This roundtable discussion was chaired by James Harris at the Coller Institute of Private Equity.
Roundtable participants:
Yael Hochberg, Kellogg School of Management
Hochberg is assistant professor of finance at the Kellogg School of Management at Northwestern University. She is a research fellow at the National Bureau of Economic Research and associate editor of the Review of Finance.
Laura Lindsey, WP Carey School of Business
Lindsey is associate professor of finance at the WP Carey School of Business at Arizona State University. She writes about entrepreneurial finance, venture capital and corporate finance.
Simon Cook, DFJ Esprit
Cook started his venture capital career at 3i Group. He negotiated the partnership with DFJ in 2007 to form DFJ Esprit. In 2009 the firm acquired 3i’s European venture portfolio. Prior to working in venture capital, Cook was a strategy and IT consultant at KPMG.
Andrew McGregor, eCommera
McGregor founded eCommera, a provider of e-commerce services, in 2007 alongside Michael Ross. The company is backed by Frog Capital, West Coast Capital, ePlanet Ventures and WPP. Before setting up the company, McGregor was marketing director at BT and web director for The Economist.
Jeff Fagnan, Atlas Venture
Fagnan is a partner in the technology team at Atlas Venture’s Boston office. Before joining Atlas, Fagnan was a partner at Seed Capital. He recently founded Tugg (Technology Underwriting Greater Good), a crowdsourced philanthropy network dedicated to social entrepreneurship.
Venture capitalists often boast of their ‘value add’ to entrepreneurs, and one of the most trumpeted aspects of the VC package is their network of industry contacts among fellow investors, corporates and management teams.
Yet might the structure and club-like nature of these networks have a detrimental impact on entrepreneurs and the emergence of new blood in the VC industry?
We discuss the effect of these networks on the market and on portfolio companies, based on the findings of three studies.
l The first finds that communities with well-developed VC networks can create barriers to entry for new investors and can lead to lower initial valuations for entrepreneurs.
l The second finds that the size of a VC network has a measurable effect on the performance of portfolio companies.
l The third shows that networks often result in strategic alliances between portfolio companies of a shared investor.
Harris: VC networks are generally considered a good thing, so why might they have an adverse impact on entrepreneurs?
Hochberg: Evidence suggests it is harder for new capital providers to enter markets with strong networks, and entrepreneurs tend to get lower valuations on their companies by an order of around 10%. In highly developed networked VC communities, VCs share deals and information with networks so outside VCs are less likely to have access to those deals.
Harris: Are VCs colluding to keep prices down and competitors out?
Hochberg: Even if you view a highly developed network as a barrier to entry, it is still possible for new entrants to increase their chances of breaking into a market. We have found that new entrants with ties to the incumbents in the specific market they are targeting are more likely to be able to enter the market than those without – and when they do, they form syndicates with the incumbents to which they previously had ties.
In fact, there is zero evidence that incumbent VCs collude in any way, even tacitly. It is not about collusion. It is about strategic choices that are beneficial to VCs but not harmful to entrepreneurs. Most VCs have a set group of partners with whom they like to work because they have similar structures, goals and timelines.
However, it also appears that incumbent VCs that work with new entrants are less likely to be invited to future deals by fellow syndicate members. This effect tends to last for up to five years, which suggests some strategic co-ordination.
Lindsey: But it might also be that markets with highly networked communities are saturated, so there is no room for more VCs. Yael Hochberg, Alexander Ljungqvist and Yang Lu address this issue as best an academic can (see box below) but it is very difficult to know for sure which effects are causal.
Cook: In my view, even if there is evidence to suggest lower valuations on entry, these matter less than the final exit. Global winners attract premium values. There can be a trade-off in value at the time of the investment, but that comes with a higher likelihood of success later.
Harris: If there is no collusion, why are new entrants often not invited to syndicates?
Hochberg: At some point there are diminishing marginal returns from having too many people active in the market. Other studies show there are positive agglomeration effects from having a greater number of VCs in a market, but at some point these seem to stop.
Harris: Is it not the case that VCs are sometimes reluctant to bring in corporates?
McGregor: Yes. There was some scepticism about the benefits that [advertising and marketing group] WPP would bring to the table when it joined the syndicate along with our institutional backers. I think some institutional investors have had bad experiences in the past with corporate investors, but we have never had any problems with WPP. For institutional investors, the emphasis is on making a return. WPP wants to make a return on its investment but it also wants to catalyse revenue streams or impart insight into other agencies.
