Do transportation network companies decrease or increase congestion? | Hacker News Reader

Do transportation network companies decrease or increase congestion?

advances.sciencemag.org

66 pointsbspn78 comments

Link to the study exactly. https://advances.sciencemag.org/content/5/5/eaau2670 Having just read the study the results make sense. The authors look at car ownership reduction. They find that Uber and Lyft are not actually helping with that. We do not observe a meaningful change in car ownership, with an average of 1.08 cars per household in 2010 and 1.10 cars per household in 2016 (36).

This is because as people move towards dense urban centers they are now sometimes electing to use Uber and Lyft instead of transit. The thing is because parking and driving is so expensive as is living in a dense prolific city, you probably wouldn't have wanted a car anyway.

The findings clearly show Uber and Lyft create traffic. The speed data used in this study confirm this trend, showing that the average speed decreases from 25.6 miles per hour (mph) in 2010 to 22.2 mph in 2016 and that the vehicle hours of delay (VHD) increase by 63% over the same period.

In addition to the 20% of TNC VMT that is out-of-service, 70% of San Francisco TNC drivers live outside the city

It makes sense that Uber and Lyft actually worsen traffic. They bring cars and congestion into the city and create public transit competitors.

I have the one for NYC here also http://www.schallerconsult.com/rideservices/unsustainable.pd...

The question is what is the best response to help the environment? It seems that supporting Uber and Lyft is contrary to sustainable objectives. Is that partially why many cities and countries do not want them to operate in their jurisdiction?

The Science Advances paper is here:

https://advances.sciencemag.org/content/5/5/eaau2670

The corresponding author told me it was the basis for an earlier report issued by the SFCTA:

https://www.sfcta.org/projects/tncs-and-congestion

>The question is what is the best response to help the environment?

invest in public transit that can transport people in a denser fashion instead of relying on cars. The fundamental limitation of the car is the low amount of passengers / area it can transport.

http://i.imgur.com/G7cAK4Y.gif

>Is that partially why many cities and countries do not want them to operate in their jurisdiction?

Among other things, yes. But arguably the bigger issue is their inability to comply with regulations, in particular their skewed relationship to nominally indepdenent drivers gives them an advantage over companies who have more obligations towards both customers and workers. This isn't unique to Uber but applies to other "sharing economy" companies like Airbnb

The story for public transportation in most large cities in the US is not awesome.

    1.  Crime
    2.  Forces you to come in contact with people you would not choose to associate with normally
    3.  Public transportation vehicles are often dirty/unclean
    4.  Long waits during off peak times, often, which forces you to revolve your schedule around public transportation
    5.  Commutes are often longer
    6.  Can't get to your destination often without a long walk, or another vehicle
    7.  Usually still requires walking in the rain/hail/snow/sleet/wind/sub-zero temps
    8.  Initial build out and future expansion are very expensive.

Unless those are addressed in new buildouts, public transportation will continue to have the reputation of public housing and public schools in many places.

I agree that many of your points present legitimate difficulties with the use of public transit in the US as it currently exists. However, I think that the first two points are just as true for ride share as they are for public transit. Ride share also has crime (see sexual assault scandals https://money.cnn.com/2018/04/30/technology/uber-driver-sexu...) and ride sharing (particularly pool-style) also "forces you too come into contact with people you would not normally choose to associate with"; some people (including myself) see this as a benefit of public/shared transit, not a drawback.

Part of the problem is that public transit scales differently from ride share. The more riders public transit has, the better it gets. The more riders ride share has (beyond a certain point) the worse it gets.

I'll disagree with you on safety in ubers/lyfts. But even so, the rest of the list is pretty terrible.

This was the number of reported assaults in 2017 in NYC subway systems:

https://www.metro.us/news/local-news/new-york/nyc-subway-sta...

"Assault and related offenses were also frequent, with 1,243 reports within the subway in 2017, as was harassment, which accounted for 1,003 crimes."

Data for assaults against passengers in ride hailing (taxi/uber/lyft) services seems hard to come by, but there were 103 reported incidents of rape nationwide in ubers from the years 2014-2018.

https://money.cnn.com/2018/04/30/technology/uber-driver-sexu...

It's worth noting that Buzzfeed said there were was a much higher number, but that was disputed by Uber directly.

