Frequency is Relative
Five years ago, I wrote a blog post about frequency-ridership spirals, mentioning as a side comment that the impact of mass transit frequency on ridership can be lumped together with the trip time. I’d like to develop this point here, and talk about how it affects various kinds of public transportation, including intercity trains.
The rule of thumb I’ve advocated for in ETA reports (for example, on commuter rail) is that the maximum headway should be no more than half the trip time. Untimed transfers reset the clock, since passengers have to wait another time every time they make such a transfer; timed transfers do not, but are rare enough that local public transportation doesn’t usually need to consider them in service planning. Intercity transportation can follow the same rule of thumb, but can also get away with worse frequency since passengers time themselves to arrive shortly before the train does; in particular, hourly trains between cities that are three hours apart are frequent enough that increasing service is valuable only insofar as it provides more capacity, and is unlikely to lead to higher ridership through shorter waits.
Wait and transfer penalties
In the literature on modeling public transportation ridership, it is universal that passengers prefer spending time on board a moving vehicle to waiting for a vehicle, walking to the station, or walking between platforms. This preference is expressed as a factor, called the waiting or transfer penalty. different models have different levels for these penalties; passengers also likely have different penalties depending on circumstances, such as familiarity with the route or how much luggage they’re carrying.
The papers I’ve seen have penalties ranging from 1.75 (in the MTA’s model) to 3 (the higher end cited in Lago-Mayworm-McEnroe). I usually model with 2, in Teulings-Ossokina-de Groot. The factor of 2 has the advantage of consistency with an assumption that passengers don’t have a wait penalty but do assume a worst-case scenario for waits, so that the generalized travel time is equal to the maximum headway plus the in-vehicle travel time.
Update 4-19: I was just alerted to a new study by Yap-Wong-Cats using London ridership, finding an out-of-vehicle penalty factor of 1.94 pre-pandemic and 1.92 post-pandemic.
The impact of frequency relative to trip time
If the elasticity of ridership with respect to the generalized travel time, summing the headway and in-vehicle travel time but not walking time to and from the station, is e, then we can compute the elasticity of ridership with respect to frequency as a fraction of e. If the current headway is a proportion r of the in-vehicle trip time, that is to say a fraction r/(r+1) of the generalized travel time, then the elasticity is er/(r+1).
In Lago-Mayworm-McEnroe, the value of e appears to be 0.8. This means that if r = 0.5, the elasticity of ridership with respect to frequency is 0.267. The paper doesn’t quantify elasticities relative to different levels of r but only relative to absolute frequencies, but 0.267 is within the range it finds for different frequencies, dipping to 0.22 for high-frequency lines. Other papers have different figures of e, often higher in the long run as passengers adjust, but those go up to around 1 or, reasoning backward from a VTPI report, a little higher.
Of notes:
- The value of r is not constant across different uses of the same line. A commuter traveling from near the outer end of a subway line to city center faces much lower r than a traveler going a short distance, within city center or within a large outlying neighborhood. In particular, r is generally lower for commutes than for non-commute trips, which is why the latter are more sensitive to frequency.
- Systems that rely on extensive transfers can have very high values of r with short in-vehicle trips. New York averages 13.5 minutes per unlinked subway trip, with many trips facing an effective off-peak headway of 10 or even 12 minutes, at which point e is high enough that increasing off-peak frequency could pay for itself through higher paying ridership (see analysis in a blog post and an ETA report). This, again, depends on the type of trip – commuters may pick an apartment or a job based on ease of travel, reducing the need to transfer, but their non-commute trips are usually a collection of irregular trips to various destinations and are likelier to involve a transfer.
- The cost of higher frequency depends on mode (it’s higher on buses than trains) and time of day (it’s very low on off-peak trains until it matches peak frequency). Together with points 1 and 2, this argues in favor of raising the off-peak frequency on urban and inner suburban trains, potentially to the point of matching peak frequency. On longer-range commuter trains, the impact of frequency on ridership is lower, and thus the marginal cost may be such that a ratio of peak to off-peak service larger than 1 is desirable.
Intercity trains
The papers I’m citing aim to fit elasticity factors to observed ridership on local and regional public transportation. Intercity rail has its own set of models, with different assumptions. Frequency again matters, but because passengers time themselves to arrive at the departing station shortly before the train leaves, its impact is reduced.
I don’t know the elasticity of intercity rail ridership with respect to frequency; Cascetta-Coppola have the elasticity of ridership with respect to in-vehicle trip time as about -2, while Börjesson has it at -1.5 for business travel and -1 for non-business travel with a rough rule of thumb trying to approximate the impact of frequency. At the level of the sanity check, the low frequency of TGV services is not visible in TGV ridership between the provinces and Paris, compared with Japan (which charges higher fares) and as far as I can tell from a few data points Germany. TGV ridership between the provinces is bad, but that involves trains with service gaps that are much larger than the trip times, reaching six hours between Marseille and Lyon. In contrast, those three-hour gaps in service between Paris and cities three hours away by TGV don’t seem to impact ridership visibly.
What this means is that intercity trains do need a certain baseline frequency. The German system of a train every two hours on every city pair is wise, in light of the typical intercity rail travel distances in a large country with slow trains. Higher frequency is warranted if the cities are bigger and therefore require more service, or if they are closer together in time through either a short geographical distance or higher speeds. New York and Philadelphia are about 1:10 apart by rail, and high-speed rail could cut this to about 45 minutes; half-hourly frequencies in the current situation are sufficient that more service would have a second-order effect, and even with high-speed rail, a train every 15 minutes is more than enough for all purposes except capacity (the current offer is 3-4 trains an hour with irregular spacing). Frequency is freedom, but this depends on trip times; what works for four-station subway trips is not what works for trips between cities 140 km apart, let alone 360 km, and vice versa.
Pedestrian Observations 
for buses where passengers on board wait for feeder bus to connect to and from, is this considered trip time or wait time?
