Corporate governance, nance, and the real sector

Transcription

1 Corporate goverace, ace, ad the real sector Paolo Fulghieri (UNC, CEPR ad ECGI) ad Matti Suomie (Aalto Uiversity, Helsiki) May 18, 010 For helpful commets, the authors would like to thak Philip Bod, Isil Erel, Michel Habib, Philipp Hartma, Campbell Harvey, Floria Heider, Roma Iderst, Matthias Kahl, Yrjö Koskie, Max Maksimovic, Holger Mueller, Marco Pagao, Gordo Phillips, Michael Roberts, Merih Sevilir, Adrei Schleifer, Raj Sigh, ad semiar participats at the Wester Fiace Associatio, the ASSA, ad the Europea Fiace Associatio meetigs, the Baca d Italia/CEPR Coferece o Moey, Bakig ad Fiace, the Secod Fiace Summit, the Workshop o Politics of Corporate Goverace at the Copehage Busiess School, the Corporate Goverace coferece i Washigto Uiversity i St. Louis, New York Uiversity, Uiversity of Zurich, Uiversity of Housto, Wharto School, Uiversity of Marylad, Helsiki School of Ecoomics, HKUST, Natioal Uiversity of Sigapore, Sigapore Maagemet Uiversity, ad the Luxembourg School of Fiace. Some of the results i this paper were cotaied i a earlier workig paper etitled Does bad corporate goverace lead to too little competitio?. All errors are oly our ow. s: matti.suomiehse. ad Paolo_Fulghieriuc.edu. Matti thaks OP-foudatio ad the Fiish Securities Markets Foudatio for acial support.

2 Corporate goverace, ace, ad the real sector Abstract This paper presets a theory of the likages betwee corporate goverace, corporate ace ad the real sector of a ecoomy. We examie a model of idustry equilibrium with edogeous etry. We show that poor corporate goverace ad low ivestor protectio geerates less competitive ecoomies, populated by rms with more isider owership ad greater leverage. The quality of the corporate goverace system ca also a ect a ecoomy s idustrial structure: better corporate goverace promotes the developmet of sectors more exposed to moral hazard, such as the high-techology idustry. We also show that etrepreeurs may have a preferece for "extreme" corporate goverace systems, where the quality of corporate goverace ad the level of ivestor protectio are either very high or very low. This suggests that etrepreeurs operatig i ecoomies edowed with a corporate goverace system of low-quality may have little or o icetive to seek (or to lobby for) a improvemet of the goverace system of their ecoomy. Fially we show that acial liberalizatios facilitate rm etry ad the adoptio of more productive techologies, promotig ecoomic growth. Our stylized model geerates predictios that are cosistet with several observed empirical regularities.

3 1. Itroductio What is the impact of corporate goverace o the acial ad idustrial structure of a ecoomy? Cosider, for example, the case of Filad. Durig the past three decades the Fiish acial markets experieced a major shift from a bak-based acial system, similar to that i cotietal Europe ad Japa, towards a Aglo-Saxo type acial system based primarily o securities markets. The stock market boomed, the bakig sector cosolidated ad the owership structure of compaies chaged dramatically, as domestic istitutios divested their shareholdigs, especially to foreig ivestors. 1 I parallel, the idustrial compositio ad acial structure of the ecoomy also chaged: earlier o, the Fiish ecoomy was domiated by highly levered compaies, mostly related to the heavy metal ad forest idustry, whereas today it is domiated by a equity aced high-tech sector. Hyytie, Kuosa, ad Takalo (00) show that these shifts i corporate acig ad the real sector followed a major chage i the corporate goverace regime of the coutry, ad argue that the developmet of shareholder protectio was a major driver i this reorgaizatio. More geerally, the quality of a coutry s corporate goverace appears to be systematically related to the degree of competitiveess of its ecoomy ad its idustrial ad acial structure. 3 I this paper, we preset a theory of the likages betwee corporate goverace, corporate ace ad the real sector of a ecoomy. By usig a parsimoious model, we study the relatioships that emerge edogeously amog the corporate goverace system of a ecoomy ad its idustrial ad acial structure, ad geerate empirical predictios that are cosistet with several observed stylized facts. We examie a model of idustry equilibrium where etry by rms ito a idustry is edoge- 1 See, e.g., Hyytie, Kuosa, ad Takalo (00) ad Karhue ad Keloharhu (001). A example of the shift i idustrial compositio is that i the year 000 the Fiish rms led domestically early twice as may patet applicatios as i 1980, at a per capita rate that was the secod highest i the Europea Uio. Today the coutry raks as oe of the most competitive ad least corrupt coutries i the world, accordig to the rakigs from World Ecoomic Forum, IMD ad Trasparecy Iteratioal. 3 I Appedix I we provide empirical evidece showig that coutries characterized by better corporate goverace ad ivestor protectio have a more accessible local equity market, a higher degree of competitio, ad a more developed high-techology sector. I a similar spirit, Agrawal (009) shows that product-market competitio i a certai US state icreases after passage i that state of Blue Sky laws, a statue aimed at ivestor protectio at state level. 1

