Public transport subsidies are a perennially controversial topic. In this post I quickly summarise the case for subsidies before highlighting why Steven Joyce is pretty much wrong about public transport subsidies in Auckland.

Public transport subsidies are basically motivated by “externalities.”  These externalities arise because individuals do not factor in the full benefits and costs of their travel decisions, which in turn results in outcomes that are economically inefficient.  In the case of transport, the big externality is congestion: Drivers simply do not consider the cost of the delays that they impose on other road users, hence more people drive than is ideal.

So by subsidizing public transport we can reduce the numbers of people who choose to drive and thereby reduce congestion on the road network (for a given level of development).  As highlighted in Josh’s earlier post, the cost of congestion is significant: The net economic benefits of Auckland’s public transport system exceeds the subsidies by 3:1 (based on data from the NZTA), i.e. PT subsidies generate benefits three times greater than their costs.

Congestion is not the sole externality that causes too much driving; the cost of parking is another.  Parking is not paid for by users and as a result more people chose to drive than should.  Because of minimum parking requirements, this demand has to be reflected in increased supply, creating costs that must then be subsumed within the wider economy.

So the general motivation for public transport subsidies is fairly simple: They correct for externalities (or market distortions) that result in more people driving than would in a situation of accurate pricing.  But what does Steven Joyce mean when he suggests that public transport subsidies are increasing too fast?  I think what he means is that the fiscal “input” (i.e. subsidies) is growing faster than the public transport “output” (i.e. trips).

This frame of thinking is illustrated in the figure below, where the black line (input in dollars of subsidy) is rising more rapidly relative to the blue line (output in trips).

This conclusion is misguided for several reasons.  First, it focuses on financial, not economic, outcomes.

Put another way,the role of government (in the long run) is less about fiscal impacts than it is about economic outcomes.  To highlight the difference, let’s say that the period from 2005-2011 saw increased congestion on Auckland’s road network (which it did).  Intuitively we can appreciate that this should justify higher subsidies than in the past. But the economic costs of congestion figure nowhere in the above graph.  Thus the above graph is economically meaningless and not worth the pixels it is written on (at least in terms of public transport’s effectiveness).

But there is a more obvious flaw in Steven Joyce’s analysis of subsidies.  That is, the number of public transport trips is not an appropriate measure of the system’s “output”.  Instead, a much better indicator of output is the total amount of kilometres travelled by public transport, which is equal to the number of trips multiplied by the average trip distance.  Kilometres travelled is a better measure of output because it is more closely linked with congestion.

And because Auckland’s public transport network is attracting more long distance travellers, the total number of kilometres travelled has increased much faster than the total number of trips.  This can be seen by considering the figure below, where I have plotted the (indexed) growth in subsidy per trip (blue) and subsidy per kilometre (red).  The latter has increased about 16%, whereas the former has increased about 32% – so subsidies per kilometre travelled have increased only half as fast as subsidies per trip.Despite using a poorly specified measure of output, Steven Joyce might look at the red line in the figure and conclude that his original conclusion still stands, because the subsidy per kilometre in 2011 is 16% higher than it was in 2005.  But again he would be wrong – because this analysis is still not complete.

The key reason it is incomplete is because the red line is not a true measure of output efficiency over time.  The problem is that “subsidies” have not been adjusted for inflation.  So to get an real idea of how the efficiency of Auckland’s public transport system has changed over time you need to adjust down future subsidy levels in line with inflation.

And in this period inflation actually ran at between 20-30% (depending on whether you use general CPI or just the transport sector).  When this is factored in the subsidy per kilometre travelled (i.e. the red line above) does not increase at all, as illustrated by the dashed red line below (which is in constant 2005 NZD).So if the per kilometre benefits of public transport increase (because of worsening congestion) while the per kilometre costs decline (or at least remain stable), then the only conclusion we can draw is that Auckland’s public transport system is in 2011 more effective than it was in 2005.  Stated differently, Joycey baby is a) mixing up fiscal and economic implications b) using the wrong measure of output and c) failing to adjust for inflation.

