Financial Contracting Theory Meets the Real World: An Empirical Analysis of Venture Capital Contracts

Transcription

1 Preliminary and incomplete Financial Contracting Theory Meets the Real World: An Empirical Analysis of Venture Capital Contracts by Steven N. Kaplan and Per Strömberg* First Draft: March 1999 This Draft: July 1999 Abstract In this paper, we compare the characteristics of real world financial contracts to their counterparts in financial contracting theory. We do so by conducting a detailed study of actual contracts between venture capitalists (VCs) and entrepreneurs. VCs are the real world entities who arguably most closely approximate the investors of theory. (1) The distinguishing characteristic of VC financings is that they allow VCs to separately allocate cash flow rights, voting rights, board rights, liquidation rights, and other control rights. We explicitly measure and report the allocation of these rights. (2) While convertible securities are used most frequently, VCs also implement a similar allocation of rights using combinations of multiple classes of common stock and straight preferred stock. (3) Cash flow rights, voting rights, control rights, and future financings are frequently contingent on observable measures of financial and non-financial performance. These state contingencies are more common in early stage financings. (4) If the company performs poorly, the VCs obtain full control. As company performance improves, the entrepreneur retains / obtains more control rights. If the company performs very well, the VCs retain their cash flow rights, but relinquish most of their control and liquidation rights. The entrepreneur s cash flow rights also increase with firm performance. (5) It is common for VCs to include non-compete and vesting provisions aimed at mitigating the potential hold-up problem between the entrepreneur and the investor. We interpret our results as most consistent with the theoretical work of Aghion and Bolton (1992) and Dewatripont and Tirole (1994). No theory, however, appears to explain the multi-dimensional nature of the allocation of control rights in these financings. * Graduate School of Business, University of Chicago. This research has been supported by the Kauffman Foundation, by the Lynde and Harry Bradley Foundation and the Olin Foundation through grants to the Center for the Study of the Economy and the State, and by the Center For Research in Security Prices. Peter Lang and Konstantin Semyonov provided excellent research assistance. Amar Bhide, Mathias Dewatripont, Ron Gilson, Rick Green, Jeremy Stein, Luigi Zingales and seminar participants at Maryland, NYU, and the Stockholm School of Economics provided helpful comments. We are grateful to the venture capital partnerships for providing data. Ted Meyer was particularly helpful. Address correspondence to Per Stromberg, Graduate School of Business, The University of Chicago, 1101 East 58th Street, Chicago, IL or at per.stromberg@gsb.uchicago.edu.

2 1. Introduction. There is a large academic literature in financial contracting theory. The papers in this literature often begin with a situation in which an investor negotiates with an entrepreneur over the financing of a project or company. These theoretical papers typically make a number of different assumptions concerning the nature of these negotiations. These assumptions concern observability of actions, contractibility of actions, the ability to renegotiate, and the nature of information and uncertainty. Given the assumptions and the models, the papers then generate predictions. For example, a key assumption in Hart and Moore (1998) is that firm output is observable by outsiders, but not verifiable. As a result, it is not possible to write contracts on output. Despite the large volume of theory, relatively little empirical work exists that compares the characteristics of real world financial contracts to their counterparts in financial contracting theory. In this paper, we attempt to inform theory by conducting a detailed study of actual contracts between venture capitalists and entrepreneurs. Venture capitalists (VCs) are real world entities who most closely approximate the investors of theory. VCs have strong incentives to maximize value, but, at the same time, receive few or no private benefits of control. In describing these contracts, we consider the appropriateness of different assumptions and predictions in financial contracting theory. In this version of the paper, we study detailed information on thirty venture capital investments in twenty-three portfolio companies by six venture capital firms 1. Future versions of this paper will utilize detailed information on venture capital investments in more than one hundred portfolio companies by at least ten venture capital firms. For each portfolio company 1 We use the terms venture capital firm and venture capital partnership interchangeably.

3 investment, we asked the VC firm to provide the contractual agreements governing each financing round in which the firm participated. We also asked the VC firm to provide the company s business plan, internal analyses evaluating the investment, and information on subsequent performance. We describe the contracts between the portfolio companies / entrepreneurs and the VCs in great detail. We then consider how well these contracts are described by the assumptions and predictions embodied in different financial contracting theories: (1) "traditional" principal agent theories such as Harris and Raviv (1979); (2) control theories in which cash-flows are verifiable but actions are not such as Aghion and Bolton (1992) and Dewatripont and Tirole (1994); (3) stealing theories in which cash-flows are neither observable nor verifiable such as Hart and Moore (1998) and Gale and Hellwig (1982); (4) agency theories in which the inalienability of human capital is paramount such as Hart and Moore (1994); (5) theories of asymmetric information such as Myers and Majluf (1984); and (6) theories specific to venture capital financing. We obtain the following findings. First, a key feature of VC financings is that they allow VCs to separately allocate cash flow rights, voting rights, board rights, liquidation rights, and other control rights. We explicitly measure and report the allocation of these rights. We believe our measurements are more comprehensive and substantially more detailed than those in any previous work. Second, while convertible securities are used most frequently, VCs also implement the same set of rights using combinations of multiple classes of common stock and straight preferred stock.. 2

