Journalists: Stop repeating phony congestion cost estimates. They’re just weak arguments disguised with big numbers.
This month The Economist has an excellent special report exploring the prospects for autonomous vehicles. They seem to be coming faster than many people anticipated, and they pose some big challenges and opportunities for cities. This otherwise very useful contribution to the conversation is marred, unfortunately, by The Economist also posting as fact the congestion cost estimates produced by traffic monitoring firm Inrix.
As regular readers of City Observatory know, we’ve pointed out serious problems with the Inrix congestion cost estimates.
It’s painful to watch an otherwise intelligent journal like The Economist uncritically reproduce the demonstrably fictitious congestion cost estimates. Carrying on the in the tradition of the Texas Transportation Institute, Inrix now annually produces some VERY SCARY NUMBERS about how much congestion supposedly costs travelers in cities around the world.
Congestion cost estimates are the horror fiction sub-genre of what we’ve called “Hagiometry.” Hagiography is flattery in prose form; hagiometry is flattery with numbers; and congestion cost estimates are designed solely to use big numbers to scare people into believing that a problem is somehow worse than it really is.
The Inrix figures are an argument—and a remarkably flimsy one at that—masquerading as economic statistics. The implicit argument is that there is some state of the world in which people could travel just as fast at peak hours than they do when there’s no traffic on the roads. The cost estimates are constructed based on adding up how many more minutes it takes to travel when there’s traffic compared to when there are no traffic delays, and then multiplying that by some value of time. The math may be right, but the assumption imbedded in that argument, that there’s some way to build enough lanes to accomodate all that traffic, is just wrong.
A clever quantitative biologist using the Inrix methodology could easily compute the number of excess derriere contusions the world’s frogs suffer every day because of their unfortunate decision to choose jumping rather than flying as a means of travel. And that would provide exactly as much insight into transportation policy as does the Inrix report.
There’s no way to build enough roads to eliminate congestion, at any price
The essence of the Inrix calculation is this: if there were no other cars on the road when you wanted to drive to work each morning, and also when you drove home each evening, here’s how much time you’d save, and what it would be worth to you? But would it be fiscally or even physically possible to build enough roadway space to give everyone the same level of service at 5pm every day as you get when you’re driving at say 2am? Of course not.
One could just as easily add up and value the total amount of time people in the world spend traveling between any two sets of points, and count that as the cost of not having “Star Trek” style matter-transporters. But, you would argue, that can’t realistically be regarded as a “cost” because such transporters don’t exist. And that’s precisely the point, there’s no way to build a road system that allows peak hour travelers to travel at the same speeds
And even if you did build a stupendous amount of additional roadway capacity (and could repeal the fundamental law of road congestion, in which the increment of additional capacity generates more peak hour travel thus replicating congestion), there’s absolutely no doubt that the cost of constructing that roadway would greatly outstrip the supposed “benefits” of eliminating congestion. We don’t have to speculate about this: when road users are asked to pay even a fraction of the of cost of building additional road capacity, in the form of tolls, they readily indicate by voting with their feet (or wheels), that they attach very little value to travel time savings. Notice, for example the case of the I-65 Ohio river crossing in Louisville. The state highway department doubled the size of the bridge from 6-lanes to 12-lanes, and then started charging a toll. Almost immediately traffic on the bridges fell from 120,000 vehicles daily, to about 70,000.
Make it stop!
Here’s the point: Anyone who claims that congestion has a cost–meaning a real, net social cost–has to propose some transportation system that would (a) eliminate all of the delays that they’ve counted, and (b) do it for less cost than the value of supposed value of time lost to congestion. If they can’t do that they haven’t shown that congestion has any real costs at all.
Credulous reporters need to start looking past the dizzying array of numbers and ask some hard questions about the assumptions behind them. Repeating fictitious claims about the supposed cost of congestion isn’t helping their readers understand, much less solve, the world’s urban transportation problems.
Cloaking a weak argument in big—but phony—numbers
Journalists: Stop repeating phony congestion cost estimates. They’re just weak arguments disguised with big numbers.
