Recent posts (here, here, and here) have touched on the issue of time-of-use road pricing.  In this post I will briefly examine its merits, before suggesting a strategy for how it might be implemented in a way that overcomes it’s main disadvantages.  The main recommendation is that rather than compelling drivers to participate in time-of-use road pricing schemes, we should instead implement a voluntary scheme – but charge a premium (through higher fuel taxes) for those who opt out.

But first let’s provide some background to time-of-use road pricing for people who missed the earlier posts.  Time-of-use road pricing has two main advantages over fuel taxes, namely:

  • Fuel taxes charge based on the distance that you drive, which provides no incentive to travel at off-peak times.  Because electronic time-of-use road pricing provides an incentive to avoid driving in peak periods, road capacity is used more efficiently (e.g. from peak spreading and car pooling) and congestion is greatly reduced – this in turn mitigates the need for expensive capacity expansions and subsidies for public transport.
  • Fuel taxes tend to penalize low-income households, which own larger, less efficient cars and drive on rougher roads; hence they consume more fuel (and pay more tax) per kilometre.  And as more high-income households buy efficient cars, fuel taxes will need to increase – further exacerbating this inequity.  The people who drive at peak times in urban areas tend to be well-off, so time-of-use road pricing is – on average – reasonably equitable.

Electronic time-of-use road pricing is therefore efficient and equitable.  And schemes have been implemented successfully overseas, with cities such as Singapore and London leading the charge.  But the schemes in these cities are blunt – they simply charge a fixed fee to enter the central city at any time of the day.

Stockholm, however, has a much more nuanced scheme.  Here it is free to drive into the central city before 630am, after which the toll applies.  The toll starts out low but increases in approximately 50 cent increments – reaching a peak of around $2.50 from 730-830am.  The toll then drops back to a residual level (of about $1) through the middle of the day, before rising again for the evening peak.  Finally the toll drops again to zero overnight.  The scheme has been credited with a 25% reduction in peak hour traffic volumes, which has basically eliminated queues.  Obviously Swedish creativity and ingenuity extends to more than just puppet chefs, ABBA, and flat-pack furniture (warning, that last link is possibly the most boring YouTube video ever).

But if time-of-use road pricing is such a winner (and it has been extremely effective in Stockholm) why has it not been implemented in many more cities around the world?  The truth is that time-of-use road pricing has some major (mainly social/political) barriers, namely:

  • Collection costs – in London these run as high as 50%, whereas in Stockholm they run at about 30%.  In either case this cost is much higher than the cost of collecting fuel taxes.  Nonetheless, it’s important to note that even with these high collection costs economic analyses of the Stockholm scheme found it still delivered economic benefits.
  • Big brother – collecting data on people’s vehicle movements does create some interesting privacy issues.  While I think these issues are resolvable (after all, your mobile phone company not only knows where you are but also who you talk to), it is nonetheless a factor that stimulates some opposition to time-of-use road pricing schemes.
  • Boundary effects – This problem arises in two primary forms.  First, by linking the time-of-use toll to a defined area, there is the risk that traffic will avoid the area and increase congestion in other parts of the city (this effect was noted in Stockholm).  Second, certain businesses within the affected cordon may suffer if they rely on customers coming into the city to shop (studies in Stockholm did not find any evidence of this).

Advocates of time-of-use road pricing (and I am one) need to be aware of these issues and strive to resolve them.  My suggested solution is this:  Time-of-use road pricing schemes should be voluntary and not be based on a cordon.  That is, drivers should be encouraged, rather than compelled, to join a time-of-use pricing scheme, and it should not focus solely on the central city (after all, there’s congestion on many roads in peak periods).

Here’s how such a voluntary scheme could work:

  1. NZTA supply GPS tracking devices which are installed in people’s cars.  These devices record time/distance/location, for which users pay differential rates.
  2. Users that sign up to the time-of-use pricing scheme would then be exempt from fuel taxes (there would need to be some verification and/or refund process).
  3. People who did not want to participate in the scheme would remain with the current system.  So fuel taxes would operate alongside the GPS scheme.

A voluntary approach would basically neutralize (I think) the disadvantages noted earlier.  First, collection costs are much lower because you avoid the need to install much of the fixed infrastructure associated with a central cordon. And because people are “signing up” voluntarily you could make them agree to efficient electronic payment (and avoid the need for costly toll booths and call centres).  Second, knotty privacy issues are soothed because the scheme is voluntary; data security remains important, but the act of simply collecting the data is not an issue.  Third, the scheme avoids the need to set-up a physical cordon with a hard boundary, or at least lessened, which allows much greater flexibility for setting charges on roads across the city (after all many roads are busy in the peak – not just the central city).

