Whom You Know Matters
Skip to content
Entrepreneurship Finance & Accounting May 1, 2007

Whom You Know Matters

The importance of networks in venture capital

Based on the research of

Yael V Hochberg

Alexander Ljungqvist

Yang Lu

Venture capitalists often acknowledge the critical role of network relationships in the success of individual VC firms, particularly in sourcing investment opportunities, management talent, and information necessary for consistent industry success. One prominent VC, Kleiner Perkins Caufield and Byers, even goes so far as to refer to itself as a “Venture Keiretsu.” Research by Yael Hochberg (Finance Department, Kellogg School of Management) and her colleagues Alexander Ljungqvist and Yang Lu (Stern School of Business, NYU) suggests that VC networks—in particular, relationships with other VC firms—can have a significant positive impact on fund performance.

Hochberg, Ljungqvist, and Lu’s report in the Journal of Finance provides clear evidence that VC firms that are well-networked within the VC community are better positioned both to source investment opportunities and add value to their portfolio companies, leading to outsized performance relative to peer funds and to more successful outcomes.

“Finance research often treats markets as consisting of arms-length, spot market transactions. Many financial markets, however, the VC market included, are characterized by strong relationships and networks,” says Hochberg. “The performance implications of these networks haven’t been looked at in prior work, primarily because they can be hard to quantify.”

The networks of VC relationships that arise through syndication, or co-investing, in portfolio companies, however, are easily observable and tractable. Hochberg and her coauthors employed data from Thomson Financial’s Venture Economics database on VC investments in the United States from 1975 to 2003, and used information on the investment syndicates in these companies to construct historical networks of ongoing relationships between VC firms. They then linked these networks—built from 47,705 investment rounds involving 16,315 portfolio companies—to data on 3,469 funds managed by 1,974 VC firms.

To test the effect of VC networks on performance, Hochberg and her coauthors created five measures of how well-connected a VC firm is in the network, using network analysis methods based on the mathematical discipline of graph theory. These measures include:
• the number of other VC firms a firm is connected to,
• how well-connected the VC firm is to other well-connected VCs,
• the number of other VCs a firm is connected to through being invited to participate in their deals,
• the number of other VCs a firm is connected to through inviting those firms to participate in its deals, and
• the extent to which the VC firm acts as a broker between other non-connected VCs.

Each network measure is meant to capture a different aspect of a VC firm’s influence in the network. The number of VCs a firm is connected to proxies for the information, deal flow, expertise, contacts, and pools of capital to which it has access. The more ties a VC firm has to other firms, the more opportunities for exchange; being well-connected to the best-connected VCs takes into account the quality of those connections. The number of VCs inviting a firm into their deals reveals the extent to which the firm expands its investment opportunity set and gains access to information and resources it might otherwise lack. In contrast, the number of VC firms a VC invites into it own syndicates measures a VC’s ability to generate future coinvestment opportunities through expected reciprocity.

Obviously, networks are not static. Relationships may grow stale, and entry to and exit from the network over time may change a VC firm’s network status. “We constructed the networks using data only on the previous five years of syndicated deals. This allowed the network measures to adjust to entry of new VC firms, changes in relationships, and other fluctuations in the size and nature of the network,” says Hochberg. “It also avoided any bias in the measures that might arise from one VC firm having a twenty-year history of syndicated deals while another younger firm had only a five- or ten-year history.”

Hochberg, Ljungqvist and Lu then related these network measures to the subsequent fraction of portfolio companies that achieved successful exit through IPO or trade sale, and to internal rates of returns on funds released through Freedom of Information Act requests. The effect they found on fund performance was substantial. In particular, being well-connected to many VCs, being well-connected to the best connected VCs, and being invited into many other VCs’ deals had the largest effects, while the ability to act as an intermediary in bringing other VCs together played less of a role.