This month The Economist has an excellent special report exploring the prospects for autonomous vehicles. They seem to be coming faster than many people anticipated, and they pose some big challenges and opportunities for cities. This otherwise very useful contribution to the conversation is marred, unfortunately, by The Economist also posting as fact the congestion cost estimates produced by traffic monitoring firm Inrix.
As regular readers of City Observatory know, we’ve pointed out serious problems with the Inrix congestion cost estimates.
It’s painful to watch an otherwise intelligent journal like The Economist uncritically reproduce the demonstrably fictitious congestion cost estimates. Carrying on the in the tradition of the Texas Transportation Institute, Inrix now annually produces some VERY SCARY NUMBERS about how much congestion supposedly costs travelers in cities around the world.
Congestion cost estimates are the horror fiction sub-genre of what we’ve called “Hagiometry.” Hagiography is flattery in prose form; hagiometry is flattery with numbers; and congestion cost estimates are designed solely to use big numbers to scare people into believing that a problem is somehow worse than it really is.
The Inrix figures are an argument—and a remarkably flimsy one at that—masquerading as economic statistics. The implicit argument is that there is some state of the world in which people could travel just as fast at peak hours than they do when there’s no traffic on the roads. The cost estimates are constructed based on adding up how many more minutes it takes to travel when there’s traffic compared to when there are no traffic delays, and then multiplying that by some value of time. The math may be right, but the assumption imbedded in that argument, that there’s some way to build enough lanes to accomodate all that traffic, is just wrong.
In the words of Raising Arizona’s Nathan Arizona: Yeah, if a frog had wings, it wouldn’t bump its ass a- hoppin’.
A clever quantitative biologist using the Inrix methodology could easily compute the number of excess derriere contusions the world’s frogs suffer every day because of their unfortunate decision to choose jumping rather than flying as a means of travel. And that would provide exactly as much insight into transportation policy as does the Inrix report.
There’s no way to build enough roads to eliminate congestion, at any price
The essence of the Inrix calculation is this: if there were no other cars on the road when you wanted to drive to work each morning, and also when you drove home each evening, here’s how much time you’d save, and what it would be worth to you? But would it be fiscally or even physically possible to build enough roadway space to give everyone the same level of service at 5pm every day as you get when you’re driving at say 2am? Of course not.
One could just as easily add up and value the total amount of time people in the world spend traveling between any two sets of points, and count that as the cost of not having “Star Trek” style matter-transporters. But, you would argue, that can’t realistically be regarded as a “cost” because such transporters don’t exist. And that’s precisely the point, there’s no way to build a road system that allows peak hour travelers to travel at the same speeds
And even if you did build a stupendous amount of additional roadway capacity (and could repeal the fundamental law of road congestion, in which the increment of additional capacity generates more peak hour travel thus replicating congestion), there’s absolutely no doubt that the cost of constructing that roadway would greatly outstrip the supposed “benefits” of eliminating congestion. We don’t have to speculate about this: when road users are asked to pay even a fraction of the of cost of building additional road capacity, in the form of tolls, they readily indicate by voting with their feet (or wheels), that they attach very little value to travel time savings. Notice, for example the case of the I-65 Ohio river crossing in Louisville. The state highway department doubled the size of the bridge from 6-lanes to 12-lanes, and then started charging a toll. Almost immediately traffic on the bridges fell from 120,000 vehicles daily, to about 70,000.
Make it stop!
Here’s the point: Anyone who claims that congestion has a cost–meaning a real, net social cost–has to propose some transportation system that would (a) eliminate all of the delays that they’ve counted, and (b) do it for less cost than the value of supposed value of time lost to congestion. If they can’t do that they haven’t shown that congestion has any real costs at all.
Credulous reporters need to start looking past the dizzying array of numbers and ask some hard questions about the assumptions behind them. Repeating fictitious claims about the supposed cost of congestion isn’t helping their readers understand, much less solve, the world’s urban transportation problems.
Related Commentary