Yes, yes they are.

The answer depends on your point-of-view. If you are a e-bike user, such as a commuter, recreational traveler, or deliverista, it must be a convenience (otherwise why do it). If you are the consumer of warmed over restaurant meals delivered by e-bikes, you probably agree. But, if you are not, the answer is less clear.

As a pedestrian, I find e-bikes being ridden on the footpaths (or “shared paths”) at a minimum an annoyance, even if the rider is legal and under 16, which in my experience they mostly are.

I find the collection of shared e-bikes abandoned on footpaths even more of an annoyance due to its frequency.

I am not alone in not being thrilled with all these new e-bikes, one only needs to see the Letters to the Editor of the local St. George Leader for examples of e-bikes being ridden in what-to-fore were pedestrian spaces to see the issue writ large.

• Close call with e-bike for Cronulla man recovering from shoulder surgery

• Stabbing victim hospitalised in Bondi Beach after alleged teen machete attack triggered by e-bike argument: The argument is believed to have been sparked by e-bikes.

I see drivers complaining about e-bikes darting in and out-of-traffic. One only needs to read the articles (and comments) in the Daily Telegraph1 or other local media to see that lots of people feel that way. The youth have organised themselves for fun days out, which not everyone appreciates.

• Hoons in death-defying stunts on Sydney motorway - A group of unregistered dirtbike riders have been caught undertaking death-defying stunts on one of Sydney’s busiest freeways, popping wheelies and weaving between motorists.

• Popular Sydney golf course vandalised by e-bikes

I think something needs to be done about the growing number of e-bikes, especially the high powered ones, on the roads and especially footpaths. I think the number of deaths and injuries per km driven on e-bikes puts them in the same arena as automobiles in terms of safety implications, and dismissing this by just looking at the numerator (fatalities), without also considering the denominator (exposure, such as km ridden) is misleading. Unfortunately, we don’t actually know the denominator, though we can surmise it is still much lower, though rising fast.

An obvious, first-best solution that most professionals and bicycle advocates support is to build more bike lanes and lower speed limits on local roads. I agree with this.

I think mandatory training and licensure is probably safer than its absence, and enforcement of traffic rules is even more critical. People with drivers licenses should not be immune from e-bike licensure, driving a car and an e-bike are not the same thing. The NSW government is doing something, perhaps it should do more.

• Regulation on e-bike power is underway:

  Under the standard, e-bikes must have a maximum power output of 250 watts and power assistance must cut out at 25km/hr.

• Age limits are probably coming in too.

• Illegal e-bikes will be seized and crushed.

Another obvious solution is to take parking spaces from cars for the corralled storage of shared bikes off the footpaths, with geofenced enforcement on the part of the bikeshare operators. Sadly, after some 8 years of shared bikes in the City of Sydney, this has yet to happen. E-bikes have popped up in Bayside Council in recent months. No serious effort has been made to corral them in Bayside either.

A third obvious solution is to give the youth some place to act out that doesn’t annoy the oldth. I agree with this, but I believe the whole point of most youth acting out is to annoy the oldth, so good luck with that.

HOWEVER, we have not actually built a comprehensive bike lane network in Sydney. The all-powerful bike lobby has yet to deliver for us. This leaves us in the realm of the second best: in the absence of bike lanes, how do we regulate this new vehicle type for the good of all in the here and now. When cars came in more than 100 years ago, we regulated pedestrian movement in the name of safety, but in reality for the convenience of cars. This greatly reduced pedestrian access. We are at risk of doing the same today, converting pedestrian-only footpaths to shared paths so that fast electrically powered bicycles driven by adults are legal in even more pedestrian spaces, discouraging more people from walking. I hope we can do better.

FYI

• BicycleNSW position

• WalkSydney Position

FIN

Read more

• Levinson, David (2026) Network Origin-Demand Estimation using Percolation (NODE). Findings. [doi]

This paper presents a network origin--destination estimation (NODE) method for trip distribution that applies a percolation-like graph search to allocate trips from origins to destinations. Expanding in cost order from each origin, NODE matches productions to available attractions, depleting destination capacities as they are filled. Multiple origins compete for the same destinations; later arrivals may be diverted to more distant alternatives. The result is an OD matrix spatially constrained by network topology and impedance, without a global gravity function or logit structure. NODE can replicate gravity results in some settings, but departs in cases of destination capacity constraints, network bottlenecks, and heterogeneous acceptance, offering a simple, network-aware alternative to conventional models.

Pointers

• AVs

  The future is messy, the future is a process. If you told 20 year old me we would have robotaxis by now, I would have believed you, I’d might have been surprised it took so long. If you told me they would be driving to suburban shopping centres and not much else had changed in urban form to accommodate them, or in response, I would have been surprised.

  • A Waymo hit a kid who entered the road from behind a tall SUV, kid’s ok Waymo’s Commitment to Transparency and Road Safety [CBS Local News video report (pretty pro-Waymo)] Reportedly Waymo’s Speed was 17 mph before braking, 6 mph at time of impact. But maybe (just spitting out some ideas here), maybe speed limits in front of schools, during school opening and closing times, should be even lower?

  • New York drops plans to legalize robotaxis

  • Hands-free driving systems [By Ford!] confuse drivers, but carmakers push for more - WSJ

• EVs

  • Even after cutting EV incentives, Norway only sold 98 diesel cars in January

• Trains

  • Amtrak is being restructured

  • Australia HSR trundles forward (A new, redacted business case was released)

    • ‘Inherent risks’: Final costs redacted from high-speed rail business case

      University of Sydney professor David Levinson, who specialises in transport engineering, said that if the high-speed rail line was built at the estimated cost of $93 billion, it would be about 1.8 times to 2.3 times more expensive per kilometre than the section of California’s high-speed rail project currently under construction.

      “This is largely driven by the requirement for 115 kilometres of tunnelling, making it one of the most expensive proposed rail projects in the world on a per-kilometre basis,” Levinson said, adding that the proposed route was also one of the smallest markets for high-speed rail globally.

    • High-speed train ticket between Newcastle and Sydney to cost $31 for one-hour journey from 2039

      Professor David Levinson, professor of transport at the University of Sydney, said the project is “very expensive” per kilometre of rail, particularly as 115km of the line would have to run through an underground tunnel.

      “It’s not surprising that it would be expensive as high-speed rail projects go, because a lot of it is mountain tunnelling, and tunnelling is classically more expensive than building at-grade or elevated segments,” Levinson said.

      “The costs are estimated assuming everything goes right, the timelines are estimated assuming everything goes right. They build in a little bit of fudge factor, but very few of these projects come in on time [and on budget].”

      Levinson pointed to the Snowy Hydro project, which has also required extensive tunnel boring, and whose budget blew out from $2bn to $12bn at the last estimate.

      Opposition frontbencher Jonathon Duniam said he supported the idea of high-speed rail but raised concerns over the price tag.

      “It’s a great project, [a] great idea in terms of being able to link the nation up,” he told Sky News. “I think the key question here is while it is important to build these things, it’s also important to be able to pay for them.”

