Saturday 21 April 2018

New York's latest interest in congestion pricing

Some might say it took long enough, but for New York Governor Andrew Cuomo it might just be in the nick of time as he is reported in the New York Times as saying that "Congestion pricing is an idea whose time has come".

Well I'd say its time had come years ago, after all the technology to do it has been proven since the 1990s, and the actual results of implementing it have been seen in Singapore, London, Stockholm and Gothenburg (although London's success has not been sustained for multiple reasons and Gothenburg's scheme is too big and was focused on revenue collection).

Few need to be reminded of the severity of New York's congestion, and also the enormous funding problem for the city's road and public transport networks.  Although much could be written on how it could improve its planning, contracting and management of its road network, in terms of maintenance (i.e. apply world class asset manegement systems, introduce predictive maintenance and fund it on an asset lifecycle), let alone its public transport system, and why it appears to cost more than anywhere else.  

It remains that there is a huge backlog in renewals for both the roads and the subway network. 

Previous plans

I've written before about the two past attempts to introduce either congestion charging or toll reforms that would have had some of the same effects.

New York has been here before, with the first attempt to introduce congestion charging seen by Mayor Bloomberg.  It failed because the state legislature was uninterested in granting him the powers to charge on state highways.  In 2012, a different proposal was floated by former senior NYC DOT official Sam Schwartz, essentially introducing tolls on untolled East River crossings, but reducing tolls on outer New York tolled routes, which would generate new revenue and be more equitable.

However, the whole idea of charging fell dead in 2013, with Mayoral candidates uninterested.

In 2014, the Sam Schwartz plan became known as the Move NY proposal, as a group of interested organisations backed it.   However, it didn't seem to get traction, as Mayor Bill de Blasio was lukewarm and at the time, state politicians were similarly so.

Indeed, they are all summarised well in the report discussed below, as seen here:

Previous New York congestion pricing proposals


In January 2018, another plan was released from the Fix NYC Advisory Panel, set up by the Governor.   Full report here (PDF).  See the bottom of this post for some thoughts on some of the "information" contained within it.

Proposed New York congestion pricing area

It recommends a series of measures in three phases. 

Wednesday 18 April 2018

Update on Germany's heavy (and light) vehicle charging systems

On 1 January 2005, Germany introduced the world's first full GNSS based road user charging system, applying only to trucks with a Gross Vehicle Weight (GVW) of 12 tonnes and above. Unlike the Swiss system (LSVA), which used GNSS technology to supplement the tachograph (primarily for verification and compliance purposes) as a means of measuring distance, the LKW-Maut (as the German system is called) uses an On Board Unit (OBU) with GNSS technology to map match to measure distance for chargeable purposes. A consortium known as Toll Collect won the PPP contract to finance, design, build and operate the system. 

From motorways to all highways

The system has expanded from the original network of around 12,500km of motorways by around 2,300km in recent years (mainly by adding segments of Federal highways that were used by trucks diverting from motorways to avoid charges), but will be expanded again on 1 July 2018 to ALL Federal highways. The charged network will then consist of over 40,000km of roads (see below)
Map of all Germany Federal Highways and Motorways
There are further discussions about expanding the scope of the system to include buses and coaches, and all vehicles with a GVW of at least 3.5 tonnes (which is the widely used definition of "heavy vehicles" in the European Union).   I expect that once the system has expanded to Federal Highways, there will be a need to further expand the chargeable network to include regional and local roads.

As part of the expansion to the whole Federal highway network, the process of charging trucks is to change.  At present, OBUs measure distance and calculate the charge applicable for the vehicle, then transmitting the charging data to the back office to bill the operator.  Charging rate table data is transmitted to each OBU as it is updated, so that all trip data is retained in the OBU.  All that is transmitted is data that a truck with a specific identity travelled X number of kilometres on charged highways over a set period of dates.   This is to change.  All trip data is to be transmitted to the back office, with the calculation of charges undertaken centrally, rather than on each OBU.

