Airports & aviation

Oct 31, 2017

What’s the problem airport rail would solve?

There needs to be a clear and focused justification for spending billions of dollars on building a rail line from the CBD to Melbourne Airport

Alan Davies — Editor of The Urbanist

Alan Davies

Editor of The Urbanist

Trips generated by Melbourne Airport air travellers are concentrated in the CBD and inner suburbs

The Committee for Melbourne held a seminar last week on the perennial idea of building a mass transit link between Melbourne Airport and the CBD. I was a panellist in the first session, which considered the question:

What is the problem we are trying to solve?

That’s an excellent way to think about the topic. I’ve previously argued we must look at access to the airport from all parts of Melbourne and Victoria, but I’ve no problem with focussing on a link to the CBD as it’s bound to be a primary component of any solution. After all, we know from previous studies that airport trips are heavily concentrated on the city centre and inner areas.

If a private investor wants to pay the full cost to build and operate a mass transit link, then I think the question answers itself. But there’s no sign that’s going to happen. The privately financed Sydney and Brisbane rail lines both struggled in their early years so it’s likely government will be called on to pay, in one way or another, the full cost of the link.

For the purposes of this discussion, I’m going to assume the solution is a heavy rail line. There are various route possibilities, although the leading contender must be the at-grade alignment via East Albion that was reserved some years ago. The likely cost, assuming an elevated station at the airport (like Brisbane’s), is circa $3 Billion. The Andrews Government favours connecting it with Melbourne Metro rather than terminating in the CBD at Southern Cross station.

That scale of expenditure demands a focused and precise understanding of why we should build it. There’s the inevitable danger the rationale gets clouded by populist appeals to non-central concerns. In my view, the immediate “problem” to be solved is to move travellers from the airport to the city centre. At present, that’s mostly done by taxis, private vehicles and SkyBus.

A rail line could be justified if one or more of the following conditions apply:

  • There is inadequate capacity to move passengers between the airport and the city centre
  • The travel time is unacceptably long
  • The travel time is excessively unpredictable or unreliable.

Of these, capacity is the key justification. That’s because SkyBus could be engineered to provide fast and reliable trips, at lower cost than building a rail line, if it were given dedicated road space as recommended by Infrastructure Victoria.

But while SkyBus could increase frequencies and the size of vehicles in order to carry more passengers, there’s a limit to this technology. There’ll inevitably come a time on the Airport’s current bullish passenger projections when buses will no longer cut it on such a busy route.

Infrastructure Victoria thinks a train won’t be required for at least 15 years. Perhaps, but demand changes so it might be sooner, or it might be later. It’s therefore pleasing the Commonwealth and Victorian governments are starting to plan for the possibility of a link.

There’s no shortage of other arguments thrown into the public arena to justify building a rail link. The main ones I hear most are:

  • It would reduce traffic congestion on the motorway network
  • It would moderate airport parking charges
  • It would be city-shaping infrastructure e.g. it would promote development of a more polycentric urban form.

I think these are secondary considerations that are in any event doubtful rationales for such an important investment.

The phenomenon of latent demand means a rail line won’t produce a sustained reduction in traffic congestion any more than building or widening a motorway will. Any road space liberated by motorists who divert to the train will eventually be taken by other motorists, most of whom aren’t going to the airport. The value of an airport train lies in providing a better alternative to driving, not in making travel by car easier. The way to deal with traffic congestion is to reduce demand for driving in the first place e.g. by road pricing.

Nor will a train have much impact on parking fees at the airport. The mode share of the train in Brisbane is just 7% and in Sydney it’s 17%; that still leaves a lot of travellers looking for a car park. High parking fees are mostly the result of monopoly pricing practices and should be addressed by regulation. Indeed, charges would increase significantly if airport parking were used as an indirect way to price road use.

The claimed city-shaping benefit of rail lines is uncertain. Transport systems have a generalised impact on land use at the metro level, but it’s much less assured at the level of specific lines or locations. There are a lot of stations in Melbourne that even after more than 100 years have virtually no development around them. Research shows transit doesn’t necessarily lead to the increases in land value, or the greater development, implied by grandiose “city-shaping” claims.