I don’t know, the meta-studies I’m citing are using single-corridor routes rather than timed connections.
I think it should be considered trip time, for two reasons.
There are undoubtably limits to this, if a hypothetical NEC HSR went DC to Providence in 3 hours, then sat and waited two hours before going to Bos, I’m sure there would be lots of complaints and bad press, not “increased ridership commensurate with reducing trip time from 6.5 to 5 hours.” But your ordinary stops to let expresses pass or stops to allow connections at takt knots or pulse points I think are perceived as trip time to passengers that do not have to transfer.
In my experience as a daily transit rider, on a system that doesn’t coordinate transfers well (Baltimore): Time on board the feeder bus should be counted as trip time, but the time waiting between buses/trains at the transfer point should count as wait time. Waiting at the transfer point for a second vehicle just replicates the experience of waiting at your origin point for the first vehicle. Its not really the same as being on the bus while it makes stops, b/c when you’re in vehicle you know with certainty you’re on your way. Waiting at a stop/station, you don’t know for sure when or if the vehicle will arrive and get you moving, until it arrives. I suspect that uncertainty and lack of motion towards the rider’s goal is what accounts for the perceptual difference between wait penalties and trip-time penalties. Rider surveys could probably uncover whether that’s so. And these time-penalty variables are just a method of quantifying riders’ subjective perceptions, yes?
Relationship between ridership and frequency heavily relys on what alternatives are there. If I don’t have much alternative options, I will be more likely to stick to an option which reduces its frequency.
I noted this years ago and concluded that suburbs need more frequent transit options than cities. In places like Manhattan driving is painful enough that people can put up with 12 minute head ways on the subway (they shouldn’t have to, but they will – plenty of others have commented on this). Out in the suburbs driving isn’t painful so if transit isn’t frequent you will just get in the (leaving out of course those few who cannot drive)
Of course the lack of density in suburbs means it is hard to afford to run the frequent transit needed to get people to ride. I don’t have a solution to this problem. (the keys are of course figure out how to build fully automated transit for a lot cheaper than even Spain can)
Seek for a lower breakeven point by reducing vehicle size, or cheaper labor, or slightly increasing the fare (provided that you keep a good frequency). Other methods to increase ridership help as well, like making use of micromibility as it could expand coverange of each transit station significantly.
Wow, if only every clueless person on the planet hadn’t been suggesting the exact same repeatedly-historically-proven-failure Innovative Ideas since, oh, 1970 or so. Maybe 1960 for all I know.
Micromobility! This time for sure!
PS you forgot “call on demand” and “flexible routing”. But one can’t expect all bad ideas to appear all at once in just one comment … unless adirondacker3k is posting, that is.
Besides the association of poverty, the big reason why people hate buses is that bus routes are easier for humans or algorithms to mess with. The more fixed infrastructure of trains doesn’t allow planners or algorithms to suddenly decide that Line A running on Corridor C for the past 25 years should suddenly be moved to Corridor G.
The problems with micromobility have been written about extensively (https://humantransit.org/category/microtransit is a good starter for those who might be unaware, the major commenters here seem to I’m just going to assume you know).
In order to make micro transit work, it needs to have the following metrics. These metrics apply regardless of group size,
From the moment someone initiates the trip (dials the first number or opens the app) until the vehicle arrives at the pick up point can be no more then 5 minutes + walking time. Note that walking time will be very different for a teen on the track team (who will use a stopwatch to try to beat their previous record to get to the stop), vs an elderly person who needs a walker to help them move – but you get the same 5 minutes + walk time for both. Also note that the teen can dial numbers and speak; or use an app much faster than the elderly person, but you start your timer from when they first make an effort to start the trip, not from when you get the complete information of where they are and where they want to go. (this means extensive UI work)
If someone takes the same trip 100 times, starting at random times of the day, their total trip time (end to end time, often called linked trip time) needs to be the same to plus or minute 3 minutes. You are allowed a few outliers (to account for emergencies), no more than 2 per individual and statistically less common than that.
Good luck meeting those. Even more luck is needed to afford to run that system. I’m setting minimum expectations, if you can’t meet them: we are talking about suburbs, most people can drive and so they will.
adirondacker3k
Is this a Mystery Science Theater reference?! If so tip of the hat to you.
They aren’t even running service every 12 minutes in the suburbs to be fair.
I don’t think it needs to be better than the city, but it does need to be decent.
12 minute service would help, but there are limits. People who work in the suburbs won’t even think about it – they don’t deal with traffic, plus their offices are probably not on the line and your transfers to the correct line mean total trip times are probably unacceptably long. People going downtown to work will accept 12 minute service – they already accept 30 minute, but work trips are easy to schedule around transit schedules (the factory can be asked to change start times to match transit, while the office will allow you to set your start/stop time based on when transit will get you there). Anyone going shopping, to church, their softball game, and so on needs tighter schedules, and 12 minutes really don’t fit. Besides, those are off peak times and so there is less likely to be traffic.
I think leisure trips can be flexible to an extent around public transport too.
And additionally particularly late night driving isn’t great as it stops you drinking and requires concentration when you are tired.
@ Matthew Hutton
The idea that you should rearrange your life around someone else’s schedule hurts transit. If you can’t figure out how to make transit convenient for humans who have things to do then admit it and tell those people transit isn’t for them, just drive. In the suburbs this might be the only possible answer (because of low density), but be honest about it. Don’t ask people to arrange their life to how you want/think they should live. Either meet them as their are or give up on them.
You can of course make dense areas where transit can run frequent better places to live. There are a number of things that make suburbs better than dense cities. Many of them are things that have nothing to do with low density build forms. Where people live is a compromise and often high density living involves a lot of compromises that make it not a useful option. Start correcting them.
