Eren Inciz

08 November 2009 Preliminary draft: Please do not cite or quote without the permission of the authors.

Abstract Control of a …rm entitles entrepreneurs to enjoy nonmonetary returns that cannot be contracted to banks. A typical venture capital …nancing contract, however, allocates considerable control of the …rm to the venture capitalist, which allows the venture capitalist to gain access to nonmonetary returns. We consider a model in which entrepreneurs di¤er in their concern about nonmonetary returns and the venture capitalist give certain importance to them while banks do not care about them at all. The model exhibits many empirical facts. In one equilibrium, banks prefer not to …nance startup …rms. In another equilibrium, entrepreneurs who care less about nonmonetary returns raise funds from a venture capitalist while the rest choose bank-…nancing. Thus, in equilibrium, bank-…nanced entrepreneurs divert more resources in a …rm to tasks that yield nonmonetary returns while venture-capital-backed …rms have higher internal rate of return. Keywords: bank, control, entrepreneurship, nonmonetary return, private bene…t, venture capital JEL Classi…cation: G21, G24, G32, L26, M13 y

The authors would like to thank Thomas Hellmann for helpful comments. Tel.: 90-216-483-9284; fax : 90-216-483-9250. Address: Sabanci University - FASS, Orhanli / Tuzla 34956 Istanbul

TURKEY. E-mail address: [email protected] z Tel.: 90-216-483-9340; fax : 90-216-483-9250. Address: Sabanci University - FASS, Orhanli / Tuzla 34956 Istanbul TURKEY. E-mail address: [email protected]

1

1

Introduction

Control of a …rm entitles entrepreneurs to enjoy nonmonetary returns that cannot be contracted to banks. A typical venture capital …nancing contract, however, allocates considerable control of the …rm to the venture capitalist, which in turn opens the gates for nonmonetary returns for venture capitalist as well. As equity providers, venture capitalists usually have seats in the board of directors. They have rights to use …rm property and actively involve in the management. They participate in forming the organizational structure and establishing the strategies of the …rm. They help …nding customers, business consultants, lawyers, suppliers, and even further …nancing. The contracts also give them the right to involve in employing or …ring key managers and other personnel. Some contracts may even give them the rights to dismiss the founding entrepreneur from the …rm and replace him with an outside manager, which they occasionally do.1 If control of his …rm allows an entrepreneur to enjoy nonmonetary returns, they should also accrue to the venture capitalist to the extent of her control in the …rm. Typical banks, on the other hand, are passive debt providers who do not care about nonmonetary returns.2 This paper incorporates this di¤erence between banks and venture capitalists in a simple model of start-up …nancing when entrepreneurs have the option to raise funds from one or the other. In the standard incomplete …nancial contracting models or in models of capital structure and control (such as Aghion and Bolton, 1992, Grossman and Hart, 1988, Harris and Raviv, 1988, and Holmstrom and Tirole, 1989), control of a …rm allows entrepreneurs to enjoy nonmonetary returns but the outside …nanciers are assumed to be passive actors who care only about the monetary returns. This approach is a good approximation when entrepreneurs …nance themselves from one type of outside …nancier, say only banks or only venture capitalists, which is often the case in these models. This is because entrepreneurs probably care more about nonmonetary returns than any outside …nancier and thus one may reasonably assume that …nanciers do not care about nonmonetary returns at all. However, when it comes to modeling the entrepreneur’s choice between two di¤erent …nancing options that coexist in an economy and if one of them cares more about the nonmonetary 1

See Gompers and Lerner (2000), Gorman and Sahlman (1989), Hellmann and Puri (2002), Kaplan and Stromberg

(2003), Lerner (1995), Sahlman (1990) for evidence on all of these. 2 See Bottazzi, Da Rin, and Hellmann (2008) on the degree of activism by investors, who …nd strong evidence that venture capitalist’s role in the portfolio …rm goes beyond the simple provision of …nance.

2

returns than the other, the weight one gives to the nonmonetary returns becomes important. In the start-up …nancing market, for instance, banks and venture capitalists (and possibly other …nanciers such as angels) coexist. What kinds of entrepreneurs …nance from venture capitalists and what kinds of them from banks? There is certain facts that lead us to think that venture capitalists care about nonmonetary returns. Venture capitalists care about not only the current deal with the portfolio company but also the e¤ect of this deal on their reputation in fund-raising from investors and in attracting further start-up projects that are promising (Gompers, 1996). Many of them, especially corporate venture capitalists, have multiple goals. Their pro…ts are de…nitely a major goal but they may also have strategic goals. For example, Intel wants to promote technologies that use computing power; university-VCs care about academic prestige of technological advancements from their schools; government-VCs care about innovation and employment (Brander, Egan, and Hellmann, 2009), bank-VCs care about future loan clients (Hellmann, Lindsey, and Puri, 2009). There is also strong evidence that venture capitalist use the information they have about their portfolio …rms to help them engage in strategic alliances with other …rms in their portfolio (Lindsey, 2008). They may even care about having seats in the board of directors of many …rms which they may view as prestigious positions for their career or as a source of individual power which may yield private bene…ts. Gorman and Sahlman (1989) report that, on average, venture capitalists spend about half of their time in monitoring nine portfolio companies and sit in board of directors of …ve of them. They also …nd that a typical venture capitalist visit portfolio companies about 19 times a year and spend 100 hours at each …rm annually, and their most frequently performed services are helping the …rm in raising additional funding, strategic analysis, and management recruiting. Moreover, Hellmann and Puri (2002) …nd that venture-capital-backed …rms experience higher degrees of professionalization than …rms raise funds from other sources, implying the active involvement of the venture capitalist in structuring and managing the …rm. As well connected individuals in speci…c industries and controlling board members in the portfolio …rm, they may even in‡uence decisions so that the …rm purchase services and inputs, employ managers and other employees, from friends or relatives. It is also well possible that a venture capitalist prescribes certain investment strategies to portfolio companies to hedge the risk in her own portfolio and these strategies are not necessarily the (monetary) value maximizing strategies for an individual …rm. 3

