In John and Williams (1985), axes are he dividend cos. 3

4 Two main criicisms are addressed o he earlier signaling models. Some of hem consider dividends and share repurchases as subsiues (e.g. Bhaacharya, 1979, Miller and Rock, 1985). Ohers fail o explain dividend smoohing (e.g. John and Williams, 1985). Mos of hese models are single signal models. Afer his firs generaion of signaling models oher models wih muliple signals were proposed (e.g. Williams, 1988, Consaninides and Grundy, 1989). They were developed o exend earlier models and/or ovoid heir caveas. These models focused mainly on he ineracion beween dividend policy and differen corporae decisions (e.g. invesmen, financing, sock repurchase). On he empirical side, mos dividend signaling lieraure was ineresed in esing wo main implicaions of he hypohesis relaed o he informaion conen of dividends : (i) dividend changes should be followed by subsequen earnings changes in he same direcion and (ii) unanicipaed dividend changes should be followed by sock price changes in he same direcion. There is a general agreemen in he lieraure abou he effec of dividend changes on he sock price. Bu his is no he mos basic implicaion. According o Allen and Michaely (003), he fundamenal implicaion of he signaling heory is ha dividend changes and earnings changes should move in he same direcion. However, his implicaion has received lile empirical suppor. Acually empirical sudies diverge abou he sign and he significance of he relaionship beween dividend changes and fuure earnings changes. Some sudies fail o find any significan changes in earnings following dividend changes (e.g. Was (1973), Penman (1983)). However, oher sudies confirm he exisence of significan increases in earnings in he year of and he year afer dividend increases (Healy and Palepu, 1988, Brickley, 1983). Michaely, Benarzi and Thaler (1997) found a significan posiive relaionship beween changes in dividends and changes in pas and conemporaneous earnings. Bu hey did no find any evidence of such relaion beween curren changes in dividends and fuure changes in earnings. Nessim and Ziv (001) used he same sample as Michaely e al. (1997) bu a differen mehodology (e.g. linear mean reversion model of earnings). They found ha dividend changes are posiively and significanly relaed o fuure earnings of he firm. Michaely e al. (004) confirm heir earlier resuls by using a non-linear mean reversion model of earnings. Using a differen sample, Verdugo (006) reaches he same conclusion as Nessim and Ziv (001). A common predicion of he mos signaling models is ha dividend changes are used o signal changes in fuure earnings or cash flows. However, his predicion has received lile empirical suppor. Acually he resuls of he empirical sudies abou he exisence and he 4

5 significance of a posiive relaionship beween curren dividend changes and fuure earnings changes are inconclusive. Allen and Michaely (003) discuss an alernaive possibiliy suggesing ha increases in dividend may convey informaion abou changes in risk raher han in performance. Thus furher ess of he dividend signaling heory are ye needed.. Informaion asymmery and corporae risk managemen In a perfec capial marke seing, shareholders are well diversified and care only abou he sysemaic risk of he firm. Thus corporae risk managemen which consiss mainly in he hedging of he unsysemaic risk (ineres rae, exchange rae and commodiy prices) is irrelevan. However, in he real world, many firms are engaged in hedging aciviies of unsysemaic risk. Moreover heir managers rank risk managemen as one of heir mos imporan objecives (e.g. Barram, 000). Several heoreical papers inroduce some marke imperfecion o provide raionales for corporae decision o hedge unsysemaic risk (see Sprčić e al. 009, for a survey). Acually hedging reduces he dispersion of operaing cash flows, which reduces financial disress and bankrupcy coss, exernal financing coss, axes liabiliies and he informaion asymmery beween managers and ouside invesors regarding he expeced fuure cash flows. Thus, corporae hedging enhances he marke value of he firm. Informaion asymmery is a key marke impacion ha makes corporae hedging poenially beneficial. However i is less frequenly examined by he risk managemen lieraure. There are few papers ha provide heoreical suppor for he effec of corporae hedging on he level of informaion asymmery. For insance DeMarzo and Duffie (1995) and Breeden and Viswanahan (1989) use he inuiion ha hedging lessens he asymmeric informaion regarding firm value by reducing he noise in evaluaion measures. They argue ha hedging reduces he exposure of he firm o facors beyond he manager s conrol and hus reduces he noise in earnings and fuure cash flows esimaes which are usually used by ouside invesors as measures of firm value. DaDal e al. (00) provide empirical evidence supporing hese heoreical sudies..3 Dividend policy and corporae risk managemen Several empirical sudies consider he dividend policy as a significan deerminan of he corporae risk managemen policy. Bu hey reach differen conclusions. Some sudies argue ha hedgers have lower dividend yields and dividend payous han no hedgers (e.g. Smih and Was, 199). Also, Ross (1997) argues ha he beer is a firm's abiliy o hedge; he more frequenly i will refrain from paying dividends. These resuls suppor he hypohesis ha 5

