Wednesday, July 12, 2006

Heartburn over daft congestion economics

Sinclair Davidson in yesterday's Age argues that while 'Many of the 'solutions' to congestion assume it is caused by excess demand for the road. But much congestion is caused by reduced road supply'.

I had to read this statement three times to try to make sense of it. Excess demand here is the demand for road travel less the supply of road travel opportunities. If this is excessive it could be due to either excessive demand or deficient supply. Indeed it doesn't make sense to pose things in these terms - it is analogous to the economist Alfred Marshall's difficulty of trying to work out which blade of the pair of scissors cuts a piece of paper.

But it is quite clear from what follows what Sinclair intends. The congestion difficulties we face in Melbourne have little to do with the large numbers of cars using the roads without paying a toll reflecting congestion costs imposed. They instead stem from the Bracks Governments' speed limits, tram stops and having 'special events' that 'forever' disrupt traffic. Attempts to internalise congestion externalities by levying tolls for travel during congested periods are just a 'revenue grab' by the Bracks Government. Those seeking the imposition of such tolls 'want an immobile population' and 'a dreary, or even abandoned CBD'.

This is unsound economics and Sinclair should know better. People who travel on roads impose congestion costs on others. As has been known for at least half a century, attempts to internalise all such costs makes a community better-off in the sense that the value of the revenues gains exceeds costs to both the 'tolled-off' (those ending road use) and the 'tolled-on' (those who continue but who must pay the toll). Speed limits will have little impact on traffic flows under congested conditions because the traffic itself is moving so slow. And the Brack's Government far from favouring congestion tolls as a source of revenue (I wish they would) initially sought to exclude the consideration of such tolls from the terms of the recent VCEC inquiry.

This kind of daft economics causes heartburn to those of us seeking to change government attitudes to road pricing to improve the quality of life in our cities. We are not anti-travel at all - we just don't want to put up with congestion that can be eliminated to the net benefit of all.

11 comments:

Sinclair Davidson said...

We are in some agreement. This is unsound neo-classical economics. Where we disagree is on the cause of the 'congestion'. You say drivers impose costs on each other. (Readers should know that Harry's view is totally accepted mainstream economics). I say, government imposes costs on motorists through 'work zones, special events, poor traffic synchronisation, and bottlenecks'. When government has reduced the marginal speed limit by 10 - 20 km/h it is not surprising that the average speed limit has fallen by 3 km/h.
If it is true that drivers do impose costs on each other (and the broader community) proponents of tolls etc. need to demonstrate that these costs have not already been internalised through existing road charges.
Harry quite correctly indicates there is a theoretical problem, and it may be solves by levying a toll. But I would like to see evidence of a practical problem and how in practice it could be solved.

Damien Eldridge said...

Existing road charges on non-toll roads do not vary by route. This means that they do very little to cure congestion. All they do is reduce the number of vehicles on the road (rego charges, GST on car sales, etc) and the number of vehicle kilometres travelled by cars that are on the road (fuel taxes). Some of this reduction will relate to trips on non-congested roads. As such, existing road user charges are a very blunt instrument for combatting congestion. A better approach might be to levy an appropriate toll on congested routes during the period of congestion. Of course, this will depend on the cost of detecting use and collecting the toll. But if these are not prohibitive, then such a toll would tend to encourage people who can vary their travel plans (either in terms of timing or route choice) to do so. Of course, the additional travel time and frustration on congested roads will also encourage this, but the congestion costs caused by an additional car on the route will be spread over both that car's occupents and the occupents of the other cars on the route.

david p said...

Hi Harry,

FWIW my interpretation is that Sinclair D. is using congestion in a looser way. If we fixed all of the things Sinclair discussed there would still be congestion - just that there would be more cars (which would mean some benefits). The effects of the things Sinclair D. highlights look like congestion but perhaps it is better to think of them as restrictions of supply (a reduction of the supply of roadspace that can be travelled at a certain speed (this is the good) at a point in time). The costs and benefits of these should be assessed (and costs of traffic disruptions minimised). However, as stated above, there will still be congestion just as even if we doubled the available roads!

hc said...

