The model

The NSW Government announced in 2011 that it would build WestConnex, a 33-kilometre motorway network linking western Sydney with the inner west, the CBD, and the airport. At the time, the project was estimated to cost approximately A$10 billion. The final forecast cost, depending on what is included, is between A$20 billion and A$45 billion.

The public contributed A$7.1 billion in grants and concessional loans before a single private dollar was invested. The NSW Government also bundled three publicly owned motorways (the M4, the M5 East, and the M5 Southwest) into the eventual sale package. Credit Suisse valued those existing public assets at A$9.2 billion.

In 2018, the government sold 51 per cent of WestConnex to a consortium led by Transurban for A$9.26 billion. In 2021, it sold the remaining 49 per cent for A$11.1 billion. Total received: approximately A$20.36 billion across two transactions. Writing in 2018 about the first sale, University of Sydney researcher Chris Standen calculated in The Conversation ‘a financial return of 34 cents for every dollar spent’, against public spending on WestConnex of more than A$23 billion ². That calculation predates the 2021 sale and does not cover it.

The Transurban CEO described the 2021 acquisition as supporting ‘free cash growth and distributions for Transurban security holders for the life of the concession.’ He was correct. The enterprise value of WestConnex, based on that transaction, was A$33 billion.

01How the concession model works

The toll road concession is a well-understood financial instrument in infrastructure finance. In its standard form, a government body offers a private company the right to build, operate, and toll a road for a defined period; the end dates of Transurban’s Australian concessions run from 2026 to 2065 ²⁰. The company bids for the concession competitively, agrees a toll schedule and escalation formula, builds the road, operates it, and returns it to the government at the end of the term.

The model has genuine logic. Governments face budget constraints. Private capital can fund infrastructure without immediate public debt. The company takes construction and traffic risk. At the end of the concession, the public gets a paid-off road.

What the model also does, in the concessions Transurban holds, is give a private company the right to toll a road until an end date set in the concession (extended for a number of roads), with escalations written in: Transurban says embedded fixed and CPI escalations cover more than 90 per cent of its FY25 toll revenue ²⁰. That this amounts to a legal monopoly is this desk’s own reading. Morningstar’s analysis rates Transurban a ‘wide moat’ company, ‘underpinned by efficient scale and other competitive advantages in existing markets’, and says the core Australian roads ‘benefit from strong barriers to entry’ ³.

Correction, 15 October 2026. This section said that Morningstar describes the ‘wide moat’ as ‘the government-granted concession that makes each road a legal monopoly’. Morningstar does not say that. Its analysis says Transurban ‘has a wide economic moat, underpinned by efficient scale and other competitive advantages in existing markets’, that the core Australian roads ‘benefit from strong barriers to entry’, and that concessions ‘grant the right to operate the roads and collect tolls for predetermined amounts of time’ ³. The passage and reference 3 now say what Morningstar says. That a concession is a government grant, and that this article reads it as a legal monopoly, is this desk’s own reading and is not attributed to Morningstar. The summary point that cited Morningstar ³ for “a legal monopoly on an essential road for 30 to 50 years, with automatic toll increases” went further than the page: it now gives what Morningstar says (strong barriers to entry; “like many natural monopolies”, the firm does not control its pricing) and marks the legal-monopoly reading as this desk’s own.

Correction, 19 October 2026. This section and the subtitle said that a concession typically runs 30 to 50 years, and this section said that Australian concessions transfer a legal monopoly, guarantee automatic toll increases regardless of traffic or cost, and provide contractual protection against competition. None of that was cited. Transurban’s FY25 Corporate Report gives concession end dates, not the length of each concession: from 2026 to 2065 for its Australian roads ²⁰. It says embedded fixed and CPI escalations cover more than 90 per cent of FY25 toll revenue ²⁰. It does not describe protection against competition, so that claim is cut, as are the “30 to 50 years” term and the guarantee “regardless of traffic or cost”. “Legal monopoly” is marked as this desk’s own reading. The subtitle now says a set term, because concession periods have been extended for a number of roads ²⁰.

02The WestConnex transaction: anatomy of a deal

WestConnex is the most documented example of how the concession model operates in practice in Australia, because it is the largest infrastructure transaction in Australian history and the level of public scrutiny was higher than for most toll road privatisations.

The public investment inputs are documented. Federal government: A$1.5 billion grant plus A$2 billion concessional loan. ¹ NSW Government: A$3.6 billion (A$1.8 billion from Restart NSW infrastructure fund plus A$1.8 billion from Consolidated Fund). ¹ The NSW Government also contributed three existing publicly owned motorways (M4, M5 East, M5 Southwest) valued by Credit Suisse at A$9.2 billion. ² A$5.3 billion of the A$9.26 billion first sale price was reinvested back into completing WestConnex Stage 3. ¹

The Conversation analysis, written in 2018 on the sale of the first 51 per cent, calculated ‘a financial return of 34 cents for every dollar spent’ ². The NSW Treasurer described the sale as a ‘very strong result.’

