In the last few years New York’s MTA has gone through multiple cycles in which a new head talks of far-reaching reform, while only small incremental steps are taken. The latest is the MTA Transportation Reinvention Commission, which has just released a report detailing all the way the MTA could move forward. Capital New York has covered it and hosts the report in three parts. Despite the florid rhetoric of reinvention, the proposals contained in the report are small-scale, such as reducing waste heat in the tunnels and at the stations on PDF-pp. 43-44 of the first part. At first glance they seem interesting; they are also very far from the reinvention the MTA both needs and claims to be engaging in.
Construction costs are not addressed in the report. On PDF-p. 53 of the first part, it talks about the far-reaching suburban Grand Paris Express project for providing suburb-to-suburb rapid transit. It says nothing of the fact that this 200-km project is scheduled to cost about 27 billion euros in what appears to be today’s money, which is not much more than $150 million per km, about a tenth as much as New York’s subway construction. (Grand Paris Express is either mostly or fully underground, I am not sure.) The worst problem for transit in the New York area is that its construction costs are an order of magnitude too high, but this is not addressed in the report.
Instead of tackling this question, the report prefers to dwell on how to raise money. As is increasingly common in American cities, it proposes creative funding streams, on the last page of the first part and the first six pages of the second part: congestion pricing, cap-and-trade, parking fees, a development fund, value capture. With the exception of congestion pricing, an externality tax for which it makes sense for revenues to go to mitigation of congestion via alternative transportation, all of these suffer from the same problem: they are opaque and narrowly targeted, which turns them into slush funds for power brokers. It’s the same problem as the use of cap-and-trade in California.
One of the most fundamental inventions of modern government is the broad-based tax, on income or consumption. Premodern governments funded themselves out of tariffs and dedicated taxes on specific activities (as do third-world governments today), and this created a lot of economic distortion, since not all activities were equally taxed, and politically powerful actors could influence the system to not tax them. The transparent broad-based tax, deeded to general revenue through a democratic process, has to be spent efficiently, because there are many government departments that are looking for more money and have to argue why they should get it. Moreover, the tax affects nearly all voters, so that cutting the tax is another option the spending programs must compete with. The dedicated fund does neither. If the broad-based tax is the equivalent of market competition, a system of dedicated funds for various government programs is the equivalent of a cartel that divides the market into zones, with each cartel member enjoying a local monopoly. In this way there’s a difference between the hodgepodge of taxes the MTA levies and wants to levy and Ile-de-France’s dedicated 1.4-2.6% payroll tax: the payroll tax directly affects all Francilien workers and employers, and were it wasted, a right-wing liberal politician could win accolades by proposing to cut it, the way New York Republicans are attacking the smaller payroll tax used to fund the MTA.
The proposals of where to spend the money to be raised so opaquely are problematic as well. There is a set of reforms, based on best practices in Continental Europe and Japan, that every urban transit system in the first world should pursue, including in their original countries, where often only some of those aspects happen. These include proof-of-payment fare collection on buses, commuter trains, and all but the busiest subway systems; all-door boarding on buses; mode-neutral fares with free transfers; signal priority and bus lanes on all major bus routes, with physically separated lanes in the most congested parts; a coherent frequent bus network, and high off-peak frequency on all trains; and through-service on commuter rail lines that can be joined to create a coherent S-Bahn or RER system. As far as I can tell, the report ignores all of these, with the exception of the vague sentence, “outfitting local bus routes with SBS features,” which features are unspecified. Instead, new buzzwords like resiliency and redundancy appear throughout the report. Redundancy in particular is a substitute for reliability: the world’s busiest train lines are generally not redundant: if they have parallel alternatives those are relief lines or slower options, and a shutdown would result in a major disruption. Amtrak, too, looks for redundancy, even as the busiest intercity rail line in the world, the Tokaido Shinkansen, has no redundancy, and is only about to get some in the next few decades as JR Central builds the Chuo Shinkansen for relief and for higher speeds.
The only foreigners on the Commission are British, Canadian, and Colombian, which may have something to do with the indifference to best industry practices. Bogota is famous for its BRT system, leveraging its wide roads and low labor costs, and Canada and to a lesser extent the UK have the same problems as the US in terms of best industry practices. Swiss, French, German, Japanese, Spanish, and Korean members might have known better, and might also have been useful in understanding where exactly the cost problems of the US in general and New York in particular come from.
The final major problem with the report, in addition to the indifference to cost, the proposal for reactionary funding sources, and the ignorance of best industry practices, is the continued emphasis on a state of good repair. While a logical goal in the 1980s and 90s, when the MTA was coming off of decades of deferred maintenance, the continued pursuit of the maintenance backlog today raises questions of whether maintenance has been deferred more recently, and whether it is still deferred. More oversight of the MTA is needed, for which the best idea I can think of is changing the cycles of maintenance capital funding from five years, like the rest of the capital plan, to one year. Long-term investment should still be funded over the long term, but maintenance should be funded more regularly, and the backlog should be clarified each year, so that the public can see how each year the backlog is steadily filled while normal replacement continues. This makes it more difficult for MTA chiefs to propose a bold program, fund it by skimping on maintenance, and leave for their next job before the ruse is discovered.
