On Incompleteness (of Networks, of Life …)

Reader Russell Bozian thinks he spies a theme coursing through recent posts.

Will partially built houses ever be energy efficient? Ones where the walls are up, but the roof has not been put on yet? If you don’t qualify for a full home loan, will banks even lend you half the money you need to build a house? Will the banks figure that you can at least have half a good life, living in a house with walls but no roof? …

Jarrett, your original post wonders out loud why Portland can spend tens of millions on transit and not, in 2010, see a much higher percentage of people commuting to work on public transport. But at the present build rate, won’t it take at least until 2050 before Portland has a comprehensive and ubiquitous public transport network, such as we see in Manhattan? Why do we starve public transport of the money it needs to finish a decent network of routes, and then pause to criticize its incomplete performance statistics? Would we ponder why the waterworks is not delivering much water to our faucets, if we only gave them enough money and right of way for each water line to stop 100 feet short of our houses?

There are some kinds of public projects, like building air traffic control towers, that you just need to buckle down and spend money on to finish–without pausing and looking at the half-finished tower and pointing out that not many planes are being guided yet. Public transport networks to replace automobiles is one of those kind of big, expensive projects that (a) many other cities in many other countries have completed successfully and (b) do not lend themselves to meaningful performance analysis in their incomplete condition.

Russell is talking, of course, about the power of network effects.  Every addition to a network makes all parts of the network more useful, so there should be a multiplying effect as a network expands. But unlike a house, it’s hard to decide when a network is complete.  Even if you do, people will still think of things you might add, just as you might add things to houses.

Few rallying-cries are as satisfying as “Complete the Network!”  But history’s cruel joke is that we never know a network is complete until we notice we’ve stopped building.

The road network looks very complete to me, as it does to Russell, but many road engineers don’t think it’s complete at all.  Look at San Francisco:  A freeway from the north coast extends for over a hundred miles, climaxes at the Golden Gate Bridge and — ends!  With no connection to the freeway on the other side of the city!  The very same thing happens if you follow Interstate 5 across the Canadian border and into Vancouver.  It just ends!  If you want to get from the freeway on one side of Vancouver to the freeway on the other, you have to use city streets.  If you’re used to the American model of continuous-welded Interstates, these places are like some kind of eternal construction detour, and generations of well-meaning freeway-trained motorists have asked: “When are they going to complete it?”

They’re not.  Or rather, it’s complete now, in its fragmentary state, whatever that is.

None of this disputes Russell’s fundamental point, which is that network effects matter.  The more places a network goes, the more useful each part of it is.  We have to keep saying that.  It’s an important point and one of the few transit-related insights that motorists can instinctively understand, because it has a direct equivalent in the motorist’s world.

The goal of completeness can be rhetorically useful, but don’t get too attached to it.  It’s death, in more ways than one.

34 Responses to On Incompleteness (of Networks, of Life …)

  1. Kyle February 1, 2010 at 4:52 am #

    If you’re expanding the network with little modal share gain perhaps there are two things that should be examined:
    1) Is the city saving operational expenses by shifting riders from motor buses to rail? If the city is saving a bundle, and other levels of government pick up the capital cost or the vast majority why not.
    2) Are the statistics meaningful? I know in Canada the statistics for modal share include school trips which might explain the different results.
    Moving on from these two things, I find it hard to believe that your commenter is trying to justify stagnant modal share on network effect. If network effect exists at all (which it does), expansion of the network should always improve share even if just a bit. You cannot say network effect does not work when you have 3 lines and add a fourth, but it does work when you have 8 lines and finish the ninth.
    I submit this for Portland: network effect is low in the city due to low frequency on the outlying lines. It is inefficient to transfer so the modifier is zero. Unfortunately this is hard for Portland to fix due to constraints on its infrastructure, but it is solely fixing it with first the second downtown loop and now a second river crossing.
    This isn’t too hard to understand – Denver at one point caused a negative network effect on its LRT and saw ridership drop. When they first opened the line to service Union Station and the sports stadia they diverted some of suburban line runs that used to run on the standard route to Union Station. Network effect isn’t always about amount of network, I guess it is about where it goes too.
    Which I guess gets to the point: should Portland have built to the airport or concentrated more on work trips?

  2. Danny February 1, 2010 at 6:10 am #

    Network effects are extremely important for transit…but that doesn’t mean that an “incomplete” network doesn’t exhibit any network effects. There should be measurable change even with small improvements in networks.