Some issues to look at: Why are high density schools nearly universally bad? Why do dense apartments rent for a lot more than a house payment for a much larger house in the suburbs? Why are dense apartments mostly 2 bedroom, and never more than 3? Why do dense parks focus on adults (art) not playgrounds (kids)? Why is entertainment in dense areas focused on young adults (bars, R rated movies, live theater, life music – though at least the symphony attracts older adults). Nothing wrong with things for adults, but dense cities are clearly attracting the rich and unmarried adults just out of college, move to the suburbs when you get married. Note that the above list is very US centric – there are plenty of examples of dense areas in Europe and Asia where dense areas are attractive to other demographics.
do agree that late night driving isn’t fun. However most transit systems don’t run late at night. Often those that do run with reduced schedules that entice people to drive anyway.
There is a system that delivers on-demand low volume transportation. Automobiles. It’s going to be automobiles in places designed with the assumption that everybody will be using automobiles.
There is a system that delivers on-demand low volume transportation. Automobiles.
Adirondacker is entirely correct. I have said several times that if you look at cars as as a type of “transit”, their salient features are 1) infinite frequency (equivalent to zero headway – the vehicle is always waiting at the ‘station’ when the passengers arrive) and 2) infinite span (runs 24/7/365).
As much as places like this blog debate things like takt and transfer penalties and stopping patterns, in many cases the largest single factor in transit ridership is simply the level of transit service run. If you run service hourly instead of every two hours, if you run the weekend service pattern on Sunday instead of nothing, if you run service until 10pm instead of stopping at 8pm, in all cases you get more riders. Automobiles in this sense “run more service” than any transit system could ever hope to, except in large dense urban areas. Large dense urban areas are also where cars move slowest due to congestion. Where each mode is most appropriate should be clear, and this is why in every country that is rich enough, cars have the highest mode share, even in countries that have excellent transit of all types or very high non-car-mode-share in urban areas.
Isn’t part of the issue with the situation that because car-centric development paradigms are so unavoidably infrastructurally intensive that, at any but truly rural scale, it’s expensive enough that people would balk at fair pricing for it? This is obviously the argument behind the “suburban Ponzi scheme” that the Strong Towns people talk about, and data by Urban3 seems to bear this out.
Yes, lower-density development built around the there-whenever-you-want-it “transit” of car ownership is nice, but people would be a lot less happy to use it if they had to pay the actual cost of it, and it’s so geometrically inefficient that it never really can be accommodated at anything more than suburban scale, anyway. Car sprawl is financed by more car sprawl until the lifeblood is sucked out of a city; it’s never really been sustainable, it’s just that a country large and rich enough can do it for a long time before the game has to end.
Ergo, the solution is to put as much as possible within a reasonable distance of the there-whenever-you-want-it “transit” of walking/cycling, and use actual mass transit modes for anything further away.
That, at least, seems to be the argument (though I can’t deny that I, personally, find it persuasive).
Isn’t part of the issue with the situation that because car-centric development paradigms are so unavoidably infrastructurally intensive that, at any but truly rural scale, it’s expensive enough that people would balk at fair pricing for it?
Auto mode share for a few selected EU countries:
Do you believe that these countries pursue ‘car-centric development paradigms’? Are these countries all ‘truly rural scale’? Do these countries subsidize car ownership and gasoline instead of taxing them heavily?
The answer to all of these questions is of course no. Yet despite many of these countries being considered paradigms of social welfare and equality (Norway, France), high population density (Netherlands), quality bike/pedestrian focused urbanism (Netherlands, Denmark) and extensive high quality transit (France, Netherlands) car is the dominant mode share by far. Only Norway might be considered truly rural at a national scale, however the Netherlands is easily the densest country in Europe that is not a city state and has almost equal car use.
The fact is that cars work at many areas beyond just ‘truly rural scale’ – rural , exurban, suburban, and many smaller or lower density urban areas.
people would be a lot less happy to use it if they had to pay the actual cost of it
Given that in almost every country money is taken by the government in the form of gasoline taxes and then given to transit in the form of subsidies, I have little tolerance for the ‘drivers don’t pay their way’ bit. There may be a valid argument that no/very few forms of transportation pay their own way, and that transportation is subsidized by other taxes that come from the economic activity transportation enables, but in most places transit does less than cars. In many places transit fares do not pay the operating cost of the service, while not contributing a cent to capital cost, while the argument against cars paying their way always revolves around trying to point out that they don’t pay the full capital cost. If transit ticket costs were raised so that farebox recovery was 100% and all line extensions were paid for by debt serviced by the transit agency I guarantee you people would be a lot less happy to buy those tickets.
the argument behind the “suburban Ponzi scheme” that the Strong Towns people talk about
Strong Towns has a number of good points:
The “suburban Ponzi scheme” argument of theirs is absolute bunk, however, for a number of reasons:
I think we all agree that mass transit doesn’t scale down to rural (ie farms) areas. While cars do not scale up to very dense places (Manhattan). The question is what happens in the middle.
There are a number of problems with cars. I won’t list them all (I’d be sure to miss one if I tried), and you make place different wights on different issues. However there are a lot of people that would find their lives improved if they could/would substitute some arbitrarily good transit for their car. So the question is what can we do to bring the benefits of transit to more people is my question.
I think that with world leading (that is beat Spain) construction costs we could extend a metro system out to the suburbs and par for it in part by all the money we save on smaller highways – but this of course has never been proved in the real world. (suburbs can use elevated metro systems they don’t need to dig expensive tunnels – but is this enough – I don’t know)
Does road infrastructure really last that long?
Maybe on paper it does, but when you take a walk around, it’s clear that road infrastructure in the US is in a very sad state of neglect and disrepair. There are pot holes everywhere even in regions with easy climates like Silicon Valley, and the remaining road surface is uneven with massive deep puddles in the rain. The sidewalks tend to be cracked, uneven, and overgrown. Snow clearing (where it applies) is poor, and snow clearing for anything other than the carriageway manages to be even worse. And everything is covered in a ton of litter.
