The ARC’s most recent Transport and Urban Development Committee meeting looked at the options for extending some services on the Western Line to Huapai. After the unfortunate (but not altogether unforeseeable thanks to the slowness of our trains) failure of the Helensville rail service, it seems as though some more realistic options are being looked at to extend some rail services from Waitakere township out to Huapai – a growing settlement to the northwest of Auckland.

The ARC report outlines a number of different options as well as their cost, for running trains to Huapai:
First things first, I think it’s very promising to see that we’re not likely to go with the failed Helensville model of only one train a day each way. That unfortunate experiment gave people only one option, and only catered for one type of passenger (someone working in the CBD). Having a number of different services should give people more confidence to use the service, as they know they have a range of options to get home.

As shown in the table below, adding more services vastly increases the number of people likely to use the train – because of the greater flexibility offered. It would be pointless to go to all the trouble to just attract 26 passengers, as estimated by Option 1:
As you can see, all the options require a fairly hefty subsidy (the costs are net of income from fares). But remember at the same time these are likely to be pretty long trips – so the benefits through reduced congestion are likely to be at the higher end of the $17 peak hour benefit to road users that typical rail trips provide. Taking that into account, the proposal might well have a positive benefit-cost ratio – and it would be interesting to see such an analysis undertaken.

The ARC report discusses the merits of the different options, and comes to a conclusion that I find myself agreeing with:
However, one thing that requires careful consideration is the fact that the Western Line has now lost its express train – which used to be non-stop between Newmarket/Grafton and New Lynn. This express train was the Helensville service, when that service operated. The lack of such a train means that rail trips from Huapai to Britomart are likely to be well over an hour in length – potentially reducing its attractiveness to commuters. It makes me wonder whether the decision to remove the Western Line express train was somewhat premature.

Looking through the minutes of the meeting, there is an interesting addition that I think deserves comment.
Resolution (c) adds an interesting new idea into the mix, the idea of a rail service that enables west-to-south trips and doesn’t add to the growing problem of a maxed-out Britomart in terms of its train capacity. As I have noted previously, by early next year (once Western Line trains are bumped up to 10 minute peak frequencies) Britomart will be at capacity, so it will not be possible to add any further capacity to any of the rail lines (in terms of more trains, we can of course still make the train longer) until we build the CBD rail tunnel. However, a Huapai to Tuakau (or wherever) service that doesn’t enter Britomart doesn’t add to this problem, so could be in addition to the 10 minute frequencies on the existing lines that we will be having by early next year.

It would be interesting to see how many people do change between the Western and Southern lines at Newmarket at the moment, to get some sort of gauge on the attractiveness of such a service. With a lot of workers in the west and a lot of jobs in the south (and vice versa to a lesser extent) I imagine there is the potential for such a service to be useful. But, as I noted above, the real benefit is that such a service can add capacity to the Western and Southern lines even when Britomart has been maxed out. With a train every 4-5 minutes between Newmarket and Britomart (western line plus southern line plus Onehunga line services) and an integrated ticketing system that doesn’t require extra fares to be paid, perhaps the thought of changing lines at Newmarket won’t be so horrible either.

It’s good to see some thought going into the long-term future of Auckland’s rail services, and a bit of creative thinking into how we might get around the Britomart capacity problem in the medium term while we build the CBD rail tunnel. It will be interesting to see whether this proposal advances.

Share this


  1. I’m not sure we can make the trains much longer… I read that Britomart can’t take trains longer than 6 carriages. Basically Britomart can only handle 2 carriages per minute in its current form I believe. Therefore,in about 3-4 years it may be at passenger carrying capacity- so even if CBD loop construction was starting now, we should be worrying.

    Someone did suggest that because of the EMU’s higher speed and acceleration, they would increase Britomarts capacity slightly…can anyone confirm this?

    1. You’re right that it will be difficult to make the trains much longer than 6 carriages. However, my understanding is that a lot of the trains currently being operated are four carriages, with some even being only two carriages long. I think that it’s only about every second Western Line train that is six carriages long at the moment – meaning a lot of possible extra capacity in the system by making all the trains that length.

  2. Another way to add peak capacity is to press the Strand station back into service and loop trains around it, perhaps to past Orakei. This would separate out the Newmarket-destination traffic (from the south), provide traffic along the Eastern Line with a more direct connection into Newmarket, and provide a station for the new build near the Strand area.

    1. Not a bad idea, actually, setting up a southern-to-eastern link that doesn’t require travelling to Britomart or Otahuhu to get on board. The tracks are already there, and usable (I was on the first train out to Onehunga, and we boarded at the Strand on a train that came through from the waterfront end), so it’d be hundreds-of-thousands (if that) rather than millions to build a loop connection between Southern and Eastern. Terminate services at Otahuhu, since that’s where the Eastern and Southern reconnect, and you’re starting to get a moderately flexible train option for getting between the inner Eastern suburbs and Newmarket/Parnell.

  3. We should at the very least be ordering more SA carriages, something which isn’t being done.

    How frequent will the Swanson-Waitakere shuttle be? Is it planned that that will only be hourly at peak?

