Fair Tolls & Transport That Works

Sydney is regularly described as the most tolled city on Earth. The independent review the NSW Government commissioned put it more carefully, and the careful version is the one worth defending: Sydney "has more toll roads than any other city in Australia and is one of the most tolled cities in the world" (Fels and Cousins, 2024). The argument does not need the superlative. It needs the arithmetic.

On modelling prepared by NSW Treasury and Transport for NSW for that review, and described by the reviewers as resting on "a conservative set of assumptions", motorists face "a toll burden estimated at $123 billion in today's dollars over the next 37 years to 2060. Of this, $64 billion is from WestConnex alone". Sydney motorists are already paying around $2.5 billion a year (Fels and Cousins, 2024). Every figure here is Sydney-specific; none of it is a national average doing duty as a local one.

One clarification belongs at the front, because two numbers circulate and they are not two liabilities. Ministers have quoted a figure of $195 billion. That is the same toll burden expressed in nominal dollars — the cash motorists will hand over decade by decade before inflation is stripped out. The review's $123 billion is that same stream in today's dollars. Quoting both without saying which is which is how a real problem starts to sound like an exaggerated one.

What the network actually costs, road by road

The review broke the burden down asset by asset in today's dollars to 2060 (Fels and Cousins, 2024):

• WestConnex M4 — $24.60 billion
• M7 — $20.05 billion
• M5 South-West — $18.87 billion
• WestConnex M5 East and M8 — $18.30 billion
• Hills M2 — $15.95 billion
• Eastern Distributor — $6.09 billion
• NorthConnex — $5.80 billion
• Sydney Harbour Bridge and Tunnel — $4.01 billion
• Lane Cove Tunnel — $3.75 billion
• WestConnex M4–M8 Link — $3.37 billion
• Cross City Tunnel — $1.25 billion

Those eleven lines are effectively the whole $123 billion, and four of them are WestConnex assets. This is not a burden spread thinly across a lot of small roads. It is concentrated in a handful of very large ones, and the largest of them are the ones the state no longer owns.

Who pays, and how much

Finding 8 of the review is blunt: "The financial impact of tolls is greatest in Western Sydney." A Department of Customer Service survey cited in the report found that residents of Blacktown, the South-West and Parramatta who used toll roads more than once a month — 37 per cent, 53 per cent and 45 per cent of those surveyed respectively — spent on average $95.90, $87.63 and $84.35 a month, against a Greater Sydney average of $60.70 among motorists who used a toll road at least once a month (Fels and Cousins, 2024).

We should be honest about the shape of that burden rather than reaching for the worst case. The review's own survey found that around 74 per cent of toll road users spend under $20 a week, and only about 3 per cent spend more than $100 a week (Fels and Cousins, 2024). Tolls costing thousands of dollars a year are real, but they describe a tail, not the typical driver. What the tail has in common is geography: between 40 and 50 per cent of toll road users surveyed in Rouse Hill–McGraths Hill, Baulkham Hills, Liverpool and Hurstville spend $20 or more a week (Fels and Cousins, 2024). For those households, the toll is not a motoring cost. It is a second utility bill attached to the job.

Why it keeps rising

The familiar shorthand — about four per cent a year or inflation, whichever is higher — is true of the biggest concession and misleading about the rest. The escalators are contract-specific, and several are harsher than the shorthand suggests. On the review's table of tolls as at April 2024 (Fels and Cousins, 2024):

• WestConnex (M4, M5 East, M8, M4–M8 Link, Rozelle Interchange) — "Until 31 December 2040: The greater of CPI or 4% per annum. From 1 January 2041: CPI per annum", on a $1.67 flagfall plus $0.62 per kilometre, capped at $11.78 for a car.
• Hills M2 — the greater of CPI and 1 per cent per quarter.
• Cross City Tunnel — CPI per quarter.
• Eastern Distributor — "Greater of (37.5% x CPI + 62.5% x AWE) and 1% per quarter". That toll is indexed mostly to average weekly earnings, which means it tracks wages rather than lagging them.

Underneath the formulas sits the structural problem the review identifies in Finding 6: "Tolls are too rigid and are locked-in for decades without options for review." Unlike a regulated water or electricity business, a toll concession has no periodic price reset at which an independent regulator can ask whether the return is still reasonable. Transurban told the Legislative Council's 2017 road tolling inquiry that "Beyond the initial agreement there is no pricing flexibility in the concession" (Fels and Cousins, 2024). A price set in 1992 or 2014 simply keeps escalating.

A price with no logic

The clearest single proof that nobody designed this network as a network is the per-kilometre spread. The review reports tolls ranging "from $3.13/km in the Cross City Tunnel to $0.25/km for the M7", a twelve-fold gap it attributes to "historical toll setting and escalation approach differences" (Fels and Cousins, 2024). Two motorists pay wildly different rates for the same commodity — a kilometre of road — because of when the contract was signed.

It is worth correcting a claim commonly made on our side of this argument, including by us. The steepest per-kilometre rates are not in the west; they are in the CBD tunnels. The review says so directly: "On a per kilometre basis, tolls are already relatively low on the M7, but the evidence was that people from Western Sydney suburbs spend more on tolls per week than people from elsewhere do" (Fels and Cousins, 2024). The Western Sydney burden comes from distance travelled and the absence of an alternative, not from a premium rate per kilometre. That is a better argument, because it is the true one.

This did not happen by accident. It happened one deal at a time.

