Entrepreneurial finance: Banks versus venture capital

https://doi.org/10.1016/j.jfineco.2007.05.004Get rights and content

Abstract

We analyze how entrepreneurial firms choose between two funding institution: banks, which monitor less intensively and face liquidity demands from their own investors, and venture capitalists, who can monitor more intensively but face a higher cost of capital because of the liquidity constraints that they impose on their own investors. Because the firm's manager prefers continuing the firm over liquidating it and aggressive (risky) continuation strategies over conservative (safe) continuation strategies, the institution must monitor the firm and exercise some control over its decisions. Bank finance takes the form of debt, whereas venture capital finance often resembles convertible debt. Venture capital finance is optimal only when the aggressive continuation strategy is not too profitable, ex ante; the uncertainty associated with the risky continuation strategy (strategic uncertainty) is high; and the firm's cash flow distribution is highly risky and positively skewed, with low probability of success, low liquidation value, and high returns if successful. A decrease in venture capitalists’ cost of capital encourages firms to switch from safe strategies and bank finance to riskier strategies and venture capital finance, increasing the average risk of firms in the economy.

Introduction

Although start-ups and venture capital finance are often linked in the public eye, bank loans are a more common source of finance for entrepreneurial firms.1 Both sources share some common features. Because entrepreneurial firms are usually small and have high risk of failure, both venture capital and bank loans require careful monitoring of borrowers. Both types of finance use covenants to restrict the borrower's behavior and provide additional levers of control in the event that the firm performs poorly. These covenants often restrict the ability of the firm to seek financing elsewhere, which ties to yet another common feature: the use of capital rationing through staged financing and credit limits as means of controlling borrowers’ ability to continue and grow their business.

Despite these similarities, significant differences exist between these two types of financing. Whereas banks lend to a wide variety of firms, firms with venture capital finance tend to have very risky and positively skewed return distributions, with a high probability of weak or even negative returns and a small probability of extremely high returns (see Sahlman, 1990; Fenn, Liang, and Prowse, 1995). Whereas bank loans usually take the form of pure debt, venture capitalists almost always employ convertible securities or a combination of debt and equity (see Kaplan and Stromberg, 2001). Banks’ monitoring and control rights are typically far less intensive than those of venture capitalists and focus on avoiding or minimizing bad outcomes. Banks mostly monitor for covenant violations, deteriorating performance, or worsening collateral quality that might jeopardize their loan. They exercise control by threatening to force default and possible liquidation. By contrast, as shown by Sahlman (1990) and Kaplan and Stromberg (2001), venture capitalists often hold seats on the borrowing firm's board and voting rights far in excess of their cash flow rights, and they could have the contractual right to replace the entrepreneur with a new manager if covenants are violated. Along with these rights, venture capitalists monitor borrowers more frequently than banks do and play an active role in most of the firm's major decisions.2 Finally, the funding structures of the two types of institutions are very different. Banks offer their investors relatively liquid investments, which in turn subjects banks to possible liquidity shocks. Venture capital funds impose restrictions on their investors, insulating the funds from liquidity shocks but forcing them to pay investors a premium for lack of liquidity.

In this paper, we develop a simple theoretical model that captures these differences. An entrepreneurial firm seeks financing from one of two institutions, a bank and a venture capital fund. Once funded, two possible conflicts arise between the entrepreneur and the institution. The first is the well-known tension between the entrepreneur's desire to keep the firm going to maintain her control benefits and the institution's desire to liquidate poorly performing investments. The second conflict is more novel: Even if it is optimal to keep the entrepreneur's firm going, there could be additional choices to be made. Should the firm expand conservatively or aggressively? Should the firm attempt an initial public offering (IPO) or settle for sale to another corporation? Again, a tension exists between the entrepreneur and the institution, because the entrepreneur could excessively prefer aggressive or risky decisions that maintain or expand her control benefits even when such decisions do not always maximize contractable cash flows. We model this phenomenon as a choice between a safe and a risky continuation strategy if the firm is not liquidated early.

If the firm is to be financed, the institution must be given incentive to monitor the firm's situation to reduce these conflicts. We assume that banks are less skilled at monitoring than venture capitalists. Banks can determine only whether or not the firm should be liquidated, whereas venture capitalists can also learn (at added cost) whether a safe or risky continuation strategy is best. The underlying idea is that the firm's optimal continuation strategy is affected by subtle details of the firm's situation that the institution can identify only through more intensive monitoring. Intuitively, venture capital funds are better at assessing the firm's strategic situation because they are more specialized and have more expertise at running firms than banks do. In our model, passive monitoring by a bank reveals only whether the firm is in a good or bad state, while active monitoring by a venture capitalist also reveals whether a firm in the good state is in a high substate, in which the risky strategy is the better choice, or a medium substate, in which the safe strategy is the better choice. It follows that the value of learning the substate so as to choose the best strategy is higher as the variance of the cash flow from the risky strategy across the two substates is higher. We refer to this variance as the firm's strategic uncertainty.

