“At the margin” is one of my favourite economic expressions.

Recently I’ve been thinking about how the City Rail Link (CRL) affects the financial performance of Auckland’s rail network “at the margin”.

  1. What is the impact of the CRL on patronage?
  2. What is the impact of the CRL on fare revenue?
  3. What is the impact of the CRL on rail operating costs?
  4. What is the impact of the CRL on cost recovery?

Now I don’t want to send the wrong message: Financial performance is not as important as economic performance.

But financial performance is important to some degree; as it defines how a given project impacts on the National Land Transport Fund (NLTF). And where a project’s operating costs exceed its revenues, then it will require on-going operating subsidies.

By extension investing in projects now that require on-going subsidies will reduce our ability to fund future projects. So if we invest in transport projects that require large on-going operating subsidies, then we are effectively making a decision to reduce the funding that is available to fund other projects in the future.

That’s why I think financial impacts are important.

It’s interesting how quickly political biases enter the discussion as soon as one mentions the words “financial performance”.

Many people on the right presume that roads are associated with good financial performance, whereas public transport – especially rail – is not. On the other hand, many people on the left presume that anyone who is concerned with financial performance is guilty of short-term “roads first” thinking. Both are presumptuous.

And I think the CRL provides a good example to challenge some of these biases. First let’s quickly make some guess-estimates in response to the questions posed earlier:

  1. Patronage impacts = The CRL is ultimately expected to generate ~20 million additional rail boardings p.a. (possibly more)
  2. Revenue impacts = 20 million x $4 per boarding = $80 million p.a. in additional rail fare revenue,
  3. Rail operating costs = let’s say we double service levels for circa $70 million p.a. in additional rail operating costs
  4. Cost recovery = Revenues / Operating costs = 114% (in a marginal sense).

Now I’m the first to admit that the numbers are a bit rubbery. While we won’t gain 20m boardings overnight, the same holds for the operating costs: We’re not going to need to double rail frequencies straight away either. We’re also ignoring capital costs of course, but that’s a different question.

Hidden in the numbers above is an interesting assumption that I’d like to discuss in more detail: That is the average fare of $4 per boarding. Why is that interesting? Well according to NZTA documents Matt has received in the past, the average fare on the rail network is currently around $2.70 per trip. So why have I assumed $4 per boarding?

This is a situation where I think it’s important to think “at the margin”. Or to put it another way, the CRL is not your “average” rail project.

Simply put, the CRL opens up vast swathes of the city centre. It seems fairly clear that the marginal passenger attracted to the rail network by the CRL is more likely to be paying a higher fare than the current network average, which is essentially dragged down by a relatively high proportion of concessionary travellers, which in turn is a function of the rail network’s extremely limited coverage of the city centre.

In contrast, trips to the city centre are far more likely to pay an adult fare. One way to get an idea of how much higher the marginal fare might be would be analyse average fares for people boarding/alighting rail services at Britomart. But I’ll leave that to AT …

Figures obtained in the past from Auckland Transport says that the average journey length on the rail network in excess of 15km. That means on average most people are travelling from outside of isthmus and based on the current fare structure would equate to a 3 or 4 stage fare depending on the line. Current adult HOP fares $4.05 for a three stages or $5.04 for a four stages. As such an average fare of $4 for passengers attracted to the network by the CRL seems about right.

On the operating cost side, it currently costs something like $90 million a year to run the rail network. A decent proportion of those cost are somewhat fixed or at least exhibit some economies of scale. One of the big advantages of electrification is that it should reduce the cost of running individual services and that means Auckland Transport will be able to run more of them than they do now for little incremental cost.

As such a doubling of future services due to the CRL should not result in a doubling of the total operating costs (please correct me if I am wrong on this).

For the purposes of this exercise I’m going to assume that the future electrified rail network costs $100 million a year to run and that adding services as part of the CRL will increase costs by $70 million per year.