It is explained as follows by Toll Collect:

This is necessary because the very large route network will give rise to changes much more frequently in the future, especially on the federal trunk road – e.g. construction sites or traffic blockades. Therefore, route information that determines the toll calculations must be adjusted much quicker than before.

In other words, with some routes likely to be closed at some times (and trucks diverted onto alternative, longer routes), it will be easier to apply exceptions to the current system by calculating charges centrally, than by applying such temporary changes through updates to over 1.1 million vehicle OBUs. 

In the early days of discussions about GNSS OBU technology, options were discussed as to whether OBUs would be a "thick" client or a "thin" client.

A thick client would have all processing undertaken in the OBU, which would then only transmit the calculated charge data to the back office for billing purposes.  There are advantages in reducing communications costs, higher protection of privacy (due to a lack of trip data being transmitted) and reducing the risk of the system not being functional due to breakdown of the centralised system (as OBUs could record trips and store charging data until the back office is ready to receive it).  However, the main disadvantage is the need for updates of charging table and maps needing to be transmitted to all OBUs, which risks updates not being made in a timely manner or not being able to be made swiftly in the event of the need for changes due to roadworks or accidents.

A thin client simply measures the chargeable events and transmits that data to the back office, which collects the data and uses the trip information against maps with charge rates to calculate the charges for each OBU.  The advantages are that the OBU can be less complex and cheaper, with it being much easier to amend change maps and rates for charging, including discounts and exemptions by location and time of day.  Unsurprisingly, the key disadvantage is the need for regular communication of charging data and that transmission of such data may be seen as tracking movements of vehicles.

Alongside this change, new enforcement infrastructure is to be installed on Federal highways.  600 poles will be installed with enforcement equipment (four metres high) to classify vehicles, read number plates and check that vehicles subject to the charge have operating OBUs (or prepaid trip passes).

Use of revenues

The net revenues from the system were originally used to fund maintenance on the motorway network and also contribute towards railway and inland waterway infrastructure improvements, but this was changed in 2011 to be fully hypothecated for Federal motorways and highways. €4.63b in gross revenue was raised from the system in 2016, with an estimate €4.66b in 2017.  About €1.45b is spent on managing and operating the system, and "harmonisation measures", and another €150m is diverted to the Federal Ministry of Finance to makeup for a "shortfall in vehicle tax revenue".  That leaves €3.2b net revenue for expenditure on the Federal motorways and highways per annum.  All net revenues are spent by a dedicated agency, VIFG (Verkehrsinfrastrukturfinanzierungsgesellschaft).  VIFG - the transport infrastructure finance company - is a Federal Government owned company which was set up to implement the heavy vehicle charge and to use the funds collected, as well as manage public-private partnerships.  It now has the function of managing the payment transactions for all expenditure on the construction, maintenance and operation of all Federal highways (and motorways) since 2016.  A total of €7.7b was spent on Germany Federal highways in 2017, so the LKW-Maut contributed to just over 40% of that spending.

€5b in revenue is estimated to be collected in 2018, but expected to reach just under €8b by 2022, due to the expansion of the charged network and growth in freight traffic.

The current tariff rates are shown here.

Change of operator and temporary nationalisation

The Federal Government is also to nationalise the encumbent operator (Toll Collect) of the country's heavy vehicle charging system (LKW-Maut) in order to resolve a 12 year old dispute as the single supplier PPP contract ends in August 2018. The Federal Government has been seeking compensation for lost revenues (worth €7 billion) due to delays in the system commencing, and has been in arbitration since 2006. Furthermore, there are concerns that as Toll Collect is the sole provider of accounts in Germany (with 1.1 million vehicles with registered accounts), that any transfer to a replacement operator could be complex if Toll Collect is not selected for the next contract.  The plan is that on 1 September (the day after the contract expires), the Toll Collect LKW-Maut business will be nationalised, and will then be transferred to the new contractor in due course.

