Post #411: Toward a cost per car-trip avoided and cost per traffic-minute avoided.

Posted on October 7, 2019

This post finishes Post #408, with a plea for taking the most systematic approach possible to implementing measures to reduce traffic congestion on Maple.  Systematic, from the standpoint of cost-benefit analysis.


In Post #408, I called for the Town to add a traffic “hold-harmless” provision to MAC zoning.  More-or-less, I asked that for every car that MAC development will add to Maple Avenue traffic, the Town should institute some new policy to take one off Maple.  That way, the Town’s decision to convert Maple to a housing district via MAC would (plausibly) make traffic no worse than it would be otherwise.

I can see where some might characterize that as an extreme position.  And it is.  But I would say that it’s just about exactly as extreme as not worrying at all about about the cumulative traffic impact of Maple redevelopment.  (Which was, more or less, the Town’s position until recently.)  Those two postures are mirror images of each other.  One position demands no change in traffic, the other accepts any change in traffic.

To be clear, I’m asking for new efforts to reduce traffic that would not have been undertaken absent MAC.  In other words, I’d like to match MAC development to a new, offsetting, traffic reduction initiative.

Is it even remotely plausible to suggest this?  Could the Town find enough ways to keep drivers off Maple (or otherwise ease traffic congestion) that they could fully offset the traffic impact of MAC redevelopment?  Or is it a fantasy to think that the Town could achieve that?  Or is it simply a matter of costing too much?

This brings me back to the fundamental problem with the Town’s “multimodal” transportation study (Post #359):   No cost-benefit analysis.  The Town’s contractor, Kimley-Horn, produced a list of projects the Town could do on and around Maple, along with approximate costs.  But provided no estimates of what the likely benefits would be — no estimate of car trips avoided, traffic waiting times reduced, or traffic-related injuries avoided.

If that report had some estimate of benefits, we would already have at least a guess as to whether a “traffic hold-harmless” for MAC was plausible.  Or whether that’s just crazy talk.  And we would know whether or not that report actually gave us some clear path forward for addressing additional traffic congestion on Maple.  Or whether it was just an empty bill of goods.  Or whether such traffic reduction is flatly impossible, or whether it’s just a matter of cost.

Any predictions (estimates) of the impact on traffic are obviously going to be pretty loosey-goosey.  For any “alternative” transit mode (bus, bike, scooter, and so on), the biggest unknown is the demand for them.  For example, the consultants who projected use of Capital Bikeshare in Reston overstated demand five-fold (see this post for details.)  Even for more concrete changes, such as improvements to an intersection, the estimated improvement relies on mathematical models whose “fit” to the actual Maple Avenue experience is not established.

But that’s not a reason to avoid doing those estimates of impact.  It’s certainly a reason to be aware of the uncertainty of the estimates.  But if you don’t have some estimate of benefit, then you are picking your traffic-related projects more-or-less at random.  So while the estimated impact on traffic, for any one project, may be poor, I’m pretty sure that, on average, using some estimate to judge cost-effectiveness is preferable to picking projects by throwing darts at a dartboard.  Which is more-or-less what you are doing if you have no estimate of cost-effectiveness.

Why am I harping on this point?  Money matters.

I’d say that the experience of suburban governments in the DC area is that getting people to use anything other than cars for local transport tends to be hugely expensive.  The Town could easily spend all the money it has allocated to “multimodal” transport changes and end up with more-or-less zero net impact.

I’ve already given the example of the (estimated) average cost of $25 per bike ride for the Tyson’s Capital Bikeshare system (this post).  And don’t laugh, because there’s a good chance that Vienna’s forthcoming Capital Bikeshare racks are headed in the same direction.

To illustrate this, let me briefly review the economics of the existing bus service in Vienna, the Fairfax Connector (Post #225).  And then the City of Fairfax City University Energysaver (CUE) buses.  The Federal Transit Administration provides a concise summary of theses systems, available at this URL.

From the 2017 summary of the Fairfax Connector, we learn that:

  • The system costs about $2 per passenger mile.
  • Bus fares cover 13% of operating costs and none of the capital costs.
  • Buses typically have five passengers (passenger miles/revenue miles = ~5).
  • Ridership has been declining, per bus-mile traveled.

From the 2017 summary of the City of Fairfax CUE system, we learn that:

  • The system costs about $1.50 per passenger mile.
  • Bus fares cover 36% of operating costs and none of the capital costs.
  • Buses typically have five passengers (passenger miles/revenue miles = ~5).
  • Ridership has been declining, per bus-mile traveled.

For comparison, the American Automobile Association pegs the cost of driving the average private vehicle at around $0.60 per mile.  And Metrobus (WMATA bus service), based on that same source, costs $1.70 per passenger mile; fares cover 20% of operating costs; buses typical run with 9 passengers; and ridership has fallen slightly in recent years.

In effect, to put one person on a bus, Fairfax County ends up paying three times the cost of driving a car, in tax subsidy.   Surely you can see that if a shallow-pocketed entity like the Town of Vienna faces the same economics, our tax dollars probably are not going to be able to provide much (e.g.) bus service.

This is not some pro-car plea.  Fact is, car travel is heavily subsidized out of general tax revenues.  In the Commonwealth of Virginia, as of 2014, all road-user fees (such as gas taxes and tolls) accounted for just 41% of total road spending (calculated from this reference).  So, in effect, the “fare” paid by car drivers (for the use of the road) is just 41% of the cost of building and maintaining the physical road network in Virginia.*  And, just like public transit, the rest of the cost mostly comes out of general tax revenues and other non-road-related tax sources.

* (But it’s not a huge deal.  That subsidy works out to about $0.06 per per vehicle mile traveled in Virginia, or about $900/year for a vehicle traveling 15,000 miles per year.)

This is really just a caution that any of the proposed improvements could end up costing a lot of money per car trip avoided or passenger-mile traveled.

That cost is so high, in fact, that it leads to some truly odd proposals for controlling traffic congestion.  E.g., at some price, it can become smarter simply to bribe people to stay off the city streets, rather than provide transportation alternatives, as has been considered in AspenWashington DC, and a handful of other places.  (That Washington DC example is not quite as crude as a bribe,  but that’s the gist of it.)  The bribe is, in effect, a negative toll.  Instead of paying to use the road, you get paid not to use the road.

And, of course, economists are huge fans of “congestion pricing”.  (E.g., like the outlandish tolls on I-66 inside the beltway, when the roadway is crowded.)  In fact, I’d say that economists are more-or-less the only fans of congestion pricing.  Thus proving that just because it’s efficient (but inequitable) doesn’t mean that people will want to do it.

Finishing up, for now.

Here’s my ask.  I would like the Town to put all the proposed “multimodal” changes into a single spreadsheet that does some sort of crude cost-effectiveness calculation.  It could be cost per car trip avoided, or cost per traffic-waiting-minute avoided, or whatever appears reasonable and feasible.  And show the assumptions behind the calculation.  Not just seat-of-the-pants, not just doing projects based on instinct, but a reasonable try at a black-and-white calculation.

And then, assess whether or not you really could offset the projected additional traffic from Maple avenue redevelopment.

And, finally, sort the list from most to least cost-effective, and spend money accordingly.

My firm belief is that the mere act of having to make that educated guess about the impact of these alternatives will do a lot to straighten out the Town’s thinking in this area.  At the very least, as they go about funding these changes, they will be forced to acknowledge what part of the decision is based on firm information, and what part is based on guesswork.