I want to focus on the positive.  It’s worth re-stating that Auckland’s public transport system has actually performed reasonably well over the last six years, in both financial and economic terms.  To make trends in subsidies clearer, the graph below shows the percentage change in rolling 12-month average subsidy.  You can see that subsidies were increasing up until about July 2006, after which time they seem to have declined in real terms.

This figure also highlights the impact of the Northern Busway, which attracted large numbers of long distance trips with relatively little increase in subsidy (NB: Although the actual impact has probably been accentuated by the linear interpolations used between data points).  This suggests the “peak” in public transport subsidies (on a per unit of output basis) was reached long before Steven Joyce waltzed onto the scene wearing a self-proclaimed hat of fiscal restraint (unless you’re talking about RONs – then his wallet is positively wide open).

It also seems likely that public transport subsidies per kilometre travelled will continue to decrease at about 1-2% per annum (in real terms).  By extension, it would also seem that Auckland’s PT system is experiencing economies of scale, which means that the marginal cost of new users is lower than the average cost of existing users.  This is good news because it means average efficiency will increase as the system grows – something that we would intuitively expect.

Of course, that’s not to say that Auckland could not look for further efficiencies from its public transport system; some potential opportunities include:

  1. Charge for park and ride at Albany and Constellation – both raises revenue and pushes NEX passengers back onto local buses.  Would probably raise a cool $1 million (assuming $3 per day for 1,000 car-parks over 48*5 = 240 weekdays, with 250,000 trips per year shifted to local buses @ $1.50 revenue per trip).
  2. Target subsidies to price sensitive demographics – such as students and beneficiaries.  A peak/off-peak differential should also be introduced, encouraging peak spreading and higher utilization at off-peak times.
  3. Maximize revenue from peripheral activities – install advertising and vending machines on trains and at stops.  Yes I know it’s ugly, but it’s a price I’m prepared to pay if it helps us fund a better public transport network.
  4. Improve competition for bus contracts – NZ Bus appears to have market power.  It’s not helped by the “net” contracting model.  Replacing this with a gross contract would shift benefits and revenue considerations to the council, while leaving the bus companies free to focus on managing costs (something they do fairly well).

Despite these opportunities I suspect the big efficiency gains will come through changes to the current network.  The reduction in subsidy levels shown earlier is an aggregate response to reasonably well-targeted capital investment, such as Britomart, the Northern Busway, and HOP ticketing.  These projects have driven sustained growth in long distance trips.  Bus corridor improvements out west and east seem to offer similar opportunities for efficiency gains, as perhaps does the City Rail Link (although I’m not yet sure as to when it should be delivered).

In the long run time-of-use road pricing and parking reforms are probably needed if we are to push the public transport network towards the point where it is financially self-supporting – which would allow us to ignore characters like Joyce altogether.  These are, ironically enough, the very externalities that justify subsidies in the first place.  There’s a nice symmetry if market prices for roads and parking would result in PT become operationally self-financing.

I suppose on one level Joyce may be right: Market pricing may well deliver a reasonably efficient transport outcome, assuming that they correct for externalities.  But it’s very unfortunate that Joyce has put the financial cart before the economic horse – he should focus on internalising these externalities (i.e. “making markets” more efficient) before he tries to squeeze subsidies for services and infrastructure that are already efficient, and increasingly so.

His current approach to subsidies is amateurish and more likely to deliver false economies than efficiencies.  And for that he deserves a serve.

Share this


  1. Nice analysis admin, I’d love to say that Joyce is merely uninformed, however, I’m quite sure that these sorts of stats are more than well known by MoT and NZTA, but as we’ve seen again and again Joyce simply abuses statistics to show what he wants to show. I also think this obsession with operating subsidies is a bit of a joke because we’re talking such small amounts of money, if MoT really wanted to save money they’d be looking at the vast profligate spending in roads, rather than squeezing PT funding.

    I do think this could make a nice article for the Herald, perhaps I’m dreaming but surely an expose of how poorly the MoT is being run would be a great Sunday Star Times article….

    1. rtc, it’s Stu Donovan doing the nice analysis here. And although it is unpopular with some, I am compelled to agree that Joyce is clearly playing a very knowing and nasty game of numerical manipulation at every possible point. He isn’t dim or uninformed but smart and nasty and political. I’ve have just heard him on the radio repeating his extreme highway bias like it is a universal truth. Like some forth law of thermodynamics that the CRL is uneconomic and the building of any highway is good for the nation. He’s clever and dismisses any other view with condescension.

