Hacker News Reader

Comment by nickles | Hacker News Reader

> When it comes to roads, why is increasing capacity to resolve the capacity problem always automatically dismissed?

There's a large body of evidence [0][1] indicating that adding road capacity does not alleviate congestion. In the short run, one tends to see immediate rises in traffic on the new infrastructure. Over a longer period, the amount of traffic utilizing new infrastructure increases until all added capacity is being consumed. This is known as induced demand [2].

Congestion pricing [3] is a more economically viable manner of reducing congestion on shared infrastructure, but it comes with its own set of concerns [4].

[0] https://www.citylab.com/transportation/2015/11/californias-d...

[1] https://www.wired.com/2014/06/wuwt-traffic-induced-demand/

[2] https://en.wikipedia.org/wiki/Induced_demand

[3] https://en.wikipedia.org/wiki/Congestion_pricing

[4] https://www.nytimes.com/2018/04/09/nyregion/congestion-prici...

View on HN

The goal of transportation infrastructure is to move people to their destinations quickly. It turns out that adding road capacity moves more people, rather than moving the same people more quickly, but that's still a success case.

If we should reject road expansion because it fails to decrease travel times, then we should certainly reject a shift towards public transit, which actively makes them much worse.

"Congestion" is a non-problem. You get no points for "solving congestion" by kicking everyone off the road. It's quite likely that a highly congested road is still delivering better end-to-end travel times and user experience than a realistic bus system.

"Solving congestion" is great for you if you can afford the congestion charge, but it's not clear why a democratic society should let the rich kick the poor off the roads for their personal convenience. Really the only angle under which it's a coherent goal is reducing carbon emissions, but then you don't even need to talk about induced demand - it's sort of obvious that more road capacity means more driving.

Why does public transit increase travel times? Good public transport is far more efficient than car-driving, for the simple fact that you can fit fifty people on a bus, and five hundred on a train. This massively alleviates congestion, is vastly more mechanically efficient, and massively simplifies planning.

In a city with bicycle lanes, bikes are a lot faster than cars for short trips, and they require far less infrastructure - since bike lanes are about a third of the width of a road, and they require no dedicated parking space. Not to mention the fact that bikes are far cheaper to own and operate.

I can understand having a sentimental attachment to cars - but in practical terms, they're a ridiculous solution to the problem of mass transit.

>Why does public transit increase travel times?

Because we don't all share an origin, destination, and schedule. Public transit can be extremely effective for the parts of the journey that it covers, but the extreme ineffectiveness of walking the other parts can cancel all that out. You have to completely rethink density and development patterns to get to a point where most origins and destinations are near stations. That would take decades even if we wanted to do it, and SB827 was just defeated in California. Then you have to run trains frequently enough that schedules don't matter. That's expensive as hell and the necessary capital projects take decades. And even if you do them, they might deteriorate and regress later if you don't keep up with the maintenance (see: New York).

When I lived in Berkeley, BART covered 90% of my commute distance, station to station, in 25 minutes. My actual end-to-end commute time was over an hour, basically a wash compared to taking an Uber over the Bay Bridge at rush hour. Bicycling to my local station got it down to maybe 50.

25 minutes is an entire commute the average American. Being just 1.5 miles away from a station doubled that. Public transit has a wickedly hard problem to solve - even 90% coverage is still next to useless.

Realistically, we share enough of an origin, destination, and schedule for the differences to be irrelevant. Most of us go to work at the same time, to the same areas, then go back to the same residential areas. That's exactly why rush-hour congestion happens. If it wasn't like this, cars would indeed be a good solution.

I live in a city with very good public transport. My average tram trip is about 20 minutes, and there is usually one that goes directly to where I want to go. I live within a hundred meters of a tram stop - and this is far from unusual. I also live within a mile of a rail station, and there is a bus stop within a hundred meters.

As I see it, there are two intertwined problems: 1. American cities are built around the car. Homes are distributed across vast areas, in the form of suburbs. 2. Many cities have very bad public transport options. This makes cars more viable, despite the inherent inefficiency of what is essentially a road-train with a hundred drivers.

