[AnthonyMouse]

> This leads us nicely to the Downs–Thomson paradox. When people in the above scenario start to take other modes of transport it reduces the load on our bottleneck. Eventually reaching an equilibrium where the speed of different modes of transport balances out (as people stop switching from one mode to the other)

The premise here is that travel time can be the only trade off, but suppose we make a different one: Stop charging fares for mass transit. Then more people take it because it costs less rather than because it's faster and it can be less expensive (and only slightly slower) even when the roads are minimally congested.

[Xylakant]

Public transport is already largely cheaper than owning a car in many places, yet people drive. One good example to study is Germanys 50 EUR Ticket - now 58 EUR. It's a flat rate for all of Germanys public transport, including regional trains. You can get anywhere in Germany with this, and 58 EUR is not even remotely achievable as monthly cost for a car. Yet, while it has increased ridership, the majority of people drive.

The problem is that transportation system quality matters more for a lot of people. The problem ends up as people owning a car for the last mile - that is from the rapid transit to their porch. And once they own a car, the calculus changes - you already incure the cost for the car.

So what you need is a reliable way to get door to door - and that requires more than slapping down a few light rail tracks. It requires connections that cover the last bit as well - and they will often run unprofitable. In the end, building such a system requires the (political) will to regard public transport as a common good infrastructure like road that gets paid from taxes and is not considered an enterprise that (could potentially) make money. In the end, this could also be made free, but free alone will not make that happen.

[AnthonyMouse]

> Public transport is already largely cheaper than owning a car in many places, yet people drive.

This is not a binary distinction. If you save $0.20 by taking public transport but it takes an hour longer, of course people drive. If you save $3 by taking public transport and it only costs you five minutes, that's different math.

> You can get anywhere in Germany with this, and 58 EUR is not even remotely achievable as monthly cost for a car.

When most people have a car you have to compare it not to the amortized cost of owning a car but the marginal cost of driving one you already have.

The majority of trips might be suitable for public transport but then people have a car because it's such an inconvenience to go to Costco and carry back everything you buy there on a bus, or they occasionally go somewhere the bus doesn't. So they get a car and then the insurance, tax, depreciation, etc. are all sunk costs and to get them to take the bus instead of driving themselves it has to beat the cost of gas.

Which it can, if you make it zero. Which in turn increases ridership, allowing you to justify more routes, which reduces latency, which causes even more people to take mass transit. By making mass transit more attractive instead of making driving less attractive.

> It requires connections that cover the last bit as well - and they will often run unprofitable.

Or you can just handle 85% of the cases that would have a justifiable amount of ridership and then let people drive a car or get an Uber in the 15% that would be mostly disused, instead of leaving it how it is now where people drive the majority of the time.

[cyberax]

> Which it can, if you make it zero.

But you can't. Transit costs A LOT, its costs are just pushed onto car owners.

Instead, we should be honest and price it at the full 100% recovery rate, with 100% capital cost return. People will then start to think: "Should I continue paying that $20 per trip on a light rail, or should I get a car?"

"bUT poOR peoPLE@@!!!" - poor people also deserve comfort. I'm all for sponsoring car purchases for poor people and/or giving them money to buy transit passes at full cost.

[AnthonyMouse]

> But you can't.

Well sure you can. We know how much it costs, the budgets are public. Completely zeroing out fares would be a single-digit percentage of the government budget. Meanwhile it would save the public money on net, because collecting the cost as taxes has lower overhead than operating a parallel fare collections infrastructure. And it benefits drivers by giving them exactly what they've always wanted -- an incentive for other people to use mass transit:

https://theonion.com/report-98-percent-of-u-s-commuters-favo...

> Instead, we should be honest and price it at the full 100% recovery rate, with 100% capital cost return. People will then start to think: "Should I continue paying that $20 per trip on a light rail, or should I get a car?"

Which is exactly the problem with your plan. If you build a rail line and set the price at $20/trip then people don't use it, so the amortized cost of the rail line becomes $30/trip because you have to pay for all the same fixed costs with fewer riders. But a $30 fare reduces ridership even more and soon there is no mass transit which in turn makes it suck to drive because there are too many people in cars and your commute is 20 miles in two hours.