      The government said the project “stacks up” from a cost benefit point of view, with the report determining a benefit cost ratio of 0.8 to 1.2. In comparison, a Sydney council project to build a bike network has a ratio of 2.68 to 1.

      Levinson said benefit cost ratios comes with “optimistic” assertions on jobs and housing opening up, but that it’s lower than it should be.

      “It’s a positive [ratio] in that it’s above zero but it’s supposed to be above one … if you were going to go forward with it, it should be well above one.”

    • Also on 2GB: ‘Shovel ready’ – $90 billion high-speed rail project targets 2028 start. (No link to my live interview (Sad Emoji))

• Bikes

  • Cyclists elated by MetroWay Announcement

    David Levinson is Professor of Transport at the University of Sydney’s School of Civil Engineering. He opined that “MetroWay is presently a bit short on details, but the basic idea is a good one.” It would provide “safe” paths for those who live between stations, and might allow “e-bikes to stay off the Metro trains altogether,” he said.
    He added, however, that it “should have been planned and delivered from the beginning, and should be proposed as part of a more complete network of separated bicycle facilities throughout New South Wales.”

  • Cyclists slam ‘nightmare’ shared path at new Sydney Fish Market

• Pedestrians

  • Sydney’s Worst Speed Hump (mistaken for a Wombat (raised pedestrian) crossing.

• Your Regular Reminder that Cars are Weapons

  • Car crashes in 99 Ranch Market in Westwood

  • Video shows moment man charged with hate crime `allegedly’ rammed synagogue [in Brisbane] … but don’t worry kids, don’t be scared, it’s not terrorism, meth was found in the car. (The perpetrator was from Australia.)

  • Car crash into Chabad headquarters in NYC is investigated as hate crime

  • Graphic Video: Islamic Republic security forces in Iran run over protestors.

  • Essendon North nail salon firebombed [and rammed]

• Crime

  • How Crime Flourishes on Turo (Rental service like Airbnb for cars)

Posts

Publications

People

• Mann, I and Levinson, D. (2026) Access for Appraisal: A Systematic Review of Estimating Transport Benefits via Real-Estate Uplift. Transport Reviews

Transport investments generate benefits that extend well beyond mobility, yet canonical cost–benefit analysis (CBA) remains dominated by travel-time savings. This paper advocates the use of real-estate value uplift as an alternative approach to quantify project benefits, arguing that increased accessibility commands a locational premium. We use a targeted systematic search to identify empirical studies that both quantify the real-estate capitalisation effect and apply it within transport appraisal. Across the fourteen studies identified, the ex post appraisals largely use difference-in-differences for causal inference, yet applicability to ex ante appraisal is ambiguous; the ex ante studies rely on cross-sectional models with limited efforts to address endogeneity. Once capitalisation effects are quantified, appraisal applications are demonstrative, aggregating property value changes while neglecting externalities, supply-side responses, and the validity of model assumptions; many report uplift magnitudes exceeding direct user benefits. We thus outline a framework and research agenda for real-estate-based transport appraisal, incorporating key elements of benefit transfer, the supply and demand of floorspace, agglomeration, hedonic price model design, and value capture. Further theoretical development and empirical work, particularly focused on causality, transfers, and land use supply shocks, are needed to build confidence for broader adoption within CBA practice.

Key Words: Transport Appraisal, Real Estate, Cost-Benefit Analysis, Accessibility, Land Value

That’s the Newcastle to Sydney to Western Sydney Airport High Speed Rail proposal.1

It looks primed to receive another quarter of a billion dollars to move from Business Case to “shovel-ready” within two years, assuming the government is re-elected. The highly redacted business case has been released, and the headline number is a benefit–cost ratio somewhere between 0.8 and 1.2.2

A ratio below 1 means costs exceed benefits. A projected ratio barely above 1 means everything has to go right. Everything rarely goes right.

Tunnels

The line requires about 115 kilometres of tunnelling.

It is not surprising that it would be expensive (about 1.8 times to 2.3 times more expensive per kilometre than the section of California’s high-speed rail project currently under construction) as high-speed rail projects go. A lot of it is mountain tunnelling, and tunnelling is classically more costly than building at-grade or elevated segments. That is before geology, contractor claims, delays, and optimism bias do their work.

The costs are estimated assuming everything goes right. The timelines are estimated assuming everything goes right. They build in a little fudge factor, but very few of these projects come in on time and on budget.

We do not need to speculate wildly. We have recent Australian examples of large underground projects evolving once construction begins. The examples of the failure to fully understand what was going on underground include the recent M6 and Snowy 2 cost blowouts.

If the project were overwhelmingly economically compelling, perhaps that risk would be tolerable. When the benefit–cost ratio struggles to clear 1 under favourable assumptions, prudence would suggest caution.

Scale and Gravity

The obsession with the one-hour Newcastle–Sydney trip is revealing. Is the main market in this region really Newcastle to Sydney? Or is it Greater Sydney to Greater Sydney, with Newcastle as an extension?

We all run gravity models in our heads. Interaction depends on mass and distance. There is simply insufficient economic mass in Newcastle for this to be financially transformative on its own.

Australia has too few mid-sized cities, so Newcastle seems larger than it is and rises in the imagination of New South Wales-based politicians. But high-speed rail globally works best in corridors linking very large metropolitan regions with strong intermediate markets: Tokyo–Osaka, Paris–Lyon, Beijing–Shanghai.

There is a serious risk that in a decade or two, the line just fails economically.3 All transit services have a virtuous circle/vicious circle problem. More riders leads to increased frequency of service, which makes the line more attractive, which leads to more services. But in reverse, it’s the death of transit. This happens more widely than is acknowledged by proponents. And while, to its credit, Australia subsidised fairly frequent train and bus services through the pandemic, ridership still has not recovered four years on, and the amount of farebox subsidy has continued to increase. There is no guarantee those subsidies will remain if economic conditions change, which inevitably they will.

The Alignment Question

The High Speed Rail Authority has already decided on an alignment that skips Hornsby, Epping, and Olympic Park.

Why build a Metro if you are not going to connect it to further destinations? No one actually wants to go only to Central Station, which is far to the east of the region’s population center. People want to go to where the other peope are. Cities are spaces not points.

If for engineering reasons, the line were running under the North Shore say, instead of the current rail alignment, one might at least ask whether stops at Dee Why or Manly would aggregate commuter demand. The current alignment appears to chase the one-hour headline travel time (each stop adds a bit of time for those on-board with deceleration, alighting, boarding, and acceleration) while foregoing providing express commuter services that could serve intermediate markets.

High-speed rail succeeds when it aggregates demand. Skipping nodes shrinks the catchment.

This pattern is not unfamiliar. There is a similar logic visible in Sydney Metro West, focusing on Parramatta to Sydney rather than the places in between.

Note that the current (presumably interim) terminus would be Western Sydney Airport. This is one more attempt “to make fetch happen”. This is one of the seemingly bottomless pit of infrastructure investments to develop Western Sydney, the new Airport ($AU11B), the new Western Sydney Airport Metro (now $AU13B presumably), and so on. More people want to live in the east of Sydney, but cannot afford to because of development regulations driving up prices. Is it so hard to change words on paper that we must build yet another train to the middle of what is now nowhere.

