We subsidize transit to spur development.

We subsidize development to spur transit ridership.

We cannot make up our minds whether we want development to drive transit use, or we want transit use to increase development. Advocates would say we want both to create transit-served high density communities because of all the good that it brings.

But if the transportation-land use cross-dependency is so strong, why do we need to subsidize both sides of the equation? If you subsidized transit to get the positive externalities which are beneficial to development, it might make sense (e.g. maybe transit is undersupplied because the private sector cannot capture the positive externalities due to transaction costs and prohibitions on value capture). If even after subsidizing transit, and there is still not enough demand for private development to proceed without subsidy, maybe you are trying to stimulate development in the wrong place.

What positive eternality is the private sector development producing which justifies subsidy? The best I can figure is transit ridership. (Which isn’t really a positive externality, but might lower the average cost of transit, which is often dominated by high fixed costs, and slightly reduce the negative externalities of other modes.) And we want transit riders and service to justify private development?

I believe that the transportation-land use system acts as a positive feedback system (but a self-limiting one), in the high growth phase of development, transportation investment does drive land development, and land development can induce construction of infrastructure. We have saturated this effect in many places, so the effects of additional accessibility are now quite small on mature networks. The marginal benefits of new infrastructure may no longer outweigh the marginal costs.

The beauty of a positive feedback system is that it takes off on its own, without interference. The subsidy is an indicator the economics of the positive feedback system are not there. The evidence is that there were such effects historically, see the Special Issue of JTLU on the Coevolution of Transport and Land Use for examples, as well as papers cited therein.

Advocates would again say: but you need to prime the pump. To which I respond, we have primed the pump for almost 40 years in the US, with a large share federal capital surface transportation funds going to transit systems that now collectively serve 1.5 percent of all person trips (according to summaries of the NHTS by Polzin and Chu).

To provide some recent evidence, federal capital expenditures on transit in 2008 (from 2010 Conditions and Performance Report, the most recent available) was $6,400,000,000, of which 76.4% was on rail. Federal capital expenditures on highways in 2008 was $34,300,000,000. Transit’s share of federal surface transportation capital expenditures in 2008 was thus 16 percent.

This is not to say any transportation subsidies are inherently justified or unjustified. (That is for another post). It is just to point out that there has been massive transit investment over a period of decades far out of proportion to the relative demand for transit, and that investment has done almost nothing to boost transit’s market share (or ridership), almost nothing to stimulate development, i.e. nothing to “prime the pump”. And if 40 years of sustained investment won’t do it, it seems reasonable to suppose another 40 years won’t do it either.

The reason for this is simple. Transit is not perceived as a better mode for 98.5 percent of trips than something else. It is not faster and/or it is not higher quality and/or it is not less expensive in out-of-pocket terms. It once was, before today’s competing modes had matured, before different land use patterns locked-in.

There are things we can do to make transit better. (And there are certainly things we can and should do to make other modes less attractive, like internalizing negative externalities with better pricing). I think it is technically and economically (though not politically) feasible to double (or even quadruple) transit ridership nationally. But spending scarce resources to subsidize private development to produce transit riders while we simultaneously subsidize transit service to induce private development is not the path to increase in ridership. It just gets us lost.

When transit was on the upsurge from 1880 to 1920, its competitor was walking and horse. It outperformed both. The US was growing fast, and it grew around new transit lines. But with the advent of automobility, paved roads, large lots and thus increased distances between homes, the suburbanization of retail to serve new suburban residences, and then suburbanization of employment, and then new highways to serve these new markets, those advantages withered.

Remember, the built form of the 1880s-1920s still mostly exists, and transit market share is low even there, where it could be much higher. We don’t necessarily need more development in those areas, certainly not development requiring subsidy. We need the residents and workers in those areas to want to take transit more. It already has the potential to be a decent choice, since the land use pattern was constructed with transit as the preferred mode.

We spend too much capital trying unsuccessfully to convince people with better choices to use transit. We don’t spend enough on those without good alternatives and who might travel more or farther, or those where transit is actually competitive, where the built environment still resembles transit’s glory years of the early 1920s.

Instead of requiring the public to subsidize land development near transit, transit should create value for land development that can be captured to help pay its capital costs. If a new route doesn’t create enough value to cover its costs or generate land price appreciation, maybe we should spend finite transit-supportive resources somewhere that does.

9 thoughts on “Cross-subsidies

  1. Jay


    The insight into the lack of positive externalities as a justification for subsidies is a particularly powerful way to think about public transportation expenditures.

  2. Brendon SlotterbackBrendon

    This: "We spend too much capital trying unsuccessfully to convince people with better choices to use transit. We don’t spend enough on those without good alternatives and who might travel more or farther, or those where transit is actually competitive, where the built environment still resembles transit’s glory years of the early 1920s."

    Why is this? Dense areas have lots of voters (although they may vote at lower rates). Is it mostly the fault of federal funding structure (only pay for new, not incremental upgrades)? Or is it political feasibility (easier to build on greenfield or railroad trench in the case of SW LRT) and unwillingness to make grand plans? How do we stop funding trains to nowhere (sprawling suburbs) and instead fund better buses to existing neighborhoods?

