I ran a Patreon poll with three options for posts about design compromises: overbuilding for future capacity needs, building around compromises with unfixably bad operations, and where to build when it’s impossible to get transit-oriented development right. Overbuilding won with 16 votes to bad operations’ 10 and development’s 13.
It’s generally best to build infrastructure based exactly on expected use. Too little and it gets clogged, too much and the cost of construction is wasted. This means that when it comes to rail construction, especially mainline rail, infrastructure should be sized for the schedule the railroad intends to run in the coming years. The Swiss principle that the schedule comes first was just adopted in Germany; based on this principle, infrastructure construction is geared around making timed transfers and overtakes and shortening schedules to be an integer (or half-integer) multiple of the headway minus turnaround time for maximum equipment utilization.
And yet, things aren’t always this neat. This post’s topic is the issue of diachronic optimization. If I design the perfect rail network for services that come every 30 minutes, I will probably end up with a massive upgrade bill if ridership increases to the point of requiring a train every 20 minutes instead. (I chose these two illustrative numbers specifically because 30 is not a multiple of 20.) In some cases, it’s defensible to just build for higher capacity – full double-tracking even if current ridership only warrants a single track with passing sidings, train stations with more tracks in case more lines are built to connect to them, and so on. It’s a common enough situation that it’s worth discussing when what is technically overbuilding is desirable.
Expected growth rates
A fast-growing area can expect future rail traffic to rise, which implies that building for future capacity today is good. However, there are two important caveats. The first is that higher growth usually also means higher uncertainty: maybe our two-track commuter line designed around a peak of 8 trains per hour in each direction will need 32 trains per hour, or maybe it will stay at 8 for generations on end – we usually can’t guarantee it will rise steadily to 16.
The second caveat, applicable to fast-growing developing countries, is that high growth raises the cost of capital. Early British railroads were built to higher standard than American ones, and the explanation I’ve seen in the rail history literature is that the US had a much higher cost of capital (since growth rates were high and land was free). Thus mainlines in cities (like the Harlem) ran in the middle of the street in the US but on elevated structures in Britain.
But with that in mind, construction costs have a secular increase. Moreover, in constrained urban areas, the dominant cost of above-ground infrastructure cost is finding land for multiple tracks of railroad (or lanes of highway), and those are definitely trending up. The English working class spent 4-5% of its income on rent around 1800 (source, PDF-p. 12); today, spending one third of income on rent is more typical, implying housing costs have grown faster than incomes, let alone the general price index.
The upshot is that cities that can realistically expect large increases in population should overbuild more, and optimize the network around a specific level of traffic less. Switzerland and Germany, both of which are mature, low-population growth economies, can realistically predict traffic many decades hence. India, not so much.
The expected growth rate helps determine the future benefits of overbuilding now, including reduced overall costs from fronting construction when costs are expected to grow. Against these benefits, we must evaluate the costs of building more than necessary. These are highly idiosyncratic, and depend on precise locations of needed meets and overtakes, potential connection points, and the range of likely train frequencies.
On the Providence Line, the infrastructure today is good for an intercity train at current Amtrak speed every 15 minutes and a regional train making every stop every 15 minutes. There is one overtake segment at Attleboro, around three quarters of the way from Boston to Providence, and the line is otherwise double-track with only one flat junction, with the Stoughton branch. If intercity trains are sped up to the maximum speed permitted by right-of-way geometry, an additional overtake segment is required about a quarter of the way through, around Readville and Route 128. If the trains come every 10 minutes, in theory a mid-line overtake in Sharon is required, but in practice three overtakes would be so fragile that instead most of the line would need to be four-tracked (probably the entire segment from Sharon to Attleboro at least). This raises the incremental costs of providing infrastructure for 10-minute service – and conversely, all of this is in lightly developed areas, so it can be deferred without excessive future increase in costs.
An even starker example of high incremental costs is in London. Crossrail 2 consists of three pieces: the central tunnel between Clapham Junction and Euston-St. Pancras, the northern tunnel meandering east to the Lea Valley Lines and then back west to connect to the East Coast Main Line, and the southern tunnel providing two extra tracks alongside the four-track South West Main Line. The SWML is held to be at capacity, but it’s not actually at the capacity of an RER or S-Bahn system (as I understand it, it runs 32 trains per hour at the peak); the two extra tracks come from an expectation of future growth. However, the extreme cost of an urban tunnel with multiple new stations, even in relatively suburban South London, is such that the tunnel has to be deferred in favor of above-ground treatments until it becomes absolutely necessary.
In contrast, an example of low incremental costs is putting four tracks in a cut-and-cover subway tunnel. In absolute terms it’s more expensive than adding passing tracks in suburban Massachusetts, but the effect on capacity is much bigger (it’s an entire track pair, supporting a train every 2 minutes), and moreover, rebuilding a two-track tunnel to have four tracks in the future is expensive. Philadelphia most likely made the right choice to build the Broad Street Line four-track even though its ridership is far below the capacity of two – in the 1920s it seemed like ridership would keep growing. In developing countries building elevated or cut-and-cover metros, the same logic applies.
The two main aspects of every infrastructure decision are costs and benefits. But we can discern some patterns in when overbuilding is useful:
- Closing a pinch point in a network, such as a single- or double-track pinch point or a flat junction, is usually worth it.
- Cut-and-cover or elevated metro lines in cities that are as large as prewar New York (which had 7 million people plus maybe 2 million in the suburbs) or can expect to grow to that size class should have four tracks.
- On a piece of infrastructure that is likely to be profitable, like high-speed rail, deferring capacity increases until after operations start can be prudent, since the need to start up the profitable system quickly increase the cost of capital.
- Realistic future projections are imperative. Your mature first-world city is not going to triple its travel demand in the foreseeable future.
- Higher uncertainty raises the effective cost of capital, but it also makes precise planning to a specific schedule more difficult, which means that overbuilding to allow for more service options becomes reasonable.
- The electronics before concrete principle extends to overbuilding: it’s better to complete a system (such as ETCS signaling or electrification) even if some branches don’t merit it yet just because of the benefits of having a single streamlined class of service, and because of the relatively low cost of electronics.
Usually cities and countries should not try to build infrastructure ahead of demand – there are other public and private priorities competing for the same pool of money. But there are some exceptions, and I believe these principles can help agencies decide. As a matter of practice, I don’t think there are a lot of places in the developed world where I’d prescribe overbuilding, but in the developing world it’s more common due to higher future growth rates.
I do not know how to code. The most complex actually working code that I have written is 48 lines of Python that implement a train performance calculator that, before coding it, I would just run using a couple of Wolfram Alpha formulas. Here is a zipped version of the program. You can download Python 2.7 and run it there; there may also be online applets, but the one I tried doesn’t work well.
You’ll get a command line interface into which you can type various commands – for example, if you put in 2 + 5 the machine will natively output 7. What my program does is define functions relevant to train performance: accpen(k,a,b,c,m,x1,x2,n) is the acceleration penalty from speed x1 m/s to speed x2 m/s where x1 < x2 (if you try the other way around you’ll get funny results) for a train with a power-to-weight ratio of k kilowatts per ton, an initial acceleration rate of m m/s^2, and constant, linear, and quadratic running resistance terms a, b, and c. To find the deceleration penalty, put in decpen, and to find the total, either put in the two functions and add, or put in slowpen to get the sum. The text of the program gives the values of a, b, and c for the X2000 in Sweden, taken from PDF-p. 64 of a tilting trains thesis I’ve cited many times. A few high-speed trainsets give their own values of these terms; I also give an experimentally measured lower air resistance factor (the quadratic term c) for Shinkansen. Power-to-weight ratios are generally available for trainsets, usually on Wikipedia. Initial acceleration rates are sometimes publicly available but not always. Finally, n is a numerical integration quantity that should be set high, in the high hundreds or thousands at least. You need to either define all the quantities when you run the program, or plug in explicit numbers, e.g. slowpen(20, 0.0059, 0.000118, 0.000022, 1.2, 0, 44.44, 2000).
I’ve used this program to find slow zone penalties for recent high-speed rail calculations, such as the one in this post. I thought it would not be useful for regional trains, since I don’t have any idea what their running resistance values are, but upon further inspection I realized that at speeds below 160 km/h resistance is far too low to be of any consequence. Doubling c from its X2000 value to 0.000044 only changes the acceleration penalty by a fraction of a second up to 160 km/h.
With this in mind, I ran the program with the parameters of the FLIRT, assuming the same running resistance as the X2000. The FLIRT’s power-to-weight ratio is 21.1 in Romandy, and I saw a factsheet in German-speaking Switzerland that’s no longer on Stadler’s website citing slightly lower mass, corresponding to a power-to-weight ratio of 21.7; however, these numbers do not include passengers, and adding a busy but not full complement of passengers adds mass to the train until its power-to-weight ratio shrinks to about 20 or a little less. With an initial acceleration of about 1.2 m/s^2, the program spits out an acceleration penalty of 23 seconds from 0 to 160 km/h (i.e. 44.44 m/s) and a deceleration penalty of 22 seconds. In videos the acceleration penalty appears to be 24 seconds, which difference comes from a slight ramping up of acceleration at 0 km/h rather than instant application of the full rate.
In other words: the program manages to predict regional train performance to a very good approximation. So what about some other trains?