4 ous. I our ecoomy, rms (etrepreeurs) are edowed with techologies of di eret e ciecy, with the more e ciet oes requirig less capital. Etrepreeurs have limited wealth ad seek acig i competitive capital markets. I the product market there is free-etry i that all etrepreeurs that obtai acig ca eter. 4 Thus, the degree of competitio is edogeous, ad is determied oly by the ability of etrepreeurs to ace their rms. The ability of etrepreeurs to d acig is limited by the presece of agecy costs i both the debt ad the equity markets. We model the agecy cost of equity as i Stulz (1990, 005), ad assume that a rm s isiders may trasform some of the cash- ow to equity (that is the rm s free cash ow, et of paymets to creditors) as private bee ts. The private use of the rm s resources is ie ciet, makig outside equity costly to the etrepreeur. We model the agecy cost of debt as a risk-shiftig problem (see Jese ad Mecklig, 1976). As it is typical i the presece of moral hazard i the debt markets, rms must maitai a certai miimum level of equity to mitigate the moral hazard problem, geeratig debt capacity. We show that corporate goverace cocers i the equity market iteract i a essetial way with the moral hazard problem i the debt market, ad joitly determie a ecoomy s idustrial ad acial structure. Whe rm isiders have a greater ability to appropriate corporate resources (that is, whe the agecy costs of equity are more severe) debt becomes more desirable, sice it reduces the eed of outside equity acig. The ability to issue debt, however, is limited by the moral hazard problem i the corporate debt market. Thus, the simultaeous presece of the agecy costs of debt ad equity determies the overall ability of rms to raise capital, ad limits etry ito potetially pro table idustries. Our model determies edogeously a ecoomy s idustry cocetratio ad the acial structure of rms as a fuctio of ecoomywide factors, such as the overall quality of the corporate goverace system, ad sector-speci c factors, such as a idustry s exposure to moral hazard. We show that ecoomies characterized by worse corporate goverace are characterized by greater idustry cocetratio (i the spirit of Raja ad Zigales, 003). 5 I tur, greater 4 The fact that acig is a major barrier to etry is re ected, e.g., i the OECD World Competitiveess Report , which lists availability of acig as oe of the most importat barriers a ectig busiess competitiveess i several coutries. 5 Thus, the causality betwee the quality of a ecoomy s corporate goverace ad its degree of competitio may ideed ru i the opposite way to the oe suggested i traditioal theory (see, for example, Alchia,

5 idustry cocetratio leads to greater idustry pro ts, higher debt to equity ratios, more isider owership, ad higher returs o assets. These results are a direct cosequece of edogeeity of idustry cocetratio i our model: bad corporate goverace reduces rms ability to raise capital, which limits etry, icreases idustry pro ts ad debt capacity, leadig to greater leverage ad isider owership. Thus, by edogeizig idustry cocetratio, our model establishes a ovel lik betwee the quality of the corporate goverace system, owership structure, idustry cocetratio, ad leverage. These predictios help to explai the stylized facts that emerge from cross coutries studies such as La Porta, Lopez-de-Silaes, Shleifer ad Vishy (1997) ad (1998), Stulz (005), amog others. We ext show that corporate goverace impacts also rms choice of techology. This happes because bad corporate goverace pealizes "equity itesive" techologies, i.e. idustries where risk substitutio is possible, such as the high-tech sectors. This meas that i coutries with poor corporate goverace low-quality techologies may crowd out, i equilibrium, highquality techologies that are more exposed to the moral hazard problem. Thus, coutries with bad corporate goverace systems may be trapped i a equilibrium i which their idustries are domiated by less pro table ad less e ciet rms. We exted our results i several directios. First, we itroduce competitive baks that, at a cost, ca reduce the extet of the moral hazard problem. I this way, etrepreeurs ca obtai fuds also i cases where they would ot be able to raise capital from idividual ivestors. We d that more e ciet rms use direct acig, while margial (less e ciet) rms borrow from baks. Secod, we examie the bee ts of covertible debt (ad similar istrumets produced by acial iovatio) as tools to cotrol moral hazard (as suggested i Gree, 1984) ad, thus, potetially facilitatig etry. Surprisigly, we d that the agecy costs of equity iteract with the moral hazard problem i a way that covertible debt may i fact icrease, rather tha decrease, the isiders icetives to take risks, thus with o e ect o etry. This happes because isiders ca bee t from coversio of covertible debt, sice coversio (by elimiatig debt) icreases the cash ow to equity ad allows more fud diversio, iducig more risk takig 1950, ad Stigler, 1958): poor corporate goverace ad ivestor protectio may i fact lead to high idustry cocetratio. 3

6 rather tha discouragig it. Third, we examie the icetives to improve the quality of the goverace system both at rm level ad for the overall ecoomy. We show that rms i idustries more exposed to moral hazard ivest more to improve their corporate goverace, geeratig a egative correlatio betwee the quality a rm s goverace system ad its leverage. We the ivestigate etrepreeurs preferece for good goverace ad, thus, their icetives to lobby for good goverace legislatio i their ecoomy. We show that the quality of the corporate goverace system has a ambiguous impact o etrepreeurs welfare. O the oe had, etrepreeurs bee t from good goverace because it reduces the cost of raisig exteral equity; o the other had, isiders are hurt by good goverace, because it facilitates etry exposig them to more competitio. We show that etrepreeurs may have a preferece for extreme corporate goverace regimes, that is for regimes with either a very high or a very low corporate goverace quality. This observatio suggests that etrepreeurs i ecoomies characterized by bad corporate goverace have little icetive to lobby for a improvemet of their corporate goverace system. It also suggests that coutries would segmet themselves ito two groups, oe with high-quality corporate goverace systems, ad oe with low-quality systems, with little trasitio from oe group to the other. Fially, we cosider the e ect of acial market liberalizatios o ecoomic growth ad rms techology choices ad prefereces for goverace. We show that acial market liberalizatios, facilitatig equity acig, iduce more etry ad the adoptio of the more productive high-quality (equity-itesive) techologies, promotig ecoomic growth. I additio, we show that liberalizatios a ect prefereces for good goverace ad ca make etrepreeurs more likely to bee t from improvemets i the corporate goverace system of their ecoomies. Our paper rests at the itersectio of several strads of literature. The rst oe is the rapidly emergig literature o corporate goverace ad its e ect o the real sector. 6 By explicitly edogeizig the market structure of a idustry, our paper shows that corporate goverace ad capital structure cosideratios iteract i a essetial way to determie the 6 For excellet surveys of the literature, see Shleifer ad Vishy (1997), La Porta, Lopez-de-Silaes, Shleifer ad Vishy (000), ad Becht, Bolto, ad Roell (00). 4