  3. David LevinsonDavid Levinson Post author

    @Brendon, Brian Taylor at UCLA haw written a lot about this. See e.g. His case is California, but it applies generally in the US. The problems are several:

    (1) Federal funding for capital not operations

    (2) The multiplicity of jurisdictions, each wanting some service if each pays (spatial equity)

    (3) Regional planning organizations that have to satisfy suburbs (80% of the region), and the cities (20%).

    (4) Socialized transit. If transit were privately provided (and/or user fees were the dominant source of revenue) and only "profitable" routes provided, those would be urban, and a few commuter buses.

    (5) The end of annexation. If the cities continued to be able to annex their suburbs, things would look at least somewhat different.

    (6) The sexiness of ribbon cuttings vs. the sexlessness of continued operations and maintenance.

    There are undoubtedly others.

    One would think more roads would substitute for less transit, but apparently that deal cannot be negotiated. Each jurisdiction wants its share of each pot.

  4. David Greene

    This is far too simpistic an analysis and I expect better from a U of M faculty member. A 1.5% aggregate percentage is meaningless. Transit is not meant to serve everyone. It is meant to serve those in very specific areas. How about the 40% or more of commuters that use transit to get to downtown Minneapolis? How about the overcrowded 16 whose popularity is actually killing the line? We need that rail investment on University Avenue to sustain service.

    Have you looked at the developments along Hiawatha Ave.? I am genuinely curious to know which received subsidies (I have no doubt quite a few have) and which have not (any?). It's an interesting case study, I think, because the window of development hit just as the economy tanked. Yet we are still building there. Developers just don't build anywhere, subsidy or not. They have to believe they'll have a decent return.

    There is plenty of space to build in the cities and inner suburbs. Remember that the streetcars of the 1880-1920 growth spurt in Minneapolis served a much smaller metro area (if you could even call it that). Uptown is a sea of parking lots yet people complain about a lack of parking.

    I believe the jury is still out on Northstar. It opened service just as thousands of people lost their jobs. It's a commuter train, after all. It looks like numbers have risen steadily lately. I certainly don't think we're ready for more commuter rail but I also think it's premature to declare Northstar a complete failure. The lack of a connection to St. Cloud is a big problem.

    This isn't an infrastructure problem. It's quite clear where we can and should build the infrastructure. It's a social problem caused by an overreliance on one mode of transportation. Fortunately the generation after mine is much smarter and is demanding that we do things differently.

  5. Alex

    Transit's commute mode share for the 50 largest metros is 5%, which I think is a reasonable return if the 16% national transit funding rate is mirrored there (my guess is that it isn't – more likely a handful of large metros have larger transit capital funding rates and the vast majority have much smaller).

    Regarding TOD subsidies, I'd guess there are few programs that are no strings attached. I know the first two programs you linked to above come out of affordable housing funding pots, and they've just added the TOD requirement, which I think is a pretty good idea as long as you're already funding affordable housing.

    1. David Greene

      That's a good point. A lot of these subsidies are to ensure that there are affordable units available. That's good public policy. We know from experience that mixed-income housing is a good way to reduce segregation and close the income and opportunity gaps.

  6. Sam

    And if we priced carbon appropriately? And roads appropriately?

    The transit-road comparison is unfair – we're only beginning to reconstruct our urban rail systems, whereas the urban highway system is built out already.

  7. David LevinsonDavid Levinson Post author

    @Sam … Carbon charges are highly speculative, if it were in the "market value" range of $25/ton, it wouldn't make a difference (immeasurably small) in highway demand. Recent prices on the CCX (The Chicago Climate Exchange ) are on the order of $0.10 per metric ton – this of course has to do with the low demand for offsets, which is due to policy, or lack thereof.) In Europe, the price has dropped to 6.15 Euros per tonne.

    We of course should still price carbon.

    Road charges would matter at a few times of day and a few highly congested location (in a "downtown" London, it makes on the order of a 20% difference in daily motor vehicle traffic (ref:… , which is the official and no doubt overstated estimate ) or 16% (… ), I would guess in Minneapolis it would be much smaller).

    We of course should still price roads.

    Reconstruction of the rail network is an interesting question. However we have an extensive bus network already, the rail network will increase ridership (and transit mode share) marginally on that. But the point isn't the car vs. transit comparison (except the side point that car drivers are subsidizing transit a lot, and it hasn't made a difference). The point is that transit should be creating value, and developers should pay to be near transit, we should not have to pay developers for the privilege.

    1. David Greene

      This is so wrongheaded I don't know where to start.

      Everyone subsidizes transit, not just car drivers. Everyone subsidizes cars, not just transit users. Everyone subsidizes every single kind of transportation because it is part of the common good. Transportation doesn't make money, nor should we demand that it do so. Its value is in its necessity.

      The rail network is going to provide a lot of value to a lot of people. Again, aggregate numbers are meaningless. Please consider the residents of North Minneapolis who have been cut off from jobs in the southwest suburbs for DECADES. Southwest LRT will open up completely new employment opportunities for people who desperately need them. Transportation is about connecting people and creating community, only secondarily about moving people from point A to point B. Guess what? A and B only exist because we subsidized the ability of communities to locate at either end.

Comments are closed.