6 risk managemen miigaes he informaion asymmery problem which is he main raionale for dividend disribuion according o he signaling heory. Oher sudies using more accurae measure of he level of hedging (e.g. hedge raio) consider he dividend policy as a measure of he balance shee liquidiy and/or he level of informaion asymmery and reach he same conclusion (e.g. Dionne and Garand 003, Haushaler, 000). However, some sudies provide argumens in he favor of a posiive relaionship beween hedging aciviy and dividend disribuion. These sudies suppor he hypohesis ha a low dividend yield could be perceived as a subsiue o hedging (e.g. Nance e al., 1993, Mian, 1996, DeAngelis & Garçia, 008). Despie he abundan lieraure on he deerminans of he dividend policy, no sudy considers he effec of he corporae hedging. Our wo main commens abou he empirical lieraure are: (i) here is no consensus abou he effec of he dividend policy of a firm on is risk managemen policy, and (ii) lack of empirical resuls abou he effec of corporae risk managemen on he corporae dividend policy. Thus, he relaionship beween he corporae dividend policy and he corporae hedging needs more heoreical and empirical suppor..4 Conribuion This paper exends he signaling model of Bhaacharya (1979) by inroducing a corporae risk managemen sraegy. I is a muliple signal model which emphasizes on he ineracion beween he dividend policy and he corporae risk managemen as subsiues policies. To he bes of our knowledge, no signaling model sudied he ineracion beween hese wo policies. In his model we predic a negaive relaionship beween hese wo corporae policies. This resul is in line wih he dividend signaling heory. According o he predicions of his heory, if corporae risk managemen affecs negaively he level of informaion asymmery, hen higher hedging level should be associaed wih lower dividend yield and/or payou. Also, i considers implicily he alernaive predicion of signaling models: dividend changes convey informaion abou changes in he firm risk. Acually if dividend changes signal decreases in he firm s risk hen high dividend yield should be associaed wih low hedge raio. In boh cases if he predicions of he dividend signaling heory are verified, a negaive and significan relaionship beween dividend levels and hedging levels should be observed. Also, he negaive relaion beween dividend increases and hedging levels may explain he phenomena of disappearing dividends documened by Fama and French (001). Moreover, i may explain he dividend smoohing. Acually, if he signaling power of dividend changes 6

7 weakens or vanishes in he presence of corporae risk managemen hen firms should become less moivaed o change heir dividend policies jus o signal heir fuure performance. 3 The model We assume ha he firm operaes in a dividend signaling world as modeled in Bhaacharya (1979). We assume ha shareholders have a single-period planning horizon and he manager operaes in he bes ineres of curren shareholders. The model is developed in erms of marginal analysis for a new projec aken on by he firm. We assume ha he manager is beer informed han ouside invesors abou he firm s fuure prospecs. Thus he manager is he only agen informed abou he disribuion of he new projec fuure cash flow (x). He aemps o signal his privae informaion via he commimen of an incremenal dividend (D). Dividends are axed a he rae while capial gains are no axed. There is a penaly (β) incurred by he firm in case of cash flow shorfall o cover he commied dividend. is he dissipaive cos ha allows signaling o occur ( e.g. he cos of exernal financing). If he cash flow (x) exceeds he commied dividend, D is paid o shareholders and he amoun of fuure exernal financing is reduced by (x-d). And if he cash flow (x) is lower han he commied dividend, D is paid o shareholders and he amoun of fuure exernal financing is increased by (x-d). In his model, he dividend increases he probabiliy ha he firms will need cosly exernal financing for is fuure invesmens. Unless he firm has really good fuure earnings prospecs i will be relucan o mimic good firms and increase is dividend. We exend he model by assuming ha i is possible for he manager o hedge a fracion (h) of he fuure cash flow using a linear sraegy in he spiri of Froo, Scharfsein and Sein (1993). The reasoning behind his exension is simple. Ouside invesors ofen use esimaes of earnings and cash flows as measures of firm value. Hedging reduces he noise around earnings and fuure cash flows by reducing he exposure of he firm o facors beyond he manager s conrol. Consequenly, hedging lessens he asymmeric informaion regarding firm value by reducing he noise in evaluaion measures. We expec ha he more willing he firm is o hedge is fuure cash flow, he less informaive he dividend changes and he lower he manager s incenives for cosly signaling hrough dividends. We make he implici assumpion ha corporae hedging aciviy is observable by ouside invesors. This assumpion is realisic given he implemenaion of many disclosure requiremen regulaions by he Financial Accouning Sandards Board (FASB) since he beginning of he 90s (e.g., FAS105, FAS107, FAS119, FAS133, FAS138 and FAS161). 7

