Ken Henry discusses road congestion pricing and appears to also get it wrong like Harry Clarke

Henry says

Most of the time, most cars impose minimal costs on other road users. However, when vehicles drive on a congested road they impose costs on other drivers. Each driver thinks of their own need to get to their destination, not considering how, by taking up space on the road, they impinge on the ability of other drivers to do so.

We’ve seen this argument before, from Harry Clarke, which is where Ken Henry may have got it, perhaps.

He’s wrong. Both John Humpheys and I demonstrated the logic/arithmetic error in this argument in a thread some time ago and Harry never came back to defend it. It appears to be wrong because they/he doesn’t appear to be counting correctly, which seems a pretty surprising error for a Treasury Secretory head to be making although i forgive Harry of course as I’m eternally gracious that way.

Here’s Harry saying pretty much the same thing in the thread by giving an example like Henry’s.

John suppose I enter a crowded road and work out that the time and petrol cost of getting to my destination is $15 (that’s the personal marginal cost of me entering traffic) but that by doing this I get a benefit of $25 so off I go.

I head off to this particular destination (’The Honey Motel’) along with 1000 other motorists who go to about the same place. I increase their travel times by 1 second per car so I cost them 1000 seconds or by about 17 minutes. If the value of their time is $50/hour I have increased their travel costs by about $14-20. That’s the larger social marginal cost of me entering traffic. It’s bigger than my personal marginal cost.

The $14-20 is the charge that should be levied on me if there was efficient tolling. So now my total travel costs would go to $29-20 (my new personal marginal cost of making the journey) so I forego the trip.

Its sad. In this deal I lose consumer surplus of $10 as a consequence of the charging since I now don’t go. But society saves $14-20 providing a $4-20 social gain. That is the basis for congestion pricing.

Without it I would have got my $10 surplus but other motorists would have lost $14-20.

Meanwhile back at ‘The Honey Motel’ my irrepressively sensuous adoring fan growls with annoyance into her mobile phone that ‘this is absolutely the last straw’. But I don’t believe her. I know that I project a well-worn, animal magnetism and that she will still come to offer me new invitations even though I valued an afternoon with her at a mere $25.

Here’s John response

The 1000 motorists lost 1 second each because of you.

And you lost 1 second because of each of the 1000 motorists.

So the cost to you of the congestion was 1000 seconds (or $14 assuming $50/hr). And the cost by your contribution to the congestion was 1000 seconds (also $14).

The problem with your example is that you shouldn’t have started with a cost to you which included the congestion cost. Start with your personal non-congestion costs… then add in the congestion cost to find the optimal decision. To work it out for society you should add $14 to the cost. To work it out for yourself you should also add $14 to the cost. So the incentives are already correct.

Here is mine.

Let’s take your example of one driver costing all the other drivers combined 1,000 seconds (1 second each). That ‘s fair enough, however you simply can’t account the cost of one driver to the rest without accruing I second from each of the other drivers to single driver.

So here’s your math equation (999 seconds-1000 seconds)+ 1 second =0. (Modified it)In other words each person costs the other drivers equally.

Road pricing shouldn’t be about getting bloody cars off the road. The priority is about pricing roads to usage and reflecting true replacement costs. it may very well happen that people do get off the road and the marginal user begins to take public transport. However as far as public policy is concerned there are also other issues involved like quality of life etc. Personal preferences made and dressed up to appear like an economics argument shouldn’t see the light of day.. Henry really shocks me.

82 thoughts on “Ken Henry discusses road congestion pricing and appears to also get it wrong like Harry Clarke

  1. Playing devils advocate I might note that your time might be worth more than my time. We kind of acknowledge this by giving way to ambulances. And the fact that a high toll will cause some people to not travel but not all people also reflects variances in travel utility. Paying for congestion in dollars may in some circumstances deliver higher utility than paying for it in time.

  2. That may be true. Terje might be a well to do IT guru, earning twice as much an hour than any economist (i.e, $100/hour).

    Terje might be catching a flight and Mark cuts Terje off. It only takes a second but because of this he misses several green lights in the process.

    The flight does not go for another two hours, and Terje misses it by seconds.

    Has Terje incurred an externality of two hours and one second (or $200.01) or is this a load of steaming tosh?

  3. Terje… you are right. I was answering Harry within the bounds of his example (where he assumed constant value on time).

    Jason made a similar argument at the pub once. But his mistake was to assume that your wage was an accurate indicator of the value you put on time — meaning that the time of high-income people is more important than the time of low-income people. There are many reasons for thinking this not the case.

    I’m all for people going out and getting rich if they like… but I don’t think that makes them an inherently better person who cares more about their life than a low-income person.

  4. Well said John. Where were you a while back when I was arguing at catallaxy against road pricing as a general solution to congestion?

    Besides, drivers don’t impose costs on other drivers in any sensible sense. They each impose costs on themselves. To argue otherwise is like arguing that a group of suicide bombers is each murdering the others.

  5. Leaving aside all the complications of whose time is lost and who values that time (and the other factors like flights missed and ER rooms arrived at in time), the trouble with John’s point is that the marginal cost of each car on the road is not 1 sec (or whatever value you use).

    The trouble is that the marginal cost of each car in traffic is different. The first car imposes zero costs on other cars, the second car imposes a miniscule cost on the first car, and so on. Basically, the marginal social cost of using a road rises with each car, so the congestion cost that each driver faces when deciding to use a road (the average of all the previous cars) is below the marginal cost of adding that person’s car to the road.

    So people aren’t facing the marginal social costs, so they consume more than is optimal. Hence road pricing, which despite its imperfections, has the great potential to bring private and social marginal costs closer together.