I tag this post under both good categories (“good transit” and “good/interesting studies”) and bad ones (“incompetence” and “shoddy studies”) because there are a lot of good ideas in the report. But none of them rises to the level of reinvention, and even collectively, they represent incremental improvement, of the sort I’d expect of a city with a vigorous capital investment program and industry practices near the world’s cutting edge. New York has neither, and right now it needs to imitate the best performers first.
The Regional Plan Association has a new study warning that Metro-North’s infrastructure is falling apart, and demands $3.6 billion in immediate spending on state of good repair. In general, my line on deferred maintenance is “you mean the agency deferred maintenance all those years and didn’t tell us?”. But in this case, despite the language, most of the proposed spending is improvements, namely rehabilitation or replacement of old movable bridges with low speed limits, rather than ongoing maintenance folded into long-term capital spending.
$2.8 billion of the proposed program is for replacing five bridges: Pelham Bay, Cos Cob (over the Mianus), Walk (over the Norwalk River), Saga (over the Saugatuck), and Devon (over the Housatonic). I believe all five should be replaced in the medium term, but the cost proposed is much higher than it should be. $560 million per bridge is quite high, and out of line with Amtrak found on PDF-pp. 29 and 56 of the Northeast Corridor Master Plan. Amtrak cites the cost of replacing the Pelham Bay Bridge alone at $100 million, and the cost of both replacing it and modifying curves on the Hell Gate Line at $500 million. It cites the cost of replacing both the Saga and Walk Bridges at $600 million.
Now, the RPA lists Saga as the easiest bridge to replace since it’s two two-track bridges, so work can be done one bridge at a time with less disruption to ongoing service, but conversely Pelham Bay is also quite cheap according to Amtrak.
But there’s a more serious problem, which is the avoidance of talking about service plans for commuter and intercity rail. If there is serious effort at adding Metro-North service to Penn Station or at raising intercity rail speeds, then the worst speed and capacity restrictions should get priority, and the infrastructure construction should be based on what promotes the desired service plans. It is very expensive and probably cost-ineffective to six-track everything from New Rochelle to Stamford, to allow three speed regimes: local, express, and intercity. I have argued before that it’s better to leave it at four tracks and bypass bad curves, around Port Chester, and make this the six-track segment. This is of course independent of maintenance issues, but suggests which bridge replacements are necessary to support these bypasses (Cos Cob) and which aren’t (the rest are less critical, especially Walk, which intercity trains should bypass on a straighter I-95 segment).
Likewise, there’s a capacity crunch west of Stamford but not one east of Stamford, and this again suggests Cos Cob as the most important priority. Finally, the slowest segment of the NEC away from immediate station areas is the western corner of Connecticut, from the state line to Stamford; Stamford’s curves are mild, while those heading out of Port Chester all the way across the Mianus are quite bad, and straightening the segment would also require straightening the bridge, which can be done easily if it’s replaced. Despite all this, the RPA and Amtrak are saying Cos Cob needs rehabilitation and not replacement, which misses opportunities to both improve reliability and speed up a slow segment.
Moreover, there is no mention of grade-separating Shell Interlocking, just south of New Rochelle. While not a state of good repair issue even in theory, the interlocking’s tight curves impose a limit of either 30 or 45 mph (so, 50-70 km/h), depending on source, in an area that could otherwise support 200 km/h or more. It is very difficult to straighten New Rochelle to sufficient curve radius for that, but 150 requires only minor takings. This may be necessary, independent of speed issues, to raise capacity enough to allow Metro-North service to both Grand Central and Penn Station. It’s possible to schedule trains through the flat junction, but this imposes an additional constraint on the schedule, on top of track-sharing with Amtrak and, in the East River Tunnels, the LIRR.
The relative costs of different technologies of transit are not fixed. Although there are some rules of thumb for the ratio of tunneling cost to above-ground transit cost, the actual ratio depends on the city and project, and this would favor the mode that’s relatively cheaper. Likewise, the ratio of operating to capital costs is not always fixed, and of course long-term real interest rates vary between countries, and this could again favor some modes: more expensive construction and cheaper operations favor buses, the opposite situations favor rail.
In general, els cost 2-2.5 times as much as at-grade light rail, subways 4-6 times as much, according to Table 6 in this Flyvbjerg paper; Table 5, sourced to a different previous paper, estimates per-km costs, and has ratios of 1.8 and 4.5 respectively.
