End of the car age: how cities are outgrowing the automobile

[BrownTown]

Originally Posted by Gordo

I get that a transit union may think that, but if we're intentionally holding improvements in transit back merely to maintain a few jobs in the transit union, well, that's absurd.

Unfortunately that's exactly what happens not just in public transit, but across the whole spectrum of the government. Another good example would be automated container ports which are used in many other countries but which the unions in the US successfully lobby politicians against. Unions always fight against automation and public sector unions are the worst because their jobs can't be shipped overseas the way that most of the old union jobs have been.

[Eightball]

Originally Posted by dchan

From what I've read, automated trains work well and quite reliably for single-line systems, but engineers and programmers can't figure out how to make them work reliably for multiple-line systems.

In the example of NYC, they're good for the L train (which has been completed automated, technically speaking*) and the 7 train because both run on their own dedicated lines. Automated train lines would be much harder program for, say, 47-50 Rockefeller Center station on the 6th Avenue IND line because the 4 lines that stop here split up and go to three different stations north of this station.

* As you inferred in an earlier post, even though the trains can be run completed automated without the need for a driver, the L train continues to use both a motorman driver and a conductor because the transit workers unions refused to allow the MTA to eliminate the motorman jobs.

Link showing that is why they kept drivers?

The red line in dc was just re built to once again utilize ATC but they kept a driver on each train in case of need of override and also to monitor the train IE yell at people blocking doors, creating nuisances, provide assistance in emergencies etc that seems like it will be a vital function in the future as well so I don't know if savings will be realized there

Going auto ensures a much smoother subway ride but its no panacea. The red line crash in 2009 was running in ATC but WMATA had installed an track circuit component which did not fit the system. There actually was an operator on the train who was able to start braking the train, reducing the speed somewhat before the crash. This likely saved lives.

http://en.wikipedia.org/wiki/June_20...rain_collision

[eschaton]

Originally Posted by BrownTown

Unfortunately that's exactly what happens not just in public transit, but across the whole spectrum of the government. Another good example would be automated container ports which are used in many other countries but which the unions in the US successfully lobby politicians against. Unions always fight against automation and public sector unions are the worst because their jobs can't be shipped overseas the way that most of the old union jobs have been.

I'm not going to pretend I know the issues in the individual countries, but Western European countries which have implemented automation of transit have much stronger labor movements than our own.

Part of the reason why the fight may be more winnable there is in many of these countries government is more serious about a "just transition" out of a declining area of the economy. Sweden in particular has been very forward thinking about ensuring as it moved away from heavy manufacturing it offered comprehensive retraining and good job options. Union density has fallen a bit, but it remains among the highest in the world at 68%.

Regardless, you cannot blame the individual labor unions for fighting to represent their workers. Not only because it's in their self-interest to do so, but also because they legally have the responsibility to represent their workers, and can face lawsuits which could cost them millions if it's found they violated the "duty of fair representation." But within a U.S. context it's pretty clear there is no clear plan where the millions of workers who will lose their jobs in transit (union and non-union) will go.

Personally I believe in another generation automation will be so widespread that the idea of "full" employment is going to need to be discarded. Either we move to dramatically shorten the work week, and/or we will move to a system where there is a basic income/guaranteed minimum income.

[Beedok]

Originally Posted by Gordo

It's an absurd reason because transit agencies should not have their primary goal be to employ as many people as possible. I get that a transit union may think that, but if we're intentionally holding improvements in transit back merely to maintain a few jobs in the transit union, well, that's absurd. The transit agency, as a part of the government, is supposed to do what's best for the population at large.

High unemployment is not good for a society. Laying off hundreds of good paying jobs isn't going to be good for any city's economy.

[Gordo]

Originally Posted by Beedok

High unemployment is not good for a society. Laying off hundreds of good paying jobs isn't going to be good for any city's economy.

