today's attack from reason.com

Tim Cavanaugh at Reason.com attacks me this morning for this post, in which I argued that redundancy in transit networks (as hailed by the Economist in their discussion of my book) often comes at the expense of overall service quantity and thus your ability to get where you're going:

... Walker speaks up on his blog, to explain that when he talks about reliability, he doesn’t mean you should actually let people provide a variety of approaches for taking customers where they want to go: 

"Massive redundancy" may be fine if you're a megacity, though even there, its effectiveness may be a feature of the peak that doesn't translate to the rest of the day.  Anywhere else, services need to work together as a network.  Even in London, New York, Paris, Hong Kong and Berlin, that's really what's happening. 

This is what happens when your mind is full of smart networks and transit-oriented growth. The proper word here is not “redundancy” but “competition.” To the owner of a taxi medallion or a member of the Transport Workers Union, minibuses, gypsy cabs, rolling chairs and pedicabs are all redundant, because you’re already providing all the service a customer could legitimately need. If some abuelita is stuck in the rain for 45 minutes waiting to make one of your smart connections, well, that just shows you need more money so the system can be more efficiently planned. 

Note the use of "competition" not as an idea but just as a mantra.  This last paragraph is so incoherent that I'm not even sure what I'm being accused of, so let me just clarify the question of competition.

The problem with encouraging multiple transit products to compete for the customer along the same path of travel is that transit's usefulness lies heavily in frequency (thus preventing 45 minute connections, for example), and frequency is an expensive resource that must be concentrated so that it can be made abundant.  To introduce competition among transit services going the same way is to undermine frequency -- as in the increasingly discredited British model where people were required to let Joe's Red Bus go by because their ticket was good only on Jim's Blue Bus.   Earth to competition fantasists: Outside of the peak commute, people just want to get where they're going now, but they want this throughout the day, which means they want frequency.  Abundant frequency arises from concentrating and organizing a single pattern of service, not encouraging lots of different services to run on top of each other.

I have never opposed private sector competition.  There's obviously nothing wrong with taxis, pedicabs, etc competing with each other, and even competing with transit.  On the peak commute, transit is usually overcrowded and can benefit from others taking up some of the load (because peak transit service is so expensive).  But outside the peak commute, where frequency matters, nothing can compete with transit at its price point, once it's built up sufficient frequency.  Taxis and pedicabs and autorickshaws can still have a role (a) at other price-points and (b) in places where the geography prevents transit from offering attractive service.

Still, we need to be more critical of cases where we are spending public transit dollars on multiple services that compete with each other instead of adding up to the greatest possible mobility.  Competition fantasists imagine that when Joe's Red Bus and Jim's Blue Bus run on the same route, the customer is being empowered.  Actually, she's just being obstructed, because in most cases, what she wants is any bus, now.

Comments

[Aaron Priven]

If you are reason.com, you don't need evidence or logic to support competition. It's always the only good answer in any scenario.

[David in Ottawa]

Jarrett,

I'm not sure you're responding to this line of criticism in quite the right way.

I don't know if your background is in economics (I don't think it is, but mine is) but what people like Tim Cavanaugh are doing is applying economic arguments developed in *economic space* to *geographic space* without realizing that there are pretty fundamental problems in doing so (and not just for transit).

Virtually all economic theory is carried out in an essentially virtual environment of "economic space" in which location simply doesn't exist (in some respects, it's like the internet). For a lot of economic matters, this is perfectly all right. Discussions on macroeconomic policy and monetary policy really don't have to consider location, for instance. Heck, even in some seemingly geographic issues one doesn't actually need to consider location: take the general observation that increasing road capacity lowers the cost of travelling by car. It seems geographic but in fact it's pure economic space, at least until you start fiddling with capacities on individual roads.

But every now and then economics by necessity veers off into geographic or real space in issues such as land pricing and infrastructure provision (including transport). Location is not something that can be ignored like it is in economic space. It takes quite a bit more economic skill and an appreciation for the limitations of economic theory in geographic space and how those theories have to be modified to actually be able to say anything meaningful about economics in these circumstances.

[Jarrett at HumanTransit.org]

David.  Very helpful comment, thanks!  

[david vartanoff]

Well said, Jarrett. Your final paragraph makes clear why "commuter rail" should be fare integrated w/ "mass transit" as for instance Metra and CTA in Chicago.