Fast endpoints look good on press releases. Connected networks move people.

Financial Sustainability

Let us conduct a very simple exercise.

The suggested fare is about $31 for a one-hour journey.4

Suppose 22.7 million riders per year paying that fare. The riders are not all going end-to-end, so this revenue estimate I am about to give is hugely optimistic. Also assume no inflation, and recognise the ridership is from 2061 forward, so it will be lower in earlier years. That yields

$31 trip * $22.7 million/trip * 100 years = $70,370,000,000.

The capital cost is at a minimum over $90,000,000,000.

Even in the best of back-of-the-envelope assumptions, which to be clear I do not believe, passengers will not pay the cost of the system. So the question then is: who will?

So this rail cannot be justified purely on farebox recovery. The case rests on time savings, housing development, agglomeration, emissions reductions. But if, after including those benefits, the ratio still struggles to exceed 1, one must ask how much optimism is embedded in the modelling.

The project has no obvious, robust value capture mechanism in place to close the gap. And one has a hard time seeing the project operator recovering the gap from land development.

Opportunity Cost

The north–south Metro to Western Sydney Airport from St Mary’s to the Aerotropolis is estimated at roughly $11 billion. The airport itself is of similar magnitude.

The market capitalisation of Qantas is in that range ($AU15B as of 2026-02-28). Sydney Airport’s market value ($AU23B) is substantially larger.

The government could, in theory, buy Qantas six times over for the price of one HSR line between Newcastle and Western Sydney Airport. This is a way of illustrating scale. We are deploying capital equivalent to acquiring significant national firms, or four international airports.

The Long Horizon

The future is uncertain.

The World of Tomorrow

What technologies will be available in 15 years? In 50? In 100, the anticipated lifespan of this investment?

Aviation will certainly decarbonise further, and smaller craft may be electric. Remote work may alter commuting patterns permanently. Autonomous vehicles will reshape travel behaviour. Automation not just from AI but from robots will change what we do and how we do it, in ways we cannot anticipate.

Big Australia/Small Australia: Demographics and Immigration

Whether this line is successful also depends on whether Australia continues to grow, which depends on whether Australia continues to permit immigrants. The current government has been at best wobbly about their support for continued immigration, with international students caps and the like being bandied about. The leading opposition in the polls at the moment, Pauline Hanson’s One Nation (PHON) is clearly opposed to more immigration from places where people want to immigrate from. Without immigration, population will quickly begin to decline in Australia as it has in most of the world, where births are lower than replacement rate. That implies less need for new housing or new infrastructure. But the flip side is, if it goes forward, the infrastructure’s thirst for demand will be used to justify higher immigration rates.

Boosters and Hustlers

The HSR crowd should be pleased. There is abundant consulting work. There are authorities to staff, designs to refine, business cases to iterate.

Hustlers will hustle, and boosters will boost.

But at some point, arithmetic intrudes.

If the best case yields a benefit–cost ratio that hovers around 1 under generous assumptions, that is not a compelling endorsement. It is a warning that the margin for error is razor thin.

High-speed rail may one day make sense in Australia. But the economics must lead the enthusiasm, not follow it.

Footnotes

• Levinson, David (2026) Seeking causality in transport research. Transportation Research Today. Volume 1, Issue 1, Article 100001. [doi]

Abstract

Transport policy asks causal questions, yet much transport research answers them with associative designs and then uses causal language. This Perspective (i) distinguishes causal, associational, and descriptive claims, (ii) offers a compact set of norms and a checklist for aligning claims with identification, and (iii) illustrates these ideas with examples drawn from pricing, operations, access, network evolution, and evaluation.

Keywords

Causality, Inference, Perspective

• Jantabadi, F., Ermagun, A., and Levinson, D. (2026) Spatial patterns of access-density mismatch reveal infrastructure gaps and strategic opportunities for new housing. npj Sustainable Mobility and Transport. [doi]

The interaction between regional access to opportunities and local density of residences has been central to questions of urban structure. However, real-world dynamics often diverge from idealized economic models, resulting in an “access-density mismatch.” This mismatch occurs when rapid population growth, inefficient transport systems, or uneven infrastructure investments lead to a disconnect between where people live and the access provided by the transport network. Access-density mismatch manifests in two primary forms: (i) high-density areas with lower-than-expected access, and (ii) low-density areas with higher-than-expected access. Here, theoretical explanations and empirical investigations are conducted to: (i) examine the association between local density (gauged by walk access to population within five minutes) and regional access (gauged by transit and automobile access to employment opportunities within thirty minutes), and (ii) identify the access-density mismatch areas and their potential correlates at the census block level across the fifty most populated metropolitan areas in the United States.

Five insights are identified.

• First, a statistically significant positive correlation exists between local density and regional access, with automobile access inducing higher residential density than transit access. A 1% increase in automobile access and transit access is respectively associated with a 0.5% and 0.34% increase in local density at the national level.

• Second, more auto-oriented and less centralized metropolitan areas exhibit a weaker association between transit access and residential density.

• Third, metropolitan areas where transit access-induced density is above the national average almost invariably exhibit a higher share of transit ridership for commuting.

• Fourth, the spatial pattern of automobile access-density mismatch is similar to the spatial pattern of transit access-density mismatch. As one moves away from the central business district, regional access tends to surpass local density. This suggests that the current state of infrastructure can support a higher residential density, and there may be market opportunities for increased residential development.

• Fifth, there is a difference in the magnitude of automobile and transit access-density mismatches. In city peripheries, the mismatch is greater for automobiles, with automobile access often exceeding the local density, leading to infrastructure under-utilization. In contrast, in core urban areas, transit access-density mismatch is more pronounced, with transit infrastructure lagging behind local residential density. This disparity reflects historical infrastructure investments, where automobile networks have expanded faster than transit systems, yet have not correspondingly increased residential density.

• Li, Hoang, Bahman Lahoorpoor, David Levinson (2026) Streetcar and Interurban Deployment in the United States: 1894-1926. Findings [doi]

We compile operating track lengths for US streetcar and interurban systems from the McGraw American Street Railway Investments Directories for 1894–1926, aggregated by state and by metro area, with a national total. The United States network reaches an observed maximum of 80{,}660 km (50,120 miles) in 1918, with a logistic asymptote of 84,208 km, inflection year 1901, and R²=0.89. We report cities and states by extent and timing, and show that most city and state series follow logistic S-curves.

ROADS FOR WHOM: THE PEOPLE OR THE GOVERNMENT; US OR THEM

When we design build streets and roads, as planners and engineers, we think of them as being for the movement of people and goods in general. But this isn’t always the case, traffic calming and low traffic neighbourhoods both aim to prioritise local residents over through traffic, to improve livability and safety by reducing speeds and flows. In contrast, while Baron Haussman’s redesign of Paris and creation of boulevards provided sanitation and light, it was designed in large part to help the authorities keep down the masses through barricade prevention, military mobility, clear sightlines for artillery, and other factors.