  3. Ted King February 1, 2010 at 10:39 am #

    1) US-101 / CSR-1 does connect decently with I-280 via Park Presidio Blvd. (PPB) and 19th Ave. The green belt that brackets PPB makes that semi-drag-strip more tolerable (my parents lived near 14th + Lake for a while). Upgrading those streets to a highway would be hugely expensive and tear the hell out of the surrounding housing.
    2) SFMuni’s pseudo-Metro is a forearm-to-fingers network with a long tail. One of the missing pieces is a connector from the L-Taraval terminus (near the zoo) to the N-Judah terminus (near the windmills in Golden Gate Park). The benefit of such a connector lies in being a fall back to the Duboce tunnel.
    The network effect of such a connector lies in the possibility of an extension north to 33rd + Geary / PLH or an extension south to Stonestown Mall and SFState. This line (B-Beach ?) could tie a major art museum (Palace of the Legion of Honor) to the zoo, shopping, and a large university. There is a bus line currently serving this route – #18-46th Ave.
    SFMuni’s Metro :
    Short tail (Ballpark / Caltrain) – N-Judah / T-Third
    Long tail – T-Third
    Forearm – Market St.
    Palm – Duboce + Church Sts.
    Thumb (folded over) – J-Church
    Fore-finger – N-Judah
    Other fingers – K-Ingleside, L-Taraval, M-Ocean View

  4. EngineerScotty February 1, 2010 at 11:20 am #

    Saying that the San Francisco road network only consists of freeways would be like saying that the New York public transportation system only consists of the subway. While it’s a good analogy, it is a flawed one, as cars crossing the Golden Gate don’t find themselves forced to turn around and head back to Marin–they can indeed, as Ted notes, proceed through the City. If they want to get downtown, they keep left after the tollbooth and head down Marina; if they want to head to the Valley, they bear right onto Park Presidio. In either case, the freeway soon ends but the pavement does not.
    On the other hand, once you get outside the core of Portland (and many other US cities), its not hard to find many places where the nearest bus stop is a half-mile away, and the bus only comes every half hour. The equivalent situation in San Francisco would be the for drivers crossing the Golden Gate on US101/CA1 to have their choice of two unimproved dirt roads, with tire-puncturing rocks and deep ruts abounding in either case. Were that to be the situation, it wouldn’t matter if the Golden Gate were its present six lanes, or magically widened to twelve–its utility as a transit link would be severely limited.
    Likewise with Vancouver–those heading north into the city from Richmond don’t find themselves on an offroad trail suitable for 4WDs only–instead they have their choice of Granville or Oak or Cambie–all wide thoroughfares which suit their automobile just fine; the only issue is they have to wait at traffic lights which distribute the traffic jam. (And if you want to get from Surrey or Richmond to Whistler, Marine Drive to Boundary Road is an excellent bypass of downtown).
    Certainly, no network is ever complete. But some networks are certainly more useful than others; in many developing lands, automobiles (especially those incapable of offroad driving) are of limited use because so many places are not reachable by pavement.
    To assist with the quandry of whether a transportation network is “complete”, I propose the following simple question: Is the network constrained by reach or by capacity? If the biggest problem faced by the transit operator is that much of the provided service is inefficient “coverage” service, it’s hard to argue that the network is, in any sense, complete. If, on the other hand, the biggest problem is handling the rush-hour load–that’s a major milestone. Road systems in the US and elsewhere frequently have the latter problem; whereas most transit systems outside places like Manhattan have the former.

  5. Ted King February 1, 2010 at 11:22 am #

    To those who want I-80 heading east –
    3) You’re supposed to have split off back in San Rafael for the Richmond – San Rafael Bridge (I-580) or back further north (Novato) for Vallejo (CSR-37). The indirect connection in San Francisco is :
    CSR-1 (19th Ave./ Junipero Sera Blvd.)
    Brotherhood Way East
    Alemany Blvd. to San Jose Ave.(CSR-82) ramp
    I-280 North
    US-101 North
    I-80 East
    . Yes, you’ve got to be a local resident to understand the asphalt and concrete spaghetti. There are other routes through San Francisco that are shorter but way more confusing.
    P.S. Looking back at what I’m posting proves two points – (a) that the highway engineer’s dream of a connected network is dead in San Francisco and (b) that San Francisco residents value their street network over a fast, highway network. It may be crazy, but it’s also fun and HUMAN.

  6. Danny February 1, 2010 at 12:35 pm #

    Umm…This just came to my mind, so I thought I would throw this out there even though it has probably already been discussed: When MAX and the streetcar came, they replaced existing transit bus lines. There was no transit network extension involved. But there was some transit network deconstruction, because of bus cuts.
    So this story might still be one of network effects, but different from how Russell said. Instead of lower ridership being explained by an incomplete network, it might be explained as lower ridership from a shrinking network.

  7. EngineerScotty February 1, 2010 at 2:37 pm #

    Depends on which MAX line you are speaking of.
    The original MAX project did result in some reconfiguration of eastside bus lines; with some outer lines being converted into feeder lines (and some express lines being discontinued). However, the project did represent a substantial increase in capacity.
    The westside MAX line added substantially to westside transit capacity; which never was very good previously. Part of the reason was the West Hills bottleneck (on the east side there’s a redundant street grid; there’s only a few through routes which pass through the West Hills).
    Not sure what bus lines changed when the Red Line went online. The Yellow Line did replace the old 5/Interstate bus, but with a higher-capacity service. The Green Line, likewise, didn’t directly replace any bus service; many bus lines cross the I-205 green line segment (or begin at Green Line stations, such as the 14/Hawthorne); these are still there. (The opening of the Green may have exacerbated systemwide service cuts as a result of the recession; but those were not planned changes in service as a result of MAX).
    Likewise, the Streetcar (which is a different matter, as it is largely funded by the City, not TriMet) didn’t directly replace any bus services.
    A common narrative is that TriMet is cutting bus to add rail; as if the former is more important or deserving of funding than the latter. In some cases, sure–the Yellow Line replaced a bus with a higher-capacity line. And in the grand scheme of things, as TriMet has had to cut services due to declining revenues; bus service has been cut.
    A question for you: Do you think that TriMet should abandon MAX (or some part of it), mothball the trains, and re-dedicate the operational monies to additional bus service? What MAX service would you reduce or eliminate, and where would you add the additional busses?