4 Third, cash flow rights, voting rights, control rights, and future financings are frequently contingent on observable measures of financial and non-financial performance. These state contingencies are more common in early stage financings. Fourth, these rights are allocated such that if the company performs poorly, the VCs obtain full control. As company performance improves, the entrepreneur retains / obtains more control rights. If the company performs very well, the VCs retain their cash flow rights, but relinquish most of their control and liquidation rights. Fifth, we find that it is common for VCs to include non-compete and vesting provisions aimed at mitigating the potential hold-up problem between the entrepreneur and the investor. These provisions are more common in early stage financings where it is more likely that the hold-up problem is more severe. We interpret our results as most consistent with the theoretical work of Aghion and Bolton (1992) and Dewatripont and Tirole (1994). Dewatripont and Tirole (1994), in particular, capture the gradual shift in control rights that we document. No theory, however, appears to explain the multi-dimensional nature of the allocation of control rights in these financings. In comparing financial contracting theories to real world contracts, we believe this paper breaks new ground. In describing venture capital contracts, our paper extends previous work by Sahlman (1990), Gompers (1997), and Black and Gilson (1998). Sahlman (1990) describes the basic deal structures used in venture capital investments and the economic rationales for them. His is the first detailed discussion of these issues that we are aware of. We extend Sahlman (1990) in several ways. First, Sahlman bases his analysis on forty stock-purchase agreements from a broad range of VCs. His analysis, however, is almost entirely qualitative. He does not present any systematic description and analysis of those 3

5 agreements. Second, Sahlman focuses on the typical set of terms and does not report or analyze the broad range of terms and contingencies. This is particularly important for considering the appropriateness of the assumptions and predictions of the theories that distinguish between cash flow rights, control rights, and liquidation rights. 2 The data and approach in this paper are perhaps most closely related to those in Gompers (1997). His paper is similar to ours in that he describes aspects of venture capital contracts. This paper differs, however, in a number of respects. First, Gompers had access to only a subset of the contracts and information that we have analyzed. He does not analyze data on ownership (cash flow rights), voting rights, liquidation rights, or descriptions of the underlying businesses (from the business plans). Second, Gompers selected his sample to consist entirely of investments in convertible preferred stock, therefore, eliminating some of the important variation that we find in these financings. Third, Gompers theoretical and empirical focus is on explaining the use of convertible securities rather than understanding the relationship between venture capital financings and different financial contracting theories. Black and Gilson (1998) consider explanations for the greater vitality of the venture capital market in stock market- versus bank-centered capital markets. In so doing, they describe different aspects of venture capital contracts. They make the argument, which we confirm as important, that automatic conversion provisions provide important non-monetary incentives to entrepreneurs because they transfer control from the VC to the entrepreneur if the entrepreneur performs well. Like Sahlman (1990), however, they do not present any systematic evidence on the contracts themselves. 2 Because much of the new financial contracting theory had yet to be written, it is not surprising that Sahlman (1990) did not address these issues. 4

6 The paper proceeds as follows. Section 2 describes the assumptions and predictions of a number of prominent financial contracting theories. Section 3 describes our sample. Section 4 presents our results. Section 5 discusses our results in relation to the assumptions and predictions of existing theories. Section 6 summarizes and discusses our results. 2. Theories In this section, we summarize a number of financial contracting theories. For each theory, we describe the key contracting assumptions and predictions. We group these theories by their general approach. The theories we study are meant to be representative, not exhaustive. More extensive overviews can be found in Allen and Winton (1995) or Harris and Raviv (1992). The security design theories interpret financial contracts as the solution to conflicts of interest or agency problems between investors and the entrepreneur. A conflict exists because the entrepreneur must transfer a portion of the profits generated by the project back to external investors in return for their financing. The entrepreneur will not take the optimal action because he does not get all the monetary benefits from taking that action, while alternative actions give other benefits to the entrepreneur. The different security design theories assume different types of conflicts of interest in choice of actions. These include: (1) not exerting the optimal amount of costly effort; (2) taking actions that yield private benefits rather than monetary benefits; (3) spending resources on perks or stealing; (4) holding up investors by threatening to leave the project. 5

