The Scale of Solyndra

Continuing with my "all Solyndra, all the time theme", I want to expand upon what I wrote in my last post:

I am a huge fan of Tim Harford's Adapt, and indeed am now attempting to write a book of my own on a related theme. But it isn't a brief for hog-wild speculation on any sort of project. Scale matters--you can't take billion-dollar fliers on too many ideas. Appropriateness matters--retirees shouldn't put their living expenses in tech stocks. And the potential payoff matters--we should not invest billions to develop a slightly better form of white-out, or into experiments that have a 1-in-a-trillion chance of developing a low-cost way to turn lead into gold.

In the last post, I addressed the potential payoff of the Solyndra venture, which seems iffy on the payoff from either a financial or an environmental perspective; their process was unproven, and their competitive advantage too easily eroded.  Today I want to talk about the scale of the investment.

A lot of Solyndra's defenders have been arguing that if the government is going to play venture capitalist, this is simply what's going to happen:  venture capitalists accept that a lot of firms in their portfolios are going to go bad.  This is true--sort of.  The private equity market is not uniform.  While there are of course exceptions, in general, the more money that is being poured into a company, the farther along it has to be towards sustainable profitability.  Angel investors may accept that the overwhelming majority of their investments will fail, but  by the time you're pulling in $1 billion worth of capital and getting enormous bank loans, you should have some pretty solid proof that you have a product that people are going to want to buy, at a price at which you can profitably produce it.

Obviously there are exceptions to this rule, but broadly, you shouldn't think about late-stage the way you think about an angel investment. Commenter Circleglider, who's in the business, had a lengthy discussion of this the other day:

While venture capital is of course a form of private capital, private equity firms and funds typically invest at later stages of a company's growth cycle (see this diagram for an example of what types of investors participate at the various stages of a company's growth cycle).

This model assumes a stepwise function where, upon significant events, increases in a company's valuation are linked to decreases in investment risk. The private capital market is segmented based upon this stepwise function, with firms investing larger sums at later stages with lower risk.

Early stage investors (angels and venture capital) must evaluate opportunities based on very little hard data. Factors such as market size and growth, management team and technology assets drive decisions. At this stage, things like five-year pro forma financials and balance sheets are pretty much meaningless. Because investing at such an early stage is a very high risk endeavor, intangibles drive decisions and relatively small sums are invested (anywhere from $100K to $10M).

Venture capital funds don't expect every deal to succeed. Portfolio return rules-of-thumb can vary from 80/20 to 60/30/10, where maybe only 10% of all investments turn a profit. For the overall portfolio to succeed, those 10% must perform spectacularly.

Later stage private equity firms invest in on-going concerns after basic risks about markets, technology, management and business models have been resolved. Over an entire portfolio, most deals are expected to make a profit. This means PE firms can invest relatively large sums ($10M to $1B), and are comfortable with much more modest returns. Because PE firms invest in relatively mature companies, they can apply more traditional analytic methods.

Normally, internal process controls help prevent bad decisions at all investment firms. However, when these controls are either ignored or non-existent, it's usually the PE firms that fail spectacularly. This is because they can invest much larger sums (and therefore loose much larger sums), and because they assume their targets are on-going concerns (and are therefore amenable to traditional analytic methods).

Green technology occupies a special niche here in Silicon Valley. Yes, some VCs invest in green deals. But most don't. This is because these deals are not viewed as being able to stand on their own merits - for any of them to succeed, market distortions are required, typically in the form of government subsidies, tax credits, mandates, regulations or funding guarantees. This means that green deals are basically political deals - whether or not your deal works depends entirely upon what happens in Sacramento or Washington. So VCs who do invest in green deals are comfortable with these distortions, and factor them into their investment decisions.

Solyndra appears to be the result of a combination of many of these factors: certain investors were comfortable with the political aspects of the deal, they were institutionally predisposed to assume the Company was an on-going concern, and very large sums were available for investment. As a class, it appears that the total exposure of VC firms in Solyndra was around $150M (data from Solyndra's S-1 filing with the SEC). Private equity firms, however, appear to have sunk nearly $650M into Solyndra.

This is a really spectacular failure. Yes, similar things have happened in private markets before. But they've usually been associated with market bubbles. Maybe there's been a bubble in green tech - if so, only a select few got the fever. And one of the preconditions of participation was a willingness to ignore fundamentals and accept the necessity of direct government action in order for any individual deal to succeed. Some may call that a wise national industrial strategy. Others call it crony capitalism. I call it stupid.
In fact, as far as I can tell, the bulk of the late stage (post DOE's conditional approval of the loan guarantee) funds were supplied by one source: Argonaut Ventures, an investment arm of the George Kaiser Family Foundation, with another $50 million supplied by the foundation itself.  Together, those two entities provided at least $388 million out of $949 million worth of the non-government debt and equity financing, or about 40% of the total.

However, that number is almost certainly higher--perhaps much higher--because the only hard information that anyone seems to have on the company's financing comes from the initial filing for the company's IPO (subsequently withdrawn when the auditors pointed out that they were hemorrhaging cash with no relief in sight). When the IPO was withdrawn, current investors stepped in with emergency funding, presumably in an attempt to salvage their prior investments. (The DOE are apparently not the only people who have trouble facing up to the reality of sunk costs.)  The consensus seems to be that Argonaut participated heavily in those rounds, but I'm not sure anyone except the principals knows exactly how heavily.

So it's not exactly like there were tons of private companies rushing to throw hundreds of millions of dollars at this thing--there seems to have been one organization that threw a whole lot of money in there, and a bunch of firms that threw in substantially less.  Maybe this is common, but still--if it weren't for one apparently very well-capitalized family foundation, it's not clear that Solyndra would even have made it to the point of getting that loan guarantee; from 2008 on, they're providing the bulk of the funds at every capital raise.  To be fair, there were well-known firms like US Venture Partners and KKR participating for much smaller (though in the case of USVP, still quite large) amounts.  But overall, the funding seems less proof of the market's judgement of Solyndra's chances than the George Kaiser Family Foundation's judgement of same.

Now, unlike some libertarians, I have no trouble with the government spending these sorts of funds on something like the Large Hadron Collider, where "We have no idea what is going to happen, but whatever it is will be cool and informative!"  But I am distinctly prejudiced against plowing half a billion dollars worth of government funds into a company to see whether they can finally get their manufacturing process to work.  The point of basic research is that any answer is in some sense equally interesting--it is useful to the world to know what is not true, as well as what is.  But a company is aimed at only one result--making a profit.  If you aren't pretty sure of that result to begin with, then you shouldn't take massive risky bets.

VC's handle the risk problem by diversifying and getting a big return on the few successes.  But when you take really large bets, you can't diversify very well.  The other investors are at least plausibly at "calculated risk" amounts (and later, the tragically common "failed to understand the concept of sunk costs" amounts.)  GKFF, on the other hand, seems to have gone all in--I have to wonder how big a percentage of their asset base this constituted.

Of course, you can argue that cleantech simply can't prove its manufacturing processes work with less than a billion or so worth of capital.  Maybe so.  But that should give us pause.  We have lots of very large industries, from auto to electricity, that managed to break even without government subsidies; we have lots of industries with huge positive externalities, like the internet, that managed to scale organically over time.  Why is it not possible for cleantech to do the same?  Because it is competing against cheap carbon, you will say, and that is true here, but it is not true in Europe, where the cost per killowatt hour is substantially higher than ours.

More to the point, cleantech has in fact received billions and billions in both front end and back end subsidies.  A billion doesn't seem to be enough.  This makes one wonder if any sum will be.