If you have been following this blog, you may have detected a theme, I don’t like the public wasting money on un-needed infrastructure. Of course no-one endorses “wasting” money, we just disagree what is wasteful and what is an investment.
I think the answer is obvious in retrospect. A successful investment had a positive “return on investment” at or above market interest rates. An unsuccessful investment (or waste) had a negative return on investment. Projects with positive but below market rates of return sit in an analytical purgatory.

In prospect, I think I can assess forecasts accurately, and project advocates are not to be trusted for a variety of well-understood reasons ranging from optimism bias to strategic misrepresentation. Unfortunately, other people also think they can assess forecasts accurately, even if we know they can’t. If we had better mechanisms for requiring forecasters to be more accurate, we could mitigate these problems.

When these projects are small, it is not terribly important. The analysis of benefits and costs should not be costlier than than the benefits from the analysis (i.e. the difference in total welfare from a build/no build or a build this vs. build that decision). But when the project proponents ask for hundreds of millions of dollars (or more), we should be paying more attention.

A large number of projects seems to fall in the category of what I will dub “Mostly Empty Syndrome”. MES projects are infrastructure that are underutilized. Mostly they are not underutilized by design, but by mis-forecast. There are facilities however, like NFL stadiums, that are in fact underutilized by design, with lots of marginal events schedule to mitigate the grossness of the structure.

These are projects that serve purported needs, but those needs don’t materialize. Or they just are insufficient to justify the cost. Or they can be met in a different way. A few examples are listed below:

• Vikings Stadium – As I suggested before, they really only need to play in a TV studio. More importantly, the facility is unused or underutilized 355 days a year. The worst aspect is that they could not somehow figure out a way to share the stadium with the college team at a university that is adjacent to the stadium site.
• Stillwater Bridge – The old bridge required a replacement. The new bridge is overkill for the actual demands on the road.
• SPUD – It is basically slated to serve 350 train passengers a day, and a few buses until lots of speculative rail lines are opened.
• Northstar – Some forecasts greatly over-estimated benefits.
• Maryland’s Inter-County Connector is failing to realize forecasts of demand.

There are some counter-examples perhaps, projects that were long considered white elephants, though eventually demand caught up with supply. Dulles Airport meets this criterion. That still does not mean it was a good idea to build it when it was built, even if it is heavily used today. The 20 years of underutilization are fixed capital that could have been better spent in some other way.

There are other counter-examples, projects that were expected to fail (by someone, though not by proponents) that exceeded expectations. The Pennsylvania Turnpike comes to mind.
What connects MES projects?

1. I believe history will show that they are designed and pushed through by politicians serving narrow interests rather than by market demands or public sentiment. That is, they are generated by top-down rather than bottom-up processes.
2. They are large and require special treatment.
3. They are backward looking, built near, at, or past the maturity of the technology they represent. Air travel was still growing when Dulles was built, and the market grew into its capacity there. Auto travel was still early in its cycle when the Pennsylvania Turnpike was built. In contrast we now have peak travel, long ago passed peak railway, and are hopefully at peak football.