vc inside rounds as rescue financing: theory and …...2 i. introduction venture capitalists (vcs)...

1

VC Inside Rounds as Rescue Financing: Theory and Evidence

Brian Broughman and Jesse M. Fried

Last Revised: July 20, 2010

Abstract

VC-backed startups often invite outside VCs to negotiate and price the terms of follow-on financing rounds, in part to reduce conflicts between the entrepreneur and current VC investors over valuation. But many follow-on financings are structured as “inside” rounds – only current VCs participate. One explanation for inside rounds is that they are used by VCs for self-dealing – to sell themselves stock at a low valuation. We consider an alternative “rescue financing” explanation for inside rounds: current VCs sell themselves inflated-price stock to reduce the costs associated with down-round financing. We test these competing explanations using a hand-collected dataset of inside and outside rounds from Silicon Valley startups sold in 2003 and 2004. We find considerable evidence consistent with the use of inside rounds for rescue financing, and little evidence of their use for self-dealing.

JEL Classifications: G24, G32, G33, G34, K12, K20, K22, M13

Keywords: Venture capital, dilution, corporate governance, opportunism.

For helpful conversations and comments on earlier versions of this paper, we are grateful to Robert Bartlett, Lucian Bebchuk, Sascha Becker, Ola Bengtsson, Effi Benmelech, Tom Chang, John Coates, Paul Gompers, Jim Greiner, Darian Ibrahim, Louis Kaplow, Kate Litvak, Edwin Miller, Eric Rasmussen, Jeff Stake, seminar participants at Harvard Law School, Indiana University, the University of San Diego, Vanderbilt, the 2008 Law and Entrepreneurship Retreat, the University of Cincinnati, the University of Wisconsin, the 2009 American Law and Economics Association annual meeting, the 2009 Law and Society Association annual meeting, the 2009 Midwest Law and Economics Association annual meeting, and the 2009 Conference on Empirical Legal Studies. We would also like to thank Albert Chang, Jake Robinson, Jennifer Su, Bruce Sun, and Fennie Wang for valuable research assistance. VentureReporter.net provided access to their database of mergers and acquisitions. This project was generously supported while the authors were affiliated with Berkeley Law School by a grant from the Kauffman Foundation through the Lester Center for Entrepreneurship and Innovation at UC Berkeley. We thank the many entrepreneurs who agreed to provide data for our research.

2

I. Introduction

Venture capitalists (VCs) play a critical role in the financing of high-risk, technology-based startup companies, investing billions of dollars annually in these businesses (Gompers and Lerner, 1999; NVCA 2009). Startups do not typically receive all of their VC financing at once. Rather, after the initial financing round, additional financing is provided in one or more “follow-on” rounds, each of which is separately negotiated and priced (Gompers, 1995).

In the dot.com era, most follow-on financings were structured as outside rounds: the largest investor in the round is a new VC, and that VC negotiates terms with the startup (Lerner, 1994). Bringing in a new investor to negotiate pricing can mitigate conflicts between the entrepreneur and existing VCs over the value of the firm (Admati & Pfleiderer, 1994; Lerner, 1994; Brander, et. al., 2002), expand the startup’s contact network (Hochberg, Ljungqvist, and Lu, 2007), and improve fund diversification (Tian, 2008).

Since 2001, however, more than half of follow-on rounds have been structured as inside rounds: only existing VCs participate (Bengtsson and Sensoy, 2009; Venture Source 2010). VCs could use inside rounds to engage in self-dealing by exploiting their power in the startup to sell themselves cheap stock (Klausner & Litvak, 2001; Gilson, 2003).1 In fact, many of the lawsuits filed by entrepreneurs allege that VCs used an inside round to dilute the entrepreneur (Atanasov, Ivanov & Litvak 2008).2 But little is known about whether the use of inside rounds for such self-dealing is widespread (Klausner and Litvak 2001).3

We consider an alternative explanation for the use of inside rounds, which we label “rescue financing.” In particular, VCs of firms whose values have declined may use inside rounds to sell themselves stock at an inflated valuation. Although VCs overpay for the shares, they also reduce the costs associated with “down-round” financing – a financing at a lower valuation than the previous round. In certain cases, the reduction in down-round costs from using an inflated valuation may be large enough to make the overpayment worthwhile.

The potential costs to existing VCs of a down-round financing are threefold. First, a down round will trigger anti-dilution rights held by the firm’s existing VCs. The triggering of these rights can substantially increase the existing VCs’ proportional ownership of the firm. While this increase would appear to benefit existing VCs, it typically leaves too little equity to properly incentivize the entrepreneur and other employees. And in some cases, it is mathematically impossible to accommodate anti-dilution rights and give the new investor the equity it requires to participate (Bartlett, 2003). Thus, anti-dilution dilution rights almost always must be renegotiated in a down

1 VCs’ power over other participants in the startup comes from their contractual rights to block financings, expenditures, and corporate transactions (Fried and Ganor, 2006), as well as their seats on the startup’s board, where they often constitute a majority (Broughman and Fried 2010, Kaplan and Stromberg 2003).

2 In one of the few self-dealing cases that actually went to trial (Kalashian v. Advent, 1997) a firm’s VCs, in control of the board and holding over 90% of the firm’s stock, conducted an inside financing round at $0.005 per share, reducing the founders’ proportional ownership from 8% to 0.007%. Several years later, the VCs took the firm through an IPO and the stock price reached $70 per share. The founders sued, and after a weeklong trial (but before a verdict was delivered) the VCs settled the case for $15M (King, 1997).

3 Litigation by entrepreneurs against VCs for self-dealing is rare. The infrequency of litigation could indicate that VC self-dealing is constrained by threat of liability or reputational harm (Black & Gilson, 1998; Gilson, 2003; Smith, 1998, 1999; Suchman & Cahill, 1996; Atanasov, Ivanov & Litvak 2008; Broughman, 2009). Alternatively, the low litigation rate could reflect the existence of substantial impediments to bringing even meritorious lawsuits against VC investors (Fried and Ganor, 2006).

3

round (Mahoney, 2002). Such renegotiation can be costly and time-consuming, particularly if existing VCs have different anti-dilution rights (Bartlett, 2006).4

Second, down rounds can demoralize the entrepreneur and other employees, reducing firm value (Lemon, 2003).5 Third, a down round forces VCs to write down the value of their earlier investments in the firm. This write down reduces the reported performance of the fund holding those shares, and can undermine VCs’ ability to raise a new fund (Lerner, 1994).

When a firm’s actual value has declined since the last round, the renegotiation and “write-down” costs associated with a down round will be borne directly by the existing VCs. To the extent down-round renegotiation and demoralization costs reduce the value of the portfolio firm, these costs will be borne indirectly by the existing VCs through their earlier investment in the firm.

These down-round costs can be reduced by financing the firm at an inflated valuation. Indeed, if the valuation equals or exceeds that of the previous round, down-round costs would be eliminated: there would be no need to renegotiate anti-dilution provisions, no demoralization, and no write-down costs. However, inflated-valuation financing also imposes a cost on the investing VCs: they overpay for their new shares.

An outside VC will not knowingly overpay for the new shares; it will walk away from the deal if the price is too high. But an existing VC is in a different position, for two reasons. First, to the extent the existing VC overpays for the new shares, it reduces the down-round costs it would otherwise bear via its pre-existing equity interest in the firm. Second, a fraction of the overpayment for the new shares is “recovered” by the existing VCs via their existing ownership interest; the remainder is captured by other shareholders (whom we collectively call the “entrepreneur”). The larger is the existing VCs’ pre-round ownership, the more of the overpayment is recovered. If the reduction in down-round costs and existing VCs’ pre-round ownership are large enough, existing VCs will be better off overpaying for the new shares than conducting an inside round or an outside round at a fair valuation.

Another way to explain the rescue financing tradeoff is that it reduces down-round costs for existing VCs but also transfers value from these VCs to the entrepreneur. In particular, the value transferred to the entrepreneur is the overpayment for the shares issued in the new round of financing, multiplied by the entrepreneur’s proportional interest in the firm before the financing. Existing VCs will thus engage in rescue financing only if (a) the resulting reduction in their down-round costs exceeds (b) the value transferred to the entrepreneur.

The possibility of VCs using inside rounds to avoid one of the down-round costs -- write-downs -- is well understood. One Silicon Valley VC put it colorfully: “The easiest, but most blatant way of hiding a write-down is to conspire with fellow co-investors to do an inside round at an artificially high valuation. … [W]hen all of the existing investors … already own a significant chunk of the company… [i]t doesn’t take much peer pressure on the average VC funded board to get everyone to go along with the plan and the common investors are usually in favor of it because they are technically getting less dilution.” (Burnham, 2004). Anecdotal reports suggest inside rounds are sometimes used to prop up valuations (Garland, 2009).

4 Consistent with the view that renegotiation of VCs’ cash-flow rights is difficult, Bengtsson & Sensoy (2009) find that “major renegotiations of previous round contracts are rare.” A firm could attempt to restore the entrepreneur’s and other employees’ incentives by issuing additional stock options to employees or creating a management carve-out plan, but such measures entail their own transaction costs.

5 Forbes et al (2005) present evidence suggesting that down rounds increase the level of conflict between boards and CEOs.

4

The rescue financing and self-dealing explanations generate different and usually opposite predictions about the about the use and performance of inside rounds: (i) rescue financing predicts that inside rounds are more likely when values decline; self-dealing makes no such prediction; (ii) rescue financing predicts that inside rounds are more likely when VCs’ pre-existing ownership interest is larger; self-dealing predicts the opposite; (iii) rescue financing predicts that inside rounds will be disproportionately down- and even (flat); self-dealing makes no such prediction; (iv) rescue financing predicts inside rounds are likely to be made at inflated valuations (relative to actual value); self-dealing predicts the opposite; and (v) rescue financing predicts inside rounds are likely to generate lower rates of returns than outside rounds; self-dealing predicts the opposite. 6

We find preliminary support for rescue financing (and little support for self-dealing) using a hand-collected dataset of 90 follow-on financing rounds from 45 Silicon Valley firms that were sold in 2003 or 2004. We first investigate how changes in NASDAQ – which we use as a proxy for startup values – affect the likelihood of inside financing. Consistent with rescue financing, we find that a 20% decline in NASDAQ between rounds increases the likelihood by approximately 15% that the next financing will be structured as an inside round. We also find that the likelihood of an inside round is higher as VCs’ pre-round ownership increases. In particular, a 10% increase in existing VCs’ pre-round ownership is associated with a 40% higher likelihood of inside financing. These effects are both statistically significant.

We then examine the valuations used in inside rounds and outside rounds. Consistent with rescue financing, we find that inside rounds are disproportionately down- and even-rounds. Relatively few are up-rounds. For each firm, we then generate a “relative valuation” by dividing the valuation used in the final financing round of each firm to the best observable estimate of the firm’s actual value at the time – the firm’s subsequent sale price. Controlling for various factors, we find – consistent with rescue financing but inconsistent with self-dealing -- that the relative valuations used in inside rounds are significantly higher than the relative valuations used in outside rounds.

Finally, we calculate the rates of returns on all 90 follow-on rounds. Inside rounds in our sample on average have a significantly lower rate of return – by about 30% to 40% - than outside rounds, a result that is consistent with rescue financing but inconsistent with self-dealing.

