Urban

Tokyo Won't Save Melbourne's Suburban Rail Loop: The Case for State-Led Land-Banking Over Passive Value Capture

Sanjay Ravichandran 20 July 2026 12 min read

Tokyo Won’t Save Melbourne’s Suburban Rail Loop: The Case for State-Led Land-Banking Over Passive Value Capture

Melbourne’s Suburban Rail Loop is being sold partly on a financial promise: that the infrastructure will pay for roughly a third of itself by harvesting the property value it creates. It is an elegant idea. It is also, in the Australian institutional context, largely untested — and the overseas precedents invoked to justify it are more interventionist, more state-dependent, and more ambiguous in their outcomes than the government’s framing suggests.

The SRL’s $34.5 billion price tag is split into three nominally equal tranches of $11.5 billion each: state government, Commonwealth, and value capture ¹. That third tranche is doing serious work. And the mechanisms assigned to deliver it — land tax uplift, a windfall gains tax, developer levies, car park levies, and revenue from state-initiated development — are passive instruments dependent on density outcomes that Victoria’s fragmented municipal planning system may quietly frustrate. The five mechanisms themselves, and the 40-year collection horizon, are described in ². The argument here is that this structural mismatch is not a detail to be resolved in implementation. It is a foundational design flaw, and the only credible remedy is a shift from passive value harvesting to active state-led land-banking: acquiring parcels before construction announcements inflate prices, then master-developing them with binding density and affordability obligations.

What Tokyo and Hong Kong Actually Did

The Asian transit-oriented development story has been told so many times in Australian planning circles that its retelling has begun to obscure its content. The canonical version runs: build a rail line, let density cluster around stations, skim the resulting property uplift to repay construction costs. It is a compelling loop. It is also not quite what happened.

In Tokyo, Tokyu Corporation financed rail construction over decades by assembling land before the lines were built, then repaying commercial and Development Bank of Japan loans directly from land sales proceeds ³. The cross-subsidy worked because the developer-operator owned the asset being appreciated. The Tsukuba Express, more frequently cited as a contemporary proof-of-concept, tells a different story: 80% of its financing came from no-interest government loans, with local governments contributing a further 14%, leaving only a residual role for private capital ³. That is not a privatised value capture model. It is a publicly subsidised land development model with a private operator attached.

Hong Kong’s MTR is the other pillar. Its property development operations did outperform rail operations in net profit terms during 2000–2007, and that record has been invoked constantly in Australian infrastructure debates. But the relationship inverted between 2008 and 2010, and it is not structurally guaranteed to hold ³. More sobering still, MTR covers only roughly half its project costs through value capture, with government-funded portions still exceeding world-average construction costs ⁴. Hong Kong’s model works — to the extent it works — because MTR operates under a rail-plus-property framework with delegated planning authority that gives it near-sovereign control over precinct development. It does not petition local councils for rezoning. It holds the land, sets the brief, and develops it.

The normative coda to the Hong Kong story deserves more attention than it usually receives in Australian discussions. The city that has most successfully monetised transit-oriented development also has the developed world’s most overcrowded housing conditions, with median floor space of just 16 m² per capita ⁴. Whether this outcome is a direct consequence of optimising transit development for developer profitability, or a product of separate Hong Kong housing and land-supply policies, is a question the available evidence does not definitively resolve. What is not in dispute is that optimising a transit network for developer profitability and optimising it for resident welfare are not the same objective, and Hong Kong’s experience is a multi-decade demonstration of what can happen when the first is treated as a proxy for the second.

The common thread in both models is prior ownership or near-sovereign planning control. Victorian authorities have neither.

The $11.5 Billion Bet

The SRL’s value capture tranche is projected to be collected over 40 years ². Land tax in SRL precincts alone is projected to deliver $5.75 billion of that total ². The government cites evidence that land values near transit projects rise by at least 7 percent as the underlying justification for these projections ².

Several analytical problems compound here and none of them are addressed in the government’s available documentation.

First, the government’s cited figure of at least 7 percent uplift is presented as an evidentiary floor for its projections. Uplift is only capturable if density permissions are granted and development actually proceeds at the permitted scale. The SRL’s stated planning target of 70,000 homes within walking distance of stations ⁵ is an aspiration; it is not a zoned entitlement. The distance between those two things is where the funding model is most exposed.