To understand the e¤ects of the di¤erence between “hands-on” contracting with venture capitalists and “hands-o¤” contracting with banks on the distribution of entrepreneurs between these two sources, we consider penniless entrepreneurs who are in a position to choose from which source to raise funds to …nance their start-up projects. A project yields both monetary and nonmonetary returns and entrepreneurs di¤er in their weight they give to nonmonetary returns of the project. Nonmonetary returns are private and noncontractible to other parties. The venture capitalist also gives certain importance to nonmonetary returns while banks are concerned only with monetary returns. We assume that both types of …nanciers are unaware of the entrepreneurs’weight on the nonmonetary returns whereas the venture capitalist’s weight is common knowledge. Loan contracts with a bank take the simple debt form under asymmetric information. When the entrepreneur chooses bank-…nancing, he hires a manager and operates his project as the sole principal. Contracting with the manager involves moral hazard as his e¤ort is not observable. Depending on his weight on the nonmonetary returns, the entrepreneur o¤ers an employment contract to the manager such that the manager optimally allocates his e¤ort between two tasks: the task that yields the monetary returns and the task that yields the nonmonetary returns, both of which use up resources. Contracting with the venture capitalist is more involved. We do not …nd the venture capital contracting process a simple …nancial agreement between the entrepreneur and the venture capitalist but an implicit bargaining process over the preferences of two contractees on monetary and nonmonetary returns of the …rm, which means that the venture capitalist e¤ectively becomes a co-principal in the project. In venture capital contracting, …rst, the entrepreneur decides how much …xed compensation and what level of ownership share to o¤er to the venture capitalist. If the venture capitalist gets any positive ownership share, she e¤ectively becomes a co-principal in the project with bargaining power (in the subsequent decisions) given by her ownership share. Thus, the venture capitalist becomes one of the controlling parties in the …rm and she may have di¤erent agenda and preferences, which is captured in the model by the weight she gives to nonmonetary returns. Then, to come to an agreement and align their preferences in the …rm, co-principals bargain over what weight to give to nonmonetary returns in designing the optimal employment contract to o¤er to the manager. If the bargaining falls apart, the project will still have some value to the venture capitalist. Thus, the entrepreneur must provide a payo¤ that is high enough to pay o¤ at least the sum of this value and the capital investment by the venture capitalist. 4

We show that the contract between the entrepreneur and the venture capitalist may take three forms. Consider …rst the cases in which the entrepreneur gives more weight to nonmonetary returns than the venture capitalist. If, in addition, the project yields su¢ cient monetary returns to pay o¤ the venture capitalist, the contract takes the simple debt form in which the entrepreneur surrender no ownership of the …rm, but if not, the contract takes the equity form in which the venture capital gets some ownership share. Thus, equity provision by venture capitalists is attributable to ex post wealth constraints of entrepreneurs. When an entrepreneur gives some ownership share to the venture capitalist, he does so voluntarily, and in the model this automatically gives the venture capitalist some control in the …rm to the extent of her ownership share in the …rm (or even higher in practice).3 Thus, in some cases, to be able to raise the necessary funds, entrepreneurs may have to relinquish control to the venture capitalists, as in Hellmann (1998). Consider now the cases in which the entrepreneur gives less weight to nonmonetary returns than the venture capitalist. In these cases, we interpret the venture capitalist as another business …rm since she attaches more weight to nonmonetary returns than the …rm. Here, regardless of the level of the monetary returns of the project, it is optimal for the entrepreneur to sell the …rm to the venture capitalist. This is more like an existing company’s acquisition of a new start-up. We …nd that there will be no bank-…nancing when the project does not have a positive value to the venture capitalist if the bargaining with the entrepreneur falls apart. This is in line with Ueda’s (2004) …nding that perfect intellectual property protection makes venture capital …nancing optimal. We also characterize one equilibrium when the project still has some positive value to the venture capitalist if the bargaining fails. In this equilibrium, entrepreneurs whose weight on the nonmonetary returns is less than a threshold raise funds from a venture capitalist while the rest choose bank-…nancing. Hence, in equilibrium, bank-…nanced entrepreneurs divert more resources in a …rm to tasks that yield nonmonetary returns while venture-capital-backed …rms have higher internal rate of return. This result shows the possibility of the conjecture Hellmann (1998, p. 71) puts forward: “[O]nly those [entrepreneurs] willing to yield control rights choose venture capitalists, while the others seek …nancing with private investors or other more passive sources of funds.” There are many papers focusing on the contracting between entrepreneurs and venture capitalists 3

In many circumstances, venture capitalists’control power is much higher than the size of their ownership share.

This would make our results even more stronger.

