Finance Social Justice Washington DC

As Congestion Mounts, Transit Agencies Consider Varying Pricing

» Washington Metro considers charging customers more to use system’s most congested stations, increasing peak-hour commute costs.

Downtowns play a primary role in organizing the daily lives of millions of Americans. Despite increasing suburban sprawl and the more recent comeback of inner-city housing, downtowns remain the single largest work centers of virtually every U.S. metropolitan area.

In the biggest of those center cities, rail rapid transit plays a vital supporting role; by hauling in tens of thousands of people to a limited number of downtown stations every morning, these systems allow the creation of dense urban cores that would not be possible were everyone to rely on private automobiles. Just as importantly, most urban rail systems would make little sense if they didn’t serve highly attractive destinations; it’s not a coincidence that almost every American urban rail line reaches the job center of its respective region. Downtowns and rapid transit are mutually reinforcing.

Why make this seemingly obvious point? Because private interests and the public sector have spent the last century working together to build these jobs centers, and the congestion now experienced daily on major rapid transit systems from New York to San Francisco is not unexpected: It was planned.

There’s nothing sinister about this fact: From a social equity perspective, there are good reasons to concentrate employment growth downtown and there are positive effects of economic accumulation that result from dense downtown cores. But that clustering — in addition to the standardized work hours enforced by most employers — ensures that rail lines and especially their downtown stations are packed in the morning and full in the evening, only to be frequently empty at midday. It’s not a particularly efficient distribution of ridership, but it’s what happens when thousands of people are working in close quarters downtown.

Facing a tough budget year and little hope of significantly more money from local, state, or federal sources, the managers of Washington, D.C.’s Metro system are considering adding a 50¢ surcharge for customers using or passing through the network’s busiest stations in the center city during peak hours. It’s an approach that has been promoted by a coalition of transit advocates and smart growth proponents who argue that some combination of additional fees would aid in the budget crisis, affect mostly wealthy riders, and reduce congestion by encouraging people to go to work during off-peak hours.

Similar schemes have been proposed over the years for a number of American transit agencies suffering from congestion at downtown stations.

Washington’s Metro already charges varying fares based on distance and hour traveled; a trip between suburban Bethesda and Union Station, for instance, varies between $3.00 at peak times and $1.95 other times. Metro requires customers to pay more for the train than the bus, likely resulting at least partially in the very different demographics of the city’s rail and bus systems. This is quite different from New York’s Subway and city buses, for instance, which charge a single, set fare at all times and for any journey, no matter the distance.

The “congested core” surcharge now under study for implementation for Washington could go into effect at the “peak-of-the-peak,” between 7:30 and 9:00 AM, and between 4:30 and 6:00 PM, when Metro is packed with riders. The mode of implementation has not yet been determined — nor has the funding device been approved at all — so I won’t get into the nitty-gritty of specifics.

Though this fee would likely reduce congestion at some stations and perhaps induce several thousand employees to change their work hours, its primary effect will be simply increasing the fares of most system users. This would provide immediate financial benefits to the cash-strapped transit agency, but it seems unlikely to solve long-term capacity problems with the Washington Metro or substantially increase the number of off-peak commuters. Most people, it turns out, still need to get home to their families at a reasonable hour, which means that most people will choose to pay the extra fare instead of changing their work schedules.

Thus the fee won’t reduce congestion dramatically — especially since the system continues to see ridership growth.

But more fundamentally, one should ask whether it makes sense for a transit system to charge extra for exactly the service it is supposed to provide best: journeys to and from the downtown core at peak hours. Should the District of Columbia push for years to increase the number of office jobs downtown if it decides to reverse the game later on and disincentivize the use of the region’s primary transit service to get there? Why penalize the people who are using the system in exactly the way that the system was designed to work?

Unlike automobile congestion zones, which are meant to increase the number of people using public transportation and other alternatives to driving (and, in turn, encourage the further densification of core land use), a transit congestion zone serves the opposite role. By increasing the cost of getting downtown by train, it degrades the value of the transit system’s primary use, which is to get from the outskirts to downtown during rush hour. It encourages car use and the build-up of areas outside of the core instead of within it.

