Venture Capital 2.0: The LP Conundrum

(The second in what will likely seem like an interminable series. First here.)

Limited partners (LPs) are investors in venture capital funds.  We raise money from them, just as companies raise money from us.  We tell them our credentials and, to a lesser extent, our business plan.  LPs usually have to make stronger commitments than VCs with even less data.  A VC who loses confidence or interest in a company can choose to cease new investments in that company.  The result is often that their investment gets diluted, perhaps massively, but it still remains.  A LP who loses confidence in a VC fund technically still faces a legal obligation to continue meeting their capital calls.  At best, they face losing all their capital.  At worst, they have no choice but to throw good money after (perceived) bad.

When we were last raising money in 2004, I heard two comments from potential investors that really captured the LP conundrum about venture capital – one from a major corporate LP with over 25 VC relationships and one from a pension fund manager for one of the largest public employee pension funds in the US.  The two comments were

We have no idea anymore what makes a top tier venture fund (corporate)

I think I should invest in smaller VC funds to get a high IRR, but I (1) have no staff, (2) can’t be more than 10% of any fund, and (3) get measured in the short term by how much money I put to work.  With $Xb to invest, I can’t write a check smaller than $50m and would prefer $100m.  (public)

The first LP was owning up to the fact that data-driven predictive models are neigh unto impossible in the LP business.  The second was saying that even with an intuitive model, he was optimizing multiple goals, not just pure IRR.   The second one is easier to solve than the first.  If you had a reliable predictive model, you could solve the back office and incentives problem by expanding staff to capture greater return from a larger number of smaller funds.  (That assumes smaller does generate better returns, bringing you back to the empirical question.) Neither invested in us, by the way.

Most LPs are driven to partnership longevity as a proxy predictive indicator – a.k.a. Roman Numeral Bias. Why are predictive models about relative manager performance so difficult?  I think it is the interaction of several factors.

• There is a lot of noise in the causal modeling of financial outcomes (public market, business cycle, technology cycles, small sample sizes, etc.)
• The lack of transparency is a serious fog for anyone trying to assess fund manager performance
• Few VC firms have institutional business strategies that transcend the immediate principals

The net result is that the prevailing model for venture investments appears to be driven by two selection criteria: 

1. Longevity as proxy for performance and
2. Current top-quartile performance as a predictor of next fund top quartile performance.

Last year Alignment Capital published a really interesting analysis of fund managers that speaks to these two criteria.  After analyzing 645 separate venture funds, they found:

1. Longevity weakly correlates with IRR, showing continuous improvement between funds I and III, leveling off thereafter.
2. Second quartile funds were nearly as likely to have a top-quartile follow-on fund as the current top-quartile funds.

So the two criteria of longevity and current performance are directionally accurate, but don’t predict future performance with any meaningful precision.  This is probably why so much capital continues to flow into the business – half the firms are in the top half.

Historical performance can only take you so far.  One needs a theory of future drivers of returns to select among venture capital managers. Being yet-another-top-half-IT fund is not enough to be considered seriously by anyone. LPs construct portfolios of all private equity and venture capital investments based on their strategies, applying sound financial principles of predicting correlations of returns to maximize risk-adjusted return. The typical VC fund attributes are size, industry, stage, and geography.  These attributes pass for strategies in most fund raising conversations. My next post will deal with why I think this is mostly flawed, for both VCs and LPs.

Comments

[jim Forbes]

Peter,
Good Lord this i a great post. I printed it and am taking 25 copies up to a class on web businesses at Cal Poly tomorrow.
Best,
Jim F

[will]

"Roman Numeral Bias" never heard of it. . . thats hilarious. . .

Seems like there is a huge agency problem in the industry that is keeping it from being efficient and for performance to become persistent (for good and bad).

LP's :"get measured in the short term by how much money I put to work"

VC's : "Few VC firms have institutional business strategies that transcend the immediate principals"

Entrepreneur's : (my color commentary) Looking for a quick exit to gain financial security or just to buy a condo in mountain veiew for $1M

Would love to dig into that data more, I wonder if the data is saying that future IRR's are simply a type Markov process and thus, only correlated to immediate pred. fund performance and not the entier fund group/history. Research seems to imply that but need more data to confirm. . .

(I think LP's need to read Moneyball)

[nick Gogerty]

you may also be interested in the issue associated with hedge fund performance and ahistoricism.

they have a short half life as well.

http://nickgogerty.typepad.com/thoughts_for_now/2006/03/hedge_fund_ahea.html

[Megan Cunningham]

Peter,
I love this post. Very provocative. The one thing I kept thinking back to was your earlier observation about how the increasing size of funds leads to an inflated dynamic in investment patterns ("Valuations are moving up – [because of] the need to invest more money – not the intrinsic value of startups.") Similarly, on the entrepreneur's side, there is pressure to raise more-more-more in a given round...as a knee-jerk requirement to be taken seriously. (This is not something we've done, or believe in, of course - but the pressure is there nonetheless.) So what's the answer there, given that there is so much capital sitting on the sidelines?