7 2.1 Traditional" principal agent theories The traditional principal-agency approach, pioneered by Holmstrom (1979), assumes that the agent s effort is unobservable to the principal. There exist, however, one or more observable and contractible signals that are correlated with effort. The optimal incentive contract ensures that the agent puts in enough effort by making the agent s compensation dependent on the outcome of the signals. In the context of a financing problem, the signal is typically output or profits. If the agent is risk-neutral, the optimal contract involves making compensation as sensitive as possible to the signals, i.e. as high pay-performance sensitivity as possible. Harris and Raviv (1979) show that with a risk-neutral principal and agent, and no wealth constraints, the optimal financing contract is to give a fixed payment to the investor and make the manager the residual claimant. The problem with this solution, however, is that the entrepreneur is generally wealth constrained, which is why he needs to borrow in the first place. Hence, contracts where the entrepreneur pays the investor in the case of bad outcomes are not feasible. Innes (1990) shows that with a risk-neutral, wealth constrained manager, and the added assumption that payouts to investors and the manager have to be monotone in profits, the optimal contract corresponds to giving the investor a debt claim. The results of this class of models are quite sensitive to a number of alternative assumptions, e.g. the number of actions and signals, and the risk preferences of the agent. With risk-aversion, for example, pay-performance sensitivity will decline with the risk of the firm. One potential limitation of these models is that they are all about cash incentives and compensation. Ownership is relevant only as it affects pure cash-flow rights. To the extent they 6

8 are important, these models cannot explain the allocation of control rights between investors and the entrepreneur such as voting rights, board membership, and liquidation rights. 2.2 "Control theories": cash-flow verifiable but not actions In the traditional principal-agent models, actions are unobservable, but signals such as output or monetary results are contractible. The control theories change these assumptions by assuming that actions are indeed observable, but not verifiable. Output and monetary benefits are contractible. As a result, control rights that determine who chooses which action to take will be important. The control theories build on the incomplete contracting literature, pioneered by Grossman and Hart (1986). Two important papers that take this approach to security design are Aghion and Bolton (1992) and Dewatripont and Tirole (1994). In Aghion and Bolton (1992), the project yields both monetary benefits, i.e. profits, that are verifiable and can be transferred to outside investors, and private benefits or actions that are non-verifiable and only go to the entrepreneur. The magnitude of these benefits, in turn, depends on what (non-verifiable) action that is taken with respect to the project. In order to finance the project, the entrepreneur gives up some of the monetary benefits to investors. This introduces a conflict of interest. If the entrepreneur is in control, he has the incentive to take actions that increase private benefits at the expense of monetary profits. In order to induce the entrepreneur to take the efficient action, the investor has to pay the entrepreneur (i.e. renegotiate the contract), rendering the project less profitable ex ante. As a consequence, the expected monetary profits might decline so much that the project would can no longer be financed. 7

9 To overcome this problem, the entrepreneur can give up control of the project to investors. The investors will then take the action that maximizes monetary profits, making the project more financially viable. The cost of this to the entrepreneur is that the investors do not care at all about his private benefits when choosing their action. Since the entrepreneur is wealth constrained, he will not be able to pay the investor to avoid this ex-post inefficiency and no renegotiation will occur. It is optimal to give as much control to the entrepreneur as possible, given that the investors participation constraint is met. When full entrepreneur control is not feasible, it is optimal to give the investor control in enough states of the world to break even, e.g. in the worst states of the world where profits are likely to be low. Aghion and Bolton (1992) point out that a debt contract that transfers control to investors in default states has this feature. Dewatripont and Tirole (1994) build on Aghion and Bolton (1992) by focusing on the optimal correlation between control rights and cash flow rights. They add a prior effort problem (as in "traditional theories"), and assume that the person in control also will have the choice between two non-verifiable actions a more risky and a less risky one. The entrepreneur always prefers the riskier action (the other one involving for example a partial or total scaling down of operations). The optimal contract calls for providing traditional pay-for-performance incentives to induce high entrepreneurial effort. In addition, however, the probability of the less risky action being taken should increase when performance deteriorates. To achieve this objective, control has to be gradually transferred, as performance deteriorates, to an outside investor who is increasingly less sympathetic to the entrepreneur and increasingly more likely to choose the less risky action. As a result, the party in control should look more and more like a debtholder (because debtholders prefer less risk) when things get worse, while more control should be transferred to the entrepreneur or to an equityholder (because equityholders prefer more risk) as 8