Interviews with the 45 entrepreneurs of the firms in our sample provide further evidence that inside rounds were not driven by self-dealing. Entrepreneurs generally reported that they were encouraged by the VCs to seek outside financing. Only three of the 24 entrepreneurs in firms using inside rounds reported that they believed the inside rounds were priced at unfairly low valuations. 7 In these three firms, however, the VCs exited at a loss. Thus even if the inside rounds were dilutive ex ante they did not hurt the entrepreneur ex post. In short, while some of the inside rounds in our sample may have been used for self-dealing, the entrepreneurs’ own accounts suggest they were a small minority.

Our sample is limited to Silicon Valley firms sold in 2003 and 2004 and consists only of firms whose entrepreneurs responded to our request for information. Factors unique to the Silicon Valley VC market, to this time period, to firms that end up being acquired in private sales, or to firms whose entrepreneurs were willing to provide data to us could limit the generalizability of our results.

6 Brander et al. (2002) offer a model in which inside rounds are used by existing VCs when they are confident in their valuation of the firm and know they are investing on favorable terms. Their model, like the self-dealing explanation, predicts that inside rounds will generally yield higher returns than outside rounds.

7 None of the entrepreneurs in the 21 firms that had only outside-round financings reported any problems with financing valuation.

5

To help address these selection concerns, we create a “shadow sample” of 102 VC-backed firms randomly selected from all US-based VentureXpert (VX) firms that were founded between January 1, 1996 and December 31, 2002 and that received at least $5M in financing and disclosed valuation data for at least one round of financing. VX does not provide the same type of consistent and fine-grained data (such as valuation data and actual round returns) as our main, hand-collected sample. However, for many of the follow-on rounds of these 102 firms one can determine whether the round is inside or outside and the valuation change since the last round. The financing patterns that emerge in this shadow sample – particularly the increasing use of inside rounds as NASDAQ declines, and the clustering of inside rounds among down and even rounds -- are strikingly similar to those of the main sample, suggesting that the firms in our main sample are representative of VC-backed firms generally during this time period, at least with respect to the use of inside rounds. Appendix B describes the VX shadow sample and compares it the firms in our main sample.

Our paper contributes to the literature on the structure of VC financings. The use of outside rounds has received considerable attention. Admati and Pfeiderer (1994) show that outside rounds can reduce conflicts over valuation between existing VCs and entrepreneurs. Hochberg, Ljungqvist, and Lu (2007) find evidence that outside rounds can increase the startup’s contact base. Our paper provides an explanation for why, notwithstanding these potential benefits to outside rounds, VCs and entrepreneurs may both prefer to use inside rounds, and offers evidence consistent with this explanation.

Our paper also contributes to a better understanding of governance within private VC-backed startups before they are sold. While economists have studied the contracts VCs use when investing in startups (Kaplan and Stromberg, 2003) and how they evolve (Bengtsson and Sensoy, 2009) little is known about how VCs actually exercise control within these firms, including the potential for conflict around financing transactions (Klausner and Litvak, 2001). Our study provides preliminary evidence suggesting that, while self-dealing may occur, VCs generally do not use their control rights to force through inside financings that dilute entrepreneurs.

The remainder of this paper is organized as follows. Part II presents the rescue explanation for inside financing. Part III describes our dataset. Part IV presents our findings. Part V concludes.

II. The Rescue Financing Explanation for Inside Rounds

In this section we describe our rescue financing explanation for inside rounds. In Part A we illustrate the core intuition with a numerical example of a startup firm seeking follow-on financing after its value has declined.8 In Part B, we describe testable hypotheses.

A. Numerical Example

We illustrate the rescue financing theory with a simple numerical example involving a startup firm (“Startup”) seeking follow-on financing after its value has declined.

Startup receives $20M of Series A financing from several VC funds for 50% of Startup’s outstanding equity; the other 50% is held by Entrepreneur. For simplicity, we assume that all of Startup’s equity is in the form of common shares. The Series A investment implies a “post-money valuation” of $40M, and a “pre-money valuation” of $20M.

8 Appendix A provides a formal model of rescue financing.

6

Several years later, Startup needs $20M of Series B financing. Outside investors would demand 40% of Startup’s equity for the $20M. In other words, the deal would be done at a $50M post-money valuation ($30M pre-money valuation). Such a financing would be a “down-round:” the Series B pre-money valuation is less than the Series A post-money valuation.

The proposed financing will impose various down-round costs on Startup and its Series A investors. For simplicity, suppose that these down-round costs are borne directly by the Series A investors and total $2M. 9 These costs would be avoided by using a $60M post-money valuation for the Series B round (that is, giving the Series B investors 33% of Startup’s equity).

Outside Series B investors will refuse to invest at a $60M post-money valuation, but the Series A investors will find it in their interest to do so if the cost of overpayment – the amount of value transferred to Entrepreneur – is less than the $2M benefit from eliminating down-round costs.

The amount of value transferred to Entrepreneur is simply (a) the amount by which the VCs overpay for the Series B multiplied by (b) Entrepreneur’s pre-round ownership interest. If the VCs purchasing the Series B shares invest $20M for a 33% stake in a firm worth $50M, they will pay $20M for an equity stake worth $16.66M ($50M/3). The Series B purchasers will thus overpay for the stock by $3.33M. Because Entrepreneur’s pre-round ownership interest is 50%, Entrepreneur captures $1.66M of the overpayment. The Series A investors will thus have an incentive to use a $60M valuation for the Series B, even though the actual post-money value is only $50M. They avoid $2M in down-round costs at a cost of $1.66M, for a net gain of $0.33M.

Another way to think about the tradeoff for the Series A investors is as follows: if the Series A investors accept outside financing at a $50M post-money valuation, they bear $2M in down-round costs. If the Series A investors conduct an inside Series B round at a $60M post-money valuation, they lose $3.33M on their Series B investment, and make $1.66M more on their Series A investment, for a net loss of $1.66M. Because this net loss of $1.66M is less than the $2M down-round costs they would incur with outside financing at a $50M post-money valuation, the Series A investors are better off doing the Series B round themselves at a $60M post-money valuation. The table below summarizes the existing investors’ payoffs under both an outside round at a $50M post-money valuation and an inside round at a $60M post-money valuation.

Existing Investors’ Payoffs in Outside and Rescue-Financing Rounds

Who Funds Series B?

Nominal Post-Money

Valuation

Actual Post-

Money Value

Existing VCs’ Return on

Series B Shares

Value of Series A Shares

Down-Round

Cost

Total Value for Existing

Investors Outside Investors

$50M $50M N/A $15M $2M $13M

Existing Investors

$60M $50M -$3.33M ($16.66M-

$20M)

$16.66M $0 $13.33M

9 A down-round Series B financing led by an outside investor may transfer value from Entrepreneur to the Series A investors to the extent the Series A investors’ anti-dilution rights enable them to convert into more shares of common without putting up any more money. Because anti-dilution rights are generally not enforced, we ignore this complication here.

7

B. Predictions

The rescue financing and self-dealing explanations for inside rounds generate different predictions about the use and performance of inside rounds:

Changes in Firm Value and Use of Inside Rounds. Down-round costs (and thus the benefits of reducing those costs) – renegotiation costs, demoralization, and write-down costs – are difficult to measure directly. But they are all likely to increase as actual firm values decline. Rescue financing thus predicts that as actual values decline, firms are more likely to receive inside financing. In contrast, the self-dealing explanation for inside rounds does not yield a clear prediction for the relationship between inside rounds and changes in firm value.

VCs’ Proportional Interest and Use of Inside Rounds. The larger is the VCs’ proportional interest in the firm before the financing, the less value will be transferred to the entrepreneur, and the lower will be the cost of rescue financing, everything else equal. Rescue financing thus predicts that, in firms whose values have declined, inside rounds are more likely as VCs’ pre-round ownership percentage is higher. In contrast, self-dealing would dilute the value of all preexisting equity (including the VCs’ own preexisting interest), and consequently the benefit to VCs of self-dealing decreases as the VCs’ pre-financing ownership interest increases. Thus self-dealing predicts that inside rounds are more likely, everything else equal, when VCs’ ownership percentage is lower.

Valuation Changes and Use of Inside Rounds. To the extent inside rounds are used to reduce down-round costs, we would expect an inside financing to be a down- or flat round. An inside down-round financing at an inflated valuation would reduce down-round costs. A flat-round would completely eliminate down-round costs, but at a greater overpayment cost. The existing VCs will choose an inflated valuation that reflects this tradeoff.10 We would not expect an inside round used for rescue financing to be an up round. An up round provides no additional down-round cost reduction relative to a flat round, but increases overpayment cost. In contrast, outside rounds could be down, flat, or up. Thus, rescue financing predicts that inside rounds will be disproportionately down or flat.

Relative Valuations and Returns on Inside Rounds. To the extent inside rounds are used for rescue financing, valuations used in inside rounds should be higher (relative to the firm’s actual value) than valuations used in outside rounds, and realized returns should be lower. To the extent inside round are used for self-dealing, the opposite is true: valuations should be lower in inside rounds than in outside rounds, and returns should be higher.

III. The Data Standard VC databases, such as VentureXpert, do not generally provide the type of fine-

grained detail necessary to test the rescue financing predictions. Such databases do not report consistent valuation data or actual returns to different financing rounds (Kaplan, Sensoy & Strömberg, 2002). Instead, we test our predictions with a hand-collected dataset of 90 follow-on rounds by 45 VC-backed Silicon Valley firms that were sold in 2003 and 2004. This section (i) explains how our data were gathered, (ii) addresses selection concerns, (iii) describes the firms in our dataset, and (iv) summarizes the terms used in follow-on financings.

10 The model in Appendix A analyzes this tradeoff.

8

A. Data Gathering

The data for this study were gathered in 2005-2007 from entrepreneurs in Silicon Valley as part of a larger empirical project on the internal governance of VC-backed firms. The study collected information on conflicts between VCs and other participants in the startups around financings, replacement of the CEO, and the sale of the firm.

We obtained from VentureReporter.net a list of VC-financed firms located in California and sold to an acquirer in 2003 or 2004. We removed firms not located in or around San Francisco, San Jose, or Oakland (broadly defined as “Silicon Valley”),11 leaving a population of 193 firms.

For each firm we sought to locate and obtain data from one or more founders/executives (“entrepreneurs”) about the firm’s life – from formation to sale. We identified current business addresses for entrepreneurs of 141 of the 193 companies and contacted them. Entrepreneurs from 57 of the 141 firms agreed to provide us with data – a response rate of 40.4%. The information obtained, supplemented by publicly filed corporate charters, covers the lifespan of each firm. Among the data gathered were the state of incorporation, detailed cash-flow rights and control rights negotiated in each VC financing round, and the terms of sale. For this study on inside rounds, we removed 12 of the 57 firms because they did not receive follow-on financing, they did not have a true exit event (one where cash or publicly traded securities was paid to the VCs), or we were unable to verify follow-on round valuations. We thus have an effective sample of 45 firms and 90 follow-on rounds of financing.12

B. Selection Issues

Our sample is limited to Silicon Valley firms sold in 2003 or 2004. Factors unique to the Silicon Valley VC market, to this time period, or to firms that end up being acquired in private sales could limit the generalizability of our results. Our sample consists only of companies whose entrepreneurs voluntarily responded to our request for information. There could be systematic differences between firms whose entrepreneurs responded to our inquiries and firms whose entrepreneurs did not. We sought to minimize such biases by soliciting data from every entrepreneur we could locate and offering confidentiality.