However, specifically in Vancouver, the premiums of elevated and underground construction appear much lower. The cost estimates for rail transit to UBC are $2.9 billion for an almost entirely underground extension of SkyTrain and $1.1 billion for at-grade light rail along Broadway, both about 12 km. Elevated construction is in the middle, though closer to the light rail end: the estimates for the two all-elevated SkyTrain extension alternatives into Surrey are $900 million for 6 km for rapid transit alternative 3 and $1.95 billion for 15.5 km for alternative 1. The under-construction Evergreen Line, which is 11 km long of which about 2 are in tunnel, is $1.4 billion.
In the rest of Canada, this seems to be true as well, though the evidence is more equivocal since the projects that are considered above-ground are often elevated rather than at-grade. The Canadian above-ground projects that Rob Ford’s Eglinton subway is compared with are not wholly above ground. Calgary’s West LRT, which with the latest cost overrun is $1.4 billion (a multiple of the preexisting three-line system) for 8 km, includes a 1.5 km tunnel, a short trench, and some elevated segments. Edmonton’s North LRT is $750 million for 3.3 km, of which about 1 km is in tunnel and the rest at-grade. But while it’s hard to find the exact ratio because of those mixed projects, the costs are not consistent with the ratios found in Flyvbjerg’s sources.
Outside Canada, those ratios seem to hold up better. American above-ground transit projects, such as the Portland Milwaukie extension and the Washington Silver Line, are as expensive as Calgary and Edmonton’s light rail, but American subways are much more expensive than Toronto’s Eglinton subway ($325 million/km, 77% underground and the rest elevated): Manhattan tunneling is more difficult, so its $1.3-1.7 billion/km cost may not be representative, but conversely, BART to San Jose’s $4 billion for about 8 km of tunnel is for tunneling partially under a wide railroad right-of-way, with no crossings of older subway infrastructure as is the case for Eglinton at Yonge.
Conversely, French tunneling costs are comparable to or lower than Canadian ones, but at-grade light rail is far less expensive than in North America. The RER E extension was at least as of 2009 budgeted at €1.58-2.18 billion for 8 km of tunnel (see PDF-page 79 here; this excludes €620 million in improvements to the existing commuter lines the tunnel will be linked with) – somewhere between the per-km costs of Vancouver and Toronto subways, but in a much denser environment with more infrastructure to cross. But the cost range for Parisian trams is much lower, about €30-50 million per km, in line with the subway:tram cost ratio of 4-6; the cost range in other French cities tends to be a little lower.
What this means is that in Canada in general, and in Vancouver in particular, questions about what mode to build should have higher-end answers than elsewhere. It doesn’t mean that the Eglinton subway is justified, but it does bias suburban rail lines in Vancouver toward elevated SkyTrain extensions rather than light rail, and inner extensions toward SkyTrain subways. For the same cost of building a subway under Broadway, Translink couldn’t build too much additional light rail; it could build two lines, say on Broadway and 41st, or maybe three if both non-Broadway routes are short, but certainly nothing like the entire network that SkyTrain opponents believe is the alternative, citing European tramway construction costs.
In response to my takedown of Reason, specifically my puzzlement at the estimates of inaccuracy in traffic forecasts, alert reader Morten Skou Nicolaisen sent me several papers on the subject. While there is past research about traffic shortfalls, for example this paper by Flyvbjerg (hosted on a site opposing the Honolulu rapid transit project), Flyvbjerg’s references are papers from twenty years ago, describing mostly subway projects in developing countries, but also rapid transit and light rail projects in the US built in the 1970s and 80s. Unlike Flyvbjerg, who posits that planners are lying, the authors of the papers he references have other theories: currency exchange rate swings, the challenges of underground construction, inaccurate forecasts of future economic growth, outdated traffic models based on postwar road traffic models. See section 6 of Walmsley and Pickett, and sections 3.3 and 4.2 of Fouracre, Allport, and Thomson (see also the range of costs for underground construction in developing countries in section 3.3).
The question is then whether things have improved since 1990. Since the first study to point out to cost overruns and ridership shortfalls in the US was by Pickrell, the question is whether post-Pickrell lines have the same problems, or whether there are better outcomes now, called a Pickrell effect.
The answer, as far as ridership is concerned, is very clearly that ridership shortfalls are no longer a major problem. See recent analysis by Hardy, Doh, Yuan, Zhou, and Button; see specifically figure 1. Cost overruns also seem to be in decline and are no longer big, although a multiple regression analysis finds no Pickrell effect for cost, just for ridership.
In particular, there is no comparison between projects from 30 years ago, most of which are underground, and present-day developed-world high-speed and urban rail lines.
When I went to an IRUM meeting nearly two years ago, the participants crowed about the possibility of restoring rail service on the Rockaway Cutoff. New York urban planner and technical activist David Krulewitch recently posted his proposal in a comment, showcasing multiple ways of reusing it for faster connections between Midtown and the Rockaway branches of the A, serving JFK and/or the Rockaways. Although the possibility has raised excitement among most local transit activists (some of whom have posted fantasy maps in the various subway forums including such an extension), I’m more skeptical.