[Daniel]

It's interesting how these folks consistently attack the inefficiencies of all public systems (government waste!) while consistently celebrating the inefficiencies of all private enterprise (creative destruction of competition!). The more reasonable approach is to recognize that all systems are inefficient to some degree and pick the one that is the least inefficient at solving the particular problem at hand.

[Tom West]

Here's an economic argument agianst having compteing transit providers:
If you have two competeing bus services providing half the frequency, the competition will not lower the price by enough to compensate for the extra wait time.

(In fact, the value of time for the extra wait is generally much more than the fare).

[Ben Allen]

One thing that Reason and Reason-oid media outlets continually miss is the value of network effects (bafflingly, they manage to miss it even when they directly mention the value of being in a big network, as in this article). This sentence:

You don’t need experts planning out more brilliant three- and four-transfer itineraries. You need more shit running more frequently to more destinations.

drives me up the fucking wall. Yup, you need more shit running more frequently to more destinations. But it needs to interoperate for users to get any value out of it, and if by any chance the best-case scenario happens and "competing" companies manage to work out fare sharing arrangements to allow for interoperation, you might as well just have one group doing it. Because really, that's what you've got.

Reason, with a capital R like that, has no idea how to deal with transit, just like they have no idea (really) how to deal with Facebook or Google or Wikipedia, even though lowercase-r reason makes it pretty clear that things like transit (or facebook, or google, or, hell, roads...) work best the more people use one system. Because Reason (unlike reason) is dumb, they don't understand that a) there is a natural monopoly here, and b) competition is remarkably bad at deciding who should have that monopoly — it's a situation where democratic rather than capitalistic decisionmaking is head and shoulders better by any reasonable (as opposed to Reason-able) standard.

[Anon256]

To be fair, the issue with British buses was essentially a failure of the regulators, for actively trying to stop fare integration when they should have mandated it. There are plenty of other cases (Tokyo subways, Singapore buses, London buses, British trains) where multiple private providers compete on parallel or connected routes but have integrated fares and reasonably complementary schedules due to the regulator mandating or encouraging coordination, rather than banning it as collusion.

[ant6n]

If you want to leverage market competition in public transit to enable cheaper service, you should let transit operators compete for the contracts to run a given system for a given amount of time. From the point of the passengers, it doesn't make a big difference, but it may allow cheaper operation and thus require less subsidies. For something like this to work, the contracts have to be pretty tight so that the private operators can't skimp on service, or delay often etc.

This way of organizing service also means that the public transit has to use standardized technology so that multiple operators can come in and bid. This is of course trivial for buses, because roads are standardized etc. For rail, this is much more important, etc. etc.

[Ben Smith]

As big an idiot as the current mayor of Toronto i, the previous mayor really screwed up things when it comes to redundancy. Currently the province is implementing a smart card fare system which is to work on all transit systems in the region, so you don't need to carry multiple fare media around to use multiple systems.

The previous mayor however did not want to get on board with this smart card, and instead pursued an "open payment" system where it would deduct the fare directly from your credit or debit card. He claimed this was the next generation, and that smart cards were outdated technology.

Not only does this ignore the purpose of the card was to create an integrated way to pay for fares across the region, but it also misses the ability to deal with the redundancy issue you point out. Because the subway doesn't go to the city limits, suburban transit system buses follow the same routing as city ones to get to the station. The result is that you can watch several buses go by before you can take one that services 'your zone.'

To the topic of actual overlapping service, say local alongside rapid or express, I see this as a very controversial topic among transit planners and enthusiasts. On one hand, you are duplicating the service along the corridor, which could be argued as a lack of funds. Should the new metro or light rail line have the same stop spacing as the bus it replaces to ensure easy accessibility to transit? On the other hand, they are two different transit services meeting different needs along the same corridor. If someone nerds to travel a short distance they can take the local line, while someone who needs to travel further can take the express line which stops less frequently. Different strokes for different folks.

