How Do Investors Accumulate Network Capital? Evidence from Angel Networks

Vijay Yerramilli is Assistant Professor of Finance at University of Houston Bauer College of Business. This post is based on a recent paper authored by Professor Yerramilli and Buvaneshwaran Gokul Venugopal.

Networks are widespread in many financial markets, and play a crucial role in the transmission of information and mitigation of agency conflicts. In the context of entrepreneurial finance, Hochberg, Ljungqvist, and Lu (2007) show that venture capital (VC) funds with higher network centrality (i.e., better-networked VC funds) deliver better future performance, in terms of the proportion of their portfolio investments that successfully exit through an IPO or sale to another company. However, we know little about how networks are formed or how some investors end up becoming central to their networks. Is network centrality itself determined by reputation gained from good past performance? Do social connections translate into future co-investment connections? Addressing these questions is challenging because most financial markets are dominated by a few large institutions that became big via a series of consolidations, which makes it impossible to examine how their networks developed over time. Moreover, since individuals often move across institutions, the true relationship between individual performance and network connectedness may not be reflected in institution-level metrics of performance and network connectedness.

In our paper, How do Investors Accumulate Network Capital? Evidence from Angel Networks, which was recently made available on SSRN, we overcome these challenges by using the angel investor market to understand how investors accumulate network capital. This market provides an ideal setting because it allows us to focus on individual investors, and to examine how their position in the network changes over time with their performance. This is crucial because, unlike institutional investors such as VC funds or private equity groups, individual angels are not endowed with large network capital to begin with, and have to build their connections from the ground up. Given the high failure rate of start-ups, we can expect that angels who successfully guide their portfolio companies to the next stage of financing will subsequently become more important within their networks.

Despite their obvious importance, angel investors have received very little attention in the entrepreneurial finance literature, largely due to unavailability of structured data. We overcome this problem by collecting data on start-ups and angels from CrunchBase (, which is the largest crowd-sourced database on start-ups and investors, AngelList (, which is the leading on-line fund-raising platform for start-ups, and a variety of other sources. Our sample comprises 4,108 individual angels who invested in 12,225 start-ups during the period 2005-2014.

We measure the performance of angel investors based on whether portfolio companies for which they acted as lead investor successfully progressed from one financing stage in their life cycle to the next; e.g., “seed” stage to “series A” stage, or from “series A” stage to “series B” stage, and so on. For each angel-year combination, we define three indicator variables for success to identify whether the angel successfully lead any of his portfolio companies to the next stage during the year (Success), whether he successfully guided any of his seed-stage portfolio firms to the series A stage during the year (Seed Success), and whether one of his portfolio companies underwent an IPO or was acquired during the year (Successful Exit). We follow the economic sociology literature to create network connectedness measures, such as Degree Centrality and Eigenvector Centrality, to capture the number and quality of connections, respectively.

Our main hypothesis is that, all else equal, successful performance by an angel investor should lead to an increase in his network connectedness relative to his unsuccessful peers. The growth in network connections should, in turn, result in better future performance of the angel’s existing portfolio companies as well as lead to new funding opportunities for the angel. We refer to this as the reputation hypothesis. Of course, successful performance is not exogenous. Therefore, an obvious concern is that an angel’s existing network connectedness or innate ability may be driving both his current success and future growth. We refer to this alternative hypothesis as the network capital hypothesis.

To differentiate between the two hypotheses, we match each successful angel (“treated” group) with several unsuccessful angels (“control” group) during the same year who are very similar in terms of their degree centrality, number of rounds invested and years of experience. Then, we estimate difference-in-differences regressions to examine how network capital growth of successful angels varies relative to their unsuccessful peers in the years before and after they experience success. As per the reputation hypothesis, successful angels should experience higher growth relative to their unsuccessful peers only after successful performance, but not before.

Our results are broadly consistent with the reputation hypothesis. Angels that deliver successful performance are rewarded with more new co-investment connections, see an increase in the quality of their network connections, and are rewarded with more investment opportunities, both as lead investors and as participants, in the following three years compared to their unsuccessful peers, although both groups are very similar before success. These results are particularly strong for angels with low existing network capital. Success also begets more success, in the sense that successful angels are more likely than their unsuccessful peers to lead their other seed-stage portfolio companies to the series A stage and to obtain venture capital financing over the next three years, although both groups are similar before.

An interesting feature of AngelList is that, just like other online communities, it allows investors to follow the activities of other investors without actually co-investing with them. Using this on-line “follower” data, we show that successful angels attract more new followers relative to their unsuccessful peers in the year after they experience success, and that an angel’s existing followers are more likely to establish a new co-investment connection with him in the year after he delivers a successful performance.

The full paper is available for download here.

Both comments and trackbacks are currently closed.