I have recently got hold of (for the second time) Paul Mees’s first book: A Very Public Solution. While it has many similarities with his most recent book, Transport for Suburbia, there are early parts in A Very Public Solution which give a good overview of the “transport situation”, setting the scene for why we need better public transport and – the main focus of the book – what better public transport might actually entail. One thing that Mees discusses, and which fits well into quite a few thoughts I’ve had recently about the way we assess transport projects, is our obsession with doing everything we can to fix congestion – and how that might not actually be the best approach to smart transport policy. This paragraph sets the scene quite nicely:
…congestion remains the focus of most professional transport planners. This focus is reflected in the cost-benefit analyses used to justify new transport infrastructure, which assess benefits almost exclusively in terms of travel time reductions. Even when transport planners attempt to consider other factors, their traditional training often constrains the results. The Environmental Protection Authority in Victoria commissioned a study of ‘transport externalities’ in 1994. The consultant road engineers who conducted the study estimated the annual cost of congestion in Melbourne as $2031 million, compared with $86 million for road noise and $45 million for cancers caused by vehicle emissions.
While cost benefit analyses here in New Zealand are slowly being tweaked to reflect more wide-ranging issues than just travel time savings, this obsession with making it quicker (or at least attempting to) for people to travel from A to B still provides the bulk of the justification for most transport projects. If your project doesn’t generate significant travel time savings, then it’s unlikely to fare well in its cost-effectiveness assessment. Indeed, Mees points out that congestion may actually have benefits for a number of people:
It may actually improve the lot of some residents, since slow-moving traffic makes less noise and is less intimidating for pedestrians and cyclists. And although there may be more accidents on congested roads, they will be less severe, owing to lower speeds.
My fundamental interest in transport comes from its interaction with our urban form. Generally the greater priority we give to shifting vehicles through a place, the more we degrade that place. Obviously we can’t have every street in the city being pedestrianised, or a shared space, so we need to find a balance between the “through” and the “in” – but the fundamental flaw of our current way of measuring transport projects is that we only look at one side of this equation: the “through”.
If we think of a project like turning Nelson and Hobson streets back into two-way streets, we would be tilting the balance away from shifting traffic through these streets and focusing more on making the western part of Auckland’s city centre a nicer place to live, work and generally be in. A careful analysis of whether this project was “worth it” would look at the general ‘urban benefits’ of two-waying the streets, analyse the impact on traffic (whether positive or negative, my assumption is that it would probably be slightly negative), then see whether the net gain is worth the cost of undertaking the works necessary to implement such a change. The problem is that our current cost-benefit analysis process – because it focuses solely on reducing congestion – would only look at one side of the issue (the impact on traffic flows) while ignoring everything else. Which means it’s fairly unlikely we’ll see this project proceed until this issue has been resolved.
In short, focusing too much on reducing congestion can kill our city through the never-ending obsession with making cars move faster and faster – inevitably degrading the urban environment to a greater and greater extent.
The other issue with focusing too much on congestion is that actually eliminating this “unspeakable evil” is nigh on impossible – and certainly impractical. Mees discusses this further:
The Victorian Transport Externalities Study defines congestion as: ‘the difference in resource costs between the road network operating under current traffic conditions, and the road network operating under ideal conditions where delays have been eliminated and traffic is able to proceed at the maximum safe speed.’
Martin Mogridge’s book Travel in Towns describes these ‘ideal conditions’ as ‘patently absurd conditions’, since in large cities they apply only in the dead of night. Delays due to other traffic are unavoidable in an urban area: congestion-free motoring is possible at all times on the Nullarbor Plain or Death Valley, but only because almost nobody lives there. Mogridge concludes that ‘the cost of congestion’ is therefore an invalid concept in an urban area’.
It is often quoted that congestion in Auckland costs our economy $1 billion a year. The NZ Council for Infrastructure Development highlighted the numbers behind such a figure in one of their recent presentations:
There are plenty of flaws in the way this has been calculated but the most obvious is the issue of “what are we comparing it with?” A roading system completely free of any delays?
Mees picks up on this point:
…congestion is usually at its worst only on part of the road system and for only part of each day. People accept congestion at theatres, holiday resorts and supermarket checkouts at times of peak demand, because they know it is wasteful to build capacity that sits under-utilised most of the time. A simple example is provided by a sandwich bar, which is quiet most of the day but crowded at lunch time. Lunch time patrons queue to be served: they do not expect simply to arrive at the busiest time of the day and be attended to without delay. Why should roads be different?
Another way to manage peak demand is obviously through some sort of pricing mechanism that relates to the time of day, but the main point is that it’s utterly stupid to try to eliminate congestion – because in order to do so you would have to build a level of capacity that is completely over the top for all other times of the day. That’s just a waste of money and a hugely unnecessary negative effect on the quality of our cities: as inevitably providing more vehicle capacity will adversely effect the quality of our urban environment through which that road passes.