I ran the same calculation on Metro-North’s M-8. Its power-to-weight ratio is 12.2 kW/t (each car is powered at 800 kW and weighs 65.5 t empty), shrinking to 11.3 when adding 75 passengers per car weighing a total of 5 tons. A student paper by Daniel Delgado cites the M-8’s initial acceleration as 2 mph/s, or 0.9 m/s^2. With these parameters, the acceleration penalty is 37.1 seconds and the deceleration penalty is 34.1 seconds; moreover, the paper show how long it takes to ramp up to full acceleration rate, and this adds a few seconds, for a total stop penalty (excluding dwell time) of about 75 seconds, compared with 45 for the FLIRT.
In other words: FRA-compliant EMUs add 30 seconds to each stop penalty compared with top-line European EMUs.
Now, what about other rolling stock? There, it gets more speculative, because I don’t know the initial acceleration rates. I can make some educated guesses based on adhesion factors and semi-reliable measured acceleration data (thanks to Ari Ofsevit). Amtrak’s new Northeast Regional locomotives, the Sprinters, seem to have k = 12.2 with 400 passengers and m = 0.44 or a little less, for a penalty of 52 seconds plus a long acceleration ramp up adding a brutal 18 seconds of acceleration time, or 70 in total (more likely it’s inaccuracies in data measurements – Ari’s source is based on imperfect GPS samples). Were these locomotives to lug heavier coaches than those used on the Regional, such as the bilevels used by the MBTA, the values of both k and m would fall and the penalty would be 61 seconds even before adding in the acceleration ramp. Deceleration is slow as well – in fact Wikipedia says that the Sprinters decelerate at 5 MW and not at their maximum acceleration rate of 6.4 MW, so in the decpen calculation we must reduce k accordingly. The total is somewhere in the 120-150 second range, depending on how one treats the measured acceleration ramp.
In other words: even powerful electric locomotives have very weak acceleration, thanks to poor adhesion. The stop penalty to 160 km/h is about 60 seconds higher than for the M-8 (which is FRA-compliant and much heavier than Amfleet coaches) and 90 seconds higher than for the FLIRT.
Locomotive-hauled trains’ initial acceleration is weak that reducing the power-to-weight ratio to that of an MBTA diesel locomotive (about 5 kW/t) doesn’t even matter all that much. According to my model, the MBTA diesels’ total stop penalty to 160 km/h is 185 seconds excluding any acceleration ramp and assuming initial acceleration is 0.3 m/s^2, so with the ramp it might be 190 seconds. Of note, this model fails to reproduce the lower acceleration rates cited by a study from last decade about DMUs on the Fairmount Line, which claims a 70-second penalty to 100 km/h; such a penalty is far too high, consistent with about 0.2 m/s^2 initial acceleration, which is far too weak based on local/express time differences on the schedule. The actual MBTA trains only run at 130 km/h, but are capable of 160, given long enough interstations – they just don’t do it because there’s little benefit, they accelerate so slowly.
Unsurprisingly, modern rail operations almost never buy locomotives for train services that are expected to stop frequently, and some, including the Japanese and British rail systems, no longer buy electric locomotives at all, using EMUs exclusively due to their superior performance. Clem Tillier made this point last year in the context of Caltrain: in February the Trump administration froze Caltrain’s federal electrification funding as a ploy to attack California HSR, and before it finally relented and released the money a few months later, some activists discussed Plan B, one of which was buying locomotives. Clem was adamant that no, based on his simulations electric locomotives would barely save any time due to their weak acceleration, and EMUs were obligatory. My program confirms his calculations: even starting with very weak and unreliable diesel locomotives, the savings from replacing diesel with electric locomotives are smaller than those from replacing electric locomotives with EMUs, and depending on assumptions on initial acceleration rates might be half as high as the benefits of transitioning from electric locomotives to EMUs (thus, a third as high as those of transitioning straight from diesels to EMUs).
Thus there is no excuse for any regional passenger railroad to procure locomotives of any kind. Service must run with multiple units, ideally electric ones, to maximize initial acceleration as well as the power-to-weight ratio. If the top speed is 160 km/h, then a good EMU has a stop penalty of about 45 seconds, a powerful electric locomotive about 135 seconds, and a diesel locomotive around 190 seconds. With short dwell times coming from level boarding and wide doors, EMUs completely change the equation for local service and infill stops, making more stops justifiable in places where the brutal stop penalty of a locomotive would make them problematic.
Successful transit cities are not alike. There are large differences in how the most expansive transit networks are laid out. It takes multiple series of posts across several blogs (not just mine but also Human Transit and others) covering just one of them, for example stop spacing or how construction contracts are let. With so much variation, it’s easy to get caught up in details that differentiate the best systems. After all, the deepest communities of railfans tend to sprout in the cities with the largest rail networks; arguing with railfans with experience with London, Tokyo, or Paris is difficult because they know intricate details of how their systems work that I am catching up on but only know in the same depth for New York. Add in the fact that London and Paris view each other as peer cities and from there the route to arguing minutiae about two cities that by most standards have good public transit is short.
But what if this is wrong? What if, instead of or in addition to figuring out differences among the top transit cities, it’s useful to also figure out what these transit cities have in common that differentiates them from auto-oriented cities? After all, in other aspects of development or best practices this is well-understood: for example, a developing country can choose to aim to be hyper-capitalist like Singapore or the US or social democratic like Sweden or France, but it had better develop the institutions that those four countries have in common that differentiate them from the third world.
Unfortunately, before discussing what the common institutions to transit cities are, it’s necessary to discuss things that may be common but don’t really matter.
The US as a confounding factor
The biggest problem with figuring out things all good transit cities have in common is that in the developed world, the US (and to some extent Canada and Australia) is unique in having bad transit. Frequent commenter Threestationsquare has a list of cities by annual rapid transit ridership (counting BRT but not infrequent commuter rail, which lowballs parts of the US); New York is near the top, but the second highest in the US, a near-tie between Boston, Chicago, and Washington, would rank #22 in Europe. As a result, some social, political, and technical features that appear to differentiate good and bad transit are not really about transit but about the US and must be discarded as confounding factors. Fortunately, most of these confounding factors are easy to dispose of since they also occur in New York.
The more difficult question concerns factors that are distantly related to the weakness of US transit but are not direct explanations. I wrote about racism as such a factor a few months ago, arguing that high US construction costs come from weak civil service, which in turn comes from the way American segregation works. The US is not uniquely racist or even uniquely segregated; the unique aspect is that it a) has a long-settled oppressed minority and not just immigrants who arrived after the characteristic of the state was established, and b) has segregation within metro areas (unlike Singapore, which has social but not spatial segregation) but not between them (unlike Israel, where the built-up area of Tel Aviv has very few Arabs). But while this can explain why institutions developed in a way that’s hostile to transit, it’s not a direct explanation for poor US transit except in Atlanta, where the white state underinvests in the black city. White people in Boston, Los Angeles, Houston, and other cities with little to no public transit do not avoid the bus or the train out of stereotypes that match typical American racial stereotypes, such as crime; they avoid the train because it doesn’t go where they’re going and the bus because it is slow and unreliable.
There are two ways to avoid confounding factors. The first is the sanity check, where available: if some feature of transit exists across major transit cities but is absent in auto-oriented cities not just in the US but also in Canada, Australia, New Zealand, Israel, and Italy, then it’s likely to be relevant. Unfortunately, clean examples are rare. The second and more difficult method is to have theoretical understanding of what matters.
London and Paris are transit cities. So are Prague and Stockholm. I’ve stressed the importance of scale-variance before: features that work in larger cities may fail in smaller ones and vice versa. Thus, it’s best to look at common features of successful transit cities within each size class separately.
In fact, one way cities can fail is by adopting transit features from cities of the wrong size class. China is making the mistake in one direction: Beijing and Shanghai have no express subway trains or frequent regional rail services acting as express urban rail, and as a result, all urban travel has to slow down to an average speed of about 35 km/h, whereas Tokyo has express regional lines averaging 60 km/h. China’s subway design standards worked well for how big its cities were when those standards were developed from the 1970s to the 1990s, but are too small for the country’s megacities today.
In contrast, in the developed world, the megacities with good public transit all have frequent express trains: Tokyo and Osaka have four-track (or even eight-track!) regional lines, Paris has the RER, New York has express subways (and the premium-price LIRR trains from Jamaica to Penn Station), London has fast regional rail lines and Thameslink and will soon have Crossrail, Seoul has a regional rail network with express trains on Subway Line 1, and Moscow stands alone with a strictly two-track system but has such wide stop spacing that the average speed on the Metro is 41 km/h. Smaller transit cities sometimes have frequent express trains (e.g. Zurich and Stockholm) and sometimes don’t (e.g. Prague), but it’s less important for them because their urban extent is such that a two-track subway line can connect the center with the edge of the built-up area in a reasonable amount of time.
And if China failed by adopting design standards fitting smaller cities than it has today, the US fails in the other direction, by adopting design standards fitting huge megacities, i.e. New York. Small cities cannot hope to have lines with the crowding levels of the Lexington Avenue Line. This has several implications. First, they need to scale their operating costs down, by using proof of payment ticketing and unstaffed stations, which features are common to most European transit cities below London and Paris’s size class. Second, they need to worry about train frequency, since it’s easy to get to the point where the frequency that matches some crowding guideline is so low that it discourages riders. And third, they need to maximize network effects, since there isn’t room for several competing operations, which means ensuring buses and trains work together and do not split the market between them.