Encouraging people to switch over is also easy: Just increase the differential between the time-of-use  scheme and fuel taxes.  This will encourage more and more drivers to switch, but always voluntarily.  The voluntary approach is really no different from many internet plans.  Think of it this way: The “fuel tax” plan offers a flat rate, but with high average costs, whereas the “time-of-use” plan offers variable rates, with low average costs – but only if you avoid driving in peak times. People who value simplicity might stick with the fuel tax plan (and what’s wrong with that if they’re prepared to pay the costs of doing so?), whereas people who are keen to take the opportunity to reduce their fuel bill would be likely to switch to the time-of-use scheme.

The main potential downside is that a voluntary scheme would lose some scale.  But then again, because a voluntary scheme is based on GPS trackers (avoiding the need for physical cordons, gantries, and all the associated infrastructure) it could be rolled out nationwide at the same time.  Personally I think a voluntary scheme (with targeted incentives to switch over) is the only way that time-of-use road pricing scheme that is palatable enough to be implemented in the near future.  More controversially, I’d also suggest that a voluntary time-of-use road pricing scheme should be implemented before Auckland even thinks about horrendously expensive capacity road/rail capacity improvements.  That includes the next Waitemata Harbor Crossing AND the City Rail Link.

Who knows how people will respond to better targeted price signals?  They could work at home, car-pool, or move – or maybe they will switch to public transport.  Whatever they do, it’s in our interests to find out before we pour money into expensive tunnels.  It’s also in our interests to “smooth the transition” so that we can monitor people’s behavioural responses and plan accordingly.  With a voluntary approach the transition will be much smoother, allowing the government to tailor their incentives to get the outcomes they want.  We may find that people do switch to rail – in which case the business case for the City Rail Link would be that much stronger and more certain.  Either way, more information and more choice can only be good news – for both the government and individuals.

N.B. My inspiration for voluntary time-of-use road pricing comes from the Dutch “Spitsmijden” experiments, which investigated the effects of offering rewards to drivers who voluntarily avoided travelling at peak times.  I have written a summary of these experiments (which are a wonderful initiative in their own right) that is available here.

Share this


  1. As I typically catch the bus at peak times and drive on the weekends I am pretty sure I’d sign up.

    A couple of issues to think about would be the cost of the GPS trackers and also issues of what you’re trying to achieve revenue wise. If we want this scheme to be revenue neutral some pretty clever maths might need to be done and quite a bit of tweaking may be necessary.

    Overall I like the idea though.

    1. I think GPS trackers are fairly cheap, I’m guessing around $100 but possibly cheaper. It’s just the same basic technology that goes into a Navman/Tom-Tom but without the screen/functionality etc.

      Probably makes sense for the NZTA to procure them – economies of scale, consistent technologies etc, and maybe recoup the costs from the scheme in some way. Not a biggie methinks.

      Revenue wise, yes some tweaking would be required – but that’s part and parcel of delivering a service. One of the benefits of a voluntary scheme is that you sign people up incrementally, so can tweak as you go.

  2. GPS trackers are quite cheap as Stu says.
    I brought this up in a comment on Saturday, but another option rather than use GPS would be to install Average Speed Cameras, and use information collected from those for time of use charging. That means you have a second stream of revenue. I was thinking mostly of a per km charge, and to have it on the motorway system. It might discourage people from taking short journeys on the motorway and on to local roads, but would certainly disincentivise long motorway commutes.

    1. It’s an interesting idea – but would it not be quite expensive? I assume you would rely on ANPR (automatic number plat recognition) technology like London, which seems to be relatively expensive compared to GPS technology. Plus the latter gives you a full road pricing system, rather than just one that serves motorways. But it’s an interesting option that I had not considered …

      1. I actually tried to find an estimate of the cost when writing my original comment, but couldn’t find the answer. However, I assume that like GPS devices the camera and OCR technology will be coming down in price. The other thing that I suggested in my original post was that if it was to be universally applied (rather than voluntary), would be that there could be a threshold below which you would not be billed, so that it wouldn’t confuse or impact occasional visitors to Auckland.
        The other part of my original suggestion, which is possibly easier to implement from a fixed infrastructure system, would be that you could make the pricing dynamic. So as well as, or instead of pricing in just peak times, you could also increase the charges when the roads are busy. What I was suggesting for the motorways is that in addition to the expected travel time signs, you could have a three colour traffic light system to indicate the size of the current per km charge.
        Average Speed Cameras have the additional benefit of hopefully making the road more free flowing, with traffic moving at a more consistent speed, and it could also be used with dynamic speed limits a well.