Sinclair, The evidence of a problem is the widespread evidence of traffic congestion at the morning and even peak periods and increasingly during the day. Restrictions on supply might play a miniscule role but the main effect is the strong sustained growth in the demand for travel by private car.

The charges such as rego and petrol costs that are in place, as Damien suggests, are not travel time and route specific and will have little effect of congestion.

The difficulty with your argument - apart from the fact that it is wrong - is that it suggests supply-side tinkering can resolve the problem. It cannot. The costs of scaling up road supplies to substantially impact on congestion is just too high because land is so intensively developed.

All this has been set out so many times by so many different authors and most recently by VCEC. I hope your unfounded views don't have an impact.

And Sinclair the consistently anti-tax attitude you have is particularly inappropriate when the taxes sought are not revenue-raising exercises but attempts to deal with significant social problems.

To describe attempts to internalise externalities as a tax grab is irresponsible. And that is what you wrote.

Sinclair Davidson said...

I’ll try to respond to your comments one by one:
'Restrictions on supply might play a miniscule role'
They may. But the US evidence indcates they constitute up to 60% of the problem. Now the equivalent stats don't exist for Melbourne (to my knowledge), you may be correct. Until those stats do exist, we cannot be sure what is causing congestion, if any. I am sure that taxpayers and motorists don’t want to pay twice for incompetence on the supply side which is what they would do if a toll were the solution to all problems.

‘the strong sustained growth in the demand for travel by private car’
Unlike the eternality approach, I view this as being a good thing. Demand for travel by car is a function of economic activity. Why would anyone want to slow down or inhibit that economic activity? To be fair, this is a variation of the old ‘sand in the wheels’ can improve things argument. Well, they probably can’t. Governments wanting to follow policies that can inhibit economic growth should have to explain that to the electorate – or have someone like me explain it.

‘The charges such as rego and petrol costs that are in place, as Damien suggests, are not travel time and route specific and will have little effect of congestion.’
I totally agree. But this goes to the point that access to roads is free. It is not free. To the extent that the externality argument relies on ‘free’ roads, it is undermined. Your argument is that roads are over-used (excess-demand) because of zero-prices. To the extent that prices are too low (but not zero) at some times of the day, they must be too high at other parts of the day. Now you might argue then for a time-adjusted pricing system. Maybe. But what about the existing pricing structure? If tolls come in, then rego, petrol levies etc. must fall. I have yet to see that proposal. As I indicate, why is the solution to all societal problems more tax? If you want efficient pricing, then I’m sure for many (if not most) consumers prices will fall.

‘The difficulty with your argument - apart from the fact that it is wrong’
How do you know it wrong? I concede that it does not conform to the externality framework in neo-classical economics, but does that make it wrong. The point here is more fundamental. While an academic can simply state an argument to be false according to some normative criteria in the class-room, there is a market for ideas. In the class-room and exam Harry would be quite correct to be scathing of my argument. Students are required to understand the externality argument. I do understand it, I don’t believe it. In the marketplace for ideas, Harry cannot simply assert I’m wrong (well he can) but should rather refute the arguments (one by one).

‘All this has been set out so many times by so many different authors and most recently by VCEC.’
The phlogiston and ether died a hard death too. The ‘rampant externality problem’ will be with us a long time. More seriously, facts are not weighed to establish their veracity. Solutions to social problems need to be convincing. First we need to be convinced that a social problem exists.

‘I hope your unfounded views don't have an impact.’
In all honesty they probably won’t. The government wants the money and most economists will argue for a price mechanism solution to preceived problems of congestion. I will continue to point out that even by the loose definitons of congestion there is little, and in any event the technology to implement a price solution does not yet exist in a cheap and reliable form.

‘the taxes sought are not revenue-raising exercises’
If this was true, and it isn’t, the government would undertake a revenue-neutral policy. That is they would reduce other taxes by an equivalent amount to the increase in tolls.

‘To describe attempts to internalise externalities as a tax grab is irresponsible.’
That is a normative statement.

hc said...

Sinclair, Do you have a link to the US report that suggests 60% of congestion is related to supply-side traffic delays.