The best example of an American city that fails in all three aspects above is Washington. While railfans in Washington lament the lack of express tracks like those of New York, the city’s problems are the exact opposite: it copied aspects of New York that only succeed in a dense megacity. With interlining and reverse-branching, Washington has low frequency on each service, down to 12 minutes off-peak. The stations are staffed and faregated, raising operating costs. And there is no fare integration between Metro and the buses, splitting the market in areas with price-sensitive riders (i.e. poor people) like Anacostia.
The political situation
While I’ve written before about what I think good metro design standards are, these standards themselves cannot separate the major transit cities from cities like Los Angeles (which has about two and a half rail trunks in a metro area larger than that of London or Paris) or Tel Aviv (which has no metro at all). Instead, it’s worth asking why these cities have no large subway systems to begin with.
In the case of Tel Aviv, Israel has had an official policy of population dispersal since independence. After independence the North and South of the country had Arab majorities, and the government wished to encourage Jews to settle there to weaken any Palestinian claims to these areas. As a result, Prime Minister David Ben Gurion rejected a plan to develop an urban rail network centered on Tel Aviv and instead encouraged low-income Jewish immigrants to move far away, either to depopulated Arab towns or to new towns (“development towns”) built at strategic points for national geopolitics. Decentralization was national policy, and with it came auto-oriented urbanism. A less harsh but equally politicized environment led to Malaysia’s auto-centric layout: Paul Barter’s thesis outlines how Malaysia choked informal transit and encouraged auto-oriented suburbanization in order to create an internal market for state-owned automakers.
In the case of the US, the situation is more complex, since there were several distinct political trends in different eras favoring cars. In postwar suburbia (and in Los Angeles going back to the 1920s) it was the association of cars with middle-class normality, and in California also with freedom from hated railroads; it’s related to the fact that American suburbanization was led by the middle class rather than by the working class as with more recent exurbanization. In Israel suburbanization was led by the working class, but the deliberate government policy of decentralization meant that the urban middle class’s demands for better transportation were ignored until the 1990s.
Without enough of an urban middle class to advocate for more transit, US transit withered. New cities in the Sunbelt had little demand for public transit, and in the older cities the middle class cared little for any transit that wasn’t a peak-only commuter train from the suburbs to the CBD. Moreover, in existing transit cities the middle class demanded that the urban layout change to fit its suburban living situation, leading to extensive job sprawl into office parks that are difficult to serve on transit. This paralleled trends in Canada, Australia, and New Zealand; Sydney in particular saw middle-class suburbanization early, like Los Angeles.
The political situation changed in the 1970s, 80s, and 90s, but by then high construction costs, NIMBYism constraining the extent of TOD (unlike in Canada), and indifference to leveraging regional rail for urban transit (as in Canada and until recently Israel but unlike in Australia) made it difficult to build more public transit lines.
Regional rail and TOD
The largest transit cities in the rich and middle-income world all make extensive use of regional rail, with the aforementioned exception of Chinese cities, where the lack of regional rail is creating serious travel pain, and New York, where the city itself is transit-oriented but its suburbs are not. Smaller transit cities usually make use of regional rail as well, but this isn’t universal, and to my understanding is uncommon in Eastern Europe (e.g. Kyiv has one semi-frequent ring line) even in cities with very high metro and tramway usage.
However, smaller transit cities that do not have much regional rail have full metro systems and not just tramways, let alone BRT. Curitiba and Bogota are famous for their BRT-only transit networks, but both instituted their systems in a context with low labor costs and both are building metro systems right now.
The other common element to transit cities is TOD. Here, we must distinguish old cities like London, Paris, Berlin, and Vienna, whose urban layout is TOD because it was laid out decades before mass motorization, and newer cities like Stockholm, Tokyo, and every city in Eastern Europe or the East Asian tiger states. The latter set of cities built housing on top of train stations, often public housing (as in the communist world or in Stockholm) but not always (as in Tokyo and to some extent Hong Kong), in an era when the global symbol of prosperity was still the American car-owning middle class.
The importance of TOD grows if we compare countries with relatively similar histories, namely, the US and Canada. Neither country does much regional rail, both have had extensive middle-class suburbanization (though Canada’s major cities have maintained bigger inner-urban middle classes than the US’s), and English Canada’s cities came into the 1970s with low urban density. The difference is that Canada has engaged in far more TOD. Calgary built up a large CBD for how small the city is, without much parking; Vancouver built up Downtown as well as transit-oriented centers such as Metrotown, New Westminster, Lougheed, and Whalley, all on top of the Expo Line. Nowhere in the US did such TOD happen. Moreover, American examples of partial TOD, including Arlington on top of the Washington Metro and this decade’s fast growth in Seattle, have led to somewhat less awful transit usage than in the rest of the country.
Most cities in the developed world are replete with legacy rail networks that can be leveraged for high-quality public transit. We see cities that aim at transit revival start with regional rail modernization, including Auckland and to some extent Tel Aviv (which is electrifying its rail network and building new commuter lines, but they run in freeway medians due to poor planning). Moreover, we see cities that are interested in transit build up high-rise CBDs in their centers and high- and mid-rise residential development near outlying train stations.
“Regional rail and TOD” is not a perfect formula; it elides a lot of details and a lot of historical factors that are hard to replicate. But both regional rail and TOD have been major elements in the construction of transit cities over the last 60 years, and while they both have exceptions, they don’t have many exceptions. In the other direction, I don’t know of examples of failed TOD – that is, of auto-oriented cities that aggressively built TOD on top of new or existing rail lines but didn’t manage to grow their transit ridership. I do know some examples of failed regional rail, but usually they make glaring mistakes in design standards, especially frequency but also station siting and fare integration.
At a closer in level of zoom, it’s worthwhile to talk about the unique features of each transit city. But when looking at the big picture, it’s better to talk about what all transit cities of a particular size class have in common that auto-oriented cities don’t. Only this way can an auto-oriented city figure out what it absolutely must do if it wants to have better public transit and what are just tools in its kit for achieving that goal.
Boston has two main train stations: South Station, and North Station. Both are terminals, about 2 km apart, each serving its own set of suburbs; as a result, over the last few decades there have been calls to unify the system with a regional rail tunnel connecting the two systems. This tunnel, called the North-South Rail Link, or NSRL, would have been part of the Big Dig if its costs hadn’t run over; as it were, the Big Dig reserved space deep underground for two large bores, in which there is clean dirt with no archeological or geotechnical surprises. The NSRL project had languished due to Massachusetts’ unwillingness to spend the money on it, always understood to be in the billions, but in the last few years the pressure to build it intensified, and the state agreed to fund a small feasibility study.
A presentation of the draft study came out two days ago, and is hogwash. It claims on flimsy pretext that NSRL would cost $17 billion for the tunnel alone. It also makes assumptions on service patterns (such as manual door opening) that are decades out of date not just in Europe and East Asia but also in New York. The Fiscal and Management Control Board, or FMCB, discusses it here; there’s a livestream as well as a link to a presentation of the draft study.
The content of the study is so weak that it has to have been deliberate. The governor does not want it built because of its complexity, no matter how high its benefits. Thus, the state produced a report that sandbags a project it doesn’t want to build. People should be fired over this, starting with planners at the state’s Office of Transportation Planning, which was responsible for the study. The way forward remains full regional rail modernization. As for the cost estimate, an independent study by researchers at Harvard’s Kennedy School of Government estimates it at about $5 billion in today’s money; the new study provides no evidence it would be higher. I urge good transit activists in Massachusetts, Rhode Island, and New Hampshire to demand better of their civil servants.
The study says that the cost of a four-track NSRL tunnel under the Big Dig would be $17 billion in 2028 dollars. In today’s money, this is $12 billion (the study assumes 3.5% annual cost escalation rather than inflation-rate cost escalation). It claims to be based on best practices, listing several comparable tunnels, both proposed and existing:
- California High-Speed Rail tunnels (average estimated cost about $125 million per km, not including overheads and contingency)
- Crossrail (see below on costs)
- The M-30 highway tunnel in Madrid (average cost about $125 million per km of bored tunnel in the mid-2000s, or around $150 million/km in today’s money)
- The canceled I-710 tunnel in California (at 7.2 km and $5.6 billion, $780 million per km
- The Spoortunnel Pannerdensch Kanaal (around $200 million in today’s money for 1.6 km of bore, or $125 million per km)
Unlike the other tunnels on the list, Crossrail has stations frustrating any simple per km cost analysis. The headline cost of Crossrail is £15 billion; however, I received data from a freedom of information request showing that the central (i.e. underground) portion is only £11.6 billion and the rest is surface improvements, and of this cost the big items are £2.2 billion for tunneling, £4.1 billion for stations, £1 billion for tracks and systems, and £2.7 billion for overheads and land acquisition. The tunneling itself is thus around $150 million per km, exclusive of overheads and land (which add 30% to the rest of the project). All of this is consistent with what I’ve found in New York: tunneling is for the most part cheap.
With the exception of Crossrail, the above projects consist of two large-diameter bores. The mainline rail tunnels (California HSR and Pannerdensch Kanaal) are sized to provide plenty of free air around the train in order to improve aerodynamics, a feature that is desirable at high speed but is a luxury in a constrained, low-speed urban rail tunnel. The highway tunnels have two large-diameter bores in order to permit many lanes in each direction. The plan for NSRL has always been two 12-meter bores, allowing four tracks; at the per-km boring cost of the above projects, this 5 kilometer project should cost perhaps a billion dollars for tunneling alone.