  2. “Steven Joyce waltzed onto the scene wearing his self-proclaimed hat of fiscal restraint (unless you’re talking about RONs – then his wallet is positively wide open)”

    We see this kind of declaration of fiscal restraint all the time, mixed with the profligacy of hypocrisy so many times from the right. The recent US debacle was a similar failure of the right, first they had their small government tax cuts for the rich rhetoric under Bush whilst wastefully burning money on the wars in Iraq and Afghanistan, and now they’re trying to push their ideas of fiscal restraint again, whilst denying their ilk is the reason for the failure.

    One could also lump the recent events in Norway together with the events in US and the cutting of PT subsidies in Auckland together and say they are all the failure of the thinking on the right. It’s not even drawing much of a long bow to do so.

    1. I’m not an expert across those topics so I can’t say but you may well be right.

      What I will say is that governments do well at making decisions that maximise overall benefits, while the private sector seems to do well at minimising costs. The reason being that benefits are dispersed across many agents operating in many different markets, whereas costs tend to be more narrowly defined (although this is not always true).

      So in the context of Auckland’s public transport, this would suggest that we need to move away from the net contracting model. NZ Bus are interested in maximising profits, not benefits, where the former may not align with what is economically optimal (indeed it won’t in the case of feeder buses to the rail network).

      A gross model is better, but better still (I think) would be a regional bus company. The goal would not be to eradicate competition, but to keep it honest. Much like Kiwibank has successfully done in the banking sector. It would also provide the council with better insight into actual operating costs etc, which they currently don’t have.

  3. Great post Stu and it confirms my thoughts that we are running a more efficient PT system these days than we were a few years back. Good to see that explained nice and clearly.

    1. Thanks and yes, we are! I think it just highlights the need for AT to shift away from “trips” as their preferred indicator, because it underplays the true contribution PT is making to serving Auckland’s transport demands.

  4. Great post, this is exactly the sort of analysis that MOT should be putting out. Not inflation adjusting this sort of thing renders it fairly meainingless. I also not on the original graph, the patronage axis starts at zero, whereas the subsidy axis doesn’t, exagerating the point that is trying to be made.

  5. Good post Stu – I think the trend for longer trips is definitely happening, I know on the trains the majority of people are getting on at stations quite far away from town and my own personal observations on the western line is that most are getting on west of New Lynn.

    1. I think you’re absolutely right there. Short distances don’t make sense in Auckland for most people. The actual trip time includes all the time it takes to walk from the front door to the stop, all the time spent waiting, all the time on the bus, and all the time spent walking from the bus to the destination. If the buses are infrequent or unreliable, and the walking distances are large, then the advantage over a bus trip is likely to be huge. Plus, the cost structure of short trips doesn’t argue for them very well.

      Longer distance trips face few of those issues, particular those made on rail or busway at peak or near peak times.

    2. Yes I agree. Another good reason why we should focus on long distance trips is because they compete less with walking and cycling. Auckland could be a fantastic city for walking/cycling if it got its act together …

      As an aside, the tendency for short PT trips to compete with walk/cycle trips is what makes me uneasy about the free City Circuit bus – from what I can see it mainly allows students to sit on their ass.

      But maybe I’m just becoming a grumpy old man (turning 30 in two months).

      1. As an aside, the tendency for short PT trips to compete with walk/cycle trips is what makes me uneasy about the free City Circuit bus – from what I can see it mainly allows students to sit on their ass.

        Perhaps. But if you’ve commuted in from another part of Auckland, another 20 minute walk on either end takes a full 40 minutes out of your day. That’s a lot of ones waking hours spent traipsing up and down Queen St and through Albert Park.

        1. Hmmm .. I’m not saying the bus should not be provided, just that it should not be free.

          The flat 50c inner-city fare seems fair. Maybe make it a day ticket for that price. With HOP the transaction costs associated with these sorts of marginal fares drops substantially …

          And then you can decide whether 20 minutes of your time is worth 50c. On a sunny afternoon probably not :).

