Monday, July 13, 2009

The case for public transportation

Daniel comments on my post about the "subsidy fallacy":
Well, in subsidizing transportation's defense, it's very difficult to provide good, alternative public transportation (I'm assuming you're referring to public transit) only through the direct revenues of riders. Nevertheless public transit is a very important system to have.
This is definitely true, and if my poor readers will forgive me for droning in lecture mode for a minute, I think it's valuable to discuss just why this is.

Certain kinds of public transportation, like subways, have very high fixed costs but marginal costs that are close to zero. In such circumstances, the efficient economic prescription is to charge near-zero fares. This unfortunately means a loss of market discipline, and it's complicated a little by the fact that using taxes to fund transportation has its own negative effects, but from the perspective of sheer economic efficiency, using tax dollars to set fares closer to marginal cost is overwhelmingly positive.

Moreover, as Daniel points out, it can be difficult or impossible to support public transit on fares from ridership alone. This does not mean that public transit is an economic loss. In fact, it's easy to explain how a transit system might be unable to support itself through fares even if it's a large benefit overall. Consider the New York City subway, which I've used every day for the last two months. The alternatives for commuting to work are slim: taxis are far too expensive, cycling is slow and miserable in bad weather, and the costs of owning and parking a car in Manhattan are prohibitive. All in all, I'd be willing to pay at least $10 for each leg of my daily commute, and there are many more like me. But if MTA jacked up its fares to $10, the vast base of other customers who are willing to pay $3 or $4 would dry up, even though they cost almost nothing to transport. It's completely conceivable that there is no single price where MTA could raise enough money to fund its operations, even though there would be an enormous surplus if it could charge each individual customer the true value of the service.

Now, as Megan McArdle points out in a different context, it's important to recognize when we're just rehashing generic arguments in favor of socialism. After all, there are a lot of superficially good arguments for socialism, but in most cases they're far outweighed in practice by its negatives. For transportation within cities, however, I think there's often a perfect storm of high fixed costs, near-zero marginal costs, and the practical need to set a single price (i.e. an inability to price discriminate) that compels public investment.

This all ties back to my earlier post about the subsidy fallacy, the commonplace argument that because we have such large subsidies for roads and other forms of traditional transportation, almost any change in spending should be welcome. This frequently degenerates into hand-wavy rhetoric that justifies new projects in the absence of any cost-benefit analysis whatsoever. As I hope this post illustrates, there are strong economic arguments available to justify public subsidies for transit, which makes it all the more important that we avoid the weak ones.


Mark said...

Your post makes and relies on some assertions that are not, as far as I can tell, supported by evidence or much analysis.

1) "High fixed costs [combined with] marginal costs ... close to zero ... means a loss of market discipline." The examples of FedEx, UPS, and (less clearly, because of extensive government regulation, subsidies, and limitation of competition) airlines would seem to prove otherwise.
2) "It can be difficult or impossible to support public transit on fares from ridership alone." While there is ample evidence that government-owned public transit requires subsidies, that is not necessarily representative of public transit as a whole. The structure of government-owned public transit is, of course, subject to political pressure rather than to other market forces. (Since I am only arguing that you have not really made your case, I not say more on this, except that has a wealth of sourced, relevant information on this topic.)
3) You say that price discrimination in public transit is impracticable. (I believe this is an accurate expression of one of your assertions.) Unlikely. Again, the business practices of airlines, FedEx, UPS, (and even USPS) suggest that price discrimination is practicable in public transit. A little imagination suggests other possible implementations, e.g., full-fare customers are guaranteed a spot on the first bus, half-fare customers may have to wait up to an hour, quarter-fare customers may have to wait two hours. Or perhaps full-fare customers get better seats -- some airline customers pay a huge premium for not much more than that. Any serious problem with implementing price discrimination is going to be political.

As far as I can see, your argument establishes, at most, that a subsidy for public transit is worth considering. You certainly have not provided any "strong economic arguments ... to justify public subsidies for transit." That would require -- at the very least! -- some actual data.

Matt Rognlie said...

Mark, I absolutely agree that we need actual data (or at least some sense of comparative magnitude of costs and benefits), and using this rhetoric alone to justify public transportation would be a bad idea. The idea here isn't to make the argument-to-end-all-arguments for public transportation, but just to point out some economic fundamentals that often get lost in the discussion.