7 competitive coditios i the idustry, i the spirit of Raja ad Zigales (003). Closely related is also Stulz (005), which argues that the agecy cost of equity limits a rm s ability to raise capital ad, therefore, to take advatage of the bee ts of globalizatio, ad Joh ad Kedia (003), who discuss the costs ad bee ts of alterative corporate goverace systems. Our paper is also related to the growth ad ace literature (see, for example, Raja ad Zigales, 1998, ad Levie, 1997, for a comprehesive survey) i that better corporate goverace ca icrease a ecoomy s growth by facilitatig rms capital raisig ad the adoptio of superior techologies. Thus, our paper provides a ew chael through which acial liberalizatios a ect the real sector of a ecoomy (see for example, Bekaert, Harvey, ad Ludblatt, 005, ad 009, for empirical evidece o the e ect of acial liberalizatios o ecoomic growth ad productivity). The secod strad of literature is the oe o the iteractio betwee acial ad market structure (see e.g., Brader ad Lewis, 1986, ad Maksimovic, 1988, amog others). These papers show that a rms s acial structure ca be used strategically to iduce a more aggressive behavior i the output market. I our paper, we rely o a di eret, o-strategic coectio betwee market structure ad rms capital structure. I this sese, our paper is close to Maksimovic ad Zecher (1991) ad Williams (1995), which focus o the e ects of agecy costs o itra-idustry variatio of techology choice ad capital structure. 7 The third strad of literature is the oe o idustrial orgaizatio ad the determiats of market structure (see, for example, Vives, 1999, amog may others). Moreover, our paper exteds i a (geeral) market equilibrium settig earlier literature that examies the impact of capital market imperfectios o product market competitio (see, for example, Poitevi, 1989, Bolto ad Scharfstei, 1990, ad Suomie, 004). Our paper is orgaized as follows. I sectio, we preset our basic model. I sectio 3 we preset the mai results of the paper. I sectio 4, we discuss our model s predictios. I sectio 5, we study the e ect of corporate goverace o the choice of techology. I sectio 6, we examie the role of the bakig sector ad the role of acial iovatio. I sectio 7, we examie etrepreeurs prefereces for good goverace. I sectio 8, we study the impact of acial liberalizatios. Sectio 9 cocludes the paper. All proofs are i Appedix II. 7 See also Riorda (003) for a discussio of this literature. 5

8 . The basic model We examie a ecoomy edowed with three types of agets: etrepreeurs, cosumers ad a large umber of small ivestors. Etrepreeurs, with o iitial wealth, are edowed with productio techologies (described below). Productio requires ivestmet of capital, which etrepreeurs obtai from ivestors. Ivestors are edowed with oe uit of cash each. Cosumers purchase the goods produced by the etrepreeurs, ad are characterized by their demad fuctios (described below). All agets are risk eutral. Etrepreeurs, idexed by i, are distributed cotiuously over the real lie, i [0; 1), ad have access to two di eret productio techologies. Techologies, idexed by fh; Lg, di er by their productio costs ad produce goods that ca be of either superior or iferior quality. Goods of superior quality are valued more by customers ad ca be sold at a greater price. The high-quality techology, = H, produces always superior quality goods, but at a greater cost. The low-quality techology, = L, produces superior quality goods oly with probability, while with probability 1 it produces goods of iferior quality. Productio is subject to moral hazard i that a etrepreeur s choice of techology is uobservable to both ivestors ad customers. The total cost of producig q i uits of output with techology by etrepreeur i is C ;i (q) = F ;i + cq i ; (.1) where c is the (costat) margial cost ad F ;i the xed cost, with F H;i > F L;i 0. Thus, the high-quality techology has greater xed cost. 8 I additio, etrepreeurs di er by the e ciecy of their techologies. We assume that more e ciet etrepreeurs have techologies with lower xed costs: F ;i = F + i, where is a measure of the e ciecy di ereces amog techologies. Thus, etrepreeurs with lower i are more e ciet. If a rm has produced superior quality goods, it ca sell its products to cosumers i the output market, where the demad for its output, x i ; is x i = p i + ep; (.) 8 We ca iterpret the greater xed cost of high quality techologies as the additioal R&D expeditures required to produce goods with superior features, ad thus of superior quality. 6

9 where is a positive costat that re ects the size of the market, is the total umber of rms i the idustry who produce superior quality goods, p i is rm i s price, ad ep the average price of the superior quality goods i the market. 9 superior quality goods, we have that ~p 1 This meas that if the most e ciet rms produce R 0 p jdj. As customary i the case of moopolistic competitio, we assume that rms are small ad therefore treat as a cotiuous variable (but we will still refer to as idicatig the umber of rms). Note that the demad schedule (.) is similar to that i moopolistic competitio, where a rm takes the other rms prices as give ad acts as a moopolist o the residual demad curve. 10 We assume that, if the rm s products are of iferior quality, cosumers are willig to pay oly the margial cost c for the goods, obligig the rm to set p = c. This implies that oly rms that produce superior quality goods ca recover their xed costs. Furthermore, throughout the paper we assume that < F L =F H, which implies that the high-quality techology is more e ciet tha the low-quality oe. Thus, the parameter characterizes the severity of the moral hazard problem: a greater value of makes it more likely that a rm usig the low-quality techology produces superior quality goods, icreasig its icetive to select such techology. Sice the value of the parameter depeds o the techologies availability to rms, which are presumably similar to all rms i the same idustry, we iterpret as represetig the exposure of a particular idustry to moral hazard. We will iitially assume that is su cietly small (or F L su cietly large) that the low-quality techology is ot sustaiable (i.e., pro table) i equilibrium: Assumptio A1: c (where c is de ed i the Proof of Propositio 1, Appedix II). This assumptio guaratees that all etrepreeurs choose i equilibrium the high-quality techology. The case i which also the low-quality techology is pro table (ad thus sustaiable) to some rms is examied i Sectio 5.1. Etrepreeurs obtai capital by issuig securities to ivestors. For simplicity, we restrict 9 Note that the demad fuctio (.) implies that, whe p i = ep (which will hold i a symmetric equilibrium) total idustry demad, ad thus output, is a costat ad equal to. I Sectio 8, we will examie the case i which total idustry demad is a decreasig fuctio of the average price ep eve whe p i = ep. 10 See, for example, Fujita et al. (1999) ad Ottavio et al. (00). Our demad fuctio is also similar to that i Salop (1979), with the di erece that i his circular city model, ep i is the average price of the two rms located closest to i. 7

