You mention the issue of free buses leading to enshitification but don’t really reckon with the degree to which that enahitification reduces the utility of buses for riders. If a bus and car are equally convenient but I run a real risk of getting screamed at by an insane junkie on a bus I’m going to drive. So if you want buses to be free and actually useful you need to price in the cost of keeping them safe and pleasant.
he does address this though. his argument is that, unlike on a train, it's difficult to prevent non-payers from boarding a bus or to evict them once they've boarded, so screening passengers with fares wouldn't be effective.
I think he is wrong about that. You can absolutely have fare gating on buses, though it does require changing how payment for buses works (basically you have to make it hard for people to get on and then pay).
im not convinced its impossible. but off the top of my head i can't think of a physical setup that doesn't require either lots of infrastructure at each bus stop (impractical) or allow the possibility of a nonpaying customer to get to a fare gate inside the bus and refuse to leave (preventing the driver from leaving).
You would most likely have a situation where non-paying customers could get in the door but not get a seat without paying. And yes, that would still result in some problems with people refusing to get off the bus. But it would also disincentive trying to fare jump since the best you could do is get a foot in the door and then not sit down, plus it would make it pretty easy for law enforcement to respond if a person has functionally confined themselves to the entrance of a bus they refuse to vacate. The system would only work if both drivers and cops treated free riding as a serious problem so there's certainly a cultural dimension to it as well.
> The system would only work if both drivers and cops treated free riding as a serious problem
But there’s no evidence that it is, and trying to remove such passengers is functionally impossible as is. So your solution is akin to saying “if we just put shit tons of resources into a problem that isnt very important and call the cops regularly, we can keep the most vulnerable people in society from skipping their fare, so long as we are comfortable with making the bus system unusable.”
The premise is that free riding is an issue because it leads to the enshitification of public services. BART pre- and post-fare gating is a pretty good example of why you might believe this to be true, but if you're committed to the notion that free riding isn't a problem and in fact letting people not pay and make public services unusable for most of the public is a justifiable form of welfare then we're too far apart to have a discussion about it. Forcing people to pay for services makes them *more* usable not less for all but the sliver of people who both refuse to pay and are unwilling to or incapable of acting in a prosocial manner once on board.
I would argue that "the most vulnerable people in society" are the victims of disorder, not the deranged junkies who are enshittifying transit. As in, riders who could be elderly, disabled, pregnant, mothers with young children - basically anybody who isn't an able-bodied adult male.
The irony beyond that charging a fare and having police do periodic check ins for fare evasion plus general prosociality would solve the quality of life problem. But Decker doesn't ride the bus, or frankly GAF about people who do.
The keyhole solution here is to try and find proxies for "is a minimally prosocial person" that are less inaccurate (to be pedantic, more sensitive without being too much less specific) than "can pay full fare". You can imagine a lot of these and probably it's worth having municipal systems experiment with different ones. For example, suppose your bus pass becomes free or nominally-priced but requires you to provide some proof of stable in-city residency which insane junkies mostly don't have the spoons to get. Then you might still get most of the incentive benefit, and as a bonus you can still make money off of charging for tourist passes.
I have ridden systems at times when they have been free and at times when they charge, yes. But as OP has noted, you can also know this by knowing whether or not busses actually refuse such riders or remove them. This sort of knowledge you’d get by riding buses and/or researching the policies and practices of transit providers. It seems that it would be up to the person claiming there is a problem to substantiate it, no?
you were previously saying that if someone were to ride the bus regularly they would know whether free fares would have any effect on the frequency of problematic riders. if your view is that fares are not enforced, and fare enforcement is the only way fares could have an effect on frequency of problematic riders, then i agree that knowing these two facts is enough to know that free fares have no effect.
but you can ride the bus regularly without knowing what happens when people can't or refuse to pay the fare. i rode the bus daily for many years without ever noticing this happen. so i disagree that your view must be obvious to anyone who rides the bus.
I like the idea of free fares in both theory and practice -- we have free buses in my hometown, and it's great -- but my main concern is that in New York City and in other places it's not clear that it's the best marginal use of transit money and political capital. There's a finite pool of money that people are willing to spend on transit, generally, and so going fare-free trades off against service improvements.