  6. How are other car drivers “third parties” to their aggregated behaviour?

    If there is a common pool problem, what matters is if the benefits of high traffic flow outweigh wait times.

    They seem to do, even if the wait time has deadweight losses. But also we would have a situation where non drivers obtain positive externalities from those who drive.

    So, should we then tax the car less?

  7. Robert C.

    As I said in the body of the thread, I really don’t have an issue with road pricing per se providing of course the objective behind that is to more accurately charge for the cost of maintaining roads and not to necessarily keep people off roads to satisfy some distorted personal preferences.

  8. Also Robert:

    Leaving aside all the complications of whose time is lost and who values that time (and the other factors like flights missed and ER rooms arrived at in time…..)

    Well you really can’t leave it aside as that is the example Henry based his comments on.

  9. JH, I’ll have another go. I am trying in the example to show how the externality arises and how efficient pricing yields a welfare gain.

    I head off the Honey motel because the benefits of me doing so exceed my private costs.

    From society’s perspective however I add more in costs to the travel plans of others in making this journey than I gain net in surplus. If I am charged for the extra costs I inflict I lose the surplus but society gains more than I lose which is the basis for seeing the unpaid costs I impose as an externality which can be internalised by charging for it.

    I think my arithmetic is 100% sound. No mea culpa’s here.

    Oh, and are you really suggesting that I implied rich people are more worthwhile than poor people because they earn higher wages? I didn’t say that.

    In the income/leisure model of labour supply the wage is the opportunity cost of not working – taking ‘leisure’. So the value of time in a car is proxied by the wage or some fraction of it. That’s all.

    If different people have different values on time then it might be the rich who benefit most from congestion pricing but this has nothing to do with the efficiency case for a toll. It is simple to show a compensation scheme can be devised (based on toll revenues) which more than compensates those ‘tolled-off’ the road.

  10. Harry:

    You’re example is wrong. The correct equation is (999-1,000) +1 =0.

    You simply cannot give away seconds to 999 other drivers without them giving you back a second. Any way you want to cut it and dice it’s simply wrong to say otherwise.

    In addition Mark’s point is that as a common pool problem it is simply more sensible to let people work things out for themselves. That is that as a group they may be perfectly happy to suffer the congestion amongst themselves.

    But to repeat your (and Henry’s)logic/arithmetic is simply wrong and the example seems to indicate that the argument is based on that.

  11. jc, according to the govt, in 2005/6, total road expenditure for all govts was 12bil and fuel excise 9.6bil. A small increase in fuel excise covers the costs of roads while ignoring that non-drivers also benefit from roads in ways not reflected in prices of goods.

    Road pricing to manage congestion biases road use to the wealthy, which is wrong for a public good.

    I think Harry’s example missing something. It is not just Harry who doesn’t go, it’s everybody whose expected benefit from the trip is less than the toll. So the saving from harry not going is less than expected because a bunch of other people also don’t go. Also driving tends to displace leisure not work.

    It is also rubbish to say that me paying a road use tax compensates anyone who also chooses to go to work in their car.

    The only real saving to society is the reduced cost of building roads because pricing is used to reduce flows. The decisions made are nothing more than an arbitrary tax that has the goal of excluding those for whom the toll is too high a price to pay. It does not internalise and externality.

    I think this is one of the areas in which economists live down to their reputation.

  12. Harry Clarke:

    Your story above is another example of personal preferences masquerading as economics. It also doesn’t negate the fact that both examples (yours and Henry’s) are conceptually and logically wrong.

    The fact that you don’t wish to confront this other than saying you’re right is not going to change it.

    Sorry but that’s the way it seems.

  13. And your values seem skewed! If I had an adoring fan waiting for me at a motel, I’d give that a higher value than the dollars you’re quoting!

  14. HC — I know you are trying to show an externality, but my response was that the “externality” impacts on you just as much as society and so isn’t really an externality.

    The congestion gives a cost to everybody else… but it also gives a cost to you. If each person contributes 1 second to the delay and there are 1000 people, then you have caused 1000 seconds delay, but you also face 1000 seconds delay by joining the road. So your private congestion cost equals your social congestion cost. So long as you factor in your personal congestion cost (1000 seconds), then your decision is optimal.

    To put it simply: congestion is it’s own disincentive.

    And no — I didn’t suggest that you implied rich people are more worthwhile than poor people. You did exactly the opposite: you assumed all time value was equal. I prefer your assumption as, while it’s obviously not true, I think it is the most appropriate abstraction. It was Jason Soon (in a separate conversation in a pub) who suggested that we should use wages as a proxy for the value people put on time.

  15. Can someone rationally explain to me why would/should anyone be concerned with congestion other than the people who voluntarily choose to become part of it.

    If roads are properly priced for replacement costs why would Harold or Henry have to become concerned if there are a lot of people using the road. People can choose to become part of it or seek an alternative.

    There’s something really disturbed with a mindset that wants to charge for congestion to simply try to price people off roads. It really is heavy handed authoritarian bullshit.

  16. Whether or not congestion charges are a good or bad idea they are pretty common. A lot of Internet providers will offer a different download allowance based on the time of day. So at 2am when link congestion is low they will let you generate traffic for a much lower fee. Off peak phone calls are nearly as old as the telephone. Phone companies can give you a congested tone (it sounds slightly different to a busy tone) but they prefer to price most congestion out of the system and I suspect consumers prefer it this way also. Likewise the transmission of electricity during congested periods will cost you more.

    In terms of the worth of peoples time we have transit lanes to acknowledge that a car with 4 people in it that gets delayed 10 minutes is of more concern than a delayed car that has only one person in it. An easy reform would be to still enforce transit lanes during busy periods but to allow any vehicle to use the transit lane if they pay an extra fee.