Sydney's toll network was assembled project by project, each with its own concession, its own pricing formula and its own confidential terms — some running into the 2050s and 2060s. The independent toll review commissioned by the NSW Government found what motorists already knew: the network has no consistent pricing logic, and the outcome is neither fair nor efficient. Historically, some concession arrangements have also included confidential clauses protecting operator revenue, which risks quietly constraining future transport decisions.

Motorists are paying these tolls for decades. They are entitled to know what was signed in their name. Yet key contract terms, traffic modelling and amendments have too often been treated as commercial-in-confidence long after the commercial sensitivity has passed.

What the Auditor-General can no longer tell you

This is not a rhetorical complaint. It is a documented loss of oversight. Auditing WestConnex in 2021, the Auditor-General recorded that "Since the NSW Government sold 51 per cent of its stake in WestConnex in August 2018, the Auditor-General no longer has the mandate to provide this assurance. We are also unable to provide any assurance regarding the performance of tolling concessions" — with the consequence that "the total costs of WestConnex, including those levied on road users through tolling, are not reported alongside the full cost of delivering the project" (Audit Office of New South Wales, 2021).

The same audit found that $4.26 billion of related work had been moved outside the reported $16.812 billion 2015 WestConnex budget: the Sydney Gateway road component at $1.76 billion beyond WestConnex's $800 million contribution, the Network Integration Program at $2.3 billion, and the Parramatta Road Urban Amenity Program at $198 million. It also found that "complete and valid records of key decision-making processes, analysis and advice were unavailable", and that Transport for NSW "often provided information that was incomplete or unverifiable" (Audit Office of New South Wales, 2021). A project whose own auditor cannot reconstruct the decisions is not a project the public can be said to have consented to.

Sold twice, and banked once

The remaining 49 per cent of WestConnex was sold in two tranches, completing in 2021, to Sydney Transport Partners — a consortium of Transurban, AustralianSuper, the Canada Pension Plan Investment Board and Tawreed — for $11.1 billion, taking the consortium to full ownership. Net proceeds were required by legislation to be paid into the NSW Generations Fund before being applied to retiring an equivalent amount of debt (NSW Treasury, 2021). Whatever the merits of debt retirement, it is worth being clear about what happened: a stream of payments from Western Sydney motorists running to 2060 was converted into a lump sum that did not have to be spent on transport, and was not.

What has actually been legislated, and what has not

Professor Allan Fels AO and Dr David Cousins AM delivered Motorists First on 16 July 2024 after a fifteen-month review, with 16 numbered findings and 42 recommendations. Recommendation 1 is that the Government adopt the review's Proposed New Tolling Principles (Fels and Cousins, 2024).

What has followed is institutional rather than pricing reform, and the distinction matters. The Transport Administration Amendment (NSW Motorways) Act 2024, passed in November 2024, created NSW Motorways — formed on 1 July 2025 — gave IPART a toll monitoring and investigation role, and provided for an independent tolling ombudsman. A NSW Tollway Ombudsman, described as the first independent tolling-focused ombudsman in Australia, became contactable from 5 January 2026, and Pauline Gatomé was appointed NSW Motorways' Customer Advocate (Service NSW, 2026).

We support those institutions and would defend them, with one caution. A monitor and an ombudsman are downstream remedies: they can investigate a price and adjudicate a dispute about a bill, but neither can reopen a concession or reset a formula that the contract says is fixed until 2040. Giving IPART a monitoring role over prices it cannot change is an improvement on nobody watching, and it is not the same thing as regulation. The test of the new arrangements will be whether what the monitor finds is published in full, whether the ombudsman's caseload is reported by operator and by cause, and whether either body's findings are capable of changing anything upstream of a customer complaint.

New South Wales has not adopted a single network toll on a declining distance basis. That headline structural proposal has been agreed only in principle; the pricing restructure itself remains the subject of negotiation with concession holders. Anyone campaigning on this — including us — should say so plainly rather than treat the scaffolding as the building.

The Australian Democrats support moving from a patchwork of one-off deals to a network-wide fair pricing scheme, negotiated with concession holders and, where needed, legislated. The direction is straightforward:

• Consistent distance-based pricing across the network, with declining per-kilometre rates for longer trips, so that the households making the longest drives are not the ones facing the largest total bill.
• Off-peak discounts that actually spread demand, and harbour crossing tolls brought inside the same whole-of-network settlement as every other road.
• A statutory public-interest test before any future toll concession, extension or amendment is signed — assessing distributional fairness, network effects and long-term cost to motorists, with the assessment published.
• Full transparency: plain-language contract summaries published for every toll concession, disclosure of pricing formulas and escalation clauses, no more secret side letters, and guaranteed Auditor-General access to concession documents — restoring the assurance that lapsed with the 2018 sale (Audit Office of New South Wales, 2021).PUBLISHED — under our plan

One thing should be said plainly before any of that is promised. Concession holders bought a contracted revenue stream, and changing it costs something — in compensation, in extended concession terms, or in a slower path to reform than a campaign leaflet would like. The review's in-principle agreement with operators exists precisely because the alternative is litigation. We are not going to pretend a fair network price is free, and we would not support a settlement whose cost to the public is disclosed only after it is signed. The published assessment we ask for below has to cover the price of the fix, not just the price of the status quo.

On harbour crossings, part of this argument has already been settled and we should acknowledge it rather than re-announce it. The Government has confirmed it "will proceed with the introduction of two-way tolling on the Sydney Harbour Bridge and Tunnel when the Western Harbour Tunnel, which will be tolled in both directions, opens in late 2028", on the reasoning that "Western Sydney motorists pay rising tolls in both directions on all their motorways year after year, while tolls on the Harbour Bridge and Tunnel are one-way and did not increase once between 2009 and 2023" (Service NSW, 2026). The review's own data bear that out: the harbour crossing toll changed just three times between 1992 and 2023 — in March 2004, January 2009 and October 2023 (Fels and Cousins, 2024). Our position is that this rebalancing has to arrive as part of a settlement that lowers the western bill, not as a standalone new charge.