THE BORDER

The question of who gets to access any place, and who decides, has evolved. Two hundred years ago there was effectively no border control in the United States, and people could just move to where they want (and whoever was already there would be forceably removed or killed). In 1855 New York State implemented immigration checkpoints at Castle Garden. In 1891, US federal immigration controls were put in place, and Ellis Island opened in 1892. The people who were here decided who could enter the club. But what about people who have entered without permission? With the creation of the Department of Homeland Security and reorganisation of immigration enforcement agencies after 9/11 (in response to terrorism by people who were authorised to be in the US, though a few had overstayed their visas), immigration enforcement not as a border control, but as an interior security function became more prominent. Non-documented immigrants, who were previously tolerated because of their great value to society as a whole, were now the most convenient punching bag of an authoritarian regime aimed at dividing and conquering as a means toward power. So the newly over-funded enforcement agencies rent vehicles at airport, stage in public parks, and stay at public accomodations in the target area, using the same networks of access as the people they target.

THE DEPORTATION PIPELINE AS A TRANSPORT NETWORK

Deportation is not merely a legal status; it is a logistical chain of linked nodes, exploiting tools developed for traffic safety and funding, civilian law enforcement, and mobility:

• Locate: Relying on movement data, license-plate readers, and surveillance.

• Stop: Often initiated through routine administrative or traffic contact.

• Hold: Utilization of local jails and detention facilities.

• Transfer & Remove: The use of vans, buses, and commercial or charter flight networks.

For interior enforcement, digital infrastructure is vital. Data pipelines, such as 287(g) partnerships where local police share booking data with federal agencies, turn routine traffic stops into mobility checkpoints. The “same channels” used for commerce are repurposed for extraction.

MINNEAPOLIS, 2026

The actions of deportation enforcement agencies lead us to shine new light on the question, asking how residents can immediately reconfigure streets to slow down traffic, impose checkpoints, and know who is traveling through the neighbourhood. This is analogous to the walled city.

The two most famous deaths of deportation protestors by ICE agents in Minneapolis took place on streets, one while the victim was in a private vehicle, another while the victim was outside. In a real transport sense, these and other protesters were imposing friction on the enforcement agents.

THE ETHICS OF MOBILITY

Transport systems are by their nature (you can travel here, on the network, and not there, off the network) regulatory systems. They contain within them futher rules of governance. When infrastructure is used for interior enforcement, mobility becomes a civil liberties issue.

The ethical question is who can move through this space without being targeted, tracked, or carried away?

Thesis Title: “Logistics Strategies for Emergency Medical Services“

Lead Supervisor: Dr. Emily Moylan.

Co-Supervisor: Professor David Levinson.

Abstract: This thesis advances Emergency Medical Services (EMS) logistics by developing innovative delivery strategies, optimizing facility locations, and designing adaptive dispatching policies to enhance response efficiency and resource allocation. Recognizing limitations in traditional EMS models, this research explores alternative logistics strategies—including rendezvous and pre-hospital models—that offer greater flexibility in high-demand, resource-constrained environments, laying the groundwork for adaptive solutions in EMS.

A comprehensive literature review identifies and critiques current EMS strategies, revealing critical gaps in facility location planning, dispatch policies, and resource management in real-time. These findings establish a foundation for the research’s models, prioritizing patient outcomes, and advancing EMS logistics to meet different healthcare needs.

To address spatial disparities in EMS access, the thesis employs facility location optimization, determining optimal resource placement for rendezvous delivery strategies. Analyzing geographic and demographic factors, these methods enhance response times by ensuring EMS resources are strategically positioned to maximize accessibility and patient medical outcomes.

Dispatch policies are explored through agent-based simulation, with a focus on rule-based assignment methods that adapt dynamically to resource types, incident severity, and varying performance metrics across diverse resources and demand types. This approach enables a nuanced allocation of resources, tailoring responses to the specific needs of each incident while optimizing overall system performance.

In conclusion, this research offers actionable insights into EMS system design and implementation, with broader implications for emergency logistics and urban health policy.

IN CASE YOU MISSED IT

• Findings Press — Newsletter (There was also a nice shout out to Findings in this CityNerd video by Ray Delahanty)

• TransportLab – Newsletter

ANNOUNCEMENTS

• Consider submitting an extended abstract for the International Symposium on Transport Resilience – (July 15-17, 2026, Lyons). Submissions DUE: February 15, 2026.

POSTS

RESEARCH

GRADUATIONS

MEDIA

Lord Mayor Calls For Cycleway Ramp On Harbour Bridge’s Southern Side

The new facility coming off the Bridge looks nice. See this video by Chris Topher to get a sense of it. A better Cycleway Ramp on the south side has risen in importance. Bikes, and e-bikes especially, coming off the bridge will collide with pedestrians and vehicles. Now, with this new northern ramp, there will be more of them.

North Sydney’s bike network presently looks pretty fragmented, which tends to make it far less effective than a connected network. The plan (link) is better, but there is no reason aside from priorities that it can’t be done in a year or two instead of waiting until 2045. The cost of the whole network ($124M) is on par with the North Sydney pool. ($122M).

BOOKS

While I finish up my next book, you should check out these:

YOUR REGULAR REMINDER, AND A NEW REMINDER

• Vehicle Ramming

  • Protesters try to attack driver after truck speeds through Iran demonstration in Los Angeles

  • Driver crashes into McNamara Terminal at Detroit Metro Airport

  • Officers fire shots as police car is rammed in Melbourne

• False accusations of vehicle ramming are used to justify state-sponsored domestic terrorism.

  • Chris Mims: “WSJ investigation: In the past 6 months ICE agents have fired at vehicles 13 times, leading to: * 8 people shot * 5 of which were U.S. citizens * 2 died * no victims drew a weapon.” Videos Show How ICE Vehicle Stops Can Escalate to Shootings
    A WSJ visual investigation found that the Minneapolis ICE killing is one of 13 incidents where federal immigration agents have used deadly force against civilians in vehicles since July.

  • AlsoJesse Ventura comments on killing of Renee Good.

  • Bluesky thread from expert on vehicle ramming attacks.

  • Ninth ICE shooting into vehicles since September

  • Previous claims that ICE agents were being rammed by vehicles thrown out.

LINKS

• Automation

  • Chinese driverless delivery vans

  • Delivery robots climbing stairs in Pittsburgh

  • Waymo uses Philippine based tele-operators

• Trains

  • Death toll from Spanish high-speed train disaster rises to 45

  • Commuter train crash in Spain

• Architecture

  • Rendering of lots of new buildings in Sydney. Minecraft architecture seems to be ending.

• Micromobility

  • Hundreds injured in Melbourne’s e-scooter trial, new data shows

  • ‘E-bike cowboys’ swarm Sydney northern beaches golf club
    

As always you can follow me on Mastodon, BlueSky, or RSS.

On the one hand, there is now evidence many good people would hide you from the Nazis; on the hand, there are Nazis.

US Federal agents are responsible for two out of three homicides in Minneapolis this year.