    Congestion is in essence a mismatch between supply (the supply of an open road) and demand. Rather than a price change solution we live with a quality change solution. A congested road is a lower quality transport solution than an uncongested one (which is why we try to bypass traffic jams by taking a different route). I’m not convinced that a temporary quality drop is better than a temporary price increase. Personally I hate wasting time in traffic and will routinely take the toll road if it lets me avoid congestion.

  17. p.s. Traffic congestion reminds me of soviet era shopping. You do a lot of waiting because the price is too low. And the producers have little incentive to increase supply.

  18. The empirical evidence that first brought the eponymous Braess’s paradox to Braess’s attention shows that road traffic congestion externalities really can happen. It is not necessarily true that “…your private congestion cost equals your social congestion cost… congestion is it’s [sic] own disincentive”. Sometimes, the numbers do not work out with a precise matching of costs and incentives.

    Of course, determining whether such things are happening, and if so what the right things to do about them are, are distinct questions needing data and theory to sort out on a case by case basis – but that’s serious stuff, not something to be ruled out on the grounds that a priori there is never a problem. At most you can say it seems a sufficiently remote possibility that you would need good reasons to incur the costs of such studies. But Steinberg and Zangwill found that “Braess’s paradox is about as likely to occur as not occur” – and that’s only one form of non-optimality.

  19. PM:

    the point is that the externality is basically internalized.

    However this is far from the point that good old Harry and NSW right wing faction ALP member Ken Henry have in mind. Their basic thrust (if I can liberal in the use of that word with these two) is that they really want to keep the cars off the road as their version of optimality is in fact their personal preference.

    In any event he discussion relates the example Harold gave in the previous blog comments and ALP member Ken used in his speech. They’re both wrong.

    PM, have you got over that angry period with me or are you still in angrified.

  20. I think its pretty simply, just bid against other drivers for a spot on the road. Your bid will be based on how much you value that spot, which will be reflected based on the current amount of traffic and how rich you are and where you are going etc. If the roads are empty there are plenty of spots, if they are conjested there are hardly no spots. That could at least be the basis on how the tolls are set for different times of day. This system would encourge employers to change shift times as it would be equivalent to people being given a pay rise. Eventually the usage of roads would even out over most of the day. A win for everyone.

  21. just bid against other drivers for a spot on the road

    When it is airplanes trying to land at an airport we call them slots not spots. The same goes when rail freight is being booked in for track time.

    In terms of Internet protocol (ie IP packets) congestion looks much the same as link failure. When I enter a stretch of road my intention is to drive on it. When I find that somebody is parked in my way I regard this as a temporarily broken transport medium. What is the good of a road you can’t drive on?

  22. why should the government worry about people making their own decision to travel on a congested road?

    The soviets didn’t worry about people queueing for sugar so why should our government be any more caring. Hmmm.

    Roads are mostly government owned. I think they should privatise them (without signing up to restriction to competition contracts) or else manage them properly.

  23. I still want to know how the originator of an externality is also a third party, especially when the information of the “externality” is known.

    It just abuses the definition of externalities to the point of uselessness.

  24. Terje — I’m not arguing against road charges, but the “externality” argument is wrong.

    PML — The only way for private costs not to equal public costs is for people to have a different value on time. That is undoubtedly true, and so the system will not be perfect. There are no perfect markets in the world. However, I find it hard to believe that a planner will be able to make an effective judgement about the diverse and ever-changing time-value that people have, and I’m certain they will then fail to arrange for proper compensation for those made worse off (creating another inefficiency).

  25. John Humphreys comment hits the nail on the head so hard it gone through the board.

    That’s it.

    The externality argument is wrong to the point where the terms has no meaning any longer, as Mark Hill says.

    How the hell is planner supposed the devine all this. For example would anyone want harry or Henry becoming the congestion planners for this racket after knowing that what we do about their personal preferences? Fme.

  26. JC seems to have suggested that a motivation (for Ken and Harry) for congestion based pricing is to reduce the number of cars on the road. I’d argue that maximising the number of cars on the road is in any case the wrong metric. The purpose of a road network is to provide transport. A parking lot can contain a lot of cars but we don’t want our road network to emulate a parking lot. Of more interest is the road networks rate of throughput.

    A very simple example of a road network is a toll road from point A to point B with a toll of $0 being a special case. A private operator of such a road makes revenue when cars pass the tolling point and obviously a high flow rate is in their interest. At times of severe congestion (perhaps due to exit bottlenecks) the rate of revenue can drop to zero as cars cease to move. Of course the simplified example won’t generally experience congestion at the exit point because the road network beyond the exit isn’t included in such a simplified model. And any queing at the entry point isn’t going to impact revenue and in fact it is probably good for revenue because it helps to ensure that the pipe is running at capacity rather than being idle. A tollway operator benefits from high input congestion (strong demand for service) and low output congestion (high client processing speed).

    To elliminate most congestion in a network requires a price signal eminating outwards from the bottleneck to discourage excess ingress of traffic and to promote alternate options.

    In computer networks the price signal is packet loss and the management process is window size and dynamic routing however the packet loss concept would not work on roads because the cost of destroying excess cars that are stationary for too long is way too high and the owners of the traffic occupy those cars rather than being separate from them (data traffic is in any case more analogous to a series of fleet operators rather than discrete motorists) and retransmission of destroyed motorists later on isn’t possible. A car network that loses cars isn’t viable. In data networks (eg the Internet) there are also generally few obstacles to adding excess capacity at choke points and profiting from the flows generated. Road networks are harder to augment due to physical spacing issues and the cost of building incremental additions.

