Finding ways to fund the construction and maintenance of major infrastructure projects will be a continuing challenge during the 21^st century.  But we must ensure, if we decide to pursue alternative funding mechanisms, that they do not have more problems than what is being replaced.  Take the excise tax on gasoline and diesel fuel as an example.  This tax has long been the primary funding mechanism for constructing and maintaining the American road system.

The Wall Street Journal recently published a story discussing some alternatives.  As the WSJ notes, the core issues are that (1) costs for building and maintaining roadways are generally increasing more rapidly than gas tax revenues, and (2) progressively improving vehicle fuel economies and increasing proportions of alternatively fueled vehicles on our highways mean that the tax revenue per mile driven is expected to decline significantly.

Point (1) receives far too little discussion.  Why have infrastructure costs increased so much in recent years?  Yes, commodity prices are generally up, but so are all the other costs of construction.  Corruption, collusion, and unnecessarily high construction standards are probably significant components as well.  Despite overwhelming public political correctness when it comes to infrastructure, the time is long overdue for all parties (especially politicians) to get to the real bottom of the rapidly rising infrastructure cost issue.  The mere claim that a mile of roadway should cost a specific amount of money to construct and maintain is in no way reasonable on its face.

We need to find ways of introducing more competition into the infrastructure arena.  Removing the monopoly of the professional engineering and technology/technician associations would be a good start.

As they say about democracy, it is the least worst system.  Based on my analysis of the various alternatives, this reasoning also applies to the current gas tax and roadway funding mechanisms (although I leave room below for a potentially superior pure general revenue approach).  The WSJ article in question presents and briefly analyzes several gas tax alternatives.  What follows is my cursory assessment of the fatal flaws in each approach.

1. Tax the Miles: As the WSJ notes, "[t]he idea that gets the broadest support is to take the user-fee piece of the gas tax to its logical conclusion: tax motorists on the miles they drive. Many economists argue that such a tax -- known as a vehicle-miles-traveled tax or mileage-based user fee -- is the fairest, most sustainable replacement for the gasoline tax."  This idea is good in theory, poor in practice.  The issue isn't accounting for how many miles were driven on which roads, when, and by whom (we can use relatively simple existing technology for that -- albeit with massive and valid privacy concerns); the issue is in assigning costs (both spatially and temporally).

If you vary the cost on a particular piece of road over time to account for congestion (or other factors), you may remove cost certainty for those needing to use the roadways.  As various pro-business advocacy groups often remind us, cost certainty is a core component of a stable economy.  Remove it, and you introduce instability.  Much of our economies is based on using roadways if and when needed, and unless we move to a highly undesirable centrally planned economy, many citizens, businesses, governments, etc. simply cannot control when and how they need to use the road system.  This cost uncertainty has not been adequately considered, especially for those entities highly sensitive to variations in travel costs.

But how do you price use of a road in general?  Different sections of roadways can cost very different amounts, even over a short distance.  Think of a city in a region of moderately complex topography and geology, with steep hills, different types of subsoil, etc.  In many cities, one even encounters topographically very different road conditions over a short 3- to 4-mile drive home from work.  The winter maintenance costs for a sloped roadway (e.g., plowing, sanding, salting, general repairs) may be much higher than that for the road at the bottom of the slope.  Indeed, weather conditions change appreciably with elevation (even over small amounts).

How do we deal with this?  Average the costs?  That isn't fair, either (and fairness -- the "user pays" model -- underlies the rationale for this approach).  Why should individuals who live in the lowlands subsidize those who chose to live in the highlands?  That makes no more sense than just subsidizing the entire national road network.  Reducing the scale of incoherence is no less irrational.

