From the comments: Public transit’s problem is overstaffing, not wages

Saying that it’s more important to reduce overstaffing isn’t really an argument against reducing overpayment, nor is the accusation that this is just spiteful union busting with little practical benefit.

Say the MTA could increase productivity enough to cut 27,000 workers and thus save $2.7 billion a year. Cutting total annual compensation per employee from $100,000 to $75,000 — which would still be way above market rates in a city where cab drivers make about $30,000 a year driving well over 2,000 hours a year — would still be an extra $500 million annually. That’s a serious amount of money, not chump change that people only wish to take to be spiteful.

I realize there’s a serious debate to be had on the costs and benefits of measures that wildly increase the pay of less skilled workers, but Levy needs to realize that those who disagree with him have reasons beyond being nasty.

His comment also indicated that NYCT’s pay is in line with international standards.

Staffing is really fucked in San Francisco, the largest and densest city to lack a rail backbone to its system. (the rail lines do not carry more passengers than any other line, they are NOT the backbone of the system and buses do not feed into them. They are also barely higher capacity than an articulated bus because they only run 2-car trains) It should be no shock that their farebox recovery is like 25%, which for the second densest major city in the US is a huge shame. It also means service at night sucks.

Does anyone know where we can find staffing levels per passenger numbers?

There are no “international standards” for paying low skill workers. There are only local labor markets. The MTA need not worry about workers bolting for Japan if it cuts labor costs.

Even if your argument stems from your feelings about “fairness,” international pay is neither here nor there because living costs vary. An effort to pay MTA workers “fairly” would make decisions based on NYC living costs, not what folks in Japan happened to make.

“You say there are reasons beyond just being nasty, but you don’t articulate what they are.”
Hundreds of millions of dollars that could be saved annually, as I illustrated above. That’s money that could go toward reducing the MTA deficit, improving service or cutting fares. You can argue that paying people well is more important than any of those things, but those things are actual goods beyond nastiness.

“I’m also curious how you argue that their pay is above market rates – on what basis are you making that claim?”
If you paid a market clearing rate, you’d get about five qualified applicants per job opening. The MTA gets several hundred. Unskilled workers, even in NYC, make about $40,000 a year in total compensation for 2k hours of labor in industries with free entry. MTA employees tend to make twice that for less work. (True, not all of them are unskilled. Train mechanics have a very specialized skill. But the vast majority of MTA employees would be considered unskilled or semi-skilled.)

“I don’t find the cab comparison particularly illustrative – you’re comparing MTA total compensation to a cabbie’s salary.”
I’m comparing it to a cabbie’s total compensation. The overwhelming majority of cabbies are freelance businessmen who pay a set amount to rent a cab for a day and take home the money they can generate, less the costs of the cab. They get no salary and no benefits. And they do exactly the same thing as bus drivers, who constitute a very large chunk of the MTA’s workforce. But they do if for a third as much money, which again speaks to the market clearing wage for that labor.

Overstaffing is not a problem on the bus drivers end of the scale it is a problem at the top end where we have managers, VPs and 3 letter names galore. 20 dollar an hour bus jockey is nothing compared to a 400k a year position that involves little more skills than could be found in a state college graduated Urban Planning major.

According to a story in this week’s New Yorker, which is about as fact checked as any publication in the world, the average NYC cabbie makes between $25k per year and $38k per year. Story: The Thin Yellow Line. Author: Lizzie Widdicombe. Issue: April 18, 2011.

Staffing per passenger is usually very easy to find; virtually every agency has publicly available passenger numbers, and many say how many employees they have. Staffing per train-hour or bus-hour is also reasonably easy to compute, but usually requires you to add up numbers from the schedules (which is what I did for Toei and Tokyo Metro). The challenges are separating labor expenses from the rest, and finding the breakdown of the workforce by position.

I’m not invoking Japan to argue that unionized workers would bolt for Japan. That would be stupid. Rather, I’m invoking Japan to argue that the work involved in running an efficient transit system is not low-skill. Train and bus drivers need to drive safely and maintain consistent schedules and headways. Cleaners and maintenance workers need to work quickly to minimize the time equipment is out of service.