  4. It’s very good to see this thinking. Option 3 is a great start.

    I’ve often looked at the tracks running past the centre of rural industrial Kumeu (twin suburb of Huapai) and thought that patronage to there would be greater than patronage to Waitakere and would be a bit of a no brainer. The disappointment of the announced schedules of the designed to fail Helensville trains around here was palpable.

  5. The fact that Helensville only have about 2500 people in it was always going to be very hard to make the extra 30km on the round trip wothwhile just to service the one town. By contrast the Kumeu-Waimaku strench is home to about 8000 people and is a lot closer to the existing terminus.

    Option 3 sounds like a good idea, and they can step it up if it proves popular.

  6. Do we really want to subsidise and encourage sprawl by running public transport out to areas that are still mostly rural? I think an urban development strategy should encourage increased density and improved services to the metropolitan area, rather than building on farm land.

  7. Well it really depends on what you consider sprawl. Kumeu-Huapai and Waimaku are relatively self contained urban areas many kilometres away from the Auckland conurbation.

    Personally I wouldn’t call directing growth into separate towns beyond the urban boundary with a rapid transit link to be classic suburban sprawl by any means. That is how London has managed it’s growth over the last half century and it is something Auckland should look at more strongly.

    1. I’m defining “sprawl” here as meaning you’re encouraging people to live in a rural area while commuting in to the city. And this causes a town to develop in the rural area where none existed before, replacing farm land or vineyards or whatever is out there at the moment and requiring roads, sewerage, and power pylons. Also I suspect that they’ll be just another suburb of Auckland in 20 years time.

      Using your London example, places like Watford and Harrow were small villages before someone ran the Metropolitan line through them. And London metro area essentially encompasses half of SE England when you include the commuter belt towns that are almost physically attached to the city, like Slough or St Albans.

      1. It’s more like encouraging people to live in a small urban area near the city rather than making the city bigger. There are already towns at Kumeu, Huapai and Waimaku, but yes I’m suggesting allowing those to get bigger while remaining separate towns is a better idea than just sprawling the city out to engulf them.
        So obviously planning comes in to play to ensure that urban growth is only allowed in the actual villages and towns, while exurban sprawl remains restricted and the urban growth boundary of Auckland stays put.

        On London that was exactly my point, although I was more talking about the towns outside the M25 than those within it. The huge megacity of London has expanded by directing growth into descrete satellite towns and villages rather than simply expanding the one urban area as an enormous amorphous blob. As a result it is a lot more sustainable and easily serviced by rail infrastructure than something like Los Angeles or Houston. The idea being that while most people may commute to the city to work, they do so by rail and otherwise shop and live mostly within their own town. It reduces car dependency and cuts down on the number of trips made overall (at the expense of having longer rail based commuter trips)

  8. As a station only has a walk-up of say a half kilometre radius per station, rail to a place like Kumeu is more like to lead to intensification around the hub than sprawl. Can’t beat motorways for inducing sprawl. See the muppet McShane; if rail promoted sprawl he’d be all over it.

    1. Patrick, While quite agreeing with your sentiment, I’m afraid it would be rather hard for Owen McShane to be ‘over’ any form of sprawl since he died in March last year. But, as you know, there are plenty of other sprawlists eager to slap motorways everywhere! In fact many of them seem to have found a particularly agreeable niche at AT.

  9. Yeah well keeping the MUL is your only defence against sprawl, and really rail transit is not the problem here. I’m all for people living in Hamilton and working in AK if there were a train… like all those people who commute from York… And yeah the ‘suburb’ was invented by the train, but a rail based suburb [or tram] is of a completely different quality to a car and motorway based one. It still has local community and liveable streets… leafy and all things good about the ‘burbs, not eight lane vileness cutting communities apart…. really in AK car dependency is the big quality of life problem, running a few trains north ain’t no thing.

  10. Satellite towns seem like one way we could potentially improve housing affordability while still enabling people to work in the big cities. However, I think in the longer term that would necessitate four-tracking of some of the railway lines, so the “inter-urban” trains can run express through the city.

  11. “the benefits through reduced congestion are likely to be at the higher end of the $17 peak hour benefit to road users that typical rail trips provide”


    The $17/passenger number quoted on pSP10-8 of the EEM (vol 2) applies to measures that increase patronage of existing services, not to new services.

    Leaving aside that problem, using the ‘Auckland’ figure from that table would only be relevant for peak passengers traveling most of the way into the city, since that figure only applies if you can assume that “during peak periods there are congested traffic conditions (where the ‘ruling’ intersection or bottleneck operate at least 80 percent capacity during the peak one hour period) and includes a factor for induced traffic effect.”

    However it is to be expected that many passengers from Kumeu etc would actually be traveling to other western stations like Henderson or New Lynn. Look at the estimated peak hour journeys from option 1: 26 per day. I presume that’s total am plus pm, so you’re actually talking about around 13 people. Let’s assume that 50% of those are going in to the city (you can pick a different number if you want to). So you would be spending about $300k to get a $17/trip benefit for about 7 people.