The economic inefficiency and dead weight caused by congestion from substandard transit is going to be orders of magnitude more important to a city's economy than a few specific jobs.

[Beedok]

Originally Posted by Gordo

The economic inefficiency and dead weight caused by congestion from substandard transit is going to be orders of magnitude more important to a city's economy than a few specific jobs.

But automating the transit doesn't really effect congestion. It might slightly improve the scheduling, but considering that most well run transit agencies get like 95%+ on time without automation which is about the same number as many automated systems. The only reason big Toronto's been thinking about it is to get better lined up at the stations to put suicide barriers up. I've never heard anything about efficiency or congestion issues being improved. Now automating personal vehicles would help with congestion, but professional transit operates (mostly) know what they're doing well enough. Laying off hundreds or thousands of people will probably make a bigger difference in the economy than shifting reliability from 95% on time to 96% on time.

[mhays]

It should be obvious that money saved on labor costs can go to other things, which might add jobs.

The tougher part is that driving jobs can be filled by people with minimal education or training (there's some training of course), and there aren't a ton of decent-paying jobs like that left.

[eschaton]

Originally Posted by mhays

It should be obvious that money saved on labor costs can go to other things, which might add jobs.

The problem is, this isn't going to happen in isolation.

Newer, low-cost manufacturing robots threaten to eliminate essentially all of the remaining jobs in this industry. Amazon and other companies are experimenting with thousands of robots in warehouses. Robots now work in hospitals dispensing and delivering medicine. Self checkout is going beyond grocery stores and into restaurants. Researchers are working on fruit-picking robots which will replace agricultural workers. Robotic systems are even being developed to take over cooking.

This is just considering the hardware side of things as well, not the software side. Internet shopping was in many ways the replacement of many retail jobs with lot of software (and a small amount of software developers) which may be part of why retail employment has been largely stagnant since 2000. Customer service AI continues to get better and better. With many companies, unless you have an unusual issue you never talk to a human being. Stock reports are already written by computer. According to Moore's Law, the processing speed of computers doubles every two years. By 2020, we should have computers with more raw processing power than the human brain, and soon surpass this by many orders of magnitude. Even if real AI isn't possible, the question becomes what cognitive skills won't a computer be able to do as well or better than a human. Graphic designer? Actuary? Lawyer? Engineer?

Sorry, I know this is a bit afield from the original post. But the fact remains that the looming crisis isn't just limited to transportation-related fields. It's across the entire economy. Thus it's not just a matter of figuring out how to shunt a few million workers into more productive fields, because it's unclear in the longer run what fields it will still be productive to be employed within.

[Gordo]

Originally Posted by Beedok

But automating the transit doesn't really effect congestion. It might slightly improve the scheduling, but considering that most well run transit agencies get like 95%+ on time without automation which is about the same number as many automated systems. The only reason big Toronto's been thinking about it is to get better lined up at the stations to put suicide barriers up. I've never heard anything about efficiency or congestion issues being improved. Now automating personal vehicles would help with congestion, but professional transit operates (mostly) know what they're doing well enough. Laying off hundreds or thousands of people will probably make a bigger difference in the economy than shifting reliability from 95% on time to 96% on time.

I wasn't just talking about congestion in terms of trains being late, but also about congestion in terms of too many people to fit on the trains at the current throughput. Full automation of the the Yamanote line in Tokyo (by far the busiest train line in the world) is in planning stages now, and is expected to allow intervals down to less than two minutes, from the minimum 2.5 minutes now. That's a huge increase in throughput that will help with congestion in stations throughout the system (especially waiting areas).

Additionally, increased capacity at little or no cost has huge effects on how many people can rely on a transit system, which in American cities could help substantially with road congestion if it caused even a small percentage of modal shift.

