Improving Both Transport Efficiency and Equity in Sydney with One Neat Trick

One of the things that attracted me to Sydney is that it is arguably the “most tolled city in the world”.¹ I did my dissertation on road pricing, and so from a research perspective this presents a huge opportunity. Road pricing has been the great white whale of transport economists for a century or more. People in Sydney are already used to paying for roads.² So the hardest step, going from no toll to toll has already been made.³

The topic is in the air again, as New South Wales’ Independent Toll Review: Interim Report came out in March 2024. My initial thoughts were in this post. My submission to that Review, organised per their advice, is repeated in this footnote.⁴ I will comment about how to implement the goals of the Interim Report itself, which I almost entirely agree with, and next steps.

Sydney’s Motorway Network

The problem is an under-regulated monopolist

Adam Smith wrote about The Duty of the Sovereign:

According to the system of natural liberty, the sovereign has only three duties to attend to . . . First, the duty of protecting the society from violence and invasion . . . secondly, the duty of protecting, as far as possible, every member of society from the injustice or oppression of every other member of it . . . and, thirdly, the duty of erecting and maintaining certain public works and certain public institutions, which it can never be for the interest of any individual, or small number of individuals, to erect and maintain; because the profit would never repay the expense to any individual or small number of individuals, though it may frequently do much more than repay it to a great society. - Smith, Wealth of Nations, V(I)(III) (1776)

As early as 1776 Smith recognised that it was a public responsibility to provide roads, that the private sector could never do so. These should be self-liquidating (pay for themselves) though:

The greater part of such public works may easily be so managed as to afford a particular revenue for defraying their own expenses, without bringing any burden upon the general revenue of the society. - Smith, Wealth of Nations, V(I)(III) (1776)

There is no problem charging users for the services they use. But the toll rates must be in public hands:

The tolls for the maintenance of a high road cannot with any safety be made the property of private persons. - Smith, Wealth of Nations, V(I)(III) (1776)

In a series of manoeuvres that would have appalled but not surprised Smith, a set of murky public-private partnerships (PPPs) whose details have not been disclosed to the public has enabled TransUrban to control in some fashion the vast majority of the blue links in the Sydney Motorway Network map shown above, and granted it the ability to raise these tolls by at least 4% per year, more if inflation is higher. The rate of tolls is in private hands, and not directly regulated. This is unwise. All other non-competitive public utilities in Sydney (e.g. water, public transport, energy) have their rates regulated.

This is inequitable (i.e. unfair) according to most definitions of fairness I can think of (Levinson, D., 2010). Some drivers pay more than others for trips of the same distance on the same type of road. People with lower incomes pay a higher share of their income to TransUrban. People in some areas are more likely to have toll roads as their only realistic option for long distance trips, others have choices. Some toll users get their tolls capped, others do not. The prices are higher than the marginal costs, so market allocation principles are not met.

But of course it is also inefficient because many of the roads are under-utilised, while parallel untolled roads are overused. For the usual reasons, discussed in my submission (safety, environment, congestion), we collectively want more traffic on motorways and less on local streets.

Unfortunately negotiating a different outcome with TransUrban given their existing contracts, will take years, require a small fortune in legal bills, and probably be unsatisfactory. As David Hensher notes:

The problem with this model is the rigidity introduced by the base case financial model, the costly process of agreeing changes and the absence of competitive tension in negotiating with an incumbent. There are also concerns around lack of transparency and loss of social legitimacy.

Under the current legal framework, TransUrban holds the cards. There is a more elegant solution, cutting through most of the lawyerly negotiations.

The Neat Trick, Buying TransUrban.

We would like transform this imperfect toll network into something that adds to social welfare rather than just augment many an Australian’s superannuation (retirement) fund.

As of today’s writing, the market value of TransUrban is AU$ 39.62B. (~US$25B at today’s exchange rate).