Make public transport remotely competitive with private transport on a time basis and people will flock to it in cities. In cities where public transport is effective (many European capitals and pretty much just NYC and select parts of DC and Chicago in the US), people of all strata of life use it because it's time effective to do so. Because "everyone" is using it, it feels (and usually is) reasonably safe, clean, and "normal".

In other cities in the US, it's so broken that people with time pressure and money in their lives don't use it (and so don't particularly support it from a policy/voting perspective either). Predictably, it's no longer "normal" and feels less clean and safe (and might actually be so).

Sadly, that ship may have already sailed in most North American communities. A this point, they're _very_ heavily engineered to favor private vehicle ownership. Many of the ones I spend time in are so sprawled out that it takes me more time to get my errands done by car than it does for me to do them on foot in my home town.

(And then you need to waste even more time going to the gym regularly in order to make up for the desperately sedentary lifestyle that's been imposed on you by your habitat. . . but I digress.)

In a place like that, public transit is just doomed. There's no density of bus stops that will work. Either everyone lives at least a 10 or 15 minute walk from the nearest bus stop (possibly without even a sidewalk to walk on), or the stops are packed so densely that it's almost faster to walk than it is to take the bus, or the routes are packed so densely that the buses end up being a sort of hyper-expensive 4-passenger vehicle, on average. And you can just forget about any rapid transit options.

Did the report account for more road work and construction between that time? Anecdotally, I've noticed more constructions now than before. Maybe that also contributed to the slow down in traffic.

I think the environment question is easy, go forwards not backwards. Let's expedite the process of converting all those Uber and Lyft vehicles to full electric. Doing this allows consumers to get what they want while also protecting the environment.

I can't speak for US cities but in the UK outside of rural areas the answer is pretty much universally public transport.

London has good public transport already. Now imagine if we doubled fuel tax / parking charges / etc and spent it on putting on more buses, building more cycle lanes, having express buses go around the A406/M25/etc.

It'd be so, so much better.

> The new findings, based on a computer model that simulated the speed of traffic with Uber and Lyft vehicles removed

Not sure if that's actually what's going on, but if it is, it's a ridiculous conclusion. Let me rephrase: "Japanese cars are making traffic in the USA worse, based on a computer model that simulated the speed of traffic with Japanese vehicles removed."

Well, no.

Japanese cars are not otherwise distinguishable by their frequent mid-block stop-and-wait actions, for instance, nor for diverting commuters from public transportation.

If you remove any amount of cars you improve traffic which is why the conclusion is weak.

The problem is that in SF those people were already using taxis in a lot of cases but people were limited by price. By making taking a taxi cheaper, you are practically guaranteed to increase the number of cars in flight at any one time.

I don't understand why they don't just replace the Uber vehicle with a "normal" driver then draw conclusions from that. How much worse do they make traffic compared to a "replacement driver"?

That's fair, but a certain percentage of Uber and Lyft customers are people who would otherwise drive their own car, so any study that removes the effects of Uber- and Lyft-operated vehicles altogether paints an inaccurate picture.

The point is not (so much) the number of cars, the point is the different behaviour. Self-drivers will go as smoothly as they can from one parking location to another. Uber and Lyft drivers will loop until they get a ride and stop disruptively at the beginning and end of each ride.

Their behavior is better than people circling endlessly looking for parking. If traffic is worse, but more people are getting the places they want to go faster, isn't that an improvement in the transportation system?

Stop disruptively? What does that even mean. Do taxis also stop disruptively according to your definition?

> Uber and Lyft drivers will loop until they get a ride

I ride Uber almost every weekday twice a day. I'd say about 80% of the time the driver has their next trip queued up before my trip ends.

What Uber has done for cities is reduce the number of people who drive in to work and then spend 15 minutes looping around looking for a place to park, possibly multiple times each day.

It really hasn't. People who drive to work park in monthly lots, usually with assigned spots. Uber replaced people driving to restaurants and bars.

Either way I'd call that a net win for society. Where's the paper that estimates how many dui-related deaths Uber et al has prevented?

There are several, including papers from Uber.

https://www.nytimes.com/2017/04/07/business/uber-drunk-drivi...

> Stop disruptively? What does that even mean. Do taxis also stop disruptively according to your definition?

Yes. But the number of taxis has not exploded in 5 years.

