Factors that Matter in VC Returns
Sebastian Mallaby's terrific study on the venture capital industry, The Power Law, makes several persuasive arguments:
- Skill matters. Namely, VC funds with better VCs will perform better than those without. That may sound obvious, but given the low probability of success of any single investment, there is a widely held perception that luck outweighs skill. Mallaby argues against this notion.
- Venture networks matter. It is easy to see venture firms cluster in certain areas, Silicon Valley most prominently. Mallaby makes the point that the funds embedded in the networks outperform. In the words of economists, there are agglomeration effects in the industry - the returns of one firm depend in part on its proximity to other firms.
The common explanation for the later argument is that once a cohort of VC firms are located by each other, it is easier for founders to find and pitch them, leading to better deal flow and then positive selection bias in the portfolio. As founders cluster by the funds, denser talent pools for hiring grow naturally in the portfolios.
As the subject of the book is the VCs themselves, Mallably spends less time on the the talents of the founders funded by the VCs, but it seems self evident that the talent of the founders also matters considerably to the eventual outcome of the fund.
Combining these these points, a simple model is that VC returns are a function of:
- a) quality of deals seen (founder talent & VC network),
- b) ability to select correctly, and gain access to the deals desired (skill)
- c) ability to add value to companies post investment (skill)
The weights of these three are debatable, and arguably different in different markets. We also know there is path dependency in the industry - Sequoia's next fund has a much better chance of good returns because of the previous success of prior funds. Sequoia sees/wins more deals, since every founder would love to be funded by Sequoia, and has demonstrated selection skill, since the GPs of the firm have shown prior ability to select winners.
What is not in this list: size of local market, tax policy, regulatory policy, proximity to research universities, depth of local capital markets.
There is a widely held belief that Silicon Valley rose because of Stanford and/or because of large government spending that helped develop the semiconductor industry. Mallaby dismisses these arguments, pointing out that Boston's universities had much deeper talent pools, and larger government research spending during the rise of Silicon Valley.
More recently, Spotify, Skype and Flywire, all developed in a more highly regulated and taxed Europe, and didn't need deep local capital markets to exit. They all listed in the US. Whether the stock trades in New York or Stockholm does not affect the return to the early stage investors in the company.
Summarizing Mallaby, the sine non qua of VC success is the skill of VC investors. Innovators like Arthur Rock and Don Valentine developed the west coast venture toolkit of taking bolder risks, traunching investments, and understanding the return potential of a diversified VC fund girded by power law returns of its portfolio. Later pioneers at Sequoia, Benchmark YC, and A16Z all developed unique, and highly profitable, approaches to investing in early stage companies as the Silicon Valley market matured.
(At the end of Power Law, Mallaby does make recommendations around tax breaks for venture investors. As a long time San Francisco tax payer, I admit I didn't find this part of the book as persuasive, as we have both excellent VC funds and high taxes, but agree that on the margin, regulations like pass through entities and low capital gains can help kickstart a new cluster of VC investors.)
I'm thinking about all this in the context of a new venture cluster. How did NYC, London, and China develop as a vibrant venture communities, with top performing funds? Why haven't more cities followed?
It seems to me the conversation among both Limited Partners and General Partners in potential emerging venture ecosystems is misguided. The discussion is heavily weighed to the size of the local market, or prospects of the local economy.
The size of the San Francisco taxi market had nothing to do with Uber's eventual success - nor did the fact that the early employees and the investors live in one of the highest tax jurisdictions in the United States. Sweden's population of 10M hasn't limited Spotify's $30B market cap.
Digital businesses naturally scale beyond local markets. Imagine trying to evaluate Coinbase by the size or growth rate of the California economy? Post-covid, these companies are also being built transnationally as well.
What matters is talent. Talent of core founding teams. Talent of local investors to find, evaluate and support these businesses. Everything else is noise.
Yes, talent can be a function of the broader local economy, but the two are not equal. LPs should focus on the talent of local VC investors, and the surrounding pool of investable founders.
Now, the harder question. What are the metrics to evaluate talent?
I'm wrestling with this. Seems like conventional wisdom here looks only at broad measures, which are rough at best - such as number of university graduates, or MBA students, or engineering degrees. These are all inadequate because they focus on a static view - what human capital the local environment is producing each year.
Dense talent networks form when people emigrate to a new cluster. Silicon Valley thrives because ambitious people move there, from all over the US and the world. And, immigration data itself is too broad, as investors would really like to consider the narrower skills sets around building high growth technology companies.
Here's the data point I want: A rolling three year trend line of graduates from the global top 50 engineering schools, top 50 MBA programs and top ~50 tech companies, that have moved/returned back to a specific city.
Anyone seen a data set like that?
Did you like this post? In one click, please let me know