I'm a bit puzzled by the safety argument against free fares, although this is admittedly because I generally feel pretty safe on public transit no matter what. But to the extent that I think about safety on public transit, it's that more riders -> more safety. I'm kind of skeptical that some of those riders being homeless/etc. could change it, unless it's 100% having effects on the homeless? And we also probably want the homeless people to have reliable transit to help them get out of homelessness! So you might be being a bit too generous to it.
> Incorporating the ability to change the network of roads or trains would make it complicated beyond our ability to answer, so we will not analyze it.
This is an unreasonable assumption, it is the central issue. The tradeoff available is not tax cars to pay for busses, it’s use limited budget for free busses over other transit improvements. On this metric the proposition fails. Free busses are a far from trivial cost to the point that the shortfall in combination with increased usage would require cuts to the network. The marginal rider is also more likely to be as you noted a problem rider or an otherwise pedestrian/biker not necessarily a driver, and those people are more likely to take shorter trips slowing the bus down more
Basic tax theory would suggest that you shouldn’t fund government expenditures by taxing products with positive externalities, and that is precisely what fares are. Second, substantial changes in the rail infrastructure are unlikely, and I am quite doubtful there are large welfare gains from new bus routes. Holding the network fixed is a reasonable decision to make.
>and I am quite doubtful there are large welfare gains from new bus routes. Holding the network fixed is a reasonable decision to make.
Network redesigns have had huge positive impacts for ridership in many cities. You fall into a local optimum for a while, as ridership patterns change you tinker around the edges for a while (reducing service on an under-used route, maybe adding an express option on a high-demand route)… until in a couple decades, the original map is just no longer well-suited to current travel-patterns.
what do you mean by substantial changes in rail infrastructure? e.g., in toronto, construction just began on an 18km extension of the commuter rail system, expected to increase annual ridership by about 5%. is that substantial? or do you mean like a major reorganization?
if the effect on rail networks from free fares is mainly from having to reallocate budget, then it's not necessary to model them while studying free fares. just compare the benefits of spending $X more on subsidizing bus routes to the costs of spending $X less on rail infrastructure (or whatever it is you plan on cutting) and decide whether it's worth it.
Mr. Decker, I appreciate that you consider the problem of "undesirables" with regard to trains. May I suggest that we centralize our national welfare system and build free housing in Kansas? Rather than trying to gatekeep services *within* the city, if we gatekeep *the city itself*, then things become simpler. Additionally, we might pass a law saying that people who have been convicted of a crime are not allowed to enter certain cities, in the same way that we prohibit sex offenders from living near schools. My hypothesis is that rural criminals do much less harm than urban criminals, because lower density means a lower risk of re-victimization. When we consider the economic effects of crime (lower social trust), criminality can be considered a kind of "pollution," just as we would think of air pollution or water pollution. By redistributing criminals from urban to rural spaces, we can generate a net increase of national welfare. I suggest turning the federally owned lands in Nevada to a giant open-air criminal resettlement zone, where they will be provided with free housing and internet services. Thank you for your consideration.
My priors reading a paper like this is that when you make something artificially cheap or free, you weaken the price signals and often end up with chronic underfunding and decay, especially because it’s politically easier to cut fares than to raise taxes or impose user charges. I try to keep in mind the asymmetry between the technocrat who can raise taxes as easily as they can subsidize costs in their model, with the actual incentives for the politicians implementing said model. How often will we actually get the concomitant tax raises to offset the cost reduction such that transit doesn’t fall into perpetual dilapidation?
I’m also skeptical of technocratic claims that people are systematically “failing” to maximize utility, since revealed preferences suggest many people genuinely value the flexibility and autonomy of driving, even if it results in more traffic.
At a deeper level, I’m also thinking about property rights and responsibility. I worry that redistribution through subsidies can erode norms of accountability by separating users from payers, and I’m not convinced that marginal efficiency gains alone settle the ethical question of who should bear costs in a system like this. Why should the people rendering a service be the ones not paying, while all the people who are *not* using the service be the ones who do? Doesn’t this just erode personal responsibility and externalize a different set of costs onto others?
I hope I’m not making a category error when engaging with empirical work like this but still asking whether those institutional and moral concerns are being overlooked. I have a lot of libertarian baggage I suppose.