Who owns it? It’s largely institutions. For instance the UniSuper superannuation fund (full disclosure, my, and most Australian academics’, superannuation fund) holds a large chunk of this, as well as being a major capital investor on motorways for which TransUrban is the lead investor. Unisuper’s Balanced Portfolio alone holds AU$1.4B of TransUrban stock. You can see the TransUrban annual reports here which provide some more disclosure.

About half of that value is their operations in Sydney, according to their annual reports.

That’s a lot of money … but also it isn’t.

Suppose New South Wales (with or without cooperation from other states or the Commonwealth) buys out the company and its partners and sells off the non-local bits, the net cost is about $20B (plus a bit more for the motorways for which TransUrban is only a partial investor). New South Wales also gets to keep the stream of future revenue using whatever tolling strategy it wishes, which can be used to guarantee the bonds issued to finance the takeover.

We can imagine lots of ways of doing this, all requiring some legal authority of course, but nothing that has not been done in the past when acquisition at fair market prices of private assets was thought to serve the public interest — and what’s more fair market than stock valuations.

Paid out of general revenue, over 20 years, $20B at $4000 per Sydneysider (at 5 million Sydneysiders and 0% interest and inflation) is $200 per year, or less than $0.55 per day. Adjust assumptions to get more precise estimates, but this is obviously a feasible policy choice. And that assumes just removing tolls from motorways, which is an unwise outcome.

Instead the bonds required to acquire the currently privately managed motorways should ultimately be compensated by the stream of road pricing revenue from all roads, first currently tolled motorways, then untolled motorways, and finally non-motorways.

The users of roads would pay for the benefits they receive, but tolls would be set at a rate to optimally manage overall congestion (on both motorways and local roads), reduce crash risk across the system, and reduce overall pollution effects on neighbourhoods. And while equity may need to be more fully dealt with in other ways, this improves overall fairness compared to toll rates set by the current unregulated monopolist.

Some users would no doubt pay more than they do now, and others less, this is unavoidable. The higher tolls for some users would only be partially offset by lower fuel and registration taxes. But with more balanced use of the entire road network over space and time-of-day, society as a whole is better off, and wastes less time delayed by traffic. The distribution of costs would be fairer, and the gains from the efficiency can be used to help compensate anyone is otherwise worse off.

Today’s non-motorway users benefit because other people who are now priced off the motorway (including maybe themselves) could use it for a lower price, saving time. Today’s motorway users benefit from a lower toll, at the cost of slightly higher travel time on the motorway (but not so much that avoidable congestion sets in), but still a lower travel time than they would have faced had the motorway not been there. People for whom time-of-day switching is feasible now have the option of lower tolls in the off-peak, benefitting themselves and others with whom they are no longer competing for scarce road space. Less traffic on local roads allows them to be modified to better serve local needs, and be repositioned properly within the Movement and Place framework where they belong.

Now, to be clear, the private sector can solve some of the problems the public sector creates (e.g. public sector road management is too sensitive to public opinion, making implementing systematic road pricing difficult, and so the perceived cost of travel is too low), and the public sector can solve some of the problems the private sector creates (i.e. privately set tolls are too high), but ideally we would like to solve both (Goldilocks tolls which are just right).⁵

Also to be clear, we should not fall into the “sunk costs fallacy.” The tolls should be set to be optimal in the short run, not necessarily to recover enough funds to pay back the costs of building them. The toll revenue collected can be used for that or any purpose, but paying back investors sooner results in tolls that are too high to be efficient.⁶ We can discover the optimal toll through experimentation. While we can estimate this toll through research, in the end, price discovery is an exploratory process.

The relationship q=kv is the fundamental equation of traffic, and above is a representation of the “Fundamental Diagram”. Output Flow (q) in vehicles per hour, increases with link density until a maximum point (maximum throughput or capacity, q*). This occurs at density (k*) in vehicles per km. This is the point at which an additional vehicle begins to lower speed (v). Until this point, speed (v) remains high, the region on the left of the diagram shows where adding traffic has negligible effect on travel speed. To the right of this point k*, density can increase, but throughput does not, and speed drops — congestion sets in. Tolls should be set to ensure density is just below k* and flow is just below q*. At times with high enough demand, lower tolls would result in increased density, decreased speed, and no increase in throughput. Higher tolls waste available capacity. I have drawn this as a trapezoid (more or less), which is typical for freeway sections. Networks as a whole are noisier and the trapezoid is more parabolic, so there is some loss of speed as the system approaches q*, leading to a trade-off between efficiency and reliability (See Geroliminis and Levinson 2009).