Second, the taxable base in a corridor dominated by existing residential land is significantly narrowed by the characteristics of the value capture instruments as currently designed. The value capture model is structurally dependent on turnover of that land to higher-density uses — which requires rezoning, development approvals, and market uptake, across a 40-year horizon, across multiple municipal jurisdictions, many of which have historically been conservative on height and density.

Third — and this is the critical fiscal question the available evidence simply does not resolve — the $11.5 billion figure is projected over 40 years, yet the available sources are silent on the discount rate and net present value assumptions embedded in that projection ¹. Whether the nominal figure represents a credible fiscal commitment or a politically convenient headline number cannot be assessed without that information. The temporal mismatch between when the debt is incurred and when the revenue flows is not a minor modelling footnote. It is the central question.

Property industry groups have already pushed back on the levy schedule ². Organised resistance to the passive instruments has materialised before a single apartment has been approved. That is a reasonable early indicator of what 40 years of political and market friction looks like.

The Council Problem

The structural exposure is sharpest when you examine the planning mechanics. The SRL corridor is subject to fragmented municipal control, with individual councils retaining statutory authority over height limits, setbacks, and permitted uses. The value capture instruments — particularly land tax uplift and developer levies — are mathematically dependent on significant increases in permissible density around stations. If councils exercise their discretion conservatively, or if community opposition produces downzoning pressure, the density underpinning the revenue model does not materialise.

This is not speculation about an unlikely scenario. It is a description of how Australian planning has routinely worked. As a general pattern in Australian planning history, state infrastructure announcements have repeatedly been followed by years of precinct-plan negotiation in which councils and community groups progressively compress development envelopes. The state announces; the councils deliberate; the apartments do not get built at the density that justified the announcement.

The Asian models avoided this through institutional design rather than collaborative goodwill. Tokyu owned the land ³. MTR held delegated planning authority. Victoria has constructed a revenue model that requires specific density outcomes from a planning system institutionally configured to resist them, and the available sources do not identify any mechanism by which the state intends to close that gap.

There is also a selection-bias risk embedded in this structure. Value capture schemes structurally favour projects in wealthier, already-redeveloping areas where property markets are deepest ⁴. If a significant portion of the SRL’s projected revenue depends on realising high per-square-metre development values, then the financial logic creates pressure — not necessarily explicit, but structural — toward station locations and precinct designs that maximise land economics rather than transport utility. The SRL route-selection rationale is not examined here, and whether that pressure has shaped the project’s design cannot be determined from the available sources. The question nonetheless warrants independent scrutiny.

The São Paulo Detour and What It Tells Us

The São Paulo CEPAC model offers a partial alternative template. The city issued tradeable development certificates — essentially monetised future density rights — raising BRL 2.9 billion (approximately $806 million USD) from a single urban operation between 2004 and 2012 ⁶. It is an inventive mechanism for front-loading value capture revenue without requiring the state to hold land. But it still depends on a functional land market and strong municipal planning powers, and the São Paulo experience demonstrated a distributional failure: 33% of the raised funds reached low-income communities, and 59% went to road infrastructure serving higher-income car users ⁶. These two categories together accounted for 92% of raised funds.

The São Paulo lesson is not that innovative value capture mechanisms cannot raise money. It is that without robust legislative safeguards, the money raised can be captured by incumbent interests and redirected away from its stated public-benefit purposes. The same risk attaches to SRL value capture revenues under future Victorian governments with different spending priorities — and the available documentation does not describe any binding mechanism preventing that outcome.

The Case for Land-Banking

If passive instruments harvest too little too late, and if the institutional conditions for the Asian models cannot be reproduced in Victoria, the logical alternative is to change the state’s relationship with the land itself.

Active land-banking means the state acquires strategic parcels around future station sites before construction or rezoning announcements drive prices beyond reach. It then master-develops — or sells to developers under binding density, affordability, and design obligations — capturing the full margin between pre-announcement acquisition cost and post-infrastructure development value, rather than a levy slice of someone else’s margin.