• Shaer, Amin, Fielbaum, Andres, and Levinson, David (2026) Detecting Dangerous Driving via Computer Vision: Linking Video-Based Indicators to Road Crashes. Journal of Transportation Safety and Security. [accepted and in press] [doi]

This study demonstrates the potential of combining computer vision with regular traffic cameras for detecting dangerous driving behaviors (DDB). We combine data extracted from 258 h of traffic camera footage across Minnesota with road crash records from 2016–2022. Using computer vision, we identify Dangerous Driving Behavior Indicators (DDBIs), including speeding, short headway, and lane violations—alongside traffic flow, truck counts, and time-to-collision (TTC) metrics. These indicators are analyzed individually and jointly to detect aggressive driving and compound aggressive driving behaviors. An Ordinary Least Squares (OLS) model examines the relationship between DDBIs and the number of instances where TTC falls below two seconds (NTTC2). A Negative Binomial Regression (NBR) model then links NTTC2 to crash frequency, while Structural Equation Modeling (SEM) explores the broader pathways through which behavioral factors contribute to crash risk. Results show that short headway, speeding, and aggressive driving increase NTTC2, which in turn is positively associated with crashes. These findings suggest that video-based behavior detection can support proactive traffic enforcement and crash prevention.

• Gao, Yang, Andres Fielbaum, David Levinson (2026) Roadspace Allocation between Autos, Buses, and Bicycles with Heterogeneous Demand. Transportation Research part A. Volume 205, March 2026, 104884. [doi][share]

The allocation of road space among different transport modes has long been a key issue in urban planning, yet it lacks solid theoretical foundations. This paper investigates the optimal allocation of road space among three transport modes: private vehicles, buses, and bicycles, for overall system performance. The travel time for each mode is determined based on travel speed derived from fundamental diagrams (FDs). Changes in bus travel time are the least sensitive to excessive demand, as the number of buses is only indirectly affected by demand. A mode choice equilibrium framework based on deterministic user equilibrium is proposed to handle cases with and without heterogeneity in passengers’ waiting time thresholds for buses. Analytical and numerical results reveal that the optimal road space allocation strategy depends on the demand level. Without considering passenger heterogeneity, the optimal strategy is a corner solution — allocating all road space to one of the three transport modes. When heterogeneity is considered, low and medium demand levels result in all space being allocated to private vehicles and bicycles, respectively. For high demand levels, the optimal solution is a non-corner solution, where road space is allocated to both buses and bicycles, and the proportion allocated to buses increases as demand rises. The initial road space share for buses significantly influences system performance. Crucially, this induces either a virtuous or vicious cycle that impacts public transport usage. The threshold for this effect is around 0.4, meaning that allocating approximately half of the road space to buses is critical, and this threshold decreases as demand increases. This study highlights the importance of tailoring road space allocation strategies to demand levels to maximize transport efficiency.

Thesis Title: “Programming Traffic: Design of Micro-decisions in Automated and Mixed Transport Systems”

Lead Supervisor: Professor David Levinson.

Co-Supervisor: Dr. Mohsen Ramezani.

Abstract: This dissertation explores microscopic decision by human drivers and builds control frameworks that optimise microscopic actions of autonomous vehicles. Existing studies often treat mandatory lane-changes as absolute, neglecting the possible option of not performing them. As such, we explore the incentive behind the manoeuvre by analysing and modelling the cost for not performing lane-changes for exiting freeways. The developed mathematical model is deterministic and can estimate the exit-missing cost with simple network-level variables. Both simulation and empirical results indicate that exit missing costs are not substantial.

Driving is a two-dimensional task, yet studies have often neglected the interdependence between the two dimensions. This study integrates longitudinal car-following with lateral lane-keeping and discretionary lane-changing behaviours to create a realistic model that replicates human-driven trajectories. Recognising the hierarchical nature of human decision-making, we incorporate a game-theoretic reasoning structure and validate the model using a naturalistic trajectory dataset. Model calibration is conducted heterogeneously through a Bayesian framework. The model is able to replicate macroscopic traffic patterns with high accuracies and can generate smooth human-like trajectories.

Building on the two-dimensional model, we design an optimal controller for autonomous vehicles operating in a lane-free freeway environment. Vehicles are assumed to be communication-free, with each making decisions independently. A game-theoretic formulation ensures that vehicles account for the potential actions of others. The full framework is evaluated at both the microscopic and macroscopic scales. Experiments demonstrate how reasoning depth and vehicle-size heterogeneity influence overall traffic efficiency. The proposed lane-free controller is also compared against a similarly defined lane-based controller.

• Cui, Mengying, Lijie Yu, Zhe Dai, Ang Ji, Keyue Liu, David Levinson (2026) Accessibility Dynamics with High-speed Railway Network Expansion: A Line-specific, Granular Analysis. Travel Behaviour and Society. Volume 43, April 2026, 101239 [doi][share]

HIGHLIGHTS

• Stepwise accessibility analysis by HSR line reveals distinct network maturity phases.

• Line accessibility impact depends more on itself than network sequence.

• HSR improves accessibility and equity in Guanzhong Plain Urban Agglomeration.

• Findings indicate potential accessibility and equity gains from future HSR expansion.

ABSTRACT

This study conducts a line-based accessibility analysis of high-speed railway (HSR) network expansion in the Guanzhong Plain Urban Agglomeration of China based on a multi-modal inter-city network. We propose an improved accessibility indicator that integrates opportunities and transport networks from neighboring regions, overcoming the artificial inequities inherent in traditional methods that restrict destination and route selections to rigid study area boundaries. The results reveal that the HSR network in the study area is still in the growth phase, with accessibility gains following an S-shaped trajectory, and has not yet reached maturity. Future investments are expected to sustain improvements in accessibility and spatial equity. More importantly, the operational sequence of HSR lines is less significant than their individual contributions to accessibility and equity. This implies few, if any, synergies between the HSR lines and emphasizes the importance of strategic line placement over construction order in fostering efficient and equitable regional connections, particularly in emerging HSR networks.

Is there any real value to subscribing to multiple LLM services (Gemini, ChatGPT, Perplexity, Claude , etc)? Or should I focus on one? Which one is best at the $20 or so per month level.

I expected them to favour themselves, but to their credit (or loss), they did not. By design, some AI content follows:

CLAUDE: (FREE VERSION):

AT THE ~$20 PRICE POINT

…

ChatGPT Plus ($20/month) is probably the most well-rounded choice for a single subscription. It has GPT-4, strong plugin ecosystem, image generation with DALL-E, and broad capabilities.

Claude Pro ($20/month) is excellent if you prioritize long conversations, nuanced analysis, or writing assistance. The context window is massive and useful for working with documents.

Perplexity Pro ($20/month) is best if you do a lot of research and want cited, up-to-date information.

Practical recommendation: Start with one (ChatGPT Plus for versatility), use it for a month, then add a second only if you find specific gaps in what you need.