To address these selection concerns, we create a “shadow sample” of 102 VC-backed firms randomly selected from all US-based VentureXpert (VX) firms that were founded between January 1, 1996 and December 31, 2002 and that received at least $5M in financing and disclosed valuation data for at least one round of financing. The shadow sample covers a broad range of VC-backed firms regardless of exit outcome. The 102 firm shadow sample was compiled by choosing every 40th firm of the 4070 firms in VX meeting these criteria. Based on the list of VC investors we code the status – inside or outside – for each follow-on round in the shadow sample. While VX does not include sufficient data on round valuations or performance to test our hypotheses directly, the

11 We used LinkSV to filter out firms not meeting these criteria <www.linksv.com>. LinkSV profiles all companies located in Silicon Valley (in or around San Jose, San Francisco, and Oakland) that received VC funding. Companies not appearing on LinkSV were removed from our sample.

12 Due to data limitations a relatively small sample size is common to studies of private (pre-IPO) VC-backed firms. The number of VC-backed firms in other projects using hand-collected data is 51 in Hsu (2004), 170 in Hellman & Puri (2002), 119 in Kaplan & Stromberg (2003), 132 in Cumming (2008), 50 in Broughman & Fried (2010), and 210 in Bengtsson & Sensoy (2009).

9

financing patterns – particularly the increased use of inside rounds after NASDAQ declines, and the clustering of inside financing in down rounds and even rounds – that emerge in the VX shadow sample are broadly consistent with the financing patterns observed in our sample. An analysis describing the VX shadow sample and comparing it to our data can be found in Appendix B.

C. Sample Description

Our sample firms are ‘high-tech’ businesses, primarily in the biotech, software, telecommunications, and internet sectors (Panel A of Table 1). The concentration of firms in our sample is representative of VC financed firms generally (Kaplan and Strömberg, 2003 at 284).

At the time of sale, the firms in our sample had received an average of $44 million in VC funding,13 and had been operating for an average of approximately five years. The average sale price was $55 million, but there is considerable variance in exit outcomes. Some firms were essentially liquidated, yielding a sale price less than one million dollars, while several firms were sold for considerably over $100 million.

Sale proceeds are divided among preferred and common stockholders. Since we know the date of each financing round, the date of sale, and how the sale proceeds are allocated, we can calculate the VCs’ internal rate of return (IRR) for each round of financing and for each firm. The average firm-level IRR for the VCs was 0.20. However, only 16 out of the 45 firms were sold for a profit, suggesting that the average IRR is skewed by a few highly profitable firms. Panel B provides descriptive statistics on the amount invested, years of operation, and sale price.

[INSERT TABLE 1 HERE]

D. Financing Round Characteristics

The 45 firms in our study received, on average, three rounds of VC financing (Panel A, Table 1). We document a total of 90 follow-on rounds of financing.14 We collect detailed information on each round of investment, including the composition of the investor group, the allocation of board control, the VCs’ cash-flow rights, and the valuation used in the financing. Table 2 defines variables used throughout the remainder of the paper and provides a correlation matrix.


Inside vs. Outside

In our dataset, 26 (29%) of the rounds were inside, and 64 (71%) were outside. Inside rounds are more common in later rounds of investment (Figure 1).

[INSERT FIGURE 1 HERE]

13 The companies in our sample also received, on average $0.15 million in entrepreneur/family financing, and $0.45 million in angel financing. This is likely to understate the extent of angel financing, since any investor receiving preferred stock in a financing rounds that include at least one VC investor is treated as a VC investment for purposes of this paper.

14 The number of follow-on financing rounds in our sample is consistent with findings of practitioner surveys. See e.g. VentureOne ‘Deal Terms Report’.

10

Valuation

We record the valuation used in each round of financing. We use and report the nominal post-money valuation.15 The nominal valuation is equal to the VC’s investment divided by the fraction of common stock that the VCs are entitled to on a fully converted basis.

If the nominal pre-money valuation (post-money valuation less the amount invested) is higher than the post-money valuation of the previous round, the new round is considered an “up round”. If the nominal pre-money valuation is lower than the post-money valuation of the previous round, it is considered a “down round.” If the pre-money valuation on the new round is the same as the previous round’s post-money valuation, the round is called “even.” Of the 90 rounds, 45 (50%) were up, 14 (16%) were even, and 31 (34%) were down.

IV. Empirical Results

We now use our data to test the predictions generated by rescue financing. First, we explore whether, as predicted, declines in firm value and larger VC ownership stakes increase the use of inside rounds. Second, we examine whether inside rounds are clustered at down and even valuations. Third, we compare the valuations assigned to inside and outside rounds to determine whether there is evidence consistent with the prediction that inside rounds are made on worse terms. Fourth, we compare the rate of return on inside rounds to that of outside rounds to determine, whether as predicted, the returns on inside rounds are lower. We conclude by describing several possible explanations for inside rounds besides self-dealing and rescue financing, and discussing the extent to which these other explanations are consistent or inconsistent with our results.

A. Declining Values, VC Ownership, and the Use of Inside Rounds

Rescue financing predicts that inside rounds are more likely to occur (i) when a firm’s value has declined since the prior round and (ii) when a firm’s existing VCs hold a large ownership interest prior to the new round. We consider both predictions here.

Declining Values: A firm’s value may decline either because of firm-specific factors (the performance of the firm’s management and engineering team, etc.) or extrinsic factors, such as falling values in the overall market. Because we are unable to observe a reliable measure of firm-specific performance, we focus on external factors that should affect a firm’s value. We use as a proxy for market values the closing price of the NASDAQ national market, for which we have daily values. For each round of financing, we measure the closing price of the NASDAQ national market

15 Nominal valuation is the standard method of valuing VC-backed startups. Nominal valuation, however, implicitly assumes that the common stock held by entrepreneurs has the same value as the VCs’ preferred stock, even though preferred stock typically includes a liquidation preference and other rights that make it more valuable than common stock (Metrick 2007). To account for the value of these rights we also report valuation on an implied basis using an option valuation technique outlined in Metrick (2007) that derives a value for the firm based on the amount the VCs pay for their stock as well as the liquidation preferences, participation rights, and other features of the VC’s preferred stock. For purposes of calculating implied valuations, we use an annualized volatility for the VC industry of 89% (Cochrane, 2005). For details on this approach see Metrick (2007) at pages 252 – 287. The method of valuation – nominal or implied – does not affect our findings reported below. We ran all the regressions in Tables 4 and 5 (below) using both methods of valuation, and we found qualitatively similar results.

11

on the date of investment (NASDAQ). We then calculate the percentage change in this measure since the date of the prior round of financing (Δ NASDAQ). Rescue financing predicts that a decrease in NASDAQ prices should increase the likelihood of inside financing.

Preexisting VC Ownership (%): We measure the fraction of cash flow rights collectively held by a firm’s existing VCs prior to each round of follow-on financing (VC Ownership %) on a fully converted basis. Rescue financing predicts that inside rounds are more likely when VCs hold a large preexisting equity interest.

We estimate, using logit, the likelihood of inside financing as a function of (i) Δ NASDAQ, (ii) VC Ownership %, and (iii) control variables. Table 3 reports our results. In model (3-1) and (3-2) we report univariate logit results for Δ NASDAQ and VC Ownership % respectively. Both have a statistically significant effect in the predicted direction.

In models (3-3) and (3-4) we add additional control variables. We include VC Control, which records whether the VCs controlled a majority of the firm’s board seats prior to the new round. The self-dealing explanation suggests that VC Control may lead to inside financing rounds on terms favorable to the VCs. We also control for the amount invested in the new round (Investment in Round (in $ millions)), and the Number of Directors sitting on the firm’s board prior to the new round. A new investor will often demand a board seat in connection with the financing. If the firm already has a large number of directors on the board the parties may be reluctant to do an outside round, since they wish to avoid increasing the size of the board. Finally, in model (3-4) we add industry dummy variables.

Table 3 reports estimates consistent with rescue financing. An increase in VC Ownership % has a positive effect (significant in all models) on the likelihood of inside financing. A decline in NASDAQ is associated with an increase (significant in most models) in the likelihood of inside financing. Declining valuations in the market for startup firms increases the use of inside financing.

Figure 2 illustrates the relationship between inside rounds and the performance of the NASDAQ market. The top portion of the diagram graphs the weekly NASDAQ closing price from 1997 through 2003. The bottom portion shows the frequency of inside rounds over the same period. When market prices were increasing (in the late 1990s) inside rounds were rarely used. However, after 2000, as NASDAQ prices fell, inside rounds increased in frequency.



Setting all variables to their median value we predict the marginal effect of a change in NASDAQ and an increase in VC ownership on the probability of inside financing. We find that a 20% drop in NASDAQ increases the likelihood of inside financing by about 15% (from 26% to 30%). While NASDAQ is certainly not the only factor driving the use of inside financing, it does have a meaningful impact.

The effect of VC ownership on the likelihood of an inside round is even more pronounced. A 10% increase in VC Ownership % is associated with a 40% increase in the likelihood of inside financing (from 23% to 33%). Rescue financing suggests that VCs’ preexisting equity reduces the cost of inside financing, and thus predicts, consisted with our data, that VCs are more likely to do an inside round when they hold a large preexisting ownership interest in the firm.

12

B. Clustering of Inside Financing in Down and Even Rounds

Rescue financing predicts that inside rounds will be disproportionately down or even. To test this prediction we examine the change in valuation since the last round of financing. We then compare valuation changes across inside and outside rounds.

Figure 3 shows that firms are reluctant to do financings that are slightly down-round. Only one of 31 down-round financings (3%) is done at a valuation between 0% to 25% below the previous round. In contrast, 14 of 45 up-round financings (31%) are done at valuations between 0% to 25% above the previous round, and 14 financings are done exactly at an even valuation. The infrequency of “slightly” down rounds and the clustering of so many rounds at an even valuation are consistent with there being significant costs to down-round financings.

The data also are consistent with the use of inside rounds to minimize down-round costs. The minimization of down-round costs can be accomplished by using an even valuation; there is no reason to do an up-round. In fact, inside rounds are rarely seen in up-rounds. The table below Figure 3 shows that only 8% of inside financings are up-rounds, while 67% of outside financings are up-rounds.

Inside rounds make up disproportionate fraction of the down rounds and even rounds. Of the inside financings in our sample 35% are even-rounds and 58% and down-rounds. This contrasts with outside rounds, where only 8% are even-rounds and 25% are down-rounds. These differences are statistically significant at the 1% level. Using a t-test for equal means we show that inside rounds are more likely to be Down or Even rounds and less likely to be Up rounds.


C. Round Valuation: Inside versus Outside

Rescue financing predicts that inside rounds will have higher valuations (relative to the firms’ actual values) than outside rounds. Unfortunately, it is impossible to observe the 'true' value of a private company at the time of financing. To address this problem, we use the eventual sale price of the firm – negotiated at arm’s length between the buyer of the firm and the firm’s investors -- as a proxy for the firm's 'true' valuation in the last round of financing.

The sale price may of course reflect events that did not occur until after the last round of financing. Consequently, the use of the sale price is not a perfect indicator of the firm’s actual value at the last financing. But we assume such distortions are randomly distributed (or controlled for in the regressions below), allowing the sale price to serve as an unbiased estimate of a firm’s value at the time of the last financing.