Some parts of a town are built on very good subsoil conditions that require much lower construction and maintenance costs than do other parts of the town built on poor subsoil (say, those with high clay contents and/or high water tables, filled in wetlands, etc.).  Why should residents who chose to live along routes with good subsoil conditions subsidize those that did not?  Once again, they shouldn't, but think about the level of spatial pricing detail needed to accurately price road travel to match road construction and maintenance costs in even a small town.  Mind-boggling.  And back to the "just average the costs over reasonably small areas" counter-argument: what scale of averaging should we use in populated areas?  A hundred yards?  One mile?  Ten miles?  None of the answers is rational, and the costs of obtaining, processing, and auditing the collected data would be large.  Perhaps better to leave things as they are.

What about less populated regions (be they urban, suburban, or rural)?  The road cost is the road cost.  If only five people live at the end of a 20-mile-long road, think how expensive their mileage taxes will be to cover construction and maintenance costs.  Well, we should subsidize them, you say?  Well, why not subsidize the whole system? I say.  It's impossible to make rational and defensible choices that everyone will agree on regarding which sections of roadways we should subsidize, how, and why.  And if we selectively implement the "tax the miles" approach to just certain urban areas or major highways, that's discriminatory as well.  I simply cannot see a coherent manner of implementing this method universally, with full adherence to true cost recovery principles for each driver.

2. Tax the Roads: As the WSJ notes, "[m]any support a more limited form of mileage-based user fees: toll roads. Relying more on tolls is already helping states make up for lost gas tax revenues; over the past decade, about a third of all new limited-access road miles have been paid for with tolls."  Fair enough, but really, tolls are practical only for limited-access roads and bridges.  For the vast majority of roads in the USA, tolling is impractical.  Why should I pay the full toll for a road if I am going to drive on only one third of it for my trip?  We cannot have toll booths every hundred yards on every road.  Thus, it doesn't get us anywhere near the answers we need.

3. Index the Tax to Inflation: As the WSJ notes, "[i]n the near term, some favor changing the gas tax so that it at least keeps up with the rising cost of construction without requiring lawmakers to cast a series of politically unpopular votes to raise the tax rate."  This doesn't address the core problem of increasing fuel efficiencies over time and the higher proportion of alternatively fueled vehicles expected in the future.  It's politically easy to sell (as easy as a tax increase can be, of course), but it fails to meet our needs.

4. Tax Oil, Not Gasoline: As the WSJ notes, "[a]nother way to fill the gap in transportation revenue and needs is to broaden the tax base, replacing the current federal tax on gasoline and diesel fuel with a levy on every barrel of oil consumed in the U.S."  This is dunce-cap ridiculous.  Why not just tax gas and diesel (the refined products of oil), as we already do?  If a real problem is that increased fuel efficiencies and more non-fossil fuel-powered vehicles reduce gas tax revenues per mile driven, and we're in some crisis because of this, how does shifting the tax from gas/diesel back down to the starting oil (much of which is used for non-transportation-related purposes) help out?  Answer: it doesn't.  The proposal is nonsense, and far less fair than the current system.

5. Tax Cars: As the WSJ notes, "Washington could also fill part of the gap in gas-tax revenue by taking a page from the states and assessing a charge on vehicle registrations.  The 2009 transportation-finance study estimated a federal fee of $2.75 for cars and light trucks and $5.50 a year for heavy trucks could raise $1 billion a year."  This takes the proverbial cake for dumbness.  And why should those who hardly ever use their vehicles pay the same amount as those heavy drivers?  We needn't say more.  This approach lacks any semblance of rationality as a potential solution (even a partial one).

Of the five options above, the "tax the miles" approach is the least problematic, but even it is fatally flawed.  As much as it may hurt some to hear, the best approach may be to remove the gas tax entirely, not pursue any of the five options above, and just build and maintain roads out of local, state, and federal general revenues.  Is this unfair and arbitrary?  Of course, but it's probably less unfair and arbitrary than any alternatives would be in practice, and it also doesn't suffer from the pretense of trying to appear rational when it isn't.  Yes, that latter point counts in policymaking (and is electorally popular).  Much of the public is smart enough to know when someone is trying to polish a taxation turd by making it sound intellectually solid when it isn't.  What they'd rather hear is the truth: there is no perfect system, and the one we have is less of a problem than any of the other ones we could give you.