« Traffic Congestion--Posner's Comment | Main | Response on Health Care-BECKER »



Feed You can follow this conversation by subscribing to the comment feed for this post.


Interesting post on the use of dynamic fees for entering inner cities to reduce road congestion. One thing I thought of while reading this was that the treatment of inner city traffic congestion ignores the very real problem of suburb-to-suburb traffic, and traffic along outer ring "beltways" that surround most cities in the United States. To take an example with my hometown, there is significant traffic in the Minneapolis, MN area generated along highways that connect one suburb to another, because many people work in other suburbs than the ones in which they live, and never enter the inner city but merely drive around it on beltways. Some of the worst traffic in the region I am describing occurs nowhere near the city core. Perhaps Minneapolis is a unique case, but I cannot see why. I'm not sure how to design an efficient variable fee system that would extend 20-25 miles out of the city center to prevent the very real congestion that occurs that far out.


in reading the section on the variable toll rate in london depending on congestion levels, using and electronics device native new yorkers are familiar with in the form of an "EZpass"... i can't help but think of one problem: notification.in order for the increased toll at higher congestion times to work, the driver would need to be aware ahead of time. before the journey even began, i'd say, since if it were in the form of signs posted outside the area... one may wince at the amount, but who would turn back to start the trip over again using a different method?that being said, and i suppose this would be the main point: how often do these levels fluctuate throughout the day? if the toll level was reported in a way that could be ascertained before the trip... would it perhaps change by the time they got there? and do they do so with enough regularity day-to-day that this can all be reasonably estimated ahead of time?and to add to the point in the previous comment about beltways... washington, DC's beltway is the same way. however, i suppose you can apply the same concept to a beltway. even though it's not a "city center" area.


In Stockholm, Sweden, tehre is a congestion charge trial at this very moment. The tax varies so that it's higher during rush hours. So far the results are lower traffic and saved time for commuters. Read more here:


"This tax on time from congestion is a very inefficient tax because it is not paid to anyone, but in effect just throws away time, the most valuable resource that people have."

I don't see how that follows. The value of one's time is not infinite -- people put price tags on both their work time and their leisure time and craft their labor supply curve accordingly.

Are you summarily assuming that the "congestion tax" would be so de minimus that everyone would value their time more than their tax money? I would consider that an unrealistic -- and dangerous -- assumption.

Dan Hill

You mention the Singapore pricing system but seem to have outdated information. It began as a rather crude cordon pricing scheme using pre-paid vouchers but since the late 90's has been fully electronic using a smart card in the vehicle which stores the value and overhead gantries which read the card and deduct the payment from it. The fee varies throughout the day. Unfortunately there is no fee on weekends which makes that by far the most congested time to drive into the central city.

Moves are now afoot to extend the system from a cordon pricing scheme to also levy tolls on major arterial routes.

As another commenter noted, the biggest limitation is informational - how do I know in advance (since this is the only way to change my behaviour) what the fee is going to be on any particular road and any particular time of day?


"Unless the government would do great damage with the revenue collected from congestion tolls..."

The most obvious use for such tolls is to subsidize and vastly improve public transportation. Chicago's CTA raised their price to $2 recently. If I travel downtown on my own, I take the train. If I have 1 or more passengers, I drive because parking is available for $8-$10. Lower the cost and I'll use the CTA more often. Chicago's Metra commuter train is a terrific option, but many people can't use it because there isn't enough parking available at the suburban stops. Add more parking and more people will use it (although parking fees have to be modest).


I think an interesting solution could include electronic systems that would allow the government to charge drivers according to individual characteristics and current traffic conditions. Germany has a GPS based toll system for trucks in place that could potentially be expanded to charge all drivers individual toll rates. This individual toll might depend on the kind of car driven, when and where the driver drives, the current traffic situation, and the income or wealth of the driver, among others. Check out the website at http://www.toll-collect.de/ for more info about the German toll collect system for trucks.