What the review actually modelled

The review did not just assert that a network toll would be fairer; it modelled it. For 2026, comparing scenarios, it put the average car toll at $9.02 under the status quo, $7.62 under a network toll restructure — 16 per cent lower — and $5.43 if the restructure were combined with a general toll reduction, 40 per cent lower. Across all vehicle classes the equivalents were $11.18, $9.11 and $6.48 (Fels and Cousins, 2024). The declining distance method it worked through charges $0.65 per kilometre for the first segment, with the rate cut 15 per cent for each subsequent four-kilometre segment — $0.55, $0.47, $0.40, then $0.34.

Three caveats travel with those numbers and should not be left off. They are modelled averages for 2026 under indicative structures, not government policy and not a forecast of what anyone will actually be charged. Part of the reduction in the average comes from spreading the same revenue across more trips through two-way tolling, which is not a saving for a motorist who was previously travelling in the untolled direction. And the review is candid about the traffic consequence: "the implementation of two-way tolling is expected to add pressure to roads nearing capacity, potentially increasing congestion on some roads. Our analysis to date highlights the M2, M7, and M5 South-West as areas of concern" (Fels and Cousins, 2024). Those are Western Sydney roads. A reform sold to Western Sydney that congests Western Sydney is not finished being designed.

What a public-interest test would have to ask

A test that is only a slogan will be passed by every deal put to it. The review's findings supply the questions a real one has to answer before a concession, extension or amendment is signed, and each of them is a question that was not asked the last time:

• Is there a reset? Finding 6 is that tolls are "too rigid and are locked-in for decades without options for review" (Fels and Cousins, 2024). A concession with no periodic price review is a price nobody can correct for thirty years. If a deal has no reset mechanism, the test should ask what the state gained in exchange for giving one up.
• Who bears the escalation risk? A formula set at the greater of CPI or 4 per cent transfers all inflation risk to the motorist and none to the operator (Fels and Cousins, 2024). That is a distributional choice, and it should be stated as one in a published assessment rather than buried in a schedule.
• Does the upside actually come back? Finding 16 shows the existing sharing regimes only return money when performance exceeds agreed levels, so extra revenue below that threshold stays with the operator — and "The benefits shared with government to date have not been significant" (Fels and Cousins, 2024). Any new sharing clause should be modelled against realistic traffic, and the model published.
• What does it do to the roads next door? The review's own warning that two-way tolling would add pressure to the M2, M7 and M5 South-West is exactly the kind of network effect a project-by-project assessment cannot see (Fels and Cousins, 2024).
• Who can check afterwards? If the answer is nobody — as the Auditor-General has said of the tolling concessions since 2018 (Audit Office of New South Wales, 2021) — the deal fails the test on that ground alone.

Relief is not reform

Recent relief measures have helped some households — but claim-back schemes are a band-aid, and the review said as much in its own findings. Finding 15 reads: "Toll reform is preferable to toll relief. The current toll relief schemes are inadequately targeted and underutilised, in part due to overly complex administration. Toll relief is not financially sustainable given the existing pattern of toll escalation and limitations on the availability of government resources to fund relief" (Fels and Cousins, 2024).

The underutilisation is measurable. Transport for NSW estimates that 35 per cent of trips eligible for the M5 South-West Cashback scheme are never claimed. For the registration-based relief scheme, an average of 82 per cent of eligible vehicles in the 1155–1504kg weight class and only 64 per cent in the 1505–2504kg class applied over five financial years (Fels and Cousins, 2024). Nor is the benefit evenly spread: the top ten postcodes accounted for roughly 38 to 41 per cent of all M5 Cashback amounts claimed between 2011 and 2023, and in 2023 postcode 2170 alone accounted for about 10 per cent of claims (Fels and Cousins, 2024).

Finding 16 is sharper still, and it is the finding that should end the argument about rebates as a substitute for pricing reform: "Concessionaires are an unintended beneficiary of the current approach to toll relief. Increased traffic and patronage of toll roads, through induced demand created by toll relief, directly benefits operators by increasing their revenues." Government's protection against that is the upside-sharing regime in the concession agreements, but the review explains that "These mechanisms only return funds to the government if a toll road's performance exceeds agreed levels. So, if a toll road is used more, but not enough to hit agreed levels for sharing, the government will not receive a share of the extra revenues/profits, even though private concessionaires may be earning more due to toll relief" — and notes that "The benefits shared with government to date have not been significant" (Fels and Cousins, 2024). Public money spent on relief partly ends up as private revenue, and the contractual mechanism meant to claw it back rarely fires.

The fiscal scale is not small. Four relief schemes — the M5 Cashback and the three toll relief schemes — distributed $1.174 billion to motorists between July 2018 and 13 May 2024, and in 2022–23 alone around 753,000 toll accounts claimed $379.56 million across them (Fels and Cousins, 2024).

The $60 weekly toll cap belongs in the same frame, and the honest criticism of it has changed. It is no longer a scheme that will expire: it became permanent on 1 January 2026 after a two-year trial in which 712,000 claims were made and $214 million paid out, with 40 suburbs receiving more than $1 million each — Blacktown, Baulkham Hills, Auburn, Merrylands, Marsden Park, Castle Hill, Quakers Hill, Lakemba, Kellyville and Bankstown the largest claimants. The permanent scheme carries new guardrails: a fair-use limit of $400 per tag or plate and a $5,000 annual limit (Service NSW, 2026). Set that $214 million over two years against the roughly $2.5 billion a year Sydney motorists pay in tolls (Fels and Cousins, 2024), and the cap returns on the order of four per cent of what is collected. That is real money to the households claiming it and we would not take it away. But it is now a permanent public subsidy propping up a pricing structure nobody has fixed — which is a worse problem than a temporary one, not a better one.