The economics are direct. The political economy is not. Active land-banking at SRL scale requires the Victorian government to compulsorily acquire land ahead of formally declared need, to override council planning controls where necessary, and to operate as a large-scale property developer. Each of those steps invites legal challenge, political attack, and the institutional risk that a government developer replicates the patronage and efficiency problems of its private counterparts. São Paulo’s experience is a reminder that even state-directed development can be captured by incumbent interests ⁶.

There is also a critical legal question that the available evidence does not resolve: whether Victoria’s compulsory acquisition laws permit land assembly at pre-announcement valuations. If existing rules require the state to pay market value at the date of acquisition, and if announcement effects have already inflated that market value by the time acquisition proceedings commence, the land-banking margin may be substantially or entirely destroyed before the first parcel is purchased. This is not a secondary implementation detail. It is a precondition for the entire proposition, and it needs to be answered publicly before the land-banking case can be treated as a settled alternative rather than a theoretical one.

The Housing Tension the Government Won’t Name

The SRL documentation presents the 70,000 homes target as a public benefit justification for the project ⁵. The value capture revenue logic and the housing supply logic are presented as mutually reinforcing. They may not be.

Maximising developer levy revenue favours high-end apartments that inflate land values and per-square-metre prices. Genuinely affordable housing supply — the kind that addresses Melbourne’s actual housing stress — tends to suppress the per-square-metre values on which the land tax projections depend. These objectives pull in opposite directions, and the government’s documentation defines neither affordability thresholds nor income-mix requirements for the 70,000 homes figure. There is no available evidence that any proportion of those homes will be social or affordable housing in any binding sense.

This is where Hong Kong’s warning lands most precisely. World-class transit infrastructure, funded through property development cross-subsidy, produced the developed world’s most overcrowded housing conditions ⁴. Whether that outcome flowed directly from the transit-developer financing model or from the broader structure of Hong Kong’s housing and land policies, it represents a cautionary endpoint. Melbourne has a once-in-a-generation opportunity to build differently. The financial model currently being used to justify the SRL creates structural incentives to repeat Hong Kong’s mistake.

What Should Change

The Victorian government should not abandon value capture as a funding contribution to the SRL. Extracting a return from the land value uplift that public infrastructure creates is defensible in principle and worth pursuing. But the passive instruments currently on the table — projected over 40 years, dependent on density outcomes from a fragmented planning system, exposed to organised industry resistance and political attrition — are not reliably sufficient to meet a $11.5 billion commitment.

Three changes would materially improve the model’s integrity.

First, publish the NPV assumptions. The 40-year, $11.5 billion projection needs to be accompanied by the discount rate, sensitivity ranges, and density-realisation assumptions underpinning it. Without this, the figure cannot be assessed against the borrowing cost it is meant to service. A government confident in its modelling should have no objection to releasing it.

Second, resolve the compulsory acquisition question and legislate land-banking powers. If Victoria’s acquisition laws do not currently permit pre-announcement assembly at pre-announcement prices, the government should seek legislative reform. The design of a dedicated SRL land acquisition authority — insulated from day-to-day political interference, with a clear mandate to acquire, hold, and master-develop station precincts — would provide both a more reliable revenue stream and genuine planning control over density and affordability outcomes. This is operationally harder than levies. It is also more likely to work.

Third, attach binding affordability obligations to the 70,000 homes figure. If the Commonwealth’s $11.5 billion contribution comes with conditions — and the available sources are almost entirely silent on this question — federal affordable housing requirements may provide an override mechanism that the state has been unwilling to impose on councils itself. The Commonwealth should use that leverage if it exists. The SRL’s housing-supply rationale is substantively hollow if it translates entirely into market-rate apartments that maximise land tax revenue while delivering nothing to households priced out of Melbourne’s existing suburbs.

The Tokyo and Hong Kong models succeeded because their operators controlled the land and the planning. Victoria’s SRL is attempting to replicate their financial outcomes without replicating their institutional conditions. The gap between those two things is not a communications problem or an implementation challenge. It is a structural design failure, and the time to address it is before construction begins and the political economy hardens around the existing model.

This piece was researched and drafted with AI assistance and reviewed and edited by a human before publication.