azeem.azhar

The very kind folk at Thomson Reuters have provided me with some of the data that Fred Wilson was asking about in his two blog posts on the VC maths problem (I, II).

Starting with US data from 1994 to 2008:

A total of $69.5 bn was raised via 998 IPOs of VC backed firms (of all flavours) in the US. The total post offer value was $396 bn.

There were 3000 M&A transactions from 1994 to 2009 YTD of VC backed firms in the US. Of these, 1460 were disclosed. The sum of  these deal values was $219bn, the average was $ 150m. On average each VC-backed firm had received $32m in financing and in-toto they had received $31m in venture backing.

The data for IPOs and M&A is sufficiently different to merit treating them differently.

On IPOs, the mean offer amount is $70m but the median offer amount is $49.50m. This means that we have a skewed distribution, as the median is lower than the arithmetic mean. I’ll hold by my previous comments that this is best approximated by a Poisson distribution, as illustrated above, than by a Powerlaw (partly because I did this analysis in 2007 on the same dataset and that was the conclusion I reached.).

The average age of deals to IPO was 7.13 years, with the median age being 5.80 years. (This once again is a skewed distribution.)

Now this data is horribly skewed by the glory years of 1999-2000 which represent 40% of all IPos in this 15 year period, and 50% of funds raised. Let’s just keep this data in.

Keeping it in, let’s reprise that VC backed businesses in the US exited as a combined IPO value of $396bn in that 15 year period.

On M&A transactions, we should make one assumption: that non-disclosed transactions performed the same as disclosed transactions. In likelihood, non-disclosed transactions are likely to be worse than disclosed ones. Only half were disclosed and represented deal values of $219bn in toto, so assume the total deal value was twice that so $440bn. Now these M&A transactions were most likely the startup acquired by  a larger firm, and in many cases substantially larger, so let’s assume that the VC-backed firms acquisition value represented 33% or so of these transactions. Total value created by these VC-backed firms is than $140bn over this 15 yr period.

As in the IPO market about 50% of this value was in the g(l)ory days of 1999-2000.

In toto

So by that estimate, over the 15 year period, VC firms generated $396 + $140 bn of value = $540 bn, or an average of $36bn per annum, of which about 50% happened in the glory years. This means that IPO and M&A activity in those insane years was seven times higher than the long run average.

If we control for those glory years and assume they were just trend line years, then the average annual exit value for IPOs drops to $14bn and for M&A to $6bn for a average of $20bn over the past 15 years.

In Fred’s original post, he argues that

So here’s the venture capital math problem. We need $150bn per year in exits and we are getting about $100bn. That $100bn produces roughly $50bn in proceeds for venture firms per year. After fees and carry, that $50bn is around $40bn

Now, by these rough estimates, we aren’t getting anything close to that. In fact we are way off. By these estimates VC in the US is only just about pushing out what gets poured in.

The top years for exits were 2o04 and 2007. A year like 2004 where total exits, using my methodology, came to $46bn , and  2007 with similar numbers tell very different stories. 2004 exits would have been funded by funds of the 1999 to 2001 vintage, which were bumper crop years where at least $150bn went into VC. (Don’t ask where the rest went but at 2% management fee about $1bn went on mortgages for houses in Woodside).

But for 2007 years, the appropriate fund vintages would more likely have been 2002-2004, or the dry years of VC investing where $30bn was raised, of which $17bn was raised in 2004. These funds will have exited starting most like in 2005 (which was a bumper year for M&A) and in most cases will still be holding some good companies today.

But to wit, to a Fred’s main question: is there sufficient value creation by VCs relative to the amount of money raised by funds? The answer seems to be that the major determinant as to whether there are adequate returns for LPs is the amount of VC raised. There appears to be an appetite in IPO and M&A markets that, irrational exuberance, aside can be sated.

The economics of that argument makes sense. An increase in the supply of private company capital will diminish returns from that capital over all. (Of course, the structure of the VC business means that there are stll decent returns to be made for a few, mostly pre-determined firms. See Private equity or Personal Equity.)

But it also appears that the industry surfs very very close to the wire between being profitable and being unprofitable over all. Despite what VentureBeat wrote some days ago, industry returns are actually much less tendentious if you control for the bubble years (which you desperately need to). Fred’s top down maths makes the chasm shockingly clear.

(You can find the Thomson Data here.)

Popularity: 23% [?]