Assumptions about average fare and operating costs aside, it seems the CRL actually performs relatively well from a financial perspective. In the long run it may even pay its way operationally, especially if AT can generate additional revenue from leasing space within/around the stations.

If this is the case then not only would we have significantly boosted rail patronage and accessibility across the city, but we would have done so while also improving the relative operational performance of the overall rail network.

The table below summarises these assumption in the context of the overall rail network, where farebox recovery improves by ~25% to what is a relatively high 79%.

CRL Financial Performance

To provide a point of comparison I have also crunched some numbers on cost recovery for Puhoi-Wellsford. These suggest that “cost recovery” for Puhoi to Wellsford is circa 50% (NB: This number is even more rubbery – I might do a separate post on this if people are interested in how the concept of cost recovery might be applied to roads).

Not that this is a post about the issues with that project, but it’s worth keeping in mind whenever anyone tries to convince you that new highways “pay for themselves” more so than new railways. While that may be true “on average” the story could be quite different “at the margin”. And it’s the performance at the margin that matters most.

The general message, however, is that the financial performance of the CRL is not too bad. To cut a long story long ;).

Share this


    1. good question – in my back of envelope analysis I assumed that OPEX = 1% of CAPEX p.a. So with CAPEX of $700 million OPEX = $7 million p.a.

      If someone can dig into the business case and find some better data please let me have it.

      I’m wondering whether the structurally intensive nature of Puhoi-Wellsford would result in higher OPEX than your normal road? Or whether the higher construction standard would result in less ongoing maintenance? Hopefully someone out there can tell us more about that …

  1. I think this is a really important point, The rail network in Auckland gets a lot of criticism for how much it costs but projects like electrification will really help that by bringing down the cost of services and by making the system much more attractive. The CRL takes that even further and there are probably not that many projects out there that perform as well as you suggest.

    One point though is I have long thought the estimates for the impact of the CRL are underestimated. I guess it’s just because we simply haven’t been used to seeing the transformational impacts projects like this have. I suspect a patronage increase of 30m, perhaps even more is possible.

  2. What’s Skytrain patronage in Vancouver? 140m a year on a similar sized network to Auckland?

    Of course Vancouver actually allows real intensification around its train stations and has a properly integrated bus network.

    1. We’ll soon have an integrated bus network and certainly on the Western line and inner city suburbs like Eden Terrace we’ll be seeing a lot of intensification.

        1. I’ve heard one estimate that on Canandian metro systems staffing accounts for approximately 70% of marginal operating costs, don’t have a source however. SO that suggests that driverless reduces marginal opex by two thirds.

          Figures from the Auckland electrification business case predicted $220k per year for “energy and maintenance” and $80k for “train crew costs”.

          So to start with staffing is about a third of all opex, however we don’t know how much of that maintenance is required anyway. Wouldn’t be surprised if staffing was over half the cost at the margin.

        2. Don’t know what Auckland is or what other Canadian cities are but from what I can find it cost Vancouver $329 million Canadian dollars to run their rail system last and for that they had patronage of 123.3 million trips. Of that $329m $65m of that was for staffing costs so only 20% of total operating costs. By comparison their bus network cost $766m to run of which staffing made up 54% of costs.

  3. Does anyone know what are the expected reduction in the operating costs with the electrification?
    I keep hearing this, but it would be interesting to know. Say a saving of 15% would be a huge savings for AT.

    1. I have been chasing up AT for this data for a while but they are being slack getting back to me. I suspect that overall expenditure will eventually end up being fairly similar with additional services using up the operational savings.

  4. Stu, I think that while your $4 a fare estimate may sound right to you, the overall effect will not be an extra $80m in the kitty revenue-wise.
    Instead it will grow, but at the same it the figure will be a result of cannibalising the revenue on the bus side of the PT equation as users transition from bus only mode journeys to multi-modal journeys where bus may form 1 or 2 parts to the journey but not all, while the train does the long haul components.