And non-road infrastructure also suffers due to how spread out everything is. The reliability of my utilities in SF was atrocious, with nicer-end-of-third-world-country tier outages of electricity and internet, and occasional outages of water and sewer. The typical year for PGE is literally worse than the typical year compared to major cities developing countries like Bangkok, and worse than absolute terrible years for cities in developed countries such as Tokyo in 2019. And the roll out of new utilities is also terrible, with the US having slow, limited, and expensive internet, both landline and cellular.
Significantly poorer countries than the US can afford to build and maintain much higher quality infrastructure, because they don’t have to build and maintain as much of it.
So in that sense, Strong Towns is right in pointing out the unsustainable nature of car oriented infrastructure, but the expected failure mode isn’t municipal bankruptcy, it’s just low quality infrastructure.
Which I guess if you ask a typical US suburbanite whether they’d prefer to live in a dense neighborhood with near pristine streets, reliable electric/water/sewer, and fast reliable internet, or what they currently have, many would choose the status quo.
@Sassy
That sounds like section bias – you are seeing problems and not noticing all the areas where infrastructure is good. All roads need maintenance, one pothole every 100 feet or so every spring needs to be fixed, but the road itself is good for 25 years if you fix them. Sidewalks crack, and so maybe on small section should be replaced here and there early, but the whole is good for 60 years. Likewise a pipe might spring a leak when the local earth moves, but the rest of the pipe will be good for much longer, but eventually the pipe breaks down and you should replace it all.
Where I live the city made the intentional decision to under build a local road when they upgraded from gravel. They know they will have to replace it in 10 years because of that – but they also know that in 5 years all the corn fields will be turned into a subdivision and so after 10 years the road will not see as much heavy construction traffic and so the road to last for 25 years today is more expensive (ignoring inflation) than the road in 10 years to last for 25 would be (they are also assuming no bus service which beats up roads worse than construction). This is sometimes the correct decision, but this road clearly will bring down an average.
If you do a proper statistical sample you discover that the infrastructure overall lasts – with proper maintenance – as long as stated. Often you can still use it pasded the end of life in degraded form. However if you just look around you won’t see the true age/state.
@henrymiller74
Small sections clearly fail earlier than others, but for the infrastructure as a whole to be good, it should have vanishingly few failed sections. Most of the sections still being good is not enough.
If a car drives by when I’m walking past a deep puddle, I get wet, regardless of whether the rest of the road was okay. One deep puddle in the carriageway has effectively crippled pedestrian usage of the sidewalk in the rain over a much longer stretch of road.
I think focusing on the failures being only in small sections makes the problem appear worse, not better. The US is struggling to even repair small broken sections of infrastructure, what will happen when the rest of it finishes crumbling?
It’s possible to fix the road surface before the surface becomes uneven enough for deep puddles, much less deep potholes. There’s just too much road surface, too much damage to the road surface, and/or too little money for that to be feasible in the US, and the US has plenty of money.
It should be possible in any developed country to keep potholes to a minimum. We certainly used to be able to manage.
And products are available to do permanent pothole repairs in inclement conditions such as https://tarmac.com/products/asphalt/ultipatch-pothole/ and https://tarmac.com/products/asphalt/ultipatch-viafix-quick/ or https://www.instarmac.co.uk/news/casestudy/kier-record-82-reduction-in-repeat-visits-thanks-to-ultracrete-tough-patch/ are available.
Certainly in the Bay area it should be very easy as it doesn’t freeze and taxes are very high.
And the roll out of new utilities is also terrible, with the US having slow, limited, and expensive internet, both landline and cellular.
According to Ookla (https://www.speedtest.net/global-index) the US has the 5th fastest broadband speed in the world, and the 9th fastest mobile internet speed, and has gone up in both rankings in the past year. Why do people keep making statements that are so easily factually disproven?
The reliability of my utilities in SF was atrocious . . . . Significantly poorer countries than the US can afford to build and maintain much higher quality infrastructure, because they don’t have to build and maintain as much of it.
Using the quality of infrastructure in SF in this conversation is illogical on two counts. First, SF’s built environment is exactly what Strong Towns recommends and the exact opposite of the suburban/car focused paradigm that will supposedly lead to underfunded and poor infrastructure. SF is a very dense city that accordingly needs less infrastructure per person as almost anywhere in the US outside of NYC.
Second, SF is fabulously wealthy by any standard, US, international, or developed country. The city of only ~800k has a budget of $14.6 Billion. SF “can afford” to build an maintain whatever it wants – if quality of services there is poor its because the city is mismanaged, not because the built environment requires too much infrastructure.
The electricity and the telecommunications are provided by private companies most places. So San Francisco’s municipal budget is only vaguely related to the quality of service provided by private companies.
We were bamboozled by the glory of the private markets and things like cell phone service and internet/data are unregulated. And you can get Green Stamps and Plaid Stamps with many of the plans offered.
“Are these countries all ‘truly rural scale”
Of course this will depend on your definition of “rural”, but I’d argue yes, because in only one of these countries is there a single city with over 3 million people in its entire metro area (Paris), and that metro area’s population is barely 1/3 the population of the entire country (France), and in that country, the core city of the next largest metro area (Lyon), there’s not even a million people. These are arbitrary standards, sure, but I’ll remind you that in much of Asia–and in the future, much of Africa–there are many, many cities which are far larger than this, and often have high mode shares for transit. E.g., about half of Japan and South Korea’s populations reside in just three cities (Tokyo/Osaka/Seoul), and these all have transit mode shares around or slightly above 50%, with cycling/walking putting the non-car share at over 75%.
Sure, I’m cherry picking an extreme set of examples, but:
a.)There just isn’t any denying that in terms of pure resource consumption per person, transit/walking/cycling development paradigms consume fewer resources than car-centric ones
b.) There’s no reason to privilege any type of mode of transit in the abstract; what can provide the most “transit” (including, potentially, cars) to the most people–sustainably, over time–is the right one to promote. Because (a), and the fact that it seems pretty clear at this point that we can’t afford U.S. or probably even EU levels of car ownership/usage for everyone in the entire world in any case, there’s no reason to privilege driving for some groups of people and not others. On the contrary, there are good reasons to penalize car use for exactly that reason, because the infinite frequency/infinite span of cars are only available to people who can afford them, which isn’t everyone, and there’s no really just way to decide who gets access.