  3. I like the idea of time-of-use tolling, but I would also like to see the charing to include the vehicle type in the toll calculation. For example, If the base peak charge was, for arguments sake, $2.00 for a less than a 1400cc powered vehicle, then perhaps another charging tier for up to a 2 litre vehicle, and then another charge tier for anything above that.

    1. Completely agree. Higher rates for larger vehicles and/or polluting ones. Just note that the rate would be per kilometre rather than just a flat fee – to discourage long distance trips.

  4. Do you realise that a GPS connected to a computer is the same as a policeman watching your speed in every car with one of those GPSs. Whilst I do not condone speeding I can tell you that in Sydney the speed past a fixed speed camera is 10kph less than what is allowed. Those with one of your congestion busting GPSs and wishing to ensure they are not fined will also do 10kph under. That will do wonders for your congestion!!! Still after a few unintentional fines combined with increased congestion PT will look much better.

    1. Good point Peter – never thought about the opportunities for speed enforcement that would come from a GPS based road pricing scheme.

  5. My thoughts are that we should be looking to move to TOU road pricing but I’m not convinced that the option you present is the best one. A couple of issues I can think of quickly are:
    1. Anything that is installed inside of the car, be it GPS or some other tech has the issue of being tampered with and while there might not be that many that would attempt it, why even go to the risk.
    2. I can see quite a few potential loopholes in having two systems, i.e. I go buy petrol from my local station in one car with a huge petrol tank, I could then siphon off that petrol to another petrol tax paying car and reduced my tax burdon.
    3. Having two types of systems are confusing and difficult to monitor and police which would actually push up collection costs further than just having one new system i.e. if there was some form of rebate then people have to be organised with things like petrol receipts etc.

    My thought would be automatic number plate reader equipment installed at various locations feeding back to a central server. Initially we would target motorway interchanges and the main arterials however if there are minor roads that start to get flooded with traffic then focus can then shift to fixing those problems with either more readers or road works to make things less attractive as a rat run. Doing the readers on a larger scale would obviously cost a bit but I imaging that most of the costs are probably from the software to read the number plates so after that is sorted out it should be pretty cheap and easy to put more cameras in and link them into the system.

    Another thing to think about is admin costs, petrol taxes are obviously easy as there is no way of escaping them where as any alternate road pricing can be harder to administer. One way around this could be to have a system whereby petrol stations have a direct link into the central database and when someone tops up with petrol they plug in the number plate, it tells them how much taxes are owing and that gets added straight onto the price people pay.

    I also think that when we eventually go to a road pricing solution we should still retain some taxes on petrol to cover the environmental impacts of using it as a way to encourage efficiency. That way you get a tax based on how much use the roading resources and one based on how much petrol you use. That way if you buy a more efficient car, or perhaps even an electric one, you can reduce the tax you pay on petrol but still pay your fair share of roading costs.

    1. Just some quick responses before I run off to work:
      1. Tampering with technology is not a problem. Honest – so many countries already use GPS trackers for various tolling situations that I’m sure the systems are secure.
      2. The loopholes would need some careful thought. But I think it’s doable – the refund option is the best. I.e. you pay the same price for fuel (including fuel taxes) up front and then at the end of the month they reconcile the difference between what you’ve paid in fuel taxes and TOU charges.
      3. Two types of systems is not confusing! This would replace RUC – so we would go from 2 systems to … 2 systems! Not so very complicated methinks.

      But your comments have prompted another idea, which may address your concerns: Just move everyone to GPS trackers and offer two pricing plans. One plan would be the same as fuel taxes: They simply record the kilometres you drive, not information on location/time, while the other would offer TOU chages. Voila!

      In terms of environmental taxes on fuel, you could achieve the same effect with a GPS tracker by simply having a base flat rate for driving any distance right?

      1. I’m not sure how many use GPS trackers but many use an automated tolken type system, I remember that Melbourne had this years ago where you basically pass under the gantry and it pings the token in your car which matches it back to an account, if it doesn’t get a response then the next gantries take a photograph and you are charged that way, there was no point in tampering with the device and with more modern technology we can do away with the token part. The issue with a GPS based solution is that the device would have to store your travel data until it could be uploaded so the risk of tampering would be higher. There would also be an issue of how to upload the data to the central system, RUC’s are already considered cumbersome and the ‘mass market’ aren’t reliable enough to ensure they would remember to do it.