10 performance improves. Dewatripont and Tirole (1994) show that the contract can be implemented either with a single outsider in control with a contingent security, or with contingent control and noncontingent securities. 2.3 Stealing theories : cash flows neither observable nor verifiable One criticism of the control theories see Hart (1995) is that changes in control in those models do not necessarily coincide with default on a contracted payment. Hence, these theories do not explain one of the distinguishing features of real-world debt contracts. These critics argue that it may be more appropriate to assume that not only actions, but also profits and cash flows themselves are non-observable or non-verifiable. As a result, there is no way to stop the entrepreneur from stealing the profits from the venture if he so wishes. There are two main strands to this literature. The first one, the Costly State Verification (CSV) models of Townsend (1979), Gale and Hellwig (1982), and others, assumes that profits are completely unobservable, unless a verification cost is paid. The other strand, represented by Hart and Moore (1998), Bolton and Scharfstein (1990), and Fluck (1998), assumes that profits are observable but not verifiable to outsiders and courts. The resulting optimal contracts are very similar for the two approaches. The optimal financial claim is a debt-like claim in which (1) the entrepreneur promises a fixed payment to the investor; and (2) the investor takes control of the project and liquidates the assets (or, in the CSV-models, pays the verification cost and takes all the remaining profits) if the payment is not made. Fluck (1998) and Myers (1998) show that an equity-like claim can also work in an infinite period setting, where the investors have the right to replace the entrepreneur (in which case they get the value of the firm without the entrepreneur, i.e. the "liquidation value") and will do so whenever the dividend is too low. 9

11 2.4 Inalienability of human capital theories Hart (1995) concedes that the assumption of unlimited stealing opportunities that underlies the models of the previous section might be a bit extreme (p.102). Hart and Moore (1994) relax the stealing assumption, while developing a model that has a similar intuition or prediction. They assume that (1) the firm s value with the entrepreneur exceeds its liquidation value and (2) the entrepreneur / manager cannot commit not to leave firm. Because the firm is worth less without him, the entrepreneur can threaten to leave the firm unless any promised payment is negotiated down closer to the liquidation value. Hence, external financing capacity is limited by the value of the firm's assets without the entrepreneur and by the amount of the going concern value that the investor would be able to capture in a renegotiation. The optimal contract will call for a debt security, where control is transferred to the investor if the promised payment is not fulfilled, in which case the firm is liquidated. The size of the promised debt payment is limited by the liquidation value of the assets and by the relative bargaining power of the investor. 2.5 Asymmetric information theories Finally, there are two other security design approaches in the literature. The first group is the theories that extend Myers and Majluf's (1984) signaling model to a security design setting. These theories assume that the entrepreneur has private information about the quality of the venture. Given the assumption that the entrepreneur designs the security to offer investors (rather than the other way around), and that payoffs have to be monotone, Nachman and Noe (1994) and Duffie and DeMarzo (1999) show that the optimal contract will be debt. The intuition is that in order to minimize the cost of mispricing, the good firms will have an incentive to issue the 10

12 least informationally sensitive security which turns out to be debt. If agents are risk-averse and not wealth constrained, entrepreneurs can also signal higher quality by retaining a larger stake of the firm, as in Leland and Pyle (1977). In these models, the greater the asymmetric information, the greater the stake that will be retained by the entrepreneur. An alternative approach that perhaps is more relevant for venture capital is to assume private information on behalf of the investors rather than the entrepreneurs. Garmaise (1998) develops a model of security design where venture capitalists who have privately observed signals about the project bid for the entrepreneur's venture using different contract provisions. The optimal strategy for the venture capitalist involves offering to take a debt claim when the private signal is low, and an equity claim when the private signal is high. 2.6 Venture capital-specific theories In recent years, a number of theories have been proposed that specifically address venture capital contracting. Most of these theories have aimed at explaining the use of convertible securities in these financings (based on the results in Sahlman (1990). Berglof (1994) introduces one of the first theories of venture capital contracting. In this paper, the financial contract is designed to maximize the ex ante expected value in a future sale of the firm to new buyer. In his set-up, cash-flows are observable but not verifiable (as in Hart and Moore (1998)); the entrepreneur derives non-transferable private benefits; and the future buyer can expropriate some of the assets in the bad state after having acquired control of the firm (at the expense of any minority shareholders). Control rights determine who will bargain with the future buyer, while cash flow rights determine how much of the future buyer s productivity gain will pass on to the holder. The optimal contract resembles a convertible security in the 11

13 sense that it gives the entrepreneur control in good state, with the venture capitalist getting all cash-flow rights, and the venture capitalist control in the bad state as well as cash-flow rights. Giving control rights to the entrepreneur in the good state enables him to get fully compensated for his private benefits in the bargaining with the future buyer in this state. Giving the venture capitalist control in the bad state allows him to get compensated for being diluted by the new buyer after the sale. Cornelli and Yosha (1998) provide a theory of convertibles and staged financings. In their paper, the entrepreneur can engage in short-term signal manipulation, making short-term profits seeming higher than they actually are. When financing is staged, the entrepreneur has an incentive to manipulate profits in order to make the firm look good enough for the venture capitalist to finance the next round. The model shows that with a properly designed convertible security, the entrepreneur no longer has an incentive to manipulate. This is achieved by giving the VC more of the cash flow rights upon conversion, the better the signal. Repullo and Suarez (1999) solve for the optimal security design in a setting where both the venture capitalist and the entrepreneur have to provide unobservable effort. In their model, the project needs financing in two stages, start-up and expansion. All (positive) cash flows occur after the expansion stage. At startup, the project s profitability is unknown. At expansion, the profitability is known, but not verifiable so that the initial contract cannot be written contingent on that profitability. VC effort is valuable only in the second period and the second period payoffs are dependent on the effort that the VC and the entrepreneur provide. In the optimal contract in their model, the first period investors who might not be the VC receive a payoff of zero if profitability is low and a fixed fraction of the firm if profitability is high. The authors argue that this is consistent with the use of convertible debt in early-stage 12