In any event, the two planned intermediate stops on the Albion route, Footscray and Sunshine, are already on rail junctions. Yet Footscray is only the eleventh largest activity centre in the suburbs and Sunshine the twenty sixth. Together, they account for circa 0.6% of all jobs in Melbourne. It’s worth thinking about how many firms would choose to locate in these centres, and thereby forego the agglomeration benefits of the massively larger city centre, just to gain a few minutes in airport travel time.

If the route took in large potential redevelopment sites at the former munitions site at Footscray and/or Essendon Airport, this would be a stronger argument, but by no means an assured one. There would be constraints on the number of dwellings that could be built on these sites and, in the case of Essendon, an alternative location – inevitably with lower aviation benefits – would need to be found for its airport function.

The longer route and/or tunnelling required if the line is to do “double duty” would add to the capital cost and/or increase total trip time. Both of these factors could weaken the Benefit-Cost ratio. Any increase in travel time due, say, to a longer route or more stops, makes the train less competitive relative to other modes and may weaken patronage. This is especially problematic as time saving is by far the dominant benefit in transport projects.

These sorts of issues, plus financing options like value-capture, have a place in assessing the case for airport rail but they’re secondary; they’re not the key reason why a rail line should be built. The benefits and costs associated with them are relevant but should be considered in subsequent phases when choices are made about options e.g. one route vs another.

The decision on if and when some form of mass transit link should be built (almost certainly rail), should turn on fundamental transport criteria like capacity, trip time, and trip predictability.

See other recent articles on the subject of airport mass transit links:

(Visited 97 times, 1 visits today)


Leave a comment

Your email address will not be published. Required fields are marked *

33 thoughts on “What’s the problem airport rail would solve?

  1. JohnC

    “Infrastructure Victoria thinks a train won’t be required for at least 15 years.”
    They along with some commenters here have not had a taste of gridlock Melbourne airport style at 6.30am or almost any time… A perfectly functional airport in the 80s but no longer fit for purpose. Our airport is our face to the world and it is no longer a pretty one.
    Never mind the sight of jet lagged visitors being abused by fat f***s in Ill fitting uniforms for not following their taxi queue directions. We need a train line for all sorts of reasons many of them intangible.

  2. Sean Deany

    The InterCity and AirTrain proposal put forward by Rail Futures recently is a good option as it separates Metro rail operations and gives a realizable scenario for future proofing the Melbourne region for a given HSR corridor. However I also believe that an Orbital LRT network with an Airport stub has an important role to play where it can be delivered sooner and would also allow for needed employee access to the precinct(s) on the approaches to the airport.

  3. hex

    Spot on Alan.
    The Airport rail link is a ludicrous waste of money, and the Age should start looking at the actual business case (and the cost!) before continuing with its simplistic campaign.

  4. meltdblog

    There are already vehicle pricing mechanisms in place at the airport, Taxis pay a surcharge and parking rates are easily varied. Extending pricing to all vehicles entering/exiting the area would be easy enough to implement though likely unpopular. The free to use private vehicle pickup area is (overly) convenient.

  5. Russell Pollard

    It needs to terminate in the city of Melbourne itself at Southern Cross. And taxi access needs to imptove from that point. If I come from overseas I do not want to have to change platforms at Footscray before finding Melbourne. I want to get to the city where I have booked accommodation, the place in the travel guides. There is a reason people travel between cities! Lots of reasons.

    And if I am going to Melbourne Airport from Mernda or Frankston or Toorak, spare me a trip to Footsceay or Sunshine. Like all Melbournians I can manage a trip to the centre of Melbourne without getting lost.

    There are all kinds of arguments but for me the most compelling is the one about daring to get it right in the first place so that it’s remains a useful part of Melbourme’s infrastructure for the next century. India is putting in rail lines at enormous cost . . . part of that cost being the failure of the British to do it many decades ago. The US neglect of rail us also coming home to roost. Mass transit and rail as a safe, comfortable and efficient part of it, allows a city to function as a city.