Venture capital's expertise comes at a cost, however. As shown by Lerner and Schoar (2004), venture capital funds impose liquidity restrictions on their investors so as to shield themselves from liquidity shocks. Such shocks would be especially problematic for these funds because of their lack of diversification and their highly information-intensive assets. To compensate for the illiquidity, fund investors demand a higher return, which in turn causes venture capital funds to require a high return from the firm. By contrast, banks’ funding structure leaves them open to liquidity shocks, but with a lower average required rate of return.

For many firms, the level of strategic uncertainty might not be high, or the cost of intensive active monitoring could be so high that the entrepreneur is unwilling to reimburse the institution for this cost. For these firms, bank finance and passive monitoring are optimal. Moreover, the optimal contract for bank finance is debt. As shown by Winton (2003), debt is less risky than equity, and so the institution's assets are less affected by its private information about the firm, reducing adverse selection problems when the institution itself needs additional funding. Because these costs are passed on to the entrepreneur in her cost of funds, she shares this preference for debt, all else equal.

If instead strategic uncertainty is high, so that the impact of the choice between risky and safe strategies varies greatly with the firm's precise situation, venture capital finance and active monitoring are optimal. For the venture capital fund to have incentive to monitor actively, it must gain greatly from having the firm pursue a risky strategy when conditions are favorable and otherwise gain greatly from a conservative strategy. In general, debt does not accomplish this in a cost-effective manner. Although the venture capital fund's promised payment can be set so high that it bears most of the firm's cash flow risk, this implies that the fund effectively buys much of the firm initially. To the extent that this is more than the firm's required investment, this needlessly increases the firm's reliance on costly venture capital. A position that combines debt with equity (either a convertible security or joint holdings of debt and equity securities) can give the venture capital fund the necessary exposure to the firm's strategic decision with a lower overall investment.

The firm must have several characteristics if venture capital is to be optimal. First, strategic uncertainty must be high. Second, expected profits from the risky continuation strategy cannot be too high. Otherwise, the institution can recoup its investment even if the firm unconditionally pursues the risky strategy, which is the manager's preference. Third, the firm's cash flow distribution must be sufficiently positively skewed; i.e., the probability of success must be low, the value of the firm in liquidation low, and the firm's cash flows in success high. Greater skewness means that the institution can recoup its investment only by taking high payments when the firm is successful, so that it gains more from active monitoring of the firm's strategic decisions. It also implies that, if a bank financed the firm, its liquidity costs would be extremely high.

Venture capital is also more likely to be optimal as venture capital funds’ cost of capital decreases or the severity of bank liquidity shocks increases. Moreover, a decrease in venture capitalists’ cost of capital not only increases the number of firms that receive venture capital funding, but it also increases the number of firms pursuing risky strategies at least part of the time, as firms that would have opted for bank finance and conservative strategies switch to venture capital finance. This has both positive and negative consequences. On the one hand, there is an increase in risky activity, including the adoption or aggressive expansion of innovative products. On the other hand, even though the venture capital funds permit the risky strategy only when conditions seem good, the risky strategy still has the potential to perform badly. Thus a boom in venture capital finance could sometimes be followed by a bust when risky strategies do not pan out. This could help account for the 1990s boom in venture capital financing followed by the bust of the early 2000s.

Although several papers consider various aspects of venture capital financing, few researchers address the issue of what determines the choice between venture capital and bank financing. In Landier (2003), an economy's entrepreneurs choose safe projects backed by bank debt and low monitoring if the stigma associated with failure is high and risky projects backed by venture capital finance and high monitoring if the stigma associated with failure is low. In Ueda (2004), the choice between bank and venture capital financing depends on the relative importance of more accurate screening and the level of intellectual property rights protection. By contrast, our model makes cross-sectional predictions on the relative use of bank loans and venture capital based on differences in the risk and returns of firms’ cash flows, and it explains the differences between banks and venture capitalists in terms of financial securities employed and exercise of control.

A paper that is closer to our model is that of Schmidt (2003), with a model in which both entrepreneur and venture capitalist can invest effort to improve the performance of the firm in different states of the world. Convertible debt gives all cash flows in the bad state to the venture capitalist and splits cash flows between both parties in the good state, which is optimal for getting both parties to exert effort in the states where their contribution is assumed to matter most. Although Schmidt provides a motivation for the use of convertible debt in venture capital, he does not model the choice between bank and venture capital finance, which is the key focus of our paper. Also, whereas Schmidt simply assumes that venture capital targets have certain features, we model the pros and cons of venture capital in terms of primitives, such as the underlying characteristics of the firm's cash flow distribution, the importance of strategic choices that follow directly from these characteristics, and the monitoring activity of the financing institution.3

Another related paper is Hellmann (2006). In his model, the start-up firm faces a choice between an acquisition and an IPO. It is assumed that both the entrepreneur and the venture capitalist contribute to firm value after an IPO, whereas they both retreat from the firm following an acquisition. Convertible securities that automatically convert to common equity upon completion of an IPO provide the best incentives to the entrepreneur and the venture capitalist when their effort is required the most. Like Schmidt (2003), Hellmann assumes that venture targets have certain specialized features, and he does not model the choice between bank and venture capital finance. In our concluding section, we discuss the contrasts between our results and those of Schmidt and Hellmann in more detail.