5

and even more on the contracting between entrepreneurs and banks. However, we know of only four papers which get to the grips with modeling the entrepreneurs choice between raising funds from a bank or a venture capital: de Bettignies and Brander (2007), Landier (2003), Ueda (2004), Winton and Yerramilli (2008). No paper we know of focuses on the importance of nonmonetary returns on behalf of the venture capitalist or outside …nanciers in general. In Ueda (2004) bank-…nancing takes place in the presence of incomplete information on behalf of the bank, which therefore asks for collateral to screen. Venture capitalists, on the other hand, have better ability to evaluate projects and thus contracting with the venture capitalist is not subject to asymmetric information. But, the venture capitalist is able to undertake the project by herself if the negotiation between the parties breaks apart. This expropriation possibility leads to the result that tighter intellectual property protection makes venture capital …nancing attractive. Entrepreneurs with little collateral also …nances from venture capitalists. If, for example, there is perfect intellectual property protection or if entrepreneurs do not have su¢ cient collateral to provide, there will be no bank-…nancing. She also …nds that entrepreneurs who raise funds from venture capitalists have higher returns. Many of our results are in line with Ueda’s results, which are themselves in line with empirical facts. In an attempt to explain the di¤erences in forms of start-up …nance across sectors, regions, or countries, Landier (2003) models the choice between bank-…nancing and venture capital …nancing. In his model, failed entrepreneurs are stigmatized and the degree of stigma can be di¤erent in di¤erent sectors, regions, or countries. A high degree of stigma worsens the outside options of entrepreneurs. In such a regime, entrepreneurs choose risky projects and in this high risk environment, venture capitalists …nance start-ups since they can closely monitor them. In a low stigma regime, however, the outside options of entrepreneurs are better which lead them to choose safe projects. Consequently, in this safer regime, banks …nance start-ups with debt contracts which requires little monitoring. To explain why banks use debt contracts with little or no monitoring whereas venture capitalists prefer convertibles with strong monitoring and exercise of control, Winton and Yerramilli (2008) also models the choice between bank and venture capital …nancing. They incorporate the di¤erences in the risk and returns of …rms’cash ‡ows to explain the relative use of venture capital and bank loans. They …nd that entrepreneurs with higher chances of good outcomes will use less informationally intensive methods of …nance such as bank loans instead of more informationally intensive methods such as venture capital …nancing. Venture capital …nancing is attractive only 6

when the entrepreneur’s returns are highly risky and skewed, with good outcomes being unlikely. Finally, de Bettignies and Brander (2007) combined the entrepreneur’s …nancing choice problem with the double moral hazard problem between the entrepreneur and the venture capitalist. Because the entrepreneur and the venture capitalist jointly provide costly e¤ort in the …rm but they do not fully bene…t from the return on e¤ort as they only own a share of the …rm, one important issue in venture capital contracting is the presence of double-sided incentive problem.4 In de Bettignies and Brander (2007), bank-…nancing involves debt …nancing which does not distort entrepreneur’s incentives to provide e¤ort in his …rm that he wholly owns. However, if he raises funds from a venture capitalist, his incentives deteriorate due to the fact that he surrenders an ownership share of his …rm. In exchange, he gets venture capitalist’s managerial input. They …nd that venture capital …nancing is superior only when the entrepreneur highly regards the venture capitalist’s managerial input or when his own e¤ort is not too important in the …rm. The paper is organized as follows. Section 2 outlines the model. Section 3 examines a bank-based …nancial system as a benchmark. Section 4 describes the details of contracting with the venture capitalist. Section 5 analyzes the entrepreneurs’choice between bank-…nancing and venture capital …nancing and Section 6 concludes.

2

The Model

We consider a unit mass of risk-neutral and penniless entrepreneurs (indexed by E). Each entrepreneur is endowed with a start-up project that requires K units of capital and a manager. Because the entrepreneurs are penniless, they have to raise funds from outside …nanciers. There are two di¤erent sources of …nance in the economy. On the one side, risk-neutral banks provide loans in a perfectly competitive market in exchange for nonnegative returns. If the entrepreneur gets bank…nancing, he shall pay back (1 + r)K at the end of the period where r

0 denotes the rate of return

that the bank demands for, which is endogenously determined. On the other side, a risk-neutral

4

Cassamatta (2003), for example, points to double moral hazard in explaining why venture capitalists are sources

of both managerial advice and …nance rather than specializing only in …nancing while managerial advice is provided by consultants independently. Inderst and Muller (2004) explores the double-sided moral hazard problem in a search model and explain short- and long-run dynamics of the venture capital industry.

7

venture capitalist (VC hereafter, indexed by V C) provides equity-like …nance which requires a …xed compensation R, a transfer of an ownership share 1

of the start-up (where

remaining to the entrepreneur), and certain control rights. Both R and

2 [0; 1] is the share

are also endogenously

determined. The speci…cation of the start-ups’production technology is based on the standard multiple-task moral hazard model of Holmstrom and Milgrom (1991). There are two tasks in our model, and if undertaken, the project yields two-dimensional, state-contingent, observable and veri…able returns drawn from a normal distribution whose variance-covariance matrix is assumed to be …xed. The …rst dimension of the returns is monetary and the second dimension is nonmonetary. Nonmonetary returns are nontransferable and noncontractible, and thus they accrue only to the principal(s) of the project to the extent of their ownership share in the …rm. In case of bank-…nancing, the entrepreneur is the sole principal and thus the sole bene…ciary of the nonmonetary returns since contracting with a bank is a simple borrower-lender relationship where one party gets the loan in the beginning of the period from the other party and pays it back at the end of the period with the required interest rate speci…ed in the terms of contract. However, in the case of VC-…nancing, the VC becomes a co-principal in the project since contracting with the VC attributes certain control rights to her which often includes allocating a seat for her in the board of directors, rights to dismiss or employ managerial team and employees, or bring in suppliers ... etc. Therefore, we assume that, in the case of VC-…nancing, both the entrepreneur and the VC enjoy the nonmonetary returns of the project. We allow entrepreneurs and the VC to give di¤erent importance to nonmonetary returns of the project. Let coe¢ cient

i,

i = fE; V Cg, be the weight that principal i assigns to nonmonetary

returns. We shall assume that entrepreneurs are heterogeneous in terms of this coe¢ cient. In particular,