Washington’s proposed fee is relatively minor and its global effects would be minimal, but there remains a conceptual gap between the idea of charging more to use center-city stations and the way in which American cities are currently designed. If we’re going to continue the concentration of center-city offices, we need to provide transit that reinforces it, not that works against it. Transit systems play an essential part in organizing regional developmental geography; their fare policies must reflect broader land use goals, not defeat them.

47 replies on “As Congestion Mounts, Transit Agencies Consider Varying Pricing”

Well said.

It’s unfortunate that Metro is in the financial position where it has to make this kind of choice – and while I’m glad they are looking at congestion pricing principles, for Metro to do so while only looking narrowly at their role within the entire transportation system is backwards.

Unfortunately, given the fiscal and political realities, the only way for the region to embrace Metro’s benefit to the entire transportation system is for the local jurisdictions to pony up more money to help close the gap – and these funds have not yet been forthcoming.

I wonder if using the system BART has been using for years would work, not just for Washington but for all subways and light rail lines? You get a ticket when you board and the system calculates your fare when you leave. You either pay then or the fare is deducted from the value stored on you ticket. Thus fares could easily be adjusted for trip length and time of day.


Washington already does this, and as far as I know always has. Many other systems with long suburban sections do as well. The trouble with implementing such a system in a place like New York is that people have to swipe/tap/insert their ticket/card to get out of the system, as well as in. In addition to requiring the replacement of a lot of turnstiles, this significantly slows down the flow of people leaving stations, making the congestion at downtown stations much more of a problem. The zoned/peak pricing system can thus do more harm than good for combating the problem discussed. It’s also more complicated to implement peak or distance-based fares in a system with monthly unlimited-ride passes (which WMATA and BART strangely lack).

Frank, DC actually uses the same system as BART. It already incorporates higher fares for rush hour commutes. This would create a two-tier peak of peak premium for certain central stations.

Ironically, NJ transit just took the opposite step, eliminating the discount for off-peak trips.

How does this differ from zoning as used on a system like the London Underground? Are central stations in Washington already within a more expensive central zone, and the surcharge will be on top of that?

In a peak/offpeak and distance based fare structure, its is creating a de facto more expensive zone.

One thing it would do in an integrated fare system would bribe some passengers heading toward the edges of the congested area to get off at the previous station and take a long walk (or bring a folding bike and ride&pedal).

In a split fare system, its a way to bribe more people to take the bus into the congested core instead – and that is probably silly, since its highly unlikely that the marginal cost of carrying bus passenger during peak is actually lower than the marginal cost of carrying the rail passenger during peak.

It seems to me that when a system is congested, as Metro is, that the pricing should reflect the actual nature of the system. There are likely a number of people using the busiest stations who could get off a station earlier or later and without much difference in their commute. Also there some people who could shift their hours earlier or later. Pricing that recognizes the actual limitations of the system and works with them will lead to reduced congestion, more revenue and probably little ridership. For that matter, Metro should look very carefully at lowering its suburb-suburb and reverse direction fares since it likely has lots of unused capacity that could be used to serve those markets (such as they area) for essentially no cost.

Washington Metro’s fare system is based on distance, not zones. So a short ride during peak hours is going to cost $1.75 (the base peak fare) throughout the system. Even if you’re traveling within DC’s center (say from Dupont Circle to Union Station) the trip costs $1.75. If you’re traveling from one suburb to another (Shady Grove to Rockville), its still $1.75. The trip gets more expensive the further you travel, not whether or not you enter DC’s core.

I wonder if people would feel differently it was framed as an off-peak discount rather than as a surcharge. Doing so exploits a bias all humans have.

Because there’s people out there who look at the picture, not just the frame. A certain trip is going to be priced differently, and higher, than other trips.

The problem is that using prices as a signal to encourage off-peak riding won’t work in the case of employees. Many of them can’t negotiate to start or end their work shifts before the surcharge period.

My alternative would be to impose the surcharge, but offer a complementary service: an express bus that costs the same or less and provides point-to-point service from the busiest originating rail station to the busiest destination station.

I’m about to generalize, and what I’m about to say doesn’t always apply – and I’m not terribly familiar with the situation in DC.