For each firm we compare the valuation assigned to the firm in the last round of VC financing (last round valuation) to the sale price. Our dependent variable is the ratio (Last Round Valuation)/(Sale Price) or (‘LRV/SP”). The self-dealing theory for inside rounds suggests that inside rounds will be undervalued. By contrast, rescue financing predicts that LRV/SP will be systematically higher in Inside rounds as compared to Outside rounds.

Figure 4 illustrates the distribution of LRV/SP for inside rounds and outside rounds. Inside rounds often have an LRV/SP ratio exceeding 2 or 3. In contrast, in outside rounds LRV/SP is generally close to one. Consistent with the rescue financing theory, inside rounds appear to be overvalued relative to the ultimate sale price, while outside rounds are not.


13

To test the hypotheses in a multivariate setting, we estimate the following equation:

Log (LRV/SP) = F(Inside, Controls) (1)

Our dependent variable in Eq. (1) is the natural log of LRV/SP. We take the natural log of our dependent variable so that our data are not left-truncated at zero. Our treatment variable records whether the last round of financing was an Inside round. We also control for other factors that might affect the sale price and the last round valuation including:

VC Control (which equals one if the VC’s control a majority of the board seats going into the new round of financing, and zero otherwise. VC control of the board may enable them to push through a financing that is more favorable to them);

California (which equals one if the firm was incorporated in California, and zero otherwise. Compared to Delaware, California provides greater protection to minority shareholders, through separate class voting and threat of litigation (Broughman and Fried, 2010). Consequently, incorporation in California may limit a VC’s ability or willingness to push for a low valuation);

Δ NASDAQ Sale (%) (which equals the percentage change in NASDAQ from the date of financing to the date of sale. This controls for factors that might affect the sale price but could not be foreseen by the parties on the date of financing);

VC Reputation (which equals one if the VC firm leading the round of financing was formed prior to the median year of formation for the VCs in our sample (1990), and zero otherwise. This controls for possible investment discounts that may be awarded to VCs with a good reputation, as documented by Hsu (2004));

Duration Fin. to Sale (which equals the period of time (in years) between the last round of financing and the sale of the firm); and

Investment (the amount invested in the last round of financing (in millions)).

Regression results are shown in Table 4. Model (4-1) reports univariate results, and in models (4-2) and (4-3) we add the above control variables and industry dummies.


In each specification we find a positive correlation between inside rounds and the LRV/SP ratio. This result is significant in most specifications. Relative to the sale price, firms are given a significantly higher valuation in inside rounds than in outside rounds.

To get a sense of the magnitude of this effect, consider model (4-2) evaluated at the median value of LRV/SP = 1.60. In model (4-2) the coefficient estimate for Inside = 0.942. We can use this information to calculate the marginal effect of inside financing – x – given by the following:

Log(1.60 + x) = Log(1.60) +. 942 = 1.41 = Log(4.09) (2)

x = 4.09 – 1.60 = 2.49 (3)

From Eq. (3) we find that inside financing is associated with an increase of 2.49 in LRV/SP. Given that the median sale price for firms receiving inside financing in their last round is $11.5 million,

14

model (4-2) predicts that a firm’s nominal valuation will be about $29 million higher (= 11.5 * 2.49) if the last round of financing was an inside round as opposed to an outside round.

Inside financing has an economically (and statistically) significant effect on valuation. These results are consistent with predictions of our rescue financing theory that inside investors will tend to offer higher valuations than outside investors are willing to provide.

Our results could be biased by omitted variables that correlate with both the use of inside financing and high valuations. Ideally we would instrument for the use of inside financing. However, a good instrument is simply not available. We attempt to minimize this bias by looking at valuation relative to the firm’s ultimate sale price (LRV/SP). Omitted variables correlated with Inside are only a concern to the extent that they may cause the investors to incorrectly value the firm relative to its sale price.

Rescue financing predicts that existing VCs may choose to do a follow-on financing at a higher valuation than is available from an outside investor to reduce down-round costs. This higher value inside financing could in principle, be a down, even-, or (slightly) up-round. But inside rounds are more likely to be done to reduce down-round costs when they are done at a down or even valuation than when they are done at a higher valuation than the previous round. In model (4-4) we thus limit our analysis to the sub-sample of 33 firms where the last round of financing was a down or even round. The rescue financing theory for inside financing applies particularly to this subsample. Model (4-4) shows a similar coefficient for inside financing. Inside rounds are overvalued in the subsample of down & even rounds. The result is statistically different from zero at the 12% level, but not at conventional levels.16

Finally, we address the possibility that outliers drive our results. Given our small sample size, one or two observations could have a large affect on the coefficient estimates reported in table 4. To address this concern we re-estimate models (4-1) and (4-2) using robust regression (‘RREG’). This approach assigns weights to each observation in order to reduce the effect of outliers and then re-estimates on the weighted observations. Our results are reported in models (4-6) and (4-7). While the precise coefficient estimates do change, we find the same basic result: inside financing has a positive and significant affect on the relative valuation assigned to a startup firm. This suggests that our results are not driven by outliers.

D. Round Returns: Inside versus Outside

Finally, rescue financing predicts that inside rounds will tend to yield lower returns than outside rounds. Figure 5 shows the VCs’ internal rate of return (Round IRR) for each round of follow-on financing based on the time of financing. Inside rounds have a lower IRR over almost the entire time period.


Using multivariate regression, we estimate Round IRR as a function of inside financing. We use the same set of control variables found in models (4-2) through (4-5). Table 5 reports results.


16 The lack of statistical significance at the 10% level or better is likely due to the small number of firms (n=33) in the down and even subsample.

15

Inside rounds perform worse than outside rounds. This result is statistically significant in most models, and the effect is economically meaningful. Round IRR is estimated to be between .30 to .60 lower in inside rounds than in outside rounds.

It is important to emphasize that the lower returns in inside rounds does not necessarily reflect the full value of these rounds to VCs. First, Round IRR does not completely capture the value created when down round costs are reduced by doing an inside round at an inflated valuation. Second, Round IRR does not include the cross-subsidy that inflated-value financing creates for the VCs’ preexisting equity in the firm. If these two benefits were included, the relative returns generated by inside rounds would be higher.

E. Alternative Explanations for Inside Rounds

We do not claim (or believe) that every inside round is driven by rescue financing or self-dealing. There could be other reasons why VCs use inside rounds. Below, we briefly describe possible alternate explanations for inside rounds. While each of these explanations is consistent with some of the results reported above, none of them can explain the data as well as rescue financing.

Over-optimism/hubris. A firm’s existing VCs may, for psychological reasons, value the firm more than an outside VC. To the extent the existing VCs are unaware of this bias, they may turn down outside offers as “too low,” conduct inside rounds at valuations that are too high, and earn poor returns on these rounds. Hubris could explain why, as we find valuations in inside rounds are higher than in outside rounds and returns are lower. It could also help explain why inside rounds are more likely if VCs’ pre-round proportional interest is higher (the fact that VCs’ pre-round proportional interest is higher may suggest that the VCs are particularly fond of the firm). However, hubris could not easily explain why inside rounds increase as NASDAQ valuations decline, and why inside rounds tend to be disproportionately down and even, rather than up.

Cost of adding new investor. While there are substantial benefits to adding a new investor through an outside round (Admati & Pfleiderer, 1994; Hochberg, Ljungqvist, and Lu, 2007), there are also costs. These costs include identifying and recruiting potential investors and then integrating them into the governance fabric of the firm. In some cases, the costs may exceed the benefits, leading the existing VCs and entrepreneur to prefer an inside round. However, this cost explanation cannot explain why inside rounds increase when firm values decline, why inside rounds are more likely as existing VCs’ ownership interest goes up, and why inside-round valuations tend to be high and returns tend to be low.

Asymmetric Information. VC-backed firms may turn to inside rounds when there is an information-asymmetry problem (Myers, 1984). In particular, existing VCs may have difficulty credibly signaling to outside VCs the value of the firm, causing outside VCs to demand discounts and making outside rounds an expensive source of capital. Asymmetric information can explain our finding that outside rounds are conducted at more favorable valuations and generate higher returns

But we are skeptical that there is significant information asymmetry between existing and new VCs. First, VCs are repeat players and syndicate many deals (Sorenson & Stuart, 2001). They have strong incentives not to “cheat” each other. Second, outside VCs can and do protect themselves against asymmetric information by insisting that existing VCs participate in the new round (Admati and Pfleiderer, 1994). Third, startups often accept outside financing before existing VCs have run out of capital (Lerner, 1994).

16

Information asymmetry cannot explain many of our results. It cannot easily explain why inside rounds increase when values decline and VCs’ existing ownership is higher, and why inside rounds are disproportionately down or flat. Nor can it explain why inside rounds are more common in later rounds; one would expect information asymmetry to decline as a firm ages.

V. Conclusion

VC-backed startups typically invite outside VCs to negotiate and price the terms of follow-on financing rounds, in part to reduce conflicts between the entrepreneur and current VC investors over valuation. But many follow-on financings are structured as “inside” rounds -- only current VCs participate. One explanation for inside rounds is that they are used by VCs for self-dealing – to sell themselves stock at a low valuation.

We offer an alternative “rescue financing” explanation for inside rounds. Our explanation for inside rounds builds on the fact that there are various costs associated with a “down-round” financing -- a financing at a lower valuation than the previous round, including the costs associated with the renegotiation of anti-dilution rights, the demoralization of the firm’s employees, and writing down the VCs’ prior investment in the firm. We show that, because of the presence of such costs, existing VCs of a firm whose value has declined can benefit the firm and themselves by conducting an inside round at an inflated valuation rather than obtaining outside financing at a fair valuation.

Rescue financing generates a number of testable predictions that are at variance with those predicted by the self-dealing explanation for inside rounds. The rescue theory predicts that as actual values decline, and existing VCs’ ownership interest increases, firms are more likely to receive inside financing. It predicts that inside rounds will be disproportionately down or flat rounds. And it predicts that inside rounds are likely to be made at inflated valuations and associated with lower rates of returns than outside rounds. In contrast, the self-dealing theory predicts that inside rounds are likely to be used when existing VCs’ ownership is low and that inside rounds will be made at bargain valuations and associated with higher returns than outside rounds.

We find preliminary support for the rescue theory using a hand-collected dataset of 90 follow-on financing rounds from 45 Silicon Valley firms that were sold in 2003 or 2004. We first show that changes in exogenous factors likely to affect firm value influence the likelihood of a round being inside in the manner predicted by rescue financing. Our results suggest that a 20% decline in NASDAQ between financing rounds increases the likelihood by approximately 15% that the next financing will be structured as an inside round, and a 10% higher VC ownership prior to the new round is associated with a 40% increase in the likelihood of inside financing. We also find that inside rounds are disproportionately down and even rounds. Finally, we show that, in our sample, inside rounds tend to be done at higher valuations than outside rounds. The inside rounds in our sample on average also have a significantly lower rate of return – by about 30% to 40% - than outside rounds.

Our analysis contributes to the literature on the structure of VC financings. While the benefits of outside rounds has received considerable attention from theorists and empiricists, our paper provides an explanation for why, notwithstanding these potential benefits to outside rounds, VCs and entrepreneurs may prefer inside rounds, and offers evidence consistent with this explanation.

Our paper also contributes to a better understanding of governance within private VC-backed startups before they are sold. Little is known about how VCs actually exercise control within

17

these firms, including the potential for conflict around financing transactions. Our study provides preliminary evidence suggesting that, while self-dealing may occur, VCs generally do not use their control rights to force through inside financings that benefit them at the expense of entrepreneurs.