E

is uniformly distributed over the interval 0;

further assume that

E

with pdf f ( ) and cdf F ( ). We shall

is privately known by the entrepreneur whereas

and is in the interval 0;

VC

is common knowledge

.

Figure 1 shows the sequence of events. At the beginning of the period, the entrepreneur privately learns his type (coe¢ cient) and decides whether to raise funds from a bank (denoted by B in the …gure) or a VC. We assume limited liability in both kinds of …nancial contracting. If the entrepreneur chooses bank-…nancing, he signs a contract indicating that he must pay (1 + r)K to the bank at the end of the period if he is solvent, and if not, that he has to pay whatever left. If he chooses VC8

Figure 1: The sequence of events …nancing, he o¤ers the pair ( ; R) and certain control rights to the VC in exchange for VC’s supply of K units of capital. The VC may accept or reject the o¤er. If she rejects the o¤er, the game ends. Otherwise, the game continues and they bargain over the allocation of returns of the project with bargaining powers given by their ownership shares of the …rm. This determines a level of

?

that

they agree on. If they do not get to an agreement, they get their disagreement payo¤s: zero for the entrepreneur and C

0 for the VC the entrepreneur (this is a standard assumption in the literature

that appeared, for example, in Hellmann, 1998 and Ueda, 2004). After the bargaining process, they o¤er an employment contract to the manager (in the case of bank-…nancing entrepreneur does this alone). The means of the two-dimensional returns are determined by the manager’s e¤ort choice, which neither the entrepreneur nor the VC can observe or verify. Hence, any contract o¤ered cannot depend on manager’s e¤ort choice. The manager decides whether to accept or reject the o¤er, and in the case he accepts, he simultaneously decides the e¤ort level he would like to provide. Finally, the publicly observable and veri…able state is realized. The wage payment to the manager (which is neither observable nor veri…able by the bank) is made, and the contractual liabilities to the …nanciers are satis…ed by the entrepreneur. What we have below is devoted to deriving four important expressions that are frequently used in the rest of the paper: the value of the project for the entrepreneur and for the VC when the entrepreneur raises funds from a VC (eqs. (8) and (9), respectively), the value of the project for the entrepreneur when he raises funds from a bank (eq. (11)), and the monetary value of the project (eq. (12)). Obtaining these pro…t expressions requires describing the production technology in detail and obtaining the optimal employment contract with the manager under moral hazard. Consider the manager’s problem …rst. The manager has two tasks to complete, task 1 yields the monetary returns and task 2 yields the nonmonetary returns. He is in a position to choose a vector 9

of e¤orts t = (t1 ; t2 ) that speci…es the e¤ort he would like to provide on each task. The private cost of providing e¤ort is given by T which is continuous and strictly convex function of t1 and t2 . We shall further particularize this cost function by assuming the following quadratic form: k1 t21 k2 t22 + ; 2 2

T (t1 ; t2 ) =

(1)

where k1 and k2 are strictly positive parameters. Given the manager’s e¤ort choice on each task, the returns are distributed with a two-dimensional normal distribution with mean We assume that

where

1 (t1 )

takes the following linear form: 0 1 0 1 (t1 ) [email protected] (t) = @ (t ) 2 2

is the monetary return,

2 (t2 )

1 t1 2 t2

:

(a) (Debt-…nancing) If

2 E

1

K + C, then

2

?

and the entrepreneur’s return is (b) (Equity-…nancing) If

2 E

1

2

= 1 and R? = K + C; 1

+

2 E

K

2

< K + C, then

(31)

C > 0. ?

is the maximum real number in [0; 1]

solving (1

? 2

)

2 VC

( ? )2

2 E

2

=K +C

R? = and the entrepreneur’s return is 2 2. Suppose (a) If

E

1

=

VC

2

2

?

E

(

?

E

+ (1

?

( ? )2

1

?

)

(32)

1

V C)

2

;

(33)

2 ):

(34)

2

> 0.

= . K + C, then ?

2 [0; 1] and R? = K + C 20

(1

?

)(

1

+

2

?

In particular, both K+C

(

return is (b) If

2

1 ?

1

2

+

2

2

+

1

= 1 and R? = K + C (debt-…nancing); and

K

C > 0.

< K + C, then

2 f 2 [0; 1] : (1 2 ? ) ?

Note that,

2

1g

K+C

2

= 0 and R? = K + C

solutions. The entrepreneur’s return is 3. (Acquisition) Suppose 2 VC

E

V C.

0.

The entrepreneur sells the project for a price of

1

+

C to the venture capitalist, i.e.

K

2

= 0 and R? =

< 0 (acquisition) are among the solutions. The entrepreneur’s

2)

2

?

?