10 the space of feasible cotracts by assumig that rms ca issue oly debt ad ew equity. 11 I particular, rm i seeks to raise F H;i by sellig to ivestors a fractio i [0; 1] of its shares, valued at S i ( i ), ad zero coupo debt with a face value B i ad a market value D i. 1 Fiacial markets operate competitively, ad all agets have access to a safe storage techology that o ers zero retur. Outside ivestors are atomistic. After issuig equity, etrepreeurs maitai cotrol of their rms, which they maage i their ow iterest. Etrepreeurial cotrol of rms geerates a co ict with outside shareholders who are exposed to (partial) wealth expropriatio from the etrepreeur, who is the rm s isider. We abstract from other sources of disagreemet betwee outside shareholders ad isiders-maagers (such as those due to di ereces i risk-aversio, as i Joh, Litov, ad Yeug, 007). I the spirit of Jese (1986) ad Shleifer ad Wolfeso (00) we model this agecy cost of equity by assumig that etrepreeurs may divert to themselves a fractio of the residual cash ow of their rms, after debt is repaid. 13 Thus, we ca iterpret the parameter as measurig the level of cotractibility of the rm s cash- ow to equity ad, i this way, represetig the extet of the agecy cost of equity. 14 We assume that diversio of rm s cash ow is ie ciet, ad a uit of diverted cash ow is worth oly < 1 to the etrepreeur (as i Pagao ad Roell, 1998, Stulz, 005, ad Almeida ad Wolfezo, 006). We iterpret the parameters ad as characterizig the quality of the corporate goverace system ad the level of ivestor protectio of the ecoomy i that they determie how e cietly etrepreeurs ca divert their rms cash ow ito private 11 We rule out the possibility of addressig the moral hazard problem by the use of optimal cotracts. While we make this assumptio for aalytical tractability, our mai results would hold as log as the moral hazard problem geerates a idustry-speci c debt capacity, eve after accoutig for optimal cotractig. This assumptio is relaxed i Sectio 6., where we allow rms to address the risk-shiftig problem by issuig covertible debt. 1 Sice, whe A1 holds, the low-quality techology is ot sustaiable, etrepreeurs i equilibrium raise F H;i = S i + D i uits of cash from ivestors to cover their xed costs for the high-quality techology, F H;i. 13 This implies that debt is a hard claim i the spirit of Hart ad Moore (1995) ad (1998). This meas that creditors rights are su cietly strog to iduce etrepreeurs to use all the available cash- ow to repay their creditors before egagig i ay cash- ow diversio. Our results will cotiue to hold as log as creditor rights are stroger tha shareholder rights. 14 This meas that, eve if rm cash- ow is potetially observable by ivestors, oly a fractio 1 is cotractible (see Hart ad Moore, 1995, ad Aghio ad Tirole, 199, where ivestors observe the state of the world, but have limited cotractibility). 8

11 bee ts. 15 For expositioal simplicity, we assume that the xed cost F H is su cietly large that, i equilibrium, etrepreeurs equity retetio is such that 1 i < for all i. This assumptio implies that all etrepreeurs have a icetive to divert the fractio of the cash ow to equity. 16 Assumptio A: c (where c is de ed i the Proof of Propositio 1, Appedix II). The timig of evets is as follows. At t = 0, etrepreeurs arrive to the capital market, aouce the target amouts of fuds that they wish to raise by issuig equity ad debt with value S i ad D i, respectively, i order to raise from ivestors the amout F H;i = D i + S i. Ivestors make acig o ers to the etrepreeurs. The capital market closes whe 0 rms have foud acig, the ivestors expect to break eve, o etrepreeur wishes to chage its proposed acial structure ad o additioal rm ca raise su ciet acig to eter. At t = 1, all 0 etrepreeurs that have bee successful i raisig F H;i of capital, i [0; ], select their productio techology, fh; Lg, ad productio takes place. At t =, etrepreeurs pay back or default o their loas. Etrepreeurs divert to themselves a fractio of the cash- ow that is left after leders have bee repaid. The residual fractio 1 is distributed to shareholders. Ivestors ad etrepreeurs cosume their wealth. A equilibrium i our model is characterized by the umber of etrepreeurs eterig the market,, ad their optimal strategies, fp i ; i ; S i ; D i ; i ; B i g, for i [0; ], such that (a) the strategy of each etrepreeur maximizes his payo give the strategies of the other players, (b) the goods markets clear, q i = x i, 8i, ad (c) the rms capital structure ad the umber of etrepreeurs eterig the market are such that o additioal etry ca occur with etrats earig o-egative pro ts. 3. Equilibrium We solve the model by backward iductio. I period t = 1; etrepreeurs that have bee successful i raisig F H;i uits of cash, choose their pricig strategy depedig o whether they 15 Evidece of large bee ts of cotrol ad associated deadweight costs ca be foud i Albuquerque ad Schroth (008). 16 If 1 i <, some iframargial etrepreeurs (the more e ciet oes) would ot, i equilibrium, divert resources for themselves. Allowig for this possibility would ot a ect our mai results, however, because the properties of our equilibria deped oly o the behavior of the margial etrepreeurs, for which i ' 1. 9

