A 1900 word finale in the series. Part 1 is here. Part 2 is here.
The separation theorem is the foundation of Modern Finance. Among other things, it is used by investors to construct optimal portfolios of investments based on underlying assumptions of risk, return, and correlation. Limited partners (LPs) use this rationale to compute allocations of public versus private equities. Within private equity alternatives they further allocate between venture capital, LBOs, distressed debt, etc. And within venture capital they allocate between early and late stage, market segment, national and local geographies, etc.
Of course, each dissection introduces more random measurement variance as the analysis gets increasingly refined. At some level, as I suggested in the prior post, the noise overwhelms the signal. Funds of funds provide a very useful function in aggregating the volatility of individual private equity firms. The theory is a Fund of funds can provide a lower beta than a direct VC fund investment. The theory relies on reliable and persistent investment strategies of the underlying components – the VC funds themselves.
Many investors (pension funds, insurance companies, unions, etc.) also want to construct their own ‘optimal portfolios’ from private equity instruments. So they overweight certain segments, such as buyouts, and underweight others, such as late stage technology venture capital funds. They, too, want the “pure plays” that portfolio theory demands.
Most VC funds, especially the newest ones, encounter an investors’ paradox. They market themselves to LPs on a specific investment thesis (stage, sector, etc.) If they sense the original investment thesis no longer is a return-maximizing strategy, they face a difficult choice – pursue a new thesis to maximize return and become accused of ‘investment drift’ or remain a pure play and forgo top-quartile status. Those who construct the portfolios (the LPs) want pure play strategies and top-quartile outcomes.
Some investors might consider this sturm und drang in VC ranks as part of the eternal Darwinian process of winnowing of investment managers. But I think there is more to it. The ballooning of the total amount of capital in private equity is changing the drivers of rates of return in venture investing. The sheer number of practitioners is at all time high. As a result, private equity markets are more efficient and more volatile. This means that sustainable investment theses based on technical parameters are not as durable as they used to be. Sometimes early is better; sometimes late. Sometimes traditional private is better; sometimes PIPEs. And venture capital is a long-only bet. But while you can’t always pick the private winner, you often can pick the public losers. The notion of a pure and enuring sector/stage theis is in jeopardy as a profit-maximizing strategy for any fund with a ten year life, or even a five year investment period.
Venture capital is evolving into two models. One is the traditional model of the long-ball home run outcomes driving the portfolio. For descriptive purposes, I refer to this as the Boutique Model. This is what we know as Venture Capital 1.0. It is the model practiced by most VCs today. This is becoming increasingly difficult for most, but not all, VCs. This model relies on VC brand as a beacon to attract entrepreneurs. The brand owners rely on their networks to expose them to and validate proposals from the best entrepreneurs. In theory, the best entrepreneurs are the tail of the distribution that can achieve that long-ball outcome.
The 100x return has always been rare. The 10x return is becoming an endangered species, and the "not great, but I'll take it" feeling associated with the 5x is now a cause for celebration. I illustrate this outcome compression in the associated graphic.
These so-called "brand firms" have always operated at the right tail of the graph. So the outcome compression from excess capital probably has minimal impact on them. But those who pursue the traditional 'long-ball" strategy of VC and fail to attract those increasingly scarce 10X+ outcomes are likely to find themselves in that investors' paradox and scrambling to "re-define" their investment thesis.
The influx of capital and capitalists is changing the definition of
the business. When the public and private technology equity markets
were less institutionally populated, the world of risk capital was
segmented into separate businesses. Early stage venture capital was a
different business from private investments in public companies
(PIPES). Late stage venture investing was different from leveraged
buyouts. Shorts and longs were as different as night and day.
I
say they were different businesses because the classic tests of
business definition all pointed to this conclusion. Reaching back to
my early days at Bain & Company, we relied on three tests to assess
if two lines of business were indeed the same business (I knew this stuff would come in handy some day).
- Do they serve overlapping sets of customers? (Shared Customers)
- Do they share overlapping sets of competitors? (Shared Competitors)
- Can you invest in one line of business and materially impact your cost position in the other? (Shared Costs)
The more the answers to these questions were answered in the
affirmative, the more likely the two initiatives were in the same
business. And the VC industry is seeing more competition from other
investment sources as
all forms of risk capital converge. These various forms of capital are
not
perfect substitutes. Angels don’t compete mano-a-mano with
hedge
funds. Seed VC funds aren’t bidding on LBOs. But there is enough
contamination at the edges in each class to cause asset inflation from
increased total competition. This convergence is bidding up prices in
all sectors.