“In many places transit fares do not pay the operating cost of the service, while not contributing a cent to capital cost, while the argument against cars paying their way always revolves around trying to point out that they don’t pay the full capital cost“
Okay, but in the most car-centric places cars don’t pay the operating cost of their infrastructure with user fees, either. There are only a handful of U.S. states which pay for road building/maintenance with at least a little bit of money from the general funds, and because of our (a) up there, there are just fewer positive externalities possible. Higher resource investment=higher bar for there being a truly positive externality, vs. an arguable external defrayment of internal cost.
“Strong Towns talks about infrastructure costs bankrupting communities, but this hasn’t happened. In all the instance of municipal bankruptcy in the US in recent years, the culprit has always been pensions, not infrastructure spending. I think in one California town from a few years back, 70% of the whole budget was going to police and fire pensions, or police and fire personnel costs (salary+benefits+pensions). If anything it goes the other way, communities have to cut infrastructure spending when other payment commitments get too high.”
First of all, I don’t think I need to tell you that “it hasn’t happened yet, therefore it’s not going to happen” is not a very logically-sound argument. Especially in a country with loads and loads of land to pull into a hypothetical Ponzi scheme of development, and a growing population which has been sold on the idea that single-family-home ownership in American car-sprawl is the “dream”, it’s entirely likely that the scam just hasn’t run out of juice yet. I can’t tell the future, so maybe it won’t, but neither can you, so maybe it will.
Second, I think that at the very least prima facie, people are generally more willing to cut infrastructure spending than emergency services/elder care (which pensions are arguably a form of) because looking for efficiencies in the latter often has direct human consequences, where the former can very easily not. As the world begins ageing evermore, it’s going to be all the more important to look for efficiencies in humane ways specifically to provide more funding for social spending. Not that no efficiencies can/should be found there, but the first recourse to enhancing or motivating enhancements of efficiency probably shouldn’t ever be pure cost-cutting. We need to figure out what money’s being wasted doing what things, rather than just saying “this number is too big; make it smaller”.
“Viewed over these lifetimes the cost to maintain infrastructure is much less and can be supported by the houses fronting it.”
The thousands and thousands of empty Rust Belt homes would like a word with you. Not to mention that with the inequitable way property taxes have worked at least in the U.S.–sub-market valuations on higher-priced, typically more infrastructurally-expensive homes leading to undertaxing those most able to pay; redlining forcing POCs to stay in neighborhoods which are assessed over market values ergo being taxed at higher rates; pooled municipal/state taxes causing higher rents in urban areas to essentially subsidize suburban services (this isn’t just Strong Towns’ prevarications; Urban3 has done the work)–incentivizing the most universalizable (i.e., “most good” to most people at lowest cost) development paradigm is the correct direction to move in.
The first Streetcar suburbs were build in the 1880s. In that time they have replaced their roads 2-3 times (best case, likely more often), added electric and telephone (both of which have been replaced at a couple times), many didn’t have sewer and water so those were added as well, though either way it is highly likely those pipes have been replaced several times as well.
They have in short been able to pay to repair and replace their infrastructure over enough history to see a pattern.
in only one of these countries is there a single city with over 3 million people in its entire metro area (Paris), and that metro area’s population is barely 1/3 the population of the entire country (France), and in that country, the core city of the next largest metro area (Lyon), there’s not even a million people.
The Randstad in the Netherlands has 8.4M people and about 50% of the country’s population. Looking at city population versus urban area/metro area is foolish since city boundaries are arbitrary. According to INSEE (the French statistical agency) metro Lyon is 2.3M people, and there are five other metros in France with 1M+. More generally, if you consider city’s/metros of 1-3M people as “truly rural scale” I’d argue you won’t find many takers, or, if so, you should not be arguing against car centric infrastructure anywhere but East Asia. If the Netherlands is “truly rural” than almost every other country is as well (Japan’s total population density is lower than the Netherlands and S. Koreas is about equal) and since everyone agrees cars are good for rural areas then naturally these countries should not invest in transit.
Okay, but in the most car-centric places cars don’t pay the operating cost of their infrastructure with user fees, either.
The link you give doesn’t refer to operation cost of road infrastructure, it refers to capital costs (the money paid to build or repair roads). Operating costs for cars are fuel, vehicle maintenance and driving, and these are paid for by car owners, not by taxes. As I have already noted, transit operating costs (fuel, electricity for third rail, driver salaries, etc.) are not fully covered by ticket costs, and most transit systems pay exactly 0% toward capital costs, compared to the 17-100% covered by drivers. And since at least four states totally 54M people (in between Spain and Italy in size) do cover road costs 100% (from high tax California to low tax Montana) it is certainly possible. Roads are cars are simply better at paying their costs by users.
As I mentioned, making any transportation “pay for itself” via user fees might not be a good policy. If you build a port and then subsidize it the compound economic benefit and associated tax revenue (from related industries like warehousing and transshipment or businesses that locate close to the port) that may leave you better off than if you tried to fund the port through dock fees and got less shipping traffic.
“it hasn’t happened yet, therefore it’s not going to happen” is not a very logically-sound argument
This is true, however, Strong Towns is not making a random prediction out of a crystal ball. They have presented a falsifiable thesis of how their so called development Ponzi scheme will lead to collapse – areas built with a high ratio of infrastructure to structures will start off profitable as the maintenance cost of newly built roads and sewers is close to zero, but will fall into insolvency when those maintenance costs climb with time beyond the tax base of the properties which have low value per acre because they are spread too far apart.
