Chapter 3 — Unpacking Alpha Generation in VC

The purpose of this chapter is to walk through my thinking on alpha generation, or outperformance, in venture capital. I think VC Performance can be attributed to a basic Beta measured by median performance in the asset class and then isolating factors that deliver top decile performance in the asset class. Without being too scientific in the approach (and discounting any impact of uncalled capital) as data issues pervade and others have done far more comprehensive work on this, my model based on the historical returns is as follows:

Photo by Annie Spratt

Unpacking Alpha

Beta in VC is composed of two core components (albeit I am isolating / looking at this in the recent cycle):

  1. R(M) VC = Cost of capital associated with value creation in early stage of technology, irrespective of market timing — a fixed effect (typically S&P / Russell 2000 + 3 -5%)
  2. Illiquidity Premium

Top Decile Performance is composed of:

  1. Strategy Timing Premium — Capturing technologies at the correct point on the inflection curve as driven by market forces (Part expert timing, and part luck!)
  2. Alpha — Comprised of certain readily identifiable, and advertised components; including investment process, portfolio management, access to deals, investment strategy and fund value add.
VC Alpha Waterfall

My issue with Alpha generation in venture capital is that it is both extremely poorly articulated and attributed (compared to the public markets) nor supported by data. I would love to perform a historical analysis across the history of the asset class to solve this problem — this is a data science problem but at present this is unachievable.

Firm Capital vs Partner Human Capital

Venture is undoubtedly still all about Human Capital. It is a game of hustle where individual partners drive superior returns over firm capital, with estimates of partner value being in the order of 6x more valuable as a predictor of future performance versus firm organizational capital. This is a business where experience matters as evidence asserts that companies funded by more experienced VCs are more likely to go public. But given the weight placed on lists like the Midas List and other individual partner rankings, it is hard to argue that this is a business where a great deal of value is tied in the firm. The brand of the individual VC trumps that of the firm, in general, save some truly “celebrity” firms like Sequoia, A16Z, Benchmark.

Components of Alpha Generation

Given the inability to attribute alpha generation with quantitative rigour, it became important for me to understand this through the practitioner’s lens. Through qualitative interviews supported by looking at a range of academic research has been conducted into the drivers of alpha or edge in venture capital, my hypothesis was to reason by way of deduction to determine which elements could be conclusively stated to increase performance at the fund level.

Components of Alpha
  1. Access

A core argument to making an allocation to long-standing VC firms is access to superior deal flow which is driven by the so-called “proprietary” networks and brands of these funds. Whilst it has been conclusively shown that network drives portfolio performance by virtue of reducing both financing and execution risk, the decline in persistence of performance over the last decades and the performance of new managers shows that it is hard to argue that firms have some structural barrier to entry and have freedom to select the best deals. Network is essentially a commodity that is attributed to partners over firms so new managers can win. Network for the sake of network is not what drives outlier returns. Deliberate and thoughtful networking is a skill that can be both learned and tested for (as we will discuss later in the series). That said, as with any walk of life, network is built and cemented over time, it can be short-circuited but the most long-standing VCs do create an aura (e.g. Ron Conway). I am not asserting that it is not an important attribute but the cold start problem is de minimus.

Source: Cambridge Associates — Venture Capital Disrupts Itself: Breaking the concentration Curse

2. Value Add

Historically, venture capital was purely risk capital and entrepreneurs were very much at the behest of their investors. Today, as the market has matured, the narrative has changed with much more active marketing to entrepreneurs about the value add of the firm and an era of “founder friendly”. The best VC’s certainly do provide more than just capital but the direct attribution of value add and ROI or impact for this is hard to isolate.

Without doubt, VCs add value through strategic introductions and business building support and future financing support. Note, this is not a commentary about emerging platform services at the earliest stages of a company’s lifecycle such as YC or other great seed fund platforms (e.g. First Round), where I do consider comprehensive, and scalable support for embryonic companies increases the probabilities of these companies staying alive. But at later stages, this is not control investing so value add is more ad hoc and the VC model is not scalable.

In my mind, the best entrepreneurs (who drive the outlier returns) would most likely have found the ways to create value without the hand-holding of their VCs and this fact has been admitted by VCs themselves (see chart below). In fact, much more negative assertions have famously been made about the value add of the broader asset class. In my mind, there is a significant conflation of the value add actually attributed to VCs. If VCs really do add significant value, why do they not advertise this quantitatively? (Some initiatives are being created)

And thinking about this as a new entrant, these services can be hired for i.e. Talent Partners and Corporate Development and within a portfolio of 25 companies, this can be provided to a sufficient level through a dedicated human resource. Remember, generalist VC really doesn’t scale as the intricacies of early stage businesses are so many — it is impossible to solve for all.

Sources: “How do venture capitalists make decisions?” (2016) Gompers, Gornall, Kaplan, & Strebulaev. / Correlation Ventures: Too Many VC Cooks in the Kitchen? / 8VC Survey / Illuminate Ventures Survey

3. Investment Strategy & Process

So by virtue of deductive reasoning in the fact that good counter-arguments exist for access and value add driving superior and consistent outperformance in venture capital, I come to the conclusion that unless it is all luck (which the long-standing performance of the tier one VCs would suggest otherwise), then venture capital, as an investment profession, must be truly about the formulation and execution of investment strategy, process and portfolio construction and management! This is not simply saying it is about effectively “stock-picking” but it is recognising that the best funds must embed systems and operations that allow them to parse, digest and prioritise the multitude of potential choices and ultimately select a portfolio of companies that is optimised for the Power Law returns of this asset class.

