Solyndra and the Dubious Benefits of Loan Guarantees

I do apologize for the all Solyndra, all the time tenor of the blog recently.  But I wanted to expand upon Friday's post a bit, because I had a blog conversation with the excellent Tom at Manifest Density which bears on the subject.

Tom argues that leaving aside the question of whether Solyndra, in particular, was a good idea, loan guarantees may represent a good way to leverage government funds in risky situations.  

All of this is a separate question from whether the government should be intervening in the economy in this manner. I think that's a conversation worth having. But it should be a conversation about programs, not transactions. To focus on one failed loan is a bit like pointing to a lottery winner and saying, "That's how they think we're going to pay for our schools -- by handing out checks? What a bunch of dopes!" But of course this ignores the bigger picture. Perhaps there is a problem with that lotto payout, but that's a different and frankly less important issue than whether it's wise for the government to run a lottery. This is why the intense focus on Solyndra strikes me and many others as a transparently partisan feeding frenzy rather than a considered discussion of energy policy.

I almost precisely disagree with this.  The question of whether or not we should subsidize solar panels should be held separate from the questions about how we should subsidize solar panels, and whether the decision making process for allocating the subsidies worked.  But as it happens, I think the questions about how we should subsidize things is important, and that one of the answers to those important questions is "not with loan guarantees".

Full disclosure: I have something of a bias against cosigning or guaranteeing loans; I think it only ever looks like a good idea because of the peculiarities of government accounting.  So Tom points out that by guaranteeing loans for four dynamite factories, each of which has a 1-in-4 chance of blowing up, you can mobilize four times as much spending as you would if you just built a dynamite factory--and that you've got a diversified portfolio of factories, to boot, instead of one factory with a 1-in-4 chance of suddenly not existing.

But as I argued last week, the capital markets have ways of dealing with risk: you spread it out over multiple people (with a loan syndicate, say) or you raise the price.  That doesn't necessarily mean charging a higher interest rate; you can also offer various sweeteners such as stock warrants, or issue equity (which is very expensive, from the point of view of existing shareholders, because it represents a permanent claim on future earnings, rather than the temporary and limited claims of a loan.)

The reason that markets won't invest is not that they think a project is risky, but that they think it is not likely to pay back its investment.  Neither private investors, nor the government, can build a winning portfolio out of a large group of investments with negative expected value.

Loan guarantees look good if you start by assuming that the government is going to spend a certain amount of money building solar panel plants, and then ask yourself how to make that money go furthest.  But let's question that assumption: why should the government build solar panel plants?  The reason that the private market will not build them is not that the individual firm has too high a risk of failure, but that the returns on the investment will not compensate for the risk:  either too few solar panels will be sold, or they cannot be sold very profitably. Neither of these problems are problems that loan guarantees fix.

However, they are problems that the government might be able to fix, by subsidizing the purchase of solar panels.  If the profits are right, lots of capital will mobilize to build plants to exploit them.  We get the same sort of "multiplier" that we got from a loan guarantee.

This has a number of attractive features, compared to loan guarantees: it doesn't require the government to pick a winning company, and more importantly, it only costs you money when the solar panels are actually installed.  By contrast, with Solyndra we have an enormous new factory that consumed resources and carbon in the building, and a warehouse full of unused solar panels.

But from the point of view of a bureaucrat, I suspect this is not as good.  With a loan guarantee program, you can point to a shiny new factory, and a hard figure of how many billions went out in loans from the private sector, which can be happily contrasted with the administration's tastefully understated spend.  With direct subsidies, you get a lot of solar panels, but you can't name the jobs that were created or point to hard figures on how much additional investment in "Green Jobs" occurred as a result of your program.

However, most of us are not bureaucrats, and we should be skeptical.