>I’m also skeptical of technocratic claims that people are systematically “failing” to maximize utility
No, this is non-responsive drivel. Each individual driver does not take into account how they slow down other drivers. This is as simple an example of externality as you can get, and requires no systematic failure to behave rationally.
I like the presentation of the model and I have mildly updated towards heavily subsidized public transit, however I still don't fully understand how the headline results could be the case.
Like there is a clear negative cost to public transportation, the cost to provide it. Shouldn't the optimal state of affairs be each mode of transport be charged equal to the total negative effects of it? E.g transit at full cost of providing the service, driving at that plus congestion charge, environmental and whatever other externalities you want to price in. Then people take each form of transit exactly when it maximizes real total social welfare. Is feasibility of accurately and frictionlessly taxing driving the restriction here? So if we assume driving can't be made optimally costly, the second best thing is to make public transit cheaper as well so that people drive less, which is a good thing because we aren't appropriately taxing driving and so its net negative?
One other thing, I'm not sure your point about trains stopping causing a spike in traffic is that strong, as clearly an acute shock reducing public transit use is going to affect traffic differently then a long term trend in public transit use, which (at least naively without being an expert in this) I would expect to resemble something like opening a new lane of roadway, where traffic maybe reduces marginally in the short term, but eventually demand eats all the new capacity.
Big difference between a boring bougie weekend stroll in a homogeneous area vs a daily commute with the multiethnic underclass. I'm skeptical you have much experience with the latter. Not having the skin in the game and pushing for radical changes that will degrade quality of life is obscene.
Moreover, as LastBlueDog noted, not only is your proposal likely to make the experience of riding buses *worse* on average by removing the one selection effect in its favor, you're underestimating just how much. You're also imagining, improbably, that people with options to drive themselves will be incentivized to ride, when the exact opposite will happen. Edge cases that tolerated the bus will flee. There's ample evidence of this both domestically and in other countries.
1. if i follow correctly the road price being simulated should be the marginal social cost of the congestion externality of driving. it seems to me that this should be an obvious kaldor-hicks improvement but eyeballing the graph it looks like road pricing + flat rebate is about neutral on consumer welfare. i don't understand how the graph can look like that.
2. in the transit pricing + budget constraint model why does top quintile welfare decrease? they're not paying more taxes because transit is self-sufficient, they're not deterred from driving because roads are unpriced, and they're experiencing less congestion. what's the catch?
3. you mention a paper finding that the welfare gains from road pricing are much smaller that you would get if everyone had homogeneous preferences. how does the inhomogeneity matter?
1. The environmental gains accrue to everyone, including non-commuters. Thus any one city will see welfare fall.
2. Lessened frequency of transit, and they have a high value of time.
3. We see that not everybody leaves the house at exactly 8 am for work. This could be that people all want to leave at 8 but are flexible, or it could be that they’re all very inflexible, they just want to leave at different times. Road prices will cause a bigger change in usage than the former.
Outside of New York City, transit systems in America are terrible. This isn't because they don't get enough subsidies, it is because they are subsidized too much. Transit agencies get so much money from taxpayers that they have no incentive to attract new passengers. Most transit systems still operate as if most jobs are downtown when in fact nearly all jobs have left downtowns. Ridership is pathetic because transit doesn't work for people who aren't going downtown. Only when subsidies end will transit agencies have incentives to serve the 95 percent of people who aren't going downtown.
Regarding the commentary at the end, I'm not convinced that the fare gates for trains/subways are really stopping that many people. I lived in DC for a while, and it seemed like TONS of people just hopped the gates. At the same time, I think that at least some potential bus fare evaders are deterred by the bus driver potentially not letting them on. I'm excited to see what happens in NYC if/when Mamdani makes the buses free.
Some factors you have not considered, and which argue against free buses:
1) Most of the cost to users of buses is not the fare, but rather the time cost of travel because they are slow. Fare or no fare, riding buses is far cheaper than owning a car, and there are few people for whom bus fare is an obstacle to productive travel.
2) Well-run bus systems have proof-of-purchase fare payment, meaning that fare collection does not add significantly to boarding time. If free fares end up inducing ridership, that actually makes travel time worse, because there is more congestion in boarding.