…

PERPLEXITY (FREE VERSION):

…

A practical strategy around $20/month
• Pick one paid plan that matches your main workflow (Perplexity Pro if research‑heavy, ChatGPT/Claude if creation/coding‑heavy, Gemini if you’re deep in Google).[photonpay +4]
• Use the free tiers of the others when you want a second opinion or a different style.[gmelius +2]
• Only add a second paid sub if you repeatedly hit a real limitation (e.g., your main tool struggles with code but is great at research, or vice versa).[dev +2]

GOOGLE GEMINI (FREE VERSION):

WHICH ONE IS BEST AT THE ≈$20/MONTH LEVEL?

Most premium subscriptions are priced around $20 USD per month (ChatGPT Plus, Claude Pro, Perplexity Pro, Gemini Advanced/Pro). The “best” choice is the one that aligns with your most frequent and important tasks.

Here is a breakdown to help you decide on a single focus:

….

CHATGPT (PAID $20/MONTH VERSION):

…

IF YOU FORCE ME TO PICK “BEST AT ~$20”

• One subscription only, general use: ChatGPT Plus. ChatGPT

• One subscription only, heavy web research: Perplexity Pro. Perplexity AI

• One subscription only, lots of writing and revision: Claude Pro. claude.com

…

CONCLUSION

They all sort of said the same thing, which was: overall ChatGPT, for research Perplexity, for writing, and for coding Claude. Which is, I guess, a summary of what the Internet says. Which is what LLMs are best at, synthesising the consensus of what everyone knows. But I guess I won’t be subscribing to Google Gemini at this point. (In any case, it will get built into Apple Siri shortly)

My own ad hoc tests, a bit biased since I’ve been using ChatGPT Plus and Pro for a while is that I prefer ChatGPT despite sometimes high level of frustrations. The free version of Google Gemini is faster, and fine for most common knowledge, but lies (hallucinates) more often. The free version of Claude also has a lying problem and is sufficiently annoying in its attempts to engage that I actually deleted my accounts from Claude. (I don’t code much, so obviously your mileage may vary).

These activities are intended to intimidate the masses and almost assuredly designed to provoke an incident from the unintimidated: once one of these agents is hurt or killed, that will justify a further crackdown, ratcheting towards the imposition of martial law, and an attempt to selectively suspend elections.

The large numbers and violence involved suggest it would be a miracle if such an incident did not occur.

This rise to power has been facilitated by the preemptive compliance of putativelyindependent institutions that have handed over their powers at the mere whiff of a suggestion that the supreme leader wants it. Every reduction in friction makes the next move easier and less costly.

We should have been trained by decades of dystopian science fiction and other media to identify fascist states. Some of the most popular movies of all time hit this very theme. And yet the majority of Americans, while perhaps not approving, just acquiesce to this not-very-slow transformation of their country.

We have been trained by centuries of imperialism to know how it comes back to bite the imperialists. Yet those who oppose “imperialism” when asserting A does it to B, are silent when C does it to D, and raise not a peep when their own country does it to E. Almost as if the imperialism wasn’t the real issue, but which power the opponents are aligned with.

We have been trained by millennia of wars to understand the nature (and disposability) of young males. From time immemorial, old men have treated young men, demographically the most aggressive group, easily organised into a hierarchical command structure like the police or army or militia or gang, as disposable tools in the struggle for power. Anyone above the age of 25 (and probably most of those younger) well understands this dynamic.

And yet here we are, dispatching armed militias of mostly young men into US cities like Minneapolis, where I lived from 1999-2017, to punish the regime’s enemies and deport those with skin color, heritage, and accents different than their own, apparently constitutionally stopped and detained solely for those reasons, arresting or killing those who get in the way.

You all were warned.³

Digging yourself out of a hole is far more difficult than never getting into it in the first place.

1. Secret police: Armed agents exercising state power while obscuring identity and agency in routine public operations ↩︎

2. https://humanrightsfirst.org/the-trump-administration-is-forcibly-disappearing-migrants/ ↩︎

3. See also George Santayana ↩︎

Australia today: According to the 2021 Census, Australia had 9,275,217 occupied private dwellings and an average of 3.1 bedrooms per dwelling. Multiply that out and you get about 28.75 million bedrooms in occupied private dwellings, excluding unoccupied dwellings, hotels, etc. The same Census counted 25,422,788 people in Australia (excluding overseas visitors). So we have more bedrooms than people, roughly 1.13 bedrooms per person.

That does not mean everything is fine. It means the problem is not “Australia has no space to sleep”. It is that bedrooms are not conveniently located near jobs, they are not on the market to be rented out, and the ones that are on the market, are priced higher than consumers would like. It is an allocation problem, and an affordability problem.

You can see the long-run drift in the background conditions: Over a century, household size has fallen from 4.5 people per household (1911) to about 2.5 (2021). New houses are still large. In the ABS floor-area series since 2002–03, the average new house sits around the 230–246 m² range, peaking in 2008–09. Our demand for (and supply of) space per capita has increased.

I tend to support the pro-development YIMBY forces (Yes in my backyard), in contrast to the NIMBY movement. We should build more housing, especially in Sydney’s CBDs and Eastern Suburbs, and other job-rich, housing-poor places. This will lower commute times for everyone.

But “build more” runs into two limits that are easy to ignore on a whiteboard.

First, the industry is not a magic tap. Construction employment and labour costs have risen strongly since 2020–21, and the ABS shows that labour-cost growth has outpaced employment growth recently, with strong demand for skilled labour driving wages. NSW’s Productivity Commission review identify “construction industry capacity” and “skills shortages” as supply constraints.

Second, migration both adds demand for homes, and it also adds labour, including in construction. Net overseas migration was 306,000 in 2024–25 (down from 429,000 the year before).

And as a corrective, consider China. Real estate can go from “needed” to “bubble” to “macroeconomic trap”. The IMF describes China’s property downturn as deep and prolonged, with housing starts down sharply relative to pre-pandemic levels. China provides a case study in what happens when a country leans too hard on housing as an engine of growth.

SIX HOUSING CLICHÉS WORTH RETIRING

Below, each “myth” has one source that states the claim (or leans into it), plus two sources that point the other way.

1) Myth: “Building more won’t make housing more affordable.” [source]

Fact: New market-rate supply tends to reduce rents, including in low-income areas, and it does so through both local effects and city-wide “moving [vacancy] chains”. [source]

2) Myth: “Filtering is a fairy tale, new ‘nice’ homes do not help anyone else.” [source]

Fact: Empirical work using address histories and register data finds moving chains that reach middle- and lower-income households and neighbourhoods. [source]:

3) Myth: “Zoning and planning are not really the issue.” [source]

Fact: The RBA estimates a large “zoning wedge” in major cities, meaning prices substantially above marginal supply costs in their framework. [source]

4) Myth: “Upzoning just bids up land, it does not reduce rents.” [source]

Fact: Large-scale zoning reform and staggered upzoning reforms produce more housing and lower (or slower-growing) rents versus counterfactuals. [source] [source]

5) Myth: “New development drives nearby rents up (gentrification beats supply).” [source] (discussing supply skepticism)

Fact: In settings where this can be cleanly studied, the net local effect around new large buildings is often rent decreases, because the supply effect dominates. [source]

6) Myth: “Airbnb exacerbate the housing crisis.” [source] [source]

Fact: Short-term rentals obviously matter a lot in some places (resort communities), but at state or national scale they are typically a small share of the dwelling stock (statistical noise), and by unlocking locked up bedrooms (remember there are more bedrooms than people in Australia), they relieve supply constraints by renting to new residents, transients, temporary workers, as well as tourists. [source][source] The complaint is that they take whole houses off the market, not just bedrooms. In the US, “a 1% increase in Airbnb listings leads to a 0.018% increase in rents and a 0.026% increase in house prices.” [source] They make houses more productive to the owners (who now can earn rent off their asset, and help pay down the mortgage), which drives up prices because it drives up value. Logically, they are also displacing some existing hotel stays and perhaps inducing some demand, creating value. But if we have a hotel-room shortage, we should build more hotels. And if hotels are now less full because of AirBnB, they can lower prices and start attracting some longer-term stays who previously would have been in the short-term rental market.