Arun Khanna

It is less expensive (administratively) to charge and bill employers for their employee's driving to work. It is fair since firms decide on location of their offices. It is more efficient since the additional costs of traffic congestion are placed on economic agents who can evaluate the marginal cost of real estate location decisions as part of their capital expenditure decisions.



A couple of problems in your analysis. 1) In competitive labor markets it doesn't matter whether the firm or the worker pays the levy, since a firm that does not compensate an employee for congestion charges will be unable to attract labor. 2) A higher tax will approximately offset a reduction in (non-pecuniary) commute costs as congestion is relieved, so the net cost to a commuter may actually be positive.

Ifti Qurashi

Professor Becker's comments about congestion, are as always, insightful and soundly backed by economic theory. Just to add some points:

It seems remarkable to me that the congestion charge in London has reduced traffic by the amount indicated (20%). Driving, I think, is probably a fairly price inelastic activity. Furthermore given the low rate of the congestion charge, 8 GBP today, and the average income of Londoners it is unlikely that the monetary cost alone would deter many people from driving. Also, some stores have offers whereby if a customer spends over a certain amount they refund the congestion charge fee. Rather than the purely monetary cost, I think that one of major factors why the congestion charge reduced congestion by so much was because it is difficult to pay. There is no easy automated electronic system. Payment has to be done in a store, over the phone or through the internet. This means that people have to remember to pay and conciously choose to do it. And while there are signs reminding commuters that they are entering the congestion zone, there is no bill. That is, the city doesn't remind people to pay explicitly. Failure to pay results in very large and immediate fines (50 to 80 GBP). This fact coupled with the inconvenience of paying (and the chance of forgetting) may play a significant role in detering casual drivers into the city. Automating the system could counter some of this effect, unless prices were raised to offset this.

Just one further point about public transport in London. My personal experience does suggest that the transport situation has improved, but I doubt that much of the revenue from the charge is being diverted towards the public transport system. Using the underground (metro/subway) is incredibly expensive. A single ticket in zone 1 for the underground costs between 1.50 GBP and 3 GBP. A return fair is double. The congestion charge (8 GBP) only applies once a day, after that every trip into and out of the zone is free. It would be interesting to see if the rise in public transport fares (instituted in January 2006) will prompt an increase in traffic.

Bill Churchill

A couple of months ago, evening rush hour traffic moved at a slower pace than usual óbut not as a result primarily of congestion. It was the result of a rather slippery snowstorm. A ìcongestion taxî would have only added insult to injury, as it would necessarily have been metered based on ìtraffic flowî (i.e., aggregate speed).

A congestion tax of sorts already exists in the form of a greater ìdown timeî for rush hour commuters. In fact, you have here a perfect case of a Nash Equilibrium at work: Everyone knows that they are in the same boat, and there is little anyone can do to improve their condition in a game that has essentially ìendedîóat least in the moment in which they are committed to any particular drive. Having already made their commitment, they have also already gauged their (collective) effect on the other players (drivers)óat least indirectly.

We need effective and personally controllable transportation for the economy to work--This is a fact. To compare us to European and Asian countries that have their entire populations in denser concentration--and who have long before addressed the various issues of ìpublic transportationî--is ridiculous. Furthermore, it is ludicrous to say such things as ìthey have a $5 gas tax in wherever (name the European ìutopiaî of your choice), as if another corruption of incentives were even able to work.

The fact is, our nation has run up its credit card too much for us to be able to effectively deal with this or many other major social problemsówe are thus consigned to letting nature take its course, while we put band-aids on the problem, such as adding the occasional extra highway lane here and there at a cost of many billions of dollars. We are now beginning to face the ìMorning After in America.î

I am not a pessimist, but I do believe that it will grow increasingly more difficult to deal with ìcongestion issuesîóespecially since we have decided to cling to the bromides of the religion of ìFree Markets,î while having forgotten how markets actually work (through a genuine availability of choices in the face of alternatives), and what they need to operate efficiently (fluidity of information). What we need instead is old fashioned conservative and liberal honesty. We need to kill the sacred cows.