14 venture financings. When the potential profitability is verifiable and can be contracted on, Repullo and Suarez find that the VCs finance the second stage if the signal is high enough and receive a fixed percentage of the company s equity. Hellman (1998) also focuses on a case where the venture capitalist has to be provided optimal incentives to provide effort. Similar to the set-up of Aghion and Bolton (1991), the project yields verifiable monetary benefits that can be pledged to investors and nonverifiable private benefits that accrue only to the entrepreneur. In addition, at an intermediate stage, the venture capitalist can engage in a costly search to replace the original entrepreneur with a new manager who might add more value. The original entrepreneur will not leave voluntarily, however, because he loses his private benefits. Hence, in order to replace the entrepreneur, the venture capitalist has to have full control rights. The optimal financing contract calls for: (1) giving the entrepreneur all equity compensation which increases his effort and decreases his resistance towards replacement; and (2) giving the venture capitalist control if the benefits from replacing management are high enough. 2.7 Assumptions and Predictions Exhibits 1 and 2 summarize the assumptions and predictions of the different theories. 3. Sample In this preliminary version of the paper, we analyze thirty VC investments in twentythree portfolio companies by six venture capital partnerships. In future versions of this paper, we will analyze investments more than one hundred portfolio companies by at least twelve venture capital partnerships. 13

15 3.1 Description To obtain this sample, we asked each VC to provide detailed information on as many of their portfolio company investments as they were willing to provide. For each of these companies, we asked the VC to provide the term sheet as well as the stock purchase and security purchase agreements for each financing round in which they participated. These documents typically include the financing terms, the firm s equity ownership investors, founders, management, etc. and any contingencies to future financing. We also asked the VC to provide (if available) the portfolio company s business plan at the time of the financing, the VC s internal analysis of the investment, and the subsequent portfolio company financial performance. Finally, we also requested that the VC provide the private placement memoranda / offering documents for the funds that they have managed over the sample period. With these data, we will be able to construct cross-sectional tests as well as to test for any portfolio company selection bias. Table 1 presents summary information for our sample. As mentioned above, panel A indicates that we have thirty investments in twenty-three portfolio companies by six VC firms. Thirteen of these investments are seed or start-up rounds. I.e., the firms receiving financing either had just begun operating or were not yet operating. The remaining investments are later stage rounds in which the firms were already operating. We have the contractual documents for all thirty investments; business plans for twelve; and some internal VC description of the investment for twelve. Panel B shows that all but two of the financing rounds were completed between 1996 and We view the young age of the sample as positive for two reasons. First, our findings will 14

16 reflect current practice in venture capital financing. Second, it is unlikely that the VCs selected many of these the companies based on the final outcome because the final outcome of many of those investments is still unresolved. Panel C shows that the portfolio companies were provided by six venture capital firms with no more than seven of the twenty-three from any one VC. Panel C also indicates that future versions of this paper will incorporate a substantially larger and richer data set that will allow for more cross-sectional analysis as well as more general conclusions. We possess (but have not yet analyzed) data on investments in more than 100 portfolio companies by twelve venture capital firms. Finally, Panel D indicates the amounts of the sample financings. The VCs committed a median of $5 million in equity in each financing round. (This amount is the total for all VCs investing in the round.) The VCs actually disbursed a median $3.8 million at the time the round closed. The mean amounts are somewhat higher than the medians because of two or three relatively large financing commitments. 3.2 Sample selection issues In this section, we discuss potential selection issues concerning our sample. Our sample of portfolio companies and financings is not a random sample in that we obtained the data from six (and ultimately twelve) venture capital firms with whom we have a relationship. We do not believe that this selection is of much concern to our results because we are not attempting to measure performance. Rather, we are attempting to characterize what contracts look like in general and, perhaps more importantly, what contracts are possible. The financings in our sample certainly show what contracts are possible. 15