12 have produced goods of superior or iferior quality. Takig as give the prices of the other rms producig superior quality goods, fp j g j6=i, a etrepreeur with superior quality goods faces a residual demad curve (.) ad maximizes his rm s total cash ow, Xi T, by selectig p i arg max p i X T i = (p i c) p i + ep : (3.1) If, istead, the etrepreeur has produced iferior quality goods, he has o choice other tha settig a price p i = c, at which it ca sell a xed quatity, x. The total cash ow accruig to a rm depeds o whether it has produced goods of superior or iferior quality, ad therefore, o the choice of techology. Give the etrepreeurs optimal pricig strategy, p, the total cash ow geerated by rm i, Xi T, is give by 8 < Xi T (p (p i c) p i ; i ) = + ~p + I i (F H F L ) with pr: 1 I i (1 ) : I i (F H F L ) with pr: I i (1 ) ; where I i is a idicator fuctio that takes the value of oe if i = L, ad zero otherwise. Firm i s cash ow is divided betwee its creditors, X D i (p ; i ), outside shareholders, X S i (p ; i ), ad the etrepreeur, X E i (p ; i ), as follows (3.) X D i (p ; i ) mifb i ; X T i (p ; i )g; (3.3) X S i (p ; i ) i (1 ) maxfx T i (p ; i ) B i ; 0g; (3.4) X E i (p ; i ) [ + (1 i )(1 )] maxfx T i (p ; i ) B i ; 0g: (3.5) Proceedig backward, at the begiig of period t = 1, after havig obtaied acig, etrepreeurs choose their techology by maximizig their ow expected payo, selectig i (B i ) arg max E 1Xi E (p ; i ); (3.6) ifh;lg where E t represets the expectatio at t o future cash ows. As it will become apparet below, the optimal choice of techology depeds o the face value of the outstadig debt, B i. The optimal acial structure is determied by etrepreeur i at t = 0 by maximizig max V i = E 0 Xi E (p ; i (B i )) (3.7) S i;d i; i;b i 10

13 subject to S i E 0 i (1 ) maxfx T i (p ; i (B i )) B i ; 0g; (3.8) D i E 0 mifb i ; X T i (p ; i (B i ))g; (3.9) S i + D i = F H;i ; (3.10) where (3.8) ad (3.9) are, respectively, the shareholders ad debt holders participatio costraits, (3.10) is the etrepreeur s acig costrait. Propositio 1 (Equilibrium): The umber of etrepreeurs that eter the market i equilibrium,, ad their choice of acig, fsi ; D i g i=0 is determied as follows: 1) I equilibrium, the rst > 0 etrepreeurs eter the market, where is implicitly determied by = p FH + + ; (3.11) where (F H F L ) 1. All i etrepreeurs choose the high-quality techology, ad produce output, qi, sold at a price, p i, give by q i = ; p i = c + : (3.1) ) Etrepreeurs ace the xed costs, F H;i, by raisig a amout of equity ad debt equal to S i = F H + i D i = (1 ) ( i); (3.13) D i = D > 0; (3.14) ad issue a fractio i = 1 ( i) (1 ) (3.15) of their shares to outside ivestors. I equilibrium, the payo to etrepreeur i [0; ], V i, is V i = + ( i): (3.16) Etry i the product market is determied by the iteractio of imperfectios i both the debt ad the equity market, captured by the parameters ad, as follows. 17 Abset capital 17 Note that is strictly icreasig i, ad hece also measures a idustry s exposure to the moral hazard problem. 11

14 market imperfectios, that is, whe = = 0, etrepreeurs ca raise i the capital markets all the fuds ecessary to ace pro table projects. I this case, etry will occur util the rets, eared i equilibrium i the product market, which from (3.1) are give by are equal to the xed costs of the margial etrat. This meas that, abset capital market imperfectios, the equilibrium umber of etrepreeurs that eter the market, c, is determied by coditio that the margial etrepreeurs ear zero (expected) pro ts, that is, by c F H c = 0: (3.17) We refer to c as the perfectly competitive outcome. From (3.11), it is easy to see c > wheever > 0. The presece of imperfectios i the capital markets reduces etry because it limits the ability of etrepreeurs to raise capital o both the equity ad the debt markets. Raisig fuds by issuig equity is costly because the etrepreeur appropriates a fractio of the residual cash ow, after the repaymet of debt, ad he ejoys oly a fractio per dollar of diverted cash ow, while the remaider 1 is dissipated. Sice ivestors ratioally aticipate the cash ow diversios, etrepreeurs ultimately bear the cost of this ie ciecy, makig outside equity expesive for the etrepreeurs. This dead-weight loss represets the agecy cost of equity. The presece of the agecy costs of equity makes etrepreeurs to prefer to raise as much capital as possible i the debt market. The amout of fuds that a etrepreeur ca raise i the debt market, however, is limited by the moral hazard problem. By choosig low-quality techology (rather tha the high-quality oe) etrepreeurs save the amout F H F L i xed costs but, with probability, evertheless obtai superior quality goods. Therefore the lowquality techology is riskier tha the high-quality oe, exposig creditors to a risk shiftig problem. 18 Sice, give A1, the low-quality techology is ot sustaiable, etrepreeurs ca i equilibrium select a capital structure whereby they have a icetive to choose the high-quality techology. Thus, the etrepreeur ca oly issue a amout of debt with face value B i satis es the icetive-compatibility coditio Bi Bi + F H F L : (3.18) 18 Our results will hold also i the case that the high quality techology produces low quality goods with some small but positive probability. that 1

15 This implies that D i = B i D ; (3.19) where D represets the rm s debt capacity. Note that (de ed i Propositio 1) represets the miimum equity value that a rm must maitai to esure that the high-quality techology is optimally chose, ad it depeds o the severity of the moral hazard problem. Debt capacity D is idustry speci c ad depeds both o the extet of the moral hazard problem ad o the level of idustry cocetratio,. Greater exposure to moral hazard icreases the miimum equity that a rm must maitai to iduce isiders to choose the highquality techology, reducig debt capacity. Coversely, greater idustry cocetratio raises a rm s ecoomic pro ts, icreasig its value ad, thus, debt capacity. 19 I equilibrium, etrepreeurs issue debt up to debt capacity, D, ad the sell equity to outside ivestors util i = 1, for the last etrat (i.e., the margial etrepreeur). Give that represets the miimum equity that all rms must maitai to satisfy the icetivecompatibility coditio (3.18), ad that the etrepreeur appropriates a fractio of it, the amout of equity that the margial etrepreeur,, issues is S = (1 ). Thus, the margial etrepreeur that ca obtai acig,, is determied by D + S = = F H; = F H + : (3.0) This coditio requires that, for the margial etrepreeur, the total value of the rm s cash ow,, after the diversio to the etrepreeur,, is equal to its xed costs, FH;. Iframargial etrepreeurs issue to outside shareholders oly the amout of equity that is strictly ecessary to raise F H;i, leadig to (3.13). Sice rms equity has a market value E M (1 ), the fractio of equity sold by etrepreeur i is S i =EM, givig (3.15). I equilibrium, the margial etrepreeur ears a ecoomic pro t which is equal to the value of the cash ow diversios,. Iframargial etrepreeurs bee t from their greater e ciecy by issuig less equity, ad 19 Note that i our stylized model debt capacity is the same for all rms i the same idustry sice, from the icetive compatibility coditios, the potetial gai from deviatig to low quality techology, F H F L, is idepedet of i. This assumptio ca be easily relaxed by assumig, for example, that more e ciet rms have also lower variable costs, which would lead to greater debt capacity. Our mai results will hold as log as there is a systematic di erece i debt capacity across idustries which is drive by heterogeeity i productio techologies. 13