  8. Danny February 1, 2010 at 4:09 pm #

    Regardless of the stupidity of a decision to install train infrastructure, it doesn’t make rational sense to abandon rail service to go back to bus, because the rail infrastructure is a fixed cost and the operation of that infrastructure is less costly than bus. Therefore, no I wouldn’t get rid of the rails.
    There is a fundamental problem that people rarely understand here. You see, rail does make sense when you understand operational costs, especially at high capacities. But it is capital intensive, and money is money. This means that you are paying huge amounts of money up front and sinking that cost, leaving you little room for flexibility with operations.
    In an unchanging world, MAX might have been an incredible investment. But government entities are subject to the business cycle too…and when funds run dry, service cuts happen. The existence of sunk cost rail almost purely dictates that when service cuts happen, its going to happen with the much more operational cost intensive bus operations.
    You may not see the cause-effect, because it is rarely visible to the outsider. I can’t tell you what I do for a living, but I can tell you that I am dealing with these types of strategic decisions on a daily basis. When you put down a lot of money up front to save money on operations, you ultimately limit the flexibility of those operations. Rail may make sense when times are good, but is the lack of flexibility worth it when times go sour?

  9. EngineerScotty February 1, 2010 at 5:11 pm #

    When it comes to publicly-funded transit infrastructure, it isn’t true that “money is money”–capital funds and operational funds are largely distinct. I’m sure you know this very well. TriMet has shown an excellent ability to go get its projects funded, and I think MAX helps rather than hurts in this regard. While no mode of transit pays for itself in Portland, MAX has a higher FRR than the bus system does. Certain bus lines may have a higher FRR than MAX in aggregate–but an (operational) dollar spent on MAX tends to provide more mobility than the same dollar spent on busses. Given that MAX ridership is quite good by US standards, I don’t see it as TriMet’s problem.
    TriMet’s fundamental problem is like that of most transit agencies–operational funds are limited, and dependent on a tax structure which is subject to business cycles. TriMet has exacerbated this problem with a few bad investments–not MAX; but the WES commuter rail line, a disasterous futures contract on fuel which has it paying above-market prices, to name two. Unlike capital funds, which can be gotten from the feds, money to pay for drivers and mechanics and gas and electricity has to come from local sources–payroll taxes, fares, and advertising.
    At any rate, the cost per boarding ride for the two systems were $2.89 for the bus, vs $1.91 for MAX. If everyone paid full-fare and didn’t transfer, MAX would actually pay for its operations. Once you factor in discounted fares, transfers, fare evasion, etc. MAX requires a subsidy, but still far less than the bus system. While the presence of fixed infrastructure limits TriMet’s flexibility in routing–MAX can only go where there are rails–the agency gets a lot more bang for its buck.
    Interestingly enough, TriMet released a report a few weeks back indicating that the recession has produced a significant decline in ridership–mostly on the bus system. (News story here, agency news release here). MAX showed a small gain, in large part due to the opening of the Green Line (subtract it out, and the remainder of the system shows a loss). Keep in mind, these numbers were for December 2009, BEFORE the most recent round of service cuts took effect.
    Now, I’m gonna beat a dead horse I’ve been beating for a while: I think the fundamental argument concerning TriMets rail investments isn’t bus vs rail–its traditional riders vs “new” riders. Its the guys who have been riding the bus for years, many of whom don’t have a car and have been loyal patrons, upset that “their” service is being cut for suburban commuters and Pearl District yuppies. Riders who live in close to town, whose service needs are met by locals, and who don’t really care about rapid transit service outside the core. Of course, reality is more complicated than that, and TriMet’s service district (and tax base) is the entire metro area, not just the city (this is a GOOD thing about Portland-area transit, as you don’t have a zillion different agencies fighting over turf). But there are probably a lot of folks who ride the #14 into town who would gladly give up MAX service to Hillsboro for higher bus frequency along Hawthorne; and who are more passionate about transit than the folks out in the ‘burbs who can always drive.
    I don’t argue that the net effect of new rail construction + decrease in operational funds = service cuts for the bus system. I disagree somewhat with the framing of the argument that many advocate (and which you may not be advancing)–that busses ought to have first priority over rail, and that rail is “poaching” moneys that rightfully belong to the bus system. TriMet is running a transit system, and needs to balance the needs of all its customers and potential customers.
    TriMet’s service does focus mostly on the inner city, in particular a roughly square region bounded by the West Hills, Columbia Boulevard, I-205, and Johnson Creek to the south–which makes sense, because that’s where the density is. But were it to ignore the outlying areas completely, it would risk putting its tax base in jeopardy.

