Monday, August 31, 2009

The man behind the Simpson Curtain

Last week I posted about arguments for free public transit, some of which refer to the "Simpson-Curtain rule." "Politically and practically, for most systems, the easiest way is to raise fares. But soon after, ridership goes down," writes Dave Olsen.

As I wrote last week, things are actually quite different, and a bit more complicated. The rule is actually the Simpson-Curtin, rule, after the Philadelphia planning firm of Simpson and Curtin that came up with it. A 1968 paper by principal John F. Curtin is often cited as providing the basis for the Simpson-Curtin rule, but it is actually concerned with extensions to the rule for various additional factors.

From the 1940s through the 1960s, Simpson and Curtin collected data on transit fares and ridership. After crunching the numbers, they found that when fares were increased, ridership went down. The ratio of the change in ridership to the change in fare is called shrinkage ratio or fare elasticity, and is explained in more detail in this PDF from the VTPI. But briefly, Simpson and Curtin claimed that for every 3% increase in fares, ridership dropped 1%, for a shrinkage ratio of -0.33. The lower the absolute value of the ratio, the less an agency would have to worry about driving riders away with high prices.

So what's wrong with the Simpson-Curtin rule? Well, it represents correlation, but correlation does not imply causation. As I wrote in the previous post, to get causation, we have to come up with an explanation that fits the facts better than any other one. The explanation that Curtin, Olsen and others have made, that raising fares by itself causes reduced ridership, is not the best explanation.

Just a few years after Curtin's paper, in 1973, Michael Kemp published a paper arguing that "transit demand is inelastic with respect to money price." He indicates a number of other factors that affect demand, including level of service ("particularly door-to-door journey time"), trip purpose, distance, transit mode, urban form, and length of measurement period. In particular, since transit riding is a choice, it is dependent on the relative attractiveness of the alternatives. He notes:
One might hypothesize that, given the initial decision to travel, transit riding will be higher when the relative prices of substitute modes are at their highest; and that under such conditions transit fare elasticities will be relatively low. It follows that one would expect ... fare elasticities to be relatively low in very large cities with highly congested central areas, particularly for those modes catering to long-haul commuter traffic.

In other words, people are least likely to abandon the bus when it's hardest to drive ("relative price," here, includes travel time and convenience). The correlation observed by Simpson and Curtin is simply due to a lurking variable that drove down ridership and consequently pushed agencies to raise prices: the massive road-building that went on in the postwar period. Of course people aren't going to stick with their old bus system when the government is building new roads and parking lots for their cars! But if the government doesn't build as many new roads and parking lots, then transit still has a chance.

That brings us back to New Jersey, and what has allowed the bus companies to remain profitable for all these years. Not only do the private Lincoln Tunnel buses have the XBL, but they also have the Port Authority Bus Terminal and a law protecting them from destructive competition from the government.

There are two factors that are even more important. The buses (a) go to Manhattan (b) under the Hudson River. Manhattan is a notoriously unpleasant place to drive a car; despite the best efforts of the New York City DOT up until a year and a half ago, it is and was still incredibly congested. It can take half an hour to go a block, "free" parking is almost always full, and private parking is very expensive.

Still, as you pointed out on Friday, the private bus companies in Brooklyn, Queens failed many years ago, even though they continued on as zombies until recently. What's the difference? I think it's very simple: cordon pricing. You can drive across the East River without paying a toll, but you can't drive across the Hudson for free.

Recently I ran into my boss on the subway here in Queens. He told me that he lives in New Jersey and owns a car, but he takes the bus to the Port Authority and then the subway and another bus to work. He's tried driving in the past, but the tolls were too expensive. Sure that's anecdotal, but it fits with the pattern that Kemp found. If it's too expensive to drive, people will take transit.

During the congestion pricing debate, Aaron Naparstek wrote a post called, It’s the Bus Riders, Stupid. Aaron is actually referring to something different from what I'm talking about (and honestly, I always hated the Clinton quote he's referencing). However, it was - and is - the bus riders: cordon pricing would have eliminated the toll-free option for crossing the East River, and thereby increased ridership on the MTA buses, bringing up revenue. With cordon pricing in place, that revenue would be more stable than the other MTA funding sources, and maybe enough to start paying down the debt.