First, the potential for JFK service is limited. The reason is that the Rockaway Cutoff only reaches Howard Beach, making it just a faster version of the A. The AirTrain is technologically incompatible with any other transit system in the region: it’s a vendor-locked Bombardier technology, of the same type used on the first two SkyTrain lines in Vancouver, in which the trains are driverless and propelled by linear induction motors placed between the tracks. This system allows trains to climb steeper than usual grades, and the maximum grade used on the AirTrain is 5.5%, considerably more than the usual for a normal subway or regional EMU (though less than the absolute maximum).
In addition, the needs of the mainline regional system and the subway are different from those of an airport people mover. A people mover needs very high frequency at all times, which is why such systems are normally driverless. In contrast, most subways are not driverless, and I do not know of a single mainline railroad that is driverless. Driverless operation requires some serious upgrades to electronics, and those upgrades are pointless if used only on a single line. If instead the JFK connection has a driver, then frequency will necessarily be very low, since there isn’t too much airport demand, and this will depress demand even further.
Although the current AirTrain system suffers from the lack of a one-seat ride to Manhattan, the situation is not too bad. Jamaica offers a very frequent LIRR connection to Manhattan at all hours, and Howard Beach offers a frequent if not fast connection to Brooklyn. This requires multiple transfers to reach most destinations, but this is not a major problem for locals who are traveling light. It’s a bigger problem for locals with luggage and even more so for tourists, but a one-seat ride to Penn Station, as proposed in LIRR connection proposals, is not too useful since most hotels are too far north. Even Grand Central is at the southern margin of Midtown proper.
For an honest estimate of how much demand there is, let us look at airports with very good transit connections. At Charles de Gaulle, 6 million passengers board at the RER station per year, 20% of airport traffic, and another 3 million use the TGV. At Frankfurt, 11% of passengers use the S-Bahn, and 15% use the ICE. Neither airport has a subway connection. Heathrow, which does have an Underground connection, has a total of 13 million Underground boardings and alightings, 20% of traffic (see data here); I do not know the ridership of the two mainline rail connections to Central London, but a thesis studying air-rail links puts the mode share as of 2004 at 9%. Assuming the train usage in Paris, New York could expect JFK to see 4.6 million boardings, or 9.2 million boardings and alightings; assuming that in London, New York could expect 13 million. The AirTrain’s current ridership is 5.3 million. Although the extra ridership would be useful at low cost, the higher cost of allowing mainline or subway trains to use the AirTrain tracks may be too high.
More importantly, 13 million passengers a year – an upper bound more than a median estimate in light of Frankfurt and Paris’s lower ridership – do not make for very high frequency by themselves, and therefore JFK could at best be an anchor rather than the primary ridership driver. Airport-only trains would be quite lonely; one of Krulewitch’s proposal’s most positive aspects is that it never even mentions premium express services such as Heathrow Express, which tend to underperform expectations as passengers prefer to ride cheaper local trains. Thus, not only would it be expensive to do an infrastructure and technology retrofit to permit direct Midtown-JFK service, but also the market for it would not be very large.
This brings us to the second possible market: the intermediate stops on the Rockaway Cutoff. They may seem useful, but in fact the development is elsewhere. Observe the land use maps of Queens Community Boards 6 and 9, which host most of the Cutoff: along the Cutoff’s right of way, the primary uses are single-family residential, with only a little commercial. Moreover, the commercial development is often very auto-oriented, for example at Metropolitan Avenue. Indeed, the only proposed station with significant dense development is Rego Park, which is on the LIRR Main Line and could be restored without restoring an entire line. Rezoning near the other stations is possible, but why not rezone near existing subway stations first?
In general, development in the US along linear corridors follows arterial roads, not railroads that haven’t seen passenger service in many decades. In the area in question, the primary north-south commercial artery is Woodhaven Boulevard; for service to the intermediate areas, the proposal should be evaluated against a light rail line on Woodhaven, providing local service from Queens Boulevard to Howard Beach and hitting multiple subway transfer points but not the airport.
The third market posited, fast service to the Rockaways, is the weakest. The stations in the Rockaways are some of the least busy in the subway system, with only a few hundreds of thousands of annual boardings each. They only support 15-minute service, with half of the A trains terminating at Ozone Park; since there are two Rockaway branches, the less busy only gets a shuttle except at rush hour, when there is enough demand for a few direct trains. Even with 15-minute service, it’s expensive to serve an area so far away with a flat fare; until a series of fare unifications, the subway charged a higher fare to stations in the Rockaways.