  10. Alon Levy February 1, 2010 at 6:38 pm #

    The part about network effects is real, but even a small system can get decent ridership. The obvious comparison here is between Portland and Calgary, which opened light rail around the same time.

  11. Danny February 1, 2010 at 6:51 pm #

    Well, I mostly agree with you, but even in the case of transit, money is money. Sure, you can get money from the feds, but most light rail systems are mostly funded by bonds and loans. I don’t know much in particular about Portland’s system, which is my my posts about Portland are mostly speculative, but I’m pretty sure TriMet is still paying for MAX through debt service. Correct me if I’m wrong.
    How many more rail projects are Portland/TriMet considering in the near future?

  12. EngineerScotty February 1, 2010 at 7:35 pm #

    Tri-Met does have some bonded debt backed by property taxes, but that’s entirely separate from the payroll tax that funds operations. It is possible, though unlikely, that in lieu of MAX construction, a payroll tax increase could have been implemented.
    As far as rail projects being worked on in Portland’s future:
    * The eastside streetcar loop (a Portland Streetcar project) is under construction, and scheduled to begin service in 2011. The first new track was laid today, actually. This isn’t rapid transit, but inner circulator service; and is being financed by the City, not TriMet. TriMet and the City will jointly fund ops when it opens.
    * The Milwaukie MAX line will start construction in 2011; this will include a MAX line from the PSU campus (where the Green and Yellow lines terminate), across the Willamette River on a “green bridge”, down to the inner-ring suburb of Milwaukie. This project is frequently referred to as the “Orange Line”, but what color it will be when it enters service, I don’t know.
    * The proposed LO streetcar project is in the DEIS phase; this is an extension of the current Streetcar along an existing abandoned RR ROW, to Lake Oswego (another inner-ring suburb). This is probably the most contentious of the projects, as it has dubious (if any) mobility benefit. Unlike the existing Streetcar and the Loop extension, both of which are new services, this project as currently proposed would replace a faster bus with a slower streetcar, and require many existing bus users to transfer rather than enjoy a one-seat ride.
    * Further out, the Barbur Boulevard (OR99W) corridor into the southwest suburbs, has been selected as the next mass-transit corridor after the Milwaukie line. This is still in the very early planning phase however.
    In addition, if and when a new bridge is built across the Columbia River, it likely will have LRT tracks and a northern extension of the Yellow Line.

  13. Pantheon February 1, 2010 at 7:52 pm #

    This argument has some merit, but public transit networks should be useful even if they are not “complete”. They are not like air traffic control towers or houses without roofs. Transit is a reasonable option for many more people in Portland than the number that actually use it.
    My problem with this argument is that it gives politicians and transit planners an unlimited number of “get out of jail free” cards when their investments don’t show results. The network isn’t “complete”, you see, so we need to wait until 2050 before judging. And what will the excuse be in 2050?
    This is the perfect argument for those who are in positions of power, because it allows them to continue to throw money into bad investments while avoiding even the slightest degree of accountability for their failed performance.
    There is a growing realization in the U.S. that the country is too dependent on automobiles, and many mistakenly hold Portland up as an example of what a city should do to get people out of their cars. The numbers suggest that Portland is in fact a great example of what not to do to get people out of their cars.

  14. EngineerScotty February 1, 2010 at 9:18 pm #

    For many people, simply being a reasonable option is not enough; transit must be THE most reasonable option. (And for some–the “transit-dependent”–transit is already).
    If anything, I think this shows that there is a limit to the effectiveness of the carrot–if you build it, it’s not a given that they will come. Some will, some won’t. In places with high transit usage, though, the stick is also present: Gasoline is expensive; parking is expensive; auto licensing is expensive; and the road system is capacity-constrained compared to potential demand. In Manhattan or Hong Kong, the bus or train is the path of least resistance.
    So far, the metro area has been reluctant to employ the stick. As discussed in a prior thread here, parking downtown is cheap (and virtually everywhere outside the downtown core and a few college campuses, free). No tolls are collected on the highways; gas taxes are dedicated to roadworks; license fees are low; and while congestion on the freeways is a significant source of complaints, it remains tolerable on most days. Whether or not there is political support for the stick to be wielded is an open question–but many motorists would be thoroughly pissed off were it to be tried. If a stick is to come about, my bet is that it will be caused by higher gas prices (transit usage spiked at $4 a gallon) or continued in-migration, assuming no further expansion of the urban growth boundary.
    So. If you are arguing that Portland’s numbers indicate a failure to make driving inconvenient, I’ll agree. No doubt about it.
    OTOH, if you are suggesting that the problem isn’t the lack of a stick but merely the shape of the carrot–that we’d see better numbers if we had invested transit dollars differently–I’m not so sure. Local-service only would mostly likely fail to attract anyone but those with no other choice out in the suburbs. Express service can help with commuters; but serves far less people than rapid transit. Use of BRT rather than LRT is an interesting thing to consider; and if you tell me the Yellow is too slow (barely faster than the bus it replaced), I’m more than willing to listen to that.
    But fundamentally, I think, the cause of Portland’s lower-than-we-would like share is not due to the particulars of the transit system; but due to the particulars of the road network. It’s easy and cheap to drive here.