If you experience transport in the context of a society built around the car, then sure, a car is going to seem viable. Vast amounts of capital has gone into the system that makes it viable. Living and working patterns are all built around it. It doesn't change the fact that fundamentally, physically, a car takes more money to build, run, takes up more space to carry fewer people, requires more infrastructure - and, even given a network of roads vastly more comprehensive than any train or tram system ever designed, works very badly.

Suburbs are also horrible from an efficiency perspective.

Land is a far more precious resource than transportation infrastructure. Relatively small expenditures on transportation yield enormous savings on housing by making more of it available.

> Because we don't all share an origin, destination

We practically used to, until cars encouraged development to spread out.

> You have to completely rethink density and development patterns to get to a point where most origins and destinations are near stations.

In other words, go back to what the world was like before cars.

> Public transit has a wickedly hard problem to solve

But not impossible. European cities seem to have a less difficult time.

Right. Those towns are designed in an era of horse and train. You see this in the U.S. on the east coast. The older the town, and the more developed it was before the arrival of the automobile, the more of the walk, horse, train design you still see. As you go west, you see a large percentage of towns and cities designed based on the automobile, and their "old town" is smaller or has completely vanished.

>European cities seem to have a less difficult time.

European cities had the good fortune to lock in a healthy level of density as "neighborhood character" thousands of years ago, so that the same sorts of structures Americans read as the rapacious excesses of a new era of unfettered capitalism and runaway inequality (i.e. 5-story apartment buildings) are instead seen as historic. They also aren't in a position of having to unwind the "mistake" of certain neighborhood communities existing.

>In other words, go back to what the world was like before cars.

The places we would invade to "go back" to, and the places we would abandon to "go back" from, are now all occupied by humans with voting rights and a preference for the status quo. Not to mention that the population is simply larger, and more concentrated in metro areas. The city centers abandoned during suburban flight were quickly claimed by low-income people of color, and also a bunch of extra people were born in (or migrated from farmland and small towns to) suburbia.

>We practically used to, until cars encouraged development to spread out.

The tenements of pre-car industrial cities were so traumatic that we still have the scar tissue today, in the form of wildly overbearing zoning and building codes. The Progressive movement's obsession with making sure buildings have enough light and air and open space, which holds back San Francisco from even European-level density today, is rooted in the outrage felt towards living conditions 100+ years ago, which the car saved us from.

Obviously there's a healthy middle ground and it's frustrating that we don't take it, but our culture has the memory of what it was like to live too close together, and quite likes the ability to spread out.

It takes riding 3 buses to get where I want to go, but only 20 minutes in a car. The car is climate controlled and plays my favorite music.

Buses are like Open Seating offices, but for transportation - just like sharing one big table at work, there's little privacy on the bus. You're going to learn about this person's skin condition because they're yelling into their cell phone.

The trunk of a car is a mobile secure locker, allowing the car to become a sort of mobile base in a remote city - you can repeatedly visit to stash purchases in the trunk, and have a cooler of food and drinks stored there as well.

  Why does public transit increase travel times?

Well, it doesn't always.

But the journey from Cambridge train station to Coventry train station takes 1h47m by car and 2h16m by train [1], as the train takes a V shaped route via London. Birmingham New Street station to Norwich station takes 3h6m by car, 3h57m by public transport [2]. Kings Lynn to Cromer, 1h7m by car, 2h27m by public transit [3].

And that's going from train station door to train station door in major cities - if you're going to somewhere out-of-town, the public transport penalty can be even greater.

Doesn't mean we shouldn't build public transport, of course. I'm not a car owner, so I rely on public transport to get around. But we also shouldn't deny the downsides of public transport.

[1] https://www.google.co.uk/maps/dir/Cambridge+Station+[CBG],+S... [2] https://www.google.co.uk/maps/dir/Birmingham+New+Street+Stat... [3] https://www.google.co.uk/maps/dir/kings+lynn+train+station,+...

That's down to the fact that there hasn't been anything more than maintenance investment in the english transport system since the 70's. It's probably the most over-priced, ineffective train system in the whole of europe. Partly because when it got privatized, most of the buyers were national train companies from other european states - who use the british train system to subsidize their own trains.

All this means is that capacity is so far behind demand that current spending doesn't come close to establishing a reasonable equilibrium. Which is what you'd expect after decades of major population growth, accompanied by infrastructure neglect.

> All this means is that capacity is so far behind demand that current spending doesn't come close to establishing a reasonable equilibrium.