Infrastructure Australia’s review of the project’s business case was critical of the lack of rigour in appraising alternatives. Modelling showed that simply tolling the existing M4 and M5 motorways would have reduced congestion, without the billions in tunnel construction. That option was not pursued.

Correction, 7 October 2026. This article said that, against total public investment of A$23 billion or more, one analysis found the government "recovered approximately 34 cents for every dollar it spent", and placed that figure after the A$20.36 billion total of the 2018 and 2021 sales; this section said the analysis covered "all public inputs". The analysis is The Conversation article of 2018 ², which calculated "a financial return of 34 cents for every dollar spent" for the sale of the first 51 per cent. It was written before the 2021 sale and cannot cover it. The opening, this section, the fact box in "The risk that isn't", the image caption and the key facts now say so. Reference 2 now gives the article's correct address (it ends -102790, not -102780) and no longer lists the 2021 sale proceeds, which a 2018 article cannot report.

Update, 7 October 2026. The Conversation article ² gives two figures: its opening summary says the sale returns 30 cents for every dollar of public money spent, and its body calculates "a financial return of 34 cents for every dollar spent". This article quotes the body's words and figure; reference 2 now records both.

Update, 7 October 2026. Reference 13 was labelled "NSW Government" but points to Wikipedia's WestConnex entry; it is now labelled as Wikipedia. No sentence in this article carries its marker.

Correction, 19 October 2026. This section said that the NSW Government funded additional road works to funnel traffic onto WestConnex tolled sections, including reducing competing lanes on Parramatta Road. It was not cited, and none of the sources this article lists carries it. The claim is cut. The public inputs in the paragraph now carry reference markers ¹².

03When the model fails: bankruptcy and consolidation

Not every toll road concession works as intended. The first owner of Sydney’s Cross City Tunnel collapsed when usage fell well short of traffic forecasts, and the owners who bought it in 2007 went into voluntary administration in September 2013 ¹⁷. Sydney’s Lane Cove Tunnel opened in 2007; its owner went into receivership in January 2010 after traffic failed to meet forecasts of 100,000 cars a day ¹⁸. Same cause: insufficient traffic.

In both cases, the concession survived the corporate failure. The roads kept operating, the tolls kept being charged, and the assets were eventually acquired by the company with the most to gain from absorbing them into its network: Transurban. It agreed in 2010 to pay A$630.5 million for the Lane Cove Tunnel, which had cost its original owners almost A$1.7 billion ¹⁹, and in 2014 bought the Cross City Tunnel from its receivers and managers for approximately A$475 million plus stamp duty and transaction costs ¹⁶.

This is the bankruptcy pathway to monopoly. A challenger enters the market with an optimistic traffic model. The model proves wrong. The company goes bankrupt. Transurban, with an existing adjacent network that benefits from the connection, buys the distressed asset at a discount. The concession rights and the toll escalation formula continue. Only the owner changes.

Transurban now controls all but three of Australia’s 21 toll road networks.

Correction, 7 October 2026. This section said the Cross City Tunnel "entered receivership in 2006, just one year after opening", that its company "had overbid for the concession", that the Lane Cove Tunnel "entered administration in 2010", and that Transurban acquired both "at distressed prices". ABC News reported that the Cross City Tunnel's first owner collapsed when usage fell short of forecasts ¹⁷ and that the Lane Cove Tunnel's owner went into receivership in January 2010 ¹⁸; no source found says the first owner overbid. The section now gives those facts and the prices Transurban paid, from its own reports ¹⁶¹⁹, in place of "distressed prices". References 6 and 8 no longer attribute this history to IBISWorld and iSelect pages, which do not carry it.

Correction, 7 October 2026. This section said that after a failed toll road changes hands "the government-guaranteed returns continue". No reference supports a government guarantee of returns, and the failures this section describes, in which the Cross City Tunnel and Lane Cove Tunnel owners collapsed when traffic fell short ¹⁷¹⁸, show that returns were not guaranteed. The words are cut; the section now says only that the concession rights and the toll escalation formula continue.