  15. anonymouse February 1, 2010 at 10:14 pm #

    Pantheon: it really depends on how you define “reasonable”. For example, most of LA is covered by a pretty comprehensive bus network, and you can get from anywhere to anywhere else. But it often takes 3 times longer than driving, and that means that a “reasonable” 45 minute trip becomes rather lengthy 2 hour and 15 minute journey, to the point where just not going is often a better option that taking the bus. And if you look closely at Portland, it seems like even the light rail generally has a faster alternative in the form of a limited access highway, and of course driving on surface streets will always be faster than buses in mixed traffic.

  16. Pantheon February 1, 2010 at 10:48 pm #

    You are making my point more eloquently than I did myself. By “reasonable option”, I mean a user has access to the network, both at their origin and destination. Therefore an “incomplete” network can be useful (unlike an incomplete air traffic control system) if it offers an acceptable level of mobility for the areas it covers. Therefore the relative incompleteness of the network can not be used to explain away zero mode change. There has to be something else at play here.

  17. EngineerScotty February 1, 2010 at 11:17 pm #

    There is–driving is still easier. Much of Portland’s land use patterns are still conducive to driving. Things would probably be worse without the urban growth boundary; but neither the city nor the surrounding suburbs have hit any geographical limits to their expansion. Unlike Seattle, Vancouver, and the greater LA basin, which are wedged between water and mountains, Portland has plenty of room to grow.
    And the effectiveness of the UGB itself may be questionable, given the lack of political will to not expand it that seems to be in evidence… Washington County wants 30,000 acres for itself for example. There’s plenty of room for infill (there are numerous hobby farms still within the UGB); but developers don’t like doing infill and neighbors don’t like it either–the pressure is on to build more subdivisions as soon as the housing market picks up again.
    THAT, more than any question of bus vs rail, is why transit share remains stagnant.

  18. Russell Bozian February 2, 2010 at 1:38 am #

    Pantheon, you’re equating “theoretically useful for moving people” with “likely to persuade anyone that their life’s going to be fine without driving”.
    No-late-night-service public transport fragments (Portland’s 50 miles of rail and 80 bus lines) while theoretically “useful” are non-competitive with COMPLETED networks (like Portland’s what, 10,000 miles of roads?) from the point of view of someone who can afford a car. Vis. EngineerScotty.
    Yes, just like a non-air-conditioned air traffic control tower in Las Vegas will not be preferred by ANY air traffic controllers in the old tower until the new one’s HVAC system is finished.
    By your logic it was foolish to let the politicians at the turn of the last century build a paved road network across the U.S. Because after all, there were 3 crazy decades of no-result (in the metric of cross-country driving preference) road spending. But because the public understood what the utility of the road network was going to be, they waited for the highway network to be done. They didn’t stop in 1910 and say “gosh, all the millions spent, and almost nobody’s bought a car yet”.
    You’ve said it allows [people in power] to continue to throw money into bad investments, but you’re the one who has effectively defined all decades-long transformative public transport network projects as bad investments.
    My basic problem with your argument is that you don’t know if a great public transport network completed in 2050 will get any riders in 2050, because you take no lesson from all the cities in the world with completed, well-used, great public transport. You are looking at a 500+ square mile region with 50 miles of light rail and 80 bus lines and saying “there’s no such thing as New York City or Paris, so how do we know that people will ever use a great public transport network in Oregon?”.

  19. Alon Levy February 2, 2010 at 4:28 pm #

    Russell, I haven’t seen any negative comparisons of Portland with New York, except by New Yorkers who want everyone to think New York is the best city in the world. The comparison I’m making is with Calgary, which has fewer route-km of light rail than Portland. Vancouver has fewer route-km of rail than Portland and several times as much ridership as well, but it uses driverless els and not LRT.

  20. Pantheon February 2, 2010 at 7:51 pm #

    Russell Bozian,
    I am equating no such thing. I am saying that if a network expands, a greater number of people have access. If a greater number of people do not choose to use the network as it expands, then there must be something wrong with the network or the alternatives are too attractive. Or both.
    I am glad that EngineerScotty and yourself have adopted my argument that driving in Portland is too easy.
    I don’t know where you get your data on mode usage from 1880 to 1910. Regardless, the frequently cited analogy between road networks and transit networks is growing tiresome, for it is an apples-to-oranges comparison. The promise of automobiles at the beginning of the 20th century was similar to that of the internet at its end. You could build roads the way you could build fiber optic cables: you knew the technology was unstoppable because it was demonstrably superior to anything that had come before. Furthermore, roads were a scalable technology – they were guaranteed to increase in value as cars became faster. Cars were also extremeley unaffordable during the period you cite, but would become more affordable later on.
    None of these advantages apply to transit. It is not demonstrably superior to the alternatives. Most judge it to be inferior and will only use it if driving becomes costly and difficult. It can’t get any cheaper than it already is. It could become faster, but only if it is designed that way: as I pointed out in a prior comment, you can get downtown from a Green Line station on any bus that crosses I-205 faster than on the Green Line itself.
    I have not defined Portland’s tranportation projects a failure. The numbers have, along with other objective criteria such as the one just mentioned.
    I take many lessons from cities like New York and Paris, but Portland is not New York or Paris, nor will it ever be. You cannot compare networks built in the pre-automotive and pre labour union age with networks built today. There are too many issues with tunnelling, labour costs, and acquiring right of way that did not exist back then. Portland isn’t even attempting to build that kind of system, and the city doesn’t have the density to make it feasible.
    What you are proposing is that we all close our eyes to reality as evinced by the mode share numbers, ignore objective information concerning the inadequacy of Portland’s transit investments and the subsidies for driving and parking, and hope and pray that if we just let things keep going as they have been, maybe in 40 years Portland will have a great transit network and 30% mode share.
    Inadequate and unacceptable transit is a greater friend to the automotive culture than anything else. True advocacy means fierce criticism, not placid acceptance, when the thing you advocate for is falling short. I do not countenance blind faith. Illusions are dangerous things.