The problem with the Rockaways is that stations are too far from Manhattan and too lightly populated for it to be otherwise. Moreover, service along the LIRR to Penn Station using the Cutoff is about 18 km long measured from the intersection of the Cutoff with the A at Liberty, and service along the R is 16 km to Lexington and 19.5 to Times Square; service along the A is about 21.5 km long to Penn Station and 22 km long to Times Square, longer but not very much so. The main advantage of the R is that it hits Midtown proper better, rather than skirting it on 8th Avenue, but there’s practically no speed advantage – about 6 kilometers of travel distance and 2 station stops, translating to perhaps ten minutes.
As appealing as sending a single local subway service from the Queens Boulevard Line along the Cutoff to serve the Rockaways and give direct service to every branch, there would be a large demand mismatch; moreover, service to Forest Hills, which has nearly twice as many riders as all Rockaway stations combined, would be degraded.
LIRR service to the Rockaways could be better, but only if it’s modernized. The way it’s run today – infrequently, not very quickly, and expensively – it has no appeal. Far Rockaway has 4,500 weekday boardings on the subway (with a travel time of 1:06-1:14 to Times Square), and 158 average of boardings and alightings on the LIRR (with a travel time of 0:50-1:00 to Penn Station). Cutting another ten minutes from the LIRR travel time to Far Rockaway isn’t going to change anything as long as operating patterns remain as they are.
But if operating patterns are modernized, is there a point in service along the Cutoff? It saves very little distance measured to Far Rockaway: 21 vs. 24.5 km. It’s more useful farther west in the Rockaways, but those are less useful areas to serve – those are the areas with the lowest subway ridership, whereas Far Rockaway’s ridership is merely below average. Although the ridership would not be as pitiful if LIRR charged subway fares for in-city service and provided reasonably high speed and frequency, and it could be studied further as a case of an in-city S-Bahn line, there are more worthwhile S-Bahn destinations on the LIRR, for examples southeastern Queens, Hempstead, Bayside, and Great Neck. The main problem is that the Rockaway Beach branch would still have too little ridership to justify high frequency, and the round-robin proposal would have the same frequency-splitting effect on the stations except Far Rockaway and its immediate vicinity as running two separate branches; each station may have frequent service, but half the trains would take too long.
Finally, the three above-described markets – JFK, neighborhoods between Rego Park and Howard Beach, and the Rockaways – cannot all be served at the same time. The intermediate neighborhoods are free, but it’s impossible to serve both JFK and the Rockaways without an additional branching, reducing frequency even further. This means that the two markets can’t be combined to create more powerful demand. It’s one or the other – either the 13 million boardings and alightings one could optimistically expect of JFK, or the 4.5 million boardings times an appropriate growth factor one could expect of the Rockaways. Neither is high by S-Bahn standards; measured in ridership per terminus excluding short-turns, the least busy RER line, the RER C, has 20 million riders per terminus.
Because of the low potential ridership of the Rockaway Cutoff, I suggest New York transit advocates look elsewhere first. Service to JFK could be beefed up with sending surplus Amtrak trains to Jamaica for an interchange, and service to the Rockaways first with modernizing regional rail and second with having it take over the Far Rockaway branch of the A if there’s demand. If there’s higher than expected growth in demand, then the Cutoff could be activated, at as a low a cost in 15 years as today. But for now, the low cost of activating the Rockaway Cutoff comes hand in hand with low benefits.
My previous post‘s invocation of Reinhard Clever’s lit review of transfer penalties was roundly criticized on Skyscraper
City Page for failing to take into account special factors of the case study. Some of the criticism is just plain mad (people don’t transfer from the Erie Lines to the NEC because trains don’t terminate at Secaucus the way they do at Jamaica?), but some is interesting:
This is what the paper says:
Go Transit commuter rail in Toronto provides a good example for Hutchinson’s findings. In spite of being directly connected to one of the most efficient subway systems in North America, Go’s ridership potential is limited to the number of work locations within an approximately 700 m radius around the main railroad station. Most of the literature points to the fact that the ridership already drops off dramatically beyond 400 m. This phenomenon is generally referred to as the “Quarter Mile Rule.”
Let’s look at WHY that is. If you live North of downtown and work North of about Dundas Street, it is probably faster for you to take the subway to work. So people aren’t avoid the commuter train because it imposes a transfer, but just because the subway is faster. Same thing if you live along the Bloor-Danforth line. Toronto’s subway runs at about the same average speed as NYC’s express trains. If one lives east or west of the city along the lakeshore, they are going to take the GO Train to Union Station and transfer to the subway to reach areas north of Dundas. I really doubt these people are actually “avoiding” the GO Train, though if there is evidence to the contrary I’d like to see it.
Toronto also has higher subway fares than NYC.