Correction, 8 October 2026. This section said the Cross City Tunnel went bankrupt 16 months after opening, citing ABC News ⁶. The ABC News page cited does not give the 16 months or use the word bankrupt. It now says only that its first owner collapsed when usage fell well short of traffic forecasts, as ABC News reported on 14 September 2013 ¹⁷, and no longer gives the year the tunnel opened, which that page does not state. The 7 October correction above now cites ¹⁷, not ⁶, for the collapse of the owners when traffic fell short, because the page at ⁶ does not say so.

04The risk that isn’t: escalations that cushion traffic risk

The standard description of the concession model says companies ‘take construction and traffic risk.’ In Australia, the traffic half of that is cushioned by toll escalations written into the concessions.

Construction risk is real: companies must complete the road on agreed terms. Traffic risk is more complex, and escalations written into the concessions cushion it: Transurban says embedded fixed and CPI escalations cover more than 90 per cent of its FY25 toll revenue ²⁰.

The escalation clauses also reduce traffic risk indirectly. If traffic is lower than expected but tolls rise faster than expected, revenue can still grow. On WestConnex, tolls rise each year by 4 per cent or CPI, whichever is greater, until 2040, and by CPI after that ²¹. The revenue model is partially self-adjusting: lower traffic volume multiplied by higher toll rates can maintain target returns.

Morningstar says traffic volumes are ‘the main unknown when concessions are agreed’, leaving the firm ‘with usage risk’, while describing demand on the core Australian roads as ‘relatively inelastic’ ³.

Correction, 7 October 2026. The fact box called 34 cents "the estimated net return to government from the WestConnex sale, after accounting for all public inputs". The source calculated it for the 2018 sale of the first 51 per cent ²; the box now says so.

Correction, 15 October 2026. This section quoted Morningstar as saying ‘households seeking to save money might use their car less’, and called that the only genuine demand elasticity Transurban faces. The Morningstar page cited ³ (21 August 2024) does not contain those words or that claim. It says traffic volumes are ‘the main unknown when concessions are agreed, leaving the firm with usage risk’ and that the core Australian roads are subject to ‘relatively inelastic demand’. The passage now says that.

Correction, 19 October 2026. This section said that many Australian concession agreements include minimum revenue guarantees and compensation if government makes competing routes more attractive, and that the NSW Government reduced competing capacity on Parramatta Road to steer traffic into the WestConnex tunnels. None of it was cited, and none of it is in the sources this article lists. The claims are cut. The passage now says what Transurban’s FY25 Corporate Report carries: embedded fixed and CPI escalations cover more than 90 per cent of FY25 toll revenue ²⁰. The next paragraph’s “CPI or 4 per cent floor” was also uncited; it is now sourced to WestConnex’s own tolling page: 4 per cent or CPI, whichever is greater, until 2040, then CPI ²¹. The heading named traffic guarantees and government backstops; none is cited, and the heading now names the escalations. The opening line, which called the standard description “partially true and partially fictional”, is softened to match.

05The concession model in summary

The concession model transfers construction risk to private operators, which is defensible. It also grants the right to toll and operate an essential urban road for a set term, with toll escalations written into the concessions: Transurban says embedded fixed and CPI escalations cover more than 90 per cent of its FY25 toll revenue ²⁰. The end dates of its Australian concessions run from 2026 to 2065 (Legacy Way), and the period has been extended for a number of roads; at the end of a concession the road is to be returned to the government, although the M5 West, whose concession ends in December 2026, will then form part of the WestConnex M8/M5 East concession, which runs to 2060 ²⁰. That a concession amounts to a legal monopoly is this desk’s own reading; Morningstar says the core roads benefit from strong barriers to entry and that, like many natural monopolies, the firm does not control pricing ³. The result in Australia: one company now controls most of the toll road network in Sydney, Melbourne and Brisbane, and earns a 75.1 per cent Operating EBITDA margin (FY25) ⁷.

If it’s a rort, we cover it.

Correction, 19 October 2026. This summary said the concession model grants a “30 to 50 year legal monopoly”, with “automatic toll escalation and various protection mechanisms against competition”, and none of it was cited. Transurban’s FY25 Corporate Report gives concession end dates, not the length of each concession: from 2026 to 2065 for its Australian roads, each to be returned to the government at the end, except that the M5 West, whose concession ends in December 2026, will then form part of the WestConnex M8/M5 East concession ²⁰. Concession periods have been extended for a number of roads ²⁰, so the term is described as set, not fixed. It says embedded fixed and CPI escalations cover more than 90 per cent of FY25 toll revenue ²⁰. It does not describe protection against competition, so that claim is cut, and the “30 to 50 year” term is cut. “Legal monopoly” is now marked as this desk’s own reading, not a description given by Transurban or Morningstar ³. The 75 per cent margin is now stated as the FY25 Operating EBITDA margin, 75.1 per cent ⁷.