        Rebates like you suggest are also extremely messy and would end up with lots of payments going around the place, I can just see the news stories on it now about how much work we are creating and how much taxpayers money we are waisting just to implement it.

        For the two payment plan suggestion, that won’t address the privacy concerns advocates but I can’t see how you can avoid that with TOU pricing. About the only way would be if you limited non TOU vehicles to set routes with manual kiosks along with higher charges to go with that but it would require much more infrastructure spend.

        For the environmental tax, no it couldn’t be achieved the same with a GPS tracker as that doesn’t take into account things like the size of the engine etc. Even if you linked the make and model to the GPS tracker with a set rate for each one, it doesn’t take into account a lot of other things i.e. people with poorly maintained vehicles, people with lead feet, people towing things will use more petrol than others in the same type of car.

        1. Only one country so far, for trucks- Germany.

          A quick browse of Alibaba sees these units at around $50-60US each, I’m sure a bulk deal for the national fleet would bring that down quite a bit. Given that car registration is $300 or so a year, it’s not a great deal. Obviously a back end system would need to be setup. Obviously any road tolling of new motorways can be incorporated into the system as well.

        2. Yes, environmental taxes could be achieved the same with a GPS tracker – because you would still be recording how much petrol was going into each car.

  6. Interested in the comparable costs of Stockholm and London.
    Stockholm is an island city as I understand,
    which may make it easier- and cheaper- in terms of enforcement costs.
    Bridges and tunnels are natural
    choke points, which means they are natural boundaries for an enforcement district.

  7. Fuel tax is so efficient it is hard to beat. Can you quantify how regressive it is? I agree completely that we want to avoid any inefficient system with a lot of physical structures like gantry everywhere. But I still think the main problem in Auckland is the lack of a quality PT alternative to driving for most… we do really need to fund improvements to the infrastructure as we change to pricing people off the roads.

    The real problem is the out of date user pays hypothecation philosophy of the fuel tax, viewing it as a transformation tax and privilege its use to funding alternatives to driving is what is needed. Heart attack time for Joyce, Liberty, and the RTF, I know, but this is where we need to go. Fuel tax as carbon tax, not only as a road builders money pot. Current policy is inconsistent with carbon reduction aims, not that the current government really have any beyond lip service.

    1. Effectiveness is the goal, no tfficiency. The difference is that “efficiency” focuses on cost minimisation, whereas effectiveness focuses on maximising net benefits (i.e. gross benefits minus costs).

      A focus on efficiency means you are (ironically) opting for the cheap solution. In Stockholm, despite the high collection costs the scheme delivered a surplus of economic benefits, even when including costs of collection). So it is effective.

      Road pricing is superior to PT because it tackles the problem at its root, rather than trying to pick winners. PT is one of the winners from road pricing, but not the only one – there are many other options that people should also be exploring before we start providing them with expensive public transport.

      1. The effectiveness versus efficiency distinction is snappy, nice. But while time of use pricing may spread the peaks out you are suggesting that you don’t see AK having a problem with disproportionate auto-de[pendency and under developed alternatives? So you [like Liberty and the Minister] and see peak time congestion as AK’s only problem? Hmmm.

        1. No, I agree that Auckland definitely has “a problem with disproportionate auto-dependency and under developed alternatives.”

          But the important question is one of priorities, i.e. what should we do first to address these issues? I’m suggesting that voluntary TOU pricing should be done BEFORE any major capital projects – road or public transport. Why? Because when faced with more precise costs drivers might respond in a lot of potential ways. Public transport is but one response of many.

          So once we have implemented voluntary TOU we can more clearly see exactly where our capital investment in alternatives to cars should be targeted – not only in terms of modes, but also areas of the city. And you would also start to see what types of land use changes occur when people pay closer to the true costs of driving.

          As an aside, as people voluntarily sign up for TOU pricing and some start to switch to PT, your operating subsidies decline and you free up money to be reinvested elsewhere – such as the capital projects.

  8. Another thought about time of use pricing is the point that admin has made before about how Auckland Transport actually discounts its parking rates for peak time drivers so is currently actively working against this. So Time of Parking Pricing is needed too.

    1. Yes you could potetially achieve similar outcomes through parking prices, although the outcomes are less clear.

      Two main weaknesses with this are:
      1. It would not catch trips that start and end at unpriced parking places, e.g. Home-school-home and commercial vehicles. Plus, councils really nas no ability to influence parking prices outside of the central city, so congestion in other parts of the city would not be affected. So GPS trackers casts the net much wider than parking prices ever could.