The issue is whether the subway and commuter rail in Toronto are substitutes for each other. My instinct is to say no: on each GO Transit line, only the first 1-3 stations out of Union Station are in the same general area served by the subway, and those are usually at the outer end of the subway, giving GO an advantage on time. Although the Toronto subway is fast for the station spacing, it’s only on a par with the slower express trains in New York; on the TTC trip planner the average speed on both main subway lines is about 32 km/h at rush hour and 35 km/h at night.
Unfortunately I don’t know about GO Transit usage beyond that. My attempt to look for ridership by station only yielded ridership by line, which doesn’t say much about where those riders are coming from, much less potential riders allegedly deterred by the transfer at Union Station. So I yield the floor to Torontonians who wish to chime in.
Update: a kind reader sent me internal numbers. The busiest stations other than Union Station are the suburban stations on the Lakeshore lines, led by Oakville, Clarkson, and Pickering; the stations within Toronto, especially subway-competitive ones such as Kipling, Oriole, and Kennedy, are among the least busy. Some explanations: the subway is cheaper, and (much) more frequent; Toronto’s GO stations have no bus service substituting rail service in the off-peak, whereas the suburban stations do; Toronto’s stations have little parking.
As a followup to my previous post about the TTI’s new congestion report, I finally did a multivariate regression analysis, with the dependent variable being cost and the independent variables being size and freeway lane-miles per capita. Such an analysis reduces the regression coefficient between freeways and congestion even more, to -42.5 from the uncontrolled -233. More interestingly, if we log all numbers (population, congestion cost, and freeways), the regression coefficient becomes a positive 0.02 – that is, adding freeways is correlated with making congestion a little worse.
Of course, it’s not literally true that adding freeways makes congestion worse. There’s a correlation if we look at the variables in some way, but it’s not going to have any statistical significance. Therefore tweaking variables slightly can make a correlation go from weakly positive to weakly negative.
In univariate regression, we can think about the square of the correlation as the percentage of the variance that is explained by the regression line. Freeway lane-miles per capita explain 3.8% of the variance in congestion (and logging either variable makes this number smaller); with 101 urban areas surveyed, it’s statistically significant, but barely so. But after controlling for population, this proportion drops to 0.7%. Thus, any sentence of the form “adding one freeway lane-mile per thousand people only cuts $42.5 from the annual congestion cost per capita” is inherently misleading: the correlation is so weak that some cities can reduce congestion without building the requisite amount of roads, or building any roads at all (for example, nearly all American cities in the last five years, congestion having crashed in the oil price boom and the recession), while others can keep building but see congestion increase (for example, Houston since the 1980s, and even today).
It goes without saying that such analysis is not going to appear in the TTI report itself. The TTI gets funding from APTA and the American Road and Transportation Builders Association. It pays lip service to congestion pricing as a solution to congestion, and instead talks a lot about building public transportation and even more about building freeways to keep up with demand. American cities may be building freeways faster than their population growth, but cities that enact no traffic restraint and just pour concrete can expect demand to grow faster than population as people become more hypermobile.
The Texas Transportation Institute has just released the latest version of its much-criticized Urban Mobility Report. Although the conclusions and recommendations made by the TTI tend to reflect its funding sources (APTA, American Road and Transportation Builders Association), the underlying data seems sound, and suggests conclusions orthogonal to those made by the report. In addition, looking at the correlations more closely suggests obvious hazards coming from any simplistic analysis of linear regression. It even showcases how we could use data dishonestly and lie with statistics. So let’s take the data that’s relevant right now and see what we can conclude ourselves.
First, the size of an urban area is a very strong correlate of its level of congestion. The linear correlation between size and per capita congestion cost is 0.71. The correlation increases to 0.8 if we take the log of population and the log of congestion, or if we consider congestion in the absence of public transportation; in both cases, it comes from the fact that New York is far below the population-congestion regression line.
Now, more freeways do not really lead to congestion reduction. There’s some correlation between freeway miles per capita and congestion per capita, going in the expected direction, but it’s weak, -0.2, and while it’s statistically significant, the p-value is an uninspiring one-tailed 0.025. Looking at a scattergram doesn’t make any nonlinear relationship obvious.
Moreover, size is a correlate of both congestion (0.71 as above) and freeways (-0.23). This is fully expected: literature on cities’ economies of scale (here is a story of one controversial example) suggests that congestion and the economic activity causing it grow faster than linearly in city size while the amount of required energy and infrastructure grows slower than linearly. I open the floor to anyone with more powerful tools than OpenOffice Calc to do multiple regression; again, the sanitized data is here.
Even without controlling for population, freeways are not a very strong correlate. The regression coefficient is -233: increasing the number of freeway miles per thousand people by 1 (the range is 0.13-1.4, with few large metros above 1 or below 0.35) reduces the congestion cost per capita by $233 per year, also uninspiring.