The problem with this thesis is that suburban style development that Strong Towns criticizes has been going on for between 60-75 years (75 if you take Levittown in 1950 as the start, 60 if you take 1965 [the year half of the interstate system was complete] as the start). This is long enough for much of that infrastructure to require replacement (even at the more reasonable figures I give, rather than an arbitrary 20-30 year depreciation timeframe). And yet suburban communities from that first wave of development (suburbs in Long Island, much of LA county, etc.) are not going bankrupt in waves, nor is the country economically moribund. On the contrary GDP and standard of living in the US are higher now than in 1950 before that model of development began.
people are generally more willing to cut infrastructure spending than emergency services/elder care (which pensions are arguably a form of)
First, if cities are choosing to cut infrastructure to pay for other services, that still invalidates the Strong Towns theory because clearly the tax base is enough to pay for the infrastructure maintenance, elected officials are just using that revenue for something else. Second, you completely misunderstand the situation of US municipal bankruptcy in the 21st Century. These cities aren’t going bankrupt because they overspent to provide world class fire/rescue operations, they are going bankrupt because of unsustainable sweetheart public sector union deals that feature some combination of double-time for all overtime at all levels (instead of time and a half for normal working hour overtime and management being fully salaried), pensions being calculated off of last year total compensation (instead of an average of that last few years base salary), retirement vesting after just 5-10 years and having a high 2-3% earned per year of service (instead of 20 years to earn retirement, and getting 50% at 30-40 years), retirement available at 50-55 years old for public safety workers (instead of 60-65 years old), fully paid government provided benefits in retirement (instead of a retirement health plan being available that requires some premiums/copayments paid for out of the pension), and no restriction on receiving a pension while working (so someone can work 20 years as a cop, then ‘retire’ and take a job as a police dispatcher, and be paid both their pension and new salary until they retire again with two pensions). As a result a firefighter who was making 100k/year working 42 hours a week (standard for public safety jobs so 4 shifts cover a whole week) can in their last year on the job work 84 hours a week, be paid 300k that year, then retire with a pension of 180k (60% of 300k) and make more each year in retirement than then ever did in base salary their whole time on the job. Paying someone in their 50’s a pension that puts them in top 5-10% of earners isn’t “elder care”.
The thousands and thousands of empty Rust Belt homes would like a word with you.
This does the exact opposite of what you think it does and only further undercuts the Strong Town theory and your position. The Rust belt cities were all built with the transit/walking/biking development paradigm you favor (because they were all founded or grew most during the late 1800’s early 1900s when cars didn’t exist and people got around by walking or later streetcar). This is also the Strong Towns development model – narrow lots, many with townhomes or small apartments, on a grid of streets instead of single family homes on cul-de-sacs, commercial areas focused on a “Main Street” model with businesses fronting on a sidewalk instead of strip mall or big box stores with a parking lot out front. Yet despite having the ‘superior’ development pattern of less linear length of infrastructure per lot, Rust Belt cities are in poor financial shape and losing population, while cities using the ‘inferior’ suburban model (Phoenix, Atlanta, Dallas, etc.) are booming.
@Onux
The data for US fixed connection internet speeds from Ookla seems suspect considering the US is below the OECD average for fiber internet connections.
The mobile might be more believable since 5G adoption in the US is actually quite high. That was surprising to learn since US cell service is expensive per GB, but at least it is expensive and fast.
@Onux
What does wide area unweighted population density have anything to do with rural scale?
@Sassy
If you have better figures for internet speed you are welcome to present them. Does the US still have below average fiber penetration? According to the OECD itself fiber growth in the US was higher than anywhere else in the world from 2020-2023 (https://www.oecd.org/digital/broadband/broadband-statistics-update.htm#:~:text=Fibre%2C%20which%20is%20the%20dominant,countries%20where%20data%20is%20available).) – over 250%.
As to what constitutes urban vs rural, that is a great question applicable on many levels. It is related to the debate from a few years back among US urbanists when the Census reported that metro LA, per their definition, was denser than metro New York. This led many people to begin using weighted population density rather than simple population density (people divided by area). There is both the question of what is the definition of urban/rural (population density? housing unit density? total paved area/building area? – to say nothing of adding suburban/exurban definitions) as well as where to draw boundaries (many political boundaries are arbitrary not functional).
Never-the-less, I don’t know of anyone who would consider France or the Netherlands overall rural, even if they do have rural areas and their cities are not as absolutely large as Seoul or Tokyo. This questions also cuts both ways. Although I somewhat assumed Norway to be “largely rural” because of how lightly populated it is overall, according to the World Bank it is actually slightly more urban (84% of pop) than France (82%) – although Norway has vast unpopulated spaces, the majority of the population there is lives in Oslo, Bergen, etc. But then these urban areas are less densely populated in an absolute sense than say Paris or Lyon. And around the questions and definitions go again…
“According to INSEE (the French statistical agency) metro Lyon is 2.3M people, and there are five other metros in France with 1M+. More generally, if you consider city’s/metros of 1-3M people as “truly rural scale” I’d argue you won’t find many takers, or, if so, you should not be arguing against car centric infrastructure anywhere but East Asia. If the Netherlands is “truly rural” than almost every other country is as well (Japan’s total population density is lower than the Netherlands and S. Koreas is about equal) and since everyone agrees cars are good for rural areas then naturally these countries should not invest in transit.
If I’m cherry-picking, so are you. The entire metro Lyon area of 2.3M people is over a much larger area than Lyon proper, and so its population density is only 500km/sqm, equal to the entirety of South Korea, which obviously includes lots of uninhabited areas (mountains, forests, etc.) and plenty of farmland (as does, obviously, the average density of Japan). Alon has a post about this, and the verdict is that the the “median” (at least 60%) Korean/Japanese/Taiwanese person lives in a city of around 2 million people, whereas e.g., the “median” French person doesn’t even live in a city the quarter the size (44% at 500k).