3) Mamdani is suggesting to make buses free while leaving a fare for the subway. This is a terrible idea, as it would cause mode shift off the subway (which is far more efficient, both for users, and per rider for the MTA) and onto buses. You would likely end up with even slower overcrowded buses, inability to pay for enough drivers to drive them, and empty subways leading to cuts in subway frequency.
Not very well. Your graph shows that the value of time is $7-8 per hour even in the very poorest part of south Chicago. For a full $2.25 bus fare, this equivalent to 18 minutes of travel time. The time lost on a trip due to taking a bus rather than driving is generally significantly more than this. For the large fraction of bus riders who get 50% or more discounts, or are fare capped, or have a higher personal value of time, the time equivalent of bus fare is much lower. So generally speaking, bus fare is indeed a small fraction of the cost to bus riders.
Transit buses have more operating costs relative to fixed investment costs than trains, the largest is probably the bus drivers, yes? That's why buses are 40 feet so that they are economical to operate. I think a key innovation for free bus transit would be driverless/automated vehicles. Keep human drivers on the main routes in the big buses with a fare, but add a fleet of automated small minibuses or large vans that will enhance service by providing more frequency, flexible routing and dropoff points to drive down walking distance and total trip time. These automated vehicles can be zero fare.
The real obvious conclusion of this piece is that we need to more widely implement pricing on driving. It works really well. Maybe we can think about reducing fares for transit at some point after we actually have infrastructure in place to put prices on driving in cities across America.
Until then, free fares mostly will induce more crowding on buses from people who otherwise would have walked, and will do little to attract drivers out of their cars. Until driving is adequately priced, the best way to compete is to have buses run faster and more frequently. Dedicated bus lanes, signal priority, more buses, and more drivers all cost money that fares help pay for, and in real politics as opposed to econ-land, removing fares will /not/ guarantee that transit agencies receive more subsidy out of general funds, so free fares lead to service cuts and a death spirals.
I live near Göteborg, Sweden. The public transit is free for pensioners, disabled people, people who need dialysis and a host of other medical conditions at not-peak times. Everbody else pays every trip. No cash fare, but you can use an app or blip your credit/debit card. Or loadup a prepaid card. Or get a 24 hour/monthly/yearly card/app payment. These can be shared. It's increased usage.
It's kinda obvious to people in so many European countries where "free bus" models exist. Take my local town Lubin in Poland, they introduced free buses for all in 2014. Few years later, the municipality found out that the cost of subsidizing buses is on pair with the cost of running the ticketing system.
But it gets better - a few years later a District Court judge sent an official letter thanking the Mayor for introducing the free buses, as it REDUCED the number of court cases by 10,000 a year - these were people being charged for ticket violations.
Demand for public transport shoot up and the town increase number of bus lines from 9 to 22 in 2024, reaching more neighborhoods and villages around....bringing new customers to town centre at a frequency of one bus every 12 min.
Town hall reports annual savings of ~$1mln per year just from operations point of view. There are tens of millions in boosted economy.
But what's most important, you people (US citizens) are unable to imagine the quality of life in such town.
> But what's most important, you people (US citizens) are unable to imagine the quality of life in such town.
Can we cool it with the casual slander of entire populations? I am a US citizen and I can readily imagine European towns with functional public transit. Happy to learn from each other without the smack talk.
We can lower the marginal cost with cheaper monthly passes. Or have each city resident get 5 free trips per month on their bus card and explore how things adjust. Seems too extreme to go all the way to free. More nuanced fares seem a better experiment first. Also- remove stops to get them faster, and allow route deviation so they can better serve riders. Bring back jitneys!
You mention the issue of free buses leading to enshitification but don’t really reckon with the degree to which that enahitification reduces the utility of buses for riders. If a bus and car are equally convenient but I run a real risk of getting screamed at by an insane junkie on a bus I’m going to drive. So if you want buses to be free and actually useful you need to price in the cost of keeping them safe and pleasant.
he does address this though. his argument is that, unlike on a train, it's difficult to prevent non-payers from boarding a bus or to evict them once they've boarded, so screening passengers with fares wouldn't be effective.