Tenders were called for a new operator from that date with four consortia shortlisted:

• Autostrade, which had won the bid for the cancelled French Ecotaxe system (which would have charged trucks on the national state highway network which competes with France's extensive network of private/commercial toll motorways)

• Skytoll. Operator of Slovakia’s GNSS heavy vehicle charging system.

• Toll Collect (T-Systems, Daimler), the encumbent operator (as a new operator).

• IBM/Continental/Abertis, a new consortium.

Whatever operator is selected should expect to share the market in Germany with new European Electronic Toll Service (EETS) operators.

Further expansion and charges for light vehicles

The German Federal Government is also examining expanding the scope of the charge to include buses and coaches, and all vehicles down to 3.5 tonne GVW. 

In parallel, Germany is developing the PKW-Maut, a time based charge for car use of ALL public roads. It has yet to be implemented due to legal challenges (including concerns from neighbouring countries about discrimination), but when implemented will apply a prepaid charge for access to those roads. Vehicle owners will be required to purchase one year, two months or 10 days of access. Prices will be dependent on a vehicle's emissions rating, ranging from €2.50 for 10 days to €130 for one year.  In parallel with its implementation, the Federal Government plans to reduce annual registration fees by approximately the same proportion.   I wrote extensively about the proposal two years ago.    It now appears this system is to be delayed until 2020 according to Berliner Morganpost (German) having previously been promised for 2019.  The light vehicle charge (known as the "car toll") is expected to raise €481m in revenue in 2020 (almost entirely from foreign users, given German car owners are expected to receive offsetting reductions in ownership based taxes).

Monday 9 April 2018

Congestion pricing for Delhi?

Delhi's congestion is chronic, as incomes have risen, car ownership has risen and by no means has Delhi been able to increase road capacity to meet this demand. Neither TomTom nor Inrix have data for congestion in Delhi, but the Centre for Science and Environment in India reports:

Average traffic speed on 13 arterial roads 50-60 per cent lower than their design speed and 35-48 per cent lower than the regulated speed of 40-50 km/hour 
No non-peak hour now on main arterial roads -- virtually no difference in time taken to travel between peak and non-peak hours 

The Hindustan Times claims that Delhi needs to replicate the success of Singapore in charging for road use.  This is a fair assessment, although I would caution simply thinking that what Singapore did could be replicated in Delhi.  For a start, Delhi needs to have a reliable motor vehicle registration database correlated to number plates for enforcement purposes.  It isn't clear that it does.  Without that, there simply cannot be a congestion pricing system at all. Secondly, there would need to be effort made to ensure that basic steps are made to ensure alternative modes are able to function appropriately.  That means making it easier to walk and cycle, as well as bus priority measures which are rigorously enforced.  Finally, there is a need to ensure that the solution for Delhi is implemented incrementally.  Don't try to replicate any other city's ideas wholesale, but look at charging a handful of locations initially at the most seriously congested periods, to see what the results are.

Yet Delhi has a network of toll roads, which use manual as well as electronic tolling.  The obvious first step ought to be conversion of all existing toll roads in Delhi to fully electronic free flow operation.  India has already mandated the National Electronic Toll Collection programme for national highway, so this provides a good starting point.  That's not to mean that the same system should be used.   There are sound reasons for thinking about competitive service delivery in terms of accounts, but if Delhi wants to move quickly, its existing toll roads provide a starting point to test charging without any barriers, and to also test charging higher at congested periods.

Delhi has tried demand management of road use

The odd-even number plate test in January 2016 was a success according to The News Minute.  This pilot meant that from 1 January-15 January, from 0800-2000 only odd-numbered cars could drive 

Delhi's Odd-Even number plate tria

Indian Express reports speeds went up 5.4%, particulate pollution decreased by a relative 10-13% on average, with notable results either side of the period the test was in force.  In other words, pollution declined after 0800, and increased again after midnight (because a new rule on truck traffic restricts much of that traffic to the midnight-0800 period).  

