Virtual Traffic Lights Could Revolutionize City Driving – [Or Not] |TechNewsWorld

I was briefly quoted in TechNewsWorld in an article about Virtual Traffic Lights, a long-discussed possibility with “Connected Vehicles”.

“Optimizing traffic lights is a great idea,” said David Levinson, a transportation analyst and professor at the University of Minnesota.

Getting them deployed any time soon, however, “is impossible, since it won’t be useful until all cars have such devices,” he told TechNewsWorld. “You can’t eliminate regular traffic lights until every car has VTL.”

Further, all the infrastructure — including hundreds of thousands of traffic lights managed by thousands of jurisdictions — would have to be upgraded, he pointed out.

“Even if all new cars have this, it will be decades before every car has this; old cars will need to be banned, retrofitted or somehow accommodated in a way that diminishes the utility of VTL,” he explained.

“Autonomous vehicles have a much more viable deployment path, and rely on the ability of the vehicle to sense conventional lights,” Levinson noted.

I really wish technology promoters would think about the deployment path instead of the end state. And discussing “carbon emissions” assumes this is deployed while gasoline is still used for powering cars, in other words, in 30 years when this may be feasible, we will still be burning fuel instead of using batteries or something else, which may be true, but would be sad.

Prices before Pavements

Alon Levy writes “The most important principle for infrastructure planning in developed countries is organization before electronics before concrete. ”

A shorter version of this is “prices before pavements”.

We cannot know how much pavement we need until we charge the appropriate price for the pavement we have. Since drivers don’t pay the full cost the externalities they generate, foremost congestion, they over-consume road transportation. Since we use unpriced usage of facilities as a guide about where to build new roads and widen existing ones, we probably generally over-build compared to the priced equilibrium (even if we still have congestion), though it is hard to say, since we don’t have a priced comparison point, and model forecasts are hard.

This phenomenon of over-construction is not surprising. It follows from micro-economic theory and we have evidence of over-consumption every day in every major American city. We have evidence of under-pricing. We have evidence that when prices rise, consumption falls (e.g. gas prices, London, Stockholm).

So where is the failure in this collective action problem?

Historically road pricing was expensive to implement, particularly compared to fuel taxes, and toll collection would create its own congestion. Today’s electronic toll collection obviates the congestion creating aspects of tollbooths. However collection costs are still higher (and in practice, nothing could be easier than a gas tax, where you collect funds at the refinery). But even with those additional collection costs, the benefits of congestion reduction in places that suffer congestion should be enough to justify a more serious effort at pricing, leave aside the general benefits from less pollution and so on.

Today, no one is incentivized to raise prices on existing facilities. Roads are governed by elected officials, who believe they are re-elected when they keep taxes down and are sometimes punished when they raise taxes. Voter/taxpayers do not trust that increased revenue will be wisely used (and they are not necessarily wrong), and believe that if government just redistributed the revenue, that still imposes a cost (which it does, but again, not so much as to negate the benefits). This is true with gas taxes (which finds that tax escalators get rolled back). It is certainly more true with a new tax that people have no experience with.

States have little incentive to internalize externalities that are imposed globally, or on residents of other states, and little incentive to raise prices much higher than the prices in surrounding states, which may be engaged in tax competition. (They do have an incentive to raise tolls on their border though – as they can engage in tax exporting.)

Moveover, there is prospect theory, which suggests people feel losses more deeply than gains. A tax or toll or user fee is a known loss of revenue for an uncertain gain in travel time savings (which we don’t really value well either). [See also this different take on the implications of prospect theory on road pricing in Singapore by Anthony Chin.]

This suggests there is a governance problem. Even if a state raises some gas taxes to pay for transportation maintenance and construction, that would not be sufficient to internalize the external costs of travel. That would be politically impossible in the current regime in the US.

Which leads to the thought, if a good idea like road pricing is politically impossible under the current road ownership regime, perhaps changing the regime changes the political calculus: I discussed this in Enterprising Roads. Roads should be more like other public utilities, which provide a service and levy a charge to pay for the costs of that service, are are regulated not directly by the legislature, but instead by a body like the public utility commission.

I don’t know which is less politically impossible.

I think evidence supports the notion that expecting widespread road pricing at anything near the full cost of travel under the current road ownership/governance regime in the US is practically impossible anytime soon.

We may get voluntary odometer charges of some kind instead of the gas tax (Oregon), particularly for electric vehicles. We may get a city like New York or San Francisco try an urban congestion charge. We will certainly get more HOT lanes and more tolling on new roads (e.g. Southern California). But these are all nibbling around the edges.

The great white whale is full cost pricing, where prices vary by time of day (very important) and location, as well as amount of emissions (very important) and crash probabilities.

Perfect pricing will never be achieved, how close to perfect can we actually achieve?

Or will some new technologies just make obsolete all our concerns about externalities, just as we never got around to full cost pricing on horses?