The actual fair wage depends on the precise mix of jobs each city has, and it’s entirely possible New York needs less skilled people than Tokyo. I don’t know – I’m not an international labor consultant (thank God). But the comparison to Toei is a lot fairer than a comparison to cab drivers.

I’m sorry to have misinterpreted your argument, but I think the argument I ascribed to you would be far stronger than the one you are making.

Actually, I can’t really believe you’re arguing that janitorial work and driving aren’t low skill jobs. They’re pretty much the definition of low skill jobs. Train driving could be done by a machine 20 years ago because it consists of almost nothing but following orders about acceleration and deceleration. Driving a cab is harder than driving a bus, even though the vehicle is smaller, because you have so many more tasks to perform: looking for passengers, planning routes, looking for addresses, listening and reacting to the scanner. And even that is still low skill work. (The maintenance workers you mention may or may not be low skill. Maintaining a train is a high skill job. Keeping graffiti off station walls is a low skill job.)

To argue that these folks are doing difficult jobs and that the quality of job performance would suffer if you did not pay high enough wages to attract truly bright folks is pretty implausible.

On the other hand, to argue that basic fairness demands there be ways to make a comfortable living, even if God didn’t bless you genetically or with parents who could open doors for you, is an argument that a lot of people are going to find compelling.

Sir:

Alon Levy fails to support his thesis about staffing levels – and I am certain that he can do a better job than he did.

First: The National Transit Database does not confirm 47,000 NYCT subway employees. NTD data for 2008:

(Annual average) Operating employees, full-time: 28,323.2.
“Capital” employees, full time: 3,978.1.

Operating employees, part-time 318.4

Unadjusted total: 32,619.7.

The ratio of annual work-hours to employees for the full-time operating employees is 1,721.6. That for part-timers is 765.1. Using these to convert the part-timers to “full-time equivalents:” 141.5.

Thus, the “adjusted total” is “about” 32,442.8; I would state this as 32,400, based on the (im)precision of the “quick and dirty” adjustment method. This total is nearly 31 percent less than the figure stated by Alon. I emphasize that Alon might well be correct – but he would need to explain the large disparity with NTD data.

6,400 employees for the Toei subway division is much too high – Toei – like NYCT – also operates buses (Toei also has a short monorail and an AGT line). I found a blog page (in Japanese) that states the following (the blog post is dated 2011 February):

Tokyo Metro: 8,379 employees, average age 38.9 years.
Toei subway: 3,420 employees, average age 42.5 years.

Next, as my grandmother might say, anyone who compares the Tokyo subway systems to New York’s should have their head examined. There are many pitfalls waiting to trap the unwary. For example: NYCT cannot achieve the same “virtual productivity” that Toei does, for several reasons. Consider, for example, that NYCT has to maintain all of its cars (… with its own employees …) – and Toei does not.

Sounds like a bit of blarney, doesn’t it?

At 2008, NYCT had 6,410 cars available for service, of which 5,288 were “vehicles operated in maximum service.”

Toei owned 1,086 cars – but “used” a significant additional number of cars from connecting rail lines. This is one of the interesting details of the Japanese practice of through-working between subway lines and connecting railways.

Each operator provides crews for all trains which operate over its tracks. Crews are usually exchanged at interchange stations between subway and railway lines. These interchange stations are usually staffed by the operator handling the larger ridership, and the other pays a fee for use of the station.

Tokyo Metro and its partners compensate each other for use of each other’s trains on a flat-fee lease basis. Toei and its partners balance the number of runs and km worked by each operator’s trains on a “no-fee” basis.

I do not have the details, but it should be clear that Toei has to maintain (and store) fewer cars than it would if its lines were “self-contained,” as is the NYCT system. Toei therefore does not need the same number of vehicle servicing and maintenance employees as it would with “self-contained” lines.