Netherlands hasn’t the barriers to universal high density like East Asian countries do, and so it could/should be denser than it is if it could be so without negatively affecting quality of life (which I admit is an extremely difficult thing to have an impartial conversation about). I’m using East Asian cities specifically because thus far, they’re providing the highest quality of living to the most people per area (not to imply that they have no QOL issues of their own); they’ve at least started to work of solving good QOL at very high density/urban populations.
I’ll make the motive for my selection plain: the “area” part of the above triangulation is obviously mostly fixed (barring things like land reclamation, space colonies, what have you), and as I’m interested in the best life (broadly defined) for the most people, East Asia has my attention. What interests are you secreting away?
“suburban communities from that first wave of development (suburbs in Long Island, much of LA county, etc.) are not going bankrupt in waves, nor is the country economically moribund. On the contrary GDP and standard of living in the US are higher now than in 1950 before that model of development began.”.
You’re conflating things that there are no intrinsic reason to believe are directly related. Suburbs in Long Island and LA remain fiscally sound because not only are these still desirable places to be (proximity to the NYC job market and Hollywood/good weather/good marketing [in essence], respectively), but they’ve also continually underbuilt their housing, for decades, keeping property values and their associated taxes high relative to their infrastructure costs. By contrast, the Midwest can’t brag of Mediterranean weather or healthy industries in denser city centers (Chicago being an obvious exception).
I understand you’re not trying to attribute higher US standards of living directly to auto dependency, merely pointing out that it doesn’t appeared to have done any harm. However, the counterfactual which is necessary to make this a meaningful thesis–a U.S. built around 21st century transit (which has obviously meaningfully advanced since the ’50s’)–doesn’t exist, so we have little more than speculation to go by. The closest counterfactuals are other countries, and obviously there are too many confounding variables for a meaningful comparison. I could point to higher-than-OECD-average U.S. government debt and lower life expectancy, but it’s not a clear-cut American loss, and again, too many variables.
“Second, you completely misunderstand the situation of US municipal bankruptcy in the 21st Century. These cities aren’t going bankrupt because they overspent to provide world class fire/rescue operations, they are going bankrupt because of unsustainable sweetheart public sector union deals”
No, I don’t completely misunderstand. In fact, I said: ‘”Not that no efficiencies can/should be found there, but the first recourse to enhancing or motivating enhancements of efficiency probably shouldn’t ever be pure cost-cutting. We need to figure out what money’s being wasted doing what things, rather than just saying “this number is too big; make it smaller”.’ A public sector that manages to extract more value than it is able to provide to the public in turn is an inefficiency, in the most literal sense. It’s a reasonable thing, I think we agree, that providing huge pensions based on an unrepresentative period of overwork is inefficient, but that doesn’t make pensions not a form of eldercare; it just means that allowing allocations of public funds based on what are still essentially private interests–“public sectors” which functionally represent only the interests of their union members aren’t necessarily working in the public interest–is probably going to be inefficient.
@Onux
Do you even read your sources? 250% is the US growth in fixed wireless access (using cellular internet as your home internet service), and has nothing to do with fiber internet at all.
As per OECD, as of June 2023, the US had a fiber internet adoption rate of 22.6%, vs the OECD average of 41.1%, and well behind world leaders like South Korea, Japan, Spain, and Sweden at 80-90%.
Re modal splits: the numbers giving France 85.1% and so on are for passenger-km, which are not a good thing to measure for this, because auto commuters drive longer distances than non-auto commuters travel by train, bike, and foot for equivalent trips. The trip to work modal split for France is 70% car, 15% transit, 15% other – and yeah, those 70% are mostly in the economic basket case parts of the country, while Ile-de-France is 43-43. The national p-km modal split mostly detects how good the intercity rail is, especially at long distances, so for example France > Germany in rail p-km modal split, but then the commute modal splits in Germany is if anything a hair higher (16% as of 2017; my French numbers are from 2014, I believe).
@Onux, @Sassy
How people get internet only matters if you feel a need to be superior to someone else for some reason. Fiber, Cable, 5g, satellite – they all work, having one does not make your better than someone who doesn’t have it. Sure there are pros and cons to each, but in the end they all work. What matters to people with a life is if their internet is “fast enough” which is purely subjective. For most of the US internet speeds have been “fast enough” for a decade or more. Find something to fight about that matters. If fiber penetration is how you feel better than the US, then go ahead and feel superior I don’t care.
I think the Strong Towns concept is a solution looking for ideal problems, rather than real problems. Detroit was 1.8 million population in 1950 (tied with LA for fourth largest US city). It went bankrupt, and now has 620k. It didn’t have an infrastructure problem — it used its infrastructure to pay excessive numbers of civil servants salaries/benefits/pensions. To exit bankruptcy, it had to agree to no longer have the city own the zoo, the Detroit Institute of Arts, the Water Department, and the streetlights (employees said they wouldn’t agree to cutbacks as long as there was a $100 million Van Gogh hanging in the DIA). In Stockton, California, employees said the city should sell the city parks to real estate developers to fund the pensions for people who raked leaves in the city parks. It takes extra effort to use Memphis as an example — for decades, Tennessee allowed cities to forcibly annex adjacent towns (when Elvis bought Graceland it wasn’t in Memphis, it was in Whitehaven). P.S. Detroit and Memphis are the two largest US cities with majority black population.
Two hourly isn’t going to get people out of their cars on to the train.
Probably half hourly is the frequency to go for for intercity services for that.
Isn’t the point of the post to say that it can be in some cases?
Once-every-two-hours-or-worse planes still get chosen in many cases over driving, even when it isn’t a multi-day drive. And even for trains it isn’t purely theoretical, as Alon points out, low frequency TGVs between Paris and small towns see decent usage, and even with lower speeds German intercity trains aren’t exactly empty.
Half-hourly is of course much better, and the relatively high frequency of even local Shinkansen services probably contributes to the relatively short average distance of Shinkansen trips, due to tons of people choosing Shinkansen for shorter trips that wouldn’t make sense at lower frequency.