But generally speaking, once a rail line exists between two points, it’s cheaper to move people between them by rail than by bus. (Think about required staffing levels per rider.) So why provide the bus service if it’s only going to duplicate the rail service?

You’re right that many employers have inflexible hours, and their employees may be stuck paying the peak surcharge. But many employers could be persuaded to permit some flexibility if in doing so they’d be effectively giving their employees a (free) raise.

Andrew wrote:

But generally speaking, once a rail line exists between two points, it’s cheaper to move people between them by rail than by bus. (Think about required staffing levels per rider.) So why provide the bus service if it’s only going to duplicate the rail service?

For the same reason Washington proposes peak station surcharges: to manage serious crowding.

It’s obviously cheaper to move so many people by train, but WMATA is imposing these fares because passenger volume exceeds capacity. Either the marginal cost of additional trains is too high, or WMATA is up against the wall in rail capacity and cannot add service.

Assuming WMATA knows the origin and destination pairs of the travel peak, a point-to-point bus can ease the ridership pressures on the train.

Chances are, the ridership might be at the right point in which buses are practical. The buses would only be to mimic the peak origin-destination pair, traffic that likely would only need buses and not a whole train.

Why is it a problem that using off-peak and uncongested zone discounts to shift total patronage around will lead to the people valuing the trip to that specific station at that specific time the most paying more for the trip?

If they want to do it, they should by all means do it as a general peak fare increase and a new uncongested zone discount on peak hour trips.

And if the result is that there is some infill development at some secondary centers outside the congested zone … what would be wrong with adding to density at a location where it is in fact cheaper to provide mass transit service?


What I was suggesting is that in addition to charging a surcharge for use of crowded stations, metro should consider charging less than the base-fare+distance formula for the underutilized parts of the system (which is most of it at any given time).

I really disagree on this one, Yonah. The realities of supply and demand in a peaked transit network are even worse than on roadways, because transit peaks impose their costs on the transit provider rather than just in the form of delays to the customer.

Service added to a network to handle peak-only overloading is very expensive service to operate. For example, drivers usually must be paid four-hour minimums for shifts that are actually as little as two hours, and the peak determines the number of vehicles that must be owned and maintained. To insist that peak service be priced equally to midday or weekend service is to argue that the riders of those off-peak services should subsidize the peak.

I have trouble even seeing a social justice argument, as the average peak-period traveller usually has a higher income than an average midday traveller (because lower-paying jobs, mostly in service industries, are less likely to start and end on the usual peak). So insisting on flat fares all day could actually be seen as regressive. The only credible reason not to surcharge the peak (or discount the midday, which is the same thing) is becuase it’s an added layer of complexity in the fare system. That’s the reason I usually use, but it only takes you so far in a network that’s groaning under peak pressure and desperately needs to encourage people to travel off-peak if they have the option.

I don’t disagree that a peak-hour transit focus is a major problem — I did point out above that it makes transportation incredibly inefficient. I should point out that the trains operating at peak hours are far more efficient per passenger in terms of labor and energy consumption than those at off-peak hours, so your formula that off-peak passengers are subsidizing the peak is not correct.

But my primary point here is not to suggest that we should continue reinforcing the primacy of commuting during the peak hour but rather that when you make a political choice to enforce land use concentration in the central business district, you better ensure that transportation is able to fulfill its role. In other words, if you spend decades building up a dense downtown as the jobs center, you can’t suddenly turn around and tell people that they’re not supposed to commute to the jobs there at a reasonable price when they need to, i.e., at peak hours.

Note that this is not nearly as big of a problem in places with multiple jobs nodes and mixed uses spread throughout a dense core. But the American model has been to push big job growth in the center cities and (mostly) nothing else.

Thus, I’m suggesting that if you create a high enough concentration of jobs in a certain area, you have no choice but to reinforce transportation so that it can handle the peak-hour loads. Forcing people who ride the trains at those hours and into the downtown to pay more is counter-intuitive.

I’m not sure whether it’s a typical example, but Paris is actually more core-dominated than any US metro area. 31% of its jobs are in the city proper. New York might match this, within much more expansive boundaries, but no other US city comes close.

The peak trains have lower energy costs per rider, but for the reasons Jarrett explained, they may not have lower labor costs. This is not as big a deal for the OPTO Washington Metro, but is brutal for FRA-regulated commuter rail.