= 0 and R? =

The entrepreneur’s return is

1

2 VC

+

(

1

2 VC

+

K

2

K

2

(36)

C) > 0:

C.

Proof. When Assumption 1 holds, the constraint set is non-empty and compact. Thus, there exists a solution due to the continuity of the objective function given in (30). We should consider three di¤erent cases

V C,

>

E

E

=

V C,

and

E

V C,

E

2

1

>

2

1

2

2

K + C) and the other

< K + C). Consider each case in turn.

V C.

~ E ( ; R) is strictly increasing and ~ V C ( ; R) is strictly decreasing in . This

implies that the participation constraint of the VC holds with equality at the solution ( ? ; R? ). Thus, ignoring the participation constraint of the entrepreneur for now, we can write down (30) as max

E

( ;R)

(

E

+ (1

)

V C)

+ (1

)

VC

(

E

+ (1

)

V C)

(K + C):

(37)

By Lemma 1, this objective function is continuous in . Moreover, the entrepreneur is solving a non-trivial utilitarian planner’s problem where the weights assigned to the principals must be interpreted as their shares of the project. Let

1

2 E

2

K +C, i.e. the monetary returns of the project when the entrepreneur operates

it is greater than or equal to the start-up cost plus the VC’s disagreement payo¤. Because that the 21

entrepreneur is solving a planner’s problem given in (37) and 0 by

E

V C,

>

E( E)

V C( V C)

2 2( E

=

2 V C)

>

the optimal solution is such that the entrepreneur is the sole owner of the project.

In expected terms, he pays o¤ the start-up cost of the …rm borrowed from the VC, K, and his disagreement payo¤, C, in full out of the monetary returns of the project given by

2 E

1

2.

This

is the unique solution of this case. Note that the participation constraint of the entrepreneur is satis…ed in equilibrium: 1

If

2 E

+

2 E

1

2

2

R? =

1

2 E

+

(K + C)

2

< K + C, then setting

2 E

+

1

(

2

2 E

1

2)

2 E

=2

2

> 0:

= 1 violates the participation constraint of the VC. As a

result, the entrepreneur tries to compensate the VC with monetary payments as much as possible in order not to distort the returns from the project (because considering the planner’s problem given in (37) reveals that the social optimum requires that the entrepreneur is allocated as much shares as possible). The remaining part is covered by ownership shares to the VC, which the entrepreneur tries to keep as minimal. Consequently, the entrepreneur gives all monetary returns he collects with his share

to the VC. Thus, for any , R =

( )2

1

2

. Consequently, the VC’s payo¤ at

should satisfy 1

( )2

+ (1

2

)(

1

+

(2

)

VC

2)

K + C:

Here, the …rst term on the left-hand side is the monetary returns that the entrepreneur earns from the project and the second term is the VC’s monetary and nonmonetary returns. Manipulating this yields ( )2

1

= =

1 1

+ (1

2

( )2 ( )2

2 2

)(

1

+

(2

+ (1

)

1

+ 2 (1

)

VC

)

VC

( )2 2

2

2)

+

K +C (2

VC

) + ( )2

K +C

At the optimum, it must be that the VC does not obtain any surplus. As a result, be the maximal real number (because recall that 2 (1

?

)

?

VC

2

K +C

2

E(

=K +C

?

2 [0; 1] must

; R) is strictly increasing in ) such that 1

( ? )2

2

:

Note that the right-hand side of this equation gives the amount of monetary returns that the entrepreneur is short on and the left-hand side gives the nonmonetary returns of the VC. Manipulating 22

this expression yields ( ? )2

1 ?

?

Finally, substituting

?

(2(1 ?

=

E

?

+ 2 (1

2

)

?

VC

?

+ (1

)

)

?

) 2

VC

(K + C) = 0

2

=K +C

1:

?

gives the following conclusion:

VC

2 [0; 1] is the

maximal real number such that ? 2

(1

( ? )2

2 VC

)

2 E

2

=K +C

1;

which …nishes the proof of case 1. Case 2. Suppose

E

=

VC

= .

Because that the two principals’interests are perfectly aligned, this case features many solutions. When

1

2

K + C, the monetary returns from the project are su¢ cient to pay o¤

2

K and C. Given

2 [0; 1], the monetary payment that should be made to the VC (for whom

the participation constraint holds with equality) is R = K + C

(1

)(

1

2

+

2 ).

the monetary payments need to be made by the entrepreneur to cover R calls for ( K+C

(1

)(

1

2

+

2 ),

If R > 0, 2

1

2)

the left-hand side giving us the monetary returns allocated to the

entrepreneur at

and the right-hand side the monetary payment needed to be made to the VC.

This implies (1

2 )

2

1.

K+C

2

Hence, every

2 [0; 1] is a solution, because then

in all of them the entrepreneur obtains the highest returns which is due to ( (

1

+

2

2)

(K + C) + (1

)(

1

+

2

2)

=

1

+

2

2

1

+

2

2)

R =

0, where the …nal

(K + C)

inequality is due to Assumption 1. Furthermore, the situation in which the monetary payments to the entrepreneur from the project is not su¢ cient to pay o¤ the VC is not possible. This is because we need to have (1 2 ) for any value of When

1

2

2

1,

0 to 13

Even though a formulation with two-sided asymmetric information is very compelling, getting tangible results

from such a model proves to be very di¢ cult.

26

establish the coexistence of bank- and VC-…nancing in equilibrium. Our goal in this case is more modest. We characterize only a monotone equilibrium in which banks play threshold strategies.