The other bill: administration fees

Alongside the tolls sits a penalty stream most people only discover by accident. In 2024–25, 46 million toll notices carrying administration fees totalling $618 million were issued in relation to unpaid tolls on the Sydney network (Service NSW, 2026). That is fees issued rather than fees collected — the figure does not tell us how much was actually paid — but it is a stream of the same order as a quarter of annual toll revenue, generated largely by people who forgot to top up a tag. Negotiations with concession holders have cleared the way for these fees to be scrapped from mid-2026 and replaced with digital reminders, phased in from July 2026 (Service NSW, 2026). It is worth asking why a fee structure of that size took a formal review and two years of negotiation to dislodge, and what else in the concessions has never been looked at.

Fair tolls only matter alongside real alternatives — and for too many people in New South Wales, the alternative barely exists.

Buses that do not turn up

The Auditor-General's January 2025 audit of the nine Greater Sydney Bus Contract regions, run by six private operators and worth close to $8 billion over the eight-year life of the contracts, concluded that "TfNSW is not effectively managing bus contracts to ensure that operators are meeting contractual performance obligations and customer needs" (Audit Office of New South Wales, 2025a). The numbers underneath that conclusion:

• Against a 95 per cent target for first-stop on-time running, operators met or exceeded the KPI in only 15 per cent of instances over the 17 months from January 2023 to May 2024.
• Cancelled and incomplete trips peaked at 5.4 per cent of all trips in November 2022, and from January 2023 to May 2024 operators failed the cancellations KPI in 79.6 per cent of instances.
• Complaints per 100,000 boardings in May 2024 were approximately double the April 2022 rate; operators met the complaints KPI of fewer than 22 per 100,000 boardings just 41.6 per cent of the time from January 2023, down from 85.2 per cent between April and December 2022.
• Customer satisfaction with buses fell to 87 per cent, from 94 per cent in 2020.

The audit also exposes a contract design flaw worth more attention than it has had. Under clause 12 of the contracts, an operator in its first two years whose on-time running sits below 95 per cent will not generally have payment changes approved for its own punctuality-improving variations (Audit Office of New South Wales, 2025a). The operators performing worst are the ones least able to propose the fix. This audit covered the nine Greater Sydney regions and not Region 6 in the Inner West, which runs under the older Sydney Bus Service Contract, so it is not a verdict on every bus in Sydney — but it is a verdict on the model the state has chosen for most of them.

Where the driver shortage actually bit

The Government's own Bus Industry Taskforce found "over 500 driver vacancies throughout NSW as of April 2023", and that "most service cancellations and other service failures are a direct result of driver shortages" (NSW Bus Industry Taskforce, 2023). Across the ten most affected regions there were 409 vacancies against 5,585 drivers needed, a 7 per cent shortfall.

The geography of that shortage is not what the debate assumes, and we have got it wrong ourselves. The worst-hit regions were the Upper North Shore (25 vacancies, 15 per cent), the Blue Mountains (13, 13 per cent), the Inner West (114, 10 per cent), the North Shore (90, 10 per cent) and the Northern Beaches (68, 9 per cent). Western Sydney's Region 1 had 23 vacancies, a 4 per cent rate — the second-lowest of the ten. Vacancies in the inner and middle rings of Sydney and in Newcastle accounted for over 75 per cent of the statewide total (NSW Bus Industry Taskforce, 2023). Western Sydney's bus problem is not a driver vacancy rate. It is coverage, frequency and the absence of a route at all in new release areas — a different failure requiring a different fix. The Taskforce was also clear that the shortage "should have been predicted and better managed", described "the lack of basic driver facilities at layover areas" as unacceptable, and noted NSW bus driver hourly rates are lower than in other jurisdictions (NSW Bus Industry Taskforce, 2023).

We would prioritise minimum service standards — a frequent, all-day network on key corridors, not just peak-hour commuter runs — with published reliability data route by route, and workforce measures, including pay and depot facilities, that fix the driver shortage rather than manage its symptoms.

Western Sydney's transport gap, measured

A NSW upper house committee has already made the finding formally. Finding 1 of the Legislative Council's April 2024 report on public transport in Western Sydney is that "There is an inequitable provision of transport options and infrastructure in Western Sydney. This is particularly the case to new growth areas in southwest Sydney, leaving many people without adequate access to jobs and essential services and entrenching socio-economic barriers experienced in some areas" (NSW Legislative Council Portfolio Committee No. 6, 2024).

The committee recorded the Government's own evidence that many employment precincts, hospitals, schools and higher-education destinations in Western Sydney remain underserved by public transport, and that "72 per cent of households in the Western Parkland City own two cars or more compared to 51 per cent in the Eastern Harbour City". Transport for NSW travel data for 2022–23, reproduced in the report, show more than half of Western Parkland City travel is by private vehicle with public transport carrying 5 per cent of trips, and that residents there travel on average 20km or 32 minutes by public transport against 13km or 21 minutes by car (NSW Legislative Council Portfolio Committee No. 6, 2024).