I think he is wrong about that. You can absolutely have fare gating on buses, though it does require changing how payment for buses works (basically you have to make it hard for people to get on and then pay).
im not convinced its impossible. but off the top of my head i can't think of a physical setup that doesn't require either lots of infrastructure at each bus stop (impractical) or allow the possibility of a nonpaying customer to get to a fare gate inside the bus and refuse to leave (preventing the driver from leaving).
You would most likely have a situation where non-paying customers could get in the door but not get a seat without paying. And yes, that would still result in some problems with people refusing to get off the bus. But it would also disincentive trying to fare jump since the best you could do is get a foot in the door and then not sit down, plus it would make it pretty easy for law enforcement to respond if a person has functionally confined themselves to the entrance of a bus they refuse to vacate. The system would only work if both drivers and cops treated free riding as a serious problem so there's certainly a cultural dimension to it as well.
> The system would only work if both drivers and cops treated free riding as a serious problem
But there’s no evidence that it is, and trying to remove such passengers is functionally impossible as is. So your solution is akin to saying “if we just put shit tons of resources into a problem that isnt very important and call the cops regularly, we can keep the most vulnerable people in society from skipping their fare, so long as we are comfortable with making the bus system unusable.”
The premise is that free riding is an issue because it leads to the enshitification of public services. BART pre- and post-fare gating is a pretty good example of why you might believe this to be true, but if you're committed to the notion that free riding isn't a problem and in fact letting people not pay and make public services unusable for most of the public is a justifiable form of welfare then we're too far apart to have a discussion about it. Forcing people to pay for services makes them *more* usable not less for all but the sliver of people who both refuse to pay and are unwilling to or incapable of acting in a prosocial manner once on board.
I would argue that "the most vulnerable people in society" are the victims of disorder, not the deranged junkies who are enshittifying transit. As in, riders who could be elderly, disabled, pregnant, mothers with young children - basically anybody who isn't an able-bodied adult male.
If the penalty for holding up a bus was a mandatory 6 months in jail or inpatient treatment facility the cops could reduce their bus-related workload.
The irony beyond that charging a fare and having police do periodic check ins for fare evasion plus general prosociality would solve the quality of life problem. But Decker doesn't ride the bus, or frankly GAF about people who do.
The keyhole solution here is to try and find proxies for "is a minimally prosocial person" that are less inaccurate (to be pedantic, more sensitive without being too much less specific) than "can pay full fare". You can imagine a lot of these and probably it's worth having municipal systems experiment with different ones. For example, suppose your bus pass becomes free or nominally-priced but requires you to provide some proof of stable in-city residency which insane junkies mostly don't have the spoons to get. Then you might still get most of the incentive benefit, and as a bonus you can still make money off of charging for tourist passes.
As mentioned, this is not something free fares would change in any way. You’d know this if you were riding buses regularly as is.
im not sure how you could know this just by riding busses unless you ride both free and fare-charging busses and can compare the difference
I have ridden systems at times when they have been free and at times when they charge, yes. But as OP has noted, you can also know this by knowing whether or not busses actually refuse such riders or remove them. This sort of knowledge you’d get by riding buses and/or researching the policies and practices of transit providers. It seems that it would be up to the person claiming there is a problem to substantiate it, no?
you were previously saying that if someone were to ride the bus regularly they would know whether free fares would have any effect on the frequency of problematic riders. if your view is that fares are not enforced, and fare enforcement is the only way fares could have an effect on frequency of problematic riders, then i agree that knowing these two facts is enough to know that free fares have no effect.
but you can ride the bus regularly without knowing what happens when people can't or refuse to pay the fare. i rode the bus daily for many years without ever noticing this happen. so i disagree that your view must be obvious to anyone who rides the bus.
I like the idea of free fares in both theory and practice -- we have free buses in my hometown, and it's great -- but my main concern is that in New York City and in other places it's not clear that it's the best marginal use of transit money and political capital. There's a finite pool of money that people are willing to spend on transit, generally, and so going fare-free trades off against service improvements.
