Passenger-Miles Are Overrated

One of the pushbacks I got about my post on road boondoggles is that I didn’t control for passenger-miles of travel, and the number for car subsidies is much lower when one divides it by the appropriate number of trillions. This is not the first time I hear people talk about passenger-miles as a measure of inherent worth, but it doesn’t make it any better.

Passenger-miles don’t vote. They’re not a unit of deservedness of subsidy. They’re one unit of transportation consumption. They’re like tons of staple as a unit of food production, or calories as a unit of consumption. We don’t subsidize food based on cents per calorie.

Even as a unit of consumption, there are flaws in passenger-miles as a concept, when it comes to intermodal comparisons. The reason: at equal de facto mobility, transit riders travel shorter distances than drivers. It’s very obvious when comparing total passenger-miles in transit cities and car cities (see e.g. page 36 here). Transit is slower than driving on uncongested roads, but has higher capacity than any road. In addition, transit is at its best at high frequency, which requires high intensity of uses, whereas cars are the opposite. The result is that transit cities are denser than car cities – in other words, need less passenger-miles.

What passenger-miles are more useful for is measuring intercity transportation. At intercity distance, mode choice has less influence over travel distance (though, even then, HSR and driving are shorter-range than flying, and thus passenger-miles can overstate the importance of flying over ground transportation). It is also a proxy for revenue, whereas on urban transit the fare is either flat or weakly dependent on distance. As a result, intercity railroads usually cite passenger-miles or passenger-km, and urban transit operators usually cite passengers.

But when it comes to local transportation, it doesn’t work very well. A country’s mode share expressed in passenger-miles is lower than that expressed in passengers, and this is going to make transit and especially walking look much less significant than they actually are.


  1. Beta Magellan

    As long as we’re on the subject of measures, one I can’t stand is ridership-per-capita—I’ve heard it used a couple of times to justify Illinois’s “let no micropolitan statistical area go untouched” approach to state-supported intercity rail—the fact that a higher proportion of people in some small town use rail for intercity trips compared to Chicago is no reason to give their needs priority.

  2. Andre Lot

    The problem in discarding passenger-miles is that, stretching the logic to its extremes, the best cities are villages in the middle of Papua New Guinea where people don’t use any vehicular transportation and barely venture out of their house of small farmh holds. I’m sure they get quite low TKU for cargo transportation as well as consumption there is woefully low and they live in extreme poverty…

    The limitation of this logic of transit users travelling shorter distances, even if not taken to those extremes, is that it assumes mobility works in a strictly functional-destination basis whereas the passenger-mile logic works on a strict area coverage basis. It’s pointless to argue the superiority of one over another.

    In any case: I don’t know numbers for US, but I do know that here in The Netherlands, the difference of passenger-mile and passenger-trip for commuting, as a % share, is not that big for cars. Considering all motorized domestic commuter traffic in the country here, is something like 73% on passenger-km and 68% on passenger-trip. In the specific case of Netherlands, a large base of students and other pass holders travelling very long distances on a daily basis (70km one-way or more) compensates for shorter trips of many others in relate to car passengers. Countries with high-speed rail like Spain and France see a small number of very wealthy commuters that 200km one-way several days a week, each trip easily accounting for the same passenger-km of 20-30 inner city dwellers. The only mode (Netherlands) that poses very different figures is biking: counting for passenger-km, it counts for something like 4%, but on passenger-trip is like 14% nation-wide.

    I’m sure if you counted walking, numbers would be even more skewed.

    But even in the context of transit itself calculations are polemic. Commuter lines tend to have higher passenger-miles count but lower total passenger counts as the “seat turnover” is rather low when everyone is departing/arriving from many place to one major central station and back in the afternoon. So some planners argue that, when prioritizing budget, a short (10km) subway line that will carry 140.000 passengers/daily is more important than a long commuter line (70km) that will carry 35.000 passengers/daily

    • Alon Levy

      Your example of commuter rail vs. urban transit is one of the related issues, of course. Although well-run urban transit operators cite passengers more often than passenger-km, there are a few in the US that justify further extensions into low-density suburbia on passenger-mile grounds. I think BART is a major offender there.

      Also, overrated != should be discarded. Passenger-km/miles are a good metric for intercity and maybe long-distance commuter transportation. The problem is that for commuter transportation, they break down. For example, in the US, transit’s share is about 1% on a passenger-mile basis and 4% on a commute mode share basis. In Switzerland, the corresponding former figure is 15%; for the latter, I’ve seen both different SBB annual reports claim 25 and 30%.