5.1

Case 1: C = 0

This section focuses on the case in which both the entrepreneur and the VC obtain zero utility when they do not get to an agreement in the bargaining stage (i.e., C = 0). This sets a benchmark for the entrepreneurs’choice between bank-…nancing and VC-…nancing when the VC obtains a nonzero utility in case of disagreement (i.e., C > 0). This is an idealistic case of perfect intellectual property rights. Proposition 5 Suppose Assumption 1 holds and consider the following conditions. ( 2 =3)

i.

1

ii.

E

iii.

1

2 E

2

< K,

VC

iv.

1

2 E

2

K,

VC

2

K,

2

=3, E,

(1 + r)K. If

2

…nances via a VC, we know from Proposition 3 that his payo¤ is

2 VC

1

VC

Now assume

2 E

1

2

2 E

1

E

and if he

K as he sells his

2

project to the VC. This payo¤ is better than his payo¤ with bank-…nancing. If all shares and thus his payo¤ is

P~ . Consider

2

VC

E,

(2

1

2 V C)

2,

then VC-…nancing is optimal; if

…nancing and bank-…nancing are equally well; and if is optimal. However,

2 E

1

1

K contradicts with

2

K < (2 1

K = (2

1

K

2 E

2 E

2 V C)

(2

2 E

2 E

2, 2 V C)

2.

Consequently,

2 V C)

2,

then VC-

then bank-…nancing 2

when

E.

VC

Thus, bank-…nancing cannot be optimal here. Next, consider, 2 E

VC

E.

and if

1

K

> (1 + r? )K if s? ( ) = B and > < 2 ? P( ) = 1 2 if s ( ) = B and > > > : 0 otherwise 30

1 1

2 2

2

(1 + r? )K

2

< (1 + r? )K :

(38)

We can now de…ne an equilibrium in this economy as follows: De…nition 1 Given parameters 1. for all

V C;

;

1;

2 ; K; C

, an equilibrium is (r? ; s? ) such that

2 [0; ], s? ( ) is an optimal choice of the entrepreneur with coe¢ cient ;

2. whenever s? ( ) = V C, the net expected return of the VC is C; 3. banks make zero pro…t under limited liability at the interest rate r? , so that Z (P ( ) K) dF ( ) = 0;

(39)

f :s? ( )=Bg

where s? ( ) = B is banks’belief. 4. banks’belief is consistent with the equilibrium. We de…ne a monotone equilibrium of this economy to be one in which there exists that all entrepreneurs with

E

r

prefer VC-…nancing while those with

r

>

E

r

such

prefer bank-

…nancing. That is, banks play threshold strategies so that a bank expects to have customers with E

>

r

in its loan applicant pool and sets its lending interest rate, r? , accordingly.

In addition to Assumption 1, which guarantees that some projects yield su¢ ciently high monetary returns to pay o¤ the VC, we shall make the following assumption so that there are entrepreneurs whose projects does not yield su¢ cient monetary returns to pay o¤ the VC. Assumption 2 The technology is such that the monetary returns of the project satis…es

2

1

2

> (1 + r? )K if 2 ( r ; ] > < 2 P( ) = 2( ; ] : 1 2 if > > > : 0 r if

(43)

As an intermediate step in the proof, the following lemma identi…es the conditions that assures

that entrepreneurs with

r

E

prefer VC-…nancing while entrepreneurs with

E

>

r

prefer

r

prefer

bank-…nancing. Lemma 4 Suppose Assumptions 1 and 2 hold. Given bank-…nancing whereas entrepreneurs with

E

r

r

, and

VC

r

, and

VC

If

E

>

r

, and

VC

2 VC

2 E)

E,

then, (

r

to

raise funds from a bank. Claim 1 Every entrepreneur with r

Due to our hypothesis,

>

>

E

V C,

r

…nds bank-…nancing optimal.

and because

E

conditions (44) and (49) of Lemma 4 are relevant. Let Note that, due to our hypothesis, for all

E

2 (

r

; ] and we have

entrepreneur with

E

2(

r

2

1 1

2

>

r

, it must be that

2(

r

; ] be such that

= K +r? K, and we have 2 E

2

E

< K + r? K for all

1 E

2 E

2

> 1

V C. 2

Then, only 2

= K + C.

> K +r? K = K +C

2 ( ; ]. Hence, for any

; ], the relevant condition of Lemma 4 is (49), which is clearly satis…ed 36

when r? = C=K. Moreover, for any entrepreneur with

E

2 ( ; ], the condition we need to check

is (44), which is also satis…ed. Next we argue that it is optimal for all entrepreneurs with Claim 2 Every entrepreneur with

r

E

r

VC

implies that

C = r? K and

V C. E,

VC

2

2 E

= K + C, >

VC

and

Then, the relevant condition of Lemma 4 is

(46), which is satis…ed when r? = C=K, and as shown above E

1

> K + C. There are two cases to consider:

Consider …rst the case in which

the case in which

to raise funds from the VC.

…nds VC-…nancing optimal.

First, note that due to the assumption that E

r

E

2 E

1

2

> K + C. Next, consider

In this case, the relevant condition of Lemma 4 is (45). Because

it is immediate to see that (

2 VC

2 E)

2

r? K = 0. Hence, it is

C

also satis…ed. It is left to establish zero pro…t for banks in equilibrium. We simply need to …nd out the

r

value that solves (42) when we substitute in r? = C=K. After substituting for the values from the de…ning equation (43) of P , (42) becomes Z

?