    Under the future proposed integrated ticketing and integrated fares, which we hope will be in place by the time the CRL opens then users will pay a set fee no matter how many actual modes they use getting from A to B. So, if the train revenue is up, it can only come in conjunction with a drop in the bus revenue.
    Well technically, this is more apportionment than anything else, the net effect is to reduce the amount of revenue earned by the bus modes and so while it grows the pie overall, the actual effect will be somewhat less than whatever $80m projected would be on the existing revenue now.

    In addition, I also believe that many of the expected additional (aka “marginal”) trips on the CRL will be at best 1 or 2 stops within the CBD area mainly as people move about the CBD via the CRL (which becomes effectively, Aucklands Tube system) as a result I don’t think they will pay $4 each trip they make. They’ll pay more like the minimum fare as it will be all within the Zone.

    It would be nice to think that these trips will all be in addition to the usual daily tidal flows in and out of the CBD, but as roads will remain the dominant mode inside the CBD for some time to come many of these trips will require car and bus users to change modes to the train to complete their journey inside the CBD.
    Either because the bus network is redesigned to make it so, or because the train users see that its quicker to park there in the CBD and get the train to their office over here..

    So, yep the extra 20 million trips figure from the CRL is not in doubt, just how much more marginal revenue they’ll generate is.

    1. “Instead it will grow, but at the same it the figure will be a result of cannibalising the revenue on the bus side of the PT equation as users transition from bus only mode journeys”

      I would beg to differ, the increased frequencies, and spread will bring many of those users to trains from cars, obviously not all, but most.

    2. Good points Greg and yes passenger diversion/cannnibalisation would need to factored into a more detailed analysis.

      Where passengers divert from buses then there is likely to be no net change in revenues, where passengers divert from cars then we (NZTA) will lose some fuel tax revenue while AT gains the fare revenue. On a per-km basis PT fares are higher than fuel taxes so that’s still a net gain.

      The other thing to consider is the degree to which rail cannabilises bus depends on the degree to which AT adapt the bus network in response to the CRL. Much of the New Network, for example, appears to have been designed in response to the opportunities offered by HOP and electrification. As such, I’d expect more “growth” than “diversion” when these rail improvements eventually come online.

      In the case of short versus long trips I’m not so sure. People typically travel further to reach the city centre because of the higher wages that are on offer there. It’s not clear to me that the inner-city trips will grow more than the outer trips.

    3. SB – Don’t disagree, but that diverison of bus passenger will mean that the $4 per trip figure (for $80m “Extra” revenue) is overstated as I said in my original post.

      SD – Yes if people divert only from cars to trains then the increase is positive all round. if people end up diverting away from a 100% bus journey as they make now then is will not be $4 per extra trip.
      And if people divert from cars to a multi-mode journey of buses and trains (because its faster, not just because its cheaper), then I expect that the resulting figure won’t be $4 per trip.
      Ideally the best outcome is these extra 20m trips a year on the CRL to/through the CBD replace 20 million car journeys into the CBD or across town from the outer suburbs (which are 3 stages or more away (i.e. $4 a trip) at todays fare levels).

      With the bus and trains integrated to each out in the outer suburbs, will that shift happen to that extent?
      Probably – but will that result in a net increase of $4 per trip to make that $80m of extra revenue? – I don’t think so.

      I see that the marginal additional $ for those 20m extra trips will be between 1 to 2 stages of revenue more (as compared to the cost of doing a 100% bus trip of the same length today)
      I think if the PT fares are too much higher than that people won’t switch modes (away from cars) so 1-2 stages in stage money is what? a max. of $2.50 per trip probably less.
      So marginal revenue will be maybe $50m at best.

      People may be forced to change modes between bus and train as part of the revamped bus service – but people won’t pay more to travel the same distance to the tune of $4 more per trip or pay more if the trip isn’t significantly faster than the alternatives (e.g. car).

      I think inner trips will become a large component of the extra CRL generated trips a year – whether they are a larger component than the longer trips I don’t know, but as a result of the existence of these lower value trips, the $4 per trip “average” for those short trips is too high and thus reduces the total revenue from those marginal trips as a result.
      Meaning $80m extra revenue is a target unlikely to be achieved.