WHERE THIS LEAVES ME

Australia does not look like my grandmother’s world of six people per bedroom. On the raw stock of bedrooms, we look like a country with enough room and rooms.

The trouble is that rooms are not the same as homes, and homes are not the same as homes near jobs, and homes near jobs are not the same as homes at prices that are affordable at a reasonable fraction of wages.

So yes, build more, especially where the jobs are. Just do not pretend the bottleneck is approvals, or AirBnB, or taxes. Also do not pretend that more supply won’t lower prices.

Yet there remains a construction sector with finite capacity. If more houses are to be built, construction workers need to be more productive, or we need more of them. Robotics aren’t advancing fast enough to increase productivity much in the short term, so we need more workers to build houses faster. And they will come either from overseas or from other domestic industries. And their labor cost will go up if their demand increases.

Notably, if we stop migration, Australia’s population would eventually decline due to low fertility rates. While that might ‘solve’ the raw demand for space, it would leave us with an aging population in large houses and fewer workers to build the infrastructure we need and serve them (us) in our retirement. The solution isn’t fewer people, it’s better alignment between who lives where and where we work.

Happy New Year to all who celebrate.

ANNOUNCEMENTS

There is a new global Association Of Sustainable Transportation Researchers. Seems better than their competitor, The Association Of Unsustainable Transportation Researchers.

I organise the Sydney Transport Advisory Board, an annual meetup of local transport aficionados who really do know better what to do than those in power, and regularly offer unsolicited advice. If you are interested in being invited to next year’s field trip in December 2026 let me know.

POSTS

• On Bondi Beach: I arrived in Sydney on April 10, 2017. My photos library date stamp indicates that I visited the iconic Bondi Beach by April 15. It’s gorgeous without the people (photos below), and there is a well-worn walk up and down the coast, most famously to Coogee, the next large beach down the Ocean, with some smaller ones in between. …

• Inner West’s The GreenWay is Effing Great. Last week I left my home in Arncliffe (at 07:00 am) to the newly opened Inner West “The GreenWay”. …

• Access for Defence: Access is usually about reaching jobs, shops, and schools. It is also about controlling who can reach us. ….

RESEARCH BY OTHERS

• A new report from the excellent Rob Bain: TOLL PRICE ELASTICITY OF DEMAND

RESEARCH

• Levinson, David (2026) When Are Impedance Choices Irrelevant? Equivalence Conditions for Hansen-Style Access Metrics. Findings. [doi]

MEDIA

TOLL ROADS AGAIN

• Two-way tolling on Sydney Harbour Bridge and Tunnel set to start in 2028

FLIXBUS

• I did an interview on FBi Radio Sydney on the FlixBus, can’t seem to find it on the Internet. FlixBus seems to be taking over much intercity bus service, e.g. they now own Greyhound in the US (but not Australia). This means lower prices for users, and hopefully economically sustainable service.

YOUR REGULAR REMINDER THAT CARS ARE WEAPONS

• Queensland: Woman dies in hospital after car ploughs into eight pedestrians.

• New South Wales (Surveillance tool of violence, rather than weapon as such). Man allegedly used Tesla car app to target domestic violence victim.

• The Levant

  • Shot: Israeli reservist rams vehicle into Palestinian man praying in West Bank

  • Chaser: Two killed in suspected Palestinian ramming and knife attack in Israel

• Maryland: ICE agents involved in Maryland shooting that injures two people
  

LINKS

• Air India loses a plane for 13 years. (No passengers were aboard)

• There will be 6.5m Sydney residents by 2045. Here’s where they’re going to live

• AVs

  • Driverless future gains momentum with global robotaxi deployments

  • Waymo Year in Review. – Waymo hits 14 million trips in 2025, more than 20 million lifetime. Still more than doubling annually. But not without hiccups …

  • The Night the Lights went out on Waymo. – Blackout in San Francisco, and no traffic signals, leads to a total, ungraceful, Waymo system shutdown, mid-lane. Updated software is coming.

  • Robotaxis could wreck parking revenue, airport official warns (Oh, no!)

• Ford retreats from EVs, takes $19.5 billion charge as Trump policies grip industry – They are moving more towards hybrids. They will undoubtedly be back.

• Why the world rejected the UN’s Road Signs – Entertaining video by Chris Spargo

• Parramatta LRT is not meeting passenger forecasts. A real chin-scratcher that one.

• The US Bureau of Labor Statistics reports the college enrollment rate of recent high school graduates was 66.2% in October 2019 and 62.8% in October 2024. … This plus a demographic cliff. Universities are in structural trouble even without the current administration.
  

THE TRANSPORTIST: LANE-FREE

As a comment, I plan to be more lane-free this year. “Stay in your lane” they say, so you don’t annoy your readers, or offend someone on LinkedIn, or whatever. But why? Am I going to be more successful by shaking fewer trees? I sort of doubt that. Will I be funded less if I speak out more? I doubt it. I have given certain local transport agencies sufficient politeness for almost no reward, on the theory I shouldn’t speak ill of past, present, or prospective sponsors, or those funding students, or at least giving us data to do research with, but if they do none of that, meh?

As always you can follow me on Mastodon, BlueSky, or RSS.

Cleaned transcript (lightly edited for clarity and flow)

Speaker H:
Now, I know this has been discussed a lot across 702 ABC Sydney today, but it is a big deal. Today, the Minns Government confirmed that, from late 2028, the Sydney Harbour Bridge and the Sydney Harbour Tunnel will be tolled in both directions.

Is this a broken election pledge? Didn’t Chris Minns say, when he came into office, that there would be no new tolls for old roads? But today it’s confirmed there will be tolls in both directions on the bridge and the tunnel.

Here is John Graham on 702 Breakfast with Chris Taylor this morning.

John Graham (excerpt):
It allows us to actually fund that $60 toll cap that’s been so important to motorists who just don’t have the option to jump on public transport and move around the city.

They do have to be in their car, it is expensive. That $60 toll cap is making a huge difference, but it is expensive to fund without some sort of reform.

But there is a second reason as well, and this is the reason why it was always the former Government’s plan to have two-way tolling on the harbour crossings. Once the new Western Harbour Tunnel, that third harbour crossing, opens, it is tolled in two directions.