Harris: Given all this, do networks benefit portfolio companies as many VCs suggest?
Hochberg: Yes. The benefit can be seen through two channels. A strong network enables VCs to source the best dealflow from a pool of investments A VC has a choice of any number of investments, but a strong network provides access to the highest-quality start-ups and the best dealflow. The second channel is through adding value and nurturing the start-ups they fund. It is not just finance that investors bring to the table. They also provide access to human capital, service providers and so forth. Each investor in a syndicate brings his own contacts, which enables a start-up to grow into something it would not otherwise be.
Cook: Networks are highly important. In fact, VCs appear sometimes to have two résumés. One involves the firm you work for and where you have a partner meeting every Monday to discuss your fund and portfolio. The other key part of the role involves board meetings with portfolio companies and working with other VCs very closely. Networking is an essential part of the job. Start-ups have to think globally from day one and the better connected VCs are, the better able they are to provide access to more finance, strategic partners and senior executives for management positions.
Harris: But what can they bring to entrepreneurs that they cannot find elsewhere?
Fagnan: Quite a bit, actually. An investor with 10 years of experience as a VC will have been involved with 40 to 50 companies. Along the way, that person would have met a number of entrepreneurs and key customer contacts.
Cook: Entrepreneurs run small companies but are looking to become global leaders in a new sector rapidly. DFJ, for example, is one of the largest venture networks in the world and each firm focuses on a geographic area – Silicon Valley, Israel, China, Europe, Brazil and so on. Each firm operates independently but shares the strong brand, IT systems and intelligence on new markets. We also share carry [performancebased profits] across the funds so we have an economic incentive. If a UK business needs an introduction to a Silicon Valley-based company, we have an internal email system and we can typically get a response within a day. Other US VCs, such as Accel, have a similar global model, and a lot of US firms have extended their networks into China and India. I think a number of US-headquartered global VC networks will emerge. This raises a separate issue, which is whether they adopt a centralised structure, where firms hire partners and report back to Silicon Valley, or a more flexible model.
Harris: Would it not be just as beneficial for entrepreneurs to go directly to corporates under a strategic alliance arrangement?
Cook: No. It really helps to have a VC involved. There is a huge interest from corporates to get to know and work with VCs because they often rely on an outsourced innovation model, which requires them to create their own dealflow and connections. Some of the larger technology companies have their own VC investment programmes but it is hard for one business to meet thousands of start-ups. VCs act as a liaison.
Lindsey: VCs do facilitate strategic alliances for their companies, but the
point is that these relationships are formed disproportionately more often within the portfolio of a shared investor. This can help mitigate contractual issues. One of the problems of strategic alliances is that they require information to be disclosed, which carries its own risks, but by having a shared investor, it can be ensured such disclosures take place when the interests of the contracting companies are aligned. Furthermore, VCs can serve as an information conduit or an enforcement mechanism. In the process of screening portfolio companies for investment, VCs collect information, which can be disclosed selectively to portfolio companies. A shared investor in a strategic alliance is also in a better position to punish a company that attempts to take advantage of its partner company by not introducing it to new technologies in the future.
Harris: But is there a temptation to form alliances within a portfolio to hide a poorly performing company?
Lindsey: There have been studies that suggest VCs pass on struggling portfolio companies to more successful ones. However, I find evidence that the alliances benefit portfolio companies and that they exit on their own. It is not the case that, on average, VCs facilitate strategic alliances to hide an investment.
Harris: So which are more beneficial – external contacts or other companies in a VC’s portfolio?
Fagnan: It is an older view of the start-up world. Today, start-ups are more global, fragmented and open. It is not important if an alliance is internal or external. It is easier to arrange meetings between portfolio company members so entrepreneurs can share tips, but it happens just as often with external companies.
Lindsey: My study does not show which is more beneficial. It shows that alliances are more likely to form when there is a shared investor. It is a model that is loosely based on the old Japanese system whereby a main funder provides capital to a number of companies, all of whom work closely with one another. At the time I wrote the paper, Kleiner Perkins linked itself to this model in its promotional material, but it has since removed any reference to this system, perhaps because it is no longer fashionable.
Cook: The Kleiner Perkins famous keiretsu effect, where portfolio companies worked together, used to be important but it is a global market now. Ultimately, VCs have minority stakes and can only influence, not control. If two DFJ companies are working together, there would have to be a clear agreement made about intellectual property and commercial rationale beforehand. Investors prefer to keep at arm’s length commercially so that when it comes to an exit, the buyer knows everything is clean.