16 thus by earig, i equilibrium, greater ecoomic pro t, give by (3.16). Fially, from (3.14), it is easy to see that, abset moral hazard (that is, with = 0), all rms would be etirely debt aced ad etry would occur util = c. Similarly, abset the agecy cost of equity (that is, with = 0) all rms would have costless access to equity ad agai, from (3.0), etry would occur util = c. It is precisely the iteractio of the imperfectios i both the equity ad debt markets, i.e. whe > 0, that limits the ability of etrepreeurs to raise capital, reducig the equilibrium umber of rms that ca eter a ew market. 4. Corporate Goverace, Fiace, ad Idustry Cocetratio Our model shows that idustry cocetratio ad rm acial ad owership structures are joitly determied by the iteractio of the quality of the corporate goverace system of a ecoomy (measured by ) ad idustry characteristics (that is, the exposure to the moral hazard problem, measured by ). I this sectio we develop predictios o the cross sectioal variatio that would be observed across idustries withi a ecoomy (that is, i the same legal jurisdictio), ad across di eret coutries with heterogeous legal jurisdictios. Propositio (Corporate goverace, idustry cocetratio ad acial structure): Ecoomies with worse corporate goverace regimes are characterized by greater idustry cocetratio, greater debt level, lower book ad market value of equity ad, for the more e ciet etrepreeurs, by greater isider owership (de ed by! i 1 i ): < 0 ; D > 0; S i < 0; E M i < 0;! i > 0 i i < i c(; ); (4.1) where i c (; ) is de ed i the Appedix. Furthermore, de ig the elasticity of etry to corpo- rate goverace as "( ; j) =, we have that "( ; j) > 0: (4.) Propositio shows that the quality of corporate goverace ad ivestor protectio a ect several critical features of the idustrial ad acial structure of a ecoomy. First, ecoomies characterized by worse corporate goverace (higher ) have greater idustry cocetratio 14

17 (lower ). This happes because worse corporate goverace regimes limit etrepreeurs ability to raise equity from capital markets, which impairs etry of ew rms ad, thus, icreases idustry cocetratio. I additio, from (4.), the elasticity of the umber of rm eterig a idustry i equilibrium,, is icreasig i that idustry s exposure to the moral hazard problem,. This meas that the e ect of the quality of the corporate goverace system o etry is more proouced precisely i those sectors where equity acig is more critical. Secod, iterestigly, low-quality corporate goverace regimes lead to greater debt capacity. This property is a direct cosequece of the edogeeity of idustry cocetratio: A worse corporate goverace regime, reducig etry, leads to greater idustry cocetratio ad, therefore, to greater rms pro ts i equilibrium. I tur, greater pro ts relax the icetive compatibility costrait, (3.18), ad icrease debt capacity. Third, worse corporate goverace, icreasig isiders cash ow diversios, reduces the cash ow that ca be pledged to outside ivestors ad, thus, leads to lower book ad the market values of equity (give by S i ad Ei M, respectively). The e ect of the quality of corporate goverace o isider owership,! i, depeds o a rm s positio withi a idustry. Less e ciet rms (greater i) rely relatively more o equity acig. Worse corporate goverace implies that these rms must sell a relatively greater fractio of equity to outsiders, decreasig isider owership. Coversely, more e ciet rms, i < i c, sell less equity ad, thus, rely relatively more o debt acig. This meas that the icrease i debt capacity that comes with a worse corporate goverace regime (as discussed above) allows these rms to issue relatively less equity to outside ivestors, icreasig isider owership. Propositio 3 (Moral hazard, idustry cocetratio, ad acial structure): Sectors exposed to more severe agecy costs of debt are characterized by greater idustry cocetratio, lower corporate debt level, greater book ad market value of equity, ad less isider owership: < 0; D < 0; S i > 0; E M i > 0;! i < 0: (4.3) Idustries exposed to a more severe moral hazard problem (greater ) are characterized by greater cocetratio. This happes because greater exposure to moral hazard reduces a rm s debt capacity. Firms, however, ca oly partially o set the reductio i debt acig with a correspodig icrease i equity. This happes because a reductio of a dollar i cash 15