However, since this is an actual resource inefficiency of increasing density through a single dense core rather than through a network of cores, if there are compensating benefits that justify the single dense core policy, then those benefits ought to be able to cover the differential transport system cost.

The direct approach is to vest some of the incremental benefit of the denser settlement in the transport systems that permit denser settlement, with an ad volorum land tax funding capital improvements, so that in the dense urban core where expensive improvements are required to effectively support the substantially higher value of land per acre, those improvements can in fact be funded.

But if that direct approach is not in place, then the transport system has to harvest some of the extra benefit it provides relative to bus and car transport in some other way.

A (1) general fare increase (2) steeper off-peak discount, and (3) uncongested zone discount allows the Metro to harvest more of the benefit it creates.

Jarrett, the Metro in DC already has higher fares for peak period trips. The idea of a peak of the peak surcharge is nice when it’s framed as a congestion mitigation measure, but since they’ve floated the idea of making it quite large (.50 on top of a max peak fare of $4.60 – you’re talking about $10.20 round trip) and last for 1.5 hours during each rush period (thus much harder to avoid), the congestion pricing aspects of it seem dubious and more as a way to whitewash fare increases that are needed for revenue purposes.

Now, I don’t think they’ll go with a 50 cent surcharge, and I’d much rather see Metro raise fares than cut service.

Again, I think that such surcharges might be counter-productive for the overall transportation system (that peak hour commuter from an outer station with a $10 round trip fare, plus $5 for parking at the Park n Ride might just drive instead), but I recognize that Metro’s in a tough spot. I wish there was a little more regional appreciation for the system that translated into dollars.

Melbourne’s suburban train network (“Metro”) has free “Early bird” rides before 7am to encourage people to travel outside the peak. Effectively it’s a 50% discount; you get a free ride into work, but have to pay to get home again.

(The morning peak is generally more crowded than the afternoon peak because in the morning white collar worker commute time matches school/university commute time.)

Certainly people use it, though the jury is out on just how effective it has been.

Conventional off-peak pricing is not currently offered, though there are discounts for travelling after 6pm, or on weekends (though availability is currently limited).

I agree with Jarrett. I think it’s time that successful transit – i.e. the Washington Metro – think of itself more like a business and less like a utility. Businesses charge more when their products are in higher demand; thus the Metro should charge a higher price. No where does it say that people should be able to travel wherever they want whenever they want at a reasonable price; many people change their work hours right now to avoid congestion, there is no reason that subway riders couldn’t also. We can make a political decision that the central city should have a lot of jobs without ensuring there is adequate capacity for everyone to arrive at 8 and leave at 5 PM; we merely need to ensure that everyone can arrive between 6 and 9 and leave between 3 and 7.

The trouble with Metro running itself like a business is that its competitors, the roadways into and out of DC, are not run as a businesses, but are heavily subsidized and cost the same to use regardless of time or congestion levels. Until there is congestion pricing for auto trips into downtown during peak, Metro introducing congestion-based pricing is likely to be counterproductive and drive users to its competitors.

I think the author is missing a major point here. The reason for the increase in price isn’t primarily congestion: it’s the budget crisis.

Given that WMATA needs more cash from its riders, how should it do that? It could raise the price of parking, it could increase fares across the board, it could target increases in various ways, or perhaps there are other options.

The problem with raising fares across the board is that it would increase the price (and thus discourage use of the system) in uncongested hours, directions, or stations when the system has significant added capacity anyway. Discouraging use then would constitute Metro shooting itself in the foot.

Think of the targeted rate increase less as a way to discourage use at peak times or locations, but more to prevent from increasing the price in off-hours and -locations when additional riders are basically free money for Metro.

So, WMATA has been overly(?) successful in attracting riders, and now wants to penalise them for choosing not to drive??? Did this intensely stupid idea come from the highway lobby or sprawlburb developers? It is WMATA’s job to extract the money from the Feds to get more cars and lengthen the platforms. Because DC is a “company town” it is up to the Feds to either fund more capacity or spread out the work schedules. Punishing those who cannot arrange flex time by double surcharging them is just wrong IMHO.

i think decongestion pricing for transit makes sense because it encourages people to take transit, and to take it at the times that are most appropriate for them. if the pricing is done right, now you’ll be able to actually ride in comfort instead of being packed-in like sardines – that will appeal to lots of folks. it was inevitable that transit agencies needed to try it, so i’m glad they are.