Whereas we focus on the choice between banks and venture capital, two recent papers focus on the choice between venture capital and angels (wealthy individuals who invest directly in entrepreneurial firms). In Chemmanur and Chen (2003), venture capitalists can add value to some of the firms they finance, but angels cannot. In Bernhardt and Krasa (2004), venture capitalists are informed investors who have control rights, angels are informed investors who do not have control rights, and banks are uninformed investors who do not have control rights. In reality, banks have some private information and control rights, as we assume.

The rest of the paper is organized as follows. Section 2 outlines our model and basic assumptions. Section 3 examines how exercise of control by the institution at date 1 depends on the type of monitoring that it engages in. Section 4 describes the bank's liquidity costs as a function of its information and liquidity needs. Section 5 examines equilibrium behavior for the bank and for the venture capital fund. Section 6 examines the firm's optimal choice between the bank and the venture capital fund as a function of the underlying characteristics of the firm. Section 7 discusses empirical implications and concludes.

Section snippets

The model

The firm: A firm operates over three dates: 0, 1, and 2. At date 0, the manager of the firm makes an investment I. Because the manager has no investable funds of her own, she must obtain the necessary funds by issuing claims to an institution. The firm yields verifiable cash flows at date 2, which we denote by X. At date 1, the firm can be in one of two states, good or bad; also, the good state has two substates, which we refer to as medium (substate i=m) and high (substate i=h). We refer to a

Equilibrium exercise of control at date 1

In this section, we begin by defining equilibrium in our setting. We then analyze the institution's exercise of control and possible renegotiation with the manager at date 1, taking the institution's level of monitoring and contractual payments as given. This allows us to state the expected value of the firm (cash flows plus control benefits) and the expected payments to the institution as a function of monitoring level and contract structure. We show that under any contract that lets the

Expected liquidity costs

With the results of Section 3 in hand, we can now specify the banks's liquidity costs Λb(S). We begin with a simple formulation and then discuss which features generalize to other possible liquidity settings. The critical feature is that these costs tend to fall as the bank's claim on the firm becomes flatter and safer.

We have already defined the frequency λ and severity β of the bank's liquidity needs. Suppose that there is also a small chance δ that the bank is erroneously thought to have

Equilibrium monitoring

In this section, we analyze the equilibrium outcomes under bank finance and venture capital finance. We examine the conditions under which bank finance and venture capital finance are feasible. Finally, we characterize optimal contract structures for both forms of financing. For bank finance, the optimal contract is debt; and for venture capital finance, the optimal contract usually has some equity-like features as well. Intuitively, active monitoring requires a minimum level of exposure to the

Choice between bank and venture capital finance

Thus far, we have shown that there are three possible monitoring and contracting outcomes to consider. Two outcomes involve bank finance, and one involves venture capital finance. We now determine which of these three outcomes is in fact chosen by the manager, given the underlying parameters that govern the firm's contractable cash flows, the manager's control benefits, and the probabilities of the various states and substates. We then examine how the manager's choice varies with changes in

Concluding remarks: implications and extensions

We suggest that venture capital differs from bank finance by greater use of equity features and by more active monitoring, particularly when the firm is choosing continuation strategies. We now discuss our model's implications for the choice between these two sources of finance, how our results compare with those of two related papers, and some extensions of our analysis.

One point that our model emphasizes is that, from the manager's viewpoint, monitoring is often a necessary evil. Thus, if the

References (26)

  • M. Carey et al.

    Does corporate lending by banks and finance companies differ? Evidence on specialization in private debt contracting

    Journal of Finance

    (1998)
  • M. Carey et al.

    The economics of private placements: a new look

    Financial Markets, Institutions and Instruments

    (1993)
  • Chemmanur, T., Chen, Z., 2003. Angels, venture capitalists, and entrepreneurs: a dynamic model of private equity...
  • Cited by (123)

    • When less is more: Information and the financing of innovation

      2023, Journal of Economic Behavior and Organization
    View all citing articles on Scopus

    We thank Thomas Chemmanur (the discussant) and other participants at the 2004 meetings of the Western Finance Association, seminar participants at the University of Minnesota, University of Michigan, and University of Wisconsin and the 2003 Summer Finance Symposium at the Said School of Business at the University of Oxford for helpful comments. This paper was partially funded by a Carlson School Doctoral Dissertation Fellowship.

    View full text