      Again the best outcome we can hope for is that these inner trips are in addition to the expected extra 20m trips from outer suburbs who are going to pay at least $4 per tirp, so the total marginal trips “pie” is larger, and whatever revenue these inner trips make is simply seen as additional to, not instead of the other trips..

      1. I suspect that a decent proportion of usage will be generated from the inner west stations due to the faster trip times but equally a lot of users will come from further out. Currently trains from Henderson might be able to roughly match a car in peak time but off peak they stand no chance on speed. With the CRL it is likely a trip to town will be in the range of 30-35 minutes. That will be much faster than driving at peak and even competitive off peak. As such I think the growth in trips will be fairly even between the inner and outer parts of the city.

        1. Greg of course you’re right but the question remains how much will the CRL grow the whole non-driving network (bus/ferry/train and Active) as opposed to simply attract an existing PT user on another mode?

          Several reasons it looks sensible to assume it will attract a significantly high proportion of new users:

          1. This was the experience with earlier and much more modest improvements to rail system, ie the growth in rail patronage this century in Auckland has been accompanied by growth in bus and ferry use too. So zero net cannibalisation.

          2. The project offers a radical transformation of the utility of the entire non-driving proposition in Auckland, not just for rail, but for every kind of connecting mode from walking, cycling, ferry, and especially buses. Especially because of the move to a connecting network and AK’s form if there are 10, 20, or 30 mil new train journey pa post CRL then there will almost certainly be just as many new journeys on other modes, esp buses, in order for those people to get to the train stations. We can expect the CRL to grow the whole pie, lift all boats; whatever metaphor you like. So not just zero net cannibalisation but further growth of bus use becausewhole system now so much more frequent, reliable, faster; better.

          3. Population is growing, driving costs increasing, access to the key destination of CBD by car not getting easier or cheaper.

          4. Currently appears we have chosen to sprawl more with this future growth. This means more people making longer journey for work, study, and play, across and between whole as well as to centre. More incentive for people to choose those longer rail trips where possible (park and ride at the fringes set to boom and should be planned for: a la Perth). As Matt note the west clearly gets a whole new deal with CRL, and the south will too with more, nicer, and faster trains (Puke will surely get electrification soon too). Also despite UP intensification is happening too where possible which is also positive for general PT uptake.

          Remember Britomart was supposed to be an expensive way to get people off buses; the reverse happened, more people got on both buses and trains, and ferries too. I am confident that the CRL will blow projections off the chart, if all else is ready to support the potential uptake, this is the ‘Killer App’ not just for PT in AK, not just for movement in AK, but for its whole urban form, vitality, and success as a place.

          Key point here though is that AT have to plan and invest to attract those drivers to try their improving network to get fully value from CRL and current improvements. Connecting stations and services, much better off peak services, and of course integrated fares. But they know that. Don’t they?

        2. Patrick,

          Yes, previous PT mode growth did not cannibalise one mode by another as we didn’t have integrated fares then and there was enough fare $ to go around so no real arguing about how big the relative size of each slice of the pie was as each slice was more than big enough already. Under the CRL model which includes Integrated fares you will get that exactly the opposite going on – arguing about who gets the biggest fare slice and/or who gets to choose the slice to take first.

          And while the bulk of PT providers (buses and ferries) are privately owned and operated, the portion of the revenue they get to keep (or not) from the fare $ is an important issue for us all, not least as the owners of said ferries and buses tend to have the ears of those meddling politicians, which the average bus, train or ferry user does not.
          So when the fare $ money starts to drain away from their previous lucrative little earners, they will act.

          And do not forget that over time all PT modes were basically eaten alive by cars and all the attendant reprioritisation from central and local governments that went on from the last 60+ years of cars over everything else. And even now the road lobby seems to have a direct line to the NZTA decision makers (and the Ministers) hence the endless road building programs we have had and sitll continue to get.