I'm a bit puzzled by the safety argument against free fares, although this is admittedly because I generally feel pretty safe on public transit no matter what. But to the extent that I think about safety on public transit, it's that more riders -> more safety. I'm kind of skeptical that some of those riders being homeless/etc. could change it, unless it's 100% having effects on the homeless? And we also probably want the homeless people to have reliable transit to help them get out of homelessness! So you might be being a bit too generous to it.
> Incorporating the ability to change the network of roads or trains would make it complicated beyond our ability to answer, so we will not analyze it.
This is an unreasonable assumption, it is the central issue. The tradeoff available is not tax cars to pay for busses, it’s use limited budget for free busses over other transit improvements. On this metric the proposition fails. Free busses are a far from trivial cost to the point that the shortfall in combination with increased usage would require cuts to the network. The marginal rider is also more likely to be as you noted a problem rider or an otherwise pedestrian/biker not necessarily a driver, and those people are more likely to take shorter trips slowing the bus down more
Basic tax theory would suggest that you shouldn’t fund government expenditures by taxing products with positive externalities, and that is precisely what fares are. Second, substantial changes in the rail infrastructure are unlikely, and I am quite doubtful there are large welfare gains from new bus routes. Holding the network fixed is a reasonable decision to make.
You can quantify the reduction in SF Muni maintenance costs when they removed free, turnstyle-hopping riders.
Free riders 5x system maintenance cost.
Soon, we'll have more data on rider rated safety. Looking forward to your intellectual honesty motivating you to quantify this impact
>and I am quite doubtful there are large welfare gains from new bus routes. Holding the network fixed is a reasonable decision to make.
Network redesigns have had huge positive impacts for ridership in many cities. You fall into a local optimum for a while, as ridership patterns change you tinker around the edges for a while (reducing service on an under-used route, maybe adding an express option on a high-demand route)… until in a couple decades, the original map is just no longer well-suited to current travel-patterns.
what do you mean by substantial changes in rail infrastructure? e.g., in toronto, construction just began on an 18km extension of the commuter rail system, expected to increase annual ridership by about 5%. is that substantial? or do you mean like a major reorganization?
if the effect on rail networks from free fares is mainly from having to reallocate budget, then it's not necessary to model them while studying free fares. just compare the benefits of spending $X more on subsidizing bus routes to the costs of spending $X less on rail infrastructure (or whatever it is you plan on cutting) and decide whether it's worth it.
Mr. Decker, I appreciate that you consider the problem of "undesirables" with regard to trains. May I suggest that we centralize our national welfare system and build free housing in Kansas? Rather than trying to gatekeep services *within* the city, if we gatekeep *the city itself*, then things become simpler. Additionally, we might pass a law saying that people who have been convicted of a crime are not allowed to enter certain cities, in the same way that we prohibit sex offenders from living near schools. My hypothesis is that rural criminals do much less harm than urban criminals, because lower density means a lower risk of re-victimization. When we consider the economic effects of crime (lower social trust), criminality can be considered a kind of "pollution," just as we would think of air pollution or water pollution. By redistributing criminals from urban to rural spaces, we can generate a net increase of national welfare. I suggest turning the federally owned lands in Nevada to a giant open-air criminal resettlement zone, where they will be provided with free housing and internet services. Thank you for your consideration.
you mean a prison. an open air prison.
no, you don’t know what a prison is. I am describing a micro-state or reservation.
Are the homeless people allowed to, y'know, leave the... reservation? Whenever they like?
is the united states a prison? most of its people do not have the right to leave
My priors reading a paper like this is that when you make something artificially cheap or free, you weaken the price signals and often end up with chronic underfunding and decay, especially because it’s politically easier to cut fares than to raise taxes or impose user charges. I try to keep in mind the asymmetry between the technocrat who can raise taxes as easily as they can subsidize costs in their model, with the actual incentives for the politicians implementing said model. How often will we actually get the concomitant tax raises to offset the cost reduction such that transit doesn’t fall into perpetual dilapidation?
I’m also skeptical of technocratic claims that people are systematically “failing” to maximize utility, since revealed preferences suggest many people genuinely value the flexibility and autonomy of driving, even if it results in more traffic.
At a deeper level, I’m also thinking about property rights and responsibility. I worry that redistribution through subsidies can erode norms of accountability by separating users from payers, and I’m not convinced that marginal efficiency gains alone settle the ethical question of who should bear costs in a system like this. Why should the people rendering a service be the ones not paying, while all the people who are *not* using the service be the ones who do? Doesn’t this just erode personal responsibility and externalize a different set of costs onto others?
