Paul, Fred and Carlos - The System Dynamics of the VC industry

Paul Kedrosky and Fred Wilson have been in furious agreement - it seemed to me - about the restructuring of the VC industry. Last year they both opined that the VC industry was too big - Paul came at it from the angle of lousy returns over the decade implying that the industry was economically ineffective, Fred noted that the industry was chasing too few good deals with too much money.

Today the both wrote new posts on the topic (Paul's here and Fred's here)

Fred's worry is that silly money is coming back in:

My concern is that investing went up in the second half of the year to a $5bn/quarter rate, which is $20bn run rate, a bit above the number that I think is optimal for the industry's returns.

2010 will be an interesting year. If VC investments go back up to $25bn to $30bn per year, then the diet didn't stick and we are back to an overfunded industry that will produce subpar returns on average.

Paul's take is that VC may just be broken more than that now.

Finally, is it possible that venture is just so badly broken that we have no idea what the right amount of capital is? Yes, it is possible that we could see 10x as many startups and find that it is a distribution with many local maxima, some of which are much more attractive than the current level of investing. To be honest, I would love it if that were true, because it would be wonderful for the economy, and this assuredly isn't physics -- annual outlays is not fixed like Avogadro's constant. Good luck making that case, however, to burned limited partners who have just watched an expanded venture industry deliver the worst returns in its history.

Its at this point that people like me wot do strategy and game theory and system dynamics and all that stuff start to get interested in What Happens Next - what is the system dynamic of the VC industry? Instead of using my own brain I asked Google, and found the model I posted above (from a rather nice little paper by Carlos Yepez, Carlton University, Ottawa).

For those who are not familiar with system dynamics, the trick is to look at the model and find out where the postive and negative feedback loops are. When this model works well, these loops are all positive - Great Entrepreneurs give VC's great opportunities (right hand loop), which allows them to create great exits (bottom left) which then informs the market for investors (top left)

So, eyeballing this you can see the following is happening:

Firstly, if all systems are positive then the overall system grows - but at some point it hits a limit. I agree with Carlos that the limit is the appetite for great exits by the public markets. This has now dried up, so they are now looking at trade sales, which by and large have lower exit values. At this point Carlos's model shows the bubble starts to unravel. But it takes time because...

Secondly, the system is not self - regulating. Just because the Great Exit loop is stopping does not mean the Entrepreneur or Investor loops do - in fact I'd guess they are picking up old information about Great Exits so they are actually accelerating - more Greedy Investors and more Eager Entrepreneurs flock into the system. The VC matchmaking market expands on the money flooding in and then increasingly desperately has to find places for the money to go, so eventually has to start to fund the Entrepreneur Supply Market itself. Enter Incubators etc etc.

Thirdly, that increases the velocity of supply of Eager Entrepreneurs, but that means that the success ratio of Great Companies available falls, so you start seeing much more investment by the Didn't Get VCs in the Me-Too companies (aka the Dumb Money).

Fourthly, this Dumb Money phenomenon happens because at this point the lower values of (trade) exits means the VC's need higher volumes to keep net surplusses the same, but this is impossible so the VC's ROI drops. And the optimistic belief is that the next one just may be The Next Google. This is still not a problem so long as the Eager Investors think there is still a great return.

I think this is where we were in the cycle till about a year ago. Today, the Startup Promotion Engine is still in full gear (TechCrunch 50's and all the rest are GO!), but the public markets are by and large closed, and the trade market is slowing - even Google can't buy everything.

So the big question is what are the Investors doing now? Eager Investors are of course herd animals, so the last of the herd is still rushing in where the Angels are scared to tread. But at some point they must decide that this is not the best game in town and the herd reverses. If that happens then supply of money dries up and you get a bust - which is what Carlos' model shows happens (He assumes a 36 month lag before the light dawns, but Paul's data implies it may be far longer).
And as you may know, busts typically overcorrect themselves after bubbles.

So what happens next?

Firstly, I think Carlos's model of 36 months lag is too low for the Investor Herd - so the bust may take longer to materialise. There is the Great Recession factor, ie some Investors have reached their funding limits, but - in my humble opinion - this collapse will tail off slowly over a number of years and could be further ameliorated by schemes for governments to "cultivate innovation as a national priority", and tax breaks for funding startups etc etc. I expect to see lots of calls for all that over the next few years

Then, if you follow the model the VC industry shrinks, but the supply of Entrepreneurs doesn't immediately. Now, if the average funding per Entrepreneur goes down (and there are a lot of signs that they need less money, for software services anyway) so the No. of deals reduces by less - but as there is less surplus in the chain the VC's have to do simpler deals to make margin, and there is likely to be massive consolidation to get operating synergies. (There is an issue here about how good VC's are at picking winners - statistically its not great, so they have to keep on with wide portfolios - but you can expect less of the obvious lemons to be funded)

One option to mitigate this is for the VC industry to manage its own decline (which is why Paul floated the Cartel word methinks) but I think it will do so in a more Darwinian manner.

The other option is to exit 'Net Tech and find cycles where the ROI stays higher - hence the rush for the GreenTech/CleanTech bubble but this too will pop. (I see Vinod Khosla recently started to talk about agriculture - does he know merchant banks were buying Ukrainian farms just before The Crunch?).

But those options require big capital investments and people who are really good at getting things right, ie the weak VCs will die but the good ones will win, and the industry will rebuild. In about 10 years time.

And then the system dynamic predicts the next bubble