Our sample firms were located in a particular area, Silicon Valley, and exited through private sales at a period of time when values were generally declining. We cannot rule out the possibility that, in other places or other periods of time, inside rounds are less frequently used for rescue financing. And because all of our firms exited through private sale, we may not be able to extrapolate to firms that exit through IPO or are dissolved. Our data and analysis do, however, show that inside rounds may have benign motives and that in Silicon Valley during this period, inside rounds were more likely to reflect rescue motives than an attempt to exploit entrepreneurs in firms undergoing the most common form of exit: private sale.

Our work suggests a number of interesting avenues for future research. It would be worth investigating financings in VC-backed firms in locations outside Silicon Valley and in other time periods to determine whether our findings generalize to other settings. Similarly, it would be desirable to gather data on round valuations and round returns from firms that exited through IPOs or were dissolved. Finally, future research might seek to determine the extent to which each of the down-round costs – renegotiation, demoralization, or write-downs – drives rescue financing. We hope our study is of interest to regulators, courts, and practitioners seeking to better understand the dynamics of VC financing and corporate governance within VC-backed firms.

18

Appendix A: Formal Model of Rescue Financing

In this appendix we formalize the rescue financing intuition with a three-period financial

contracting model.

At Date 0, a wealth-constrained entrepreneur (“E”) forms a business (“Startup”) to pursue a new venture and obtains first-round (Series A) financing from a VC investor (“VC1”). E and VC1 divide monetary returns in a linear manner, to VC1 and to E, where .17

At Date 1, Startup requires second-round (Series B) financing so that it can continue operations and be sold for a price V at Date 2.18 Series B financing is provided either by (i) VC1 (an “inside round”), or (ii) an outside VC investor (“VC2”) (an “outside round”). The second round financing contract gives the investor – either VC1 or VC2 – a fraction of Startup’s equity in exchange for the investment of .

The price for which Startup can be sold at Date 2 is known at Date 1. It depends on the state of nature (s) revealed after Date 0 and before Date 1.19 We will shortly describe these down-round costs and the bargaining process by which the Series B investor is chosen.

At Date 2, Startup is sold to an acquirer and proceeds are distributed according to parties’ cash-flow rights. All parties are risk neutral. There is no discounting.

The time structure of the model is summarized in Figure A1.

Figure A1: Timeline

17 VCs typically hold convertible preferred stock (Kaplan & Stromberg, 2003) rather than common. The central point of our model, however, applies even if we relax the linearity assumption. As long as some portion of any surplus created by a new round of financing may be captured by an earlier round of VC investment, our basic intuition holds. The assumption of linearity makes it easier to specify how any gains from financing are distributed, but it is not necessary to create the basic dynamic.

18 For simplicity, we assume that absent such financing Startup will be worth $0.

19 Following the incomplete contracting literature, we assume the Date 1 state of nature is sufficiently complex such that it cannot be contracted over at Date 0 (Grossman and Hart, 1986). However, at Date 1 the parties can observe s, and the second-round financing contract is priced accordingly. This assumption lets us focus on the second-round contract. It is also a realistic description of staged contracting practices in VC-backed firms (Gompers, 1995).

Date 0

•First round financing

•State of nature is revealed

Date 1

•Second round financing

Date 2

•Firm sold. Cash flow rights paid out

19

Down-Round Costs and Startup’s Value

We now specify the potential down-round costs associated with the Series B financing and the price paid for Startup at Date 2. A down round occurs when the pre-money valuation at Date 1 is less than the post-money valuation at Date 0. On Date 0, after the Series A financing, the post-money valuation of Startup is . On Date 1, after the Series B financing, the pre-money valuation of Startup is . The Series B is thus referred to as:

an up-round if a down-round if an even-round if

We assume that there costs to a down-round, and that these costs may affect the sale price at Date 2.20 We define the sale price as

(A1)

Where is the value of Startup in state of nature, s, absent down-round costs, and the term represents costs associated with down-round financing. We assume is positive (and increasing in ) whenever the Series B is a down-round, but zero otherwise. Formally stated:

To ensure a unique interior solution, we assume d is continuous and differentiable in x,

for all

. We assume that (absent down-round costs) financing always

creates a surplus, and thus that for all states of nature.21 For ease of notation we sometimes refer to simply as V.

Series B Bargaining Process

To predict when inside financing will occur in the Series B and on what terms we specify the bargaining process. Startup’s board controls the bargaining process and seeks to maximize value for its existing equity holders.22 Startup first solicits a financing offer from VC2. If VC2 offers financing, Startup seeks a better offer from VC1. If VC2 declines to invest, Startup seeks an offer from VC1.

Following the financial contracting literature (Aghion and Bolton, 1992) we assume perfect competition among outside investors at Date 1. Consequently, VC2 will offer a contract if there is a valuation that enables it to yields zero profit in expectation. VC1 will make a counteroffer (or offer, if VC2 declines) only if doing so makes VC1 better off. We consider three cases: (1) VC2 offers

20 Unlike the numerical example in the text, which assumes that down-round costs are borne directly by the existing VCs, we assume that the down-round costs are borne directly by Startup and thus only indirectly by the existing VCs (here, VC1). This assumption does not materially affect the analysis.

21 Startup is worth $0 if it does not receive second round financing. Thus, the difference between and can be seen as the social surplus of second round financing.

22 Our results do not depend on the allocation of board seats, or other control rights, at Startup. The results would be identical if VC1 or Startup’s other shareholders controlled the bargaining process, as long as Startup could not be forced to accept worse terms from VC1 than would be offered by VC2.

20

up/even round financing; (2) VC2 offers down-round financing; and (3) VC2 declines to offer any financing.

Case 1: VC2 Offers Up/Even Round Financing

There are no down round costs [d(x) = 0] when VC2 offers financing at an Up or Even valuation. Given perfect competition, VC2 will set the post-money valuation at Date 1 equal to . VC2 will offer to invest for cash flow rights ( ) such that .

Proposition 1: VC2 will offer an Up or Even round if and only if .

Proof: Assume . By substituting for we see that , which is by definition an Up round or an Even round. By contrast assume . Substituting gives us , which by definition is a down round.

Proposition 2: If VC2 offers financing at an Up or Even valuation, VC1 will not make a counteroffer.

Proof: If VC1 were to make a counteroffer it would be at a higher valuation23 [ VC2’s proposal would give VC1 a payoff equal to If VC1 were to make a counteroffer it would receive Proof by contradiction: assume VC1’s net payoff from inside financing is greater than its payoff if VC2 provides outside financing. It follows that:

(A2)

Substituting for into Eq. (A2) and combining terms we find that:

(A3)

Eq. (A3) is a contradiction since and Thus, we conclude that VC1 will not propose a counteroffer if VC2 offers financing at an Up or Even valuation.

Remark: Proposition 2 shows that when VC2 offers to invest at an Up or Even valuation, VC1 cannot improve its position by making a counteroffer. The basic intuition is straightforward. Because of perfect competition, the surplus from an outside round is divided entirely between VC1 and E – Startup’s preexisting equity holders. If VC1 were to offer a higher valuation, it would receive a negative expected return on its Series B investment. To be sure, a higher valuation would increase the value of VC1’s Series A investment. However, the increase in the value of the Series A investment would always be less than the overpayment for the Series B shares because E captures some of the value transferred to current shareholders in the overpriced round, and a higher valuation does not reduce down-round costs. The net return to VC1 from offering a higher price would be negative.

Case 2: VC2 Offers Down-Round Financing

VC2 takes into account the cost of down round financing [d(x) > 0] when pricing its offer. In this case, VC2 will set post-money valuation at Date 1 equal to . VC2 will offer to invest for cash flow rights ( ) such that .

Proposition 3: VC2 will offer a down round whenever and there exists such that

Proof: Assume and there exists such that Since it follows that , which is by definition an Down round. From Proposition 1 we

23 We assume startup will not accept a lower valuation.

21

know that if VC2 will offer an Up or Even round. And if there is no such that VC2 will not offer financing at all.

When VC2 offers a Down round, VC1 may propose a counteroffer, depending on the magnitude of down round costs. Because of its Series A interest, VC1 can participate at a higher valuation than VC2, and when down round costs are large, VC1 can increase its total payoff by making a counteroffer at a higher valuation.

To see this, let represent VC1’s combined payoff – from the Series A and Series B equity – if it proposes a counter-offer at Date 1:

(A4)

By definition h is a concave differentiable function over the range [0, 1].24 Thus, we can solve for which maximizes by taking the first derivative and solving for zero. Let

be defined such that:

(A5)

The cash flow rights represented by yields the highest payoff for VC1.

This allocation of cash-flow rights is important since it lets us identify when VC1 will make a counteroffer and on what terms. If VC2 demands cash flow rights greater than

, VC1 will make a counteroffer at a post-money valuation equal to

. The valuation represented by

effectively defines a tipping point. It is the point at which a change in valuation has the same marginal effect on VC1’s existing equity as its new equity.

Proposition 4: If VC2 offers Down-round financing, VC1 will:

i. make a counteroffer if and only if , and

ii. if VC1 proposes a counteroffer it will set valuation equal to .

Proof of part (i): VC1’s payoff function, h, is concave and takes its maximum value at . If

VC1 would decrease its payoff by offering to invest at a higher valuation. If

VC1 would increase its expected payoff by offering to provide inside financing in exchange for cash-flow rights in the range

. Both results follow immediately from the concavity of h.

Proof of part (ii): This counteroffer valuation follows immediately from VC1’s payoff function. Provided

, technically VC1 could improve its position by offering to provide inside financing in exchange for any cash-flow rights in the range

By definition, however, the point would give VC1 the highest

payoff, and since this is also the highest valuation for Startup there is no reason to invest on any other terms.

Remark: Proposition 4 specifies when VC1 will make a counteroffer and the terms of inside financing. In some down-round settings [when

] VC1 cannot improve its payoff by offering to invest at a higher valuation. Our analysis predicts that Down-rounds led by an outside investor will sometimes occur. However, in other down-round settings [when

] VC1 can improve its payoff by proposing a counteroffer at a higher valuation. Our analysis predicts that inside financing is more likely to occur as the costs of down-round financing increase.

24 To see this note that and by assumption . It

follows that .

22

Case 3: VC2 Declines to Offer Financing

If the costs of down round financing are sufficiently large, there may be no allocation of cash-flow rights that would give VC2 its required return. VC2 will refuse to invest if there is no such that

Even though VC2 declines to offer financing, VC1 may be able to provide inside financing. This is possible because VC1’s Series A claim – worthless without Series B financing – captures some of the value created by second round financing. VCV1 can invest at higher valuations and avoid some down-round costs that an outsider faces.

Proposition 5: Inside financing is feasible even though outside financing is not whenever

for all .

Proof: By definition of the expression

for all . It follows that there exists such that

for all Consequently, inside financing may be feasible even though outside financing is not.

Predictions.

Our model leads to several testable predictions. First, our model predicts that a firm will not receive inside financing if its value has not declined since the last financing round. However, if an outside investor proposes a down-round the firm’s insiders may be willing to make counteroffer at a higher valuation. Thus, inside rounds are more likely as values decline.

Second, our model predicts that VCs are more likely to do an inside round when their preexisting ownership percentage (π) is higher. Furthermore, the extent of overpricing in inside rounds increases when the existing VCs have a larger preexisting ownership fraction (Prop. 4).