[Miles Bader]

Well if reason.com's attacking you, it's a sure sign you must be doing something right! ~~

[Andrew]

In a very high density area, it makes perfect sense for several services to compete because they are all overcrowded. For example, in Paris between Châtelet-les-Halles and Gare de Lyon there are no less than 4 rail lines (Metro 1, Metro 14, RER A, RER D) all of which are packed. In a low density area, a common mistake of transit operators is to have several routes with low frequencies running parallel to each other to make all residents have a short walk to a bus stop but none of them have good service. The solution tends to be to cut service on the minor route to very low frequencies or cut it entirely, while reallocating buses to the major route. Suburban bus services in the Greater Toronto Area have a big problem with this. For instance I think that York Region Transit (bus system in northern suburbs in Toronto) should cut frequencies on minor routes in residential areas with low ridership, while increasing frequencies on main routes like Major Mackenzie, 16th/Carrville/Rutherford, Jane, Bayview, Leslie etc. all of which have poor off peak frequencies particularly on weekends; since choice riders will not use routes with infrequent service, ridership will increase as a result.

[Chris]

Anon256 - Blaming regulators for the British bus mess is a bit unfair, as they're doing exactly what they are supposed to be doing under a system of unfettered competition.

The sad joke is that what this 'free market' system has produced is several large companies (effective monopolies in most areas) whose dominance of the market makes any attempt to regulate them impossible; as they simply threaten to withdraw service or offer unrealistically high prices for commissioned services knowing that no other company in the area has the vehicles or drivers to put in a competitive bid.

[Rob]

I was amused that the blogger went on about transit advocates who don't use transit (not sure I know any of those), and then picks on you of all people. But then he admits he's reviewing two authors, neither of who's books he's read! Irony is truly a lost art.

I was also amused to see that the column is called "hit and run," and they punctuate that by not providing an opportunity to comment. That may just be the philosophy there! I know there are people at Reason who take their rationality seriously, and I'd hope they would find this embarrassing.

[kamil]

The creative destruction of competition is not inefficient by definition. It happens when a less efficient company is destroyed by the more efficient competition.

It is only inefficient if the inefficient company could somehow figure out its goods or service were not needed because of future more efficient competitors and not invest in production capacity that will be destroyed. This would require a time machine.

The understanding of economics would have to increase as much as Reasons understanding of how geometry and geography affect economics to be able to pick which combinations of systems to solve a particular problem at hand efficiently.

The Reason article is ignorant ignoring long understood economics such as network effects.

[EngineerScotty]

Rob,

The Reason article accepts comments--at least it did when I looked it at. Unfortunately, the comments contained little discussion of substance (but quite a bit of spam left behind by pornographers...)

[MaxUtil]

Reason's complaint (at least the coherent one) largely stems around the fact that connections are not well integrated and timed so they add significantly to travel time. So his solution is LESS planning for the system? Got it. If poor Jarrett had ever actually ridden a bus in his life, he would understand this!

[Alon Levy]

You need to understand that for Reason, all government planning is definitionally bad (the45-minute connection canard), and if government planning happens to be good, then it's stifling freedom and needs to be removed anyway. There are people pushing for dismembering Switzerland's rail system through EU-style separation of infrastructure and operations, but so far they haven't gotten anywhere, and SBB's line on the matter is that what they have now works.

[Max Wyss]

To Alon Levy: Infrastructure and operation are legally separated in Switzerland. They may still bear the SBB label, but that's it. This separation also lead to the situation that most of the local traffic in the Canton of Bern and surroundings is operated by bls, but on tracks owned by SBB, which are alsoused by SBB (Passenger unit) for interregional and intercity traffic.

And the system works. The system works because competition (or better said, artificial competition) of service is stupid and actually costs (the taxpayer) more than an integrated fare system.

[Alon Levy]

So, just to get this straight, the sort of privatization that SBB was complaining about is not EU track separation, but British privatization of services?

[Max Wyss]

I would need the source of the SBB complaint to interpret it accordingly.

There was a whole deal concerning the Canton of Bern; SBB handed over the regional train operation to bls (that's where the Canton of Bern (and the other cantons concerned) order and pay for the service. On the other hand, bls (or at that time, BLS) handed over the intercity services to SBB (essentially (Bern)-Thun - Spiez - Interlaken and -Brig ). There was also an exchange of rolling stock involved.

I can see some reasons why the SBB is not happy with the divisionalization; it requires a proper allocation of rolling stock (primarily locomotives) to an operating unit. As the SBB has always ordered multi-purpose locomotives, usable for any service, the efficiency goes down the drains, if there is a lot of paperwork etc. to be done to use a "Cargo" locomotive in "Express" duty, and vice versa. And it is even worse with the availability of drivers; there are assignments where the (driver spends more time deadheading than actually operating a locomotive…