Guest Post: Congestion Charging

A guest post by David Lupton:

So Wayne Brown wants to introduce congestion charging.  Good on him.  There will be the usual complaints on this blog by those who would think it absurd to have a system where you queue for hours for subsidized bread but think it normal to queue for hours for subsidized roads. Congestion pricing for Auckland has been discussed for years but no-one has had the balls to introduce it.  At one time the excuse was that we didn’t have the technology.  That is no longer the case.  

But will it work?  Ah that’s a question.  The answer depends on what you define as working.  It will definitely reduce congestion, the scheme proposed will not eliminate it.  And it does depend on how you define congestion.  For a road that is operating at capacity – ie maximum sustainable flow – the traffic speed will be about 75% of the free flow “three o’clock in the morning” speed.  It seems to me that is the sensible situation to be aiming for.  Engineers refer to the situation where speeds are more than 75% of free flow as “normal” flow and only when the speed drops below that as “congested flow”.  This is how Singapore defines it, and how the value lanes in the US operate.  The congestion charge is varied to keep the traffic flowing at about 75% of the free flow speed.  

Is that what Wayne Brown is proposing? Apparently not.  From what I have read so far, it sounds like a fixed fee is proposed that is invariant with time and place.  This is the same as London and a few others.  Yes it will reduce congestion, and yes it is better than nothing at all, but because it will be too high at some times and places and too low at others (it may be right twice a day like a stopped clock) it will not fully decongest the roads and it will cause some unnecessary hardship.  

To understand what a better congestion charging system would look like, we have to realize that the way congestion affects us is not always intuitively obvious.  First of all a congested road carries less traffic than if it is uncongested. Ok you knew that, but the corollary is that an effective congestion pricing scheme means more people can travel at their preferred time rather than less.  You don’t get priced off, you get priced on.  But note that I did say an effective scheme.  To achieve this you need a toll that is high in the peak and low in the off-peak like airline charges.  A fixed toll doesn’t work because it provides no incentive for people to go earlier or later to benefit from the reduced rate. 

But we already pay for our roads I hear you cry, and yes we do  – or at least we did until the last government started syphoning the money off into pet schemes.  New Zealand had one of the best user-pays schemes in the world and it worked well, at least as far as the national highway system was concerned. Heavy trucks, cars and buses all pay an objectively assessed allocation of the total land transport budget based on the costs they impose on the system.  But it only works on average. The motorist driving in suburban Auckland pays about the same fuel taxes per kilometre as the peak hour motorway commuter and yet the cost of providing for the latter is much higher than for the former. 

What would an effective charge look like? I have already alluded to the scheme in Singapore and for value lanes in the US.  In the latest version of the Singapore scheme, the toll is varied dynamically to ensure a target speed is maintained.  If traffic starts to slow, the toll is increased, if it speeds up, the toll is reduced.  The US value lanes work the same way.  Value lanes are simple to manage with a single variable toll.  In Singapore we are taking a city-wide scheme and they have opted for a distance-based scheme.  The ideal for a city-wide scheme is actually time based.  You can show mathematically that the optimum toll is proportional to the difference between the actual time and the free flow time.  The rate per minute is set Singapore style to ensure that the traffic flows at a target speed, but unlike the distance charge, a charge per excess minute adjusts automatically to deliver the ideal toll at all times.  Even though the toll is variable, it is predictable.  We already have signs that tell us the expected time for our trip and Google can predict the best time to travel tomorrow. 

But don’t we need better public transport first? Not really.  Given adequate notice, the private companies that provide our bus services will buy more vehicles. Bus services will benefit from uncongested roads, reducing trip times and the fleet requirement.  Our huge subsidies to public transport were justified (pre climate change) in the belief that getting people out of their cars will reduce the need for more urban roads. Sorry folks but if you have congestion pricing for roads you don’t need subsidized public transport any more – the market will be more than happy to provide it unsubsidized. 

What about the poor?  You will hear that cry a lot if the plan goes ahead.  Not from the poor but mainly from the usual suspects.  As I noted earlier, we already pay for roads on average.  A fully dynamic pricing scheme that is designed to be revenue neutral shifts the burden from those who travel outside peak hours or in the suburbs to those who need to get to the office at 8:30am – generally not the poor.  Because with dynamic charging you can adjust the timing of your trip to minimize the cost, this form of pricing should be welcomed by those who like to speak for the poor.  Because it replaces rationing by queueing by rationing by price, it should also appeal to the economists amongst us.  A revenue neutral scheme would return the revenue by abolishing the Auckland petrol tax and reducing the fuel levy generally.  Similarly, a large proportion of the costs of local roads is met by property taxes.  Congestion pricing could shift some of the costs from the ratepayer to the commuter. 

What about the technology? Well you already have the required technology in your pocket.  What would be needed would be an app on your phone that would function very much like the Uber app  – this could be combined with number plate recognition and pre-and post payment schemes to address privacy and other concerns.  A potential model would be a cordon pricing scheme with a fixed charge with the option of signing up for the phone-based scheme that would be designed to deliver a lower charge.  If that were the plan, Wayne Brown’s scheme might be the first step.  Lets hope so.