Big data should be used for problem solving, not propaganda and promotion

Cue the extreme telephoto shots of freeways!

Wallow in the pity of commuters stuck in traffic because of all those other people!

Wail that congestion is getting worse and worse!

It’s that time of year again:  The big traffic congestion report is out.

Yesterday, Inrix released its annual traffic scorecard for 2017. It’s based on some of the biggest of big data out there–more than 500 terabytes of records of vehicle movements in cities around the world. It purports to rank how bad traffic is in every large city on the planet, and offers a stupendous, but entirely fictional estimate that congestion is somehow “costing” travelers in the US, UK and Germany $467 billion per year.

Time and again, we and others have debunked these congestion cost claims.  They’re based on the unrealistic assumption that there’s some way that one could build enough capacity so that every trip could travel at the same speed in peak hours as it does in off peak times.  (Thanks to induced demand, that’s literally impossible). The congestion rankings completely overlook the differences in trip lengths among metropolitan areas, awarding low congestion rankings to sprawling places where people have to drive further.

This time, we’re not going to dignify the congestion report by echoing any of its rankings. Suffice to say, the current report is essentially indistinguishable from any of those that have gone before it. We’ve been down this road (sorry!) many times before, so we’ll just remind you of some of the points we’ve made about the flaws and absurdity of this and previous congestion reports.

  • We’ve tried humor and parody: witness our Cappuccino Congestion Index, which applies the “cost of delay” model to coffee shops, to show how ridiculous it is.
  • We’ve tried deeply academic copiously footnoted white papers:  Read, for instance, our 62-page report Measuring Urban Transportation Performance.
  • We’ve shown how longer commutes, not levels of traffic congestion, are the primary determinants of whether people are satisfied with their region’s transportation system. Going faster doesn’t make you happier, it just makes you sprawl more.
  • We’ve calculated how the sprawling development patterns (that are correlated with “good” congestion scores) impose a multi-billion dollar tax on travelers in metro areas around the country.

It’s a tragic waste of potentially useful data.  Once upon a time, Inrix, the company that produced its report, made an honest effort to make its data transparent and regularly available, so that one could track progress. The Inrix data showed that nearly all cities made big progress in reducing traffic congestion between 2010 and 2012–when gasoline prices were considerable higher than they are today. That data has been deep-sixed, along with the insight that, if we making driving somewhat more expensive, we’ll experience less congestion.

Of course, alarmist congestion reports have been with us for decades. Inrix is carrying on in the tradition of the Texas Transportation Institute, which for years produced a similar report. At least the TTI showed a bit more of its work and revealed more about its methods, and made a rough stab at trying to say how we might tackle the problem. The Inrix report is all about generating heat, not shedding light, on the nature and solutions to urban transportation problems.

The good news is that many in the news media are wising up to the fable that’s being spun here.  The Oregonian‘s Eliot Njus dutifully reported the Inrix headline claims, but then added the caveat that:

There are plenty of criticisms of these kinds of reports, chief among them is that they set a baseline of “free-flow” speeds. That can mean driving the posted speed limit is considered an effect congestion. Congestion studies also tend to understate the effects of shorter commutes in small and compact metro areas such as Portland’s.

We have more data about urban transportation system performance than we’ve ever had. But because it’s chiefly being used for propaganda (traffic is horrible!) and promotion (Inrix!), we actually know less than we used to.  A really useful report would focus in on careful collection of time series data within a market, and would track travel distances, as well as speeds, recognizing that shorter trips are a boon to both travelers and cities. There is a small glimmer of hope here: in the blog post accompanying the report, Inrix economist Graham Cookson acknowledges that the congestion problem is about the imbalance between supply and demand, and that pricing roads can play a key role in bringing these in to balance:

But the fundamental reason is an imbalance between the demand and supply for roads. Managing demand for road space is critical. That includes smoothing demand through flexible working, avoiding peak hour trips through remote working, ride sharing and encouraging the efficient use of our roads through wider adoption of road user pricing. (Emphasis added).

Rather than being buried in a blog post, that should be the headline of any congestion report: The reason you’re sitting in traffic is because the price that’s being charged to road users doesn’t reflect either the cost of building and maintaining the roadway, nor the effects of your decision to travel at peak hours. If we want to eliminate or reduce congestion and its attendant costs, the only effective solution is to get the prices right.