Agree with Peter/Jarrett/et al. First off, this is the basic economics of supply and demand. Second, given that driving/parking during the peak-of-the-peak is an even bigger nightmare, this will shift trips to outside that peak-of-the-peak, relieving some of the overcrowding as Peter noted.

What percentage of riders are visitors to DC? If it’s significant, and if the point is really to reduce congestion, then a system similar to that in London might work.

In London, pass-holders pay only the weekly/monthly rate and can travel at any time. People purchasing just a one-day pass have the option to purchase a cheaper pass that’s valid only after 9.30am. Likewise, if they’re using Oyster (pre-pay) for single trips, it’s more expensive during the morning and evening peaks. Cash fares are expensive at all times. (summary of TfL tickets and fares)

The main effect is to push casual users to use the network after the morning rush is over. Regular commuters are not discouraged from travelling when they need/want.

Actually, passes valid only after about 9AM on weekdays are pretty common all over the world. In some areas, you have the choice between personal or non-personal versions (which are a bit more expensive, but interesting for families, for example).

Actually the Swiss Federal Railways are thinking about such a general network pass, in order to get the ones who don’t need to be on the morning peak hour trains onto later ones.

I’m not sure if labor costs are a limiting factor for short-turn trains? Generally, an operator deadheads to a mid-point station , then boards a turnback train, then takes it to a point, and hands it off to another operator. Often, this is only a few hours, and, as pointed out, there is a minimum report time of four hours.

However, this is a minimum, that means the operator can have an additional train, before or after the short turn train. This would result in a six or seven hour shift, or something like that.

Hours, shift minimums and rules are all negotiated between the unions and management, a give and take operation. It can (and has) been done.

One thing in every transit system is that, if you create a weird shift (say 5-5-10-10-10), there is someone (or a few) that will absolutely love it.

Turnback trains are usually packed to the rafters, at least in one direction, making them immensely profitable. So, even if near empty for the reverse run, they still makes money.

“so your formula that off-peak passengers are subsidizing the peak is not correct.”

I go along with this. Even if running trains back to the reverse-commute direction is expensive; even if shifts are longer than the peak; there’s no way that in most cities the load of peak commuters and their fares aren’t making up more than the difference.

While adjusting pricing for demand is important if you want to be more efficient, even more important is understanding what your market will do when you do that – in this case, the people currently paying your highest fares are also the most likely to bolt to other options if you piss them off. (IE, the off-peak services are full of transit-dependent people getting discounted fares and monthly passes – while the peak service is more likely to have people paying full price but who have a car parked in their driveway they could always go back to).

While off-peak riders can be more transit-dependent as individuals, peak riders are the most transit-dependent for their work trip. At rush hour, roads get clogged, forcing drivers onto trains. Making transit relatively worse by hiking the fare will make a few people decide to bail and drive, but most others will just shill out the extra money.

You must be using a different definition of the term “transit-dependent” than I am; for me, a passenger who has a car parked in his driveway at home is the opposite of transit-dependent; they can switch to driving at a moment’s notice.

Alon, no, that’s the exact definition of ‘choice’ – the dependent rider doesn’t have the option of the 90-minute drive if the 40-minute drive gets too expensive or inconvenient. (Not that even DC traffic is that bad, I’d bet).

I know that in the literature it’s called choice rider. I’m not disputing that term. I’m just saying that because of traffic congestion, peak-hour choice riders often don’t have a realistic highway choice. They’re choice riders in principle, but at rush hour the circumstances are such that they’re transit-dependent.

couple of other comments:

Thus the fee won’t reduce congestion dramatically — especially since the system continues to see ridership growth.

if the goal is to reduce congestion dramatically, i’m absolutely certain they could do it. but that’s not the goal – stated or otherwise.