Ayn Rand and Karl Marx warned us about this outcome decades agoónow we are beginning to live the (Malthusian) dream. After all, you canít get something for nothing and no one escapes injustice when they turn a blind eye to its happening to ìthe man behind the tree.î At times, we have mistakenly opted for an atomistic ìindividualismî that requires us to eat our fellows and think we can get away with itówe canít. At other times, we have decided to make far too many things into ìpublicî goods for the public goodówe canít swing to that extreme either.

The religion of Markets is as bankrupt as the religion of ìCollectivism.î What we need is social sobriety. Our leaders and thinkers have let us down. We need less spin and more actual scienceóthis is as true of economics as it is of anything else. We donít need another Easter Island scenario on a global scale.

Posted by Bill Churchill (02-13-06)

Richard Mason

For one year, I commuted every day from one end of Los Angeles to the other. I was struck by what seemed to be high variability in traffic congestion, even after controlling for time of day and weather. In fact, I did not mind the one-hour mean commute time that much, but I did mind the two-sigma days when it took several hours to get home.

A high standard deviation of travel time, which wrecks schedules and leaves travelers in a state of uncertainty, may cause as much annoyance and economic harm as a high mean value of travel time.

Although time-shifting traffic from peak hours to off-peak hours would presumably be somewhat beneficial on average, traffic jams can occur in Los Angeles at any time of day, perhaps because Angelenos are already trying to time-shift themselves away from peak hours as much as they can.

As several other commenters have observed, if a road pricing scheme is based upon actual congestion (as opposed to an easily knowable proxy variable, such as time of day) the question is how drivers can be informed of congestion levels early enough for them to make rational decisions about their travel. Indeed, if drivers could be provided with congestion information early enough, the road pricing scheme itself might not be necessary.

Perhaps devices like the one from www.trafficgauge.com will work.

Dan C

I agree with others that if the pricing system gets too complicated, commuters will have a hard time planning. I think a simple car tax from 6 AM to noon would capture most of the commuters and may mitigate the impact on shoppers or tourist who may visit later in the day.

However, I would suggest a higher tax on Fridays when roads always seem more crowded. Or in the snowbelt you could announce a snow emergency 24 hours in advance, with increased rates.

I also agree that this plan may not help as much in cities like Los Angeles or the triangle area in North Carolina. They don't have a real central hub.

I do wonder about the long run impact of the tax. Could it lower the value of central property because of the tax or could it raise the value because of less congestion? In Toronto, it seems that housing close to the subway lines goes for a premium.

In Chicago would you see more spur cities with people avoiding the commuter tax? And the congestion goes down in the city but increase in the suburbs.

Bernard Yomtov

Ifti Qurashi makes an interesting point about the London tax. It is not the tax per se that reduces traffic, but rather the relatively large penalty associated with failure to pay it, as well as the inconvenience involved in paying.

Maybe we could take advantage of this by charging not a uniform tax but a random one. Most of the time you drive in for free, but occasionally you are picked out and made to pay a large fee, maybe $500. If your chance of getting nailed were 1%, say, the average fee would be $5, but I bet that such an arrangement wold reduce traffic more than a standard $5 fee.

Jonathan Bydlak

I normally find myself in agreement with Mr. Becker's well-thought out analyses, but this time I find his thoughts to be a little incomplete. A few issues, as I see them, are the following:

First (and relatively minor, in my opinion) is the issue of "fairness." Like sales taxes, the congestion tax that Mr. Becker outlines is unfairly regressive, as people of lower incomes must bear a disproportionate cost relative to their incomes. In the event that the lower classes disproportionately clog America's roads, Becker's congestion tax likely would be a "just" solution to the problem. However, if I had to guess, I'd expect that car pooling decreases with income. As such, no congestion tax could serve realistically as a disincentive to wealthy drivers without first pricing the less fortunate off the road at peak times.