The stations are typically the hard part. However, NSRL has always been intended to use large-diameter tunnels, which can incorporate the platforms within the bore, reducing their cost. Frequent commenter Ant6n describes how Barcelona used such a tunnel to build Metro Lines 9 and 10, going underneath the older lines; the cost of the entire project is around $170 million per km, including a cost overrun by a factor of more than 3. Vertical access is likely to be more difficult in Boston under the Big Dig than in Barcelona, but slant shafts for escalators are still possible. At the worst case scenario, Crossrail’s station costs are of an order of magnitude of many hundreds of millions of dollars each, and two especially complex ones on Crossrail 2 are £1.4 billion each; this cost may be reasonable for Central Station at Aquarium, but not at South Station or North Station, where there is room for vertical and slant shafts.
It’s possible that the study made a factor-of-two error, assuming that since the mainline rail comparison projects have two tracks, their infrastructure is sized for two urban rail tracks, where in reality a small increase in tunnel diameter would permit four.
Researchers at the Harvard Kennedy School of Government came up with an estimate of $5.9 billion in 2025 dollars for a four-track, three-station NSRL option, which is about $5 billion today. Their methodology involves looking at comparable tunneling projects around the world, and averaging several averages, one coming from American cost methodology plus 50% contingency, and two coming from looking at real-world cost ranges (one American, one incorporating American as well as rest-of-world tunnels). Their list of comparable projects includes some high-cost ones such as Second Avenue Subway, but also cheaper ones like Citybanan, which goes deep underneath Central Stockholm with mined tunnels under T-Centralen and Odenplan, at $350 million per km in today’s money.
But the MassDOT study disregarded the expertise of the Kennedy School researchers, saying,
Note: The Harvard Study did not include cost for the tunnel boring machine launch pit and only accounted for 2.7 miles of tunneling (the MassDOT studies both accounted for 5 miles of tunneling), and no contingency for risk.
This claim is fraudulent. The Kennedy School study looks at real-world costs (thus, including contingency and launch pit costs) as well as at itemized costs plus 50% contingency. Moreover, the length of the NSRL tunnel, just under 5 km, is the same either way; the MassDOT study seems to be doubling the cost because the project has four tracks, an assumption that is already taken into account in the Kennedy School study. This, again, is consistent with a factor-of-two error.
Moreover, the brazenness of the claim that a study that explicitly includes contingency does not do so suggests that MassDOT deliberately sabotaged NSRL, making it look more expensive than it is, since the top political brass does not want it. Governor Baker said NSRL looks expensive, and Secretary of Transportation Stephanie Pollack is hostile as well; most likely, facing implicit pressure from above, MassDOT’s overburdened Office of Transportation Planning scrubbed the bottom of the barrel to find evidence of absurdly high costs.
Massachusetts really does not want or understand electrification. Even some NSRL supporters believe electrification to be an expensive frill that would sink the entire project and think that dual-mode locomotives are an acceptable way to run trains in a developed country in the 2010s.
In fact, dual-mode locomotives’ weak performance serves to raise tunneling costs. Struggling to accelerate at 0.3 m/s^2 (or 0.03 g), they cannot climb steep grades: both the Kennedy School and MassDOT studies assume maximum 3% grades, whereas electric multiple units, with initial acceleration of 1.2 m/s^2, can easily climb 4% and even steeper grades (in theory even 10%, in practice the highest I know of is 7%, and even 5% is rare), permitting shorter and less constrained tunnels.
As a result of its allergy to electrification, MassDOT is only proposing wiring between North Station and the next station on each of the four North Side lines, a total of 22.5 route-km. This choice of which inner segments to electrify excludes the Fairmount Line, an 8-stop 15 km mostly self-contained line through low-income, asthma-riven city neighborhoods (source, PDF-pp. 182 and 230). Even the electrification the study does agree to, consisting of about 30 km of the above surface lines plus the tunnels themselves, is projected to cost $600 million. Nowhere in the world is electrification so expensive; the only projects I know of that are even half as expensive are a pair of disasters, one coming from a botched automation attempt on the Great Western Main Line and one coming from poor industry practices on Caltrain.
A more reasonable American budget, based on Amtrak electrification costs from the 1990s, would be somewhat less than $2 billion for the entire MBTA excluding the already-wired Providence Line; this is the most familiar electrification scheme to the Bostonian reader or planner. At French or Israeli costs, the entire MBTA commuter rail system could be wired for less than a billion dollars.
Another necessary element is conversion to an all-EMU fleet, to increase performance and reduce operating costs. Railway Gazette reports that a Dutch benchmarking study found that the lifecycle costs of EMUs are half as high as those of diesel multiple units. As the MBTA needs to replace its fleet soon anyway, the incremental cost of electrification of rolling stock is negative, and yet the study tacks in $2.4 billion on top of the $17 billion for tunneling for vehicles.
A miscellany of incompetence
In addition to the sandbagged costs, the study indicates that the people involved in the process do not understand modern railroad operations in several other ways.
First, door opening. While practically everywhere else in the first world doors are automatic and opened with the push of a button, the MBTA insists on manual door opening. The MassDOT study gives no thought to high platforms and automatic doors (indeed, the Old Colony Lines are already entirely high-platform, but some of their rolling stock still employs manual door opening), and assumes manual door opening will persist even through the NSRL tunnels. Each train would need a squad of conductors to unload in Downtown Boston, and the labor costs would frustrate any attempt to run frequently (the study itself suggests hourly off-peak frequency; in Paris, RER lines run every 10-20 minutes off-peak).
Second, capacity. The study says a two-track NSRL would permit 17 trains per hour in each direction at the peak, and a four-track NSRL would permit 21. The MBTA commuter rail network is highly branched, but not more so than the Munich S-Bahn (which runs 30 at the peak on two tracks) and less so than the Zurich S-Bahn (which before the Durchmesserlinie opened ran either 20 or 24 tph through the two-track tunnel, I’m not sure which).
Worse, the FMCB itself is dumbfounded by the proposed peak frequency – in the wrong direction. While FMCB chair Joe Aiello tried explaining how modern regional rail in Tokyo works, other members didn’t get it; one member dared ask whether 17 tph is even possible on positive train control-equipped tracks. My expectations of Americans are low enough that I am not surprised they are unaware that many lines here and in Japan have automatic train protection systems (ETCS here, various flavors of ATC in Japan) that meet American PTC standards and have shorter minimum headways than every 3-4 minutes. But the North River Tunnels run 24-25 peak tph into Manhattan, using ASCES signaling, the PTC system Amtrak uses on the Northeast Corridor; the capacity problems at Penn Station are well-known to even casual observers of American infrastructure politics.
A state in which the FMCB members didn’t really get what their chair was saying about modern operations is going to propose poor operating practices going forward. MassDOT’s study assumes low frequency, and, because there is no line-wide electrification except on the Providence Line and eventually South Coast Rail (where electrification is required for wetland remediation), very low performance. MassDOT’s conception of NSRL has no infill stops, and thus no service to the bulk of the contiguous built-up area of Boston. Without electrification or high platforms, it cannot achieve high enough speeds to beat cars except in rush hour traffic. Limiting the stop penalty is paramount on urban rail, and level boarding, wide doors, and EMU acceleration combine to a stop penalty of about 55 seconds at 100 km/h and 75 seconds at 160 km/h; in contrast, the MBTA’s lumbering diesel locomotives, tugging coaches with narrow car-end doors with several steps, have a stop penalty of about 2.5 minutes at 100 km/h.
The presentation makes it very clear what the value of MassDOT’s NSRL study is: at best none, at worst negative value through muddying the conversation with fraudulent numbers. The Office of Transportation Planning is swamped and could not produce a good study. The actual control was political: Governor Baker and Secretary of Transportation Pollack do not want NSRL, and both the private consultant that produced the study and the staff that oversaw it did what the politicians expected of them.
Heads have to roll if Massachusetts is to plan good public transportation. The most important person good transit activists should fight to remove is the governor; however, he is going to be easily reelected, and replacing the secretary of transportation with someone who does not lie to the public about costs is an uphill fight as well. Replacing incompetent civil servants elsewhere is desirable, but the fish rots from the head.
Activists in Rhode Island may have an easier time, as the state is less hostile to rail, despite the flop of Wickford Junction; they may wish to demand the state take lead on improving service levels on the Providence Line, with an eye toward forcing future NSRL plans to incorporate good regional rail practices. In New Hampshire, provided the state government became less hostile to public investment, activists could likewise demand high-quality commuter rail service, with an eye toward later connecting a North Station-Nashua-Manchester line to the South Side lines.
But no matter what, good transit activists cannot take the study seriously as a planning study. It is a political document, designed to sandbag a rail project that has high costs and even higher benefits that the governor does not wish to manage. Its cost estimates are not only outlandish but brazenly so, and its insistence that the Kennedy School study does not include contingency is so obviously incorrect that it must be considered fraud rather than a mistake. Nothing it says has any merit, not should it be taken seriously. It does not represent the world of transportation planning, but rather the fantasies of a political system that does not understand public transportation.