The regression number alone can be used as a dishonest trick when arguing on the Internet. If we overinterpret weak correlations, we can declare that the only way to decrease congestion is to build an unrealistic number of freeways, and thus declare the problem unsolvable. Of course, for most cities we can find other cities of comparable size with much less congestion and without enormous amounts of asphalt – this is why the correlation is so weak. But a good hack should not bother himself with such caveats to talking points.
So if making an urban area larger makes it more congested, independently of and much more strongly than all else, should we give up on cities? Well, no. Assuming no change in traffic policy, congestion results from more economic activity. It then becomes straightforward to institute congestion pricing. It’s no different from how big cities can use their resources to hire more cops to deal with the crime that could result from extra interactions between people. On top of this, in very large cities, mass transit becomes a serious option: this not only reduces the amount of congestion per capita, but also removes many people from the highways to the point that congestion becomes irrelevant to their daily lives, except perhaps through higher transportation prices, which they can fully afford given the extra wealth.
Another thing to consider is that most American cities have added more freeways than people since 1982, the first year for which TTI data is available, while also becoming much more congested. If a simple relationship between freeway miles per capita and congestion held, it would be robust to these changes over time. Of course, traffic has grown even faster, leading the main report to showcase on PDF-page 21 how congestion increased the fastest in regions where road demand outgrew supply the most. But this raises the question of whether the main issue is one of demand, rather than one of supply. This is not just an issue of size: the log-log regression coefficients with cost and time is 0.42, i.e. doubling an urban area’s population will raise its per-driver congestion cost and travel delay by a factor of 2^0.42; since 1982, the average urban area on the list has seen its population grow by a factor of 1.46 and its travel delay per driver grow by a factor of 2.85 = 1.46^2.77. Cost has grown even faster, because of higher value of time.
That said, quantity of freeways does not equal quality (from the drivers’ perspective, of course, rather than the city’s). On paper, Greater New York has added freeway lanes about 9% faster than people over the last thirty years. In practice, none has addressed the major chokepoints within and into the city itself, where traffic is worst. Of course, commutes involving Manhattan are overwhelmingly likely to be done on public transportation, but diagonal commutes within the city are more likely to be done by car than on transit.
On a parenthetical note, the units of comparison here are TTI-defined urban areas. TTI’s belief about urban area population growth trends is sometimes at odds with that of the Census Bureau, but the raw population numbers are close enough. More important is the question of what to do about urban areas that are really exurbs of larger areas, such as Poughkeepsie-Newburgh and the Inland Empire. My first instinct was to lump them in with their core metro areas, but their congestion level per capita is not high. Their commutes are long, but not very congested for their size. Finally, although most correlations here are with congestion cost, the correlation numbers with travel delay and excess fuel consumptions are very similar; the one exception I’ve checked, for which I have no explanation, is log-log congestion-fuel correlation (0.84, with regression coefficient 0.73).
The political transit bloggers are talking about the new RPA/America 2050 report on high-speed rail published by the Lincoln Institute, which recommends a focus on the Northeast and California. Unfortunately, this is not an accurate description of the report. Although it does indeed propose to start with the Northeast and California, that’s not the focus of the report; instead, the focus is to argue that HSR is everything its boosters claim it is and then some more, and demand more money for HSR, from whatever source.
Look more closely at the section proposing to focus on New York and California. Although the authors say the US should prioritize, minimal argument was offered for why these are the best options. The report shows the map from the RPA’s study on the subject, which proposes a few other priorities and isn’t that good to begin with (it grades cities on connecting transit based on which modes they have, not how much they’re used). But it says nothing more; I’d have been interested to hear about metro area distribution questions as discussed on pages 113-5 in Reinhard Clever’s thesis and pp. 10-11 of his presentation on the same topic, and alignment and regional rail integration questions such as those discussed by the much superior Siemens Midwest study, but nothing like that appears in this report.
The report then pivots to the need to come up with $40 billion for California and $100 billion for the Northeast Corridor, under either the RPA’s gold-plated plan or Amtrak’s equally stupid Vision. The RPA first came up with the idea of spending multiple billions on brand new tunnels under Philadelphia, which was then copied by the Vision, and wants trains to go through Long Island to New Haven through an undersea tunnel. Clearly, cost-effectiveness is not the goal. Since the methodology of finding the best routes is based on ridership per km, offering a gold-plated plan is the equivalent of trying to connect much longer distances without a corresponding increase in ridership, which goes against the original purpose of the RPA study.
Together with the neglect of corridors that scored high on the RPA’s study but have not had official high-speed rail proposals costing in the tens of billions (the SNCF proposal and the above-mentioned Siemens report are neither official nor affiliated with the RPA), the conclusion is not favorable. The most charitable explanation is that the RPA was looking for an official vehicle to peddle its own Northeast HSR plan but actually believes it has merit. The least charitable is that the RPA wants to see spending on HSR megaprojects regardless of cost-effectiveness.