Another consequence may be a decrease in overall roadway safety. Suppose, as Mr. Becker posits, that the majority of people who switch to other modes of transportation choose to ride buses rather than their cars. While it is true that buses are the safest mode of transportation for those actually on the bus (usually under 20 deaths in a given year), would more buses on the road relative to cars be safer for the people who still choose to drive cars? Likewise, if more shoppers wait until after rush hour to go into the city and do their shopping, is a greater number of pedestrians walking along and crossing roads at night more desirable?

Becker also argues that one long-run benefit of a congestion tax would be the relocation of businesses out of major traffic areas (read: cities' downtown areas) into less crowded suburbs. Ignoring the developmental impacts that this could have for many cities (it is already difficult for suburbanites to find an excuse to travel into downtown Detroit, for example), it is not clear to me that such a relocation is as beneficial as Becker assumes. As Posner notes, decentralization of major business/shopping centers could (perhaps) be bad for sales, as consumers would be less likely to stop into nearby stores while shopping for other things. Similarly, if I were working in New York City, I'd like to have all stores open before rush hour (as opposed to being staggered), so I can run my errands before I get into work.

Yet beyond this convenience cost there is further value to having businesses all be in one location: people spend less time on the roads traveling from place to place. Traffic does not just stem from having too many cars on too few roads. It also occurs when the same number of cars are forced to spend an increased amount of time on those road. For this reason, the beneficial reduction of traffic in a few concentrated areas may be far outweighed by the detrimental effect of more traffic in other areas (as cars are forced to be on the road for longer periods of time).

Mr. Becker also states that fees should be higher during times of intense wind, rain, and/or snow. This again seems to be slightly illogical, because although road conditions worsen when the weather is bad, it is also true that these are the times when public transportation and other substitutes are least likely to be able to handle the increased demand. Therefore the substitution of commuters away from cars and into other modes of transportation actually may represent a net efficiency loss.

Mr. Becker also has not dealt with many of the practical concerns that would come with the imposition of such a policy. Would new computer systems be necessary to keep track of peoples' charges? For sure, special devices akin to EZ-Passes most certainly would be required. Would the costs associated with implementation and enforcement be outweighed by a 3 mph. increase in traffic speed?

Who would distribute these electronic devices? If the policy were instituted at the city-level (as Mr. Becker seems to favor), how would cities deal with people traveling from outside the area (how feasible would toll booths be in this situation? Would they not potentially back-up traffic even more?)? If toll booths were not used, would there not then be an incentive for people living within the central city to register / park their cars outside the city's limits?

Then there is the flip-side of the enforcement issue: how would we as consumers possibly be able to prove when/if we were incorrectly charged? I would bet that Mr. Becker is concerned about government accountability in many other contexts, so how does he propose to deal with it in this case?

Then there is the issue of setting the actual tax. How would the government account for short- and long-term construction projects in our pricing scheme? What about spur-of-the-moment accidents?

Would the city be forced to keep track of how long people are on the road? What privacy concerns would such devices create? The enforcing of tolls with EZ-Passes still make many people uncomfortable - do we really now want a device to keep track of when we're traveling, in addition to where?

Some of these points are minor, because they are more philosophical questions that can pertain to any general tax, not just one seeking to limit congestion on roadways. But they still do raise many practical concerns in this context, and it would be worthwhile for Mr. Becker to consider them.

Jonathan Bydlak

Arun Khanna

I think Professor Becker's analysis has missed one aspect of this issue. The cost of longer commute time is borne not just by the employee but rather the employer is likely to bear a bigger part of the cost. The employer bears the costs in terms of lowered employee productivity due to frazzled nerves, sub-optimal early morning start times by employees and employee need for leaving on a set time for their evening commute.
That?s one of the reasons (besides the capital budgeting aspect of office location choice and lowered collection costs for the city) why my earlier post focused on collecting congestion charges from employers.

Devil's Advocate

I am not sure how it affects the analysis, but no one has referred to the local, state, and federal gas taxes. Maybe I am completely off base...