    Of course you could just offer to write each of those 7 people a cheque for $40,000 a year if they’ll find a reason not to make those peak hour journeys.

  12. Couple of issues with your analysis there Mark:
    First of all you are comparing the annual operating cost to the *daily* benefit. If you want to do that you’ll need to divide you costs by 365 to get a daily BCR.
    Secondly the $17 benefit refers to boardings (i.e two boarding per journey) while the journeys above refer to new passengers (i.e. return trips) So you should be doubling the benefit, not halving it.
    Thirdly I wouldn’t discount trips because they aren’t going to downtown, there is plenty of congestion in Henderson and New Lynn to meet the assumptions.
    Fourthly, you have left out the PT user benefits from your equation, those are an additional $13 of benefits at peak and $8 off peak.

    So if we take those 26 journeys and assume the are half peak and half off peak:
    13 peak journeys x 2 trips x ($17.27 motorist benefits + $13.18 PT user benefits) = $792
    13 off peak journeys x 2 trips x ($1.65 motorist benefits + $8.8 PT user benefits) = $271

    Daily benefit = $1,063
    Annual benefit = $387,995
    Annual cost = $294,000

    So even the worst option has a BCR of 1.3, and we could pay each of those travellers three grand a year to take the train and still be better off!

  13. Thanks Nick. Good input, though let me clarify some of my comments…

    I wasn’t actually trying to compare annual cost with daily benefit. I was highlighting the size of the annual cost relative to the number of people for whom the peak hour benefit applies, a benefit that accrues regardless of how you take their vehicles off the road. I take your point that the 26 journeys per day are return journeys – thanks for clarifying. That does change the calculation, but nonetheless I don’t think it negates the core point…

    As you point out, the total benefit accrues to two different set of people – the broad community of road users, arguably a proxy for the community as a whole, and the rail service passengers. My point was that in this case the later group are remarkably small in number. Option 1 only deals with peak hour travel, but as I suggested I think only a subset of the 26 are making journeys that qualify to make $17 in traffic reduction benefits. But let’s call it 26 for the sake of argument and assume that it’s the same 26 vehicles every day (of course the real world is a little more complicated).

    Those 26 are each getting a private benefit of ($13.18 x 2 x 365) = $9621 per year from the ratepayer, while the benefit to the ratepayers is ($17.27 x 2 x 365 x 26) = $328k per year. (side note: I have followed your lead and taken a year as 365 days. When I first looked at this I assumed that the 26 peak hour passengers only applied for 5 days a week, but perhaps those numbers have been averaged over 7 days).

    So you spend $294k to get $328k of public benefit. Good deal! But… you get that public benefit regardless of how you take those cars off the road at peak hour. So look at it another way – you are spending $294,000 per year to take 26 vehicles off the road each day, or $11,307 per year for each vehicle. I just think that’s remarkable, and that’s the issue I was trying to highlight. As I say (facetiously obviously) you could just write each of those people a cheque – true, not $40k, but more than the already generous private subsidy of $9621 they would get otherwise.

    1. Mark – concur – why don’t they just run a bus from Huapai to Kumeu and down to Waitakere! Far cheaper than running a train which would be next to empty until about Swanson. My guess is that a lot of the people who’d use the train are already driving down to Waitakere anyway, or perhaps Swanson.

  14. Mark, one thing with these figures is they are monetarised intangibles, i.e. putting a dollar figure on things that don’t have an actual market price. While this is useful for cost-benefit comparisons it means you can’t actually translate the outcomes into real dollars to pay people, even talking about subsidy and profit is a bit of a stretch.

    Ross, the answer to that is that you probably wouldn’t get the same uptake or the same level of benefits. In particular bus travel has lesser external road user benefits because buses still occupy road space and contribute to congestion. Bus travel is given a lower benefit price for that reason.

  15. Nick R – the congestion savings of running a train along that corridor (I grew up round there in the late 70s, so I know it fairly well) will be next to negligible. Taking twenty cars off a particularly congested bit of motorway does have real benefits in terms of traffic flow; taking twenty cars off the road between Kumeu and Waitakere would barely be noticed.

    I appreciate that running a train would have a higher takeup, but the marginal *benefit* of that extra takeup is well short of the marginal *cost* of that extra takeup. My guess is that by running a bus, you would achieve about two-thirds of the benefits of an extended train service – but at about a quarter of the cost.

  16. @Ross Clark

    “My guess is that a lot of the people who’d use the train are already driving down to Waitakere anyway, or perhaps Swanson.”
    Not really possible when the park and ride at Waitakere is miserably small and already full every weekday morning.

  17. Interesting conversation.

    One aspect not considered is the growth plan for Huapai/Waimauku areas. Auckland Regional Council earmarked the region to allow the population to grow by another 9,000, I think, over the next decade (I can’t quite remember the numbers that were produced at the meetings considering the awful Cornerstone proposed developments). SH16, with only 1 NW bound passing lane between Westgate and Helensville is already congested and the increased population is going to make that any better.

    That population growth will happen and we really don’t want every one driving everywhere (there is a pretty lightweight bus service running into to town which doesn’t seem that well patronised, I could be wrong).

Leave a Reply