    Centralised ownership (control) of the entire road network (or big chunks of it) is probably the a workable way to properly integrate price signals and congestion management. This approach is common in private data networks where management is centralised. Of course for road networks this would also enable monopoly pricing. As such there is an argument for public ownership of road networks on an efficiency basis. Whether Ken or Harry are the right people to manage these networks is another matter.

    Congested roads may not entail an externality but they do entail collective costs. In theory price ought to be a viable management tool. Working out the details may be tough but I would be very cautious about dismissing this idea.

    One easy area for improvement without centralised changes is in the wider adoption of GPS systems with integrated real time traffic reports and dynamic route calculation. If used widely I suspect that these will provide a lot of benefit. In terms of any public investment I think it is probably best spent in enabling such technology via better real time data collection and distribution and in enabling GPS equivalent signaling in tunnels. At the end of the day this will only facilitate improvements in terms of localised congestion and won’t really resolve generalised congestion, except perhaps by allowing people to know ahead of time how long a journey will take.

    A further refinement would be to integrate electronic toll operators (eg E-Toll, E-Way etc) with the GPS service providers (Navman, TomTom etc) and for them to then negotiate terms with road wholesale operators (tollways) and provide a packaged solution that allows the retail consumer to simply trade off time and cost for the complete end to end journey. I suspect that this would allow for more meaningful aggregation, management and routing of traffic.

  27. Terge:

    Their motivation is not congestion as both have said. Their major motivation is global warming. If their major motivation isn’t road congestion but their personal preferences the entire thing is corrupted.

    In so doing you could end up causing great hardship to people.

    The point of the thread was to show that their example is wrong and it is wrong from a logic/numeric point.

  28. If you begin with the premise that you want cars off the road because of global warming and not congestion you’ve already screwed up the entire thing.

    Yes, you could put all sorts of contraptions on a car and who pays for them. Who pays for the additional infrastructure required? Are they accounted for in a proper c/a?

  29. Consumers are already paying for GPS systems and for tolling systems (a bit of stick has been involved in the latter) and getting benefits from both. An integrated solution could be deployed by the private sector without compulsion. In fact I expect it will be soon enough if the government stays out of it. At that point the government could elect to make all roads toll roads with a distance and time of use fee structure. The Dutch are headed in this direction with a more interventionist approach. Either way it seems like a logical technological evolution with the potential to solve a lot of problems. I would agree that there is also scope for social engineers to get on the band wagon but this isn’t a reason to halt progress. Intelligent networks with dynamic demand management via pricing are the way of the future for both roads and electricity. To me this represents a correction to old semi-socialist pricing structures.

  30. I don’t disagree with your points Terje, however I would love to see how you are going to keep the personal preferences dressed up as economics out of it.

    I also think that a lot of this congestion stuff is self contained. In other words all the users are screwing each equally and if they are willing to live with it than so be it.

    there’s also the other issue too. Most people, not all, work 9 to 5 so it’s not as though they want to be on a congested road depending where they live they have to be and in some cases the cars reduces the hardship of otherwise having a 3 hours commute per day.

  31. Terje:

    Why is congestion a socialist signal as that has really got me beat.

    Congestion occurs because people want to use the road at the same time going to and from work for the most part. This isn’t a signal of socialism at all. In fact the opposite would be true where through artificial means such as a congestion tax it would be limiting people’s choices in modes of transport.

    Going to and from work with others at the same time is not a sign of the socialist menace. Sorry that just doesn’t wash.

  32. I think Terje’s point would be that if roads were privately supplied then the owners would probably charge for the privilege of driving on their roads… and if there was high demand (ie many cars) they could charge more… and if they could charge more then there would be an incentive to build more roads… and so a system of private roads may have less congestion.

  33. Yes, thanks John. However the problem is that roads leading to the CBD are quite possibly the closest thing that you could get to a monopoly for obvious reasons. It would be impossible in some cases to build alternative routes. Even if the current roads were privatized you would also require some form of regulatory supervision (and even that could be privatized I think)

    Of course there is the case of substitution and in this case is public transport which gets to my point. Public transport could cause people great hardship especially those not close to a hub.

    The more I think about it the more I think that congestion taxes are a rotten idea. This is altogether different to making costs more aligned to road usage of course.

    There’s nothing socialistic about a lot of people trying to get to work by 9 am in the morning and leaving at 5 pm for home.

    And why stop only with congestion tax for roads. If you one was a equal opportunity congestion taxer you would also tax congested public transport at peak hour too, as after all this is congestion too.

    And this is where Harry Clarke and Ken Henry’s idea hits a wall. If you hate congestion, tax it all not just one particular form of congestion. Tax public transport too.

    This is really a form of social engineering.

  34. JC – In the soviet union prices were generally fixed and as a result there were frequent shortages and at other times gluts and a lot of waiting around in queues. It is my contention that we currently price roads like the soviets priced sugar and milk and we suffer the same types of consequences. Price is an efficient way to moderate demand. And as John suggested it can also signal for an increase in supply.

    In terms of a congested road to the CBD where we perhaps can’t build more roads we can still respond to a price signal by travelling at a different time, car pooling, bus catching, riding a motor bike or push bike or getting a job outside the CBD or a home inside the CBD so we can walk to work. There are many ways we can respond to prices even if directly increasing supply isn’t one of them.

    Free roads, and even fixed price toll roads are representative of a semi-socialist price structure. The price is nice but you do a lot of queueing and the quality is poor.

  35. Terje

    Getting to and from work at the same time is not socialistic behavior. As humans we follow a time clock.

    Roads aren’t free as we’re paying for them directly through registrations and indirectly through tax.

    If congestion on the roads is suddenly a problem then why isn’t congestion on public transport at peak hour.

    And if roads are socialistic then why isn’t public transport?