Third, our model predicts that the valuations used in inside rounds will be different than in outside rounds. Inside investors are willing to overpay for stock to reduce the costs a down-round financing imposes on their existing interests in the firm. Thus, when firms’ values have declined, the valuations assigned to firms in inside rounds will thus be higher, everything else equal, than the values assigned to firms in outside rounds.

Fourth, inside rounds will have a lower rate of return than outside rounds. This prediction is of course related to the second prediction – that inside rounds will occur at inflated valuations. If valuations used in inside rounds are too high, the returns on these rounds should be lower.

23

Appendix B: VX Shadow Sample

The following appendix describes the VX shadow sample and compares it to the sample firms used in the paper (the “main sample”). VX does not provide sufficient data on round valuations or returns to test most of the rescue financing predictions directly in the shadow sample. Nonetheless, we find evidence in the VX shadow sample that mirrors the results from the main sample and is broadly consistent with the use of inside rounds for rescue financing.

Before proceeding, we highlight four key results:

1. Inside Round Frequency. The frequency of inside rounds in the shadow sample is similar to that in the main sample, particularly when limited to California-based firms.

2. NASDAQ and Inside Rounds. As in the main sample, regression analysis indicates that the likelihood of an inside round increases when NASDAQ has declined since the prior round.

3. Clustering of Inside Financing in Down and Even Rounds. As in the main sample, inside financing in the VX shadow sample is clustered in down rounds and even rounds.

4. Inside Round Performance. The frequency of inside rounds is lower among VX sample firm that have exited via IPO or private sale than among defunct or active firms. To the extent exit outcome is a proxy for VC returns, this suggests that inside rounds are associated with a lower rate of return than outside rounds.

A. VX Shadow Sample Description

The VX shadow sample is composed of VC-backed firms randomly selected from all US firms in VentureXpert (VX) that were founded between 1/1/1996 and 12/31/2002, received at least $5M in financing, and disclosed valuation data for at least one round of financing. 4070 firms in VX met the above criteria; we selected every 40th firm from the VX database, giving us a sub-sample of 102 firms. Unlike in the main sample, which consists entirely of California-based firms that exited through private sale, the shadow sample covers a broad range of VC-backed firms in and outside California with a variety of exit outcomes: IPO, private sale, defunct, or still active (no exit yet).

Most of the VX shadow sample firms (94 out of 102) received follow-on financing. We exclude follow-on rounds that occur within 3 months of the previous financing; even though they are reported as “rounds” in VX they are unlikely to be separately negotiated (Bengtsson & Sensoy, 2009; Cumming & Dai, 2010). We also exclude rounds that do not include a VC investor. We are left with a total of 312 follow-on rounds in the VX shadow sample. We coded the round status – inside or outside – for each of the 312 follow-on rounds.25 There were 186 (60%) outside rounds and 126 (40%) inside rounds.

Table B1 provides firm-level statistics for the businesses in the VX shadow sample. The firms are primarily in the internet, medical/health, and software sectors (Panel A of Table B1). The concentration in high-tech businesses is similar to the main sample. However, industry classification for the main sample is provided by LinkSV while industry classification for the shadow sample is provided by VX. Industry definitions may vary between the two sources. Approximately 30% (31/102) of the VX shadow sample firms are in the medical/biotech sectors, versus 13%

25 A round is defined as “outside” if at least one of the VCs participating in the new financing did not invest in any of the firm’s prior rounds of financing, and “inside” if there are no new VCs investing in the new round.

24

(6/45) of the firms in the main sample. But we have no reason to believe that rescue financing applies any differently to medical/biotech firms than to firms in other industries.26

As of June 2010, VX shadow sample firms received on average $50 million in VC funding, compared to $44 million in the main sample.27 The mean shadow sample firm received initial financing in 1999, while the mean main sample firm received initial financing in 1998.

Table B1: Shadow Sample Firm Data (n=102)

Panel A: Industry Distribution of Shadow Sample Firms

Internet Specific 33

Medical/Health 22

Computer Software 17

Biotechnology 9

Communications 9

Other IT 12

Panel B:Overview

Shadow Sample Main Sample

Variable Obs. Count Mean Obs. Count Mean

Total Invested ($M) 102 50.477 45 43.999

Number of Rounds 102 4.069 45 3.00

Year of First Financing 102 1999.402 45 1998.38

California Location 102 47 .461 45 45 1

IPO 102 14 .137 n/a n/a

Acquisition 102 40 .392 45 45 1

Active 102 28 .275 n/a n/a

Defunct 102 20 .196 n/a n/a

While all of the firms in the main sample were based in California, slightly less than half of the VX shadow sample firms (46%) were located in California. As documented by Bengtsson & Ravid (2009), there are several differences between VC contracting practices in California and the rest of the US. In particular, VCs’ cash flow rights (liquidation preferences, participation rights, etc) tend to be “harsher” (less friendly to the entrepreneur) outside California. In the VX shadow sample, California-based firms have fewer inside rounds (see Table B2 below) and are more likely to reach an IPO (19% versus 9%) than firms located outside California.

Table B1 reports data on the current exit status – IPO, Acquisition, Active, or Defunct – for the shadow sample firms. Just over half the shadow sample firms had achieved an exit event by mid-

26 In the VX shadow sample, firms in the medical (40.2%) and biotech (40.9%) industries use inside rounds with the same frequency as firms from other industries (40.4%).

27 Some of the VX shadow sample firms – those that are still “active” – may receive additional financing in the future.

25

2010, either an IPO (14%) or private sale (39%). The remaining firms are either defunct (20%) or remain active (27%). Since all firms in the shadow sample received initial financing in 2002 or earlier, the “active” firms are likely to be “walking dead” or “zombie” firms that will not yield large returns to the VCs. Since the main sample is limited exclusively to acquired firms, we cannot compare the distribution of exit outcomes between the two samples. The VX shadow sample, however, does provide information about the frequency of inside rounds among firms with exit outcomes other than private sale.

B. Shadow Sample Financing Rounds

The shadow sample includes 312 follow-on rounds, of which 126 (40%) are inside rounds and 186 (60%) are outside rounds. Firms in the shadow sample receive more follow-on and more inside rounds than the main sample. Shadow sample firms received on average 3 follow-on rounds, compared to 2 in the main sample (Table B1).

Table B2 shows the frequency of inside and outside financing in the VX shadow sample and the main sample sorted across a number of dimensions.

Table B2: Shadow Sample Follow-on Round Data (n=312)

VX Shadow Sample Main Sample

Inside Outside Inside Outside

(n= ) # (%) # (%) (n= ) # (%) # (%)

Full Sample 312 126 40% 186 60% 90 26 29% 64 71%

Sorted by State

CA 143 47 33% 96 67% 90 26 29% 64 71%

Non-CA 169 79 47% 90 53% n/a

Sorted by Exit Status

IPO 48 10 21% 38 79% n/a

Acquisition 118 40 34% 78 66% 90 26 29% 64 71%

Active 103 57 55% 46 45% n/a

Defunct 43 19 44% 24 56% n/a

Sorted by Valuation

Up 73 9 12% 64 88% 45 2 4% 43 96%

Down 26 8 31% 18 69% 31 15 48% 16 52%

Even 13 7 54% 6 46% 14 9 64% 5 36%

Missing 200 102 51% 98 49% n/a

Sorted by Round

2 94 21 22% 73 78% 45 9 20% 36 80%

3 75 32 43% 43 57% 28 7 25% 21 75%

4 55 25 46% 30 55% 13 8 62% 5 38%

5+ 88 48 55% 40 46% 4 2 50% 2 50%

26

C. Key Results

We now describe the four key results from the VX shadow sample.

Result 1: Inside Round Frequency

The frequency of inside rounds in the shadow sample is similar to that in the main sample. In the main sample, which consists entirely of California-based firms, the frequency of inside rounds is 29%. In the VX shadow sample, the frequency of inside rounds in California-based firms is similar, 33%. This suggests that the firms in the main sample were not atypical of California-based VC backed firms. The overall frequency of inside rounds in the VX shadow sample is 40%, reflecting the fact that that inside rounds were more common (47%) in VC-backed firms based outside of California (Table B2).

Result 2: Inside Rounds and NASDAQ

As in the main sample, firms in the VX shadow sample are significantly more likely to rely on inside financing when NASDAQ– a rough proxy for startups’ value – has declined since the prior round. Figure B1 displays, for each year in the period 1997-2004, the frequency of inside and down round financing in both the shadow sample and the main sample as well as the NASDAQ index. The top portion graphs weekly NASDAQ closing prices from1997 up to 2004. The middle and bottom portion illustrate the frequency of inside and down rounds in the main sample (middle portion) and shadow sample (bottom portion) over the same time period. In both samples, as NASDAQ rose in the late 90s inside rounds and down rounds decreased in frequency; when NASDAQ began falling in 2000, the frequency of inside rounds and down rounds increased.

As in the main sample, regression analysis suggests that the likelihood of inside financing increases when NASDAQ has declined since the prior round of financing. VX provides the date of each of financing for each round. We can thus re-estimate model (3-1) in Table 3 using the VX shadow sample. We find that Δ NASDAQ (% change in NASDAQ since last financing) has a coefficient estimate of -.598 (significant at the 1% level) in the shadow sample, similar to the estimate produced by the model using the main sample.

We reported in Table 3 that, consistent with rescue financing, inside rounds are more likely when VCs’ ownership stakes are larger. We cannot use the VX data to directly test whether inside rounds are more likely when VCs own more of the firm’s equity because data on VC ownership is not provided. However, it is worth noting that inside rounds in the VX sample are more common in later rounds, when VCs are likely to own a greater fraction of the firm’s equity. Table B2 reports that only 22% of 2nd round financings were inside rounds, compared to 43% of 3rd round financings, 46% of 4th round financing, and 55% of 5th or later round financings, similar to the firms in the main sample. To the extent that stage of financing acts as a proxy for VC ownership percentage, the shadow sample results are consistent with rescue financing: as VC ownership increases, the firm is more likely to use inside financing.

27

Figure B1: Shadow Sample Inside Financing by Year

Result 3: Clustering of Inside Financing in Down and Even Rounds.

As in the main sample, inside financing in the VX shadow sample is clustered in down rounds and even rounds. In the main sample, approximately 92% of inside rounds were down or flat (vs. 33% of outside rounds). In the VX shadow sample rounds for which sufficient data are available, 64% of inside rounds are down or flat (vs. 37% of outside rounds). However, a disproportionate number of inside rounds are missing valuation data (Kaplan, Sensoy & Stromberg, 2002). Poorly performing firms may be less likely to report valuations to VX. We believe that the percentage of inside rounds that were down or flat would be even higher in the VX shadow sample if all of the data were available.

Figure B2 below shows graphically the relationship between the use of inside and outside rounds and valuation changes since the last round of financing in those VX shadow sample firms for which sufficient valuation data are provided. The table under Figure B2 confirms that, as in the main sample, inside rounds and outside rounds have different valuation-change distributions. Using a t-test for equal means we find that inside rounds in the VX shadow sample are significantly more likely to be Even rounds and less likely to be Up rounds.

