As I noted yesterday, I am taking some time to respond to David Levinson’s incredibly interesting piece on How to Make Mass Transit Sustainable Once and For All.

Yesterday I took up the institutional structure Levinson suggested.

Today I’ll start in the seven points. The first is competitive tendering:

The transit utility can lower costs by competitive tendering for routes. Just as bus companies today don’t manufacture their own vehicles, there is no economic requirement they run the buses themselves. The London model of bus franchises is instructive. Private firms bid to provide service on routes (and collect revenue) for a franchise period. If the route earns profits, they bid a positive amount. If the route loses money, they bid on how much subsidy is required for them to be willing to operate it. Transport for London monitors quality, collects fares (via the universal Oyster card), determines routes, and manages stops, stations, signage, and branding, so it appears as one unified system to riders. Bus ridership in London has risen significantly since competitive tendering.

Competitive tendering works in many city services, so the idea that urban transit providers can contract out for service would not be unprecedented. Many cities, including my own beloved Los Angeles, contract out a million things, including sanitation services, and it all seems to work fine enough. Levinson’s description is so beautifully written and straightforward that I wish I’d written it so I could include it in the chapter on public transit I just wrote.

That said, Levinson does a sloppy here, and he’s a super-genius so he knows better. Yes, ridership in London has risen significantly, but so has population. We can’t conclude that contracting out has affected ridership without controlling for population change and other possible contributing factors, like immigration. As it is, this is post hoc ergo proper hoc arguing, and we need better analysis to conclude that privatization has actually resulted in the sort of service quality improvements that prompted higher ridership over what would have happened anyway. But it does show that contractors haven’t just made a deal and completely gutted service, which is one of the big fears. (And there are precedents for that worry: plenty of Santiago’s bus contractors act like big jerks, rather routinely. Capitalists are seldom worried about labor or customers; they worry about other capitalists.)

Also, London is a pretty exceptional transit market where you have a strong combination of population scale, density and origin-destination proximity. That means there is a transit market there, and lots of reasons for competitive bidding to actually exist. A better analogue for many American transit markets might be paratransit services, many of which are competitively tendered now, ostensibly. I say ostensibly because the efficiency gains you get from competitive bidding result by prompting private sector bidders to discipline themselves on costs, and that happens only if you have truly competitive bidding, and in many markets, you don’t. Again returning to the question of returns to scale I brought up yesterday, you don’t just decide at a moment’s notice you are going to bid on providing bus service to London. You need a fleet. You need operators. You can’t just promise you are going to get them. And that means there are a limited number of legitimate competitors where bidding involves high barriers to entry, such as with high capital costs, like transit.

One of the big controversies over high speed rail construction in California has been that a select group of contractors preparing a bid for the gigantic sugar ocean of HSR construction contracts will receive $4 million stipends just to prepare a bid. You would think that a shot at hundreds of billions of dollars coming in your direction would get companies in the race. But there are strong reasons for paying competitors to compete if you are worried that you won’t get enough bidders because if you don’t get enough bids, and you are pressed to deliver something, that bidding turns into a lop-sided negotiation where the granter does not necessarily hold much leverage. DARPA, for example, occasionally pays stipends to those bidding on particularly complicated or capital-intensive projects.

And yes, while there are good reasons to do it, the arrangement gets back-scratchy in ways that should worry us. So it’s pretty safe to say that London’s competitive bidding would be very different from, say, the bidding that occurs in Boise, Idaho, along with the worries about back scratchiness. Plenty of paratransit customers can tell you: the service can range from terrific to terrible, and just about all of it is expensive. (With apologies to paratransit operators in Boise, who may be wonderful for all I know.) In paratransit’s case, the service is still supplied because it’s required by law. But that’s not the case with contracted buses, and it’s possible to wind up with less service this way.

Second, going to back to Levinson’s own public utility example yesterday, cable companies are a terrific example of franchise bidders who treat their individual customers like crap once they get franchises in place. So the cable guy may, or may not, come between 9 am and 3 pm. Maybe. If he feels like it.

And that’s a really the key point for competitive tendering and service quality gains you hope to achieve: if you are going to to do this, you need to be clear on service expectations. The reason the cable guy gets to treat you like crap is that’s not part of the franchise agreement which centers on channels and rights for particular sports events–not customer service response times. You want cable, you wait for the guy. It’s not like you can go with a better company; you are not really the customer they have to compete for.

In a not-particularly-competitive bidding environment, private sector providers may not treat customers all that well, either, if they have the option of naming the price they want (negative bid) for supplying service the city wants. And the city may not award bids to the best customer servers: think of a company that a) is an easy-peezy contractor, lots of goodwill and hiring each other back and forth, and super-easy contract to manage but b) treats customers badly versus a company that is harder to manage because it’s always demanding things (“my customers need better stops!”, etc) but treats passengers well. The research on public administration suggests a competition between those two does not necessarily go the public’s way. Why? Yes, customers have political oversight and they can complain, but that may not be enough to overcome the internal constituency of the contract managers in the bureaucracy. Watch all seasons of “Yes, Minister.”

Finally, it may be easier from the private company’s perspective to simply bill the city than compete for customers. In theory, and with sufficient competition, they would have the incentive to compete on multiple dimensions: costs, contractual ease, customer service. But in noncompetitive environments in a winner-take-all bid, just one of those can win out, and it doesn’t have to be the first or the last.

If you are interesting in contracting, the University of Maryland’s Hiro Iseki has a lot of papers on the subject.