Sunday, December 20, 2009

Open Urban Planning Degree 2: Carpool Fail

In an earlier post I explained my efforts to make my urban planning papers accessible to a knowledgeable, interested, general audience by imagining that I'm writing them for this blog. An added benefit of this exercise is that by posting my papers here, I make available some planning related ideas. The questions, "What is urban planning?" and 'What constitutes (or should constitute) an urban planning education?" are highly contested, and I certainly cannot, at this point, answer them. But a sort of practical answer may emerge as I pass on these papers. And I like to think that this could be of some use to someone considering going to planning skool, or to folks who do planning-related work, like community organizing, local politicking, etc., (see - I write etc. because the bounds around urban planning are perpetually implacable). Like MIT's OpenCourseWare does, I'll use the internet to make knowledge available outside the ivory tower, except unlike MITs OCW, I'll do it in a tiny, unsanctioned - let's just say it, meek way.

I wrote this paper in response to an "economic naturalist" prompt. It says, go out into the world, observe some urban phenomenon, and analyze it with economics. This assignment evinces the imperative in urban planning to address what is practical. Economics for urban planning only has worth inasmuch as it can explain some real behavior.

A final sidenote: I've read a lot of Chomsky and Naomi Klein, and had always thought of economics as an ideologically broken discipline, focused on currency and unquestioning of capitalism. This class sort of converted me. I want to use concepts such as opportunity cost, marginal cost, and externalities to create organizational systems (maybe markets - not necessarily "free" markets) that place value on the environment and distribute benefits equitably. I no longer believe that a microeconomic analysis of choice is incompatible with my ideological leanings toward anarchy and social democracy. Perhaps that's all for another post. Now, here's the paper.

Why are there so many empty seats moving (slowly) down the freeway? Wouldn’t commuting be faster and cheaper if we filled up all the vehicles?

Most cars seat 4 or 5 people, but most traffic consists of solo drivers. That’s a lot of underused capacity moving down the road. During the morning and evening rush hours, solo drivers are stuck in a sea of empty seats, yet nobody attempts to take advantage of all that potential mobility.[1] Drivers don't sell their empty seats to passengers who want to go the same places, and we don't see would-be passengers lining up to bid for rides.

The absence of such a market is most puzzling when you consider that the road is very congested, which means a lot of people want to travel along the same route. An economist would say that demand for travel on this freeway is high.[2] Whenever there's some demand for a good, anybody who can supply that good has an opportunity to charge a price for it. Then we have a seeming economic paradox. A service (mobility) is in high demand, and the ability to provide this service (empty passenger seats) is under private ownership and available for profit. This same scenario exists on the LA Metro, and there is a market: used fare tickets, still good for a few more hours of travel, are sold at informal (and illegal) markets around the stations. Why isn't there a market in casual carpooling?

In order to answer this question, let's imagine how such a market would have to work. Such a thought experiment addresses some logistical questions, like how would carpoolers match up with drivers going to their destination? And what about safety concerns? Let's just imagine the most low-tech solution possible. Let's say I am one of the many Angelenos who does not travel to work by car.[3] If I want to get to UCLA from the Koreatown neighborhood of Los Angeles, I walk to the nearest on-ramp to the 10 W freeway. I stand there with a sign saying “405 N – Wilshire” which represents where I want to get off the freeway. In a robust market, I'd find another passenger hoping to go my way so that we could get in a car together, because traveling in twos alleviates safety concerns for the driver. With all the rush-hour activity at every on-ramp, eventually someone going my way would roll up. The two of us would get in and we would pay the going price. The driver would drop us off shortly after the freeway exit, and we would be on our own to reach our final destinations.

It would seem that everybody gains in such an arrangement. That's the talent of private markets: they match up buyers and sellers who are both better off when the transaction is made. In this case, my fellow passenger and I are better off because we can get to work without owning a car. We save a lot of money, even if we spend some of it on the commute. The driver, who has already sunk a lot of money into the cost of owning and maintaining the car, gets to recoup some of this while getting to her final destination in nearly the same amount of time. So let's return to the question, why doesn't such a market form?

First of all, the market depends upon a level of robustness, or else passengers look like crazy commuter hitchhikers. The whole thing has to be established at many locations and well-known. This means there's a significant start-up cost. Whoever invested the time and energy wouldn't receive any special benefit, and everybody else who participated in the market afterward would be a free rider on the founder's original effort.