Read the actual paper. They seem to have used two methods to make the alternative ("counterfactual") scenario more realistic than just "remove Uber cars":

- They compare increase in congestion and Uber/Lyft activity for short road segments, and show that these two changes correlate

- They compare reality to a model that does not include ride sharing, but does model increased population etc. This (generally well-regarded model) shows lower increases in congestion in a world without ride sharing.

Increases in congestion at popular drop off / pick up sites are not necessarily bad. It could be that dropping people off is the new bottleneck because the rest of the road network is less congested, and now there are more trips (i.e., the road is providing more value because it is more efficient).

A study that measured door-to-door time, and the number of passenger-miles per hour would be much more useful.

I’m not sure why traffic engineers refuse to use metrics that are directly related to the quality of the experience for the people using the roads, or the economic value delivered.

A bit unfair, as you neither have read the original paper, nor even given the benefit of the doubt to the possible deep meaning of "computer model that simulated". It's not like academia is this stupid or peer reviewed papers are that low quality.

There is a link in the article to the original paper, here is if you want to investigate more:

https://advances.sciencemag.org/content/5/5/eaau2670

Material and methods chapter might be of interest for you.

It's not just the number of cars causing issues; the number of roads with ad hoc lane closures due to ride sharing pick-ups and dropoffs makes a huge difference. Many three lane streets get choked down to one lane if a delivery truck blocks a right lane and a ride share blocks the left. It's very visible every rush hour in SF.

In HK I believe that city centers have designated areas around town for taxi pickups for this reason. It seems to work quite well.

Anything that makes it more convenient for people to get from A to B in a _car_ will, of course, create more traffic. Convenience increases usage.

Uber/Lyft may make it unnecessary for some folks to own a car, but a reduction in car ownership does not mean a reduction in car usage.

One thing Uber/Lyft do for sure is reduce the need for parking spaces, which are at a premium in SF.

Not having to worry about parking would therefore be an incentive to go somewhere by car (Uber/Lyft) instead of public transportation.

Not necessarily. Making car-pooling easier can reduce the numbers of cars while also making it more convenient to get places by car.

Obviously. They are making it worse by exactly the number of trips that would otherwise have been done by transit, bike, walking or not at all.

edit: or now scooters

Actually, you take that number, PLUS the number of empty Uber cars, PLUS the number of empty taxis that would not have been empty if there wasn't Uber...

Empty taxis drive around looking for a fare. Empty ubers park and wait.

In my experience empty Ubers identify the nearest hotspot (usually from experience) and drive in that direction.

That just isn't true. They often do drive around and look for fares or go back to a place where they can collect a fare from where one did not exist. Example you pick up Passenger A in dense urban center drop them off in suburb no fares exist there. You need to drive back to the urban center to pick up another Passenger. You have worsened traffic and created extra carbon pollution.

They're empty sometimes. Even if they park at some spot.

I don't think it is so simple. Eg: people now taking extra trips because it is easy and cheap to do so. Eg: "I'll just Uber the 6 blocks to my next shopping spot" instead of walking. This happens in my friend group all the time. Plus people traveling to the city at all BECAUSE Uber makes intracity travel easy. Eg: I live in the East Bay and hate driving within the city, but am happy to be escorted around it on a whim but would have otherwise stayed more local.

The comment specifically mentioned non-consumption, i. e. "Not at all".

If the city was denser you wouldn't have to uber from one shop to the other. But the city was built for cars.

Oh for sure. I see so many areas in the city where I reimagine as pedestrian only and basically weep inside for the lost opportunity.

Clearly a trip that is taken has utility to the person taking the trip, so the "not at all" outcome is a bad one.

People will use the best option they have out of the ones available to them. Transit is usually faster than Uber for trips to downtown or the Mission, but try getting to Miraloma with it.

Every trip has a cost. It's a transaction. If it costs me $1,000,000 to get to the park (and it's fairly likely I won't make $1M at said park), I will maximize utility by not taking the trip.

In addition, every trip by you, has various costs to everyone else, namely the traffic congestion we're talking about here, plus many other things we're not talking about like pollution, publicly-subsidized police/EMT/fire for cleaning up your car wrecks, etc. Transport modes that allow you to take up less space on the roads per person, or that use separate parallel networks, or that don't pollute or involve wrecks, cost everyone less on these particular axes.