To have two-way tolling on that Western Harbour Tunnel and one-way tolling on the Sydney harbour crossings would cause traffic chaos. That’s the advice from traffic experts as they look at the network.

If we didn’t do that, we’d simply have many, many drivers rerouting through the city. It would bring the traffic network to a halt. That’s why it was always the plan.

Renee Krosch:
“Hmmm, always the plan.” John Graham speaking on 702 Breakfast this morning. So, tolling in both directions, it’s happening at the end of 2028 when the new Western Harbour Tunnel opens.

What do you think? 1300 222 702. If you’ve got thoughts about this, the text line is 0467 922 2702.

Professor David Levinson is Professor of Transport at the University of Sydney. He comes on regularly here on evenings, and here he is again tonight. Hi David.

Professor David Levinson:
Hi.

Renee Krosch:
What do you think, David? Do you think it’s fair and reasonable to toll the Sydney Harbour Bridge and the tunnel in both directions?

Professor David Levinson:
I think it’s a bit inefficient. You could have one toll that was twice as much in one direction, and you wouldn’t have to pay tolls twice, and that’s what we used to do historically on lots of bridges. You would need half as many toll collectors. Now it’s automated, but you still have to build gantries and there’s still toll collection equipment.

The problem is that not all of the harbour crossings are tolled. If you’re willing to go farther west to the Iron Cove Bridge, Gladesville, or even farther west out to Rhodes, you’re not paying a toll when you’re crossing the river. So some people would cross in the non-tolled direction one way and avoid paying the tolls entirely.

So you have to decide whether you’re going to use the Parramatta River and the harbour as a toll cordon all the way through, with a toll northbound or southbound everywhere, or if you just want to do it in the more urban areas, in the city, where you have the tunnel, the new tunnel, and the bridge. Then you could put tolls on both directions.

But they made a choice to put tolls in both directions on the Western Harbour Tunnel, and now they’re using that as an excuse to put tolls in both directions on the other two facilities, which benefits people who collect tolls.

Renee Krosch:
It does benefit people who collect tolls. Am I right that the people who collect the tolls are getting two and a half billion dollars every year from Sydney motorists?

Professor David Levinson:
I don’t know the exact number, but they’re getting a lot of money from Sydney motorists who are paying tolls on various facilities. The current harbour crossings are not Transurban facilities, but the other toll roads are, and those tolls are collected at a pretty high rate from Sydney motorists.

Renee Krosch:
We’re the most tolled city in the world, we hear this a lot. We’ve got more kilometres of toll roads, and more toll points, than any other city.

Has the public private partnership model to build infrastructure in this city ultimately let down motorists?

Professor David Levinson:
I think it’s let down taxpayers in a lot of ways. We have all of these tolls, and then we put toll caps on. The toll caps are paid for by general revenue, everybody is paying into it, so that people who are driving on the toll roads don’t have to pay more than $60 a week.

You can say, that seems fair. We have a cap on transit fares, so if you ride Sydney trains, the light rail, and ferries, you’re not paying more than $50-something a week.

But we want people to use public transport and we want fewer people to drive. Instead, we’ve created an incentive to drive more kilometres. After $60, you’re not paying tolls anymore. We’re increasing people’s incentive to drive more. I think that’s distorted, and it’s not what we want as overall public policy.

Renee Krosch:
So the amount currently to cross the Harbour Bridge is about $4 or something, is that right?

Professor David Levinson:
I think $4.41 is what the newspaper article said in peak periods, and $3.20 in off-peak.

Renee Krosch:
So you double that to go to work and come home if you live in that part of the city. Let’s call that $8 a day, $40 a week. You’re not getting anything back on the toll relief.

I understand the extra money collected on the Harbour Bridge and the tunnel is going to help fund the toll relief policy, which is a good thing. The toll relief is a good thing.

Professor David Levinson:
Well, it might be a good thing. It’s a good thing if you’re getting the toll relief. It’s not a good thing for the people who aren’t getting the toll relief, who don’t drive that much.

It’s not a good thing for transit users. It’s not a good thing for people in eastern Sydney who aren’t driving enough to get the toll relief, and who are paying for other people’s toll relief.

It’s a transfer. We can decide we want to benefit one set of people over another, but that’s a policy choice. As you say, we’re doubling the tolls on one set of people so that another set of people don’t have to pay as much.

Renee Krosch:
I understand the equity issue, if you’re living in parts of Sydney, particularly western Sydney, there are no options for public transport and you do have to drive a long distance to get to the CBD.

However, I don’t know why and how we got to this point where this city is the most tolled city in the world.

Professor David Levinson:
There was a set of decisions that they didn’t want to pay for motorways out of current revenue. They wanted to pay for them out of future revenue.

So the question becomes, how do you do that? You need financing. How do you get financing? You ask somebody else to come up with a large sum of money so you can build the road, or you promise you can sell the thing you built, like WestConnex. We sold a large share of it to Transurban.

In exchange, future motorists have to pay for it.

The alternative, which a lot of places have used historically, is you could use fuel taxes to pay for all of the road construction, and we wouldn’t have tolls at all. That would likely lead to more congestion on motorways because people wouldn’t be discouraged from using them by high tolls. Again, that’s a policy choice.

It’s better from an economic point of view that everybody pays proportional to the amount that they use the facility, but we’ve created a system where some people have paid motorways and other people have free roads and aren’t paying tolls on those roads. In a sense, they’re getting an advantage.

Even with this rebate, you have to go online and apply. You have to fill out paperwork to get it, even though it could be done automatically. We do it automatically for public transport, but we don’t do it automatically for motorways.

One effect of not automating it is that people who don’t file the paperwork don’t get paid back. That saves money for the operator or the government, in this case, from having to provide the subsidy. You can say you’re providing a subsidy, and if you didn’t get it, it’s your fault for not filling out paperwork. But we didn’t have to make you fill out paperwork in the first place, that was a policy choice.

Renee Krosch:
Some people are saying, whatever happened to free travel after the bridge was paid for? That has been the promise, hasn’t it, from the Minns Government, at least. We wouldn’t put new tolls on old roads.

Someone else is saying, if you live on the North Shore, if you live in the East, you can afford it. I don’t know if that’s right, Rob, who’s just texted in. This assumption that because you live in a particular part of the city, you can, or can’t, afford to pay for things, particularly road tolls.

Professor David Levinson:
Some people can, and others can’t. Average income in the East is higher than average income in the West, but it doesn’t mean everybody in the East is wealthier than everyone in the West.

We want more people to be able to live in the East, to take the jobs that are in the East, so there isn’t as much west to east commuting in the morning.

Public policy is trying to increase housing around public transport stations, and apartment units. Those units are not only for the wealthiest people. They’re for people who can afford to live there.

Listener (unidentified):
We’re working class people, David, we are…

Renee Krosch:
We’re out of time. David, thank you so much for being here. Professor of Transport at the University of Sydney, David Levinson.

Well, it’s happening. Confirmed today by the Minns Government. Two-way tolling.