Harris: What evidence do we have that VC networks actually improve company performance?
Hochberg: Our research indicates that the effects of a strong network are economically significant. If you were to take the average VC with the average level of networking, and change the extent of the network by one standard deviation, that would increase performance on an annual basis by about 2.5%. If you consider that the mean net internal rate of return [IRR – a measure of annual performance] for a venture firm on an annual basis is 15%, deviation moves the IRR from 15% to 17.5%, which makes a huge difference to limited partners [investors]. Strong networks also mean portfolio companies can find finance faster and benefit from a broader range of contacts. The better the network, the more likely the portfolio company is to survive and ultimately succeed. It is beneficial at the fund level, but also at the portfolio company level.
Fagnan: Sometimes VCs overplay the importance of their network in how much they can introduce companies to customers and potential acquirers. At the earliest stage, companies focus on developing products, so often the most beneficial thing is to introduce good developers that have produced tangential types of products. To me, the importance of a network depends on where the company is in its lifecycle. It is only when a company matures that it needs relationships with customers and advisory board members. It is up to the VCs to be proactive and introduce the right people. The best entrepreneurs and start-ups know how to mine every single one of their investors and board members.
McGregor: We have a diverse investor base comprising institutional VCs, such as Frog Capital and ePlanet Ventures, and a corporate investor, WPP. With regard to the institutional VCs, we do not necessarily expect them to facilitate introductions. I am sure they have those connections but that is not necessarily their remit. Their main benefit is pattern recognition. They have experience of navigating fast growth and international expansion, and they can provide a real insight into technology markets.
Harris: How are VC networks evolving?
Cook: I have worked with VCs all over the world and the standard Silicon Valley model is followed everywhere. The key issue is that portfolio companies are operating globally at the earliest stages and they need investors that
can help them globally, for example, expanding in the US or finding partners in China. Market competition, even for start-ups, is global from day one.
This article first appeared in the Coller Institute of Private Equity: www.collerinstitute.com
The research
l Networking as a Barrier to Entry and Competitive Supply of Venture
Capital, by Yael Hochberg, Alexander Ljungqvist and Yang Lu, looks at
whether strong networks among VCs in local markets restrict entry by
outside investors, thereby improving incumbents’ bargaining power over
entrepreneurs.
The authors analyse the interconnectedness of US VC networks and compare it with fund data between 1975 and 2003. They find that new entrants are at a disadvantage because of the cost and time involved in establishing visibility, credibility and access to information in densely networked markets. A one standard deviation increase in density of networking reduces the number of entrants by a third.
In addition, incumbent VCs are less likely to invite new members because they are then less likely to be invited to subsequent deals by their fellow syndicate members. However, the research finds that one way to overcome this barrier to entry is by providing incumbent VCs with reciprocal access to the entrant’s dealflow in its home market. The study also finds that more densely networked markets have significantly lower valuations for VC deals.
l In Whom You Know Matters: Venture Capital Networks and Investment Performance, Hochberg, Ljungqvist and Yang analyse VC fund performance data in the US between 1980 and 2003 to determine the relationship between fund and portfolio company performance and the extent of VC networks. The authors assess networks using graph theory, which provides a tool for measuring the relative importance of each actor in a network. The research indicates that VC funds with a larger network are more likely to be invited into other VC syndicates and have significantly better performance as measured by successful exits. A one standard deviation increase in the quality of a network will lead to 2.5% rise in exit rates. Better networks also improve the likelihood of portfolio companies raising subsequent rounds of funding and increase the chances of survival. For a one standard deviation increase in network quality, the survival probability in the first round of funding increases from 66.8% to 72.4%.
l Laura Lindsey in Blurring Firm Boundaries: The Role of Venture Capital in Strategic Alliances examines the extent to which VCs facilitate strategic alliances for their portfolio companies.
The research finds that firms are more likely to form an alliance if they share a common VC investor. In addition, this effect is strongest in cases where contracting problems are more pronounced, such as for private companies. Data also reveal that these relationships create economic value as measured by successful exits through an initial pulic offering or acquisition.
As well as providing funding and improved governance, VCs can create value by facilitating strategic relationships, especially for companies where information asymmetries and expropriation concerns may prevent such alliances.