In my writings about metro network design I’ve emphasized the importance of making sure every pair of intersecting lines have a transfer. Moreover, I’ve argued that missed connections often come from having very wide stop spacing, because large metro networks have very closely-spaced lines in the core, and if the stop spacing in the core is too wide, as in Moscow, then lines will frequently cross without transfers. In contrast, in Paris, where the Metro has very closely-spaced stops, there is only one missed connection on the Metro, between Lines 5 and 14. However, what’s missing from this discussion is what to do on lines that, due to network design, have to run express and miss some connections. This question mattered to most RER lines and currently matters to Crossrail and Crossrail 2, and will be critical in any New York regional rail plan.
I claim that the most important connections to prioritize should be to,
- The busiest lines.
- Lines that are orthogonal to the newly-built express lines.
But before explaining this, I’d like to go over the scale of the underlying problem of prioritizing transfers. For a start, look at the Underground in Central London:
Crossrail is the dashed gray line. Between Paddington and Liverpool Street, it intersects seven north-south lines, including five in rapid succession on the West End; stopping at all of Bond Street, Oxford Circus, Tottenham Court Road, and Holborn would slow down too much what’s intended to be an express relief line to the Central line.
Stopping between two stations and having transfers to both is possible – look at Farringdon-Barbican and at Moorgate-Liverpool Street – but results in very long transfer times. The RER has opted for this solution at Auber, which is located between the Opera and Saint-Lazare, with a transfer stretching over three successive stations on Line 3, leading to legendarily labyrinthine transfers between the RER and the Metro:
Observe that in contrast with the RER A’s convoluted transfer at Auber, the RER B simply expresses between Chatelet-Les Halles and Gare du Nord, missing the connection to the east-west Lines 3, 8, and 9 and the north-south Line 7, and only connecting to the circumferential Line 2 via a long underground passageway. The reason for this is that a transfer station at Bonne Nouvelle or Sentier would be very expensive to construct; the RER’s stations were all extremely costly, and the RER A’s record of $750 million per km for the Nation-Auber segment remains unbroken outside the Anglosphere. On Crossrail (the recordholder in cost per km outside the US, soon to be overtaken by Crossrail 2), it’s the stations that drive up costs as well, and the same problem is even more acute in New York.
The tension is then between the network effects of including more transfer points, and the costs and slowdowns induced by stopping more often. The first point in my claim at the beginning of this post follows immediately: it’s more valuable to stop at transfer points to busier lines. The RER A misses Line 5 entirely, as does the express Line 14, because Line 5 is so weak that it’s not worth it to detour from Gare de Lyon through Bastille to connect to it; the oldest plans for the RER A had a stop at Bastille and not at Gare de Lyon, but under SNCF’s influence the system was redesigned to connect to the train stations better and thus Bastille was replaced.
Whereas the RER A in theory connects to every north-south one except the weakest (although the second strongest after Line 4, Line 13, has an even longer connection than at Chatelet), Crossrail does the opposite. The busiest station in London excluding mainline stations is Oxford Circus, thanks to the three-way transfer of the Bakerloo, Victoria, and Central lines; the Victoria line is the busiest in the system per km (although the longer Northern and Central lines have more riders), and it’s certainly the busiest north-south trunk line. However, plans to have a transfer to both Bond Street and Oxford Circus were rejected in favor of a connection to Bond Street alone. The reason is that London’s low-capacity passageways get congested, and TfL’s hamfisted solution is to omit critical transfers, a decision also made at the Battersea extension of the Northern line, which will miss a connection to the Victoria line at Vauxhall.
This brings me to the second transfer priority: it’s the most important to connect to orthogonal lines. The reason is that parallel lines, especially closely parallel lines, are less likely to generate transfers. New York’s four-track subway lines have very high volumes of local-express transfers, because those are easy cross-platform interchanges; as soon as any walking between platforms is required (for example, on the Lexington Avenue Line at 59th and 86th Streets), transfer volumes fall dramatically. In Paris, transfers between Line 1 and the RER A happen, but usually for longer-distance travel; I find it faster to take Line 1 from Nation to Chatelet than to take the RER A, even without any transfer, purely because it’s easier to get between the street and the Metro platforms at both ends.
This issue was never really in contention when Paris built the original RER system. The one place where the RER prioritized a transfer to a same-direction Metro line over an orthogonal one, Gare du Nord, is such an important destination for commuter and intercity trains that it’s obviously justified to prioritize it over an easier connection to Line 2. However, more recently, the RER E has seen this issue surface with the location of the infill Rosa Parks station. The RER E could have sited a station at the intersection with Line 5, but Line 5 goes northeast and serves much the same area as the RER E, so the network effects from an interchange would be weak. Instead, the station is sited to interchange with the circumferential T3 tramway, which opens up a connection toward Nation and eventually toward Porte d’Asnieres.
In London, the same question is critical to the central route of Crossrail 2. The current plan has three Central London stops: Victoria, Tottenham Court Road (with a transfer to Crossrail), and Euston-St. Pancras. But Victoria itself is not much of a destination, and of the two lines served, the District and the Victoria, the Victoria line is parallel to Crossrail 2 rather than orthogonal to it. The purpose of Crossrail 2 is to add north-south capacity through the West End to decongest the Victoria line and reduce the shuffle at Victoria station between mainline trains and the Underground; to this end, there’s no need to stop at Victoria station itself.
To this effect, Martha Dosztal proposes moving Crossrail 2 to Westminster or possibly Charing Cross. Instead of spending $2 billion on a station at Victoria, London would need to spend probably a comparable amount on a station that interchanges with lines that go northwest-southeast like Jubilee or Bakerloo rather than on the parallel Victoria line; moreover, Westminster and Charing Cross both have connections to the District line, so Crossrail 2 would still connect to all three east-west Underground lines.
Finally, the application to New York is the most delicate. New York’s scores of missed connections come from deliberate indifference on the IND’s part to transfers with the older lines rather than any systematic attempt at prioritizing important interchanges; the older IRT and BMT systems have between them just two missed connections (3/L in Brooklyn, 4-5/R-W in Lower Manhattan). But including better connections in the event the city builds more rail lines remains critical. Second Avenue Subway gets this right by having a cross-platform transfer to the east-west F; there’s no transfer to the north-south Lexington Line, but this is less important given Second Avenue’s role as a Lexington relief line.
Regional rail transfers are especially circumscribed in New York given the system’s nature as a few short tunnels: new tunnels across the Hudson, and ideally a connection between Penn Station and Grand Central. This is why there is little room for improving connectivity between the subway and what I call Lines 1-3 of New York regional rail. However, the priorities I’m advocating in this post suggest two important things about Penn Station: first, it’s important to reopen passageways to Sixth Avenue to allow connections to the NQRW and BDFM trains; and second, it’s not important to have a connection to the 7 at Hudson Yards, as IRUM proposes.
On more speculative lines involving longer tunnels, the same priorities point to my proposed stopping pattern in and around Lower Manhattan. What I call Line 4, a north-south line from Grand Central to Staten Island stopping at Union Square and Fulton Street would intersect the east-west subways: the 7 at Grand Central, the L at Union Square, and PATH and most Brooklyn-bound trains at Fulton Street. The only missed subways – the F/M at Houston Street and the N/Q at Canal – go mostly north-south (except the M, which has a same-platform transfer with the J/Z, connecting at Fulton). Likewise, what I call Line 5, connecting from Pavonia to Atlantic Terminal, would connect to most north-south subways at Fulton Street.
Ideally, it’s better to make every interchange, and subway builders around the world should aim for very long-term planning in order to prevent missed connections in the future. However, when the inevitable changes happen and missed connections are unavoidable, there are emergent rules for which are more important: busier lines are more important than less busy lines, and less obviously, lines that are orthogonal to the new line are more important than ones that are parallel. These priorities make it possible to build express lines that maximize regional connectivity with minimal loss of travel time due to making local stops.
Continuing from last week’s post about signaling costs, here is what I’ve found about electrification costs.
Like signaling, electrification usually doesn’t make the industry press, and therefore there are fewer examples than I’d like. Moreover, the examples with concrete costs are all in countries where infrastructure costs are high: the US, Canada, the UK, Israel, New Zealand. However, a check using general reported French costs (as opposed to a specific project) suggests there is no premium in Israel and New Zealand over France, even though both countries’ urban rail tunneling projects are more expensive than Parisian Metro and RER extensions.
In the UK, the recent electrification project has stalled due to extreme cost overruns. Finding exact cost figures by segment is difficult in most of the country, but there are specific figures in the Great Western. Financial Times reports the cost of the Great Western project at £2.8 billion, covering 258 km of intercity mainline (mostly double-track, some four-track) and what I believe to be 141 km of commuter rail lines in South Wales, working from Wikipedia’s graphic and subtracting the canceled electrification to Swansea. In PPP dollars it’s around $10 million per km, but the cost may include items I exclude elsewhere in this post, such as rolling stock. For reference, in the late 2000s the project was estimated at £640 million, but costs then tripled, as the plan to automate wire installation turned out not to work. Taking the headline cost as that of the last link, £1.74 billion, the cost is $6.1 million per km, but there have been further overruns since (i.e. the Swansea cancellation).
In the US, there are three projects that I have numbers for. The most expensive of the three is Caltrain electrification, an 80 km project whose headline cost is $1.9 billion. But this includes rolling stock and signaling, and in particular, the CBOSS signaling system has wasted hundreds of millions of dollars. Electrification infrastructure alone is $697 million, or $8.5 million per km. The explanations I’ve read for this high figure include indifference to best practices (e.g. electrification masts are spaced 50 meters apart where 80 meters is more common) and generally poor contracting in the Bay Area.