      2. Their effectiveness depends on how elastic the demand is, and where it goes once prices are increased. No point council putting up its prices if people just go park down the road at Wilsons. Then you’re just increasing profits for the latter and (perversely) enabling them to invest in more commuter parking. So careful analysis is needed …

  9. Great ideas and discussion Stu & other folks. There are pluses and minuses of any system and ultimately we have to find the most efficient and equitable that won’t just allow those with the most money or the company car to buy priority on the network while imposing congestion and environmental costs on everyone else. All the same, so many of our traffic ‘problems’ should be seen as social problems and therefore potentially more easily and efficiently solved by applying education, appropriate (and full) pricing signals, modal choice & other incentives. Congestion is a social problem, as we know not easily solved by just building wider and longer motorways that just clog up as soon as they’re built. That’s like advocating for publicly funded liposuction to address overeating. Technology offers great advantages for voluntary TOU charging,(in addition or as an alternative to fuel taxes); applying the right incentives and signals and giving people a choice on how they can be most efficient and cost effective with transport mode and timing is a great idea which must underpin any modern transport system.

  10. Just for clarification, this article has a few factual errors.

    It is factually wrong about Singapore’s Electronic Road Pricing system. It is not a simple cordon, it actually is a multiple point charge scheme which has differential pricing at each gantry in each direction, with different times of operation, and pricing varying by increments as small as 5 minutes in some cases. It is quite sophisticated.

    Collection costs in London are not 50%, they are closer to 20%, and in Stockholm are now approaching the same. In Singapore they are significantly lower reflecting a system that has been in operation for 13 years.

    For a useful report that may be more applicable to NZ are the trials Oregon undertook in how to replace fuel tax, they are on the ODOT website.

    Finally, I don’t tire of reminding those who talk of road pricing that NZ HAS got distance based pricing, for which over half a million vehicles are liable – called RUC. Why that isn’t talked about as a platform to move forward is astonishing, and perhaps shows how short sighted and ignorant too many commentators on the road sector are.

    1. Liberty – thanks for the correction re: Singapore’s scheme. Although London’s collection costs are way higher than you suggest (at least according to the data I can find) –

      Stockholm’s are also than you suggest, based on the last information I read (which was a cost benefit analysis undertaken by researchers administrating the scheme. Maybe your numbers don’t include capital depreciation?

      And I’m really confused by your RUC comment. Yes it is distance based, but so are fuel taxes (more or less). That’s the crux of the problem: We don’t want to charge people based on distance, but time-of-day. Apologies if I’ve misinterpreted this part of your comment …

  11. London and Singapore are absurd examples for AK. London had an extremely high mode share for PT in the centre of the CBD which was priced, well over 90% as well as an extremely mature PT network- whatever silly fantasies neoliberals spout about PT making no difference to congestion. Really, how would London’s streets function with all the PT removed, especially the rail on it’s various separate ROW? The charge there was designed to cross-subsidise from those for whom price is of no matter to those for whom it is of great importance: in other words the opposite of the user pays dogma in play by authorities here. Tax the selfish and rich and build PT for all. Is that what you’re proposing then NeoLib? Put in by Red Ken, but interestingly maintained by Boris.

    Well I’m certainly all for building what is needed with money harvested from road users; we should use the NLTF to transform our society and economy not simply build more of what there is: The idea that road taxes [either RUCs or petrol tax] should only be spent on roads is like insisting that all alcohol excise must be spent on building new bars: daft.

    1. Patrick, Singapore and London are not absurd examples. No-one is likening their urban form/PT to Auckland, but merely observing the merits of the technologies that they have used and how it has been implemented.

      Not sure who you’re calling “NeoLib” (I suspect it’s LibertyScott). Personally, I welcome comments from people even if I disagree with their ideological positions.

      Their comments often provide insight that can be used to refine and fine-tune our ideas. And he did, quite rightly, identify some factual errors in my post WRT Singapore. Do you think you’re being a bit harsh?

  12. London and Singapore as models for road pricing for AK? I’m sorry I can’t agree that they’re relevant. I’m not suggesting we don’t discuss them but just pointing out that they that’s not anywhere near apples and apples; not even fruit and meat. Nice disenginuity about neoliberalism, I agree it’s an insult, but the man is happy to so self describe, so what can yo do? And who said anything about not welcoming comments?; it’s good to hear from the brighter end of the ‘business as usual’ crowd, always interesting to see how the world can look so different, and refreshing to not just be in an echo chamber of views. But where there’s nonsense it’s got to be called nonsense. Anyway, you seem obsessed by the tone of my post, and for that I am sorry, as you haven’t addressed its ideas.

Leave a Reply