[Beedok]

Originally Posted by Gordo

I wasn't just talking about congestion in terms of trains being late, but also about congestion in terms of too many people to fit on the trains at the current throughput. Full automation of the the Yamanote line in Tokyo (by far the busiest train line in the world) is in planning stages now, and is expected to allow intervals down to less than two minutes, from the minimum 2.5 minutes now. That's a huge increase in throughput that will help with congestion in stations throughout the system (especially waiting areas).

Additionally, increased capacity at little or no cost has huge effects on how many people can rely on a transit system, which in American cities could help substantially with road congestion if it caused even a small percentage of modal shift.

Going from 2.5 to 2 minutes is only really an issue for the most absolutely mega-used lines. 2.5 minutes is good enough for basically anywhere that isn't Tokyo (maybe a couple of other mega-cities have a line or two that could use it, but that's basically a non-issue for cities in general). Considering that Vancouver isn't much cheaper than anywhere else in Canada for transit I don't think there's major savings. The cost of buying the systems (especially for anything other than a totally grade separate rail line) and then buying new ones every few years, isn't a huge savings on people (especially with the more expensive technicians being hired to program and whatnot). Then there's the issue that most transit systems run at a lost so automation probably won't result in an increase in service or decrease in fares, just a slightly lower loss for the city assuming it really is cheaper.

[mhays]

Maybe Vancouver has a better fare recovery percentage in terms of operation costs. With a relatively new system and no drivers that would make sense.

[Gordo]

Originally Posted by Beedok

Going from 2.5 to 2 minutes is only really an issue for the most absolutely mega-used lines. 2.5 minutes is good enough for basically anywhere that isn't Tokyo (maybe a couple of other mega-cities have a line or two that could use it, but that's basically a non-issue for cities in general). Considering that Vancouver isn't much cheaper than anywhere else in Canada for transit I don't think there's major savings. The cost of buying the systems (especially for anything other than a totally grade separate rail line) and then buying new ones every few years, isn't a huge savings on people (especially with the more expensive technicians being hired to program and whatnot). Then there's the issue that most transit systems run at a lost so automation probably won't result in an increase in service or decrease in fares, just a slightly lower loss for the city assuming it really is cheaper.

Going from 2.5 to 2 minutes would be massive for San Francisco, because the chokepoint for the whole system is the transbay tube and the first couple Market St stations and approaching max capacity at just over 2:30. It would be huge for several NYC lines. There are several places in DC where going to that frequency at peak times would be helpful. Even if it wouldn't help anywhere else in the US/Canada, we're already talking about millions of riders from those three.

Vancouver may not be much cheaper, depending on how you measure it, but it is much more frequent than other lines with similar levels of ridership. Quality of service for the users has to matter for something, especially on a government-provided service where users are paying fares + taxes. Who knows how much quality of service of SkyTrain has helped increase development along the lines and/or support for increased funding for expansions. Each person spending a few minutes less waiting for a train does count for something too, if we're talking about metro-wide productivity.

[st7860]

The transit system in vancouver has the lowest cost per km of most major Canadian cities. The labour cost of drivers is considered one of the primary considerations that is why in vancouver on the automated main line headway can be often as little as every 60 seconds during rush hours. Of course light rail systems can't achieve such a headway , like the light rail system in Calgary , however if it wasn't for the cost of drivers , the headway in Calgary could still be improved quite a bit .

In some parts of the vancouver system, you can see trains come as often as every 30 seconds when they bunch up in downtown. So the headway in vancouver is sooner than anything in Tokyo but the trains are shorter

[BrownTown]

Originally Posted by Beedok

High unemployment is not good for a society. Laying off hundreds of good paying jobs isn't going to be good for any city's economy.

It certainly IS good for the economy to lay off hundreds of high paying jobs if those people aren't being productive since they are doing a job a computer could just as easily do. The opportunity costs of paying them is that the city can't pay other people who would be even more productive. If the city laid off the transit workers it would just mean they have more money to hire more teachers or maybe use the extra money to build new transit lines which would employee construction workers etc.