NYCT has a relatively “high-labor-cost” infrastructure that is “built in” to the system. For example:

Toei: 106 stations, 109.0 km of route = 1.0 stations per km.

NYCT: 468 stations, 398.2 km of route = 1.2 stations per km.

(Stations are counted in identical fashion: each station on each line is counted separately).

The fractional difference looks unimpressive – until you consider that if NYCT had stations spaced like Toei, it would have 81 fewer stations. That represents a significant number of staff positions.

NYCT also has a relatively “low labor productivity” traffic pattern – which reflects land-use patterns and consumer-choice issues (anyone who believes that U.S. transit consumers – “even” New Yorkers – would tolerate the type of peak-period crowding that prevails in Tokyo – should definitely have their head examined).

Consider that during 2008, an average of 38.0 million people traveled over each km of NYCT system length – while 53.8 million people traveled over each km of Toei system length.

Conservatives and libertarians, in general, do not recognize the fact that rail transit provides economy of scale. But, as the little boy once said, “This is reality, Greg.” Up to a point, the rail system with the higher passenger traffic density will provide greater cost efficiency, and its staff members will be more productive.

Now consider a measure of infrastructure use that was known, and calculated, more than 100 years ago in Europe, but Americans seem clueless about today: “annual vehicle-km per km of system length:

Toei: 109.2 million veh-km / 109.0 km = 1.0 million veh-km per km of system length.

NYCT: 577.9 million veh-km / 398.2 km = 1.5 million veh-km per km of system length.

Hmmm!

NYCT carries less than 71 percent as many passengers over each km of system length – but operates 150 percent as many veh-km per unit of system length to do so. There is probably some room for improvement, but again: anyone who believes that U.S. transit consumers – “even” New Yorkers – would tolerate the type of peak-period crowding that prevails in Tokyo – should definitely have their head examined. In other words, NYCT operates relatively more service (annual veh-km) than Toei because it “has to.”

A fundamental measure of transit labor productivity is “annual passenger-km per employee work hour.” No data for Toei, so I’ll use the data above to calculate annual pass-km per employee:

NYCT: 2,428.3 million passengers * 6.6 km (average travel distance) / 32,400 employees = 494,000.

Toei: 840.6 million passengers * 7.1 km / 3,420 employees = 1,745,000.

The apparent ratio favoring Toei is more than 3.5 to 1. However …

Remember the difference in passenger traffic density (1.4 to 1) and in annual “service density” (1 to 1.5). Those factors are essentially immutable, absent dramatic – and long-term – changes in land-use and consumer-choice factors. Also remember that, between two rail systems, the one with the smaller “peak-to-base ratio” will be the more efficient. With reference to service, “peak-to-base” ratios for Japanese rail systems are significantly lower than what NYCT provides. After all, Japanese subway systems do not have four-track trunk lines.

Oh, yes, let’s not forget the increased costs and productivity losses which occur as the result of keeping the system open, “24 / 7 / 365.”

Leroy W. Demery, Jr.

Look, I know those numbers can be quibbled with. That’s why I’m giving a more explicit measure of productivity in terms of revenue train-hours per train driver. I’m not blaming New York for not being as overcrowded as Toei; I’m blaming it for having too many employees for the amount of service it’s providing. That’s why I’m using train-hours (or train-km, which will give you the same numbers since the average speeds are the same) as the standard and not passenger-km, which is a measure of crowdedness and not just productivity. Vehicle revenue-hours is a measure that does not depend on peak hour flows or any of the other things you’re mentioning.

Station spacing is irrelevant. New York actually has fewer station staff than Toei per station – a figure I could support with links until a few months ago, when Toei overhauled the English website and eliminated the relevant factsheet. But it doesn’t matter, since the ratio of train-km per station is the same.

NYCT claims 47,000 employees on its website, though the union member commenters on Second Avenue Sagas tell me the agency has since laid off 1,000 people. The 6,400 employees figure for Toei comes from the same now-dead factsheet I’m using for the driver and station staff counts.