8 Equaion (1) illusraes he new uncerain cash flow resuling from he linear hedging sraegy: x1 hx0 1 h x (1) where 0 h 1 is he hedge raio and x 0 he expeced cash flow. The incremenal par of he objecive funcion of curren shareholders is given by equaion (). The four erms in he equaion are respecively: (i) he rise in he firm s liquidaion value V(D); (ii) he afer-ax promised dividend; (iii) he expeced gain when he hedged cash flow is greaer han he commied dividend; and (iv) he expeced loss when he hedged cash flow is lower han he commied dividend: W x V ( D ) (1 ) D ( x0 h (1 h) x D ) f ( x) dx 1 x D ( D, h) () x 1 r D (1 ) ( x0 h (1 h) x D ) f ( x) dx x where r is he afer-ax ineres rae; V(D) he response liquidaion value of he firm resuling from he commimen and he paymen of D, and x D he minimum cash flow needed o pay he promised dividend wihou penaly. Is value is given by equaion (3): x D D hx 0 (3) (1 h ) 3.1 Opimal dividend for a given hedge raio Following Bhaacharya (1979) we assume ha he fuure cash flow is uniformly disribued over [0, ]. Thus he maximizaion problem is reduced o: The firs order condiion solves: 1 1 D h max W( D, h) V( D) D (4) D 1r 8 1h V * D h * ( D ) (5) (1 h) A he opimum, he marginal profi from he dividend increase (he increase of he firm value) equals is marginal cos (axes and expeced cos of exernal financing). The second order condiion is given by equaion (6): 1 W( D, h) V( D) (6) D 1 r (1 h) 8

9 Since V(D) is increasing and concave, h lower han one is a sufficien condiion for he second order condiion o be saisfied. The hedge raio h lies beween 0 and 1 given ha speculaion and over hedging are no considered in his model. 3. Signaling equilibrium The signaling equilibrium demands ha V(D*()) mus be equal o he rue value of fuure cash flows for he projec whose cash flows are signaled wih dividend D ( Bhaacharya (1979), p. 64). Under he assumpion of a saionary dividend, he equilibrium funcion V(D * ()) is given by equaion (7): * * * 1 D h VD K D 8 1 h (7) where K=1/r. Differeniaing equaion (7) wih respec o and subsiuing for V (D * ) from equaion (5), we obain: K 1 D h dd K h D (8) 1h 1h d 81h 1h Assuming a linear soluion for he firs order differenial equaion (8), D () A, we obain he following quadraic equaion: K K 1 K A 1h ha 1h h 0 (9) K 1 K 1 4 K 1 The posiive roo of he quadraic equaion is equal o: hh K K 1 1 A K K 1 1h h K K K 1 h 41 h (10) When h=0 we obain he corresponding values for (9) and (10) in Bhaacharya (1979). A is he incremenal dividend payou. I also illusraes he sensiiviy of dividend increases o earnings prospecs. For reasonable values of ax rae ( 40%) and exernal financing cos (β 0%), A is decreasing in he hedge raio. In Figure 1, we show he funcion wih some 9