Also every driver I meet in SF has driven from somewhere else to "work" in the city.

There's also a larger impact for cars that block traffic by stopping in the lane in front of a pickup or drop address where there isn't room to pull over.

And there's an impact from out-of-town weekend drivers that don't know where they're going and take a dumb route because of that unfamiliarity.

Funny you say that because most of my Uber/Lyfts consist of me going to and from public transit for my intended destination.

Although it is more complicated than that, as I often take a shared ride home because Bart shuts down too early on weekends to be useful on the way back.

Additionally the math gets complicated when I have multiple people with me, as often times it's cheaper to uber/lyft somewhere to our destination than it is for all of us to individually pay for a Bart ticket.

This is called multimodal transport. I'm unsure how big the effect is for Uber/Lyft where booking the ride would seem slightly more involved than, say, using a bike. But it's an area of great interest especially for public transit planners.

I'm only comparing what happens in the world we have, where Uber & Lyft exist, vs. what would happen if they didn't, or would've happened back when they didn't. Obviously some of this is guesswork kind of like "Gee what would've happened if I'd married my high school crush" or whatnot. But in general if you would've walked or biked to transit before, and now you take Uber & Lyft, that makes congestion worse. If you would have driven to transit, using Uber/Lyft has no effect. If you would have driven the whole way, it's a partial improvement (by riding transit for that one segment).

And, following this, these different effects probably have different values for society, i. e.:

- Replacing public transit = unmitigated bad

- Replacing walking, bike = even worse, bikes are unbeatable for health, space efficiency, energy use, noise, and really any other category

- Replacing using own car = close to neutral, maybe slightly positive, especially when it allows people on the margin forgo car ownership entirely

- Non-consumption = likely positive, taking a trip you otherwise wouldn't take is a sign of increased freedom/participation in society and would tend to outweigh environmental concerns in my subjective judgement

So, we consider people being able to do something in a way they would much rather do it as a negative? If you have the choice between walking and an Uber, and you really want to walk, then just walk. If you have the choice between walking and walking, but you really don't want to walk, then you're worse off.

Yes, there are cars that are performing transport and some that are part of traffic that are waiting to be hailed and not performing transport - those are certainty congesting. There is also the effect of rideshare drivers coming from different areas and thus not knowing the city they are in. Reliance on GPS/map apps helps but there will still be “I don’t know where I am going and am not accustomed to city traffic” effect that compounds the others drivers who exhibit that behavior. I’m sue we’ve seen the effect of a single driver desperately trying to cross lanes to make their turn, scale that up.

Cities should really get this data from the companies to figure this out. That should be a necessity for operation in the city.

Also drivers circle a block every time they need to add a passenger. The one way streets may make left turns safer, but they increase the time a car has to be on the road in any given area.

I found it ironic that Lyft/Uber has been a godsend to the Phoenix area. Cabs were always horrible / impractical here.

When I'm in SF and other major cities, I tend to lean towards cabs more.

Anytime there is no passenger in the vehicle, it's an extra vehicle on the road. Normally, our cars don't continue to drive around after we've arrived at our destination.

Your San Francisco parking experience is very different from mine.

Many of the U/L trips happen during commute hours, where someone might have a dedicated parking spot on either end -- either at home, or a monthly garage or lot.

Good to see this. Earlier assumptions that rideshare would reduce traffic ignored how the transport system would change. Also similar to the way Airbnb can make housing more expensive or unavailable.

I live in San Francisco. Public transport is a joke. Buses are so dangerous that I won't take them (took them twice when I moved here and almost got beaten up the first time, witnessed someone getting mugged the second time (near tartine)).

I went from exclusively using public transport (previously living in UK/France) to exclusively taking uber/lyft. Thanks humanity my work provides shuttle to do the commute to south bay, I would never be able to do that with the caltrain. Mad respect to those who have to do it.

You can probably tell I'm salty about the state of things here. As a matter of fact I just passed my driving writting exam today because for the first time of my life I feel like I need a car.

I feel just as safe on Muni as I have on the handful of other American transit systems I've lived near. Feel free to read that either way :)

Polls have been pretty constant at least https://www.sfmta.com/reports/customer-rating-feeling-safe-a...

I remember when Uber was “supposed” to be a ride sharing service where you could make extra money along your intended destination. It wasn’t touted as a taxi service.