      As for the Papuan village example, those villages are poor by many metrics. Of course, passenger-km and ton-km are both correlates of wealth, but they themselves don’t indicate wealth, especially when comparing different developed countries. Let’s go back to my post’s food analogy. Calorie consumption per capita is a very strong correlate of GDP per capita; the same is true of red meat consumption. Of course, we know from basic nutrition science what the healthy level of food consumption is, so that we can pronounce American food consumption as too high, especially with regards to red meat. The state of urban studies is not as developed, so we don’t know precisely how much mobility is too much; but the qualitative concept of hypermobility does exist, and can be seen in today’s exurban America.

  3. D. P. Lubic

    Didn’t the freight railroads have a measure to equalize this out? I think it may have been something like ton-miles-per-hour, or ton-miles-per-train-hour; the idea was to get a consistent measure that would equalize the workings of a railroad with short, fast, light trains, like a Richmond, Fredericksburg & Potomac or a Nickel Plate Road, vs. a railroad with slow, heavy trains, like a coal-dominated Norfolk & Western.

    • Alon Levy

      Probably. JR East has a comparable measure for passengers: one of the metrics it cites is passenger-km per (I think) car-hour.

  4. Danny

    I agree with your comments, but I would like to point out something else that bothers me coming from the same people that you are talking. It is the idea that transit subsidies and road subsidies can be compared side by side, not just using passenger miles, but by using any metric at all. They are fundamentally different in their cost structure, and therefore their subsidies aren’t comparable.

    Take this imaginary scenario, which has been simplified for illustration purposes: A single road runs through a city, and a single rail system runs through exactly parallel to the road. Both are exactly breaking even operationally, with gasoline taxes at .80 cents per gallon and transit load factor being 60% and fares of $2.00 per ride. Mode share is exactly 50-50 split among the cities 10,000 residents.

    One day the government decides that transit is a virtue, and decides to subsidize the transit. They do this by dropping fares by $0.25. This adjusts mode share to a 45-55 split. Operationally for the transit agency, they are now losing money, but not as much as the 12.5% reduction in fare suggests…because now they have more riders and load factor has increased by a few percent. But for the road system, there is no impact at all. Less people use the roads, meaning less people pay road taxes, but road wear goes down proportionally.

    Now take the reverse scenario…one day the government decides that cars are a virtue and decides to subsidize roads. They do this by dropping gasoline taxes to $0.70 cents. Once again, we get a mode shift, now 55-45. Operationally, the road is now losing money, because more people are driving and they are paying less of their contribution to road damage. But transit is also losing money. Despite the fact that fares have not changed, the changed mode share has changed the load factor. Now, in order for the transit system to break even, they either have to raise fares or reduce service levels…both of which result in transit users paying an additional cost simply because roads are subsidized.

    Now in the US, we have subsidized roads. Since mode share has shifted dramatically, transit systems became less sustainable. We have responded by subsidizing transit. And for every additional cent of subsidy that we give to roads, we end up needing to give more to transit to maintain the same consistent price/value offering to transit users. And the result is this increasingly dangerous subsidy Jenga, where we correct subsidies with other subsidies.

    Speaking in econ terms, transit and cars are natural competitors. They are different in the sense that transit is an inferior good and cars are a normal good (no I’m not calling transit inferior, look up the economics definition if you don’t believe me). Furthermore, costs for car use runs approximately proportional to the number of users…but transit costs run inversely proportional to the number of users. The outcome is entirely predictable, and any microeconomist worth her salt would be able to tell you exactly what would happen when you subsidize one but not the other. Simply put, the outcomes are different. The subsidies can not be compared.

    • Andre Lot

      I agree with your classification of trains and cars as goods, however, there are further complications.

      Wear and tear on roads are minimal if only personal cars run over them. Just look at the few parkways or tunnels where trucks or even vans are not allowed. The structural damage to the pavement is related in a 3.5-4.0 power of the axle load (further compounded by the “stitch” effect of heavy loads crawling a high ascent through, or going on unbalanced curves at speeds lower then that optimal for the cant).

      If only cars used roads, that wouldn’t be so much of a problem except on very light maintenance items like lane paintings, cutting grass, replacing sings etc. Cars impose costs mostly in terms of capacity constraints (= road needs to be widened / new roads need to be build). Trucks cause great damage, to the proportion of one single container truck (loaded) causing incidental damage as large as 500 sedan cars. But, on the other hand, trucks rarely congest roads alone, they don’t have peak holiday demand pumping 4x the usual flow etc.

      Your calculations assumed all-fixed costs for trains and all-variable costs for roads, which is why the paradox would work as you proposed. Fair enough, there is more linearity on, ceteris paribus, road flow x road costs than transit ridership x transit costs on a narrow window, by many reasons starting with the fact road transportation (of passengers ) is far less professional labor-intense and (passenger and cargo) the scale of vehicles means usually they have not much spare capacity (again, simplifying calculations assuming a car is meant to be ridden by people of one household only). However, roads have fixed costs and transit has variable costs.