17 The contracts also represent financings by more than the six VC firms that provided data. The twenty-three companies in our current sample received VC financing from twenty-four additional VC firms either in the financing round in our sample or in other financing rounds. A total of thirty different VC firms, therefore, invested under the terms of the contracts in our sample. This suggests that the financings in our sample are likely to be representative of VC contracts in general. The more likely bias in our sample is that we have selected VC firms (or have chosen to give us data) that are better than average and that the contracts in our sample may be above average in some sense. If this is so, we believe this strengthens our results because we are more likely to have identified sophisticated, value maximizing principals. In future versions of this paper, we expect to be able to test for VC quality effects by comparing the contracts to measures of VC performance. 4. Results In this section, we describe the contracts between the portfolio companies / entrepreneurs and the VCs in great detail. We first describe the securities issued. We then describe how these contracts allocate cash-flow rights, voting rights, board rights, and liquidation rights. Last, we consider in more detail the contingencies involved in allocating those rights. In the analysis, we distinguish between early stage financings and later stage financings. Early stage financings are financing rounds that provide seed or start-up financing. These financing rounds occur, respectively, before the company has been formed and in the first year of a company s existence. 3 Later stage rounds are all other rounds. This distinction is an 3 See Sahlman (1990). 16

18 interesting one because uncertainty about viability, inalienability, and verifiability of the company should be greater in early stage than in later stage rounds. This will be important in distinguishing among financial contracting theories. In future versions of this paper, we also will distinguish whether a round is the first one in which the company utilizes VC funding. This distinction is an interesting one because asymmetric information between the VC and the founders should be greater in the first VC round than in subsequent rounds. Again, this will be important in distinguishing among financial contracting theories. 4.1 Securities Panel D of table 1 reports the types of securities used in the thirty financing rounds. Consistent with Sahlman (1990) and Gompers (1997), convertible preferred stock is the most commonly used security, appearing in twenty-five of thirty financing rounds. Panel D also indicates, however, that VC financings do not always use convertible preferred stock. In fact, three of the thirty financing rounds (and three of the twenty-three portfolio companies) do not use any form of convertible security. Instead, they use multiple classes of stock or a combination of straight preferred and common stock. While the VC financings utilize different types of securities, the financings are similar in that they allow for different allocations of cash flow, voting, board, and liquidation rights. For example, in the financing that uses multiple classes of common stock, the VCs receive a different class of common stock than the founders who receive two classes of common stock. The VC class of common stock has voting, board, and liquidation rights that are different from the founder classes of common stock. The cash flow rights of the classes of common stock also 17

19 differ in that the two classes of founders stock vest under different conditions than the VC class (which vest immediately). 4.2 Cash flow rights Table 2 presents our results on cash flow rights. By cash flow rights, we mean the fraction of a portfolio company s equity value that different investors and management have a claim to. Measuring cash flow rights is not trivial, however, because many of the cash flow rights accorded to founders and management are contingent either on subsequent performance (through performance vesting) or on remaining with the company (through time vesting). Table 2, therefore, present three measures of cash flow rights. The first minimum VC ownership measures cash flow rights under the assumption that management meets all performance and time vesting milestones or contingencies. The second maximum founder and employees vesting measures cash flow rights under the assumption that all non-performance / time-vesting stock and options vest. The third maximum VC ownership measures cash flow rights if management does not meet any performance or time vesting milestones. Under each of the three measures, VC%, founders% and other% are, respectively, the percentage of cash flow rights owned by the venture capitalists, the founders, and others. Founders include the founding management team. Others include employees and previous non-vc investors. Panel A indicates that the VC controls roughly half the cash flow rights on average; founders roughly one-third; and others roughly one-sixth. These amounts should be of great interest to practitioners and academics. It suggests that substantial equity ownership on the part of founders / managers is desirable. On the other hand, it also indicates that founders / managers give up a large fraction of ownership. 18

20 Panels B, C, and D indicate that founders retain significantly larger ownership stakes in later stage financings compared to early stage ones. We suspect that this result is a function of the fact that founders of companies that receive VC funding for the first time in a later stage retain more equity ownership than founder of companies that receive VC funding for the first time in an early stage. Table 2 indicates that there are substantial state-contingencies built into the cash flow rights. The VC stake is a median of 7.9% lower (average of 11.5%) lower under full vesting and good performance compared to the minimum vesting, bad performance state. For early-stage companies, the average and median are roughly 18%. The state-contingency (i.e. the use of performance benchmarks and vesting) is significantly higher (at the 2.5% level) in earlier stage financings compared to later stage ones. This is consistent with the traditional principal-agent theories, because agency problems are presumably higher for earlier-stage companies. Payperformance sensitivities, therefore, should be greater for these companies. This is what we observe. 4.3 Voting rights Table 3 reports post-round voting rights. Voting rights measure the percentage of votes that investors and management have to effect corporate decisions. Most decisions are based on majority rule. As such, voting rights provide one measure of control rights. Board rights, described in the next section, provide another. In table 3, minimum (maximum) VC votes represents the minimum (maximum) votes the venture capitalists control based on subsequent management performance and stock vesting milestones or contingencies. % VC, % Founder, and % Neither control are, respectively, the 19