  21. Russell Bozian February 3, 2010 at 12:51 am #

    Pantheon, we are in surprising agreement. I don’t want to close my eyes as you claim to “objective information concerning the inadequacy of Portland’s transit investments and the subsidies for driving and parking”. We must be having a misunderstanding, because I too feel that Portland’s transit investments are completely inadequate, and I shake my head over Portland’s subsidies for driving and parking.
    And I certainly don’t want to, somehow in my constant criticism of our car culture, become the greater friend to the car culture. Nor do I or most readers hold blind faith, illusions, or ill-advised placid acceptance in high esteem.
    My question is, why do we read the tea leaves of modal shares on incomplete public transport networks, to deduce the wisdom of completing them? When nobody would seriously torture the analogous car traffic numbers on an incomplete circle freeway, to try to prove that it’s a waste of public money to finish closing the ring.
    I don’t cavil with your criticism of Portland’s transport decisions, heck maybe they are all way wrong and foolish. I more narrowly object to your invoking what I consider to be unhelpful, sad numbers for ridership shares. I am claiming that ridership shares are going to be crummy no matter what in Portland until either the gas or parking apocalypse, or the public transport network gets much less unhelpful, sad and incomplete. Hence don’t use the inevitable small ridership these days as a guide to much of anything.
    And finally am puzzled that the freakish century of cheap gas, that let Americans abandon dense, efficient cities, and wastefully move to the suburbs, over decades, gives you no confidence that expensive gas will give Americans enough wisdom to wastefully abandon the suburbs, and move back to efficient cities, over decades.

  22. Alon Levy February 3, 2010 at 2:22 am #

    You cannot compare networks built in the pre-automotive and pre labour union age with networks built today.

    Vancouver is a newer city than Portland, and Calgary is even newer. Toulouse has a pre-industrial core, but it grew slowly until the middle of the 20th century. Singapore and Hong Kong were both middle-income countries until the 1970s. Perth’s growth is almost entirely postwar (Perth may have Kotkin’s approval, but like all other Australian cities, its transit share is high by US standards).
    Some of the above cities are dense, but Calgary is less dense than Portland, and Perth has only a few neighborhoods denser than Portland as a whole. All have substantially higher transit use than Portland.

  23. EngineerScotty February 3, 2010 at 10:31 am #

    What Russel and I are saying (I think)… is that some of the criticisms you level at TriMet–are barking up the wrong tree. While they may be valid (more on that below); the main problem isn’t LRT vs bus, or that MAX runs slow downtown; the main problem is that gas is cheap, parking is cheap (and free in most places), auto registration/licensing is dirt cheap, ocngestion is tolerable, and the freeway network is comprehensive and continuously-welded. While Portland doesn’t have the ten-lane concrete monstrosities found in places like Seattle or Phoenix or Atlanta; neither is Portland like SF or Vancouver or Calgary, where many trips REQUIRE surface streets if you want to drive. Until that goes away, there’s little TriMet can do to significantly increase mode share; and the argument is over a few percentage points.
    That said, there is indeed room to criticize TriMet, Metro, and the various municipalities. Here’s a question for you–and its kind of an interesting transit Rorschach test: Which LRT system do you prefer: MAX, or Seattle’s Link system? Judging by your comments, my guess is Link–despite using LRT technology, it’s a fully grade-separated, metro-scale system; whereas MAX has signficant sections where it runs like more like a streetcar than a subway. I too wish MAX were more Link-like; but don’t expect that to happen for a while.
    Many urbanists, though, love MAX and loathe Link. Why? For some, it seems, the more intimate scale of MAX (street-level boarding, frequent stops) is of utmost importance–and the separation of Metro stations is undesirable. And unfortunately, many such urbanists don’t give a damn about efficient cross-town mobility, as they hold anyone who doesn’t spend their entire lives downtown in utter contempt.
    I don’t know how much these attitudes affected the design of MAX–the “tram-train” design of the original downtown-Gresham line made perfect sense when the line ended downtown, and only became a problem when the westside extension opened. A lot of the design decisions were no doubt made to save money–if you can run on the surface its cheaper than building a subway or el. And it seems every business downtown along the line wanted a stop within a block or two.
    The main issues with MAX. I think, can be fixed by future expansion of the system. One key planning step which is now in progress is the Portland Plan, which will address an important issue ignored by the recent High Capacity Transit System Plan–what to do about downtown. Your observations about the Green Line being slower from 92nd and Powell to downtown than the #9 bus would no longer be true if the Lloyd-Center to Pioneer Square stretch (a slow surface route with 8 stops or so) were replaced with an underground route with fewer. And, of course, MAX is presently useless for most of the southern and southwestern suburbs, which is what the next two projects (Milwaukie and Barbur) will address.
    But even then–that’s only creating the necessary conditions for increased transit share. It won’t occur until pulling the car out of the garage becomes expensive or inconvenient.