Given enough speed and long enough distances, at least some people will get out of their cars for a once-every-two-hours train. And even every half hourly might not be enough to get people out of their cars for a one stop trip on a rural DMU that tops out at 65km/h.
The people with good regional train usage (which doesn’t really apply to the French) are running service every 30 minutes or less.
Yeah less than that is OK for fast services into the capital where people in the regions really don’t want to drive.
If you are competing with a drive of more than maybe 5 hours or so where you have to stop perhaps even multiple times then yeah 2 hourly is probably OK.
It’s relative to trip length – the six-hour trips of the original InterCity trains here were such that two-hour frequencies were enough. If I travel to Munich, I have hourly frequencies for a four-hour trip, but I will try to pick a train that is cheaper or more direct (most trains reverse-direction at Leipzig and take 4:30; a few trains per day, planned to run hourly under the D-Takt, skip Leipzig and instead stop at Halle with no change in direction and do it in four hours flat). Half hourly frequency is what you should aim for if you have a lot of cities that are about an hour apart; in Germany these days this is indeed the case, due to our high population density, and the same is true in the UK for the same reason, but France is less dense and those gaps in service for passengers connecting from random secondary towns to Paris are not the end of the world.
Berlin-Frankfurt is similar. The Sprinters taking just under 4 hours run every two hours (under normal conditions, trips are slower at the moment due to track maintenance), but there are multiple other connections that take 4-4.5 hours, e.g. via Leipzig, or via Hannover, so you end up with a fairly high-frequency service, just with uneven headways and highly inconsistent routes and stopping patterns.
Given railways have huge fixed costs and very low variable costs you really should be aiming for half hourly service pretty much everywhere.
Trains can be run with only two cars on more rural lines.
This is a very English opinion, where nobody except in Cornwall lives more than 2 hours from the major cities.
In places where your can’t justify more than 1-2 tph you’re not going to convince locals to ditch their for within-locality trips, but depending on what the speed gets you they will ditch it when going out of the locality.
Wilmington is halfway give or take a bit. NY-Philadelphia has to be quicker. Amtrak is aiming for quicker. I don’t remember or care if that is NEC-o-the-Future version 192 or whatever that goes to Philadelphia City Hall…. because railfans from the hinterlands don’t understand how big cities work. If it takes 45 minutes to get between NY and Philadelphia that means they have to make Philadelphia to DC 45 too. …. Wilmington is halfway….
The trains railfans call the Clockers were every half hour-ish, 18-ish hours a day. Without any stops between Newark and North Philadelphia. There were other trains – for instance the Boston – Washington trains – and many more. I’ve never have and never wlll have the gumption to collate the pages of schedule in the Official Guide I have. The Reading-Central of New Jersey were running competing trains. Very handy if you are starting out or destined to – Jenkintown…
…. The express to Washington D.C. leaves at :00 and :05 having a Philadelphia train that is cheaper? lusher? both? skims Philadelphians off freeing ups seats for Baltimore and DC. Leaves at :25 and :55? Hmm. So does one that toddles along the West Trenton line… Make railfan trainspotters gush foam! trainspot the southbound at North Philadelphia and the same train northbound at North Broad. Satisfies railfans from the hinterlands that don’t understand how big cities work by stopping at Suburban. For, insert a deep sigh here, City Hall.
There are 13 return trips by TGV on week days between Lyon Part-Dieu and Marseille St-Charles with first departures before 7 am and last ones after 8 pm with a two-hour gap at most.
There are 17 trains from Birmingham to Leeds each day with the first at 6am and the last leaving at 9pm – they run on an hourly clock face schedule with one extra train in the evening peak.
There are 31 trains from Birmingham to Manchester each day with the first at 6am and the last at 10pm, they run half hourly until 7pm and hourly after that.
And there are plenty of French examples that are worse than the one you give. Lyon to Toulouse has a worse service than Glasgow to Fort William with 3 trains a day rather than 4 – although with similar first and last trains in each direction.
I was not giving an example. I was just correcting Alon’s mistake.
UK trains are indeed more frequent. French rail network is more than twice as large (35000 vs 16000 km) and less used (391 vs 461 million passenger train-km/year). The numbers are reversed when looking at actual traffic. France has about twice the UK’s traffic (101 vs 53 billion passenger-km; 35 vs 15.7 billion ton-km). Contrary to what you believe, this is not due to the RER (London and South East passenger-km are actually slightly above Ile-de-France). French Regional traffic is about double the UK’s, and Long Distance about quadruple.
While frequency is great in all cases because it increases flexibility, people are willing to adjust their schedule if the travel time is reduced.
Caltrain is probably the best example of this where many people will adjust their schedule in order to save 5-25 minutes by taking a faster train.
Faster trains have the additional benefit by becoming a viable option to people traveling further from the station – something that no amount of frequency will be able to accomplish.
These words: do they convey or contain meaning of some sort?
They do seem grammatically viable – something that a surprisingly small amount of training on a natural human language corpus training will accomplish.
Yes they do, and you already know what they are, because you (correctly) complained about them elsewhere plentifully. Namely:
– Either a massive overestimation of speed gains achievable by stop-skipping, or an assertion of strong passenger preference for “speed theater”, i.e. superficially higher speeds, and the features traditionally associated with higher speeds.
– An explicit claim that passengers schedule everything else around the transportation, that frequency is merely a matter of “flexibility”. Not only is there no “show up and go” behavior, but even the start and end of shifts/meetings/etc. adjust to fit transportation schedules, therefore there is no such phenomenon as “assuming this train is on time, I’ll be late by 5 minutes, therefore I have to take the previous train, and then spend 25 minutes twiddling my thumbs at the destination”.
– These assumptions would make sense for air travel, but are clearly wrong for a subway line. And Caltrain is close enough to the subway end of the spectrum that these assumptions are wrong for it.