Two figures from submissions to that inquiry sharpen the point, and they should be attributed to the councils that made them rather than to the committee or to Transport for NSW. Campbelltown City Council submitted that average bus travel time in its area is nearly double driving — 3 minutes 15 seconds per kilometre against 1 minute 43 seconds — producing what it called a "self-perpetuating cycle" of car dependency. Blacktown City Council submitted that a three-leg multi-modal trip can exceed $15, or $30 a day and $600 a month (NSW Legislative Council Portfolio Committee No. 6, 2024). Households facing that are not choosing the car. They are being priced and timetabled into it, and then tolled for the choice.

The fare side of the ledger

If the answer to a toll is a train, the fare has to make sense too. IPART's 2024 review of maximum Opal fares records that "The weekly cap has not changed since it was set at $50 in 2019, down from $63 in 2018", and recommends "recommencing escalations of weekly caps (taking into account the context of rebates and prices for private passenger transport such as toll relief caps)" (IPART, 2024). The tribunal put its finger on the awkwardness directly: "While the current NSW Government motorist toll relief rebate scheme caps toll spending at $60, increases to the weekly Opal cap may be perceived as poor value as it approaches the weekly motorist toll cap" (IPART, 2024). When the state caps what a driver pays and freezes what a commuter pays, the two caps end up in the same conversation whether or not anyone intended it.

The caps also reach fewer people than is often assumed. Approximately 2 to 3 per cent of cards reached the weekly cap between 2020 and 2023, rising to 7 per cent in 2024; 7 per cent of adult Opal and contactless passengers reach the daily cap, and 11 per cent of concession card users reach the weekly cap. The Gold Opal $2.50 daily cap has not changed since 2005 and the $2 transfer discount has not changed since 2016 (IPART, 2024).

IPART's determination allows a 2 per cent average real increase to maximum fares from 1 January 2025, followed by CPI each 1 July to 2028, across rail, bus, ferry, light rail, metro and on-demand services in Greater Sydney, the Blue Mountains, the Central Coast, the Illawarra and the Hunter: most bus and light rail maxima unchanged apart from a 1.6 per cent rise in the 3–8km band, train and metro maxima up between 0.7 and 4.1 per cent by distance band, Sydney ferry maxima up 3 to 4 per cent, and the Newcastle–Stockton ferry unchanged (IPART, 2024). On time-of-day pricing the tribunal decided "not to set peak and off-peak fares in our Determination but made a recommendation" that Transport for NSW "review current peak and off-peak arrangements (including times, fares, demand)" (IPART, 2024). Because IPART sets ceilings rather than prices, Transport for NSW is already free to discount below them. The instrument for spreading demand off the peak exists and is not being used. It should be — and the responsibility for that sits with the Government, not the regulator.

Trains that have not arrived

The state's record at buying rolling stock is its own argument for publishing business cases. The Auditor-General concluded in October 2025 that "TfNSW did not effectively procure the New Intercity Fleet (NIF) or the Regional Rail Fleet (RRF)... TfNSW did not effectively scope or estimate the full costs of the NIF or the RRF... it significantly underestimated those costs for both" (Audit Office of New South Wales, 2025b).

On the Regional Rail Fleet — announced in August 2017 as a public-private partnership to replace the XPT, Xplorer and Endeavour fleets, with a new maintenance facility at Dubbo — "At contract award in 2019, the RRF project capital cost was estimated to be $1.5 billion". The audit continues: "It was anticipated that the RRF would be introduced from early 2023. At the time of this report, no RRF trains have entered passenger service, and TfNSW estimates the project capital cost to be $2.3 billion." Enabling works alone were budgeted at $44.5 million at contract award in February 2019; by 2021 the project team had identified works at up to 120 stations, six stabling locations and other sites across three states, and in January 2023 the Government approved an additional $826 million. The New Intercity Fleet rose from $2.9 billion at contract award in 2016 to $4.5 billion, including a $1 billion funding top-up approved in 2022, taking the two fleets to a combined estimated capital cost of $6.8 billion (Audit Office of New South Wales, 2025b). A dispute with the private contractor Momentum Trains was ongoing at the end of September 2025; the audit expressly takes no view on its merits, and neither do we.

Regional passengers have been riding forty-year-old carriages while the replacement bill nearly doubled. That is the cost of announcing before costing, and it is the discipline we argue for in the first of our directions for debate below.

Country roads and the country road toll

Regional New South Wales pays for city motorways through its taxes and gets the worst of the state's road trauma in return. In 2025, 355 people died on NSW roads, 28 more than in 2024. Of those, 241 died on rural and regional roads — the highest regional toll since 2017 — and 134 died in crashes where speeding was a contributing factor, almost four in ten (NSW Government, 2026).

That concentration is not new, and the Auditor-General had already documented why it persists. The 2023 regional road safety audit found that "Around one-third of the state's population lives in regional NSW, but deaths on regional roads make up around two-thirds of the state's road toll"; that "There is a disproportionate amount of trauma on regional roads, but there are no specific road safety plans or trauma reduction targets for regional NSW"; that "The proportion of road fatalities and serious injuries in regional NSW is almost the same as ten years ago"; and that "There is no regional implementation plan to assist TfNSW to target the Road Safety Action Plan 2026 to regional areas" (Audit Office of New South Wales, 2023). The same audit identified a forecast $104 million underspend of the Community Road Safety Fund in 2022–23, and that only 52 per cent of regional councils took part in the Local Government Road Safety Program that year. Money set aside for road safety went unspent while the regional toll stayed flat. We support a safety-led program with a published regional implementation plan, regional trauma reduction targets, and the Community Road Safety Fund spent on what it was hypothecated for.

