Sooner or later: Cost disease and Canadian transit.

Back in 2015, in the wake of Metro Vancouver’s failed referendum campaign for a 0.5% sales tax to fund an ambitious list of regional transit upgrades, I argued that the big-ticket items on the list should be scaled back or postponed while we focussed our limited dollars on improving bus service. Subways and LRTs are great, I said, but a couple big rail projects will eat up all the money we could instead use to make the whole network faster, less crowded, and more enticing to commuters.

I still think my argument makes sense. I occasionally take the bus during rush hour, and I seethe at being stuck behind lines of idling cars when I can see how a simple bypass lane, costing a paltry few million bucks, would save thousands of straphangers five or ten minutes out of their commute each way, every day. Many roads wouldn’t even have to be widened – simply sacrificing a few on-street parking spots would do the trick. There must be dozens of such chokepoints around the region, and they could all be unchoked for a fraction of the cost of putting a subway down Broadway. Though I’d like us to build the subway too.

But a post last month on Slate Star Codex – Scott Alexander’s estimable blog, to which I lately find myself linking with unseemly frequency (see here and here) – makes me wonder if my sensible, fiscally-prudent argument was in fact completely wrong.

Alexander discusses something called Baumol’s cost disease, a phenomenon in economics where increasing efficiency in one industry counterintuitively leads to increasing costs in an entirely unrelated industry.

Suppose new manufacturing methods save an auto plant part of the cost of building a car. Profits rise, allowing the company to boost its workers’ wages by a couple bucks an hour. Meanwhile the meat processing facility across town hasn’t seen any improvement in productivity, but if they don’t offer an equivalent wage hike they’ll lose their best workers to the auto plant. They pass the higher costs along to their customers, and suddenly the price of meat goes up because the cost of manufacturing cars has gone down.

That’s Baumol’s version of cost disease, anyway. Alexander wonders whether it’s a sufficient explanation for the perpetually increasing costs in four different sectors of the U.S. economy – education, health care, housing, and public transportation infrastructure. Even after adjusting for inflation, costs in these four sectors have gone up in my lifetime by factors of two, five, even ten, without commensurate improvements in outcomes. Life expectancy is flat. University grads are as semiliterate as ever. Apartments aren’t appreciably nicer. And subway tunnels are pretty much the same as the ones our forefathers dug for a fraction of the cost.

Alexander is American, and in his brief section addressing ever-pricier subway construction he restricts himself to American data. In fact one of the questions he asks is why the U.S. seems to be more susceptible to cost disease than other countries. So my first question was – does this disease afflict Canadian public transportation infrastructure as well?

Let’s look at the costs – adjusted for inflation – for a half-century’s worth of rapid transit projects in the two big cities I know reasonably well, Vancouver and Toronto:

(Click on images for data and sources.)

Some caveats and observations:

• With so few data points to work with, the trendlines are susceptible to being skewed by one or two pricey outliers, like Toronto’s bonkers Line 1 extension to Vaughan.
• Reported final costs are questionable, since governments tend to find ways to obscure overruns. Vancouver’s 2016 Evergreen extension, for instance, is known to have blown past its budget, but we’ve been assured that the unanticipated costs will be eaten by the contractor. The true cost, therefore, is higher than the figure shown.
• The graphs are to the same scale, but the cities’ rapid transit systems shouldn’t be compared directly since they use totally different technologies. Vancouver’s light, high-frequency, mostly-elevated SkyTrain permits smaller stations and (for the 20% or so of the system that’s underground) narrower, cheaper-to-build tunnels than Toronto’s heavy-rail subway. (I’ve left Toronto’s SkyTrain-like Scarborough RT and under-construction Eglinton light-rail project out of the analysis for this reason.)
• Even within each city, these aren’t apples-to-apples. In the Toronto graph there are visible discontinuities between the early cut-and-cover subways in the city core, with stops every 500-600 metres and relatively low per-station costs; the 1970s extensions into suburbia, often at surface level and with fewer, more widely-spaced stops; and more recent bored tunnels where the costs shoot into the stratosphere.

In any case, both graphs show a discernible upward tick since the 1990s or so, suggesting that cost disease may indeed have spread to Canada. But if so, what are the causes?