          And yes, thats the situation AT must seek to reverse quickly by demonstrating to all how EMUs and the CRL together form the key solution to the future problems of Auckland commuter traffic. Solutions to problems that many have not even anticipated.

          But this does beg the question of how soon before the political masters of AT start to get cold feet about doing a proper job which must include integrated fares?
          Lets not forget that the board of AT is Government appointed, not elected by you and me, or even appointed by local politicians elected by you or me.

          We already see signs of that “cold-feet”-ery now with endless HOP rollout delays for this and that reason, poorly thought out and communicated fare changes, no off peak fare structure, no capped daily, weekly or monthly spends on your HOP cards. No real reason to use HOP in fact.

          I suspect that the main reason for these delays and seeming ignorance of the other elephants in the room is that that the bus operators (or one in particular) is not willing to have its rice bowl broken by AT any time soon with such changes. And no matter how many times Minister Joyce et al say “If they don’t, they’re off the run”, the truth is they’re too big economically and politically to have that done to them as easily as AT seems to believe they can.

          To put it another way, If the opera ain’t over till the fat lady sings, then right now, that fat lady hasn’t even entered the building, let alone done anything else like warmed up her voice.
          And anyone who thinks otherwise is in for a rude shock in a few years.

          Lastly, you say that “Key point here though is that AT have to plan and invest to attract those drivers to try their improving network to get fully value from CRL and current improvements. Connecting stations and services, much better off peak services, and of course integrated fares. But they know that. Don’t they?”

          Can’t argue with that.

          But actions speak louder than words and AT is dithering in the extreme over this issue as you can see by the way they’re not even considering or talking anything to do with integrated fares – and thats with the PTOM model changeover under way and continuing over the next 3 years.
          So you have to wonder deep down, if AT really does know that and actually believes it to be so?
          Certainly slapping a few AT logos on the buses may look like they’re on the case, but you gotta wonder, are they really?

          I get the feeling that AT are setting themselves up mentally towards becoming the “default” mode provider – the provider of last resort – the one that you use only because you have no other mode choice. Instead of being the the premier “first choice” mode – the one you consider using first.above all other modes, including cars that is must become.

        3. Greg you are comparing the contracting model now with what we will have as part of the new network. The introduction of PTOM changes things and unless an operator runs a route commercially (including services at all times of the day) then the operators will effectively be on gross contracts so no arguing over how the farebox is split.

        4. If rail does cannabalise some bus patronage chances are that those buses will be filled up by new people catching the bus, who were previously put off by the bus being packed out. The new bus network will radically improve the reach of the rail network. The amount of people that catch bus train is extremely low, while in other countries is very high. With integrating ticketing and new network lots more people will be able to access the rail network, thus a big boost in trips.
          As for whether the trip length will increase in hard to tell. I suspect lots will catch the train from Britomart to K Road, seeing the amount the people that get on the City Link which takes 25 minutes to do that trip, and the CRL will take 5 minutes. But these trips will be good revenue makers, especially as no one minds standing on 5 minute trips, so won’t create demand to add more services.
          Of course will be really big boosts from West Auckland due to slashing of journey time, though the North-Western Busway may cannabalise some of these trips for those coming from Te Atatu South/Lincoln Road/Massey depending on the timing of that.
          Overall the CRL will result in much fuller trains off-peak (more at 6cars rather than 3), and if more trains added at peak will only be done if existing trains packed out, so these will be good revenue earners too. That sounds like a great way to lower average journey cost to me.

        5. There is no way that the CRL could actually cannibalise patronage of buses. The new network removes most long suburban buses from the rail corridors and reallocated them to high frequency rail feeders and local circulators. The remaining buses on non-rail corridors are about to go crazy with the new frequent all day every day service model.

          If rail network patronage skyrockets it’s only because bus patronage is skyrocketing to get people to the train in the first place. Sure in the long term we’ll see land use changes with more people living near rail, but those will be new trips from new population.