    Not everythimg that poses as new infrastructure gets an automatic pass just because it was preceded by a hodge podge mess. The NBN is a great example of what was a bare pass but with potential, until someone did a short term business case for it and turned it into a massive infrstructure fail getting worse by the day.

    Planners have a responsibility to think way ahead and spare us the lemons that are so easily collected along the way. This means that the business case and the short term business cycle should not dominate. In the capital city that is in the state that is marketed as the sum of its parts, airport and the trip to and from it need to reflect quality . . . a quality part to a quality holiday or business meeting or international conference. Let’s not shoot ourselves in the foot pandering to every profiteer’s argument to skew a rail line along their preferred business model. Let’s just get one that people speak about because it is efficient, clean, safe, cheap, easy as buggery to use and a feature of travel to and from Melbourne.

  6. Airport

    “There are a lot of stations in Melbourne that even after more than 100 years have virtually no development around them”

    Does the zoning permit development around them?

    “The longer route and/or tunnelling required if the line is to do “double duty” would add to the capital cost and increase total trip time.”

    Both Brisbane and Sydney trains to the airport stop at intermediate stations. Although it does reduce trip time to the airport, it also collects patronage from all other stations, which should improve the BCR because more people can use the service and make a time saving from their home station (combined markets).

    Indeed, for the Brisbane Airtrain operations, that train used to run express past a number of stops. That has now been changed to an all stops Airport service.

    I still think that light rail options to the airport should be considered instead, that would plug easily into the CBD tram network and could potentially allow the AirTram to do a loop of the CBD picking up passengers.

    1. Airport

      Correction, meant to write “reduce trip time SAVED”

    2. Alan Davies


      I expect all those under-developed stations haven’t had the right zoning, but that’s my point; there’s much more to it. You can’t assume a railway line will automatically induce development when there are huge obstacles like residents who oppose development.

      The number of stops is a trade-off between patronage and time. It’s hard to compare one line with another because of unique circumstances like the competitiveness of other modes. I note though that there’re only two stops between Sydney Airport and the CBD, both of which are major redevelopment areas. Brisbane’s train has four stops between the airport and the Valley, although it performed poorly for many years and even now only has 7% mode share. Perhaps the owners have changed their strategy?

      The problem with light rail is speed; patronage would be low because other modes would be faster most of the time.

      1. Airport

        Yes, but almost every other rail line in Melbourne that does not go to the Airport has no problem getting decent patronage. The total patronage from all sources – not just the airport – is what the project will be assessed on.

        A train may only spend 60 seconds at a station, so at four stops that might be the difference between a 30 minute (only Airport passengers) and a 35 minute journey (airport passengers PLUS the patronage generated from four stations).

        In addition to this, if four stations to Melbourne Airport are included then any private partner only needs to pay for the *increment* of rail after the last “public” station on the line, not the whole line. That would mean lower infrastructure cost (for the private proponent), possibly lower cost for tickets as well.

        7% mode share in Brisbane – the line is profitable and paying for itself. In addition, it is a miracle really given that labour costs are the highest components of operational delivery and both Sydney and Melbourne require two staff to operate the train. Melbourne would only need one staff to operate the train.

        Light rail can work just as the bus can work. Speed is determined by stop spacing, not vehicle type, and also by the level of priority given to the vehicle (mixed traffic, own lane, own corridor). LRT would be able to run in a more direct route in the median of the Tullamarine Freeway (either at surface or elevated or a mix), which would give it a distinct advantage over the current rail proposals that suggest to repurpose the Jacana freight line.

        1. Alan Davies


          Why light rail? If it’s laid in the median from the airport to the top of Royal Pde it’ll cost a lot to build because much of it would need to be in structure. Moreover, it would require removal of one lane in some parts of the route to provide space for supporting columns. There are some engineering challenges too e.g. how do you deal with the existing columns in the median that support bridges?

          The route study done by the Napthine government for heavy rail assumed a 30-minute trip to CBD North in the base case, with stops at Sunshine, Footscray and Parkville. SkyBus says it’s service takes 30 minutes off-peak. I can see that light rail would be fast along the motorway sections but once it “plugs easily into the CBD tram network” it’s speed will be limited by other trams and intersections.