The other two US projects are more remote, in two different ways. One is California High-Speed Rail: with the latest cost overrun, the projected electrification cost is $3.7 billion (table 4, PDF-p. 14). The length of route to be electrified is unclear: Phase 1, Los Angeles to San Francisco with a short branch up to Merced, is a little more than 700 km, but 80 km of that route is Caltrain, to which the high-speed rail fund is only contributing a partial amount. If the denominator is 700 km then the cost is $5.3 million per km.
The other remote US project is Amtrak’s electrification of the New Haven-Boston segment of the Northeast Corridor in the late 1990s. Back then, the 250-km double-track route was electrified for $600 million, which is $2.4 million per km, or about $3.5 million per km adjusted for inflation.
In Canada, Toronto is in the process of electrifying most of its regional rail network. The current project includes 262 route-km and has a headline cost of $13.5 billion, but according to rail consultant Michael Schabas, this includes new track, extensive junction modification, unnecessary noise walls (totaling $1 billion), and nearly 100% in contingency just because on the original budget the benefit-cost ratio seemed too good to be true. In a 2013 study, the infrastructure cost of full electrification was estimated at $2.37 billion for 450 route-km in 2010 Canadian dollars. In today’s American dollars it’s about $4.5 million per km.
In France, a report that I can no longer find stated that a kilometer of electrification cost a million euros, in the context of the electrification of a single-track legacy branch to Sables d’Olonne, used by some TGV services. While trying to find this report, I saw two different articles claiming the cost of electrification in France to be a million euros per double-track kilometer. The latter article is from 2006, so the cost in today’s money is a little higher, perhaps as high as $1.5 million per km; the article specifically says the cost includes bridge modification to permit sufficient clearances for catenary.
In Israel, the majority of the national network is currently being electrified, and I’ve argued elsewhere for a completist approach owing to the country’s small size, high density, and lack of rail connections with its neighbors. The project has been delayed due to litigation and possibly poor contractor selection, but a recent article on the subject mentions no cost overrun from the original budget of 3 billion shekels, about $750 million, for 600 km of double-track. This is $1.25 million per km and includes not just wire and substations but also 23 years’ worth of maintenance. This may be similar to the Danish ETCS project, which has been severely delayed but is actually coming in slightly under budget.
In New Zealand, the one electrification project recently undertaken, that of the Auckland regional rail network, cost $80 million in infrastructure. This is New Zealand dollars, so in US terms this is closer to $55 million. The total length of the network is about 80 route-km and 200 track-km, making the cost about $700,000 per km. But the project includes much more than wire: the maintenance facility, included in the Israeli figure, cost another NZ $100 million, and it is unclear whether bridge modifications were in the infrastructure contract or tendered separately.
The big takeaway from this dataset, taking French costs as the average (which they are when it comes to infrastructure), is that Israel and New Zealand, both small countries that use extensive foreign expertise, do not pay a premium, unlike the US, UK, and Canada. In the UK, there is a straightforward explanation: Network Rail attempted to automate the process to cut costs, and the automation failed, creating problems that blew up the budget. Premature automation is a general problem in industry: analysts have blamed it for Tesla’s production problems.
In the US and Canada, the construction cost problem is generally severe. However, it’s important to note that at NZ$2.8-3.4 billion for 3.4 km of tunnel, Auckland’s tunneling cost, around US$600 million per km, isn’t much lower than Toronto’s and is actually slightly higher than the Bay Area’s. My explanation for high costs in Israel, India, Bangladesh, Australia, Canada, New Zealand, Singapore, and Hong Kong used to be their shared English common law heritage, but this is contradicted by the lack of any British premium over French costs in the middle of the 20th century. An alternative explanation, also covering some high-cost civil law third-world countries like Indonesia and Egypt, is that these countries all prefer outside consultants to developing public-sector expertise, which in the richer countries is ideologically associated with big government and in the poorer ones doesn’t exist due to problems with corruption. (China and Latin America are corrupt as well, but their heritages of inward-looking development did create local expertise; after the Sino-Soviet split, China had to figure out how to build subways on its own.)
But Israel Railways clearly has no domestic expertise in electrification. The political system is so unused to this technology that earlier this decade I saw activists on the center-left express NIMBY opposition to catenary, citing bogus concerns over radiation, a line of attack I have never seen in California, let alone the Northeastern US. Nor is Israel Railways good at contracting: the constant delays, attributed to poor contractor choice, testify to that. The political hierarchy supports rail electrification as a form of modernization, but Transport Minister Israel Katz is generally hostile to public transit and runs for office with a poster of his face against a background of a freeway interchange.
What’s more likely in my view is that Israel and New Zealand, with no and very little preexisting electrification respectively, invited experts to design a system from scratch based on best industry practices. I’m unfamiliar with the culture of New Zealand, but Israel has extensive cultural cringe with respect to what Israelis call מדינה מתוקנת (“medina metukenet”), an unbroken country. The unbroken country is a pan-first-world mishmash of American, European, and sometimes even East Asian practices. Since the weakness of American rail is well-known to Israelis, Israel has just imported European technology, which in this case appears easy to install, without the more particular sensitivities of urban tunneling (the concrete side of the electronics before concrete maxim). In contrast, the US is solipsistic, insisting on using domestic ideas (designed by consultants, not civil servants). Canada, as far as I can tell, is as solipsistic as the US: its world extends to Canada and the US; Schabas himself had to introduce British ideas of frequent regional rail service to a bureaucracy that assumed regional rail must be run according to North American peak-only practices.
All of this is speculation based on a small number of cases, so caveat emptor. But it’s fairly consistent with infrastructure construction costs, so long as one remembers that the scope for local variation is smaller in electrification and systems than in civil infrastructure (for one, the scope for overbuilding is much more limited). It suggests that North America could reduce its electrification costs dramatically by expanding its worldview to incorporate the same European (or Asian) companies that build its trains and use European (or Asian) standards.
Most of my thinking about public transit comes from large, dense cities, especially New York. In those cities, transit ridership is not a problem; only cost is. When such cities have decent cost control, they can build massive expansion programs, as Paris is. But most of the developed world is not New York, Paris, London, Tokyo, or other transit cities. A large and (thanks to differential national population growth rates) growing share of the developed world lives in fast-growing, low-density city regions with no public transit to speak of, such as the American Sunbelt and its counterparts in Canada and Australia.
I’ve had to intellectually grapple with public transit in two American Sunbelt cities in which current transit usage is a rounding error and the built form is wholly auto-oriented: Orlando (which I was asked about by a Twitter mutual) and Nashville (which just voted against a flawed light rail plan by an overwhelming margin). In those areas, there is no chance for any public transit, provided the urban form stays as it is – but fortunately such cities can leverage their high growth rates to change their urban form, as Calgary did in the 1980s and 90s.
Density versus growth
A few months ago I made this chart:
The density and growth demand axes are not meant to come from a single quantitative metric; density is a subjective mix of residential and job density, whereas growth demand refers to either population growth or the demand for more housing as expressed by price signal. In San Francisco, most likely the richest metro area in the world, density is middling, and growth demand is higher than even in New York and London; in the American Rust Belt, density is fairly low and there is also little demand for more; in some cities on the margin of the first world there is little demand for more growth but high preexisting density. It goes without saying that it’s easier to build new rapid transit lines on the upper right corner than on the lower left one.
The situation of the American Sunbelt, most of which goes in the bubble of Texas and Georgia, is difficult. Residential density is extremely low, so the ridership base near potential rail lines is low. Moreover, streets are usually designed exclusively around auto use, so passengers are unlikely to walk a kilometer to the train station the way they routinely do in transit cities. At the destination end, things aren’t much better – American cities have high-rise CBDs, but few jobs locate there or in surrounding dense neighborhoods.
The Orlando CBD has about 80,000 jobs, in a metro area of 2.5 million people. Disney World adds another 37,000, but is not surrounded by any serviceable residential neighborhoods, and has to be at the end of any reasonable transit service coming from the CBD. Nashville, a metro area of nearly 2 million, has a CBD with 36,000 jobs. The medical center to the southwest adds another 33,000, and this time it could plausibly lie on a rail trunk, but most of the useful urban arterials converge on the CBD and not on the medical center. In contrast, Washington, with 5.5 million people, has 280,000 people working at the CBD (from the Green and Yellow Lines to just beyond Dupont Circle and Foggy Bottom), 77,000 in the Rosslyn-Ballston corridor, and 33,000 in Crystal City and at National Airport and the Pentagon. Both the percentages and the absolute numbers (including job density) count: there is a great mass of people who would be interested in taking rail to Washington CBD jobs but not to Orlando or Nashville CBD jobs.
Can regional rail work?
High-growth areas are likely to have been small a few decades ago. For the most part the metro areas in question were too small in the heyday of rail transportation to have inherited a large legacy rail network. Even the ones that did, including Atlanta and Perth, have less legacy rail than older cities of comparable size – compare Atlanta with Philadelphia, or Perth with Brisbane. And most North American boomtowns are not Atlanta. Miami has two north-south mainlines and a handful of east-west connections, none at the right place for commuter rail. Orlando has a north-south trunk with a branch to the north, and Nashville a few branches, but they’re surrounded by industrial land use and not by the sort of suburbs that developed around commuter rail in the Northeast.
