Megan McArdle concludes from “lowly MBA” analysis that it is impossible to get a net surplus from public investment. If lowly MBAs think this, then lowly MBAs faked their way through Introduction to Microeconomics which IIRC was in fact designated Econ 101 at my lowly public university:
…I do not think of spending money to build the bridge as a net increase in the country’s wealth. We exchanged money for a bridge worth the money we spent (or so we light-heartedly hope). One could argue that the bridge would generate more economic value than it cost in taxes and deadweight loss; one could also argue that it will generate less (and Japan has quite a few bridges of this description). But this is an argument for the bridge, not for bridge-as-stimulus.
[Added excerpt — TB:] Assuming that the real interest rate approximates the time preference of said American Taxpayer, and that the bridge is worth exactly what we paid for it, the net economic benefit is zero. [Emphasis added.]
I will grant that public investment demand curves may not always slope downwards, but let’s assume arguendo that they can. That is, you can in principle be a good technocrat, take the set of available public investment projects, and order them by decreasing return on investment.
Now let’s consider what projects should be pursued. If you consider funding only the most beneficial project, the good technocrat may find that the project returns, say, 30 percent annually, while someone may be willing to lend the government money for almost nothing. [Added TB:] This project is, in effect, worth more than society paid for it. So the project is approved and we move on to the next project. That project is by construction less beneficial, and if the supply curve for funds slopes upwards, the funding will cost more, too. Still, chances are the next project will pass muster, too. We go on and eventually come to a project where the rate of return is just equal to the cost of funds. So the marginal project indeed has no (or minimal) net benefit, but all of those “inframarginal” projects have positive net benefits even at the interest rate that equates marginal benefits and marginal costs.
The “economic surplus” is in fact the whole area between the demand and supply curve to the left of the point where marginal benefit equals marginal costs. The public benefits to the tune the returns on the inframarginal projects above the interest rate, and lenders benefit as a group because most of them would have been willing to lend money for less than the market interest rate. This is one of the fundamental results you should learn in introductory micro. The graph McArdle should have seen at some point is e.g. here.
Now the advanced economics point, and a practical policy consideration behind this post, is that the bridge-as-stimulus to fill an output gap can be justifiable as a pure helicoptering of money on firms and individuals involved in building the bridge. Greg Mankiw said so! (*) The bridge as bridge is icing on the cake.
(*) Possibly to his current regret.