Michael Ewens is Associate Professor of Finance and Entrepreneurship at California Institute of Technology. Matt Marx is Associate Professor of Strategy & Innovation at Boston University Questrom School of Business. This post is based on their recent paper. Related research from the Program on Corporate Governance includes Carrots & Sticks: How VCs Induce Entrepreneurial Teams to Sell Startups, by Jesse Fried and Brian Broughman (discussed on the Forum here).
It is well accepted that venture capital (VC) is a “hits” business. In a sample of over 22,000 VC-funded startups founded between 1987 and 2008, 75% had a liquidation value of zero while 0.39% had an exit value of $500 million or greater (Hall and Woodward 2010). Research indicates that returns are enhanced by investor skills, which one might group by (1) initial selection of investment targets and (2) post-investment intervention. Recently, scholars have shown that post-investment intervention by “activist” investors can improve outcomes for portfolio companies (Bottazi, da Rin, and Hellman, 2008; Bernstein, Giroud, and Townsend, 2016), but the specific mechanisms by which this is achieved remain unidentified.
Gorman and Sahlman (1986) list three nonfinancial areas where investors spend time, ostensibly in the interest of improving performance. First, VCs assist with strategic and operational planning. But whether such assistance improves outcomes remains in question, especially as Kaplan, Sensoy, and Strömberg (2009) see little change in business plans among VC-backed startups that achieve initial public offerings (IPOs). Second, investors may make introductions to customers that facilitate sales and drive revenue growth. But customer introductions are difficult to observe empirically, so constructing a clean test of this mechanism is difficult. The third category—recruiting managers—is easier to observe. Nevertheless, it is not straightforward to conclude a direct link between such actions and performance. Recruiting can be purely additive or it can involve replacing existing personnel. Additive recruiting is unlikely to be controversial whereas founders may resist being replaced.
Drawing inferences regarding the impact of founder replacement on performance is challenging because founder replacement is endogenous. Founders may decide to leave the firm voluntarily, either because they see their startup’s prospects as dim or because they prefer to be involved only in the early stages and then depart to start another venture. Replacement may instead be involuntary. Control rights afforded investors via contracts as well as voting rights on the board of directors enable investors to force founders to relinquish their role. They may replace a founder when the business is struggling but they may also elect to replace when the startup is growing quickly yet the investors doubt the founder’s ability to scale up the company (as Wasserman (2003) suggests). In addition, the quality of the person hired to replace the founder may be endogenous. It may be harder to attract strong executives to struggling startups.
It might seem self-evident that replacing founders would help firm performance. If investors are rational and add value by monitoring the firm, then they should not replace founders unless doing so is beneficial. However, if investors think that they are better informed than the founders but are often incorrect, replacement could be generally detrimental. For example, investors may underestimate the influence of a founder and thus the negative impact of their departure. Moreover, even if investors are correct that the founder should leave the company, removing a founder may have unintended negative consequences such as if loyal-to-the-founder employees become disenchanted and leave. Investors sensitive to that risk may instead try to retain the replaced founder in a different role, but the founder may thwart this plan by refusing to stay once replaced.
To address the issue of the effect of founder replacement on firm outcomes, in our paper we construct a novel database of VC-backed founders and their replacements. The data builds on the VentureSource repository of entrepreneurial firms, financings, investors and executives. A multi-pronged data collection identified founders as well as their replacements for firms founded between 1995 and 2008. Some 15% of the 11,929 firms have at least one founder replacement in our sample period, almost 40% of whom appear to stay at the startup after they are replaced. The first major question that we ask is whether these replacements correlate with startup firm exit outcomes.
Naive regressions show a negative correlation between founder replacement and liquidity events. The negative correlation between replacement and subsequent performance could be explained by selection if investors choose to replace a founder when a startup is in trouble or because a highly qualified replacement is hard to find (including when founders relinquish their role voluntarily). We then instrument for founder replacement using a plausibly exogenous shock to the supply of executives who might serve as suitable replacements: changes in the enforceability of employee non-compete agreements. Non-compete agreements have frequently been shown to restrict the mobility of workers, especially technologists and executives in the sorts of high-potential industries VCs tend to invest in. Thus the ability of an investor to attract a qualified replacement may depend on the extent to which non-compete agreements are enforceable. We assess the impact of non-compete agreements by exploiting staggered changes in 14 states, some of which tightened enforceability while others loosened enforceability. Founders are less (more) likely to be replaced when non-compete enforceability has been strengthened (weakened).
Instrumenting for founder replacement with these policy changes shows that replacement increases the likelihood of achieving a high-quality liquidity event such as an IPO or attractive acquisition. Decomposition of the instrumented results reveals that replacing founders with C-level is more consequential than replacing founders in lower roles. Interestingly, replacement appears to help more when replaced founders leave the startup after relinquishing their role.
Taken together, these findings point to the role of venture capitalists in professionalizing their portfolio companies by replacing founders with more experienced executives. Insofar as venture capitalists play an important role in both the decision to replace founders and identify their replacements, the positive causal effects we find establish a mechanism by which VCs add value to their portfolio firms.
The complete paper is available for download here.