The libertarianish Megan McArdle waves away Alexander’s data on the rising cost of subways as merely “union featherbedding combined with increasingly dysfunctional procurement and regulatory processes”. Maybe those are worsening the problem – I don’t know enough to comment – but off the top of my head I can think of four other possible contributing factors:

1. The cost of land acquisition goes up at a rate faster than inflation (because they keep making people but they aren’t making more land).

2. The ground beneath and alongside city streets is ever more crowded with pipes, cables, parking structures, and so on, which must either be relocated or awkwardly worked around.

3. An increased emphasis on worker and bystander safety slows and complicates construction. (Some of this probably falls under the definition of “union featherbedding” as mentioned by McArdle.)

4. Projects now include the expenses of mitigating environmental damage, preserving historic neighbourhoods, averting noise pollution, accommodating the handicapped – all that touchy-feely stuff previous generations didn’t give a rip about.

Doubtless there are other causes I haven’t thought of, but I’ll stop at those four because they pair off neatly into two groups I’d like to examine a little more closely. When you think about it, all four are side effects of growing wealth:

• Causes 1 and 2 – the rising cost of land and the build-up of clutter along possible transit routes – accelerate as a city becomes more populous and its taxpayers demand more and better services.
• Causes 3 and 4 – worker safety and the mitigation of environmental and social externalities – might be thought of as perks, which previous generations were willing to forego in their pursuit of progress but which we in our prosperity don’t mind splashing out on.

I think the “perk factor” actually explains much of the cost disease in the sectors Alexander identifies. As we’ve grown wealthier we’re willing to spend more on things that are orthogonal to the actual missions of health care, education, housing, and public transportation – things like prioritizing the physical and mental well-being of our workforces, or ensuring that their gender and ethnic compositions are representative of the wider population. These perks require added layers of administration that do nothing to improve the outcomes we’re attempting to measure. Those layers aren’t failing – they’re doing what they’re meant to do – but those things aren’t captured in graphs like the ones in this post.

As a taxpayer I suspect we could afford to do without much of this extra padding. But I don’t want construction workers risking their necks, or rivers recklessly diverted, or noisy trains rattling people’s cupboards, just to save a few bucks. I know next to nothing about health care or university administration or housing construction, but I suppose the people who are familiar with those matters have equally strong objections to cutting what may strike me as frivolous perks.

In any case, sticking to transit infrastructure, there’s no reason to suppose we’re likely to care less about safety, or the environment, or architectural heritage in the future. In fact those concerns will almost certainly grow, making construction ever less affordable.

To return to causes 1 and 2 – land costs and infrastructure clutter – it should be possible to mitigate cost disease through better planning – say, through more farsighted property acquisition, and coordinating with other agencies to ensure that future transit corridors aren’t obstructed. But I assume we’re already trying to do those things, and my suggestion of “Okay, well, just do them better” is not too helpful.

Sooner, or later?

One way to avert cost disease might be by preemptive surgery – building rapid transit today, at today’s comparatively reasonable prices, in anticipation of tomorrow’s needs. Metro Vancouver’s overall outline is pretty well established by geographic barriers like mountains and rivers, and more recently by the imposition of an urban containment boundary meant to preserve nearby farmland.

Source: Metro Vancouver. (Click for original.)

Therefore we can assume that the Vancouver of the future will be much the same shape as the Vancouver of today, only a lot denser. We should be able to predict with fair accuracy where future demand for transit will lie, and build in anticipation of that demand.

However, of the four causes of cost disease mentioned above, preemptive surgery really only targets the first one – rising land costs. It sidesteps the cost of infrastructure clutter only by transferring that cost to future generations who will have to spend more to build around the clutter we create today. And while it might seem thrifty to thwart the next generation’s opportunity to waste money on what we consider silly perks – by using construction methods that they’ll see as barbarously unsafe, maybe, or by bulldozing some architectural monstrosity before it’s declared a heritage monument – who are we to say what the future’s priorities should be?

I’m grateful for much of the infrastructure earlier generations of headstrong builders bequeathed me, but I wish they’d been more cautious about what they smashed in the process – like the whole blocks of Vancouver’s Strathcona neighbourhood destroyed in the sixties to make room for the Georgia viaducts, which planners are now preparing to remove. Building preemptively means we risk building unnecessarily, as the future evolves new habits of getting around that we can’t anticipate.

Still, being made aware of cost disease has tipped me in favour of building rapid transit now, while it’s still barely affordable, rather than putting it off as demand grows and grows. Spending sooner rather than later may actually be the sensible, fiscally-prudent thing to do.

M.