A commuter rail-based network can still work, but only with extensive greenfield lines. Disney World is not on any legacy rail line, because it developed long after rail stopped being a relevant mode of transportation outside large urban areas. But even then, gaps in coverage are unavoidable, as the dense neighborhoods of such cities did not develop around legacy rail.
Can transit-oriented development work?
The big question about TOD is, who is it for? In Nashville specifically, the far left opposed the light rail plan, essentially because it would cannibalize funding that could go to public housing. Now, public housing could be used to beef up density along rail corridors. Stockholm built public housing simultaneously with the subway, placing housing projects on top of rail branches, and as a result has per capita ridership today that’s not much lower than the level of Paris, Berlin, or Munich.
The problem is that public housing is horrendously expensive. A house in low-cost American cities costs around $150,000, but apartments cost more, so $200,000 per household is more likely even with some economies about size. Most of this cost is impossible to recover through rent – if low-income households made enough money to pay market rent in nice apartments, they’d just rent these apartments on the open market. The American Sunbelt does not lack for developable suburban land.
Market-rate housing is much easier to construct – for one, developers make a profit on it, and so are eager to put up their own money. The problem is that in cities like Nashville and Orlando, the middle class has close to 100% car ownership, and a large majority of households have one car per adult. The real estate industry is not going to spontaneously build housing with less parking or pedestrian-oriented retail.
In San Diego, developers build more parking than the minimum at University Avenue and 30th Street, according to Duncan McFetridge of the Cleveland National Forest Foundation. University is a bus corridor and not a light rail corridor, but the bus frequency there isn’t terrible, and the area is pretty walkable for a low-density city. In Los Angeles, I’ve read analysis that blames the region’s falling transit ridership on gentrification, explaining that in gentrifying inner neighborhoods like Boyle Heights, the middle class drives whereas the working class takes transit. It’s not like here or in New York, where recent gentrifiers rarely own cars.
How did Calgary make it work?
Calgary is a metro area of somewhat more than a million people. Its economy is based on oil, and when oil prices were higher earlier this decade its average income was comparable to that of San Francisco; its politics is thoroughly conservative, which means there is no progressive impetus for walkability or green transit. Nonetheless, it built light rail lines that get about 100 million annual riders today. Its transit mode share is 16%, higher than that of any American metro area except New York (or, in the most restrictive definition, San Francisco). This is with no residential TOD to speak of: the vast majority of housing in Calgary is single-family and low-density, and from what I’ve seen there’s almost no dense residential development near the stations.
The big thing Calgary did was develop its CBD to be high-rise. In the early 1980s Calgary was a small, monocentric city, and since then it’s grown more monocentric, developing downtown parking lots as high-rise buildings. When I visited it had a more prominent high-rise downtown than Providence, a bigger and older metro area, and walking between the high-rises was reasonably pleasant.
In low-density cities with demand for more growth, the best opportunity appears to be centralizing jobs in the CBD. The straightforward application involves developing parking lots, as in Calgary, and relying on the private market to do the rest of the job. In both Nashville and Orlando, there are also more proactive approaches, specific to their urban layouts. In Nashville, the high job density at the medical center calls for developing a continuous corridor from the CBD, about 3 km long. This corridor could plausibly get an east-west subway, in contrast with the north-south subway in the rejected light rail plan. In Orlando, the Disney World cluster calls for some residential upzoning and sprawl repair around that area, which would strengthen the case for building a rail line between that area and the CBD.
Growing cities can use their growth to support more auto-oriented development (as the big American cities did in the postwar era) or to support more public transit. This is understood in cities that already have a transit-oriented core, but it’s equally true in cities that don’t really have any public transit, like the entire American Sunbelt. Calgary, starting with very low population, managed to build a decent if not great public transit network centered on its light rail system, and the same should be doable in American cities of comparable size and age.
There are workplaces where most employees are high-income, for example office towers (or office parks) hosting tech firms, law firms, or banks. There are workplaces where most employees are working-class, for example factories, warehouses, and farms. Does this lead to a difference in commuting patterns by class? I fired up OnTheMap two days ago and investigated. This is American data, so it stratifies workers by income, education, industry, or race rather than by job class. I generated maps for New York and saw the following:
There are three income classes available, and I looked at the bottom and top ones, but the middle one, still skewed toward the working class, looks the same as the bottom class. The biggest observation is that Midtown is dominant regardless of income, but is more dominant for middle-class workers (more than $40,000 a year) than for low-income ones (up to $15,000, or for that matter $15,000-40,000).
The colors are relative, and the deepest shade of blue represents much more density for middle-class workers, even taking into account the fact that they outnumber under-$15,000 workers almost four to one. Among the lowest-income workers we see more work on Queens Boulevard and in Williamsburg, Flushing, and the Hub, but these remain tertiary workplaces at most. The only place outside Midtown, Lower Manhattan, or Downtown Brooklyn (which includes all city workers in Brooklyn due to how the tool works, so it looks denser than it is) that has even the third out of five colors for low-income workers is Columbia, where the low-income job density is one-third that of Midtown, and where there is also a concentration of middle-class workers.
The same pattern – job centers are basically the same, but there’s more concentration within the CBD for the rich – also appears if we look at individual neighborhoods. Here is the Upper East Side versus East Harlem:
I chose these two neighborhoods to compare because they exhibit very large differences in average income and are on the same subway line. Potentially there could be a difference between where East Siders and West Siders work due to the difficulty of crosstown commuting, so I thought it would be best to compare different socioeconomic classes of people on the same line. With the East Side-only restriction, we see two Uptown job centers eclipse Columbia: Weill-Cornell Medical Center in Lenox Hill at the southeast corner of the Upper East Side, and Mount Sinai Hospital at the northwest corner.
One place where there is a bigger difference is the definition of Midtown. Looking at the general job distribution I’d always defined Midtown to range between 34th and 59th Street. However, there are noticeable differences by income:
For the middle class, Midtown ranges from 34th to 57th Street and peaks around 47th. For the lowest-income workers, it ranges from 28th to 49th and peaks in the high 30s. My best explanation for this is that Midtown South and Union Square are more retail hubs than office hubs, featuring department stores and shopping centers, where the rich spend money rather than earning it.
In a deindustrialized country like the US or France, the working class no longer works in manufacturing or logistics. There are a lot of truck drivers today – 3.5 million in the US – but in 1920 the American railroad industry peaked at 2.1 million employees (source, PDF-p. 15), nine times today’s total, in a country with one third the population it has today and much less mobility. Manufacturing has plummeted as a share of employment, and is decreasing even in industrial exporters like Germany and Sweden. Instead, most poor people work at places that also employ many high-skill, high-income workers, such as hospitals and universities, or at places where they serve high-income consumption, such as retail and airports.
Since the working class works right next to the middle class, the nature of bosses’ demands of workers has also changed. Low-skill works now involves far more emotional labor; in Singapore, which makes the modern-day boss-worker relationships more explicit than the Anglosphere proper, there are signs all over the airport reminding workers to smile more. Nobody cares if auto workers smile, but they’re no longer a large fraction of the working class.
With the working class employed right next to the middle class, there is also less difference in commuting. For the most part, the same transportation services that serve middle-class jobs also serve working-class jobs and vice versa. This remains true even across racially segregated communities. The patterns of white New York employment are similar to those of middle-class New York employment, and those of black, Hispanic, and Asian employment are similar to those of the working class, with small differences (Asians are somewhat more concentrated in Flushing, and blacks in Downtown Brooklyn, reflecting the fact that blacks are overrepresented in public employment in the US and all city workers in Brooklyn are counted at Court Square).
This is true provided that opportunities for transportation are available without class segregation. This is not the situation in New York today. Commuter rail actually serves working-class jobs better than middle-class jobs, since Penn Station is closer to the department stores of 34th Street than to the office towers in the 50s. However, it’s priced for the middle class, forcing the working class to take slower buses and subway trains.
When I posted the above maps on Twitter, Stephen Smith chimed in saying that, look, the poor are less likely to work in the CBD than the middle class, so everywhere-to-everywhere public transportation is especially useful for them. While Stephen’s conclusion is correct, it is not supported by this specific data. In the $40,000 and up category, 57% of city jobs are in Manhattan south of 60th Street, compared with 37% in the $15,000-40,000 and under-$15,000 categories. It’s a noticeable difference, but not an enormous one. The reason Stephen is correct about how rides crosstown transit is different: people who can afford cars are very likely to drive if the transit option is not good (which it isn’t today), whereas people who can’t are stuck riding slow crosstown buses; in contrast, for CBD-bound commutes, the subway and commuter rail work reasonably well (especially at rush hour) and driving is awful.
Instead of trying to look specifically at low-income and middle- and high-income job centers, it’s better to just plan transit based on general commute patterns, and let anyone take any train or bus. This doesn’t mean business as usual, since it requires transitioning to full fare integration. Nor does this mean ignoring residential segregation by income, which in some cases can lead to transit segregation even in the face of fare integration (for example, the crosstown buses between the Upper East Side and Upper West Side have mostly white, mostly middle-class riders). Finally, this doesn’t mean relying on middle-class transit use patterns as a universal use case, since the middle class drives in the off-hours or to off-CBD locations; it means that relying on middle-class transportation needs could be reasonable. It just means that the rich and poor have substantially the same destinations.