18 ow paid out to creditors results oly i 1 fractio of the rm s cash ow is diverted by the etrepreeur). dollars of added equity capacity (sice a Therefore, worse moral hazard impairs rms overall ability to raise fuds, leadig to less etry ad greater idustry cocetratio. Furthermore, etrepreeurs i equilibrium substitute debt acig with equity acig, leadig to greater book ad market value of equity ad less isider owership. Propositios 1-3 geerate predictios o the cross sectioal variatio that would be observed withi a coutry (that is, withi the same legal jurisdictio), ad across coutries (that is, i legal jurisdictios that have potetially di eret corporate goverace ad ivestor protectio regimes). 0 We cosider the e ect of the three parameters fi; ; g o several key ratios determied edogeously i the model. First, withi a idustry, for each idividual rm i [0; ] we cosider the debt-to-equity ratio, Di =S i ; the book-to-market ratio of equity S i =EM i ; the degree of isider owership,! i = 1 i ; ad the retur o assets: ROA i = X i =F H;i. Secod, we compare these same key across idustries ad legal jurisdictios. 1 Tables 1-a ad 1-b below preset the sig of the partial derivatives of the ratios with respect to the relevat parameters. By cotrastig tables 1-a ad 1-b, it is easy to see that the correlatio betwee leverage ad rm pro tability withi a ecoomy ca di er whe measured withi the same idustry or across idustries. I our model, rms i the same sector di er oly by the e ciecy of their techology, while rms i di eret sectors of a ecoomy di er also by the severity of the moral hazard problem ad, therefore, by their debt capacity. Withi a give sector, more e ciet rms require less capital ad eed to issue less equity tha more ie ciet oes. Thus, more e ciet rms, have greater retur o assets ad issue relatively less equity, geeratig a positive relatioship betwee leverage ad pro tability. 3 Iterestigly, this result is cosistet with the 0 Note that i our model, rms heterogeeity origiates from three di eret sources. First, withi a give idustry, rms di er by their level of e ciecy i, with more e ciet rms eedig less capital. Secod, across idustries i the same ecoomy, di eret sectors have di eret exposure to the moral hazard problem, ad thus di eret values of : Third, across coutries, di eret ecoomies are characterized by di eret quality of their corporate goverace system, ad therefore have di eret values of. 1 For tractability, we cosider the aggregate ratios for the idustry, rather tha the averages of the ratios for all rms i the idustry. The proofs are omitted, but they are available from the authors upo request. 3 Note also that this result is robust to alterative speci catios of the source of heterogeeity withi a idustry. For example, rms i the same idustry may require the same xed assets, F H, but may di er by the margial costs, c i = c + i. Eve i this case, more pro table rms would have greater debt capacity, geeratig 16

19 dig i Mackay ad Phillips (005) that, withi idustries, ew etrats (correspodig to our margial rms) have less leverage ad are less pro table tha icumbet rms. The relatioship betwee pro tability ad leverage is reversed whe we compare across sectors. Firms i sectors more exposed to moral hazard have lower debt capacity ad leverage. I additio, these idustries are more cocetrated ad, therefore, are associated with greater pro ts ad retur o assets. Thus, greater exposure to moral hazard leads at the same time to less levered, more pro table rms ad to greater idustry cocetratio, geeratig a egative relatioship betwee leverage ad pro tability, ad betwee leverage ad idustry cocetratio. Table 1-a: Withi idustry cross-sectioal variatios D i S i S i E M i i +! i ROA i 1-b: Cross sectioal variatio across idustries ad legal jurisdictios D S id: id: S (! ) id (ROA ) id E M A plus (egative) sig idicates a positive (egative) partial derivative of the ratio or variable with respect to i, or ; respectively. Parameter i represets rm e ciecy, with a greater i correspodig to a less e ciet rm; parameter represets a techology s exposure to moral hazard, with a greater correspodig to greater moral hazard; parameter represets the quality of a coutry s corporate goverace framework, with a greater correspodig to a lower level of ivestor protectio ad corporate goverace quality. The egative correlatio betwee leverage ad pro tability is a direct cosequece of the edogeeity of idustry cocetratio of our model. This implies that a static trade-o model of the determiatio of a rm s capital structure (such as the oe discussed here) ca geerate a egative correlatio betwee leverage ad pro tability. This result depeds crucially o the agai a positive correlatio betwee leverage ad pro tability. 17

20 egative correlatio betwee debt capacity ad pro tability that is geerated by edogeous etry. 4 Across ecoomies, our model predicts that ecoomies characterized by worse corporate goverace systems (that is, by higher ) are also characterized by greater idustry cocetratio, higher debt to equity ratios (whe equity is measured either at book or market value), greater isiders owership, ad greater returs o assets. These results are agai the direct cosequece of the edogeeity of idustry cocetratio ad debt capacity: worse corporate goverace regimes reduce a rm s ability to raise capital, which limits etry ad, i tur, leads to greater debt capacity (ad, leverage) ad greater isider s owership. Thus, by edogeizig idustry cocetratio, our model establishes a ovel lik betwee the quality of the corporate goverace system, owership structure, idustry cocetratio, ad leverage. These results are cosistet with several the stylized facts that emerge from cross coutries studies. For example, La Porta, Lopez-de-Silaes, Shleifer ad Vishy (1997 ad 1998) d that coutries with worse corporate goverace have more debt relative to equity acig, lower market values of rms (compared to GDP), ad larger owership by isiders. More recetly, Stulz (005) ds that coutries with worse corporate goverace are characterized by a smaller fractio of widely held rms ad, thus, greater isiders owership. Demirguc-Kut ad Maksimovic (1998) ad Hail ad Leuz (006) d that coutries edowed with a better legal eviromet are characterized by a lower retur o capital. Klapper, Laeve ad Raja (004) documets the bee cial e ect of regulatio, that is aimed at a better developmet of acial markets, o the etry of ew rms, especially i idustries with high R&D itesity or idustries that have greater capital eeds. 5 A further implicatio of our paper is that the quality of the corporate goverace system of a ecoomy has a idepedet impact o the acial structure choices of rms, beyod 4 I this way, our model helps explaiig the apparet puzzle give by the egative relatioship betwee profitability ad leverage that is documeted i several empirical studies such as Titma ad Wessels (1988), Raja ad Zigales (1995), Fama ad Frech (00), Demirguc-Kut ad Maksimovic (1988), ad Booth, Aivazia, Demirguc-Kut, ad Maksimovic (001) amog others. 5 I a similar vei, Fa, Titma ad Twite (003) documets a egative correlatio betwee leverage ad the stregth of a coutry s legal system. The paper also shows that the presece of high-quality auditors (as measured by the market share of the Big- ve accoutig rms) is egatively related to leverage, especially i developig coutries. 18