These to me seem like user-fees, on the fed level at least, since by law the gas tax goes to transportation (highway and public) projects. Unless you are off-roading, you are paying for the road when you drive on it thru the gas tax. It seems likely (certain?) that the gas tax does not internalize all the externalities of cars driving.

However, it seems like the posts assume you have have a communter tax using EZ tags, and cordons, and toll booths. Toll booths seem particularly inefficient, since you have to stop or at least slow down at most.

You have to stop for gas, why not pay the commuter tax there? Now of course the 50 cent gas tax is not exactly politically popular, especially among conservatives, but on the fed level at least, it sure looks like a road user fee, which isnt too much different than a "commuter tax."

The more you drive, the more you pay. And the current federal scheme does actually move money from states that use almost all driving (donor states) and subsidize public transportation in others (donee states). So, another punishment for driving which causes congestion and rewarding places without congestion.

As for NYC, I dont think they need a cordon. No one has to drive to lower manhattan, there is a plethora of public transportation. So let them pay the tax in their time if they want to. It's poor cities like my hometown Houston (see our numbers on the Tx Tran Instit annual reports!) where people have no choice but to drive (well they can vote with the feet and move into the darn city instead of living in their bedroom communities in the pine trees and clogging my freeways during rush hour!

(though they are paying thru the gas tax...)

dan C

Many interesting posts.

I don't see the big advantage of the London system. Don't high parking fees have the same impact? While I agree that it is hard to fine tune the parking fees, I assume that collection and enforcement are much easier with parking fees. It may be a blunt tool, but it has the advantage of being cheap to operate. (I of course assume that congested areas have a relative shortage of parking.)

Also London, thanks to the IRA, has a system of cameras that can track cars. Most cities don't have this infrastructure.

The Singapore system seems effecient but I don't know how you get everyone to participate. You need one national card reader system, or how do you deal with people from other tax jurisdictions. How do you catch violaters? What is the cost of enforcement? I'm not sure that a very mobile population like the United States could be forced into compliance without some trouble. As another poster mentioned, the cost of setting up such a system could be nontrivial and what would be the gain. A billion on infrastructure to reduce congestion in some cities by 20-25%.

I can understand a GPS toll system for trucks. Federal regulation could force compliance and sharing of information with local taxing groups. But individual drivers are harder to track. I doubt you could get most Americans to agree to a system where the government knew where their car was at all times.

Dan c

Left this off previous post

As others have mentioned, if you want to have this flexible commuter tax, you must make sure that people can react to the changes. You should have a public transportation system that can effectively deal with extra commuters. (Remember how Amtrak fails to be an alternative form of transportation during difficult weather). If people must deal with a poor public transit system, they may just bite the bullet on the tax or seek an area (for work or shopping) where planning a commute is easier.


For more updated news of the Singapore system, refer to http://www.lta.gov.sg/motoring_matters/motoring_erp.htm.

To touch on some points raised in the discussion, all Singapore motorists know the exact charge for using certain roads and the variations by timings. As all cars in Singapore have to install the electronic unit responsible for collecting the toll, enforcement is not a big issue.

The only improvement I can see to the system is to adjust the payment rate not strictly by timing, but rather, as suggested, by the severity of road congestion at any given time.


Prof Becker as usual has posted an interesting and thought provoking piece. Happily most of the comments have been on point and correctly reasonsed. However, a couple of comments have noted the "regressive" nature of such a system. The argument is that such a system will impose costs on poor people. In areas where such proposals have been floated (such as HOT lanes in Wahsington DC) this incorrect argument has been used with great vigor. It is normally framed as "why should rich people be able to pay to avoid traffic." This argument presupposes that the present system is just. However, under the current system a ppor single mother who does not use the interstate nevertheless pays for it through her sales taxes, property taxes gas taxes etc. Her taxes may be subsidizing a rich lawyer who drives all the time... While this example is obviously somewhat tongue-in-cheek the point is that the current system is not particularly just. A ssytem where users pay for what they use seems more just.

