Sydney Closed a Railway for 25 Months. When Do Passengers Get Their Time Back?

Sydney's Southwest Metro is due to open on 30 October 2026. For passengers along the former Bankstown railway, it promises a train every four minutes during weekday peaks, more capacity and faster connections across the city. But the improvement has had an unusual price: the operating railway between Sydenham and Bankstown closed on 30 September 2024, leaving passengers on replacement buses and other routes for 760 days. [1][2]

That raises a question the project's promised capacity increase cannot answer. How long must the new railway operate before the time it saves passengers exceeds the extra travel time they spent getting around the closure?

In one transparent illustrative scenario, our model accumulates 7.87 million passenger-hours of additional travel time during the closure. It then produces 1.63 million passenger-hours of net savings in the first full operating year. Under an assumption that comparable travel demand grows by 1% annually, the cumulative account crosses zero in July 2031. This is a scenario result, not an observed break-even date or a finding that the project has paid for itself financially.

Two different comparisons, one railway

Calculating a passenger-time payback requires a consistent reference. During construction, we compare journeys on replacement transport with the journeys passengers could have made if the former T3 railway had remained in service. After opening, we compare the new metro with that same continuing-railway counterfactual.

Imagine a journey that would take 40 minutes on the old railway, 60 minutes using the closure-period alternatives and 30 minutes on the new metro. Each closure-period journey imposes 20 additional minutes. Each journey after reopening saves only 10 minutes. The saving is not the 30-minute difference between the replacement journey and the metro journey: once construction ended, the old railway would no longer have been replaced by buses in the counterfactual.

There is another complication in Sydney. The Metro City line between Chatswood and Sydenham opened in August 2024, before the Bankstown conversion began. Some passengers could already have combined the former T3 service with the City Metro at Sydenham. The Bankstown extension cannot be credited with benefits already delivered by that earlier investment. [3]

Turning travel changes into a time account

The available public evidence provides an operating benchmark, not a complete passenger history. In December 2025, Transport for NSW reported that Southwest Link's three bus routes were carrying around 40,000 passengers daily, with more than 16 million passenger trips since September 2024. These are bus boardings, not an origin-destination matrix of affected journeys or a count of distinct individuals. [4]

Our first model therefore makes its assumptions explicit. It applies the 40,000-a-day reference throughout the 760-day closure, yielding 30.4 million modelled bus boardings. It assumes that 85% correspond to journeys comparable with the former railway, and adds other affected journeys equal to 20% of that group. These proportions have not been measured. The six illustrative journey groups produce a weighted 16.68 extra minutes per eligible bus journey; the additional journeys are assigned eight extra minutes.

That construction creates the 7.87-million-hour time debt. It is composed of approximately 7.18 million hours on modelled eligible replacement-bus journeys and 0.69 million hours on other modelled affected journeys. These are calculations from a provisional demand bridge, not Transport for NSW estimates.

For the period after opening, the model retains 90% of the comparable demand, or 36,720 journeys per day, and uses an illustrative 7.30-minute net saving against the continuing old-railway network. That produces about 1.63 million hours saved per full year before demand growth. The weighted saving is deliberately smaller than the unadjusted published journey improvements because the model attempts to avoid crediting the Bankstown conversion with benefits from the pre-existing City Metro. Those adjustments still need timetable-level validation. [1][3]

The same set of scenario assumptions underlies all four results. These are not observed passenger-time totals or statistical confidence limits. Source: Rokin Research model v0.1, using NSW Government and Transport for NSW benchmarks. Evidence cut-off: 28 October 2026.

Figure 1. Illustrative cumulative passenger-time account. Scenario assumptions and source information are printed in the figure.

Why the answer could move by years

The arithmetic is simple; its inputs are not. We do not yet have a complete public station-to-station dataset showing where former T3 passengers travelled during the closure. The replacement buses were free, so they did not create the usual Opal tap history. Published service times also describe schedules, not necessarily the traffic, waiting and transfer times passengers experienced. [2][5]

The model's alternative scenarios are deliberately wide. If closure penalties were lower and post-opening time savings higher, payback could arrive materially earlier. If closure delays were worse, new-metro benefits smaller and comparable patronage weaker, it could arrive years later. The model's flat-demand illustrations span approximately 2029 to 2035 under selected inputs; this is not a probability range and should not be described as a forecast interval.

The next measurement task is therefore specific: obtain pre-closure origin-destination volumes, monthly Southwest Link boardings and alightings by route and stop, observed replacement-bus journey times and, once available, actual metro journey times and comparable passenger demand. The case will be strongest if it can show not only the system-wide crossing, but also which station pairs repay their time debt quickly and which do not.

What the break-even means for infrastructure investment

A railway conversion can create permanent capacity and reliability gains, but some passengers pay the transition cost years before those benefits arrive. A time-debt account makes the timing visible. It can help compare construction staging, the quality of temporary services and the value of completing a project sooner, provided the alternatives are technically feasible.

It is not a full social cost-benefit analysis. The account excludes construction expenditure, operating costs, crowding, changes in accessibility, newly generated trips and many other effects. A metro could have a strong overall investment case even if the passenger-time debt takes longer to repay, or vice versa. Nor does recovering passenger-hours prove that the people who experienced the closure are the people who gain most afterwards.

So when will passengers get their time back? We cannot yet give an empirically measured date. July 2031 is what one clearly labelled scenario produces, using the published operating benchmarks and assumptions that need to be replaced by journey-level evidence. The more durable result is the method: a new railway earns back the time cost of its construction only through incremental gains over the service that would otherwise have existed.

Research note. The analysis is a prospective, undiscounted passenger-time scenario prepared on 28 October 2026, before the scheduled metro opening. Its main demand, origin-destination mix, closure-time penalty and net post-opening saving are not directly observed. It does not establish an actual payback date or a financial return on the project.

Sources

[1] NSW Government. Southwest Metro to open Friday 30 October 2026 (22 October 2026). opening announcement; service frequencies, journey forecasts and capacity. Official source

[2] NSW Government. T3 Bankstown line to close from Monday (25 September 2024). start date, SW1/SW2/SW3 service pattern and free bus travel. Official source

[3] NSW Government. Historic day for Sydney as Metro comes alive under the city (19 August 2024). City Metro already operating to Sydenham before conversion. Official source

[4] Transport for NSW. Southwest Link update (December 2025). page 3; ~40,000 boardings/day and >16 million cumulative bus trips. Official source

[5] NSW Parliament. Question 106480: Southwest Link SW2 weekday AM journey times (13 January 2026). mean timetable journey, Bankstown–Sydenham, weekday 06:00–09:00; scheduled, not observed. Official source