So there's your magic formula for transit profitability:

1. Give transit its own right-of-way and good terminals
2. Make it hard to use cars
3. Make it expensive to use cars
4. Profit!


fpteditors said...

Fares exist to inhibit use. There is no free market involved. The market is controlled by the big players. See the movie:
Taken for a Ride

Cap'n Transit said...

I have seen the movie, thanks, and I highly recommend it. I don't think it proves what you say here. But I'll get into that more in another post.

Jarrett at said...

No question that you could make transit profitable by that method. Are you sure you WANT transit to be profitable?

Isn't its loss-making nature part of what keeps it under government control, and thus accountable to the city? Sure, a profitable carrier is good to its customers, but it's not always a very good partner for government's efforts in infrastructure development and planning.

Alon Levy said...

Jarrett, I don't think the profitable transit companies in Japan and Hong Kong have a pernicious effect on planning and development. On the contrary, they promote the development they can serve the best, which is dense and transit-oriented.

Jarrett at said...

Yes but, for example: I want people to own fewer cars, for reasons that go beyond how that helps the profitability of a transit operator. A profit-oriented operator will tend to skimp on late-evening service, for example, but you need really good late-evening service, even if unprofitable, to make lower car ownership viable on a large scale. It would take a sophisticated operator to understand this as a good outcome purely from a profit motive, especially if, like most businesses, there's pressure to deliver short term. Not saying it's impossible, but I wouldn't say that level of sophistication is common in the private operations market today.

Cap'n Transit said...

That's interesting in theory, Jarrett, but most government operations don't understand that either. For example, Dave Olsen happens to mention that the wonderful Island Transit has no Sunday service. On weekdays, its last run is 7:50 PM.

Meanwhile, the privately operated DeCamp buses in New Jersey run buses until midnight, including hourly service on Saturday and Sunday.

The widespread lack of Sunday bus service around the country is just baffling. How could you expect people to live without a car if there's no bus on Sundays? The surprising thing is how many people live without cars in these areas anyway.

Alon Levy said...

Jarrett: on the contrary, the profit-motivated transit operators in East Asia have more weekend and evening service, not less. Usually they just bundle all of it into off-peak. For example, read this chart of train frequency in Hong Kong, showing that the busier lines run every 4-5 minutes in the evenings and the less busy lines every 10 minutes.

Tokyo is similar, but has no combined timetable that I know of. You can search individual lines on Hyperdia, though - the Yamanote and Chuo Lines run every 4 minutes between 11 pm and midnight, and the Marunouchi Line runs every 5 minutes.

Christopher Parker said...

In fact, going after the off-peak market is a key part of any profitability a transit service might hope for. One of the large expenses is paying for the vehicles, so you want them running as much as possible (see Southwest airlines). In addition split shifts to accommodate rush hour are more expensive and less productive than regular hours. Examples: the Chicago & North Western Railroad in the sixties managed to achieve profitability on their commuter trains in part by increasing off-peak service to hourly. Metra and Metro-North and services in London have all found that going after the off-peak market helps the bottom line.

Jonathan said...

I was thinking about the ex-private bus lines in Queens this morning. They don't run to Manhattan, so how would cordon pricing help them?

Cap'n Transit said...

Yes, Jonathan, many of them do run into Manhattan. Two of those, the Q60 and the Q102, are local, but the rest are express.

George K said...

Sorry to post so late, but I just wanted to mention that he is probably talking about the local buses run by MTA Bus.

Personally, I can't see why congestion pricing, or at least East River tolling hasn't been implemented. It is already hard to drive in Manhattan traffic and parking is extremely expensive (even if you are just driving through, that traffic causes you to burn up excess gas).

Some kind of cordon pricing would actually help the people who feel that they must drive to Manhattan, by reducing traffic congestion.