From a political economy perspective, just transferring the privatised motorways to state management, with all of the political interference that results, would probably not be progress. I support the direction that the Independent Toll Review is going towards a publicly owned, but independently regulated StateTollCo, which is generally consistent with the Road Enterprise model I have talked about previously. This report is a necessary first step.

While the political interference in state transport departments in Australia has increased since I wrote the Road Enterprise report (published in 2013), the idea of state road enterprises remains valid, and the StateTollCo seems a mechanism to move towards a road utility model.

The new road utility should take over at a minimum both the tolled and the untolled motorways, and ideally all state roads — pricing should not be limited to motorways lest non-motorways be used to skirt tolls — to provide some consistency and fairness, and implement a broader road pricing reform (replacing fuel taxes and registration fees with distance charges and off-peak discounts). Concessionaires, companies who operate and maintain roads, should provide a fee for service, just like today’s bus companies, rather than have pricing power as they do now.⁷ A rump TransUrban might be one of those concessionaires.

1. And toll saturation has set in for many users …

• Hensher, D. A., Ho, C. Q., & Liu, W. (2016). How much is too much for tolled road users: Toll saturation and the implications for car commuting value of travel time savings?. Transportation Research Part A: Policy and Practice, 94, 604-621. https://doi.org/10.1016/j.tra.2016.10.012

2. Sure the rates are too high and no one is happy about tolls and the main operator TransUrban neither shares data nor supports independent research, but as an academic I could see the potential.

3. This did not come for free, and in a world of complete information, investors may not have financed the roads, as the forecasts were far off reality. A 2010 review found actual traffic 45% lower than forecast. Li, Z., & Hensher, D. A. (2010). Toll roads in Australia: an overview of characteristics and accuracy of demand forecasts. Transport Reviews, 30(5), 541-569. https://doi.org/10.1080/01441640903211173

4. My original submission to the Independent Toll Review:

Introduction

(A1) The State of New South Wales has established a new toll road review, which will examine the patchwork of tolling rates on different tolled motorways. This is an excellent effort. Road pricing represents the single most significant step society could make towards a more efficient, sustainable, and accessible transport system. It is also one of the most challenging to implement politically. Charging a fee for the use of roads during peak hours has the potential to dramatically improve mobility and reliability and reduce congestion and pollution, benefiting both individual travellers and society as a whole.

(A1,C8) Road pricing introduces a price signal that reflects the actual demand for road space. By doing so, it encourages drivers to shift their travel to off- peak hours, use alternative modes of travel (public transport, walk, bike), telecommute, shop online, or travel less altogether. This, in turn, leads to a more efficient and sustainable use of the road network, benefiting both individual travellers and society as a whole.

(A4, C2) For example, [with time-of-day pricing] if a driver travels during peak hours, they will pay a higher fee than if they were to travel during off-peak hours. This provides an incentive for drivers to adjust their travel patterns, reducing peak period congestion and improving mobility.

Use of Toll Revenue

(B4) Road pricing provides a source of revenue that can be reinvested in transport infrastructure, maintenance, and public transport systems, further improving the overall accessibility and sustainability of the transport system. By introducing a market-based solution to congestion, road pricing has the potential to ensure that the benefits of these new technologies are realised, and that the transport system remains accessible, efficient, and sustainable for all.

(B4) Tolls on existing users should not be used to widen roads for new users, as that has adverse effects of increasing automobile use in an era when public policy says we want to reduce it for a variety of reasons, including environmental sustainability and community liveability.

