Did VC Fund Size Cuts Work?
[UPDATED] Coller Capital last week released a survey of limited partners, in which 65% of respondents said that they expect VC fund terms to become more LP-friendly over the next two years. This was portrayed by some as a sea change in the LP/GP power dynamic, although I suggested that “LPs would have more credibility if they began insisting on changes to funds that have already been raised.”
My dissent was rooted in the memory of late 2001 and early 2002, when LPs fought back against the obesity of VC funds raised during the dotcom bubble. Sixteen U.S.-focused venture capital firms “volunteered” to cut their fund sizes by an aggregate of nearly $4.5 billion, including brand names like Accel Partners, Battery Ventures and Kleiner Perkins. The argument for such cuts – which were technically agreements not to call down a certain percentage of committed capital – was twofold: (a) Valuations and investment opportunities had fallen post-bubble, so less money was needed; and (b) VC firms tend to perform better when they focus on earlier (read: smaller) investments, so this was a return to successful knitting.
LPs were happy with the resulting management fee reductions, and GPs were happy that their LPs were happy (butter ‘em up for the next fundraise). All hunky dory, except we’ve never actually looked to see if fund reductions produced better returns. So let’s do that now:
I have obtained 12/31/08 IRRs for 15 of the 16 funds that cut their size, via a variety of sources (public filings, conversations with LPs and research provider Preqin). These funds are: Accel Partners VIII, Atlas Venture VI, Austin Ventures VIII, Bay Partners X, Battery Ventures VI, Charles River Ventures XI, Kleiner Perkins X, MDV VII (which was being “annexed” at last check), Mobius Venture Capital VI, Redpoint Ventures II, Pacven Walden Ventures V, Sevin Rosen Funds VIII, Trident Capital V, Trinity Ventures VIII and Worldview Technology Partners IV.
I was unable to obtain 12/31/08 IRRs for Meritech Capital Partners II, so did not include it in my calculations. [Note: This post originally used just 13 of the 16 funds, but I obtained IRRs this morning for Kleiner X and Pacven Walden V and recalculated]
What I found was a mixed bag. The 2000 vintage of “reduced funds” easily outperformed industry benchmarks from both Cambridge Associates and Venture Economics. The 2001, on the other hand, did not. Here is data for the 2000 VC vintage, from which nine of the reduced funds came:
Mean IRRs for 2000 vintage
Reduced Funds: -1.33%
Cambridge Associates: -1.39%
Venture Economics: -2.1%
Median IRRs for 2000 vintage
Reduced Funds: 1.7%
Cambridge Associates: -3.37%
Venture Economics: -2.9%
The best performing of the 2000 “cut” class was Charles River Ventures XI, with a 9.93% IRR (according to Mass PRIM). The worst was Sevin Rosen Funds VIII with a -25.2% IRR (according to CalPERS).
Now here’s data for the 2001 vintage, which had six reduced funds with available IRRs:
Mean IRRs for 2001 vintage
Reduced Funds: -4.88%
Cambridge Associates: 0.49%
Venture Economics: 1.3%
Median IRRs for 2001 vintage
Reduced Funds: -6.65%
Cambridge Associates: 1.07%
Venture Economics: -0.10%
The best performer in this class was Austin Ventures VIII with a 3.2% (according to CalPERS), while the worst was Bay Partners X with a -11.2% IRR (also according to CalPERS).
I’m not a huge fan of fence-straddling, so I combined the mean IRRs for the 2000 and 2001 vintages:
Mean IRRs for 2000 & 2001 vintage
Reduced Funds: -2.75%
Cambridge Associates: -0.91%
Venture Economics: -1.04%
In other words, the average “reduced” VC fund underperformed VC funds as a whole (both cut and uncut). This isn’t to say that these particular funds would have performed better (or worse) had they stayed pat, but simply to say that the act of fund size reduction, on its own, did not spark outperformance.
The median IRR for the reduced funds was 1.2%, but I couldn’t make a benchmark calculation out of the publicly available Cambridge or VX data. Now feel free to rip apart my calculations…
Related posts:
Ken said on June 25, 2009
You might want to change all the “x” to “%” since you are commenting on IRRs and not multiples.
Dan Primack said on June 25, 2009
You’re right Ken. Brain cramp. Fixed
Michael said on June 26, 2009
The reason the returns were not better nor particularly different from each other could be because they were all still funding the a lot of the same types of companies – just with smaller checks. None of these “top-tier†VCs was doing anything very different, so why would their returns be different? It might be useful to look at what these groups were NOT doing and see of there were successful groups (without the pedigree) in those spaces. Cleantech, SaaS, and new media come to mind as possibilties.
Originality may be more important than size.