This is my core observation about venture capital. It is no longer
just a pure and separable business. It is a feature of the overall
business of providing risk capital. It is a product in a risk capital
product line. Like many of my VC brethren, I am an early stage venture
capitalist; I am a product line manager. As a principal in an early
stage firm, I am a general manager in a single product company. Single
product companies can be highly profitable or abject failures. Lack of
diversity is neither an asset nor a liability. It is simply an
attribute. But the boutique segment is winnowing down to a very small
handful of successful brands, representing a very small portion of the
market. The best brand-driven VC 1.0 firms will continue to succeed. However, the sheer amount of total risk capital available will continue to make it difficult for many firms to build the very brands that drive predictable and continued success in VC 1.0.
If the basis for competition in the boutique segment is brand,
what is the basis for competition in the larger business of risk
capital?
This brings me back to my opening remarks about the
separation theorem. The separation theorem assumes that individual
investments have risk/return properties which are affected by the
construction of the portfolio. An early stage investment's alpha and
beta are not affected by the inclusion of a distressed debt
instrument. Now suppose the insight about how to price the distressed
debt arose from knowledge you gleaned from the early stage investment.
Or suppose a profitable public market short position arises from the
insight acquired from chasing, but ultimately being outbid in pursuit
of a 'hot' late stage investment. Or suppose experience with a PIPE
provides a source of market and customer diligence to generate ideas
about potential new Series A companies to form and fund. I refer to
this as a Crossover Model of risk capital. The crossover model is a model for Venture Capital 2.0.
It is best to illustrate with a hypothetical. Suppose I believe
Apple is developing a Skype-like Ipod for release in 2007 and I believe
it will successful. I could simply communicate that to investors and
let them construct portfolios of longs and shorts around this thesis (the separation theorem in action).
Alternatively, I could use this to construct what I believe is a
risk/return maximizing set of investments around the insight. I could
place early stage bets (communication services), midstage bets (new
battery technologies), and public bets (Apple) on the thesis. If I go
long on all bets, I have a better return than if I simply picked a
market basket of early, midstage, and public technology companies. If I
hedge my bet by shorting , instead I can perhaps construct a better
risk/return than a pure long, early stage VC position. If I
communicated the "Skype Ipod" insight to my investors, they could
construct their own portfolios. But by communicating the thesis runs
the risk of disclosure and loss of advantage. The risk of disclosure
is a real transaction cost. And the separation theorem
assumes zero transaction costs.
The basis for competition in the Venture 2.0 Crossover model is a
focus on markets, independent of stage, geography, and risk capital
instrument. The Boutique model relies on a brand-generated magnet
status to find outlier long-bets. The Crossover model would rely on a
full-spectrum view of private-to-public to generate an appreciation for
inefficiencies in an otherwise fairly efficient market for risk
capital. Like everyone else in the industry, it faces the risk of
competing for deals with marginal competitors who have excess cash
burning a hole in their pockets. However, unlike everyone else, the
crossover firm can leverage the insight in multiple, non-competitive
forms.
So Venture Capital 2.0 is increasingly about applying capital to
market insights, across the continuum of private and public, early and
late. Move money where the insight-driven opportunities are. This
cuts directly against the grain of most LPs who want to combine
investment products from great firms with different stage foci. But in
doing so, they give up the inherent advantage of cross-leveraging an
insight with either hedges (to reduce risk) or additional longs (to increase
return). Financial and product markets are no longer independent. The
hedge can be to go long in both the private upstart and the public
incumbent. The short can be to short the incumbent. Finally, if the
startup loses to an incumbent, the lost capital may be recovered by
shifting assets to the incumbent before the broader market fully
appreciates the dynamic.
Venture Capital 2.0 is happening.
The industry is consolidating into larger firms. These larger firms,
while still calling themselves VC funds, look nothing like the VC funds
of 15 years ago. They have multiple products (early, late, PIPE),
multiple geographies, affiliate or satellite entities for distribution,
etc. Few are formally affiliated with hedge funds, but I believe that
is inevitable. The multi-year partnership structure is probably not in jeopardy -- investment managers will still need to finance private companies over several tranches. However, the venture model of capital calls and distributions will likely give way to more evergreen or closed end funds in which capital can be recycled across investments. This has its own valuation and liquidity issues. Nothing is perfect.
I began this series with a flippant remark about Venture Capital
2.0. I don't think traditional venture capital is going away any time
soon. There is too much momentum, i.e., money. But I do think there
is a real alternative model for the practice of venture capital
investment as a product within a larger risk capital portfolio. It
won't emerge soon. LPs decision models are still based on a demand for
venture products packaged to look like they conform to portfolio
construction models. However, there is enough "investment thesis
drift" afoot these days that some private equity firms are going to
realize that they can use this "drift" to their advantage and drift
head-on into crossover forms of investing.
Thanks for taking the time. I hope you found it worthwhile. I don't claim to be a financial economist. So if you are, and you think my reasoning is flawed, please let me know.