  24. EngineerScotty February 3, 2010 at 12:31 pm #

    One further thing to note, since I linked to the Portland Plan website:
    Portland’s 2030 goals for commute mode share (this is just the city, not the metro area):
    * Drive alone–25% (in 2008, was 65%)
    * Telecommute 2.5% (not included in 2008 data)
    * Transit, including park-and-ride–25% (in 2008, 15%)
    * Bike: 30% (in 2008, 8%)
    * Carpool: 10% (in 2008, 8%)
    * Walk: 7.5% (in 2008, 4%)
    While the goal is to increase transit’s share by 10%, there is a larger goal to increase people-powered mobility (walking and biking) by nearly 25%. Doing THIS not only requires additional bike and pedestrian infrastructure (which Portland plans to spend money on), but better land use–only the most dedicated cyclist will pedal from a house in Lents to a job downtown.
    Interesting stuff.

  25. Russell Bozian February 3, 2010 at 10:23 pm #

    It is a privilege to discuss things with all of you, writers knowledgeable and concerned about public transport.
    I do want to take issue with Pantheon’s statement that “transit. It is not demonstrably superior to the alternatives. Most judge it to be inferior and will only use it if driving becomes costly and difficult”. Without implying that Pantheon is wrong about most things he says, I don’t share his difficulty in seeing many things about public transport that are demonstrably superior to private cars. It may even be true that the average American can’t see what is demonstrably superior about public transport, but that would be far more proof of the need for our body politic to wake up and look around at the world, than it would be evidence that public transport networks aren’t just better for business already.
    Let’s start with the 65 million Americans over the age of 12 that don’t have a driver’s license, making public transport a demonstrably cheaper mobility aid than cars and taxis for at least some cohort. We could also nod our head towards the 45,000-odd dead bodies, and hundreds of thousands of physically ruined lives, that our automobile culture is demonstrably more efficient at producing than public transport. The 60+ billion hours of unpaid chauffeur-work at least 125 million people volunteer each year to keep the car network humming is an inferiority of that network’s nature so enormous as to be invisible, but is nonetheless demonstrable.
    But the biggest-ticket demonstration of a completed public transport network’s superiority is in the bald fact that Barcelonans pay about $100 dollars per year, per person, to foreign countries for their essential gasoline. Car-centric Atlantans will pay $4,000 dollars this year for their gas at $2.50 per gallon. The oil-shock-proof aspect of Barcelona’s business climate, the fact that their lifestyle would barely notice a $4.00 uptick in gas prices, is demonstrably superior to the fragility of Atlanta’s economy, which is ready to go haywire should their gas costs rise by $10,000 yearly dollars per resident.
    In this case I don’t have to define the car culture as a failure at producing a clean, world-class cost competitive, anxiety-free lifestyle, and will say little more on the matter here. Because the numbers have defined the failure of cars for us.

  26. dejv February 4, 2010 at 1:10 am #

    European smaller sensitivity to gas price shocks are primarily caused by high fixed taxes that work as dampers.

  27. Pantheon February 4, 2010 at 1:53 am #

    I think you misinterpreted my statement. What I meant was that transit has no built-in speed and convenience advantage against the competition, as the automobile did against the horse and buggy. In fact for speed and convenience, transit is almost always at a disadvantage. Therefore we cannot assume that a complete transit network will garner increased mode share, the way road builders at the turn of the 20th century could assume that a complete road network would gain mode share against horses and buggys.

  28. Pantheon February 4, 2010 at 2:06 am #

    I am not criticizing TriMet in this thread. This is about a larger issue. What I am saying is, if Portland has not had a significant reduction in auto use in the last 20 years, something is wrong. Maybe there is something wrong with the transit investments, maybe it is the subsidized parking, maybe it is the freeway network. Maybe all of the above. But something is not working. And that means people should examine the Portland situation with a critical eye, not just assume the mode share numbers are meaningless because the network is incomplete.
    As for Link vs. MAX, I agree with the analysis in the link below, though personally I prefer Skytrain.

  29. Pantheon February 4, 2010 at 2:27 am #

    Sorry for the multiple comments, but there is a lot to respond to here. Russell, I am glad we generally agree. I am not using these numbers to argue the network should not be completed. I am saying we need to take a hard look at everything Portland has done, look at what has worked but be honest about what hasn’t worked so well.
    Regarding the effect of high oil prices on future land use patterns – my answer is I don’t know. Maybe someone will invent a car that runs entirely on some other substance. It would be unwise to take for granted that high oil prices will kill the car culture.
    Alon, thank you for the research. Those cities make much better models for Portland, and clearly show us that a city Portland’s size can and should do much better than it is doing.
    Scotty, those 2030 projections are just bulls**t PR. I guarantee the people who made them up don’t even believe them. I am more interested in their plans for the next 20 years than in their silly made up projections.