10 feasible parameers. We observe a negaive effec of he hedge raio on he incremenal dividend payou. This resul is inuiive. I is in line wih sudies suggesing ha cash flows predicabiliy decreases he marginal gain from cosly signaling hrough dividends (e.g., Chang e al. 006) and he asserion ha corporae hedging decreases cash flows volailiy and hus increases heir predicabiliy A Hedge Raio Figure 1 The curve illusraes A for 0<h<1, = 40%; β = 0% and r = 5%. The sraigh line illusraes A for h=0 (Bhaacharya, 1979). Noice ha A is no sricly decreasing in he hedging level for all values of ax raes and exernal financing coss. For high bu less feasible values of > 40% and β> 0%, A firs increases and hen decreases in h. For all values of he parameers, A always converges o 0.5 when h is near 1. 10

11 3.3 Opimal hedge raio Anoher way o emphasize he ineracion beween he dividend policy and he corporae hedging policy is o maximize he incremenal shareholders wealh in (4) wih respec o he hedge raio. The firs and second order condiions along wih he signaling equilibrium condiion provide he following opimal hedge raio (See appendix for deails): h * D (1 ) (11) The opimal hedge raio is decreasing in he dividend payou raio. This resul is in line wih empirical sudies suggesing a negaive effec of he dividend policy on he hedge raio (e.g., Dionne and Garand, 003). I is also in line wih dividend signaling heory inuiion. I suggess ha managers wih higher expecaions abou fuure performance of he firm disribue higher dividends while reducing heir hedge raio. Thus, in a signaling world shareholders are beer off when he firm deviaes from he full hedging sraegy. Finally, (11) indicaes clearly ha D and h are inerdependen decision variables. In fac we can verify ha * 0h 1 1 as 1 D. 4 Conclusion The findings of his paper reconcile dividend signaling heory wih risk managemen heory. We conribue o he dividend signaling lieraure by emphasizing he ineracion beween corporae risk managemen policy and dividend policy. The ineracion beween hese wo corporae policies has received less aenion in he lieraure despie heir common link o informaion asymmery. Using an exension of Bhaacharya s signaling model, we find ha he hedging of fuure cash flows reduces he sensiiviy of dividends o fuure earnings. A sraighforward implicaion of his resul is ha he informaional conen of dividend changes decreases wih he hedge raio. I leads o he empirical es of wheher corporae risk managemen reduces he power of dividend changes o predic fuure changes in earnings. I hus represens a new es of he dividend signaling heory. 11

12 References Aharony, J., Swary, I., Quarerly Dividend and Earnings Announcemens and Sockholders Reurns: An Empirical Analysis. The Journal of Finance 35, 1-1. Allen, F., Michaely, R., 003. Payou Policy, in: G.M. Consaninides & M. Harris & R. M. Sulz, eds., Handbook of he Economics of Finance, Vol. 1, Chaper 7, Barram, S.M., 000. Corporae Risk Managemen as a Lever for Shareholder Value Creaion. Financial Markes, Insiuions and Insrumens 9, Bhaacharya, S., Imperfec Informaion, Dividend Policy and The Bird in he Hand fallacy. Bell Journal of Economics 10, Breeden, D., Viswanahan, S., Why do firms hedge? An Asymmeric Informaion Model. Working Paper, Fuqua School of Business, Duke Universiy. Brickley, J., Shareholders Wealh, Informaion Signaling, and he Specially Designaed Dividend: An Empirical Sudy. Journal of Financial Economics 1, Chang, C., Kumar, P., Sivaramakrishnan, K., 006. Dividend changes, cash flow predicabiliy, and Signaling of Fuure Cash Flows. SSRN online library N o Consaninides, G., Grundy, B., Opimal Invesmen wih Sock Repurchase and Financing as Signals", Review of Financial Sudies, DaDal, P., Gay, G., Nam J., 00. Asymmeric Informaion and Corporae Use of Derivaives. The Journal of Fuure Markes, DeAngelis, D. and R. Garcia (008). A Mulirisk Approach o Measuring Corporae Hedging and is Deerminans. SSRN online library. DeMarzo, P., Duffie, D., Corporae Incenives for Hedging and Hedge Accouning. Review of Financial Sudies, Fall, Dionne, G., Garand, M., 003. Risk Managemen Deerminans Affecing Firms Values in he Gold Mining Indusry: New Empirical Evidence. Economics Leers 79, Fama, E., French K., 001. Disappearing Dividends: Changing Firm Characerisics or Lower Propensiy o Pay?" Journal of Financial Economics 60, Froo, K., D. Scharfsein and J. Sein (1993). Risk Managemen: Coordinaing Corporae Invesmen and Financing Policies. The Journal of Finance, 48, pp