    My point is that if congestion or rather human time clocks should be taxed if people drive cars, then the whole thing ought to be taxed. That’s absurd obviously which shows the whole idea of taxing road congestion is a crock of shit.

    It’s just another of Harold’s personal preferences and now he’s joined in that by the wombat whisperer.

  36. Getting to and from work at the same time is not socialistic behavior.

    I did not say it was.

    Roads aren’t free as we’re paying for them directly through registrations and indirectly through tax.

    Registration is just a tax so what you are saying is that our taxes pay for roads. No argument.

    If congestion on the roads is suddenly a problem then why isn’t congestion on public transport at peak hour.

    It is a problem. Have you never caught public transport during peak hour? I would note however that trains in Sydney do now charge a different fee based on the time of day. Travel at off peak times and you pay less.

    And if roads are socialistic then why isn’t public transport?

    Road pricing is what I was talking about not roads. I’d say they are both socialistic because both are generally owned and operated by governments using fee structures that ignore fluctuations in demand.

    My point is that if congestion or rather human time clocks should be taxed if people drive cars, then the whole thing ought to be taxed. That’s absurd obviously which shows the whole idea of taxing road congestion is a crock of shit.

    I don’t see your conclusion as following from your statement.

    It’s just another of Harold’s personal preferences and now he’s joined in that by the wombat whisperer.

    I’ll let Harry speak for himself.

  37. Terje says:

    Registration is just a tax so what you are saying is that our taxes pay for roads. No argument.

    Terje says:

    Free roads, and even fixed price toll roads are representative of a semi-socialist price structure.

    Are roads free or not?

    Clarke’s objective is to try and get people off roads as it supports his alarmist version of AGW. The wombat whisperer also agrees.

  38. Most roads are free to the user. Yes the roads are paid for via taxes but those taxes are not much correlated with time and place specific road demand. Fuel taxes do some what correlate to time on the road but it’s a crude form of pricing.

  39. People who think that congestion pricing can be used to encourage travel at different times need to, umm, get a life. Terje, the internet example is not really very useful. School starts at the same time everyday, whereas you can conveniently defer your wank until late in the evening.

    Here is SEQ there is an annual furore about the lack of daylight saving and the disruption to business. I think that evidences just how likely it is that congestion pricing will materially change working hours anywhere.

    London to a brick the vast majority of people with discretionary travel times already avoid peak hours.

    “In the soviet union prices were generally fixed and as a result there were frequent shortages and at other times gluts and a lot of waiting around in queues. It is my contention that we currently price roads like the soviets priced sugar and milk and we suffer the same types of consequences. Price is an efficient way to moderate demand. And as John suggested it can also signal for an increase in supply.”

    This is true until you apply it to roads. You will notice that the Soviets had prices as well, yet their prices did not seem to lead to a match between supply and demand. Hmmm, what was missing that stopped prices having their usual signalling role? I wonder if the same effect would happen here with roads?

  40. Getting rid of planning restrictions and privatising already publicly owned roads would see a user pays system. Virtually all of the tax paid via excise and registrations would be unjustifiable.

    Privatisation would beat a Government business enterprise hands down. For example, a road in a residential area need not be for profit, but maintained under community title and strata type fees.

    What is missing from the pricing now is not only a profit motive but any chance of any profit taking being reinvested in the service. No privatisation, no meaningful increase in supply to meet demand.

  41. Mark – yes but then again maybe no. If you privatise an expressway which simply links point A to point B the incentive for the private profit seeking owner is to have high input queuing (to maintain utilisation) and low output queueing (to maximise flow rate). The incentive is not to mitigate the congestion problem but merely to shift where it occurs. A private owner of an entire road network would have the right incentives, however they would also have a monopoly. I’m not convinced that a private profit seeking road monopoly would be such a bad thing but it would certainly be a hard concept to sell politically.

  42. *If you privatise an expressway which simply links point A to point B the incentive for the private profit seeking owner is to have high input queuing (to maintain utilisation) and low output queueing (to maximise flow rate).*

    If you mean you transfer congestion to multiple lanes per each direction, plenty of off ramps and congestion spread out over a greater distance at a higher travel time…

    The M5 for example can be slow for a toll road, but if it wasn’t built, the Princes and Hume Highways would be enough to send you postal by now.

    The reason why we don’t have more like the M5 is because of planning restrictions etc. The level of willingness to pay is not linked to fully formed incentives to build/supply roads and earn profits.

    It is like you’re judging the efficacy of markets in general in Norway. You should be looking at Hong Kong.

  43. The incentive is not to mitigate the congestion problem but merely to shift where it occurs.

    How so? How would you change people’s working habits away from 9 to 5?

  44. Obviously if you got rid of registration, excise, company tax on coal liquification, and fired lots of people….. then supposing you had peak-time charging, well this would alter a lot of peoples working hours.

    “How so? How would you change people’s working habits away from 9 to 5?”


    How do YOU think Lamcambria? You don’t think do you. Its called price-rationing.

  45. JC – you change where people live relative to work, you change the number of occupants per car, you change the incentive to catch buses where viable, you encourage people to work different hours. And you don’t solve congestion you mitigate it. In any case I’m not offering one particular solution just trying to be analytical about alternate scenerios.

  46. They just closed the one, mark 🙂


    Given that people can make their own decisions about how they get to work, why would you tax them even more? Clarke hasn’t said he wants to remove the excise tax and neither has the other whisperer. In fact if i recall he’s said he wants to keep it there obviously to promote his social engineering fiasco.