21 percentage of instances in which voting control is held by the venture capitalists, the founders, or neither. Switch in control indicates the percentage of instances in which voting control can switch based on subsequent performance. Table 3 indicates that VCs have a voting majority in 60% of all financings in the minimum contingency case. VCs control a majority of votes in 77% of the early stage financings (in the minimum contingency case) versus 47% for later stage financings. In the maximum VC vote contingency cases, VCs control a voting majority in 75% of the financings comprised of 92% of early stage and 65% of later stage rounds. The differences between early and later stage rounds are statistically significant at the 10% level. These results indicate that we do observe state-contingent control rights (i.e. not only in case of default on a debt payment), consistent with Aghion and Bolton. In 16% of the financings, in fact, we see voting control switching depending on state-contingencies. 4.4 Board rights Board rights and board seats also have a large effect on the rights to control corporate decisions. While they tend to be related to voting rights, they need not be identical. We distinguish between normal board rights that reflect the board rights or composition at the completion of the financing from adverse state board rights that reflect board rights or composition if the portfolio company performs poorly or reaches an adverse state. We distinguish between three kinds of board members VCs, founders, and outsiders. VC seats are board seats that are reserved for or controlled by venture capitalists. Founder seats are board seats that are reserved for or controlled by the founders / entrepreneurs. Outsider seats 20

22 are board seats that are to be filled by individuals mutually agreed upon by the VCs and the founders / entrepreneurs. Table 4 reports the board results. The boards have an average of 6.3 members and a median of 5 members. These boards are appreciably smaller than those of public companies. 4 Overall, the VC has the majority of the board seats in 40% of the cases, the founder in 20% of the cases, and neither in 40% of the cases. Interestingly, the VCs are less likely to have board control than they are to have voting control. VC board control is somewhat more common for early-stage financings compared to later stage, although the differences are not statistically significant. State-contingent board provisions (i.e. the VC gets full control of the board in the bad state) are present in about a quarter of the cases. This provides another important example of state-contingent control rights. 4.5 Liquidation rights Much of the theoretical security design literature stresses the importance of liquidation rights. Being able to liquidate, or threaten to liquidate, the firm s assets if the firm defaults is often seen to be the main way for an investor to ensure repayment. Table 5 describes the liquidation rights in our venture capital financings. VC financings utilize two types of liquidation rights. The first type consists of the rights a VC has in liquidation or bankruptcy i.e., seniority rights. The second type consists of the rights a VC has to force the company to repay or liquidate the VCs investment In all our financings, the venture capitalist has seniority in case of liquidation or bankruptcy. The liquidation rights, on average, equal the amount of funds that the venture 4 For example, see Yermack (1998) or Gertner and Kaplan (1996). 21

23 capitalist has invested (excluding cumulative dividends, as discussed below). In this sense, the importance of liquidation rights is supported in the data. The venture capitalist does not always hold the most senior claim, however. Claims senior (typically debt) or at par with the venture capitalist average 41% of the investment of the venture capitalist. Such claims are significantly more common for later stage financings than for seed and start-up financings. This is mainly due to the fact that debt is more common in later stages compared to earlier ones. Even though all financings give liquidation rights to the venture capitalists, there are some cross-sectional differences in how strong these are for different deals. One common way of making the liquidation rights stronger is by giving the investor cumulative preferred dividends. Even though these are dividends, and strictly speaking do not have to be paid out, they will accumulate and be added to the liquidation claim. Cumulative preferred dividends are present in 62% of our cases. They are slightly more common in early stage financings, although not significantly so. Optional redemption and put provisions also are commonly used to strengthen the liquidation rights of the venture capitalist's investment. These provisions give the venture capitalist the right after some period time to demand that the firm redeems the venture capitalists claim, typically at liquidation value (or occasionally, at the maximum of the liquidation value and "fair market value"). This is very similar to the right to repayment of principal at the maturity of a debt claim. Without this provision, the liquidation right loses much of its bite because there are no other contracted payments to the venture capitalist that the firm could default on. Optional redemption or put provisions are present in 86% of our financings, and are slightly, but not significantly more common for early stage companies. The maturity of these provisions averages 5.4 years. 22