Anonymous said on June 26, 2009
As someone who has spent over a decade in VC, I think you are missing perhaps the biggest reason GPs reduced fund sizes after the 2000 bubble…self interest, not the well being of their LPs.
Let me explain. For example, if you were an established firm and raised a $900MM (for demonstartion purposes only) fund that was 1/3 invested at the bursting of the bubble, it made sense for the GP to cut it in half. Why? Simple. The GPs knew the first $300MM went into crappy companies that showed little promise of return. If the fund was not cut, the next $600MM had to go up by 50% (absolute, not IRR) before any carry was paid to the GPs. Even if the GPs believed they could return those IRRs, they were better off starting over; take an LP loss on the original fund. Sure the management fees got cut, but they were (almost) made up with the subsequent $450MM fund. However, do the math on the carry. If the original fund was not cut and the first $300MM was marked to $0 and the remai! ning $600MM was eventually marked to $900MM, the carry to GPs was $0.
With the reduced funds: If the first $300MM was marked to $0 and the remaining $150MM (for a $450MM cut fund) went up the 50% to $225MM, the carry was still zero. HOWEVER, using the same increases, if the subsequent fund of $450MM went up 50%, that means $225MM of profit ($45MM of carry at 20%…management fees need to come out, but showing you the direction of the decision process). If the investments have a higher return, the math gets more favorable for the GPs. Thought you might be interested.
Again, this only works if you were an established VC (look at the roman numerals after the fund names) and had a high degree of confidence and/or cockiness to believe that your “goodwill” would lead LPs back for the next fund.
Anonymous said on June 26, 2009
Dan, I’m amazed that you continue to miss the boat on what was really going on with the fund size reductions in 2001. In all these cases, the deal offered by the GP to the LPs was to cut the size of a bubble-era fund in exchange for immediately committing to a brand-new fund.
For example, a $1B fund might be cut to $500M but simultaneously a NEW $500M fund would be created with the same LPs. Why was this happening? Because the GPs knew that their bubble funds were so far under water that they would never get to carried interest. By turning the un-called capital into a brand new fund they got to walk away from their previous train wreck and start anew with a brand new fund and brand new high water mark, all without taking any hit on management fee!
This was done with a lot of bullshit PR about how this was an ethical thing to do and the “right decision†for everyone but the reality is the GPs were screwing their LPs and the LPs sucked it up because they wanted to maintain their relationships with these GPs that had previously made them a lot of money.
So if fund size reduction is a hot topic again today, let’s at least set the record straight about what was really going on the last time around.
Signed,
A GP whose firm doesn’t believe in screwing LPs
VC Firm CFO said on June 26, 2009
I think there are a few things going on, none of which fully explains/accounts for the differences in performance between reduced funds and the larger sample size.
On the one hand, your sample size represents funds that have more typically been what we would consider “top tier†funds and hence we would expect them to be at least in the top half if not top quartile for their vintage years as compared to the larger sample sizes of both Cambridge and Venture Economics (more so for the latter). So this could sort of explain the 2000 vintage year comparison.
Another thing in the mix is how these groups define vintage year, with Cambridge typically basing it on first close (with respect to the LPs they track) and Venture Economics basing it on first drawdown. Thus timing of the underlying funds could account for some differences.
Finally, by reducing fund size these funds may have hurt their chances of digging themselves out into positive territory, which may account for why the 2000 funds didn’t perform “much†better and why the 2001 funds under-performed versus the benchmarks.
Our fund from this era bottomed out in 2003, but our investments in the latter half of the investment period (through early 2006) by far out-performed the bubble year investments, and allowed us to get into the top half and close to top third for our vintage year, whereas if we had cut it off in 2001, it would have most likely remained bottom quartile forever.
But again, this is where having a name brand fund helped them, it sounds like the LPs wanted the reduction, they got it, and then the GPs got to raise another fund to take advantage of better opportunities and restart the carry calc on a new fund.
Anonymous said on June 26, 2009
I would like to point out the elephant in the room: these returns are awful. 9.93% IRR is the best return from the entire sample- am I missing something here? I suspect that much of the VC field has been able to survive because LP’s, especially university endowments, feel a social responsibility to invest in it. However, just like the ballooning US deficit, things that seem unsustainable are generally in the long run unsustainable.
just.a.guy said on June 29, 2009
Clearly these IRR numbers include a lot of unrealized returns. If Austin Ventures VIII is sitting at a 3.2% IRR after ~8 years, that’d be a 1.29x return, and if memory serves that was an $830 million fund. So that all implies returns to GP’s of $1.07 Billion.
Can anyone point to any exits from Austin Ventures VIII that lend any credibility at all to that IRR figure? My guess is there are a lot of living dead zombies propped up in that fund and that it will eventually collapse in value. Hopefully for them that happens after some *real* returns emerge from funds IX and X.