A sample text widget

Etiam pulvinar consectetur dolor sed malesuada. Ut convallis euismod dolor nec pretium. Nunc ut tristique massa.

Nam sodales mi vitae dolor ullamcorper et vulputate enim accumsan. Morbi orci magna, tincidunt vitae molestie nec, molestie at mi. Nulla nulla lorem, suscipit in posuere in, interdum non magna.

Driverless car technology and congestion

A traffic engineer told me the other day that driverless car technology would largely replace the need for expanding road systems in urban centers.   The general idea is that if car travel can be automated that there is plenty of road space for cars to make synchronized journeys on so congestion will just disappear.  There are, however, now many studies that question the claim that driverless cars will reduce congestion – the latest is in today’s AFR.   Let’s back up a bit and sort out the basic logic here.

Roads should be expanded, if congestion and other costs are appropriately priced,  whenever roads display appropriate profitability – the exact definition of “appropriate” here depends on the “returns-to-scale of the road technology. Simply expanding roads, when they are unpriced, is unlikely to be sensible because the effort reflects the attempt to meet an excess demand for travel that is being sold at a zero financial price.  Intuitively what happens here is that there are latent, unrealized demands for travel – people leave their cars at home simply because car travel is too time expensive and inconvenient/uncomfortable. When a new road is constructed these latent demands are released and congestion tends to return to its original level.  There is a small welfare gain – new commuters get to travel from the pool of latent demands – but the level of congestion itself is unaddressed.  Anthony Downs (“Stuck in Traffic”, Brookings, 1992) described this release of latent demands as “triple convergence”.  The only way of avoiding this offset and dealing effectively with congestion is to appropriately price car travel to reflect as marginal congestion costs and, if the road makes an appropriate profit, then just as with a private firm, its scale should be expanded.

With driverless cars two things happen to a car journey.  First it becomes more pleasant – the opportunity cost of the time involved goes down because the commuter does not have to drive and can read, work or indeed, watch a Lady Gaga video if they like.  This stimulates the demand for journeys since the disutility cost of a journey declines.  Second, the immediate time taken for travel will itself decline and more cars will be released onto the road on this account.   This stimulation of the demand for travel will be quite strong because the time savings are also accompanied by an improvement in the convenience of the journey.  Indeed, accounting for this latter effect equilibrium levels of congestion might be expected to increase beyond initial levels because commuters will ensure a bit more congestion given that their journeys are more pleasant.

As with an unpriced road expansion, there is a welfare gain – more people will get to travel and their journeys will be more convenient but the effect on congestion will be to worsen it.  The net welfare effect is plausibly positive (and therefore welfare-increasing) but from the viewpoint of an individual driver that will not appear to be the case.  His/her journeys may take longer but he/she can spend more time working or recreating in the car on these longer journeys.

The upshot: Even with driverless cars the best solution to urban congestion problems is to price it at marginal cost.  Then desired road expansions and the development of driverless car technologies will avoid triple convergence issues and make commuting more convenient and pleasant in our large cities.


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.