    So you’re supporting two hits, the excise tax and congestion tax. And a congestion tax would be imposed on people that basically internalize their costs on each other.

    you change where people live relative to work,

    We’ve been doing that. According to Alan Moran only 15% of the morning peak hour traffic is heading to the CDB. Work travel these days is very diffuse.

    you change the number of occupants per car,

    Doesn’t work for the most part.

    you change the incentive to catch buses where viable,

    Yea, how? By subsidizing them?

    you encourage people to work different hours.

    As an employer why would you want to?

    And you don’t solve congestion you mitigate it. In any case I’m not offering one particular solution just trying to be analytical about alternate scenerios.

    Cars a comfortable and people are still prepared to use them despite the traffic. Why is it the government’s business if the congestion as internalized?

    The US has pretty liberal labor markets and peak hours still exists there as much as here. So does HK.

  47. JC – I’m offering my own views so what Harry thinks isn’t relevant to what I’ve said. If Harry and I agree then I’ll live with it and if we don’t agree then likewise.

    If roads are going to be owned by the government then I think we would agree that these roads should be paid for by motorists rather than taxpayers in general. At the moment motorists pay for roads via fuel taxes but there is no law of the universe that says it must be so. Electronic metering isn’t automatically worse.

    I’m not in favour of subsidising public transport. Like roads it should be funded via user pays.

  48. Okay, you’re bringing in points that isn’t what the thread is about. The thread is about Harry&Henry basing their premise on a logical fallacy.

    Harry has been told about it and refuses to believe it. Henry hasn’t been told about it or otherwise I wouldn’t expect him to be making stupid assertions based on a false premise.

  49. Terje, the fuel excise currently collects the large majority of total road spending anyway. Increasing the tax on roads by congestion charging is for a different purpose and has different effects. Mainly, it benefits the relatively wealthy when it comes to access to a public good.

    Privatising all the road network is nonsense on stilts.

  50. *Privatising all the road network is nonsense on stilts.*

    Why would you say that?

    Roads generally get built on developer’s money in residential areas. All that would need to happen is they get linked in as part of community title.

    They would be communal – but private – simply not nationalised.

    OTOH, It is fairly agreeable that private toll/arterial roads are better than publicly funded roads. I don’t think you need convincing here.

  51. Ok, for a start, privatising the road network means selling off the existing roads to private buyers who will only be interested in buying them if they can earn a return. If you look at the street outside your door I think you will stuggle to identify an opportunity for construction of a competing road to your driveway. You may then be concerned about the pricing opportunity for the person who buys your street.

    Yes, new minor roads and some bigger roads get built by developers, but they are often connected to other roads and so are part of the total network instead of a branch servicing only the new estate. As such, expecting a body corporate consisting of the lot owners in that estate to own the road is to put on them a burdern that is unfair because they are liable to all of the road users for the state of the road. Mr and Mrs average would end up getting sued for every accident on their private property road, even though the road user might have been on the way to somewhere else.

    Major roads are built on a long term planning horizon by the relevant authorities using tax money and developer infrastructure contributions. A road at X might be built with money provided by developers of estates within a 20 mile radius.

    I don’t think that toll roads are “better” than public roads and I think in many cases they are unfair because one part of town has an existing good quality public road network and other areas end up having to pay tolls for a decent road. As far as possible, taxpayers should be treated equally in the provision of public goods that do not have a welfare element.

  52. I would actually give roads away in residential areas. Effectively property owners have paid for them in development fees anyway.

    You might have a point about risk, but community title associations could buy insurance – rather than being *insured* through the taxpayer.

    Has anyone sued the provider of the road yet? – I thought the driver always had a duty of care to drive to the conditions, familiarity and presence of children etc.

  53. Mark, how does giving roads away deal with the problems? How much do you think insurance might cost for the public liability of ownership of a road on private land that is used by the public and has to be maintained by the private owners? Plus, you can’t retrospectively impose community title on existing freehold suburbs.

  54. If “I’m the Government” I can give away my own property as I please.

    Can’t I???

    If you give it away, you don’t need to sell it to firms.

    IRL I am insured (fairly specialised) for $5 mln of public liability and $5 mln professional indemnity and it only cost me $190 per year. I’m not getting the best price either. I think entire estates could get very good prices and from a bigger pool.

    In some places people pay a lot more in taxes than they’ll ever receive for services. Country areas are usually typical of this.

    Tighter cost control and more competition would see maintenance costs fall.

  55. I know the thread has veered away from the economics of road pricing (I only just checked for responses to my orignial comment).

    But the arithmetic of JC and John is wrong. I tried to summarise it by saying that when you join the road, the marginal cost of what you do (because you make congestion higher for everyone else) is higher than the average cost that you face.

    Optimal road pricing ensures that the last car into the road pays the marginal social cost of that trip, so that trips with less value than that aren’t taken. It does get cars off the road, but the real benefit is that people save a lot of time.

    My point about disregarding those complicating factors (which we all did in are arguments) was that the actual social cost of time lost depends on a lot of random factors, so really the best we can do is work out some kind of expected social cost…

    Anyway, if you’re the 1000th car to join a road, the congestion adds 1000 seconds to your trip. That’s the average congestion cost for every car joining the road. However, because congestion rises (for everyone) with every car, suppose the next car faces a 1001 second congestion delay. Fine, this new car has made everyone pay 1001 sec, or 1 extra second per car. But look at how much time has been lost by that extra car joining the road: 1001 seconds for the new car plus 1 sec for each of the 1000 cars already on the road. That extra car causes time lost on the road to go up by 2001 seconds, but the car only pays 1001 seconds. That’s an externality that causes roads to be more congested than is socially optimal.

    Of course, in considering road pricing, the deadweight losses of government intervention need to be set against the benefits, though if the revenue raised is used to offset other taxes, (a big if, of course), this doesn’t have to be such a big worry, at least in theory…

    Maybe I’m defensive on this because, as an ex-Treasury officer, I have the utmost respect for Ken Henry and his analytical insight on economic issues.