An even bigger implication relates to questions of redevelopment. There have been periodic complaints from the left about gentrification of jobs, in which working-class job sites are turned over to high-end office and retail complexes. For example, Canary Wharf used to be the West India Docks. In New York, Jane Jacobs’ last piece of writing before she died was a criticism of Greenpoint rezoning, in which she specifically talked up the importance of keeping industrial jobs for the working class. But since the big deindustrialization wave, developments brought about by urban renewal, gentrification, and industrial redevelopment have not had any bias against providing employment for the poor. It’s not the factory jobs that the unionized working class still culturally defines itself by, but it’s industries that are hungry for low-skill work, and in many cases are serious target of unionization drives (such as universities).
The Regional Plan Association has a detailed regional rail proposal out. It’s the same one from the Fourth Plan that I’ve criticized here, on Streetsblog, and on Curbed, but with more explanation for how the service should run, with stopping patterns and frequency.
There are some good aspects there, like a section about the importance of electrification and multiple-units, though it stops short of calling for full electrification and replacement of locomotives with EMUs; the focus on off-peak frequency is also welcome. There are also bad ones, like the claim on p. 32 that it’s difficult to impossible to provide through-running using the existing Penn Station tracks used by New Jersey Transit. Foster Nichols told me that there are some difficulties with grades but they should be doable if NJT commits to an all-EMU fleet, and reminded me that the ARC studies judged through-running using these station tracks and new tunnels feasible. What he expressed to me as a difficulty turned into a near-impossibility in the report, in order to justify the $7 billion Penn Station South project.
But I want to focus on one particularly bad aspect of the proposal: the stopping patterns. The RPA is proposing three distinct stopping patterns on pp. 32-45, with three separate brands: Metro, in the city and some inner suburbs; Regional Express (RX), in the suburbs; and Trans-Regional Limited (TRL), providing intercity service to New Haven, Ronkonkoma, Philadelphia, and other major stations outside the built-up area. Even as the plan talks about the importance of making sure suburban trains serve urban stations in order to give them frequent service through overlay, the stopping patterns suggest the opposite.
The proposal involves trains from the suburbs expressing through most city stations (including the infill) even on two-track lines, like the Port Washington Branch. Metro trains would make all current stops plus additional infill to Bayside, and RX trains would only serve Willets Point, Flushing, and Bayside, and then run from Bayside to Port Washington. A similar pattern happens from Jamaica to Valley Stream, resulting in the Babylon, Long Beach, and Far Rockaway Branches all having to share a track pair. Moreover, the RX trains may themselves be divided into local and express trains, for example on the New Haven Line.
This is bad practice. On a two-track line, there’s no real reason to skip a handful of inner stations just to guarantee the outer ones express service. If anything, the need to schedule trains on the same tracks would lead to more fragile timetables, requiring more schedule padding. My analysis from 2.5 years ago found that the LIRR Main Line is padded 32% and the Babylon Branch is padded 19%: that is, the scheduled travel time on the Main Line (up to Ronkonkoma) is 32% more than the travel time imputed from line speed limits and current fleet acceleration performance. Patrick O’Hara, who the RPA study even quotes as a source elsewhere, investigated this issue separately, looking at best-case timetables, and found that some runs are padded 40-50%.
In Switzerland, trains are padded 7%, and I’m told that in Japan, after the Amagasaki accident showcased the safety problems of overly precise schedules, pads are about 5%. Express trains and locals mixed on the same line make it harder to maintain tight enough reliability for low schedule padding; this way, on an all-local line, trip times may match those of express trains on mixed lines, as they do in my analysis above. The best analogy is the RER B going to the north: the express trains are 4 minutes faster than the local trains, skipping 9 stops. The stop penalty on the RER B is higher, closer to 7 minutes over 9 stops, but the shared tracks with local trains (and with the RER D between Gare du Nord and Chatelet-Les Halles) means that there’s a fudge factor in the schedule, so it’s not possible to reliably do better than 4 minutes, and the trains end up visibly crawling on the mainline.
The reader familiar with technical transit advocacy in the Bay Area may interject, what about Caltrain? Clem Tillier has no trouble proposing timetables mixing local trains, express trains, and high-speed rail on the same track pair with timed overtakes, and a 7% pad. So why am I down on this concept in New York? The answer is line complexity. Caltrain is a simple two-track back-and-forth, and HSR is generally more punctual than legacy trains because it runs for long stretches on high-quality dedicated tracks, so it’s unlikely to introduce new variability to the line. In contrast, the RPA plan for regional rail in New York involves extensive branching, so that train schedules depend on trains elsewhere on the line. In this case, introducing more complexity through local/express sharing is likely to require more schedule padding, erasing the speed advantage.
In general, my questions to establish guidelines for where express trains are warranted are,
- How long is the line, measured in the number of stations? More stations encourage more express trains, because more stations can be skipped. In higher speed zones, stop penalties are higher, but at equal line length measured in km, higher speeds and fewer local stations reduce the benefit of express trains.
- How frequent are trains? At low frequency, local stations need more frequency, so express runs are less useful. At very high frequency, there may not be capacity for different stopping patterns unless the line has four tracks. On a two-track line, the optimum frequency for a local/express alternation is about 6-12 trains per hour, 3-6 local and 3-6 express, with a single mid-line overtake. Multiple overtakes on a single line are possible, but more fragile, so they are a bad idea except in special circumstances.
- What is the demand for travel? Express trains work best if there are a few distinguished stations at regular intervals, or else if the line is long and there is strong demand at the far end; if the inner stations are very strong then it’s more important to give them higher local frequency. When performing this analysis, it’s important to make sure station ridership levels reflect genuine demand rather than service. For example, Caltrain express stops have high ridership in large part because of their better service, not nearby density, as shown in Clem Tillier’s analysis. The LIRR Main Line has far more ridership at Mineola and Hicksville than the other stations on the trunk and also far more service, but Patrick explains that this is due to better highway access, so it’s genuine demand and not just a reflection of better service.
Caltrain needs express service because it has about 20 stations between San Francisco and San Jose, depending on the amount of infill and anti-infill desired; a target frequency of 8-10 peak trains per hour; and strong demand on the outer stations, especially for reverse-peak trips. In New York, none of the two-track lines meets the same standard. Some are too short, such as the Port Washington Branch. Others are too busy, such as the Harlem Line, Babylon Branch, and LIRR Main Line. Yet others have too much demand clustering in the inner stations, such as the Erie lines and the North Jersey Coast Line.
On four-track lines, it’s always easier to run express service. This doesn’t mean it should always be run: the upper New Haven Line is a strong candidate for relegating all commuter trains to the local tracks, making all stops, giving the express tracks to intercity trains. The Northeast Corridor Line in New Jersey is a dicey example: past Rahway there are four tracks, but intercity trains could run at very high speeds, making track sharing on the express tracks difficult. My service pattern map has express trains skipping Edison and Metuchen, but it’s just two stations, making it better to just run local beyond Rahway to clear the express tracks for high-speed rail.
It’s tempting to draw proposals involving intense metro-style regional rail service only serving the urban and inner-suburban stations; I’ve had to argue against such plans on some MBTA lines. The problem is that trains from the outer suburbs are still necessary and still going to pass through the inner suburbs, and in most cases they might as well stop at those stations, which need the frequency more than the outer suburbs need the few minutes of speedup.
Thanks to the invitation of Adina Levin’s Friends of Caltrain, I came to the Bay Area for a few days, giving two talks about regional rail best industry practices in front of Friends of Caltrain, Seamless Bay Area, and SF Transit Riders. My schedule was packed, including a meeting with a world-leading expert in transportation systems in which I learned about the Barcelona bus service changes; a Q&A about both general and Bay Area-specific issues for Bay City Beacon subscribers; and several other meetings.
My two talks were in Mountain View and San Francisco, and used the same slides, with minor corrections. Here are the slides I used in San Francisco; I consolidated the pauses so that each page is a slide rather than a line in a slide. This was also covered in Streetsblog, which gives more background and gives some quotes from what I said during the presentation, not printed in the slides.
Unlike in my NYU presentation last year, I did not include a proposed map of service improvements. The reason is that in the Bay Area, there are more questions than answers about which service should use which piece of infrastructure. This makes fantasy maps dangerous, as they tend to fix people onto one particular service pattern, which may prove suboptimal based on decisions made elsewhere in the system. This is not a huge problem in New York or Boston, where the alignments naturally follow where the stub-end commuter lines are, with only a few questions; but in the Bay Area, the situation is more delicate, because big questions like “can/should Caltrain get a trans-Bay connection from San Francisco to Oakland and the East Bay?” can go either way.
The role of redevelopment in the area is especially important. Unlike New York, Paris, or other big transit cities, San Francisco does not have much density outside the city proper, Oakland, or Berkeley. Moreover, there is extensive job sprawl in Silicon Valley, which contributes to a last-mile problem for public transit; usually first-mile access is a bigger problem and high-end jobs tend to cluster near train stations. But conversely, the high incomes in the Bay Area and the growth of the tech industry mean that everywhere TOD is permitted in the core and the suburbs, it will be built. This impacts decisions about the total size of transit investment.
At present-day development patterns, San Francisco proper really doesn’t need more rail construction except Geary and the Downtown Extension, depending on construction costs. But general upzoning makes Geary worth it even at $1 billion per km and opens up the possibilities of four-tracking Caltrain within the city (which means expanding some tunnels), giving the N-Judah dedicated tracks and a dedicated tunnel under Mission, and extending the Central Subway to the north and northwest. The land use in the Richmond and Sunset Districts today supports a fair amount of transit, but not new subways if there’s no cost control.