> The first element, that’s the one that used to be present but isn’t anymore, is a large public sector, because you need a large enough bureaucracy that can supervise all the contractors.
If a large public sector is a prerequisite for lower-cost transit, then the cost of that large public sector needs to be included in calculating the cost of transit. I suspect that would change the analysis quite a bit.
Probably less than you'd think. Levy is delivering a punchy oneliner here; what the Transit Costs Project research indicates is that what's often missing in countries with high construction costs isn't so much bureaucrats as it is in-house engineering and technical program management resources. A lot of the design and program management work they do needs to be done by someone, but in countries with high construction costs this work mostly tends to be done either by consultants or by the contractors themselves, and that leads to misaligned incentives.
They do get into these things a little bit later in the interview, but the surprisingly simple conclusion is that when an organization is buying something, the people working out the deal need to understand what it is they're buying, or the organization gets ripped off. It sounds simple to the point of being silly when I put it like that, but it's actually true. And this should feel familiar to a lot of people in tech; how many times have you had to use or implement some very stupid system that got purchased because of buzzwords or because the people deciding what to use aren't the people using the system? (Education is particularly plagued by these problems.) What the Transit Costs Project finds is that this is a thing in public works as well: if your civil service has insufficient in-house technical expertise in its field, or if it doesn't trust its own technical expertise, that tends to make costs go up. It's far from the only contributing factor, but it's an important one.
When cost overruns for these types of projects can routinely be measured in billions of dollars, I think it's not that hard to justify hiring a few dozen civil engineers to keep things under control. The "big" civil service Levy talks about doesn't necessarily need to be big in terms of team size, but it needs to have weight.
This topic is so politicized that the transit lovers don't feel free publicly acknowledging all of the costs. For instance, one manager of the local bus company said that while new city buses cost $800k+++, they're a good deal when spread out over 12 year and all of the passengers. So I asked him to include the cost of the driver-- and the driver's benefits-- and that totally changed the equation. It can reach $1m/year to fully staff a bus over a year. Suddenly, the hardware isn't the biggest expense over the lifespan of the bus.
50k per driver per year sounds low. This page [0] suggests the low end is 60k and that's just what the driver makes, not the full loaded cost of the employee with benefits. I've seen the estimate for a loaded cost to be 25-40% more than the salary. So it ends up being more like 75k per driver.
Your estimates are fine, but you're only talking about the basic salary which in the US is only part of the equation. You've got to include benefits including health insurance which is easily $10-20k/year in the US for a family. Add in pension costs and it's a good rule of thumb to double the quoted basic salary. If you do this, the $350k of your estimate becomes $700k. Add in more for two to three layers of management, and you're easily over $1m.
Note my comment below, parents quotes expenses aren’t grounded in actual audited finances.
I’d just note that in the US at least, bus services generally operate at a loss and run many routes due to political reasons rather than efficiency/profitability.
The interstate system is almost profitable for the government. Fuel taxes raise somewhere around 30 billion to 40 billion a year, and the amount spent in interstate system maintenance every year is about that amount. If they increased the tax slightly, and stopped siphoning money off of it to fund non-road expenditures, it would be in the black.
I wasn’t intending to make a statement on whether or not public transit should be independently profitable, just providing context.
I didn’t even touch on ridiculous regulations like “buy America” (aka all buses have to be manufactured in the US) that drive up the cost of transit in a way that private drivers would never accept.
Precisely the right question to ask! Profit vs. loss has no relevance when one is providing a public service. The relevant question is cost: capital investment / creation cost & operating cost.
> Profit vs. loss has no relevance when one is providing a public service.
I beg to differ. While profit shouldn’t be the only indicator under consideration, when you are operating a state owned enterprise under a fee for service model, profit is certainly a valid metric to consider.
Cost and benefit. You could spend £10b a year providing a 12tph high speed train from Wales AK to Prudhoe Bay, but there wouldn’t be a reasonable benefit from it.
No offense to said manager, but I’m guessing they aren’t doing the accounting, so it’s not like they are the best source.
Edit: a better source would be the national transit database [0], which tracks the fully audited costs of every transit agency in the US. A very cursory analysis of this file [1], dividing total costs by VOMS (aka vehicles used to provide peak service, not even the size of the whole vehicle fleet) for directly operated bus service (mode=MB, TOS=DO). Gives a cost of roughly $250,000 per year on average across the US. This is fully loaded, including operations, vehicle and facility maintenance, general admin, etc. basically the $1 million per year is little more than a rumor.
If a large public sector is a prerequisite for lower-cost transit, then the cost of that large public sector needs to be included in calculating the cost of transit. I suspect that would change the analysis quite a bit.