          I don’t think it’s obvious that light rail would be faster or cheaper than heavy rail. On the other hand, we do know it wouldn’t have anything like the capacity of the base case (which links to Melbourne Metro) or the reach into Sunshine, Footscray, Parkville, Swanston St, Domain, and stations on to Dandenong.

          From what the officials are saying, the Federal/State study kicking off now will look at every possible option, including light rail. So that should give us a more definitive picture of the virtues of the various options.

          1. Airport

            Light rail is great because it is direct and can take advantage of the Tullamarine freeway alignment. You would need the train to run at V/Line express speeds to reach Melbourne Airport along the proposed freight spur line.

            The engineering requirements for light rail and even busway are less stringent than for heavy rail. Slower speed in the CBD is not so much of an issue as it has to pick up passengers near where they stay.

            Remember, to get to Southern Cross station, passengers need to walk or catch a connecting tram to the station anyway. Have planners accounted for that properly, or did they just assume that people just appear at the station immediately after leaving their hotel or workplace front door? So the idea that the train would be faster might not be true.

            Indeed, the City Loop has trains loop around the entire Melbourne CBD wasting several minutes for all passengers on board because planners decided that it would be better to pick up workers near where they worked and waste a little bit of time on the overall journey, rather than have them walk or catch a tram to collect at one or two very large stations.

            The image you linked showing columns supporting a bridge isn’t an obstacle to light rail. The image shows a four lane motorway, so the solution is obvious – take away the right most lane away from cars and give it to the LRT.

            I agree with you that more clarity will appear when ALL options are considered in a proper study.

          2. Airport

            PS: I also notice that in the same Google Maps image that there is a sign that says “City Tulla Widening Project”. I can also see an additional lane being constructed for cars.

            So the cars are already getting an added lane to their freeway, a perfect opportunity to do a 1-for-1 swap and remove the car lanes closest to the centre median for a LRT or even busway service.

  7. Peter Hill

    The problem of justifying a high QOS rail link between the CBD and Melbourne Airport is familiar to me, having conducted a feasibility studies of the Bracks government link policy in 2001. Alan Davies has got it right. Another concordant insight is given by transport planner Jarrett Walker in his article “Keys to Great Airport Transit” published on the website “Human Transit” at
    Another question is how can trip time reliability of the Skybus service be improved (by dynamic ITS-based motorway lane prioritisation or reservation methodologies, and a cost much lower than the CAPEX for a high QOS rail link? (Hint: it can be done)

  8. Horst (Oz) Kayak

    Any discussion on a liveable urban environment needs to include the performance indicator covering travel time patterns between most of the intensely used activity centres in the urban region being modelled.
    The activity centre approximations in their aggregated form need to primarily reflect the travel patterns of the people living in and visiting the region.
    Few people would disagree that the CBD and MEL airport would be in the top ten origin and destination category in any transport system models covering the 31 LGAs of the MSD.
    Therefore the first associated question on possible and practical politically acceptable transport mode provisions should be, “is the airport required to be reliably accessible to 80% or less of the MSD population within 20 to 30 minutes?”

  9. Woopwoop

    Trip predictability is a real issue.
    On a Monday morning, traffic entering the airport can be at a gridlock, causing may minutes delay and probably causing some to miss their plane.
    Of course, it’s not helped by the constant stream of people in front of the terminal crossing the pedestrian crossing to the car park.

  10. Lyndsay Neilson

    As Secretary for the Department of Infrastructure 2004-2007 I recommended against the Airport Link on this same basis of costs vs benefits, and recommended the investment in upgrading SkyBus that the Government then financed. Far more cost-effective. With the potential in the near future for electric-powered autonomous vehicles to add a new MAAS (Mobility as a Service) option we need to revisit the issue from a technology perspective as well. Lyndsay Neilson.