The reason we want to reduce car travekl is that there are external costs associated with it. This is not being anti-travel but anti-theft.

Access to roads is generally unpriced - it is not free. You pay fopr access financially and through waiting in traffic jams.

Most road pricing proposals include a revenue neutraqlity idea - to make it transparent that no revenue grab is sought.

It is irresponsible to say that efficient pricing roads is a tax grab since it never is. It is an attempt to make motorists bear the full social costs of their travel.

Sinclair Davidson said...

I will send you the reference for the 60% stat (probably Monday). On the social costs, we disagree on the extent of the social cost and internalisation. My argument is that a lot, if not all, of the social cost is already intenalised for most, if not all, drivers. Further, while I'm sure your intentions are noble, I have no doubt government intentions are less so.

Uncle Milton said...

Sinclair, most congestion occurs on suburban arterial roads during peak hours, Monday to Friday. Special events, which occur on the weekend and nearly all in thr inner city, and over sized tram stops, have nothing to do with it. And traffic synchronisation measures, like the timing of traffic lights, may not work perfectly, but but least they are working in the right direction. Or are you saying that the traffic authorities are on balance making things worse? In Melbourne, new roads to relieve congestion are being built all the time, the latest one being Eastlink from Mitcham to Frankston. It will be tolled, but the tolls won't vary by time of day, so defeating part of the benefit of the road. Not everyone who travels on that road will need to do so during the peak times (shoppers, for instance), and if they were to pay lower roll during non peak times they might be incentivised to travel at those times.

Your argument that in practice governments will just take the extra revenue and not reduce other taxes may or may not be right, but it doesn't make the economics unsound.

The technology to implement a price solution does exist in a cheap and reliable form. It is called an e-tag. As I write, drivers are driving through tolling points on certain roads, their e-tags are beeping, and they are being charged. The charge can be made time-of-day dependent at the flick of a switch.

The price solution to exteranlies has been known since 1920, when Pigou published his Economics of Welfare. Of course, there are theortetical alternatives to the externality problem, like assigning property rights, a la Coase. Mostly, this is completerly impractical. There are places in the world where you can buy access to special transit lanes, giving you superior property rights of a sort, but this just amounts to pricing the exteranlity in a different way.

Sinclair Davidson said...

The e-tag simply shifts the problem onto non-tolled roads. I would think that the pricing solution would involve payment for all road usage not just particular roads (at particular times). This would be some sort of satillite tracking technology.

As I keep saying, I have no problem with understanding the externality economics, or even Pigou. But I am reluctant to impose 'blackboard economic' solutions into the 'real world' until I'm convinced that there is a problem and that the solution is appropriate. The VCEC draft report demonstrates neither of these criteria.

The message I take from Coase is the reciprocal nature of externality. It is discussed in the last few pages of a very long paper. I also like Buchanan on externality. I am unconvinced by Pigou.

Uncle Milton said...

"I would think that the pricing solution would involve payment for all road usage not just particular roads (at particular times)."

Sinclair,

the problem only exists for particular roads at particular times so there is no need for a satellite solution. If it was possible for people to take alternative routes to avoid priced-roads, they would be taking them now to avoid congested roads, and the congested roads wouldn't be congested anymore.

Take Melbourne's Westgate Bridge. Congested during peak hours, free flowing at other times, and there is no realistic alternative route. Likewise, the Tullamarine tollway.

Sinclair Davidson said...

'Take Melbourne's Westgate Bridge. Congested during peak hours, free flowing at other times, and there is no realistic alternative route.'

Actually, I drive over the Westgate to work. On the days that I see it to be very busy, I jump off at Williamstown road and get into the city on Footscray road. So there is a viable alternative.

I have to say that often while travelling (at speed - but never speeding) on the Westgate I hear the news report saying how terrible the traffic on the Westgate is. The reports are often worse than the reality. The traffic does slow when people stop on the bridge (but that is usually associated with a break-down, or unfortunately a jumper). Most of the difficulty I have on the Westgate is when the authorities shut down a lane - as I indicated most traffic difficulty is a function of the traffic authorities and the choices they make.