But more fundamentally, one should ask whether it makes sense for a transit system to charge extra for exactly the service it is supposed to provide best: journeys to and from the downtown core at peak hours.

let’s say we agree with the premise, that rush-hour work commuting is supposed to be the best-provided service — the fundamental question i would have is, should it be? having these ginormous downtowns just seems to me to be such an extravagant waste of resources in so many ways, that now is the time to re-examine the goals for the system as we go forward.

further, peak-hour service is fundamentally a different service than off-peak-hour service, so charging differently for it is to be expected. you eat off the lunch menu, a plate is $5. you eat the same plate off the dinner menu, that same plate is $10. time of day is part of the product/service.

finally, since ‘peak of the peak’ is ‘the peak’, what is the ‘peak of the peak of the peak of the peak’, and how much is it peaking? in other words, the phrase ‘peak-of-the-peak’ makes no sense — unless we’re talking about two distinct levels of peaks with two distinct levels of pricing.

I don’t disagree that perhaps we should question whether having big downtown job concentrations is a good idea — polycentric, mixed-use cities probably work better. But like it or not, Washington has created for itself exactly that huge employment center, so we have to adapt to its needs now even as we may want to plan for its replacement in the future.

I DO disagree – absent the era of subsidized suburban sprawl in the US, human history has shown that the economies of scale produced by job concentrations in core cities are dramatic.

While a large monocentric downtown creates challenges for transit operators, the alternative is far more inefficient. Imagine if all of the meetings and lunches that make DC (and most other industries in the developed world) work required their participants to make trips halfway across a sprawling polycentric region. The capacity strain on infrastructure would be almost as great, the number of trips (and therefore total operating and environmental cost) would be even greater, and everyone would spend far more of their time just getting across town (or not bother, in either case decreasing productivity). There really are enormous efficiencies to be obtained by having everyone work as close together as possible, even if there are some challenges in getting them there. This is the point of cities, which developed on their own long before downtown-promoting government policies.

The problem is that a driving tax is like a cigarette tax – – it will just lessen the number of miles driven, thus collecting fewer dollars.

Gen Y already does not want to drive. And Boomers will soon be driving fewer miles as they hit the magic drive-less age of 65.

So taxing driving will generate fewer and fewer dollars as people cut back on driving, “driven” in part by the tax de-motivation.

Cigarette taxes aren’t meant to generate revenue; they’re meant to improve public health. The same is more or less true for driving taxes.

But even on a revenue basis, consider a case in which Congress raises the gas tax by $1/gallon, a factor of 6 increase. We already know VMT won’t drop by a factor of 6, because in 2008 we saw what VMT reductions come from a $1/gal increase in gas prices, and they’re not that drastic.

I have to disagree with yonah and agree with Bruce. One note here, downtowns are relatively new American inventions dating to the 19th c. prior to that, people did business in cities, but there wasn’t one downtown. and saying paris is more concentrated by using city statistics doesn’t really prove that point. given the infrastructure that DC has built, it would seem that downtown DC has met it’s theoretical maximum in terms of moving people to one area. Is it really more efficient to move people to one center? no. not even from a transportation point of view. It would be better for the metro is there were multiple destinations along the route, and especially at the ends of the routes generating two way traffic and rider turnover. To Bruce’s point, perhaps DC needs to consider encouraging development at other stations. As for premium pricing, it makes a lot of sense and is in no way counterintuitive. it’s demand pricing, it’s completely normal to charge more at peak demand times. If WMATA is out of capacity at peak times and out of money, it’s perfectly fine to charge extra (particularly if the money is being used for Metro service rather than subsidizing an empty bus somewhere). more important than greater operational subsidies is a real capital plan to address these problems. sure, there’s a light rail plan, but where’s the plan for adding four track sections or beefing up regional rail to divert some riders? one more note: NY also has multiple areas, notably downtown, midtown, newark, and hoboken to name a few.

@Anon256: Just one small thing: Many people work in just one building in the downtown area, so they really wouldn’t be affected by the city’s polycentrism because they’d just come in the morning and leave at night like they normally do.

For the people that do have to make those trips, the polycentrism makes it easier, as the transit service is cheaper (because of less peaking) and more frequent.

@eldondre: Well, you could say that all of those downtowns are just part of one big downtown. People are still commuting from the suburbs.

Now if say, Trenton or Paterson were a large employment center, we could call ourselves polycentric.

Leave a Reply