Whereas if you set the price to zero, the actual cost per trip which is now being covered by taxes comes out to $4/trip, because at lower cost you get higher ridership and more usage to spread the fixed costs over. Which in turn means less traffic congestion on the roads for the people in cars.

> I'm all for sponsoring car purchases for poor people and/or giving them money to buy transit passes at full cost.

You're all for subsidies as long as they're paying the full cost? Subsidies are the thing where they're not paying the full cost.

Moreover, you want the same incentive for everyone -- if a free fare would get someone at the 70th percentile income to take the subway instead of a car, give it to them so they do that.

The converse where you use means testing is not only bringing in high administrative costs, it creates a poverty trap where making a little more money causes you to lose the subsidy and thereby removes your incentive to do it. Means testing is effectively a scheme to impose high marginal tax rates on the poor.

[cyberax]

> Well sure you can. We know how much it costs, the budgets are public.

Nope. A realistic public transit network can NEVER be as efficient as a car network in a city. It's mathematically impossible, unless you sabotage your city so much, it's a hellscape (e.g. Manhattan).

> Completely zeroing out fares would be a single-digit percentage of the government budget.

So would be giving everyone a (cheap) car.

> Which is exactly the problem with your plan. If you build a rail line and set the price at $20/trip then people don't use it

Good, then don't build it! Easy peasy. Price is a GREAT signal. Subsidies, hidden fees, misplaced incentives and other crap lead to suboptimal outcomes.

You basically have a circular argument: transit is needed because it allows density, and density is good because it allows transit. And since we need transit, it must be cheap.

> But a $30 fare reduces ridership even more and soon there is no mass transit

Great, we need exactly that.

> which in turn makes it suck to drive because there are too many people in cars and your commute is 20 miles in two hours.

Nope. People will adapt and start to move out office space out of Downtowns.

And yes, this can work even at a gargantuan scale. Greater Houston Area has comparable population to New York City, yet it has faster commutes and far better living conditions.

Ideally, though, cities should stay reasonably small. 300k seems to be the sweet spot from the efficiency standpoint.

> Which in turn means less traffic congestion on the roads for the people in cars.

Nope. It just doesn't. Research shows that more transit use does NOT decrease traffic, except in very narrow cases (on arterials immediately parrallel to fast transit). Moreover, over time it leads to MORE traffic, as transit brings in density, and density results in more traffic.

> You're all for subsidies as long as they're paying the full cost?

I'm OK with giving poor people money so they can THEMSELVES decide on what they can use it, instead of trying to social engineer them by giving them "free" rides. When each ride costs $20 just in op-ex (true cost for Seattle, btw).

Richer people should pay the full cost of rides. This also applies to cars (although in my state car user fees already pay for 98% of all road maintenance and construction).

> The converse where you use means testing is not only bringing in high administrative costs, it creates a poverty trap where making a little more money causes you to lose the subsidy and thereby removes your incentive to do it.

That's exactly what transit is achieving. It keeps people trapped in poverty, by reducing their economic choices.

[AnthonyMouse]

> A realistic public transit network can NEVER be as efficient as a car network in a city. It's mathematically impossible, unless you sabotage your city so much, it's a hellscape (e.g. Manhattan).

If you have a city without any mass transit, there will be traffic congestion. Beating the "stuck in traffic" time for a car is not hard at all. Beating the car's time when there is no traffic is harder, but an express train can certainly match it, and anyway how do you intend on preventing traffic congestion in a city with no mass transit?

> Price is a GREAT signal.

Price is a great signal for incremental costs. If you're going to burn a gallon of gas, another gallon of gas has to be produced, so you only want it to happen if someone is willing to pay the incremental cost.

It works poorly for fixed costs, because the price then deters usage even though the fixed cost is fixed and deterring usage saves nothing. This is why you should charge for gas but not for roads or for occupying otherwise-empty space on a subway car.