On a totally separate note, the one thing I'd like to hear more about (perhaps coming in your next post?) is why the attributes you list here --size, industry, stage, and geography-- are not important to define to be successful. Of course merely deciding on these attributes may not be the equivalent of developing an investment strategy, but I don't believe they are unimportant either. At some point the market place of entrepreneur/VC matchmaking needs clearly defined criteria to make the fundraising process as efficient as possible, don't you think? And I would imagine the same thing is required for the LP/VC raise. Some of these attributes may not be incorporated into a model, or linked to achieving a better IRR. Yet, on a practical level, the ability to attend a board meeting without flying half way across the country seems to be necessary if a VC wants to be truly "value-added" in so many fast-growing companies simultaneously. Having said that, you seem confident that these are not smart ways to differentiate one's fund, so I'm curious how you picture the investment landscape in a more optimal organization.

[Peter Rip]

Megan:

Pazienza. L'anticipazione è tutto.

PR

[leafar]

LP should hire staff to realize analysis of underlying sectors.
Geography in IT & LS is not anymore a real strategy (still is for Buyout).
Stage is nothing if you don't get who passes the baby to whom ....
Studies shows lot of stuff but mainly what LP are looking for is cool brain.
Why choose VCs with communication revolutionary technologies and buy Buyout strategies in telecom. Ie: Why invest in Mangrove or Index (Skype) if you're also buying Carlyle ?

These guy (and I used to be one of them) are lookig for nothing except put cash @ work, they don't get what's at stake at the end of the day beacause they try to optimize/rationnalize IRR, Multiple ... J curve and diversification with the less pain in the ass possible.

Few of them are Entrepreneur or ex Fund Manager, they have a learning curve which is corelated to accident. But it takes them years to face one.
The distance with the ground on which your cash is burned is everyhting.

Brilliant Post, I love the serie, waiting for the sequel.... pazienza.

[steve]

LPs should simply fund VCs the way VCs fund portfolio companies -- in tranches, and with no commitment to continue funding beyond initial or segment inflow. Keep funding if your happy, stop if you're not. And for heaven's sake, only pay fees on invested capital, not committed!

I admire your attempt to rationalize the biz, but VC = Hollywood. And as William Goldman famously said about the movie business, "Nobody knows anything." In a year or three, the vast majority of LPs are going to take a huge painful shellacking (caused by a dearth of succesful exits.) At which point the VC industry will shrink down to a much much smaller size, as befits such a high risk, totally uncorrelated, inexplicable, irrational asset class, and as can possibly provide the type of returns (as an asset class) which justifies existance. And the vast majority of now-cool-guy VC partners will go back to earning honest, albeit smaller, livings, where secretaries are not provided hot lunches every day at investors' expense, and shit thrown at the wall is not mistaken for work by Jackson Pollock.

Oh yeah -- 99% of VC blogs will disappear also.

[leafar]

Evergreen is not the best solution Steve.
Illiquidity has a sense in this industry even if it creates some huge bugs.

[Je]

"...and shit thrown at the wall is not mistaken for work by Jackson Pollock."

Makes you wonder why LPs let themselves be snaked into thinking this was a good idea?

[Samir]

What makes a good VC? It's someone with industry/technical knowledge or senior level operating experience who has some context for sizing up markets and like people. But there are a surprising number of people out there putting money to work without either. Some of these people have access to money and a good sense of which smart guys to follow. I'm going to assume you like the Vulcan VC from your last blog because of this because I cannot for the life of me see what value he brings to the table other than Paul Allen's megabucks. Not that access to a very a big bank account isn't important, that's THE most important qualification for a VC, isn't it? I think people are beginning to realize that the days of poseurs in the Silicon Valley are over. When I see VCs without operating experience or something like experience with capital markets (which is only marginally additive), I am reminded of all the annoying, full of hot air, full of themselves biz dev, MBA types who flocked here not so long ago. Just what value do these sorts bring as "partners?"

[Peter Rip]

Samir:

Your comment about Steve is unfair and I can only assume you don't know him. I have had the opportunity to get to know *him*. He is much more than a face on a wallet. You may be cynical about VCs, but if you want to paint a specific person with the broad brush of generalization, get to know them first.

[Lee S.]

Kudos to the blogger for not deleting a post he did not like, but why don't you explain further what a venture capitalist without work (specifically start-up) experience brings to the table other than money? By definition, someone who hasn't really worked has not managed people, brought together a team, or launched a product. Isn't venture all about the people, the people and the people?

[Peter Rip]

Fair question.

Operating experience, while valuable, is not the only valuable experience. Look at the most successful VC in history, L. John Doerr. John was in Intel marketing and product management before KPCB. Yet, he brought an unprecedented (and I think unduplicated) appreciation for the confluence of technology, product markets, and financial markets to this business. A Board should have a variety of skills, not just more operating experience than the CEO.

So one important skill that many startup CEOs do not have is an appreciation for what is called the "long term cost model" of the business. As companies mature, the choices you made early on will often constrain or determine the long run cost model. The cost model determines profitability. A person who has analyzed many businesses on behalf of investors -- esp. public market investors -- can bring a perspective about how to think about the business that an operating guy does not. It is all about what experience you bring to the conversation. If you see enough successes and failures, you can start drawing inferences about causation in early companies. So, while you may not have the skills to assess whether the sales forecast is going to be achieved this quarter, you may have the skills to provide guidance about whether the sales process is repeatable enough to justify increasing the sales force, i.e., adding resources with positive contribution market. Just one of many examples of how VCs can add value without having been in-the-shoes.

There is no one-size-fits-all VC. A Board should be a constellation, not four-of-a-kind.