Opportunity cost of a parking garage

Auburn (Google Maps)

Dan's report on the ever-increasing cost of Sounder parking garages demands a little rough math. The per-space arithmetic is damning enough. But the opportunity costs are the practical reason to hope for something better.

The most expensive of these is Auburn Station, which is now to cost $120m in year of expenditure (YOE) dollars for 675 spaces (555 net new). Generous assumptions about inflation [1] make this $107m in current dollars. If this money were instead spread over 30 years for bus routes to Sounder, it would allow 10 additional buses to serve the Auburn area during Sounder operating times [2]. A similar, though slightly less dramatic, story plays out at the other stations. The potential of these buses to improve several different outcomes is an exercise for the reader.

In my earlier days, this would be an occasion to slam the stupidity of the people in the charge. But in my thirteenth year of doing this, I know that things like this happen because of incentives:

• These projects have increased considerably in cost, making the math much worse than at the time of decision.
• Without any additional outlay, parking is there for Sunday Seahawks trains, Puyallup Fair trains, and whatever else brings people into central Auburn.
• Because the law insists on voter approval, we often select projects not based on cost per rider, or least-cost alternative, but what will win over the marginal voter. It is very plausible that the marginal voter likes the idea of driving to Sounder, but not taking a bus there. The election literature, both for and against, hesitates to point out that the parking will be full.
• A quick glance at the road grid (see above) suggests that many marginal voters are right to be skeptical that a bus will serve them effectively.
• Because Metro has a wonk-approved, metric-driven formula for service allocation, it is reasonable to think that putting more money into the bus system results not in carpeting Auburn with buses, but distribution to more places in need across South King County, or the County as a whole. Little wonder that leaders in Auburn prefer the garage: bus service is centrifugal, while capital investments are concentrated. And the state has granted cities practical veto power over projects through permitting.

The 2008 election is over, so vote-maximizing calculations are irrelevant. There is still a duty to deliver what voters approved. But riders are fungible; fewer parking spaces may mean certain people don't use Sounder, but the alternative bus service can bring many more others. The ridership benefit of a parking space is about 2 times the carpool minimum of that space, but the cost-equivalent bus can be much more. I would hope that in an environment of uncertain revenues and spiraling costs, Sound Transit would find cheaper ways to get the same result.

[1] Assume all the money is spent when the garage opens in 2024, and an annual inflation rate of 3%.

[2] At Metro's current rate of about $150 per service hour, $107m would fund roughly 700,000 hours of service, or about 15% of all Metro service in 2018. But the garage lasts for a long time, if not quite forever, while there is little to show for operating expenses once spent.

It would be far more reasonable to spread those hours of a longer period. It's not clear what the useful engineering life of the garage will be. Furthermore, without succumbing to tech hype, widespread autonomous cars would make on-site garages largely obsolete. Perhaps 30 years is a good guess for robot cars or a big garage renovation, leaving us about 23,000 service hours per year. For convenience, we'll assume the increasing cost of service mirrors the interest savings of deferring most of the spending.

Sounder serves Auburn Station over a 7-hour span each weekday. We'll assume buses start an hour before the first train and end an hour after the last one, for a total of 2,250 hours per year. This implies that forgoing the garage would allow 10 additional buses to serve Auburn Station simultaneously during the Sounder service window.