  30. Russell Bozian February 4, 2010 at 7:05 am #

    Pantheon, Alon, Scotty, Danny, Ted, Kyle, DejV: Where were you and Jarret when I needed you 5 years ago, writing a funding request for researching the next big direction in public transport technology? (Sorry Anonymouse, it was a human-only shop.)
    Pantheon it gives me no great pleasure to note that I overstated the cost of gasoline to Atlantans by 100%–800 gallons times $2.50 is $2,000 per year, not $4,000. That kind of mistake goes some way towards explaining why Jarrett’s job is probably safe until about 2030.
    Alon you do make global comparisons look really easy, perhaps you have access to some secret transit/time machine version of Wikipedia. Maybe it’s in Esperanto? What’s the URL?
    So, Pantheon, your concern is that my original complaint, about Randal O’Toole-type misuse of mode share statistics to blame the victim, to fleece Americans into thinking that public transport is pointless, has its own potential to help shield bad transport projects from critical thinking and review.
    Hmm, the best response and conclusion would be to upgrade my rants in some way to reflect that:
    –we should embrace and use mode share change histories to help divine the most rewarding way to expand public transport networks,
    –but keep in mind that absolute mode share comparisons with the (2.5 trillion per year + 60 billion volunteer hour) car industry do not prove much of anything, about the economic wisdom of trying to cling to the car-centered life until the last oil company lobbyist has retired.

  31. EngineerScotty February 4, 2010 at 7:41 am #

    I’m not endorsing the goals (they are “goals”, not projections, if that makes a difference) as realistically attainable, just stating what they are. My suspicion is that without a major oil price shock, they’re probably not attainable–the City of Portland is probably on board with those numbers, but many of the surrounding suburbs are all about car culture. A lot depends on what happens with the next proposed UGB expansion.
    A lot also depends on future demographic trends; there’s an assumption that recent population growth trends will continue. Of course, if 10% unemployment continues, all bets are off.
    One thing that has hurt Portland is a relatively small industrial and capital base for a city its size. Only 2 Fortune 500 corporations are presently headquartered in the metro area (Nike and Precision Castparts); we used to have more, but quite a few were acquired in the great wave of consolidation that has occured over the past few decades. Other than Nike CEO Phil Knight, there aren’t any billionaires in town–and Knight’s philanthropy is mostly directed towards Eugene and the U of O, not to anything in the metro area. (We have to borrow Paul Allen from Seattle). Many different people will tell you many different reasons–those on the right will complain about taxes and regulations; those on the left will complain about inadequate financing of education and infrastructure. At any rate, city government certainly has some ambitious plans, but how to pay for them is always a nagging question.

  32. Alon Levy February 5, 2010 at 7:05 am #

    Russell: a lot of the stuff I say comes from Wikipedia. The part about Calgary is a case study presented in Wikipedia’s article on light rail; the article mentions Calgary as an extreme case of low costs, and I followed the links to Calgary Transit’s explanation for why it works so well. In some Asian cities, there’s a good Wikipedia “Transportation in City_Name” article.
    The other examples I bring up are really just Googling, or occasionally looking at government statistics websites. I wanted to see how successful transit was in Canada so I looked up Canada’s travel to work statistics. The Australian data I literally got from Googling variants on “City_Name public transport mode share.”

  33. Mike February 16, 2011 at 1:03 pm #

    The issues I have heard about Portland’s MAX is that it is slow.
    It takes something like an 45-60 minutes to go from Gresham to downtown Portland. By car, the trip takes 20 minutes.
    I have also heard that the % of trips taken on transit actually declined on the east side, after MAX opened, because the popular express buses that offered a speedy trip were cancelled.I don’t know how true this is. But I have heard that comment a number of times.
    There is no doubt Portland does great in the American context.

  34. EngineerScotty February 16, 2011 at 2:42 pm #

    It takes something like an 45-60 minutes to go from Gresham to downtown Portland.
    That part, you’ve heard correctly. Actual travel time between the Cleveland Avenue station, at the end of the line in Gresham, to Pioneer Square (the “center” of downtown) is about 50 minutes–here is a Blue Line timetable for your perusal.
    By car, the trip takes 20 minutes.
    Only if you make the trip in the middle of the night, and are willing to outrun the cops. 🙂 In normal, non-rush daytime traffic, it takes about 26-27 minutes (route here), and in rush hour the train is easily competitive.
    I have also heard that the % of trips taken on transit actually declined on the east side, after MAX opened, because the popular express buses that offered a speedy trip were cancelled.
    Whatever business was lost on the express trips has long since been made up. Since MAX opened in 1986, overall system ridership has more than doubled. Commute mode share hasn’t seen the same growth–part of that is an increasing population, part of that is a significant number of trips being used for non-commute purposes, and part of that is significant telecommuting and use of “active transportation” (biking or walking) to get to work.
    But other than a recent downturn correlated to the recession, ridership has been on a steady ascent.