  56. Robert,

    I still want to know how the originator of an externality is also a third party, especially when the information of the “externality” is known.

    I think this needs consideration before the maths is considered.

  57. Robert, that only makes sense if you think that bagsing or first dibs apply to road use. But if I get in the car in the morning I know my trip will be affected by those who left before me and those who leave after. Plus, roads don’t have starts and finishes so time of joining is even less relevant to the claim.

    Mark, picture yourself as an insurance company. A bunch of average people own their street and have to maintain it and keep it safe. They want insurance for their public liability, at least $20mil for any one accident ($50mil might be better). How do you think you’d go assessing that risk?

    How does giving away my street stop me and the neighbours being charged for using it?

  58. Robert.

    What ever time is taken away from you by each driver has to be given back as well. The math is wrong. Not mine, theirs.

  59. pedro,

    1. Very easily. They are specialised to do that.

    2.1 Easements.

    2.2 The fact that residents won’t want to pay and that charging tourists would be very unpopular, at least for local businesses.

    2.3 It doesn’t happen now.

  60. Mark, my point about insurance is that if you have a bunch of people with limited means in charge of maintaining a relatively small section of road then you probably impose a high premium for the high risk.

    Ok, so we will privatise roads and not allow use to be charged? Why are we privatising them again? Is the goal just to distribute the maintenance costs to small organisations poorly equipped to deal with it?

  61. Pedro, as a minarchist, I have no trouble with the road-owner, which should be the local council, setting conditions for the use of the roads, and taxing petrol, and requiring licences. Local should be the only tier, meaning Canberra would turn into a convention center.

  62. Pedro – then the problem is merely an initial problem of working back from a nationalised position to what it would be without nationalisation. The best you can do is to mimic private enterprise. The road sections would become economic parcels.

    I think 2.1-2.3 show why people won’t be charged a toll to visit their friend in the next suburb.

  63. Neither do I Nuke, leaving aside the silly idea of privatising roads, my only beef is with pricing poor people off roads so that rich people get a quicker trip to work.

  64. Mark, 2.1-2.3 are just assertions and are not evidence of anything. Also, the easements idea is wrong, but that is a thing perhaps only lawyers experienced in property would know. I think 2.2 and 2.3 are easily disposed of.

    2.2 I don’t want to pay for nice wine, say, but people charge for it any way.

    2.3 People don’t pay tolls for roads now because the government would lose the next election if they tried it. People do pay to use private facilities and a private road is no different.

    Stop being stubborn. We’re all wrong sometimes.

  65. Pedro,

    Show me a development that has more private management/community title than your usual development. They don’t charge tolls. This is what I meant by it doesn’t happen now.

    Roads aren’t even strictly a public good. They’re just publicly provided in general.

    How can 2.2 happen in a mutual company if nearly everyone is against it?

  66. Mark

    Ask them about their body corporate levies. Mind you, they don’t own the main road, just the internal roads.

    Well, do you think that the mutual company might want to charge through travelers a toll. Kinda like the way shopping centres charge a parking fee for non-shoppers.

    But we have descended to nit picking when the big picture is so much more important.

    If a public road is not a public good then what is? I know that this discussion started from the premise that there are times when road use if affected by others. But at any given time the vast majority of roads are able to be used by X without material prejudice to Y.

  67. jc – “whatever time is taken away has to be given back as well”?? As a throwaway line it sounds clever, what on earth do you mean by it?

    Mark – you’re a third party to the congestion everyone else has caused. But you also originate the externality (on other road-users) by joining a road and making it worse for everyone else. You generate external costs for them. But you are the third party of the external costs that everyone else has imposed – and that’s the information that you know about and act on. You care about the effect of everyone else on you, but you don’t care about the effect you have on everyone else.

    Pedro – the idea of joining a road in order is just a way to simplify the analysis of a very complex problem. But the central point is that your choices to join the traffic are on how the traffic will affect you. The externality is that you don’t care how much worse you make if for everyone else.

  68. Robert, I think that’s an abuse of the term “externality” to the point where it is useless.

    The idea that everyone imposes costs higher on everyone else than what they incur from everyone else and this increases as there are more road users is highly questionable. Take for example another common pool problem: common pastures. It doesn’t make sense with a village of equal livestock ownership or concentrated ownership. I don’t think the maths here doesn’t seem to support the Henry/Clarke/Cotgrove idea. Particularly when you can only drive one car at a time – as opposed to grazing more livestock than your neighbour.

    Even if you get the maths right, you’re only halfway there to showing an externality. People make these decisions in a fully informed manner and expect to incur externalities voluntarily.

    Where ever is the assumption that in the basic examples of say, land pollution or pollution of groundwater near a lake that third parties expect to incur externalities?

    “The externality is that you don’t care how much worse you make if for everyone else.”

    No that’s never right. The externality is the cost you impose on others that has not been paid for explicitly.

    However if you think roads are public goods, your solution of more taxes implies that you think more roads should be built or that people should change their schedules (going back to the common pasture, commoners could either go elsewhere or have better management practices). I think it is easier for Governments to build more roads than to dictate a change in working hours to get an optimum level of congestion.

    That said, some roads are uncongested and others are not. I think this points to a lack of planning or a lack of profit motive, or both.

    Furthermore, the current fuel taxes are in the lower range, but definitely within the range of congestion & pollution charges Harry was suggesting.

    Frankly I think this is a case of privatisation and relaxing development rules. The M7 works a lot more smoothly than the roads it replaced/superseded.