I hope I’m not making a category error when engaging with empirical work like this but still asking whether those institutional and moral concerns are being overlooked. I have a lot of libertarian baggage I suppose.
>I’m also skeptical of technocratic claims that people are systematically “failing” to maximize utility
No, this is non-responsive drivel. Each individual driver does not take into account how they slow down other drivers. This is as simple an example of externality as you can get, and requires no systematic failure to behave rationally.
The rest is fine.
I like the presentation of the model and I have mildly updated towards heavily subsidized public transit, however I still don't fully understand how the headline results could be the case.
Like there is a clear negative cost to public transportation, the cost to provide it. Shouldn't the optimal state of affairs be each mode of transport be charged equal to the total negative effects of it? E.g transit at full cost of providing the service, driving at that plus congestion charge, environmental and whatever other externalities you want to price in. Then people take each form of transit exactly when it maximizes real total social welfare. Is feasibility of accurately and frictionlessly taxing driving the restriction here? So if we assume driving can't be made optimally costly, the second best thing is to make public transit cheaper as well so that people drive less, which is a good thing because we aren't appropriately taxing driving and so its net negative?
One other thing, I'm not sure your point about trains stopping causing a spike in traffic is that strong, as clearly an acute shock reducing public transit use is going to affect traffic differently then a long term trend in public transit use, which (at least naively without being an expert in this) I would expect to resemble something like opening a new lane of roadway, where traffic maybe reduces marginally in the short term, but eventually demand eats all the new capacity.
A guy who doesn't ride public buses in a major city with mixed populations wrote this.
I ride buses in Boston and San Francisco. What do you have in mind by mixed?
Big difference between a boring bougie weekend stroll in a homogeneous area vs a daily commute with the multiethnic underclass. I'm skeptical you have much experience with the latter. Not having the skin in the game and pushing for radical changes that will degrade quality of life is obscene.
Moreover, as LastBlueDog noted, not only is your proposal likely to make the experience of riding buses *worse* on average by removing the one selection effect in its favor, you're underestimating just how much. You're also imagining, improbably, that people with options to drive themselves will be incentivized to ride, when the exact opposite will happen. Edge cases that tolerated the bus will flee. There's ample evidence of this both domestically and in other countries.
three questions
1. if i follow correctly the road price being simulated should be the marginal social cost of the congestion externality of driving. it seems to me that this should be an obvious kaldor-hicks improvement but eyeballing the graph it looks like road pricing + flat rebate is about neutral on consumer welfare. i don't understand how the graph can look like that.
2. in the transit pricing + budget constraint model why does top quintile welfare decrease? they're not paying more taxes because transit is self-sufficient, they're not deterred from driving because roads are unpriced, and they're experiencing less congestion. what's the catch?
3. you mention a paper finding that the welfare gains from road pricing are much smaller that you would get if everyone had homogeneous preferences. how does the inhomogeneity matter?
1. The environmental gains accrue to everyone, including non-commuters. Thus any one city will see welfare fall.
2. Lessened frequency of transit, and they have a high value of time.
3. We see that not everybody leaves the house at exactly 8 am for work. This could be that people all want to leave at 8 but are flexible, or it could be that they’re all very inflexible, they just want to leave at different times. Road prices will cause a bigger change in usage than the former.
Outside of New York City, transit systems in America are terrible. This isn't because they don't get enough subsidies, it is because they are subsidized too much. Transit agencies get so much money from taxpayers that they have no incentive to attract new passengers. Most transit systems still operate as if most jobs are downtown when in fact nearly all jobs have left downtowns. Ridership is pathetic because transit doesn't work for people who aren't going downtown. Only when subsidies end will transit agencies have incentives to serve the 95 percent of people who aren't going downtown.
Cool paper!
Regarding the commentary at the end, I'm not convinced that the fare gates for trains/subways are really stopping that many people. I lived in DC for a while, and it seemed like TONS of people just hopped the gates. At the same time, I think that at least some potential bus fare evaders are deterred by the bus driver potentially not letting them on. I'm excited to see what happens in NYC if/when Mamdani makes the buses free.
