P.S. We’d still do better with a failed solar company than a hedge fund manager. When a tech company fails, you get scads of engineers who learned a lot of valuable lessons both about how to do things and how not to. Tech recruiters call this scar tissue, and serious tech firms want engineers with scar tissue. The space wasters, like Dilbert’s home company, go for spec matching. It’s this non-book learning that pushes things in a way that tech transfer from universities to the field cannot. A good example was Henry Ford, who worked on electric cars at Detroit Power, then decided to make Otto cycle driver automobiles having learned the business at DP’s expense.

P.P.S. Don’t overrate the risk taking abilities of the capital market. Private companies, even venture capital firms, are amazingly risk averse. EVERY major new technology has required government backing. Who else but the government would have come up with the idea of burying perfectly good food in the dirt and found agriculture?

My point then is that in a fictitious world of perfect shared information (i.e. a world where market prices are efficient) the total cost to the government of the dynamite is the same regardless of how the government procures it – by subsidizing the risk of building new plants, by building its own plants, etc. Your example doesn’t recognize this because you don’t add up all the costs of procuring the dynamite. Under COA A you consider the cost to the government of one strategy for supporting the building of factories, but the government is still left needing to pay additional money to buy the dynamite itself. Under COA B you consider the cost to the government of a strategy for building a fraction of the capacity it needs, but not the full capacity, and not the costs of operating the plant(s). In each case the total costs to the government will be the same once you carry the analysis to the end (to dynamite in hand), unless there are information inequalities (which there always are).

Given information inequalities, the relative imbalances in knowledge determine the proper risk acceptance postures of the relevant players (government and industrialists).

Providing illustrations (more allegory than metaphor, I think) is valuable – keep it up. But I think that your dynamite-factory example needs work.

There is a key difference between course of action A (guarantee four loans) and course of action B (build a dynamite factory) besides the three – vs – one (or ¾) issue that you focus on. The other difference, and this is a whopper, is that under COA B the government owns the factory. Under COA A the factories needed to satisfy the government’s appetite for dynamite are built, but the government has no claim on their output. Under COA B the government can satisfy only 1/3 of its need, but at least it can do that much.

Given efficient pricing, the total cost to the government of securing the dynamite it needs will be the same regardless of how the government goes about doing it – guaranteeing the loans and then buying the output of the new factories, or building and operating factories for itself. The question of whether to do it one way or the other turns on presumed inefficiency of pricing: does the government have inside information that lets it price some item of risk, or investment, or spending, differently than the market? If so, then the government should act so as to secure for itself the benefit of the price difference, thereby incidentally sharing its inside information (i.e. affecting market prices).

If nobody has any inside information, then there is no reason for preferring one COA over another. But what kinds of inside information might exist that would allow the government to benefit from a pricing inefficiency? The most obvious inside information that the government has is information about its appetite for dynamite in the first place. This inside information has nothing to do with the probability of a dynamite factory blowing up, and therefore shouldn’t affect the price of risk associated with a loan guarantee, but it does affect anticipated future pricing of dynamite – a new substantial consumer is known to the government to be about to enter the market. Under these circumstances, the smart money (in this case the government) would benefit from positioning itself as a dynamite supplier before the market learns of the government’s intentions and the prices of dynamite precursor chemicals and dynamite factory components rise.

On the other hand, it is also possible for the existing dynamite industry to have inside information. For instance, perhaps only dynamite companies have the technical expertise to realize that because of new EPA regulations banning explosive-stabilizing additives, newer “green” dynamite factories are going to be more likely than legacy factories to blow up, perhaps 1 in 3 instead of 1 in 4. Under these circumstances, the government will mis-assess the expected cost of loan guarantees, and the dynamite companies can secure financial advantage by suckering the government into such guarantees based on analysis of historical but misleading loss rates.

So what is government to do in the general case? In general, sound investments are distinguished by the most knowledgeable parties being willing to accept the greatest risks (in return for a favorable negotiated share of expected returns), and having the smart money shun risk in a shared investment is a sign that the whole enterprise is probably a negative NPV proposition. So the government should use loan guarantees when it thinks it knows the risks better than the investors it is dealing with, but shun them otherwise. Government information superiority will rarely be the case when loaning money to specific companies like Solyndra, which are too small for the government to have any special knowledge about them at all.

I understand what you’re saying, but bear in mind that literally everything the government does distorts the economy in one way or another. Even a nightwatchman state will be subsidizing markets for guns, badges and silly-looking hats. You can have this conversation about everything the government does. I bet you and I have different opinions about what the government should be doing, but I’m not really interested in getting into the specifics of all that.

I don’t want to ponder the best way for the government to create a dynamite surplus. I want you to explain why the government should create a dynamite surplus in the first place.

It’s true you make the initial assumptions a bit tongue-in-cheek, but you follow them to conclude seriously (I think) that the government should output money by creating surpluses. I think this means the assumption and its context are valid considerations.

If the government wants to output money through inefficiencies such as a surpluses then many approaches, including loan subsidies, will “successfully” accomplish it. You’ve simply defined success down until it matches actual results.