An odd-even policy isn't a very good idea though, primarily because it rewards those who can afford two cars and can encourage those with one car to buy a cheap, old, more highly polluting vehicle to avoid the restriction.  However, the policy did prove that influencing traffic demand in Delhi can improve results.  (More details on the results of the test are available here PDF). 

Criticisms need to be addressed

An article in Outlook India by Dinesh Mohan (Honorary Professor at IIT Delhi) basically indicates scepticism that it is the solution for Delhi.

quotes a number of academics to put some doubt about the efficacy of congestion pricing, yet none of these offer any evidence.  He quotes Professor Peter R. Stopher of the University of Sydney as saying “that charging motorists a politically acceptable amount will probably still not make significant impact on overall system congestion, while the potential for serious impacts on the economy become large if the charges are made sufficiently high or the area covered is made sufficiently large.”  Yet this seems to presume that charging means some sort of cordon or area charge.  There has never been a scheme to date that charges congestion on a system-wide basis.  Singapore is the closest, but is still far from being a network charging system.  What IS clear is that there are positive results where charging has been applied appropriately, such as in Singapore and Stockholm.

He then quotes Brian Taylor from UCLA as saying congestion is a "sign of success", which is not what many road users think.  The analogy with restaurants is hardly appropriate when people have choices between thousands of restaurants in a city, and the restaurants that are most successful can either raise prices to increase profits, or expand locations to cope with demand.

Mohan's claim that "there has been little consensus among thinking traffic experts on how to think about urban traffic congestion and how to deal with it. This is why there are very few cities the world that had the courage to experiment with the concept of congestion charging" is quite wrong.  There is widespread consensus that pricing of road use would reduce congestion, the reason pricing has been applied in few cases is a mix of political will, lack of innovation in the development of options for pricing and the inability to communicate to a sceptical and untrusting public that paying to drive on some roads at peak times will improve conditions for them.

Certainly Delhi is different from all other cities that have introduced congestion charging, but the claim that it is very expensive to implement is simply wrong. The costs of Automatic Number Plate Recognition (ANPR) technology have dropped significantly in the past 15 years and improved in reliability.

Delhi needs a congestion management strategy

I don't doubt that there are many other ways to address congestion in Delhi that need to be explored as well.  Enforcement against illegal parking and considerable work to improve the design of existing roads could help ease congestion, along with enforcement of behaviour that promotes congestion (such as poor lane behaviour).  Delhi needs a congestion management strategy, this would consider a wide range of measures including:
  • Design of existing road infrastructure, including traffic signals, roundabouts, intersections, lanes, parking restrictions. 
  • Enforcement of traffic and parking offences.
  • Provision of adequate facilities to ensure walking and cycling can be undertaken safely and easily for short trips.
  • Public transport provision, including priority for public transport.
  • New road capacity where appropriate.
  • Road pricing.
Road pricing should be a part of this, but objectives need to be clear. Delhi still has a significant population that does not drive, but may be expected to want to own a car in the next 10-20 years. Although it's unreasonable to expect this desire in owning a car to be curtailed, that isn't a reason to not ensure that low cost alternatives are not made as reasonably attractive as they can be (walking and cycling), and that corridor space is set aside for bus transit to be developed (as Delhi also grows its underground metro system).  Delhi may also seek to wrestle control of its commuter railway system from Indian Railways, so that money can go into enhancing system capacity, whether by track, signalling or rolling stock (and to try to capture the growth in fare revenue to pay for these improvements).

For road pricing, Delhi should get its road management right more generally, and after converting existing toll roads to free flow tolling, with peak charges to help spread demand, pilot charging on critical parts of the highway network where diversion can be minimised.  Singapore does have an effective approach to reviewing and updating charges, and incrementally Delhi can develop a network of charging, that generates revenue that will be needed to fix intersections, corridors and other parts of the road network that need modernisation.