(1 + r )KdF ( ) +

r

Z

1

2

2

dF ( )

K

Z

dF ( ) = 0:

r

Substituting in r? = C=K and incorporating the uniform distribution assumption yield ! ! ! r 3 3 + r (K + C) + K = 0; 1 2 3 and solving this for

r

gives r

5.3

3

(K + C) +

1

=

C

3

3

2

K :

Some start-up …nancing arithmetic

Let us now make some back-of-the-envelope calculations. Consider a unit investment; thus, set K = 1. We assume C = 0:687 which implies a lending interest rate of r? = C=K = 6:87%. This interest rate is in line with lending interest rate published in Federal Reserve’s statistical releases in 1990s. We assume that the coe¢ cient terms is distributed between zero and 4.2; that is, 37

= 4:2.

Dyck and Zingales (2004) estimate the average value of private bene…ts as 14%. We shall employ this as a rough measure of nonmonetary returns. So, if

E[ which yields

1

E[

E

1

+

2] E

2]

( =2) 2 1 + ( =2)

=

= 12:9. From (40), we …nd that

industry is active only in the 0:37=4:2

2

r

= 1, then = 2

1

can be found from

14 ; 100

= 0:37. This means that the venture capital

8:81% of the economy. Is this a reasonable number?

By using the National Survey of Small Business Finances (NSSBF), Berger and Udell (1998) …nd the estimated distributions of equity and debt in the small business …nance. In their estimations, venture capital’s share is 1.85% while the commercial banks’ share is 18.75%, which means the relative size of the venture capital industry is 8:98%, which is pretty close to our back-of-the-envelope calculation here. They are also consistent with de Bettignies and Brander (2007) who mentions that VCs’contribution to entrepreneurial …nance is much less than commercial banks’. Figure 2 shows the market’s split between VC-…nancing and bank-…nancing in our numerical example.

Figure 2: The structure of …nancing in the start-up market If there were only bank-…nancing, the equilibrium would be given by (21). Making the necessary calculations shows that the lending interest rate in such an economy would have a lower lending interest rate: r = 6:12%, as expected.

6

Conclusion

Unlike banks, venture capitalists are not passive investors. They are specialized …nancial intermediaries who provide managerial advice in addition to …nance. Most venture capital contracts allocate considerable control rights to the venture capitalist. Protective provision terms in contracts allow them to veto transactions that are unfavorable to them and board control gives them the ability to initiate new transactions.

38

The existence of nonmonetary private bene…ts on behalf of the entrepreneur is now a standard feature of venture capital …nancing models. The basic insight we tried to underline in this paper is that if control allows the entrepreneur to enjoy nonmonetary bene…ts, it should also allow the venture capitalist. Thus, venture capitalists must be equity …nanciers who value nonmonetary returns (at least more than banks). Under this conjecture, we focused on the entrepreneurs’decision between bank-…nancing and VC-…nancing and …nd an equilibrium in which those who value nonmonetary returns more …nance from banks while the rest go to VCs, implying the fact that those …nanced by VCs have higher internal rate of return. There are couple of important features of venture capital industry which we do not focus in this paper: intensive monitoring by the VC (Gompers, 1996); staging of funding (Bergemann and Hege, 1998, Cornelli and Yosha, 2003); usage of special …nancial instruments such as convertible securities (Marx, 1998, Repullo and Suarez, 2004, Schmidt, 2003); syndication of venture capital investments (Admati and P‡eiderer, 1994, Brander, Amit, and Antweiler, 2002, Lerner, 1994); and the exit route of VCs (Black and Gilson, 1998, Cumming, 2008, Cumming, Fleming, and Schwienbacher, 2006, Cumming and Johan 2008, Schwienbacher, 2008, 2009). A nice venue of future research is to interact all of these features with the entrepreneur’s choice between bank- and VC-…nancing to better understand the market split between the two.

References [1] Admati, A., and P. P‡eiderer. 1994. Robust …nancial contracting and the role of venture capitalists. Journal of Finance 49, 371-402. [2] Aghion, P., and P. Bolton. 1992. An incomplete contracts approach to …nancial contracting. Review of Economic Studies 59, 473-494. [3] Becker, R., and T. Hellmann. 2005. The genesis of venture capital: lessons from the German experience. in C. Keuschnigg and V. Kanniainen, ed., Venture Capital, Entrepreneurship, and Public Policy, Chapter 2 (MIT Press), 33-67. [4] Bergemann, D., and U. Hege. 1998. Venture capital …nancing, moral hazard, and learning. Journal of Banking and Finance 22, 703-735. 39