Councils and the roads they cannot afford

Councils maintain most of the state's road network on budgets that were stretched before the flood years and are worse now. The Auditor-General's 2024 audit of council road asset management puts numbers on the drift: statewide, an average of 4.4 per cent of road assets were in poor condition and 2.6 per cent in very poor condition in FY2021–22, worsening to 5 per cent poor and 3.2 per cent very poor in FY2022–23. Against a gross replacement cost for council road assets of around $102.2 billion, total spending on road asset maintenance in FY2022–23 was around $1 billion; roads, bridges and footpaths absorbed around 17 per cent of total council expenditure in FY2021–22 and around 21 per cent in FY2022–23. Sixty-four councils spent less on road maintenance than their own plans said was required in FY2021–22, and 48 councils did so in FY2022–23 (Audit Office of New South Wales, 2024).

The Auditor-General's local government report for 2025 adds that councils spent $6.1 billion renewing and acquiring infrastructure, property, plant and equipment in 2024–25, up from $5.9 billion the year before, while 63 councils had inaccurate and incomplete fixed asset registers (Audit Office of New South Wales, 2026a). A council that cannot say accurately what it owns cannot plan what to renew, and cannot make a persuasive case to the state for the money to renew it.

We support multi-year, formula-based road maintenance funding for regional councils — certainty they can plan works around instead of competing annually for grant rounds — coupled with support for asset condition data good enough to justify it. Funding certainty and asset transparency are the same reform argued from two ends.

Level crossings

Level crossings are the place where the road and rail networks fail together, and where the policy response is most often built on an assumption the evidence does not support. Transport for NSW puts the state's stock at approximately 3,800 railway level crossings, of which about 1,500 are public road over public railway, spread across State, Regional and Local Roads — noting that while crash numbers are relatively low, the potential for fatalities and serious injuries in a crash is high (Transport for NSW, n.d.). Those counts should be treated as approximate: the page carrying them has not been updated in some years.

The national picture matters for design. In 2024–25 the national rail safety regulator recorded 35 collisions between a passenger or freight train and a road vehicle at level crossings across Australia, causing four fatalities and two serious injuries, four collisions between a train and a person causing one fatality and one serious injury, and 492 near hits with vehicles and 410 with persons. Critically, "Three of the four fatalities occurred at active level crossings", and across five years 57 per cent of vehicle collisions and 69 per cent of near hits occurred at actively controlled crossings (ONRSR, 2025). These are national figures, not NSW-specific ones, but the design lesson carries: installing boom gates is not the end of the problem, because most collisions already happen where controls exist. Passive crossings remain over-represented for heavy vehicles, which made up 2.39 per cent of registered motor vehicles in 2025 but 16.9 per cent of near hits with vehicles — a share ONRSR attributes to Bureau of Infrastructure and Transport Research Economics data (ONRSR, 2025). A crossings program should be prioritised on that evidence, published route by route, rather than on which crossing last made the news.

What ties this to the toll gate

None of this is a separate subject from tolls. A toll is only a fair price for a road if declining to pay it is a real option, and the evidence in this section is a catalogue of the ways it is not: a bus network that met its punctuality target in only 15 per cent of measured instances over seventeen months (Audit Office of New South Wales, 2025a), a regional fleet ordered in 2017 that has not carried a passenger (Audit Office of New South Wales, 2025b), a weekly fare cap frozen since 2019 while the toll cap was legislated into permanence (IPART, 2024; Service NSW, 2026), and a committee finding that the fastest-growing part of Sydney has been left without adequate access to jobs and services (NSW Legislative Council Portfolio Committee No. 6, 2024). Toll relief that arrives before the alternative does is not cost-of-living policy. It is compensation for a service the state has not delivered.

Every walking or cycling trip is one less car in the queue. New South Wales has good active-transport strategies on paper; the gap is delivery, and the delivery gap is a funding gap of about an order of magnitude.

The state's main active transport grant program, Get NSW Active, drew 333 applications from 113 councils in its most recent completed round, of which 85 projects were funded. Of those 85, 46 were in regional and outer metropolitan areas, 28 in the Eastern Harbour City and Central River City, and 11 in the Western Parkland City. A total of $60 million was made available for the following round — $50 million for broader active transport projects and $10 million for projects enabling children to walk or ride to school, including $500,000 for the Active Routes to School initiative — with applications open from 5 November 2025 to 12 December 2025 and successful applicants announced in April and May 2026 (Transport for NSW, 2026). Transport for NSW estimates more than 1.5 billion walking and bike riding trips are taken across the state each year and has stated a goal of doubling that within twenty years; that is an estimate and an aspiration respectively, not a measurement and not a commitment.

Councils are asking for roughly four times what is on offer, and the fastest-growing part of Sydney is receiving 13 per cent of the projects. The mode share data tell the same story from the other direction: walking accounts for 10 per cent of trips in the Western Parkland City and 13 per cent in the Central River City, against 20 per cent in the Eastern Harbour City (NSW Legislative Council Portfolio Committee No. 6, 2024). People walk where walking is possible.

There is a fiscal argument here that active transport advocates undersell. A separated cycleway or a safe walking route to school is the cheapest thing on this entire page: it is measured in millions where a motorway is measured in billions, it requires no thirty-year concession, no escalation formula and no rebate scheme, and it removes trips from the congested network rather than adding capacity that fills again. It is also the only intervention on this page that does not need to be negotiated with a private counterparty. That it is funded through an annual competitive grant round oversubscribed four to one, while the toll network carries a $123 billion liability to 2060 (Fels and Cousins, 2024), is a statement about how New South Wales makes decisions, not about what works.