      • EngineerScotty

        Of course, it ought to be mentioned that one of the nastiest vehicles for pavement damage is the good old two-axle 10m city bus. Most trucks with a GWV over 5 tons or so add additional axles to distribute the load, as do many motor coaches used in long-haul service, but the operational demands of urban public transport (low floors, easy to maneuver into stops and around corners in dense neighborhoods) produces the result that vehicles with a gross weight of 20 tons only have two axles on which to distribute that load–and busses stop far more freqeuntly than do trucks.

        • Danny

          This is why I think that costs for BRT corridors are consistently understated. Most BRT proposals assume 15-20 years between major resurfacing…but reality has shown to be more around the 8-11 year range. That assumption alone could tilt cost/benefit decisions in the wrong direction.

      • Alon Levy

        Serious, partially related question to the gallery: does anyone have figures for comparative maintenance costs of car-only roads (for example, New York City parkways) and all-vehicle roads?

        • Danny

          I don’t have an answer to your question, but I can imagine it would be hard to research. Many roads that are restricted to trucks happen to be restricted because they weren’t engineered to handle trucks….typically built on weaker substrates where no effort was put into reinforcing them. I know this is the case with I-580 in the bay area and almost all truck-restricted local roads. It would be pretty hard to find a scenario where the road surfaces could reasonably be compared.

          • Jeffrey Jakucyk

            There’s a few no-truck “parkways” right near where I live, and they’re still maintenance problems. You don’t see the problem of asphalt “shoving” at intersections that you get on roads with lots of trucks or buses, but the pavement still seems to deteriorate just as much as any other street. Problems like everyday cracks, reflective cracking of concrete, simple wearing away of the road surface, water intrusion, and loss of binders seem to be just as bad as on any other road. Of course, differences like the particular asphalt mix the plant used that day, or other conditions can have just as much effect on the longevity of the pavement too.

            On roads with a lot of heavier vehicles, yes you’ll see more of the aforementioned shoving problems, and probably more issues with base failures and alligator cracking. I think the reality is that asphalt pavement simply doesn’t last long at all under any circumstances. Concrete tends to perform much better, and I bet there you’ll see bigger differences in longevity relating to vehicle weight. There’s several city streets here in Cincinnati that were completely rebuilt, along with their streetcar tracks, in the mid 1920s with thick concrete. These roads are still in place, though with one or two thin lifts of asphalt on top that serve as a newer wear surface. So while the concrete surface may not be usable, they’re still 100% structurally sound even after almost 90 years. I have little doubt the lack of truck traffic plays a part in this.

        • D. P. Lubic

          I don’t have a maintenance figure, but I do recall that John Kneiling, who used to be the “Professional Iconoclast” in Trains, once stated that a road in New Jersey that was originally intended to be automobile-only was revised as a typical car and truck road, and that the resulting beefing-up that was necessary to handle trucks doubled the initial construction cost.

      • Danny

        It was a simplification to prove a point…I didn’t have time to provide a comprehensive bases-covered answer. You are right, reality is a bit more complicated, but the point still stands about approximate proportional vs inversely proportional cost structures. Since they are competitors, they will influence each other’s mode share, and while that doesn’t mean much for roads, it means a great deal for transit.

      • Danny

        BTW, Andre…I’m surprised you used such a conservative estimate for truck vs car road wear. I used to drive a truck for a member of the American Trucking Association, and their literature is far more biased than most independent research and yet much higher than your estimate.

        I remember reading an article where they tried to debunk the “misconception” that a single truck does 9300x the damage of a standard size car…according to them, reality was a far more reasonable 5900x…and if roads were designed “correctly”, it was more like 1000x.

        • Alon Levy

          I think you misunderstood Andre’s number. A 3.5th-4th power law is much stronger than a factor of 3.5-4; comparing an 18-wheeler with a small US car or a medium-sized European car, a 4th power law corresponds to a factor of about 10,000, and a 3.5th power law corresponds to a factor of 3,000.

          • Danny

            I didn’t do any calculations on the axle load. I was referring to this statement: “Trucks cause great damage, to the proportion of one single container truck (loaded) causing incidental damage as large as 500 sedan cars. “

  5. Pingback: » Should Public Transit Agencies Strive for Profitability?
  6. payton

    Todd Litman of VTPI calculates that each non-motorized trip displaces seven miles of driving. Measuring by trips isn’t perfect, either — obviously comparisons of inter-city trips should rely more on passenger-miles — but as far as local travel goes, a trip is a trip.

  7. Pingback: Vehicle Miles Travelled are Bad - Seattle Transit Blog
  8. Pingback: Is “per passenger km” the right metric for comparing modes? « The Melbourne Urbanist
  9. Pingback: Weekend link dump for October 2 – Off the Kuff
  10. Pingback: June, 2014 Newsletter - DFL SD 48 ~ Democrats of Southern Minnetonka and Eden Prairie

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.