> You basically have a circular argument: transit is needed because it allows density, and density is good because it allows transit. And since we need transit, it must be cheap.

The argument is that transit is good because it allows density and density is good because it makes more efficient use of a scarce resource (land).

> People will adapt and start to move out office space out of Downtowns.

The buildings in the downtowns are not going to cease to exist. Something is going to be in them.

> And yes, this can work even at a gargantuan scale. Greater Houston Area has comparable population to New York City, yet it has faster commutes and far better living conditions.

Houston Metro has a population around 8M. NYC metro is 20M, is the largest in the US, and the NYC government has been corrupt and incompetent for decades.

In particular, one of the things Houston does well is to have less restrictive zoning than most other cities, which allows for mixed-use construction that in turn lets people live closer to where they work. But that's something that helps regardless of what you're using for transit.

Moreover, the level of traffic in Houston is not good. In spite of their zoning advantage their average commute is worse than the national average.

> Ideally, though, cities should stay reasonably small. 300k seems to be the sweet spot from the efficiency standpoint.

There is nobody dictating how many people will live in a city nor should there be. You can tell from the geography of the continent that a city in the position of New York is going to be a massive port, and so it is. Houston, Los Angeles, San Francisco and Chicago are big cities for the same reason. Nevada is mostly desert so Nevada is mostly empty; Las Vegas metro was created from the combination of cheap power from Hoover Dam and proximity to California with legal gambling and it's three quarters of the state population. Inland states are disproportionately farmland.

You don't get to decide where the beachfront property is. The question is only, given that 38M people live in California, how are they supposed to get around?

> Research shows that more transit use does NOT decrease traffic, except in very narrow cases (on arterials immediately parrallel to fast transit).

That's where the traffic congestion is!

> Moreover, over time it leads to MORE traffic, as transit brings in density, and density results in more traffic.

This is the BS "induced demand" theory. What it's really describing is that if you have an otherwise desirable area (e.g. it's within an hour of the ocean) but the local government is mismanaging the area so it's full of traffic congestion or crime or excessive bureaucracy or whatever else, and you do anything whatsoever to make it not suck as much, more people are going to move there.

Fixing the problem doesn't induce demand, the demand was there the whole time and was being suppressed by mismanagement. And you get the same result from anything that fixes the problem. The only way to prevent more people from moving in there is to keep the place a hellscape so people don't want to move in.

Compare this to building a subway in Wyoming where nobody lives and the presence of a subway is obviously not going to cause a population boom there, which is what would happen if "induced demand" was actually a thing.

> I'm OK with giving poor people money so they can THEMSELVES decide on what they can use it, instead of trying to social engineer them by giving them "free" rides. When each ride costs $20 just in op-ex (true cost for Seattle, btw).

The issue is that $20 is the amortized cost, not the incremental cost. The bus costs the same to run whether it has 5 people on it or 40, but if it has 40 then the cost per passenger is 8 times less. And if it currently has 10, the incremental cost of making it 11 -- or 30 -- is zero, so that's what you want the fare to be. Which would in turn cause more people to take the bus and lower the cost per passenger.

> Richer people should pay the full cost of rides.

The incremental cost is still zero regardless of your income level.

> This also applies to cars (although in my state car user fees already pay for 98% of all road maintenance and construction).

It does also apply to cars, but we usually get it right for cars -- the roads (i.e. the fixed cost) are free but you pay for your own gas.

> That's exactly what transit is achieving. It keeps people trapped in poverty, by reducing their economic choices.

The proposal is that you'd have free mass transit, paid for by taxes, which are predominantly paid by rich people. There is nothing prohibiting you from buying a car, which would cost the same as it does now, all it does is make one of your options less expensive than it is now.

Reducing your choices can only come from making one worse than it is now, so that it takes that option off the table. Making an alternative cheaper or more convenient can't do that -- you still have the option to do the other thing and then only reason you wouldn't is if the new option is going to make you better off than the status quo which is still available.

[Mawr]

Oh hey, I actually agree! By all means, let's compare fairly and price in all costs and externalities of car ownership vs public transport. You may not like the result, but that's life.