Toll Rates

(G) The existing tolled motorways are under-utilised because the tolls are too high, and as a consequence local roads are over-used, compared to a social optimum. The tolls are too high because the toll-road operator, through a series of opaque contracts, is allowed to operate in a profit-maximising way, rather than required to price in a welfare-maximising way, more in line with other public utilities. This is well established in economics and should not be a surprise (Zhang and Levinson 2009)

(C1-C4,F) Tolls should be set on a consistent basis, system-wide. There should be a small access fee, and variable distance charge, with time-of-day discounts for off-peak periods. (C4) Work in Minnesota on High-Occupancy Toll lanes (Janson and Levinson 2014) suggests that real-time congestion varying tolls (rather than time-varying tolls based on typical congestion) may have perverse effects, as the higher price acts as a congestion signal that may attract people to the motorways.

Road-Space Reallocation

However, tolls cannot just be lowered on motorways without some consideration of local roads. Lower tolls will induce demand for additional travel by automobile, with all of the concomitant negative externalities. Instead, the lower tolls on motorways should be seen as an opportunity for road-space reallocation on major arterial roads, like Parramatta Road, Victoria Road, Princes Highway, etc. that parallel these high capacity facilities. Reallocating space towards transit lanes and streetscaping and wider footpaths, will help restore these roads to their historic function as serving local communities rather than for high volumes of through traffic. Along with this, speed limits should be lowered on local roads, to help encourage travellers to take motorways.

Heavy Vehicles

(D, C7) Long Distance Trucks in particular do not belong on local roads, but high tolls on motorways, and the absence of similar charges off-motorways, creates incentives to avoid the motorways. Lower tolls on motorways, along with regulations requiring long-distance trucks to use motorways when able to, would help improve the safety and air quality of local communities, the safety of trucks, and traffic flow in general. Heavier vehicles should pay higher tolls due both to pavement damage as well as higher overall emissions. This applies to heavy cars as well as trucks.

Public transport

(E) Buses, because of the social benefits they provide to all other transport users, should not be tolled, but should instead be subsidised and prioritised, with lanes converted to bus lanes as needed on toll roads at no cost to the bus operators.

Electric and Autonomous Vehicles

(C8) Without effective demand management strategies, the deployment of electric (EVs) and especially autonomous vehicles (AVs) will lead to even more congestion and reduced mobility. EVs don’t pay fuel taxes. AVs won’t have to pay for parking, and in the worst of all possible world, in the absence of road charges, would just drive around the block empty waiting for their owner to finish their business.

The previous NSW administration passed a law that said that EVs will have to pay an odometer tax by 2027, or once they reach 30% market share of new vehicles, whichever comes first. This is a good policy, which introduces road pricing one vehicle at a time.[] As EVs eventually come to dominate the market (while Australia is lagging other countries, there is no reason not to expect this to happen over the next two decades), more and more vehicles will be paying distance-based road tax, without the major controversy that would arise trying to do this all at once, on every vehicle. Enhancing this with an off-peak discount, to encourage more road use in the off-peak, and less in the peak, is also good policy.

This needs to be extended to AVs which are now coming online overseas, and will eventually make their way to Australia.

Data (B1,H1, H2)

Setting tolls in the absence of good data is tricky. Toll operators should be required to publicly share spatially and temporally detailed data on toll road utilisation and rates, including by type of vehicle. Every counting station, every 30-seconds, we should know the number and types of vehicles passing, and their speed.

Experiment (B1,H1, H2)

However in addition to that, the public should fund (compensate the toll road operators for) a series of widely publicised toll rate adjustments, which can be treated by rate-setters as experiments, setting lower tolls, or even zero tolls, on selected toll roads (or the system as a whole) for a period of 8 weeks, so that the elasticity of demand on the toll roads with respect to toll rates can be accurately ascertained. We need a longer window (at least 8 weeks) to allow traffic to adapt to a new equilibrium. These “toll holidays” will produce valuable information for future rate setting.

References

Janson, M., & Levinson, D. (2014). HOT or not: Driver elasticity to price on the MnPASS HOT lanes. Research in Transportation Economics, 44, 21-32.