Figure B2 also indicates that, as in the main sample, shadow sample firms are unlikely to do slightly down-round financings. Only 5 of the 26 Down round financings occur at a valuation

10

00

30

00

50

00

NA

SD

AQ

0.1

.2.3

.4.5

.6.7

Main

Sam

ple

0.1

.2.3

.4.5

.6.7

VX

Shadow

Sam

ple

1997 1998 1999 2000 2001 2002 2003 2004Year

Lowess Fit Fraction of Inside Rounds

Lowess Fit Fraction of Down Rounds

28

between 0% to 25% below the previous round. On the other hand, there are 13 rounds of financing done at an exactly even valuation, more than half of which are insider rounds. The relative infrequency of slightly down rounds and the clustering of many rounds at an even valuation provides evidence that there are costs to down-round financing.

Figure B2: Inside Rounds and Valuation Changes in the Shadow Sample

Shadow Sample: Follow-on Rounds with non-Missing Valuation Data

(n= ) Up Even Down

All Rounds 112 65% 12% 23% Outside rounds 88 73% 7% 20% Inside rounds 24 38% 29% 33% t-test for Equal Means -3.34*** 3.13*** 1.32 * = 10% significance; ** = 5% significance; *** = 1% significance

05

10

15

Num

ber

of

Rounds

-100 -75 -50 -25 Even 25 50 75 100 150 200 300 400 500

Percent Change in Valuation

Outside Rounds

Inside Rounds

29

Result 4: Inside Round Performance

In the main sample, inside rounds are associated with lower returns (Table 5). While VX does not provide information on round returns, it does indicate the exit status of the firm: IPO, private sale, defunct, or active (no exit yet). To the extent that IPOs and M&As are associated with higher rates of return for VC investors than active (i.e. walking dead) or defunct firms, we can use exit outcome as a proxy for VC performance. We find that shadow sample firms receiving inside financing are less likely to make it to a successful exit – an IPO or an M&A. Table B2 reports that IPO firms receive 21% of their follow-on financings as inside rounds. Acquired firms receive 34% of their follow-on investments as inside rounds. This compares to 55% for active firms and 44% for defunct firms. While we cannot measure IRR directly for the shadow sample firms, the clustering of inside rounds in firms that have bad exit outcomes is consistent with the use of inside rounds for rescue financing.

30

Bibliography

Admati, A. P. Pfleiderer, 1994, "Robust Financial Contracting and the Role of Venture Capitalists", Journal of Finance,

49, 371–402.

Aghion, P., Bolton, P., 1992. An incomplete contracts approach to financial contracting. Review of Economic Studies

59, 473-494.

Atanasov, V., Ivanov, V., Litvak, K., 2008. The Effect of Litigation on Venture Capitalist Reputation, Available at SSRN: http://ssrn.com/abstract=1120994

Barry, C., Muscarella, C., Peavy III, J., Vetsuypens, M., 1990. The role of venture capital in the creation of public companies: evidence from the going public process. Journal of Financial Economics 27, 447-471.

Bartlett, R, 2003. [add]

Bartlett, R., 2006. Venture capital, agency costs, and the false dichotomy of the corporation. UCLA L. Rev. 54:37–115.

Bartlett, Joseph M. and Kevin R. Garlitz, 1995, Fiduciary duties in burnout/cramdown financings, Journal of Corporation Law 20, 595-626.

Bebchuk, L.A., Cohen, A., 2003. Firms’ decisions where to incorporate. Journal of Law and Economics 46, 383-425.

Bengtsson, O. & Sensoy, B. 2009 Changing the Nexus: The Evolution and Renegotiation of Venture Capital Contracts,

SSRN: http://ssrn.com/abstract=1475347.

Bengtsson, O., Ravid, A. 2009. The importance of geographical location and distance on venture capital contracts.

Available at SSRN: http://ssrn.com/abstract=1331574

Black, B., Gilson, R., 1998. Venture capital and the structure of capital markets: banks versus stock markets. Journal

of Financial Economics 47, 243-277.

Brander, J., R. Amit, and W. Antweiler. 2002. “Venture Capital Syndication: Improved Venture Selection vs. Value-

Added Hypothesis.” Journal of Economics and Management Strategy, Vol. 11 (2002), 422 − 452.

Burnham, B. (2004), “How to Hide a VC Write-Down”

http://billburnham.blogs.com/burnhamsbeat/2004/07/how_to_hide_a_v.html

Cochrane, J., 2005. The risk and return of venture capital. Journal of Financial Economics 75(1), 3-52.

Cumming, D., 2008. Contracts and exits in venture capital finance. Review of Financial Studies 21, 1942-1987.

Cumming, D., Dai, N., 2010. Why do entrepreneurs switch venture capitalists? Available at SSRN:

http://ssrn.com/abstract=1425620.

Daines, R., 2002. The incorporation choices of IPO firms. New York University Law Review 77, 1559-1610.

Forbes, Dl, M. Korsguaard, and H. Sapienza, Board Conflict in the Aftermath of a Market Crash (working paper,

2005).

Garland, R., 2009., “Later-Stage Valuations Tumble as VCs Reset Expectations”

http://blogs.wsj.com/venturecapital/2009/06/01/later-stage-valuations-tumble-as-venture-investors-reset-

expectations/

Gilson, R., 2003. Engineering a Venture Capital Market: Lessons from the American Experience, STAN. L. REV. 55,

1067 - ___.

Gompers and Lerner, 2004. The Venture Capital Cycle. MIT Press, Cambridge MA.

Gompers, P., 1995. Optimal investment, monitoring, and the staging of venture capital. Journal of Finance 50, 1461–

1489.

Gompers, P., 1996. Grandstanding in the venture capital industry. Journal of Financial Economics 42, 133–156.

Hellmann, T., 1998. The allocation of control rights in venture capital contracts. The RAND Journal of Economics

29(1), 57-76.

Hellmann, T. and M. Puri. 2002. Venture capital and the professionalization of startup firms: empirical evidence.

Journal of Finance, 57, 169–197.

31

Hermalin, B., Weisbach, M., 2003. Boards of directors as an endogenously determined institution: a survey of the

economic literature. Federal Reserve Bank of New York Economic Policy Review 9, 7–26.

Hochberg, Y., A. Ljungqvist, and Y. Lu. “Whom You Know Matters: Venture Capital Networks and Investment

Performance.” Journal of Finance, Vol. 62 (2007), 251 − 301.

Hsu, D., 2004, What do entrepreneurs pay for venture capital affiliation? Journal of Finance 59, 1805-1844.

Kaplan, S., Sensoy, B., Strömberg, P., 2002. How well do venture capital databases reflect actual investments?

Working paper, University of Chicago .

Kaplan, S., Strömberg, P., 2003. Financial contracting theory meets the real world: an empirical analysis of venture

capital contracts. Review of Economic Studies 70, 281–315.

King, K. 1997. Warning: Rescue may Raise Risks. The National Law Journal, Nov. 24. 1997

Klausner, M., Litvak, K., 2001. What economists have taught us about venture capital contracting. In: Whincop, M.,

(Ed.), Bridging the Entrepreneurial Financing Gap: Linking Governance with Regulatory Policy, Ashgate.

Lee, P., Wahal, S., 2004. Grandstanding, certification and the underpricing of venture capital backed IPOs. Journal of

Financial Economics 73, 375-407.

Lemon, J. Don’t Let Me Down (Round): Avoiding Illusory Terms in Venture Capital, Texas Journal of Business Law,

Vol. 39, 1-50.

Lerner, J, 1994, “The Syndication of Venture Capital Investments.” Financial Management, Vol. 23(3) 16 − 27.

Lerner, J., 1995. Venture capitalists and the oversight of private firms. Journal of Finance 50, 301–318.

Mahoney, D. 2002. Down Round Financing,: How to Cope With Lower Valuations for Your Client Company, Business

Law Today.

Megginson, W., Weiss, K., 1991. Venture capital certification in initial public offerings, Journal of Finance 46, 879–

903.

Metrick, A., 2007. Venture Capital and the Finance of Innovation, John Wiley and Sons, Inc.

Myers, S. 1984. The capital structure puzzle. Journal of Finance 39, 575-592.

Sahlman, W., 1990. The structure and governance of venture capital organizations. Journal of Financial Economics

27, 473–521.

Smith, D.G., 1998. Venture capital contracting in the information age, Journal of Small & Emerging Business Law 2,

133-___.

Sorenson, O., and T. Stuart. “Syndication Networks and the Spatial Distribution of Venture Capital Investments.”

American Journal of Sociology, Vol. 106 (2001), 1546 − 1688.

Staiger, D. and Stock, J., 1997. Instrumental Variables Regression with Weak Instruments, Econometrica, Vol. 65(3),

557-586

Suchman, M., Cahill, M., 1996. The hired gun as facilitator: lawyers and the suppression of business disputes in

Silicon Valley. Law & Social Inquiry 21, 679-712.

Tian, X. (2008) “The Role of Venture Capital Syndication in Value Creation for Entrepreneurial Firms” on SSRN at

ssrn.com/sbatract=054188

Utset, M., 2002. Reciprocal Fairness, Strategic Behavior & Venture Survival: A Theory of Venture Capital- Financed

Firms, WIS. L. REV. 2002, 45-__.

VentureOne Deal Terms Report, 3rd Edition (Dow Jones, 2005)

VentureSource, 2010 [table reporting frequency of inside financing].

Woronoff and Rosen, 2005 … Effective vs. Nominal Valuations in Venture Capital Investing

32

Table 1: Description of the Dataset This table provides descriptive statistics for a sample of 45 Silicon Valley VC-backed firms sold in 2003 or 2004. Panel A shows industry distribution, based on sector classification provided by www.linksv.com. Panel B reports summary statistics for the for the sample firms

Panel A: Industry Distribution of Companies

Sector Biotech/

Medical Telecom Software Internet Other IT Sample Firms (n=45) 6 11 11 9 8

Panel B: Firm Overview

Obs Count Mean Med. SD Years of Operation 45 5.13 5 1.65 Year Founded 45 1998.38 1998 1.77 Number of Financing Rounds 45 3.00 3 1.01 Total Invested (millions $) 45 43.99 31 37.65 Sale Price (millions $) 45 55.22 22 108.95 Firm IRR 45 0.20 -0.13 2.03 Profitable 45 16 0.36 0 0.48 Number of Founders 45 2.31 2 1.24 Serial Entrepreneur 45 20 0.44 0 0.50 Prior Experience with VC 45 23 0.51 1 0.51 California Incorporation 45 14 0.31 0 0.47 Delaware Incorporation 45 31 0.69 1 0.47

http://www.linksv.com/

33

Table 2: Variable Definitions and Correlation Matrix This table defines the round-level variables for a sample of 45 Silicon Valley VC-backed firms sold in 2003 or 2004. The following correlation matrix shows the mean, and pairwise correlations among the variables in our sample. Data are only presented for 90 follow-on rounds of financing. First round financing is excluded since it does not apply to many of the variables below.

VC Control equals 1 if the VCs control the board immediately preceding a round of financing and 0 otherwise; Inside equals 1 if a new round of financing is provided entirely by a firm’s existing investors and 0 otherwise; VC * Inside is an interaction term equal to VC Control * Inside; Participation Right equals 1 if a round of financing includes participation rights for the VC’s preferred stock and 0 otherwise; LP multiple equals the liquidation preference multiple (i.e. 1x, 2x, etc.) in the current round of financing; Up equals 1 if a round of financing has a higher pre-money valuation than the post-money valuation of the previous round and 0 otherwise; Down equals 1 if a round of financing has a lower pre-money valuation than the post-money valuation of the previous round and 0 otherwise; Flat equals 1 if a round of financing has the same pre-money valuation as the post-money valuation of the previous round and 0 otherwise; Valuation Change equals the pre money valuation of the current round minus the post money valuation of the previous round (recorded on an implied basis); Investment equals the amount invested (in millions) in a round of financing; Round IRR equals the internal rate of return for each round of VC financing; Post Bubble Round equals 1 if a round occurred in 2001 or later and 0 otherwise; Round Number equals the number of the current round; Duration Financing to Sale equals the number of years between the current round of financing and the sale of the firm; Last Financing Round equals 1 if the current round is a firm’s last round of financing before the sale of the business.