We would back a connected network of separated cycleways in Sydney and regional centres, safe walking and riding routes to every school as a default of new road projects rather than an afterthought funded by competitive grant, and planning rules that put daily needs within walking distance in new suburbs instead of a twenty-minute drive away. We would also fund the program to a level that matches the demonstrated demand from councils, and publish the assessment scoring so an unsuccessful council knows why.

The thread through all of this is the same one that runs through our whole platform: transparency and fair dealing.

• Publish the contracts.
• Test future deals against the public interest.
• Price the network fairly.
• Fund the basics in the regions.

Transport policy does not need to be clever — it needs to be honest.

1. Publish the cost-benefit analysis before the announcement, not after it

Require every major transport commitment in New South Wales — motorway, rail line, bus program or toll concession — to carry a completed business case with a published cost-benefit analysis before it is announced, independently assured, and a published comparison of final actual cost against budget once it is delivered. This is not a hypothetical discipline. Auditing the Newcastle Urban Transformation and Transport Program, the Auditor-General found that "the analysis of the benefits and costs of light rail was prepared after the decision to build light rail had been made and announced", and that "light rail was announced publicly in June 2013. There was no business case or other preliminary planning work done before this announcement" — noting that "our previous reports have emphasised the importance of completing thorough business case analysis that sets out the expected scope, benefits and costs before announcing infrastructure projects" (Audit Office of New South Wales, 2018). The cost of ignoring that advice is measurable: "the State wrote off $907 million of capital project costs over the past 3 years", and the Audit Office now recommends that the NSW Government "report total actual cost compared to budgeted costs for completed projects, either in the Infrastructure Statement or agency annual reports", and tells agencies they should "conduct structured post-delivery evaluations and compare actual outcomes against business case expectations" (Audit Office of New South Wales, 2026b). The rolling stock record in section 4 is what the absence of that discipline costs in cash: a regional fleet announced in 2017, costed at $1.5 billion at contract award and now estimated at $2.3 billion, with no train yet carrying a passenger (Audit Office of New South Wales, 2025b). This extends the statutory public-interest test promised in section 3 beyond toll concessions to everything else the state builds — and it has to bind our own commitments too. The bus service standards and the regional roads program in section 4, and the separated cycleway network in section 5, should each have to show their working on the same published basis, and be redesigned or dropped if they cannot. A party that demands to see the toll contracts cannot exempt its own spending from the same daylight.

2. Price the whole road network, not just the strips of it that carry a concession

Begin the technical, legal and political work of replacing the toll patchwork and the eroding fuel excise with a single distance- and congestion-based road user charge across the network, set by an independent regulator and revenue-neutral for the average motorist rather than stacked on top of what they already pay. The economic case is not marginal. The Productivity Commission's Finding 5.3 is that "road pricing is the most effective way of tackling road congestion", because road user charging "is the most direct means of addressing congestion because it can target congested routes at the right times"; it put avoidable road congestion in Australia's cities at "an estimated $24 billion in 2018-19", growing "by an estimated 45 per cent by 2029-30" unless countered (Productivity Commission, 2021). The Commission is blunt about what a toll is and is not: "tollways are a road funding measure, not a congestion tax", the price "is the same regardless of traffic flow or time of day", and because "most roads are not priced, the effect of tollways is to add pressure to non-tolled roads, exacerbating local congestion" — which is why it concludes that road pricing "should apply to the whole urban road network, not just the segments of it where governments and private parties see an opportunity for a public-private partnership" (Productivity Commission, 2021). Modelling for Sydney and Melbourne suggests the price signal does the work: a $5 charge to enter the CBD in the weekday morning peak and another $5 to exit in the afternoon peak would see traffic heading into the cordon "drop by about 40 per cent in the morning peak" (Terrill, Moran and Ha, 2019). This has to be weighed honestly against sections 1 and 3, because it is in real tension with both: section 1 says Western and South-Western Sydney households are already paying too much, and pricing the whole network means a charge on roads that are free today. The Commission's own evidence sharpens that objection rather than softening it — on Infrastructure Australia figures it reports, 71 per cent of middle-suburb residents but only 44 per cent of outer-suburb residents live within walking distance of medium-to-high frequency public transport (Productivity Commission, 2021). A charge levied on people who have no alternative is not a price signal; it is a tax. So our position is conditional and sequenced: road pricing only as a replacement for tolls and fuel excise, never as an addition; the bus network in section 4 delivered first; and the declining per-kilometre structure promised in section 3 carried across into it, so that the longest drives are still not the dearest. The review's own modelling warning belongs here too — that two-way tolling is "expected to add pressure to roads nearing capacity", with "the M2, M7, and M5 South-West as areas of concern" (Fels and Cousins, 2024). Any pricing reform that pushes congestion west has failed its own test.

3. Make the alternative real before pricing the road, and enforce the contracts we already have

Before New South Wales asks any household to pay more to drive, it should be able to show the bus turned up. The evidence says it did not: operators met the 95 per cent on-time-running target in only 15 per cent of measured instances between January 2023 and May 2024 and failed the cancellations KPI 79.6 per cent of the time, on contracts worth close to $8 billion over eight years, under a regime the Auditor-General found Transport for NSW "is not effectively managing" (Audit Office of New South Wales, 2025a). We would publish route-level reliability and cancellation data monthly; enforce the existing performance regime rather than renegotiating around it; fix the clause 12 trap the audit identified, under which an underperforming operator in its first two years cannot readily get payment approval for its own punctuality-improving variations (Audit Office of New South Wales, 2025a); and treat driver pay, rostering and layover facilities as service delivery rather than industrial background noise, which is what the Bus Industry Taskforce found when it reported over 500 vacancies statewide, unacceptable driver facilities and NSW hourly rates below other jurisdictions (NSW Bus Industry Taskforce, 2023). Two honesty clauses. First, the audit did not cover Region 6 in the Inner West, so this is a finding about most of Sydney's bus network, not all of it. Second, this direction costs money and competes with the toll reduction argued for in section 3 — a network settlement that lowers tolls and a bus network built to a minimum service standard cannot both be free, and we would rather argue that trade-off in public than pretend it does not exist. A committee of this Parliament has already found the provision of transport in Western Sydney inequitable (NSW Legislative Council Portfolio Committee No. 6, 2024). Fixing the service is the precondition for every pricing reform on this page, not a companion to it.