Levinson, D. M. (2002). Financing Transportation Networks. In Financing Transportation Networks. Edward Elgar Publishing.

Levinson, D. (2010). Equity effects of road pricing: A review. Transport Reviews, 30(1), 33-57.

Zhang, L., & Levinson, D. (2009). The economics of road network ownership: an agent-based approach. International Journal of Sustainable Transportation, 3(5-6), 339-359.

5. Are tolls too high: Are we better off with profit maximising tolls or no tolls?

[From my 2023 post]

“While the tolls are expensive, they are only on selected facilities. As a result, many drivers avoid the motorways. So local roads are more congested than they should be, and the motorways are largely underutilised, especially compared to a social optimum, though I suspect they are also underutilised compared to the profit maximising case — one normally has to assume a toll road operator would know how to profit-maximise, I don’t think that’s the case here, if they lowered tolls, they would attract more traffic, perhaps enough to make up for it immediately, but definitely over the long run as people adapted to the use of toll roads. As it is now, they are avoided like the plague and often empty. This is not good for anyone except the wealthiest drivers.”

Tolled motorways are already effectively express lanes compared to local streets, a higher price for a higher speed. But if they could price discriminate further, and somehow “sell” their excess capacity for a lower price without losing existing customers they might do better.  This includes lower off-peak tolls, or perhaps having two classes of lanes (express and regular) or both. Paris Metro pricing (first class fares were twice second class fares, so wealthier people could buy more space on the Metro, is an analog for this.)

However we don’t want to get too complex about this, as customers care about transaction costs, including mental transaction costs, when making these decisions. Ideally this is not the only choice {no tolls, profit maximising tolls} and we can find the right tolls. See

• Levinson, D., & Odlyzko, A. (2008). Too expensive to meter: The influence of transaction costs in transportation and communication. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 366(1872), 2033-2046. https://doi.org/10.1098/rsta.2008.0022.

• Levinson, D. (2017). The political economy of private roads. In Street Smart: Competition, Entrepreneurship, and the Future of Roads (2nd Edition) (Roth, G, editor) (pp. 79-96). Routledge.

6. There are theoretical assumptions, like no scale economies, under which the toll required to pay back investors and optimal welfare maximising tolls would be identical, but they don’t generally hold, and self-financing would generally require tolls that are too high.

• Verhoef, E. T., & Mohring, H. (2009). Self-financing roads. International Journal of Sustainable Transportation, 3(5-6), 293-311. https://doi.org/10.1080/15568310802259940.

• Zettel, R., & Carll, R. (1964). The Basic Theory of Efficiency Tolls: The Tolled. The Tolled-Off, and the Untolled. Highway Research Board Record, 47. https://onlinepubs.trb.org/Onlinepubs/hrr/1964/47/47-004.pdf

• Geroliminis, N., & Levinson, D. M. (2009). Cordon pricing consistent with the physics of overcrowding. In Transportation and Traffic Theory 2009: Golden Jubilee: Papers selected for presentation at ISTTT18, a peer reviewed series since 1959 (pp. 219-240). Boston, MA: Springer US. https://doi.org/10.1007/978-1-4419-0820-9_11

7. It is probably also worth discussing Public Private Partnerships. I’ll just say I have become increasingly skeptical of thinking about “partnerships” rather than “contracts” or “concessions”. Partnerships imply sharing of risk and reward. The way many of these are structured privatises the benefit and socialises the cost and risk. And the problem with contracts and concessions rather than public ownership is that so much that is unanticipated must be specified in advance to avoid problems, and the legal costs of negotiating this, and then the transaction costs of implementation is just inelegant compared to a cleaner (public or private) ownership structure. I’ll note the attached for at least a framework for thinking about this.

• Levinson, David, Garcia, Reinaldo, and Carlson, Kathy (2006) A Framework for Assessing Public Private Partnerships. In Institutions and Regulatory Reform in Transportation (ed. Piet Rietveld and Roger Stough) Edward Elgar Publishers.