The treatment of other issues surrounding HSR is in line with a booster mentality, in which more is always better. Discussing station placement, the report talks about the development benefits that come from downtown stations and the lack of benefit coming from exurban stations, as nearly all stations on LGVs are. It does not talk about the tradeoff in costs and benefits; others have done so, for example the chief engineer of Britain’s High Speed 2, who also talks about other interesting tradeoffs such as speed versus capacity versus reliability, but the report prefers to just boost the most expensive plan.
More specifically, the report contrasts CBD stations, suburban stations, and exurban stations. In reality, many stations are outside the CBD but still in the urban core with good transit connections, such as Shin-Osaka, Lyon Part-Dieu, and 30th Street Station, but those are implicitly lumped with beet field stations. This helps make spending billions on tunnels through Philadelphia, as both the RPA and Amtrak propose, look prudent, when in reality both Japan and France are happy to avoid urban tunneling and instead build major city stations in conveniently urban neighborhoods. In fact, Japan’s own boosters and lobbyists crow about the development around such stations.
In line with either view of the report’s purpose, the literature it studies is partial. Discussing the effect of HSR on development, it quotes a study about the positive effect of HSR on small towns in Germany on the Cologne-Frankfurt line, but not other studies done in other countries. For example, in Japan, the effect of the Shinkansen on the Tohoku and Joetsu regions was decidedly mixed. The report also quotes the positive story of Lille’s TGV-fueled redevelopment, which was not replicated anywhere else in France, where cities just passively waited for infrastructure to rescue them. But instead of talking about Lille’s program of redevelopment, the report contrasts it with failed development cases in cities with exurban stations, never mind that no city achieved what Lille did, even ones with downtown stations, like Marseille. It’s not quite a Reason-grade lie, but it’s still very misleading.
Finally, the section about how to fund the $100 billion Northeast system and California’s $43 billion starter line has suggestions that are so outlandish they defy all explanation. The authors propose the following:
1. Raise the gas tax by 15 cents a gallon or more. Several cents could be devoted to passenger rail.
2. Add a $1 surcharge on current passenger rail tickets to produce approximately $29 million annually.
3. Shift from a national gas tax to a percentage tax on crude oil and imported refined petroleum products. RAND estimated that an oil tax of 17 percent would generate approximately $83 billion a year. Five billion dollars of this amount could be dedicated to passenger rail.
Of these, proposal #2 is by far the stupidest. Amtrak receives subsidies; to tax tickets is to propose shifting some change from the left pocket to the right pocket. Why not go ahead and propose to reduce Amtrak’s subsidy by the same amount and require it to raise fares or improve efficiency?
But proposals #1 and 3 are equally bad. Wedding train funding to a steady stream of gas taxes has been the status quo for decades; the result is that APTA is so used to this unholy marriage that it opposed a climate change bill that would tax gas without diverting the funds back to transportation. (That by itself should be reason for good transit advocates to dismiss APTA as a hostile organization, just one degree less malevolent than Reason and Cato and one degree less obstructionist than the FRA.) And if it were a wise long-term choice, if it were politically feasible to add to the gas tax just to build competing trains, the US political climate would look dramatically different, and instead of talking about focus, we’d be talking about how to extend the under-construction Florida HSR line.
A report that was serious about a mode shift from cars to cleaner forms of transportation would not talk about 15 cents per gallon; it would talk in terms of multiple dollars per gallon, as gas is taxed in Europe and high-income Asia. The best explanation I can think of for the funding mechanisms is that the RPA has internalized the tax-as-user-fee model of ground transportation, one that has never worked for cars despite the AAA’s pretense otherwise and that won’t work for anything else.
The overall tone of the report slightly reminds me of Thomas MacDonald’s Highway Education Board, with its industry-sponsored “How Good Roads Help the Religious Life of My Community” essay contests. It reminds me of Thomas Friedman’s “win, win, win, win, win” columns even more – which is unsurprising since I think of Friedman as the archetypal booster – but when this boosterism applies not to a policy preference but to spending very large amounts of public money, I begin to suspect that it’s advertising rather than optimism. Friedman for all his faults crows about American and Indian entrepreneurs inventing new things rather than about extracting $100 billion from the Northeast to pay for unnecessary greenfield tunneling.
Therefore, good transit activists should dismiss this report, and avoid quoting it as evidence that prioritizing is necessary. This was not what the RPA was preaching back when it thought it could get away with proposing more, and the rest of the report is so shoddy it’s not a reliable source of analysis. There may be other reasons to focus on those corridors, but the RPA did not argue them much, instead preferring to literally go for big bucks.
While reading a thesis about tilting trains, I saw a comparison of passenger comfort on different modes of transportation. This includes the following graph (p. 30), which the thesis sources to a study of motion sickness in US children and teenagers:
The scale is originally 0-3: this study polled a sample aged 9-18 and asked whether they feel nauseous on any of the above modes, where 0 is “never” and 3 “always.”