13 Haushaler, D. (000). Financing Policy, Basic Risk and Corporae Hedging: Evidence from Oil and Gas Producers. Journal of Finance 55, Healy, P., Palepu, K., Earnings Informaion Conveyed by Dividend Iniiaions and Omissions. Journal of Financial Economics, John, K. and J. Williams (1985). Dividends, Diluion, and Taxes: A Signaling Equilibrium. The Journal of Finance 40, Lang, L. and R. Lizenberger (1989). Dividend announcemens: cash flow signaling vs. free cash flow hypohesis? Journal of Financial Economics 4, Lang, L., Lizenberger, R., Dividend Announcemens: Cash Flow Signaling vs. Free Cash Flow Hypohesis? Journal of Financial Economics 4, Mian, S., Evidence on Corporae Hedging Policy. Journal of Financial and Quaniaive Analysis 31, Michaely, R., Benarzi, S., Thaler, R., Do Changes in Dividends Signal he Fuure or he Pas? The Journal of Finance 5, Michaely, R., Benarzi, S., Thaler, R., 004. Dividend Changes Do No Signal Changes In Fuure Profiabiliy. SSRN online library N o Miller, M. and K. Rock (1985). Dividend Policy under Asymmeric Informaion. The Journal of Finance 40, Miller, M., Modigliani F., Dividend Policy, Growh, and he Valuaion of Shares. The Journal of Business 34, Miller, M., Rock, K., Dividend Policy under Asymmeric Informaion. The Journal of Finance 40, Nance D., C. Smih and C. Smihson (1993). On he deerminans of Corporae Hedging. Journal of Finance 48, Nessim, D., Ziv, A., 001. Dividend Changes and Fuure Profiabiliy. Journal of Finance 61, Penman, S., The predicive conen of earnings forecass and dividends. Journal of Finance 38, Ross, S., The Deerminaion of Financial Srucure: The Incenive Signaling Approach", Bell Journal of Economics,

14 Smih, C., and R. Was (199). The Invesmen Opporuniy Se and Corporae Financing, Dividend, and Compensaion Policies. Journal of Financial Economics 3, Sprčić D. M., Tekavčič M. and Šević Ž., 009. A Review of he Raionales for Corporae Risk Managemen: Fashion or he Need? SSRN online library N o Sulz, R., Opimal Hedging Policies. Journal of Financial and Quaniaive Analysis 19, Verdugo, A., 006. Dividend Signaling and Unions. SSRN online library N o Was, R., The Informaion Conen of Dividends. Journal of Business 46, Williams, J., Efficien Signaling wih Dividends, Invesmen, and Sock Repurchases. The Journal of Finance 43 ( 3):

15 Appendix The firs order condiion of he maximizaion of (4) wih respec o h equals: 1 1 (1 h )( D h) ( D h) W ( D, h) 0 h 1 r 8 (1 h) ( D h)( (1 h) ( D h)) 0 The second order condiion wih respec o h is equal o: 1 1 1h 8D4 h h 8D W( D, h) 3 h 8 1 r (1 h) The maximizaion problem has wo complemenary soluions: D D 1 if * h D D 1 1 if When we subsiue for h= D/ in he equilibrium payou raio (10) we ge: A D h 1 K ( K 1) I is sraighforward o show ha A is bigger han 0,5. Noice ha A=D*/ and when i is bigger han 1/ he second order condiion of he maximizaion of shareholders wealh wih respec o h is no saisfied so we do no reain his poenial soluion. When we subsiue for h= (1-D/) in he equilibrium payou raio (10) we ge: A D h 1 K 1 (1 K) I is sraighforward o show ha for feasible values of β and τ, A is bigger han 0,5 which saisfies he second order condiion of he maximizaion of shareholders wealh wih respec o h. 15