          Both bus and rail patronage is going to shoot up soon, especially because of integrated ticketng. Bus patronage is not going to decline, and without a decline there can be no cannibalisation.

        6. I was only thinking about the New North Road (Rosebank via St Lukes), and potentially Sandringham Road services. People going to university and midtown will currently catch the bus (no change likely with new network), but with the CRL those trips will go to rail. May well see people on the Sandringham Road services changing to rail at Kingsland for a quick trip into-town. This will benefit bus patronage as well as rail.
          Of course rail will generate lots of new trips, and the buses will gather lots more short trips, so patronage of both will shoot up. I certainly wasn’t intending cannabalising to mean no nett gain in trips, and believe CRL will lead to huge trip growth across bus and rail, but there are lots of lille places specific changes that will be interesting to monitor.

      2. Fair point Greg.

        Personally I feel that we are likely to see a large shift from cars, and less from buses, but then I also think we will be looking at 40m trips p.a. in 20 years…..

        1. SB,
          You are 100% right – in 20 years (whether thats 20 years from now 2013, or 20 years from when the CRL opens), there will be 40 million more CRL trips undertaken, and by then the minimum “inner stage” fare will be $2 per trip, so you will get that $80m of revenue “extra” by the end of that 20 years.

          Of course, in todays money that will be worth probably half what $80m buys today.
          But that doesn’t seem to bother NZTA (or politicians) when they look at roads like PuFord.and expected benefits.

    1. As Stu points out the focus is on looking at what the long term implications are to operating expenses. Capital costs obviously shouldn’t be ignored in a full economic analysis but once the infrastructure is in place we need to be able to afford to maintain/run it. The cost of maintaining a road like Puhoi to Wellsford will be enormous with the tricky terrain, 7 viaducts and many other structures like road bridges, culverts, embankments etc. There is unlikely to be the traffic volumes to pay for the upkeep of it and the existing route which will still be in use where as in the case of the CRL, it is likely to be the opposite.

    2. Yup it is a large capital investment. The merits of which depend on a whole load of non-financial impacts, such as consumer surplus (user benefits), travel-time/reliability benefits to other road users, and agglomeration benefits.

      What I’m trying to demonstrate here is that the CRL does not seem to increase subsidies required for operating rail services, hence the long-run financial impacts are fairly negligible. Another way to look at it is that the service investment required as a result of the CRL are paid for from additional fare revenue, hence they are not a drain on the NLTF in the long run (remembering that it will “re-charge” over time from fuel excise taxes). That’s not to say that the upfront capital investment is worthwhile, just that it does not bleed the NLTF dry in the long run – which is a positive thing methinks.

      Of course if the economic benefits < economic costs then we'd be better to 1) not spend the money on the CRL and 2) instead lower fuel taxes instead. But that's expanding the question slightly.

  5. Dammit. M. Williamson was just on the News blah blahing “Auckland pop growth nothing like len’s prediction, good luck getting your train tunnel”

    Whatever our growth rate; it should have no effect on the CRL timing. We need it already, even if not one more person moves to Auckland. SMH!

    1. Exactly, population growth is largely irrelevant, for example Vancouver has the same size rail network as us and only 30% more population… But 1300% the rail patronage.

      …but maybe we’re not growing as fast as predicted because we’re not building the right kind of infrastructure.

    2. Completely agree. And the rate of growth is less important to the CRL than the distribution of the growth that does occur.

      I would not be surprised if the recent changes to Unitary Plan, which have resulted in more population growth out west and south, have turned up the dial on the CRL’s benefits.

      Someone needs to tell Maurice that when you chuck a whole load of people at the end of Auckland’s two busiest rail lines that it tends to increase the case for rail investment, rather than decrease it.

    3. Detailed data is not out until the 15th is it not? So what is he basing this on? In any case if cancelling the CRL is the sound response to decreased growth (should it actually be the case), then obviously we would be cancelling other projects based on a high rate of growth. And that would be just about any major roading project in the area.

Leave a Reply

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