Mean 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 VC Control .23 - 2 Inside .29 .23 - 3 VC*Inside .10 .60 .52 - 4 Up .50 -.13 -.54 -.26 - 5 Down .34 .15 .31 .15 -.72 - 6 Even .16 -.02 .34 .16 -.43 -.31 - 7 Valuation Change -6.93 -.04 -.35 -.11 .58 -.48 -.17 - 8 Investment 16.46 -.06 -.20 -.13 .27 -.16 -.16 .34 - 9 LP Multiple 1.44 .22 .41 .35 -.42 .43 .01 -.55 -.08 -

10 Participation Rights .60 -.03 .12 .12 -.14 .16 -.03 -.20 .17 .22 - 11 Round IRR -.04 -.01 -.01 .00 -.09 .03 .07 -.28 -.11 .14 .14 - 12 Post Bubble Rd. .60 .18 .27 .20 -.50 .50 .04 -.46 -.22 .36 .17 .33 - 13 Round Number 2.74 .22 .27 .18 -.24 .38 -.16 -.30 .18 .19 .13 .08 .26 - 14 Duration Fin to Sale 2.63 -.27 -.26 -.25 .35 -.40 .04 .35 .03 -.35 -.15 -.44 -.75 -.34 - 15 Last Financing Rd. .49 .14 .26 .12 -.44 .51 -.05 -.44 -.19 .37 .25 .46 .66 .34 -.69 -

34

Figure 1: Inside versus Outside Financing sorted by Round Number Figure 1 illustrates the use of inside and outside financing sorted by the round number. The data are from a sample of 45 VC-backed firms sold in 2003 or 2004, covering 90 follow-on rounds of investment. This graph illustrates that inside rounds are more likely if the firm has already received several rounds of financing.

010

20

30

40

2nd Round 3rd Round 4th Round 5th Round

Number of Inside and Outside Financings for each Round

Outside Rounds Inside Rounds

35

Figure 2: Inside Financing by Year Figure 2 illustrates the use of inside rounds and down rounds based on the time of financing. The data are from a sample of 45 VC-backed firms sold in 2003 or 2004, covering 90 follow-on rounds of investment. The top portion of the diagram graphs the NASDAQ weekly closing price from the beginning of 1997 to the end of 2003. The bottom portion of the diagram uses a Lowess curve to plot the likelihood of inside financing and down-round financing over the same time period. Below the graph is a table reporting the fraction of the follow-on rounds that were inside and down-rounds for each year.

Year of Financing

≤1998 1999 2000 2001 2002 2003

Observation 6 12 18 24 19 11

Inside 0.33 0.00 0.17 0.29 0.47 0.45

Down 0.17 0.00 0.06 0.42 0.63 0.55

10

00

20

00

30

00

40

00

50

00

NA

SD

AQ

0.1

.2.3

.4.5

.6.7

Late

r R

ound F

inancin

g (

%)

by Y

ear

1997 1998 1999 2000 2001 2002 2003 2004Year

Lowess Fit Fraction of Inside Rounds

Lowess Fit Fraction of Down Rounds

36

Figure 3. The Relationship between Inside Rounds and Valuation Changes

Figure 3 illustrates the relationship between the use of inside rounds and valuation changes since the last round of financing. The data are from a sample of 90 follow-on rounds of financing from 45 VC-backed firms sold in 2003 or 2004. The horizontal axis indicates the amount of valuation change since the last round. “Even” means no change in the valuation. “-25” means that the valuation change was negative but not exceeding -25%; “25” means that the valuation change was positive but not exceeding 25%. The table below Figure 3 reports the percent of outside and inside rounds that are Up, Even, and Down. The table includes a difference of means t-test comparing the distribution of inside and outside rounds.

Follow-on Rounds (n= ) Up Even Down All Follow-on Rounds 90 50% 16% 34% Outside rounds 64 67% 8% 25% Inside rounds 26 8% 35% 58% t-test for Equal Means -6.01*** 3.34*** 3.08*** * = 10% significance; ** = 5% significance; *** = 1% significance

05

10

15

Num

ber

of

Rounds

-100 -75 -50 -25 Even 25 50 75 100 150 200 300 400 500

Percent Change in Valuation

Outside Rounds

Inside Rounds

37

Table 3: Market Conditions and the Use of Inside Rounds This table reports Logit regressions on a sample of 45 VC-backed firms sold in 2003 or 2004, covering 90 follow-on rounds of financing. The dependent variable (Inside) equals one if the round was an inside round, and zero otherwise. Our treatment variables measure (i) the percent change in NASDAQ market prices from the previous round to the current round of financing (Δ NASDAQ), and (ii) the percent of equity held by existing VCs prior to the new round of financing (VC Ownership %). All other explanatory variables are defined in Table 2. The unit of analysis is the follow-on round of financing, and consequently there are multiple observations for several firms. Robust (White, 1980) standard errors, clustered at the firm level, are reported in parentheses below each coefficient estimate. We use a two-sided test for statistical significance. Logit (3-1) (3-2) (3-3) (3-4)

Treatment Variables Δ NASDAQ -.694* -.791 -.877*

(.401) (.482) (.479)

VC Ownership % 3.271** 5.419*** 6.177*** (1.309) (1.869) (2.295)

Control Variables VC Control 1.402 1.391 (.932) (.902)

Investment in Round -.081** -.079** (.034) (.033)

Number of Directors -.206 -.310 (.232) (.261)

Industry Dummies Y

Constant -1.179 -2.897 -2.530 -3.069

Wald Chi2 2.99 6.25 18.03 20.14 Prob>Chi2 0.083 0.012 0.003 0.017 No. of Observations 90 90 90 90 No. of Firms (clusters) 45 45 45 45

*= 10% significance; **=5% significance; ***=1% significance [2-sided test]

38

Figure 4: Last Round Valuation over Sale Price Figure 4 illustrates the ratio of the (Last Round Valuation)/(Sale Price) (‘LRV/SP’) from a sample of 45 VC-backed firms sold in 2003 or 2004. Data are sorted based on whether the last round was an inside round or an outside round. The diagram below illustrates the distribution of LRV/SP for inside rounds and outside rounds using a standard box plot (with outlier values excluded).

02

46

Last

Roun

d V

alua

tion

/ S

ale

Pri

ce

outside inside

excludes outside values

39

Table 4: Regression Analysis of Valuation: Inside versus Outside Financing This table reports ordinary least squares (OLS), weighted least squares (WLS), and robust regression (RREG) estimates on a sample of 45 VC-backed firms sold in 2003 or 2004. The dependent variable is Log Last Round over Sale, which measures the log of the ratio of the last round valuation (nominal)28 to the sale price. The unit of analysis is the last round of financing, and consequently there is only one observation for each firm. The treatment variable record whether the last round of financing was an inside round [Inside]. All remaining explanatory variables are defined in Table 2. The WLS estimates reported in model (4-5) are weighted based the amount invested in the round. Robust (White, 1980) standard errors are reported in parentheses below each coefficient estimate. We use a two-sided test for statistical significance. OLS WLS RREG (4-1) (4-2) (4-3) (4-4) (4-5) (4-6) (4-7) Treatment Variable

Inside .912** .942** .800* .898 1.127* 1.002*** 1.078***

(.447) (.461) (.469) (.568) (.577) (.350) (.352)

Control Variables VC Control 1.563*** 1.760*** 1.953*** 1.441** 1.005**

(.580) (.591) (.659) (.674) (.392)

California .980** .983*** .688 1.089** .669*

(.419) (.349) (.461) (.408) (.393)

Δ NASDAQ Sale (%) -.213 -.352 -.146 -.795 .443

(.693) (.626) (.980) (.731) (.545)

VC Reputation -.394 -.052 -.053 -1.050* -.433

(.455) (.404) (.541) (.595) (.356)

Duration Fin to Sale .540** .541** .790** .451* .501*

(.216) (.207) (.305) (.253) (.260)

Investment .009 .017 -.001 .013 .008

(.024) (.018) (.021) (.020) (.017)

Industry Dummies YES

Constant .237 -1.337 -1.182 -1.793 -.960 -.085 -1.323 (.286) (.524) (.549) (.693) (.719) (.222) (.571)

R2 .09 .43 .56 .44 .51 - - F-stat - - - - - 8.19 3.26 No. of Observations 45 45 45 33 45 45 45


28 The results reported in this table are unaffected by the method of valuation. We find similar results if we use implied valuations (Metrick, 2007) instead.

40

Table 5: Regression Analysis of Round IRR: Inside versus Outside Financing This table reports ordinary least squares (OLS), weighted least squares (WLS), and robust regression (RREG) estimates on a sample of 45 VC-backed firms sold in 2003 or 2004, covering 90 rounds of follow-on financing. The dependent variable is Round IRR and the treatment variable is inside. All remaining explanatory variables are defined in Table 2. The unit of analysis is the follow-on round of financing, and consequently there are multiple observations for several firms. Robust (White, 1980) standard errors, clustered at the firm level, are reported in parentheses below each coefficient estimate. We use a two-sided test for statistical significance. OLS RREG WLS (5-1) (5-2) (5-3) (5-4) (5-5) Treatment Variable

Inside -.377* -.401* -.306 -.309* -.602** (.215) (.215) (.290) (.172) (.226)

Control Variables VC Control -.187 -.243 -.304 -.365** -.287

(.304) (.286) (.454) (.180) (.239)

California .265 .160 .696 .099 .268

(.263) (.228) (.455) (.158) (.222)

Δ NASDAQ Sale (%) .711** .492* 1.070* .489** .680**

(.323) (.271) (.586) (.223) (.372)

VC Reputation .105 .058 -.133 .085 .229

(.169) (.172) (.263) (.148) (.204)

Duration Fin to Sale -.295*** -.367*** -.340* -.198*** -.338***

(.085) (.090) (.171) (.059) (.091)

Invested in Round -.004 -.008 -.011 -.003 -.004

(.004) (.005) (.022) (.005) (.004)

Industry Dummies YES

Constant .887 .893 .942 .574 1.004 (.375) (.398) (.716) (.232) (.346)

R2 .29 .39 .31 - .41 F-stat - - - 4.08 - No. of Observations 90 90 45 90 90 No. of Firms (clusters) 45 45 33 - 45


41

Figure 5: Round IRR over Time of Financing Figure 5 shows the IRR for each follow-on round of financing in a sample of 45 VC-backed firms sold in 2003 or 2004, covering 90 rounds of follow-on rounds. Outside rounds are indicated with blue dots while inside rounds are indicated with red dots. The graph also shows a Lowess curve plotting the IRR over time for outside rounds (green curve) and for inside rounds (orange curve).

-10

12

Round I

RR

1997 1999 2001 2003Date of Financing Round

Outside Financing

Inside Financing

Lowess Fit Outside Financing

Lowess Fit Inside Financing

vc inside rounds as rescue financing: theory and …...2 i. introduction venture capitalists (vcs)...

Documents