    1. Peter Hill

      Lyndsay, I agree. That was my conclusion when I conducted feasibility studies of the Melbourne Airport rail link in 2001 for Skybus. Road-based AV technologies offer the potential for more diverse forms of bus public transport at very low operating cost, e.g. demand-responsive small and medium shuttles feeding into existing metro rail stations or “semi-express” Skybus-type line-haul bus services to the Airport (and other concentrated trip generating zones)

    2. Chris

      “With the potential in the near future for electric-powered autonomous vehicles to add a new MAAS (Mobility as a Service)”

      So basically your plan is do fuck all and just hope it works out?

  11. Anthony

    Just a thought – but if this airport link were to eventually include the train lines from Bendigo, Shepparton and Albury, it could provide even more justification in terms of capacity as these regions, while not that large now, could be helped in their development to take more pressure off Melbourne.

    1. Alan Davies


      That’s essentially the Rail Futures Institute proposal I discussed briefly here. It would require building a lot of extra track to route regional lines via the airport. It also hasn’t been established that diverting metro growth to regional dormitory cities would bring net benefits, but even if it would, they’d mostly lies a long time in the future i.e. discounted. There should be some benefit from removing regional trains from metro lines, but how big that would be isn’t clear. Also important in assessing these sorts of ideas is whether the projected growth over such a long time frame (and it is a projection) will materialise according to the assumed timetable.

      The Rail Futures Institute proposal also requires construction of a tunnel exit from Southern Cross station and extension of the metro electrified network to Clarkefield and Wallan. The latter can be viewed as an opportunity to extend the size of the metro network, or as a device to deflect criticism that sections of the existing country rail line would be a stranded asset. Relative to the base-case (i.e. connecting to Melbourne Metro) this option doesn’t have the benefit of taking passengers to Flinders St station or providing a one-seat journey between the airport and stations on the Cranbourne/Pakenham line.

      1. Tom the first and best

        The Clarkefeld-Wallan proposal has the benefits of serving the airport with luggage friendly trains, providing single seat journeys to the Bendigo and Shepparton lines (and if it allows standard gauge, the Albury line as well) and separating the Bendigo and Shepparton lines from the suburban network.

      2. John.s

        Bendigo/Seymour makes sense. A RRL 2.0

        VLine trains are more suited to Airport rail journeys, with luggage racks, as well as providing the government with an easier sell for a more expensive fare.

        This also benefits the Sunbury and Craigieburn lines, no longer sharing trains with VLine services, leaving only the Dandenong lines to share with Gippsland trains, something which should have been taken care of with CD9.

        Southern Cross v Flinders St? Does it make a huge difference? Once trains start to be through-routed following the City Loop reconfiguration, I don’t see this as an issue. All trains will go via Footscray or Southern Cross. Including all VLine trains.

        1. Alan Davies

          One of the earlier studies found extending service to Flinders St significantly increased patronage.

          1. John.s

            Interesting. As someone who grew up in the Country, Spencer St has always been my “launching point” into the city, however since moving to living near the city I have noticed how much popular Flinders St is, however did not think it would be so drastic. With part of the old WTC now demolished, would we see a 4th track pair viaduct built between Flinders St and Southern Cross for VLine services?

  12. Jason Murphy

    Another good piece.

    A second way to frame the question might be to list great big destinations in Melbourne that aren’t served by rail and rank the airport among them.

    Should we serve the airport with rail before Monash University? Before Chadstone shopping centre? Before Doncaster?

    I fear the airport rail line will be built simply because other cities have one. But if we can leverage that to show that other destinations are even more deserving, that will be useful.

    1. Tom the first and best

      For an effective Chadstone rail link, the Alamein line needs to be extended to Oakleigh, via East Malvern and Chadstone. This means a tunnel from Waverley Rd to Oakleigh ($$$). This means a long lead time, due to all the extra planning and preparation a tunnel under private land requires.

      Monash is relatively easy once the Metro Tunnel opens in 2026. Just rebuild Huntingdale with sky rail (vastly improving that pedestrian and passengers hostile mess) and have sky rail along the median of North and Wellington Rds. It could even go further than Monash as the wide median continues well into Rowville.

      Doncaster is easy to get rail much closer to, via median strip rail to Bulleen, but the final stretch of actually getting the railway there requires tunnelling.