Moreover, if the program were revenue neutral as Becker suggests, the increased revenues could be used to offeset particularly regressive taxes (such as food tax or sales tax). Alternatively it could be used to fund public transportation used by poor people. Now we have the rich paying the poor to stay off the roads. Seems much more palitible if one's concern is regressive taxation.

B Wilson

The Illinois toll road system toyed with the idea of charging trucks more during rush hours to relieve congestion. Upgrades in toll booth traffic configuration and increased use in "IPass" transsponders have brought this on. Not sure if they actually went through with it (yet).

Joe Merchant

If people truly valued their time as their most precious commodity (perhaps this is more of a mature perspective), they would already be time-shifting their commutes, moving their employment and residence closer together, seeking work-from-home arrangements, and other means of reducing or eliminating the average 53 minutes per day they are spending on the road.

My observation of the commuter animal is that they do not value their time - they rather enjoy spending time in their car every day, there's even some sort of social satisfaction at work being marooned in crawling traffic for 30 minutes or more - options exist in life to avoid these situations, yet the throngs choose to participate on a regular basis.

If you want to fix a financial center so that traffic flows freely again, a tax is certainly an effective blunt instrument to achieve that end. Making the tax variable can sharpen the effect somewhat, but as discussed above, if it is too complicated, the majority of commuter-animals won't really "get" it and it will not have all the desired effects.

If anyone has seen the movie Freejack, its depiction of "two sides of Manhattan" could easily be explained by a severly high cost of entry to the "better" side of town.

Technology exists today to allow Orwell's big brother to track our every move, and based on that tracking information we can penalize speeders, stop sign and red light runners, double parkers, tax road users based on miles driven, location those miles are driven, times those miles are driven, season, weather, sign of the zodiac, and any other scheme imaginable. I personally would like for cell phone calls to or from driver only vehicles to be taxed at a rate of several dollars per minute, multiplied by a factor that increases dramatically for erratic, unpredictable, or slower than the flow driving... but I digress. Most people seem to prefer simple schemes, like toll roads, that don't take advantage of extra information which is available to do things like penalize speeding. To a small degree, our democratic society seems to be keeping technology from being applied in "distasteful" fashion to support invasive forms of taxation and traffic law enforcement.

My personal choice has been to locate my residence within a 10 minute commute of my place of employment - which severely restricts my choice of residence for a given location of employment, but does shorten my work day by an hour or so from the norm. My place of employment is in the suburbs - more by luck than design - but this has the effect of keeping me and my family out of the big city traffic mess most of the time - I find it highly annoying to spend 10 minutes covering 1/4 mile when we venture near the interstate / commercial zone on the weekends. I don't need a monetary tax to keep me from adding to the traffic jams of the city - and I expect most of those who participate in those traffic jams would still do so if there were any small anti-congestion tax in-place. Pocket change is not a big stick for behavior modification. If the taxes become large enough that some people simply cannot afford to drive into certain areas - then you are starting down the road to a Freejack style dichotomy in your cities (people on the top, people on the bottom, and noone inbetween), and that doesn't look like a bright future for the majority.

Ahamd Seyf

Two points here. First, when it is suggested that demand for free goods always expands beyong supply, causindg queues, church services are always free, but there is no queue! One can give more examples. Are you not exaggerating the relationship here. Second, the recent research on London charges- the results appeared in Metro, London daily published by the London Underground network about two weeks ago- showed that the speed in the area has in fact declined, which is a little strange. But this was what the report suggested. While this charging system may be fine, but, what about a policy to invest in public transport, hence, killing the need for the use of private cars. Frankly in the UK- where I live- public transport is really bad- expensive, irregular, not clean, and yet government ministers complain why people are not shifting to public transport. Given the current state of the public transport, why should the public shift! To pay more, and to be less comfortable, that can not be rational, can it?


مركز تحميل

The comments to this entry are closed.

Become a Fan

May 2014

Sun Mon Tue Wed Thu Fri Sat
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30 31