I'm afraid my lack of clarity is responsible for your first quote. I didn't mean that high fixed costs combined with low marginal costs mean a lack of market discipline -- I meant that the policy response of subsidizing fares entails a decline in market discipline (and I think you'd agree with that!)

It's possible that some clever administrators could establish price discrimination in mass transit, and your suggestions about buses are interesting, but I'm still skeptical of the possibility of meaningful price discrimination in many settings. I can't imagine, for instance, how subways could price discriminate without fundamentally changing the nature of their service.

Intercity rail transportation, on the other hand, may be able to price discriminate at least as well as aviation, and accordingly this post isn't such a good argument in favor of subsidizing long-distance rail. In fact, I'm something of a "train skeptic," and although I didn't get into it in this post, part of my ultimate goal here is to show that the most compelling arguments in favor of public transportation don't extend very well to that setting.

Mark said...


Thanks for your response. I certainly agree that the "policy response of subsidizing fares entails a decline in market discipline."

Your skepticism about price discrimination for subways is understandable. My gut-level reaction in this (and most) such things is "someone will figure out a way to profit by supplying that service." Of course, my gut is sometimes wrong, and in the case of subways we might never run the test.

happyjuggler0 said...

Regarding subway price discrimination, I'd say that surely if there was a will then there would be a way to engage in a "pre-boarding class" of fares, especially during rush hour (or "post-theater" where that may apply).

I grew up in the Boston area and during certain times, it wasn't much of an exaggeration to say that it "was first push, first serve" if you wanted to get on a train (the T), and sometimes you had to wait a considerable amount of time.

Show me a line and I'll almost always be able to show you inefficient pricing.

There is also the possibility of peak travel hour pricing, but that gets a lot trickier to implement when people can use prepaid tokens or passes. Not only would this raise more money during rush hour, but it would also likely rearrange marginal travelers travel timing, thus filling in far emptier cars during off peak hours, losing less money during those time periods.

On another note, I'm tempted to say that "stupid union contracts" would be less common if the system could somehow be privatized, but considering the cases of GM, Ford, Chrysler, and the legacy airlines, I am far from sure about that.

Anyway, I'd love to see a government at least try to privatize their subway system, even if it means selling it for a dollar with strings attached.

happyjuggler0 said...

Oh yeah, I almost forgot, do you think you would still be unable to afford taxi fare in a system without a medallion ceiling?

How about if NY had a constitution that made rent control illegal? This would improve both rent (civilian definition!) and likelihood that you could find a place to live near work.

John Farragut said...

I think with a bit of cleverness price-discrimination is definitely possible for the subway system. The easiest is mentioned by happyjuggler: charge different prices at different times of day. I know the way the MetroCard system is this is somewhat unfeasible (e.g. with the unlimited passes), but it could definitely be implemented with relatively minimal effort. In fact, I'm somewhat shocked that they don't do this; I live in Ithaca (during the academic year) and you pay different cab fares on different days/times of day for the exact same ride. Other possibilities (I have not thought much about these) would be to have a sort of time-locked card, not valid until an hour (or whatever) after purchase, which would let you separate the "people in a rush" market from the "people with time to spare" market. There are some obvious problems to this, but it's an idea. They could have different student/senior rates, though this would require some sort of verification mechanism not currently in place.

Anyway, I think price discrimination is definitely possible--the easiest being charging different prices at different times of day--and much preferable to the current arrangement, where taxpayers (does anyone know which ones specifically?) are billed for other people's subway use.

It seems the subway system is an excellent example of a "natural monopoly"; any common fixes for those could almost surely be implemented here.

grandmute said...

I hope this is not too pedantic, but most public transit systems already practice some form of price discrimination. For one, second-degree price discrimination is near-universal (discounting the per-use cost as more uses are prepaid for). Also, third-degree price discrimination is fairly common, as exhibited by discounts extended to children, the elderly, and other groups. Granted, it is obvious that these instances of price discrimination have not solved congestion problems in public transit.

Douglas Knight said...

While I agree that price discrimination is difficult for subways, "Paris Metro Pricing" is a standard example of price discrimination, namely where the purchased services are identical, but the expensive car is better simply because it is less crowded (and lacking poor people).