Clearly, many other factors affect VC performance. The empirical tests developed by Hochberg and her coauthors control for a variety of performance determinants—vintage year effects, fund size, investment opportunities, competition, industry and stage focus, past performance of the VC firm’s funds, and various aspects of a VC’s experience. Says Hochberg, “It isn’t just the case that networks are simply a proxy for experience—many experienced firms have had the opportunity to build strong networks, but some as a matter of strategic choice rarely syndicate deals. Once we control for the extent to which a firm is well-networked, the extent of a firm’s investment experience has little to no effect on performance.”

Similar effects on performance are seen at the portfolio company level. Being well-connected to other VCs has a positive and significant effect on the probability that a portfolio company will survive to a subsequent financing round or exit successfully through trade sale or public offering.

The analysis also suggests that the relationship between being well-networked and better performance is causal, rather than a result of superior performance merely improving a VC firm’s network position. In fact, Hochberg and her coauthors found little evidence that past IPOs and trade sales of portfolio companies helped enhance how well-connected a VC firm was in the network. Instead, what appears to be key in improving a VC firm’s network position is demonstrating skill in selecting and adding value to investments, and engaging in reciprocal sharing of deal flow.

How does networking affect performance? Is it simply through access to better deal flow, or does networking also contribute to the VC’s ability to add value to its portfolio companies? Hochberg and her coauthors performed a number of tests to ascertain the extent to which networking’s benefits can be attributed to each of these channels. Their analysis suggests that while access to deal flow is important, a strong network appears to affect a VC’s ability to provide value-added services as well. Networking appears to boost performance much more precisely when the VC does not enjoy access to better deals through its network.

While much of the analysis resulting from the research implicitly assumes that every VC in the United States can potentially have a network tie to every other VC, VC networks tend to be geographically concentrated, or involve only VCs specializing in a certain industry. For example, a VC headquartered in Silicon Valley may be very well-connected in California, but not well-connected in a network that includes VCs from the East Coast and the Midwest. Similarly, a given biotech VC firm may be very influential and connected in a network of biotech VCs, but may lack connections in the overall network of U.S.-based VCs across all industries. Hochberg and her coauthors repeated their analyses using networks derived from industry-specific networks and the network of California VC firms. In every case, they found a large effect of being well-networked on subsequent fund and portfolio company performance, and the magnitudes of the effects were frequently strengthened.

The results in the study have clear ramifications for entrepreneurs, VC firms, and the institutional investors that invest in their funds. “The finding that better-networked VCs enjoy superior performance has clear implications for limited partners choosing which VC fund to invest in,” says Hochberg. “In addition, our research provides institutional investors with a deeper understanding of the drivers of cross-sectional VC performance, beyond the models used up until now. For VCs, enhancing ones network position should be an important strategic consideration, while presenting a potential barrier to entry for new VCs. For entrepreneurs, it is another consideration to take into account when deciding on an investor.”

In Hochberg’s opinion, many central questions still remain for future research, in particular, how networks affect the entry of new VCs, the valuations obtained by entrepreneurs for their startups, and the competitive supply of venture capital. Hochberg, Ljungqvist and Lu’s new work on this topic, “Networks as a Barrier to Entry and the Competitive Supply of Venture Capital,” was recently presented at the annual meetings of the American Finance Association. “Networks are a central feature of the industrial organization of the venture capital industry,” says Hochberg. “So far, we’ve only touched the tip of the iceberg in our understanding of this aspect of the VC world.”

Further reading:

Hochberg, Yael V., Alexander Ljungqvist and Yang Lu (2007). “Networks as a Barrier to Entry and the Competitive Supply of Venture Capital.” Working paper, Kellogg School of Management, Northwestern University and New York University.

About the Writer
Yael Hochberg, Assistant Professor of Finance, Kellogg School of Management, with the assistance of Brian Yoon, Kellogg Insight Entrepreneurship Scholar and 2007 MBA candidate, Kellogg School of Management and Joseph Dwyer, 2008 JD/MBA candidate, Kellogg School of Management.
About the Research
Hochberg, Yael V., Ljungqvist, A. and Yang Lu (2007). “Whom You Know Matters: Venture Capital Networks and Investment Performance.” Journal of Finance, 62(1): 251-301.
Add Insight to your inbox.