At the inception of the fund, fund sizing and investment strategy (ownership, stage, reserves) must be defined and tied to the current market opportunity. Funds that have deliberately capacity constrained themselves have continued to perform in the top quartile. The very nature of VC where opportunities are tougher to scale, means that when funds become larger, the marginal returns to capital are going down, even at the best funds. As funds grow, managers shift to offer a variety of strategies but returns are diluted reflecting the AUM creep that is seen in other asset classes.

The Operations of the VC Investment Model are:

1. Investment Process: Managers face multiple, simultaneous, and interdependent decisions, possibly including a continuous choice set so investment process is key. That said, investing is a data problem. Leveraging systematic decision processes allows investors to boost the accuracy of their decisions and ultimately the success of the firm’s portfolio. Following a range of conversations with top decile GPs, top performing firms comprise investment partners with an articulable, iterative, scientific and repeatable investment process. I will detail the human capital elements of the best venture capitalists in more detail later as I believe it is truly a mindset!

2. Portfolio Construction: Portfolio construction is a well documented challenge in venture capital and there are well-articulated opinions on optimal sizing. Please note, some fantastic investors have provided amazing insight on this, including notably Seth Levine and Jerry Neumann. Ultimately, this is a function of fund sizing and then a strategy which must be designed to time a market. It is not a question of one size fits all but for an early stage investment fund, the historical data shows that a concentrated number of investments (<25) provides the highest potential upside. In addition, an understanding that there is a targeted and optimal loss ratio is critical.

Source: PitchBook Data, Inc. (All funds in sample)

3. Portfolio Management: Reserves management is critical in this asset class where the returns on investments must be concentrated in the winners and ownership percentages maintained. Pro rate rights are the drivers of returns (as evidenced by my earlier analysis on the Uber valuation growth).

Portfolio management has been empirically shown to result in fewer portfolio company bankruptcies as a result of clear of being able to focus time and resource on the correct company. There is debate around concentration of a VCs time, with the rational, non-emotional investors recognising that so-called “loser” investments should be abandoned and time and resource focused on the “winners”. This is empirically accurate but given this is a network.

4. “Strategy: Market Fit”

Historical fund performance analysis indicates that smaller funds in aggregate perform better on a net multiple and net IRR basis. Smaller funds have greater long-term persistence than larger funds. At the inception of the fund, fund sizing and investment strategy (ownership, stage, reserves) must be defined and tied to the current market opportunity and historically different sizes of funds have flourished at different times. I call this “Strategy: Market Fit”.

The most notable recent example of this was the growth of so-called “Micro-VCs” (sub-$100m in size) in the late 2000’s, during a period where it was becoming increasingly cheaper to start a company and GPs could build a dedicated strategy to provide less capital to better entrepreneurs in an environment that enjoyed a much more favorable valuation and competition as this was at the time too small for so-called traditional VCs or “Series A” investors. Funds such as Lowercase and Felicis generated stellar returns.

5. Differentiation

There has been limited differentiation amongst VC funds and has been little innovation in the asset class.

The maturation of the asset class is now evident in the rise of specialist and thematic firms (by sector and geography) in the last decade but as yet there is no strong evidence that VC returns are enhanced by specialisation. Certain specialised funds have delivered outperformance based on industry timing (Ribbit Capital at the advent of Fintech and clearly Life Sciences investing has been a separate asset for venture capital investors) but the CleanTech bubble showed that over-specialisation is not supported if market timing is incorrect. In fact, specialization at the fund level has no significant impact on success rates whereas stage diversification is accretive to performance (note: I consider this reflects the benefits of follow-on reserves).

Ultimately, generalist firms that can rapidly appraise and capture the inherent value in technological breakthroughs have historically captured the outlier returns. Specialisation again manifests at the partner level so when the individual investment professionals are highly specialized themselves, the marginal effect of increasing overall firm specialization is much weaker.

Whilst there are some new interesting models emerging such as scouting, scalable micro-services, capital-as-a-service and quantitative investing) I consider the only true innovation in the asset class has been by firms seeking to build that platform approaches of firms such as YCombinator, First Round and a16z. I believe these firms looked at the asset class top-down and from first principles. Whilst not having definitive fund performance results, evidence suggests these firms have developed top quartile strategies and funded some outlier companies.

I am excited about watching the asset class continue to evolve and there are some truly innovative approaches, I still believe the correct approach is for a capacity constrained approach and looking to identify investors with what I consider the winning mindset.

6. Auto-correlation

Early success in venture investing yields better deal flow in subsequent investments. VC as an investment class demonstrates clear auto-correlation via the now-established virtuous circle of strong subsequent fundraising driven by LP risk assessment weighting persistence of performance but the link to subsequent exits is not clear. Better performing partnerships are more likely to raise follow-on funds and larger funds.

Source: “The Persistent Effect of Initial Success: Evidence from Venture Capital” (2017) Nanda et al

VC ultimately shows mean reversion. Initial success matters for the long-run success of VC firms, but that these differences attenuate over time and converge to a long-run average across all VCs. Incumbents have advantages but tend to decay over time if they don’t stay hungry as evidenced by the declining fund persistence. A cold start problem is de minimus providing investment team member level quality is evident. LPs decisions to reinvest in poorly performing funds are a result of poorly aligned incentives, whereby they are not incentivized by performance, but instead incentivized to make allocations.

Disclaimer: All the opinions are my own and do not reflect those of Jetstone Asset Management (UK) LLP. This document is provided for informational and discussion purposes only. It is not a solicitation or an offer to buy or sell any security or other financial instrument. Any information including facts, opinions or quotations, may be condensed or summarised and is expressed as of the date of writing. The information may change without notice. It may not be reproduced either in whole, or in part, without my permission. This document is not marketing material or is not used for the purposes of marketing. Copyright Ahmad Butt © 2018.