      1. John.s

        Does Monash need to be heavy rail? What is Stopping an extention of Route 3 via Princess Highway (removing the short stub on Waverly Rd), via Chadstone SC, run it down Huntingdale rd and up North Rd to Monash, if not rowville. At the same time, Alamein line could be converted to Light rail, and connected at the old railway crossing between Hughesdale and East Malvern on Princess Highway and routed to Monash Uni as well, providing another link from Glen Waverly line to Chadstone SC and Monash.

        Post Metro Tunnel 2, Williamstown to Footscray could also be converted to Light rail.

        1. Peter Hill

          Better still, convert the Alamein branch rail line to true high-speed light rail (HSLR), then extend beyond Alamein, via East Malvern, then (via an admittedly conventional tramway section) along Belgrave Rd (joining to it an extension of tram route 3), then past Chadstone Shopping Centre via Dandenong Rd, Warrigal Rd, thence past Oakleigh station as HSLR past Huntingdale station, then along North Rd and Wellington Rd (in median reservations), past Monash University, then all the way to Stud Rd shopping Centre. This proposal (dubbed the “Monash Light Railway”) is featured in
          the TCPA Bulletin Vol. 18 no. 1 of 2011 and published in the TCPA website at

          1. Tom the first and best

            Heavy rail has faster trips and higher capacity. Converting the Port Melbourne and St Kilda lines was a mistake. The same mistake should not be made for the Alamein line.

            Heavy rail would be more competitive for the Alamein line-Chadstone passengers and Pakenham/Cranbourne and Glen Waverley lines-Glenferrie (Swinburne)/Camberwell/Box Hill passengers and is more competitive for Alamein line-Glenferrie (Swinburne)/Camberwell/Box Hill passengers.

            1. Peter Hill

              Tom-T-F-A-B, what is your evidence for your claim that is relevant to the middle-to-outer eastern suburban corridor between Camberwell and Rowville via the Monash region? Also, consider the severe geometric and thus infrastructure requirements , thus CAPEX required to retrofit heavy rail into this corridor. For the spatial trip densities expected in a low-to-medium urban density region, underground or overhead rail infrastructure would not be justified.

              1. Tom the first and best

                I shall split this into the 2 two separate sections.

                Of Camberwell-Oakleigh, Camberwell-Alamein already has heavy rail (with only Ashburton-Alamein needing duplication), Alamein-East Malvern already has a rail corridor (some serious work around the Freeway would be needed) and thus only East Malvern-Oakleigh needs a tunnel and that tunnel would connect 3 lines to a major centre of employment, shopping, service, leisure, etc. and that would directly a significant proportion of the people who use Chadstone as well as people travelling from the Between the Belgrave/Lilydale and Pakenham/Cranbourne lines for work, study (especially Swinburne and Monash) and other purposes (with the Merto tunnel diverting the Pakenham/Cranbourne line away from Richmond, the rail trip via the inner-city will lengthen significantly). Parking levies may need to be applied in several locations (Chadstone, Box Hill Camberwell, Hawthorn (including Swinburne), Richmond, Oakleigh) to create the mode shift to justify the infrastructure.

                The Rowville corridor has the Monash-Mulgrave employment precinct and the outer sections would be little different to the Mernda rail extension.

    2. Alan Davies

      Clayton/Monash is by far the largest concentration of jobs in the suburbs, well ahead of Tullamarine and Kew/Hawthorn. It’s developed without good rail access. There’s a prima facie case for investigating the case for building a loop/spur off the Dandenong line, or for improving feeder services to the main line. It’s spread over a large area so two or three stations, including at the university, might be necessary.

    3. Peter Hill

      Jason, a good point by you. The popularity of the airport rail link idea, and the willingness of politicians to latch onto it for electoral popularity, is what I view as a “temple-building” syndrome – with other people’s money (taxes). Build the temple, and then the masses will bow down reverently in awe.

Leave a comment

Your email address will not be published. Required fields are marked *

Share this article with a friend

Just fill out the fields below and we'll send your friend a link to this article along with a message from you.

Your details

Your friend's details