I remember before that, when it was touted as the more convenient and cheaper alternative to black car service (which was already cheaper than taxis for airport runs!).

Maybe that’s true of Lyft, but Uber started with black cars and originally had the tag line “everybody’s private driver”. Hell, the original name was UberCab. UberX cane pretty late in the game.

Presumably this is the standard problem with Uber/Lyft:

Uber/Lyft make the vast majority of their income in areas that are already well-served by taxis. Consequently, they make the problem worse.

Presumably, in the cities not well-served by taxis, they improve the situation. However, they don't generate anywhere near the income.

NYC area driver here... TLDR the entire article but I'll give my first hand perspective to say that "T" (taxi) and "L" (livery) plated vehicles on the highways in the NY metro area DO significantly add to congestion as the drivers keep their speeds well under the limit while traveling in passing lanes. Frequently I see these non "cab" "T" and "L" vehicles that I assume are "transportation network" company contractors driving unusually slow because they're either on their mobile devices (looking for pickups?) or driving so under the limit that I assume to not break any GPS speed monitoring rules enforced by these "transportation network" companies, or just generally inept.

I'm not advocating speeding by any means, but after driving 500k+ miles in the NYC metro area you can spot the ineptness and/or driving under fear of speed monitoring of these "T"/"L" plated vehicles. All it takes is for one to be driving 45 mph in a 55 zone on a two lane road with an already congested merge ahead, or they travel in the left passing lane at speeds well under the limit preventing other motorists from advancing, or overall poor driving, then it's a chain reaction of braking and delays mounting behind them.

I've always assumed that the monitoring of these drivers adherence to speed limits is the reason why they travel noticeably slower than the average driver in normal traffic. If that's the case it unfortunately causes the average capable driver to become indirectly part of the "transportation network"'s speed monitoring resulting in delays for everybody.

In the relative absence of these "T" and "L" plated vehicles, I've driven in moderate to heavy volumes where the average speeds are in the 60-65+ MPH range. However, when the livery vehicles start to increase within similar volumes of traffic the average speeds seem to drop and the congestion related delays seem to increase.

So if you were to ask me "Do transportation network companies decrease or increase congestion?" My answer would be "yes". Why? Based on my observations, I suspect it's related to the speed monitoring of the "transportation network" company and a tendency of the drivers of these vehicles being inept relative to other non "transportation network" drivers.

The Uber/Lyft driver strike a couple days ago was a good demonstration of this; my drive to work was significantly easier that day, as was the drive home, with noticeably fewer cars on the road.

Why can't they just teach drivers to leave a bunch of space in front of them and help improve the traffic? Use the app to reinforce it...

They certainly make Manhattan traffic worse.

On the flip side, car owners make traffic worse for Lyft and Uber users.

I'd rather we penalize normal car owners than penalize Lyft and Uber. I use Uber Pool every day to get to my bus stop, and there's no good alternative.

Alternative: More Parking

Alternative: Public Transport. Works well in large cities. Subways easily achieve double the average speed of cars in dense areas or rush hour. 30km/h vs 15km/h are typical figures.

Please no

Url changed from https://www.sciencemag.org/news/2019/05/uber-and-lyft-may-be..., which points to this.

> Uber and Lyft have long said their services may relieve urban congestion by facilitating access to public transit

This was always questionable for a place like SF or any high density city.

Few people take a ride share to get to a bus or a MUNI light rail stop in SF. If you are paying for the Uber/Lyft, why would you pay again for a bus ticket, especially given the inconvenience of then waiting for the bus. You'll just stay in the car to your destination, after all San Francisco isn't that big.

The exception is perhaps people in SF getting to BART to go to a suburb, but that's not a very significant use case.

Edit: removed potty language, and superlatives.

I use it this way all the time. Bart into the city from a suburb and then if it's not in walking distance and there's no convenient bus, I take a Lyft to my destination.

So do I, but that's a different use case than people within SF using ride-sharing.

I said few journeys go from ride share to bus/Muni, excluding BART, which goes a much longer distance than bus/Muni and therefore serves different types of journeys.

And journeys originating from within SF probably still account for the majority of ride shares in the city, not journeys from outside SF.

Many people at my company take Uber/Lyft to get to BART and Caltrain.

Reply on news.ycombinator.com