Obviously we can’t ignore congestion in our transport policies. We need to be able to get around our cities relatively easily in order for them to function well. But I like Mees’s conclusion on the issue:
It may well be that the best approach to congestion is to relax. There will always be congestion in large cities, but at least in cities with well-developed alternative transport modes people will be able to choose whether or not to endure it…
…The challenge for planners is not to eliminate congestion but to plan for an optimal level of congestion, bearing in mind environmental, economic and social goals.
Next time you’re stuck in a traffic jam on the motorway, have a think about whether you think Auckland would really be better off with that motorway (and every other motorway in the region to avoid bottlenecks) being twice its current width, as well as how much that would cost to construct. I think you’d take the few minutes of delay.
Processing...
A minor thought:
Another reason why trying to ‘fix’ congestion often doesn’t work, is something called peak spreading. If I know that my journey is going to be heavily delayed if I leave home at 8am, it will be quite sensible to see if I can arrange my day to leave home at 9am, or sometimes 7am. If the 8am congestion is fixed, then I’ll simply change my travel pattern back to an 8am departure and a lot of the benefits of the ‘fix’ will be lost.
Peak spreading is also seen in other transport modes, like railways (= avoiding the trains which are known to be too busy). People will change the time at which they travel somewhat more easily than they will change the mode on which they travel.
I agree Ross. I was going to talk about triple convergence (it’s the next few pages of Mees’s book) but I thought it was a bit too much for one post.
Mees is an economic illiterate, and he dismisses congestion because his pet solutions (obsessed with public transport) don’t deliver congestion reduction. He is inept on highways management and pricing, so tends to dismiss solutions that don’t fit within his remit. A a result, he is not thought of highly in many government circles. To equate a network to a restaurant queue is a standard of analysis that is simply vapid. Restaurants operate in highly fluid markets, where people have multiple choices, where price can easily be used to manage demand, and extra capacity has little external impact on others. If you took his logic, then if you open a small very successful restaurant, you’d never expand because “queuing is ok”. Indeed, you’d say if you expanded, you’d “induce demand” making it pointless in the first case. He would be better arguing about other networks, such as with intercity transport, or airports, or telecommunications, but he doesn’t know enough about those to make a credible argument. Funny how he doesn’t apply his logic to urban public transport where “capacity that sits under-utilised most of the time” is exactly what he supports building.
I’ve heard acolytes of his quote that congestion is a “good thing” because it makes road transport unattractive. The banal notion that vehicles sitting wasting fuel, creating emissions and preventing an expensive piece of capital from functioning properly (and especially imposing enormous costs on users with no alternative, like freight and emergency services) is a good thing should be consigned to the nutter bin.
However, I’ve yet to see a credible transport economist argue that congestion should be eliminated. There is an economically efficient level of congestion and the best example of seeing that working in practice is Singapore, where the Electronic Road Pricing system prices roads to ensure traffic flows at a minimum speed. It isn’t the speed limit or full free flow, but a throughput speed that is considered optimal. When traffic gets slower, the price rises, when it gets a lot faster, beyond another threshold, the price lowers.
Congestion shouldn’t be eliminated, but reasonable minimal standards of service should be expected on roads. The way to ensure is a mixture of expenditure on capacity, network management and pricing. The same should apply for public transport.
What do you say about the Downs-Thomson paradox then Liberty?
It is valid only so far as the usage of the highway in question is by users predominantly making a similar O-D trip as is offered by the parallel segregated public transport link. The obvious NZ example is Hutt-Wellington in the AM peak, (but even that is flawed because as much as 25% of the traffic on the Petone-Ngauranga link is actually heading north on SH1). I doubt whether there are too many examples in Auckland given how decentralised most trip patterns are.
I actually used it in arguing against the addition of a fourth lane on the Wellington Urban Motorway between Ngaurange and Aotea Quay unless it was tolled, and could be funded from the toll.
There are other factors of course, one of them is the price of motoring, which at the time that work was done would have been far more stable than today. High petrol prices make a difference. Parking pricing/availability is also very relevant.
So it has value when the road and public transport are direct competitors, but it is rather simplistic to apply across the board. I’d argue you need to think on a corridor by corridor basis based on OD patterns.
Trips from the North Shore to central Auckland seem a reasonable similar example. We’re now seeing 40% of trips over the harbour bridge at peak time being by bus. One would assume that as very few buses from the shore go anywhere but the city centre, PT must have an incredibly high modeshare for North Shore to City Centre trips.
So one could assume that adding road capacity on the North Shore to city centre corridor would result in everyone ending up worse off, through the Down-Thomson paradox?
Yes it seems reasonable, if the vast majority of trips over the harbour bridge terminate in the CBD. I don’t know if that is true, but I think you would be right about North Shore-City trips.
Adding additional, uncharged, capacity would only result in everyone being worse off if it induced mode shift from the bus users, which if speed of commute is paramount (not the price of parking or fuel), would be true, if the shift was more than the added capacity. I suggest the effect of added capacity on that route is probably to induce increases in property prices as people relocate to the Shore, happier about the commute from a “generally considered to be desirable” part of Auckland.
However, if the capacity is tolled, with peak tolling, then it wouldn’t be the case, as the congestion would not eventuate. If the extra capacity is fully toll/road user funded, doesn’t result in congestion, then the paradox does not occur.