  69. Robert — the “external” cost that you create (small cost * many people) is the same as the “internal” cost you face from the congestion (small cost * many people).

    You could reword this as “by joining the traffic you create an externality and also face an externality”. This is where the current poor state of welfare economics gets in the way. Currently, people think “if externality exists, must have tax or subsidy”.

    But nearly all human action creates externalities, often both negative and positive, so this criteria gets us nowhere (or, more realistically, gets us everywhere, allowing us to justify anything). What matters is if the “net externality” from an action is large and clearly in one direction.

    The net externality from congestion is unlikely to be zero because different people have a different value on time… but it is likely to be near zero and it’s almost impossible to work out what it is.

  70. Nicely put John.

    I think it is silly to think in terms of externalities when people make a decision to join a crowded road. If both people in an elevator fart, neither has a right to complain. Further, taxing this particular “externality” does nothing to compensate people for the cost they allegedly have imposed on them and which they impose on others. The effect of a congestion tax is to reduce congestion for the benefit of those prepared and able to pay the tax. You might think this a good thing, but if so you probably aren’t relatively poor and dependent on your car.

    Mind you, I have no objection to user pays, but we have already established that the fuel tax covers road costs pretty well.

    The only way in which a road tax could be said to compensate for an externality is if the money was used to build sufficient roads for the traffic levels desired by road users. And you can see where that logic leads you on the subject of road taxation.

  71. Just because the average and marginal costs both involve small costs and many people does not mean they’re equal. Next time you’re in a traffic jam, work out if the externality cost is anywhere near zero.

    When I said “you don’t care about the costs you impose on others”, that’s just another way of saying that the person who imposes the cost does not pay explicitly for it. A road tax ensures that the only trips that get taken are ones that are worth more (to the driver) than the marginal cost of them to society.

    As for poor people being dependent on the roads, why provide welfare indirectly by not requiring them to face the true costs of their road use?

    It doesn’t matter if someone chooses to join the road or not – if the congestion puts them off making a trip they would otherwise make, then society has lost out. It doesn’t matter if

    Finally, the presence of an externality doesn’t mean the government has to act, John. The whole point of libertarian ideas is not to claim (as you are) that there is no (significant) market failure in road use, but that government intervention is unlikely to improve social welfare.

  72. I ride a motorbike, so my externality (both given and taken) is quite low. 🙂

    But your point about the existence of gross externalities misses the point. It is the net externality that matters.

    You simply assert that a road tax improves efficiency, but that claim rests on the idea that the road tax will correctly adjust for an imbalance in net externalities. You have not shown the direction or size of that net externality, you have simply shown that a single externality exists. This is the problem with simplistic welfare economics, as I described above. You need to factor in all externalities, running in all directions.

    Your second last paragraph makes my point nicely. You note that some people will see the existence of congestion as a disincentive to join the road. Yes! That’s exactly what I’m saying. The congestion itself is the “tax” that you want. Though why you end that sentence with an assumption of social loss isn’t clear.

    I agree that welfare policy is best done elsewhere.

    I agree that an imperfect market doesn’t necessarily justify intervention.

    (And even then, as Mark says, it’s not clear that “externality” is the right term, because people agree to enter the road knowing the possibility of congestion. This is like entering a bar and finding out your favourite chair is taken… sucks for you, but hardly an externality. The word “non-financial cost” would be more appropriate.)

  73. Some parts of the world simply do not warrant this type of thinking. Consider the beach. There are plenty of times during summer in which a swim between the flags is a nightmare of avoiding the other members of the horde escaping the heat. I guess we could fix that by charging a fee for swimming in the flags, but who really thinks that makes sense for society? The beach is there for everyone and so is the road. We all own them both.

    A tax to ration the use of public resources does not compensate anyone for the externality allegedly caused by other users.

  74. Actually, I should clarify that I’m not against the idea of charging for road use. I am just taking issue with the use of the externality argument.

    Unfortunately, externality argument are often abused. In first year economics people are taught the simple story of “perfect markets are good, but externality makes them bad, so if you see an externality, try to fix it”. The problem is that nearly all actions have multiple externalities, positive and negative, and so a simple-minded externality approach allows you to justify anything.

    When externalities are properly understood, it becomes clear that it is difficult to determine even the direction of the externality, let alone the size.

    There are negative externalities from education. There are positive externalities from noise pollution. There are both positive and negative externalities from walking down the street. The externality argument is perhaps the most abused tool in the economic tool kit… and I think the congestion issue is another example.

    But, as I said, that doesn’t mean road pricing is necessarily wrong. A pub doesn’t introduce a cover charge because of a “congestion externality”… and likewise a road company could legitimately introduce a road charge without resorting to bad economics.

  75. I agree that externality arguments are easy to generate, and can be misused. They’re so hard to precisely measure, there are so many examples throughout society, and in any case the likelihood of the Government being able to more closely align private with social marginal costs and benefits is extremely limited.

    That said, I remain convinced that road congestion is one of the best examples of negative externalities out there. And John, I did actually identify the size and direction of the externality – it’s negative, because we not only pay the average cost but we add to everyone else’s cost too, so the total social cost of our trip is higher than what we pay. Short of writing a model, I don’t know what else I can do on that score.

    By the way, no sensible pub owner would want to levy an congestion charge because a crowded bar generates huge positive externalities – people like going to places that other people go to. Doing so would indeed be bad economics.

  76. Next time you’re in a traffic jam, work out if the externality cost is anywhere near zero.

    But isn’t that the point. It is close to zero if the costs and benefits are properly allocated out.

  77. Pubs etc do charge cover charges, because they create network externalities, not in spite of them.

    Robert – what do you call a bad example of an externality then, and how does it differ quantitatively from road congestion?

Comments are closed.