Some factors you have not considered, and which argue against free buses:
1) Most of the cost to users of buses is not the fare, but rather the time cost of travel because they are slow. Fare or no fare, riding buses is far cheaper than owning a car, and there are few people for whom bus fare is an obstacle to productive travel.
2) Well-run bus systems have proof-of-purchase fare payment, meaning that fare collection does not add significantly to boarding time. If free fares end up inducing ridership, that actually makes travel time worse, because there is more congestion in boarding.
3) Mamdani is suggesting to make buses free while leaving a fare for the subway. This is a terrible idea, as it would cause mode shift off the subway (which is far more efficient, both for users, and per rider for the MTA) and onto buses. You would likely end up with even slower overcrowded buses, inability to pay for enough drivers to drive them, and empty subways leading to cuts in subway frequency.
1. This is explicitly addressed.
Not very well. Your graph shows that the value of time is $7-8 per hour even in the very poorest part of south Chicago. For a full $2.25 bus fare, this equivalent to 18 minutes of travel time. The time lost on a trip due to taking a bus rather than driving is generally significantly more than this. For the large fraction of bus riders who get 50% or more discounts, or are fare capped, or have a higher personal value of time, the time equivalent of bus fare is much lower. So generally speaking, bus fare is indeed a small fraction of the cost to bus riders.
Transit buses have more operating costs relative to fixed investment costs than trains, the largest is probably the bus drivers, yes? That's why buses are 40 feet so that they are economical to operate. I think a key innovation for free bus transit would be driverless/automated vehicles. Keep human drivers on the main routes in the big buses with a fare, but add a fleet of automated small minibuses or large vans that will enhance service by providing more frequency, flexible routing and dropoff points to drive down walking distance and total trip time. These automated vehicles can be zero fare.
The real obvious conclusion of this piece is that we need to more widely implement pricing on driving. It works really well. Maybe we can think about reducing fares for transit at some point after we actually have infrastructure in place to put prices on driving in cities across America.
Until then, free fares mostly will induce more crowding on buses from people who otherwise would have walked, and will do little to attract drivers out of their cars. Until driving is adequately priced, the best way to compete is to have buses run faster and more frequently. Dedicated bus lanes, signal priority, more buses, and more drivers all cost money that fares help pay for, and in real politics as opposed to econ-land, removing fares will /not/ guarantee that transit agencies receive more subsidy out of general funds, so free fares lead to service cuts and a death spirals.
I live near Göteborg, Sweden. The public transit is free for pensioners, disabled people, people who need dialysis and a host of other medical conditions at not-peak times. Everbody else pays every trip. No cash fare, but you can use an app or blip your credit/debit card. Or loadup a prepaid card. Or get a 24 hour/monthly/yearly card/app payment. These can be shared. It's increased usage.
It's kinda obvious to people in so many European countries where "free bus" models exist. Take my local town Lubin in Poland, they introduced free buses for all in 2014. Few years later, the municipality found out that the cost of subsidizing buses is on pair with the cost of running the ticketing system.
But it gets better - a few years later a District Court judge sent an official letter thanking the Mayor for introducing the free buses, as it REDUCED the number of court cases by 10,000 a year - these were people being charged for ticket violations.
Demand for public transport shoot up and the town increase number of bus lines from 9 to 22 in 2024, reaching more neighborhoods and villages around....bringing new customers to town centre at a frequency of one bus every 12 min.
Town hall reports annual savings of ~$1mln per year just from operations point of view. There are tens of millions in boosted economy.
But what's most important, you people (US citizens) are unable to imagine the quality of life in such town.
> But what's most important, you people (US citizens) are unable to imagine the quality of life in such town.
Can we cool it with the casual slander of entire populations? I am a US citizen and I can readily imagine European towns with functional public transit. Happy to learn from each other without the smack talk.
We can lower the marginal cost with cheaper monthly passes. Or have each city resident get 5 free trips per month on their bus card and explore how things adjust. Seems too extreme to go all the way to free. More nuanced fares seem a better experiment first. Also- remove stops to get them faster, and allow route deviation so they can better serve riders. Bring back jitneys!