More importantly, the marginal cost per mile of driving is very low. Economists distinguish between fixed costs, which must be paid upfront to engage in some activity, and marginal costs, which vary with the level of use. In the case of driving, there are huge fixed costs: the cost of ownership and maintenance. Even the cost of a tank of gas is a fixed cost when you're behind the wheel. You already paid for the tank of gas, so the next mile is free. The distinction between fixed and marginal cost doesn't necessarily defray the driver's general incentive to recoup his costs. But a basic precept of bargaining (and markets) is that everybody tries to get the best deal possible.[4] In this case, consumers know that the marginal cost to producers is zero, and they will try to bargain them down that low. Why should I pay you to give me a ride when I know that you're going there anyway, and you already bought the car and filled the tank? I should get this ride for free.

Relatedly, even if the market were already established (perhaps by some government wanting the road to function better) and even if the marginal cost per mile of driving were higher, driving alone may actually confer a significant benefit on some people. Perhaps they really enjoy rocking out to music in the morning.[5] Alternatively, drivers may perceive the risk of inviting strangers into their car to be very high. Either way, for these people picking up a paying hitchhiker would entail a large transaction cost. Whenever transaction costs are higher than the going price, a market will not form.

In the places where we do see casual carpooling markets, some of the above criteria do not hold. In the Bay Area, casual carpools form to cross the Bay Bridge. In this case, the market's logistics are simple: there is one pick-up spot, and one drop-off spot. This solves the free rider problem that comes with establishing the market. In addition, the marginal cost of driving is much higher than normal. Single occupancy vehicles pay a toll of $4 to cross the bridge; carpools of three or more can bypass it. Now that there is a monetary benefit, more drivers are willing to take on the transaction cost of picking up strangers.

If casual carpooling were widespread, freeways would carry more people, and overall we might spend less money on commuting and less time in traffic. Unfortunately, because no one wants to do the work to coordinate all these carpools, and because driving another trip to work is basically free, all those empty seats remain an untapped resource.

[1] Except the government, which does provide a meager incentive in the form of a less congested carpool lane.

[2] Technically, they would say that the quantity demanded is very high because the price is very low. The term demand on its own refers to a dynamic function of price.

[3] About 20% of Los Angeles’ workforce, according to this document:

[4] In economic jargon, consumers try to maximize their consumer utility.

[5] Surveys suggest that this benefit is nontrivial. When people were asked to name their ideal commute time, the mean response was 16 minutes, not “no commute.” This is chronicled in the 2008 book Traffic by Tom Vanderbilt, p. 139.


ramonchu said...

I don't know bout the fixed cost of driving argument; I think when presented with a commuter hitchhiker, the situation takes on more of an opportunity cost scenario. Drivers now calculate what its worth to take a passenger, considering also what the passenger is willing to pay, versus as you said blasting the radio. I think the main reason these things don't develop is that everyone already owns a car because our built environment is so degraded as to make the totally walkable trips to 40% of our end-points deadly, dirty, and discouraging (hitchking, though not necessarily for commutes, was common when this wasn't the case even 20 years ago), leading to an enormous social factor that keeps people with cars from relating to those visibly without (especially those wanting to share/explore that disconnect). People already link up through craiglist both for travel and commute trips, though how effectively we cannot say. Your example is truly the solution, charge commuters 4 dollars to enter downtown LA and watch an entrepreneurial miracle in the number of people coordinating and using rideshare operations, as well as abstaining all together from driving to riding bicycles and transit. I'd like to read academic articles on hitch-hiking, craigslist rideshare (and craigslist in general)and the failures of carpooling (tom vanderbilt's book talks about why HOV actually increases congestion...can't remember why, latent demand I think, that wildest of wild cards)

The Flexible Carpooling Guy said...

Hello Herbie
Timely and interesting post. I keep a Google Alert for any mention of casual carpooling, so yours came up.

Our business is flexible carpooling. We classify casual carpooling as an informal flexible carpooling example, similarly the 'slug-lines' in Northern Virginia. You can look at our website for details on how we hope to formalize the system to reap the gains for other communities.

In early January we are holding a workshop in Washington DC called 'reinventing carpooling'. For the good of energy security, greenhouse gas emissions, costs of commuting, and the economy in general, we need a quantum increase in the number of people who share rides. Otherwise traffic congestion is just going to degrade. My pitch is for strategies to get 50 million new carpoolers by 2020, a four-fold increase over 2010. A big challenge.

I recently released an article calling for a new space program. If you are interested it is at

MU said...

A thoughtful and well done piece about a potential market with real implications.

But I more wanted to commend you for seeing the benefits of using economics concepts and tools for evaluating a far wider range of topics. The more mathematical end of economics suffers from the need to reduce most things down to a "price" in order to make functional formulas. But the degree to which the broader concepts really help explain and predict behavior is quite remarkable and useful. Like many tools, it is ideologically neutral, capable of good and harm. As you use these tools more, I predict you'll be surprised how useful and universal they are and how often they can explain why certain policies work while others fail, even when the cause is not intuitively clear.