I'd hesitate to specify that certain equilibria are reasonable while others are not. Building more roads to reduce congestion does not achieve that goal, in part because of the latent demand you describe (see [0] "Effect in transportation systems").

Ultimately, if your goal is to decrease time spent traveling, you will see more benefit from alternative policies. Such policies may include:

* increasing cost of using the roads (e.g. congestion pricing)

* constructing alternate, denser forms of public infrastructure (e.g. subways)

* increasing the density of passengers in vehicles (buses, carpooling, etc.)

* funding technological advances that increase the density of vehicles utilizing the road (e.g. self driving cars)

[0] https://en.wikipedia.org/wiki/Induced_demand

There’s lots of ways to approach defining a ‘reasonable’ equilibrium. You could set a target for average time spent communing, but to address a problem like the one described in the article, you’d just need to build enough freeways to ensure that using a freeway is on average faster than winding through suburban streets.

You could do this by decreasing the efficiency of the non-freeway roads, or increasing the cost of using the roads, which both only ‘solve’ the problem by reducing general quality of life (especially for poorer people, who might just get even poorer if they can’t commute affordably).

Or you can increase road capacity to keep up with population growth, and invest in alternative transport that people actually want to use. That means no cycle lanes or buses, because people really don’t want to use them (buses are only really efficient for short trips, if you have to change buses to get to work then it ends up taking so long you may as well just be inefficiently sitting in the comfort of your own car).

> Or you can increase road capacity to keep up with population growth

This doesn't work because in a city you'll eventually you end up with more roads and parking lots than buildings. US cities were built like this, and the more road capacity you build, the more sprawl you get (everything is further apart), which increases how much people have to drive which increases the traffic... It's an evil spiral of congestion.

In dense city environments, it's just physically impossible to build enough capacity to meet the potential demand for car rides at peak hours without congestions. You may hate public transportation, but than is the only solution that demonstrably work, together with scooters and yes, even bikes where geography and weather make them viable.

Exactly. I live in a small town. We have no traffic jams, and never have. I believe that proves it’s possible to build sufficient infrastructure to eliminate or reduce traffic jams. It’s just that most large cities elect not to.

I don't think the cost of that is justified in large cities. LA already dedicates more land area to automobile infrastructure than anything else, and the region is still plagued by traffic jams. At the end of the day, a single rail line can bring an amount of people in and out a city equivalent to multiple lanes of traffic. That's a lesson Americans forgot after 1945 for whatever reason, and now we have to re-learn it from the rest of the world now that we've hit the limits of car-oriented development.

Dense large cities are economically unviable if you try to accommodate car traffic. The tale of them being cheaper is true only if you give up large living spaces and personal transport.

Or perhaps for a large enough city, you can't build enough roads, because you'd need like >70% of all area for roads and parking.

The scaling doesn't work.

For another similar example, think of how much percentsge space you need in a 2 story building devoted to elevators (almost none) vs a 200 story building (a significant percentage of the middle core of the building is nothing but elevators). Take that and scale it even worse, and you have roughly the small town vs big city traffic issue.

Yep. Induced demand is a nuanced concept that's been perverted by a certain community of activists into "science says you can't build enough roads so let's not build any!!!!1111".

The truth is that you can, in fact, build enough roads to reduce congestion. It may be too expensive. The road land use fraction may be too high. There may be political problems. All of these factors have to be taken into account before deciding to build more roads.

Nevertheless, it is physically possible to build enough roads, and anyone who doesn't acknowledge this basic reality isn't having a serious conversation.

> Nevertheless, it is physically possible to build enough roads

Sure it's probably possible, but by the time we've built that many roads we'd need to cycle from the carpark to the building anyway.

What's your population density compared with a large city?

If roads took up zero physical space, I'd agree with you, but the perverse thing is that as population density goes up, you not only have more people on the roads (requiring more road), but you also have more need for living space, stores, and office buildings (all which take up space, giving less room for roads).

As it is, comparing it to a small town is meaningless.

But: isn't it a good thing?

More traffic is what those roads are for. More traffic means more activity. Focusing only on the capacity / congestion couple is myopic. If you really want 0 traffic just remove all roads: not sure the side effects will be as easily dismissed.

Reply on news.ycombinator.com