[5] Berger, A, and G. Udell. 1998. The economics of small business …nance: the roles of private equity and debt markets in the …nancial growth cycle. Journal of Banking and Finance 22, 613-673. [6] Berglof, E. 1994. A control theory of venture capital …nance. Journal of Law, Economics and Organization 10, 247-267. [7] Black, B., and R. Gilson. 1998. Venture capital and the structure of capital markets: banks versus stock markets. Journal of Financial Economics 47, 243-277. [8] Bottazzi, L., M. Da Rin, and T. Hellmann. 2008. Who are the active investors?: evidence from venture capital. Journal of Financial Economics 89, 488-512. [9] Brander, J., E. Egan, and T. Hellmann. 2008. Government sponsored versus private venture capital: Canadian evidence. in J. Lerner and A. Schoar, ed., NBER conference volume on “International Di¤erences in Entrepreneurship”. [10] Brander, J., R. Amit, and W. Antweiler. 2002. Venture capital syndication: improved venture selection versus value-added hypothesis. Journal of Economics and Management Strategy 11, 423-452. [11] Casamatta, C. 2003. Financing and advising: optimal …nancial contracts with venture capitalists. Journal of Finance 58, 2059-2086. [12] Chan, Y., D. Siegel, and A. Thakor. 1990. Learning, corporate control and performance requirements in venture capital contracts. International Economic Review 34, 365-381. [13] Cornelli, F., and O. Yosha. 2003. Staged …nancing and the role of convertible securities. Review of Economic Studies 70, 1-32. [14] Cumming, D. 2008. Contracts and exits in venture capital …nance. Review of Financial Studies 21, 1947-1982. [15] Cumming, D., and S. Johan. 2008. Preplanned exit strategies in venture capital. European Economic Review 52, 1209-1241.

40

[16] Cumming, D., G. Fleming, and A. Schwienbacher. 2006. Legality and venture capital exits. Journal of Corporate Finance 12, 214-245. [17] de Bettignies, J., and J. Brander. 2007. Financing entrepreneurship: bank …nance versus venture capital. Journal of Business Venturing 22, 808-832. [18] Dyck, A. and L. Zingales. 2004. Private bene…ts of control: an international comparison. Journal of Finance 59, 537-600. [19] Gompers, P. 1996. Grandstanding in the venture capital industry. Journal of Financial Economics 42, 133-156. [20] Gompers, P., and J. Lerner. 1998. Venture capital distributions: short-run and long-run reactions. Journal of Finance 53, 2161-2183. [21] Gompers, P., and J. Lerner. 1999. The Venture Capital Cycle (MIT Press, Cambridge). [22] Gompers, P., and J. Lerner. 2000. Money chasing deals?: the impact of fund in‡ows on the valuation of private equity investments. Journal of Financial Economics 55, 281-325. [23] Gorman, M., and W. Sahlman. 1989. What do venture capitalists do?. Journal of Business Venturing 4, 231-248. [24] Grossman, S., and O. Hart. 1988. One share-one vote and the market for corporate control. Journal of Financial Economics 20, 175-202. [25] Harris, M., and A. Raviv. 1988. Corporate governance: voting rights and majority rules. Journal of Financial Economics 20, 203-235. [26] Hellmann, T. 1998. The allocation of control rights in venture capital contracts. RAND Journal of Economics 29, 57-76. [27] Hellmann, T., and M. Puri. 2000. The interaction between product market and …nancing strategy: the role of venture capital. Review of Financial Studies 13, 959-984. [28] Hellmann, T., and M. Puri. 2002. Venture capital and the professionalization of start-up …rms: empirical evidence. Journal of Finance 57, 169-197. 41

[29] Hellmann, T., L. Lindsey, and M. Puri. 2008. Building relationships early: banks in venture capital. Review of Financial Studies 21, 513-541. [30] Holmstrom, B., and J. Tirole. 1989. The theory of the …rm. in R. Schmalensee and R. Willig, ed., Handbook of Industrial Economics, Chapter 2 in Part 1 (Elsevier Publishing B.V, Amsterdam). [31] Holmstrom, B., and P. Milgrom. 1991. Multitask principal-manager analyses: incentive contracts, asset ownership and job design. Journal of Law, Economics, and Organization 7, 24-52. [32] Inderst, R., and H. Muller. 2003. The e¤ect of capital market characteristics on the value of start-up …rms. Journal of Financial Economics 72, 319-356. [33] Kaplan, S., and P. Stromberg. 2001. Financial contracting theory meets the real world: an empirical analysis of venture capital contracts. Review of Economic Studies 70, 281-315. [34] Landier, A., 2003. Start-up …nancing: from banks to venture capital. Unpublished working paper, University of Chicago, Chicago, IL. [35] Lerner, J. 1994. The syndication of venture capital investments. Financial Management 23, 16-27. [36] Lerner, J. 1995. Venture capitalists and the oversight of private …rms. Journal of Finance 50, 301-18. [37] Lindsey, L. 2008. Blurring …rm boundaries: the role of venture capital in strategic alliances. Journal of Finance 63, 1137-1168. [38] Marx, L. 1998. E¢ cient venture capital …nancing combining debt and equity. Review of Economic Design 3, 371-387. [39] Repullo, R., and J. Suarez. 2004. Venture capital …nance: a security design approach. Review of Finance 8, 75-108. [40] Sahlman, W. 1990. The structure and governance of venture-capital organizations. Journal of Financial Economics 27, 473-521.

42

[41] Schmidt, K., 2003. Convertible securities and venture capital …nance. Journal of Finance 58, 1139-1165. [42] Schwienbacher, A. 2008. Innovation and venture capital exits. Economic Journal 118: 18881916. [43] Schwienbacher, A. 2009. Venture capital exits. in D. Cumming, ed., Companion to Venture Capital, Robert W. Kolb Companion to Finance Series (Wiley/Blackwell). [44] Thomson, W. 1981. Nash bargaining solution and utilitarian choice rules. Econometrica 49, 535-538. [45] Ueda, M., 2004. Banks versus venture capital: project evaluation, screening, and expropriation. Journal of Finance 59, 601-621. [46] Winton, A. and V. Yerramilli. 2008. Entrepreneurial …nance: banks versus venture capital. Journal of Financial Economics 88, 51-79.

43