Publish the numbers before the announcement, and price the road rather than the gantry.

Audit Office of New South Wales. (2018). Newcastle Urban Transformation and Transport Program. Performance audit, 12 December 2018. Sydney: Audit Office of New South Wales. https://www.audit.nsw.gov.au/our-work/reports/newcastle-urban-transformation-and-transport-program

Audit Office of New South Wales. (2021). WestConnex: changes since 2014. Performance audit, 17 June 2021. Sydney: Audit Office of New South Wales. https://www.audit.nsw.gov.au/our-work/reports/westconnex-changes-since-2014

Audit Office of New South Wales. (2023). Regional Road Safety. Performance audit, 30 November 2023. Sydney: Audit Office of New South Wales. https://www.audit.nsw.gov.au/our-work/reports/regional-road-safety

Audit Office of New South Wales. (2024). Road asset management in local government. Performance audit, 21 November 2024. Sydney: Audit Office of New South Wales. https://www.audit.nsw.gov.au/our-work/reports/road-asset-management-in-local-government

Audit Office of New South Wales. (2025a). Bus contracts in metropolitan Sydney. Performance audit, 29 January 2025. Sydney: Audit Office of New South Wales. https://www.audit.nsw.gov.au/our-work/reports/bus-contracts-in-metropolitan-sydney

Audit Office of New South Wales. (2025b). Rail rolling stock procurement. Performance audit, 21 October 2025. Sydney: Audit Office of New South Wales. https://www.audit.nsw.gov.au/our-work/reports/rail-rolling-stock-procurement

Audit Office of New South Wales. (2026a). Local government 2025. 28 January 2026. Sydney: Audit Office of New South Wales. https://www.audit.nsw.gov.au/our-work/reports/local-government-2025

Audit Office of New South Wales. (2026b). Capital projects 2025. 18 March 2026. Sydney: Audit Office of New South Wales. https://www.audit.nsw.gov.au/our-work/reports/capital-projects-2025

Fels, A., & Cousins, D. (2024). Independent Toll Review: Motorists First — Final Report. 16 July 2024. Sydney: NSW Treasury. https://www.nsw.gov.au/departments-and-agencies/nsw-treasury

Independent Pricing and Regulatory Tribunal of New South Wales (IPART). (2024). Maximum Opal fares 2025–2028 — Final Report and Determination. October and November 2024. Sydney: IPART. https://www.ipart.nsw.gov.au/reviews/transport/public-transport/maximum-opal-fares-2025-2028

NSW Bus Industry Taskforce. (2023). First Report. July 2023. Sydney: Transport for NSW. https://www.transport.nsw.gov.au/projects/bus-industry-taskforce

NSW Government. (2026). 355 lives lost on NSW roads in 2025. Media release, Minister for Transport and Minister for Regional Transport and Roads, 2 January 2026. Sydney: NSW Government. https://www.nsw.gov.au/media-releases

NSW Legislative Council, Portfolio Committee No. 6 – Transport and the Arts. (2024). Current and future public transport needs in Western Sydney. Report 21, 29 April 2024. Sydney: Parliament of New South Wales. https://www.parliament.nsw.gov.au/committees

NSW Treasury. (2021). NSW Government finalises sale of WestConnex. Media release, October 2021. Sydney: NSW Treasury. https://www.nsw.gov.au/departments-and-agencies/nsw-treasury

Office of the National Rail Safety Regulator (ONRSR). (2025). Rail Safety Report 2024–25, Areas of Focus — Level crossings. Adelaide: ONRSR. https://www.onrsr.com.au/publications/corporate-publications/rail-safety-report

Productivity Commission. (2021). Public transport pricing. Commission research paper, 15 December 2021. Canberra: Productivity Commission. https://www.pc.gov.au/research/completed/public-transport

Service NSW. (2026). Lasting relief for motorists as ongoing $60 toll cap begins. Media release, 1 January 2026. Sydney: Service NSW. https://www.service.nsw.gov.au/news/lasting-relief-for-motorists-as-ongoing-60-toll-cap-begins

Terrill, M., Moran, G., & Ha, J. (2019). Right time, right place, right price: a practical plan for congestion charging in Sydney and Melbourne. Grattan Institute Report No. 2019-11, October 2019. Melbourne: Grattan Institute. https://grattan.edu.au/report/right-time-right-place-right-price/

Transport Administration Amendment (NSW Motorways) Act 2024 (NSW). Assented November 2024. Sydney: NSW Government. https://legislation.nsw.gov.au/

Transport for NSW. (2026). Get NSW Active. Program page, reviewed 9 March 2026. Sydney: Transport for NSW. https://www.transport.nsw.gov.au/projects/programs/get-nsw-active

Transport for NSW. (n.d.). Railway Level Crossings. Local Government Relations grant programs page. Sydney: Transport for NSW. https://www.transport.nsw.gov.au/industry/local-government-relations