21 rm-speci c characteristics. Thus our model provides a explaatio for the digs of Booth, Aivazia, Demirguc-Kut, ad Maksimovic (001), which shows that coutry speci c factors (such as its legal framework) are as importat as rm-speci c factors i determiig a rm s capital structure decisio. 5. Goverace ad Techology Choice The quality of the corporate goverace system a ects also rms choice of techology ad thus, through this secod chael, the idustrial structure of a ecoomy. We ivestigate this possibility i this sectio by cosiderig the parameter regio where Assumptio A1 does ot hold, so that the low-quality techology is potetially pro table. We maitai the assumptio that the high-quality techology is more e ciet that the low-quality oe. Propositio 4 (Corporate goverace ad techology choice): Let ( c ; F L =F H ) ad < (where > 0 is de ed Appedix II). I equilibrium 0 ad: > etrepreeurs eter the market i) the rst 00 (0; 0 ) of these choose the high-quality techology, ad raise D of debt ad F H;i D of equity; ii) the remaiig 0 00 > 0 etrepreeurs choose the low-quality techology ad ace their xed costs etirely with debt by borrowig D i (icreasig) fuctio of : = F L + 0 : Here 00 ( 0 ) is a decreasig I equilibrium, both low-quality ad high-quality techology coexist whe <. Etrepreeurs that choose the high-quality techology, i 00, issue rst debt up to debt capacity, ad the issue all the equity ecessary to cover the xed costs, F H;i. Their umber, 00, is determied by the coditio that the margial etrepreeur is able to obtai acig, that is 00 + ( 0 00 (F H + 00 ) 0; (5.1) ) ad by the coditio that he prefers to raise F H; 00, ad select the high-quality techology, rather tha to raise F L; 00 ad select the low-quality techology, that is (1 ) 00 + ( 0 00 (F H F L ) (1 ) 0: (5.) ) 19

22 Etrepreeurs icetives to choose the high-quality techology rather tha the low-quality oe ca be see by examiig the three terms i (5.). The rst term re ects the fact that the high-quality techology produces superior quality goods with certaity, while the low-quality techology produces superior quality goods oly with probability. The secod term represets the di erece i the xed costs of the two techologies, F H F L. The third term represets a goverace cost, ad is due to the fact that the high-quality techology ca be adopted i equilibrium oly if the etrepreeur is aced by equity i the amout of (so that the icetive compatibility coditio is satis ed), while the low-quality techology ca be aced etirely by debt. Sice equity acig geerates a e ciecy loss, the adoptio of the high-quality techology is costly to the etrepreeur ad leads to a loss of value equal to (1 ). Etrepreeurs that choose the low-quality techology, that is i ( 00 ; 0 ], ca ace their xed cost F L;i etirely by debt. This happes because their ivestors are ot exposed to moral hazard, ad the etrepreeurs optimally choose debt to avoid the dissipative cost of equity. The umber of etrepreeurs that eter the market with the low-quality techology is determied by the coditio that the margial etrepreeur is just able to raise the xed cost F L; 0. The presece of the low-quality techology limits the ability of etrepreeurs to adopt the high-quality oe. From the acig costrait (5.1) it is easy to see that, all else equal, a icrease of the umber of low-quality rms that eter the market, that is a larger 0, has the e ect of reducig the umber of etrepreeurs with high-quality techology that ca coexist i equilibrium, 00. Coversely, a decrease of the umber of high-quality rms that eter the market, that is a smaller 00, has the e ect of icreasig the umber of etrepreeurs with low-quality techology that ca be sustaied i equilibrium, 0. Thus, easier access to the capital markets that facilitates etry by etrepreeurs adoptig the low-quality techology (for example, by improvemets i credit markets, o which the low-quality techology is relatively more depedet) displaces, i equilibrium, the high-quality techology. A additioal implicatio of Propositio 4 is that the umber of rms that choose the highquality techology is lower i ecoomies where the quality of the corporate goverace system is of worse quality. This happes because a icrease i makes the icetive costrait (5.) ad the acig costrait (5.1) tighter, leadig to a lower 00. Similarly, sectors more exposed to the moral hazard problem, that is, with a greater, are characterized by a smaller umber of 0

23 rms with high-quality techology. Iterestigly, whe the quality of the corporate goverace system is su cietly low, it is quite possible that either (5.1) or (5.) is ot satis ed for ay i 0. This implies that the high-quality techology caot be sustaied i equilibrium; we refer to this pheomeo as oe of goverace crowdig out. Propositio 5 (Goverace crowdig out): The high-quality techology caot be sustaied i equilibrium, that is, 00 = 0, i a idustry with moral-hazard, if > 0 (where 0 is de ed i the Appedix II). Furthermore, lim 0 = F L F H!0 : These observatios imply that the quality of a coutry s corporate goverace system has a impact o the choices of techology made by rms operatig i its jurisdictio ad thus o the idustrial structure of its ecoomy. 6 I particular, our model suggests that coutries with a low-quality of corporate goverace system may ot be able to sustai more e ciet rms i capital itesive idustries that are more exposed to moral hazard, such as, for example, the high-techology ad pharmaceutical sectors. Thus, these coutries will be at a competitive disadvatage i developig such more advaced sectors. These are ew ad testable predictios Goverace ad the Structure of Fiacial Systems 6.1. Goverace ad Bak Fiacig Baks ca reduce the agecy costs of debt by moitorig rms ad thus mitigatig the etrepreeur s icetives to take excessive risks (see, for example, Diamod, 1991, amog others). Assume ow that the ecoomy is edowed also by competitive baks ad that, by icurrig a xed moitorig cost, c b, a bak ca decrease the extet of etrepreeurial moral hazard. The bee t of bak acig is to lower the miimum equity that a rm must maitai from to, say,, thus reducig the agecy costs of equity ad icreasig debt capacity. The moitorig cost is charged up frot to the etrepreeur whe he borrows from the bak, icreasig the cost 6 Note that Propositio 5 implies that whe the e ciecy di ereces betwee techologies become small,! 0, ad whe! F L =F H ; 0 approaches zero. I this case, the high-quality techology is ever chose i equilibrium 7 I Appedix I, we provide evidece cosistet with our model s predictio. Speci cally, i Table A1, part c, we show that ecoomies with higher ivestor protectio are also characterized by a better developed hightechology sector. 1