24 Finally, 31% of our contracts give liquidation (or redemption) rights for other events than default. The most common practice is to give the right to redeem at liquidation value at a sale, merger or IPO, as an alternative to converting the claim to equity if the share price is too low. A small number of contracts, however, give redemption rights at other events, such as upon the termination of the firm's founder. 4.6 Contingencies As mentioned previously, different theories make different assumptions concerning what it is and is not possible to write contracts on. For example, it is common among some financial contracting theorists to assume that firm output is observable by outsiders, but not verifiable. For example, Hart and Moore (1998) assume that the entrepreneur and outside investors can observe firm output, but they cannot write contracts on that output because courts cannot verify the firm output. As a result, it is not possible to write contracts on output. In this section, we report the extent to which contracts between venture capitalists and entrepreneurs are written contingent on subsequent output, performance, or actions. We also detail the types of output that such contracts are written on. Table 6 reports specific examples of contingencies in the VC financings in our sample. Panel A shows that one-third of the financing rounds include provisions that are contingent on subsequent financial performance. In one financing round, the VCs contractually obtain voting control from the entrepreneur if the firm s EBIT -- earnings before interest and taxes -- falls below a mutually agreed upon amount. In another financing round, VCs obtain board control if a firm s net worth falls below a threshold. Net worth, in this instance, is a measure of a company s cumulative cash flow. These examples indicate that VCs are able to write (and presumably 23

25 enforce) contracts in which control rights are contingent on subsequent output quite independently of cash flow rights. Panel B shows that thirty percent of the financing rounds include provisions that are contingent on subsequent non-financial performance. In one instance, share vesting is contingent on product functionality or performance. In several others, vesting is contingent on FDA or patent approvals. The disbursal of committed funding also can be contingent on non-financial performance. For example, one financing was contingent on successfully completing clinical tests. Panel C reports that over one-third of the financing rounds include contingencies based on certain actions being taken. For example, in different rounds, the disbursal of committed funding is contingent on hiring new executives, developing new facilities, and completing a new business plan. Presumably, these actions are both observable and verifiable. Finally, panel D indicates that contingencies based on the sale of securities are included in thirty percent of the financings. In particular, ownership and vesting are commonly linked to a subsequent initial public offering or sale of the company. Overall, table 6 generates two strong results. First, investors (VCs) commonly write (and presumably enforce) contracts in which control rights are contingent on subsequent measures of financial and non-financial performance or output. Second, there is a great deal of variation in the contingencies in these contracts. The contingencies appear to be related to the performance measure that is most important to the investors and the company. Table 7 quantifies the qualitative information on contingencies in table 6. Panel A indicates the contingencies based on subsequent financial or non-financial performance, actions, or sales of securities are used in sixty percent of the financings. Interestingly, state contingent 24

26 contracting is more common (although not significantly so) in early stage financings than in later stage one. Table 7 also indicates that 40% of the sample financings themselves are partially contingent on the attainment of some milestone. In these financing, the VCs provide only a portion of the total funding commitment at the closing or signing of the financing. Additional funding is provided contingent on subsequent performance and actions. In two financings, the VC provided only 5% of its total commitment at closing with the rest being contingent. 4.7 Other terms VC financings include a number of additional terms and conditions. Bartlett (1995) and Levin (1998) detail many of these. In this section, we describe several of the terms and conditions that we believe are relevant to the financial contracting theories Vesting and non-compete clauses The inalienability of human capital theories of Hart and Moore (1994) assume that the entrepreneur cannot contractually commit to stay with the firm. Even though it is not possible to write enforceable contracts that force the entrepreneur to stay with a firm, there are contractual provisions that make it more costly for the entrepreneur to leave. In real-world contracts, two methods are commonly used to make it costly for the entrepreneur to leave the firm. First, the entrepreneur s shares can vest over time. This means that the company receives or can buy back any unvested shares for some low value if the entrepreneur leaves. The earlier the entrepreneur leaves, the more shares are still unvested. Second, the VCs can require the entrepreneur to sign a non-compete contract with the firm that prohibits him from working for 25

27 another firm in the same industry for some period of time in case he leaves. Both of these provisions improve the bargaining power of the VCs if the entrepreneur tries to hold up the VC. Table 8 shows that the VC financings in our sample commonly utilize both founder vesting and non-compete clauses. Non-compete clauses are used in 60% of the portfolio companies while founder vesting is in just fewer than half of the financing rounds. Moreover, both vesting and non-compete clauses are used significantly more frequently in early stage financings than in later stage ones. For seed and start-up firms, vesting is present in 70% of all cases and non-compete clauses in 80% of the cases, while for later stage companies the corresponding numbers are 25% and 40%, respectively. If we exclude California companies, where non-compete clauses are arguably not enforceable (see Gilson, 1998), non-compete clauses are used in 90% of all start-up financings. Hence, for early stage firms, where it is more likely that the entrepreneur is crucial for the firm s operations and the potential hold-up problem is more severe, the contracts practically always include contractual provisions mitigating entrepreneurial hold-up Automatic conversion It is common for securities in venture capital financings to include automatic conversion provisions. Under these provisions, the security held by the venture capitalists convertible preferred stock, convertible debt, or a class of common stock automatically converts into common stock under certain conditions. These conditions generally, although not exclusively, relate to an initial public offering (IPO) and require the IPO to exceed a designated common stock price, dollar amount of proceeds, and / or market capitalization for the company. 26