        2. “if you’ve commuted in from another part of Auckland, another 20 minute walk on either end takes a full 40 minutes out of your day.”

          Take a folding bike on the bus/train. It’s cheaper, faster, funner, healthier and less polluting.

          Your extra 40mins becomes 10.

  6. About half of the dashed line is fuel price subsidies. I wonder (strongly suspect) about whether the funding arrangements have made it less sensible for bus companies to invest in highly fuel efficient stock, including diesel electric hybrids, rather than simply maintaining a simple diesel fleet. Research shows hybrids to use between 25%-50% less fuel, depending on route, driving style, and other factors. I think telling bus drivers not to have such a heavy foot would also go some way to reducing that black line. I assume that they have little direct incentive to save fuel, and thus drive like lunatics half the time because keeping to schedules is given higher priority.

    Who covers the capital costs? What arrangements are there for fuel use? Which parties have incentives to lower operating costs? Could fuel use incentives be written into new contracts? What other inflationary costs are hidden in operating arrangements? These are questions that I feel would illuminate that first graph a little.

    I know that there were problems with the gas-electric hybrids, but hybrid technology is now very mature (and continually improving). A very large number of cities around the world use them, including London. They’re not a toy for ‘greenies’ (although actually saving the planet we live on should be given _some_ consideration by the wonks at AT and in Wellington). Unless we make the unsustainable assumption that oil prices will drop in the face of sustained demand and static production, this has to be a priority.

    As for trains, they’re being electrified, so that should (insy’allah) lower operating costs. Knowing Auckland I wouldn’t be surprised if for some reason it didn’t and ended up costing us more than the rest of the world. We live in hope.

    1. Whoops, approximately half the increase over the last decade, not the future predictions.

      If fuel prices were to increase substantially again, the subsidy line could look very ugly.

    2. Yes you may right re: fuel price subsidies creating a perverse incentive for bus companies to not upgrade their rolling stock. Although it’s worth noting that fuel efficiency would affect their tender price for new routes, so they are at a disadvantage relative to their competitors. I’m not sure how it washes out in the end …

      Re: capital costs, bus companies cover them. That’s one reason why (up until now) it’s been hard to compare bus/rail subsidies, because rail subsidies don’t include the capital cost of rolling stock while bus subsidies do. In this respect Steven Joyce is creating a level playing field – by getting the loan servicing costs of the new EMUs added to the pot of money used to fund rail services.

      Re: hybrid technonology – yes it is getting better. One option I’m interested in is a trolley bus with a battery. So you run overhead wires on the busy sections of your route, with buses reverting to batteries as the get further out and the routes diverge. That way you avoid the need for two engines (e.g. gas/electric) which is quite expensive.

    3. P.s. Fuel price escalation should be captured by the inflation adjustment, especially if you use the PI for the transport sector, rather than general consumer PI.

  7. @ George. Dunno if it would actually. Bc if fuel prices rose dramatically you’d probably get more people using PT and so the PT system would become more efficient (less buses with only 10 passengers).

    But I agree that our current diesel buses are revolting (for air pollution and health reasons as well as fuel consumption ones) and that investment into making them more efficient and less polluting is direly needed. This is the kind of private bus companies are very unlikely to do bc, as Stu says, they’re focused on increasing profits, not overall benefits.

    1. I’d suggest that SOME buses are revolting. Auckland’s new buses (e.g. Link) are better than those found elsewhere, whereas Auckland’s old buses are worse. So there’s greater variation in Auckland’s bus fleet, i.e. a bigger gap between the best and worst performers. But the rate at which it is improving is quite encouraging, so don’t despair too much!

      Some of the increase in subsidies may in fact be reflecting the cost of upgrading bus fleets, because the bus companies will try to recover that cost. To some degree higher subsidies is a natural consequence of more stringent emissions standards on buses and AT’s expectations around better buses.

      1. Yes there do seem to have been a number of new buses that have come on stream over the last 5 or so years, the new ones about to start on the new link services appear to be pretty good if the marketing hype is correct, also the ones used on the B Line services look quite nice (from the outside at least)

        1. I actually don’t think the bus fleet in Auckland is too bad. There are some dungas, but it’s balance between “bright and sparkly” versus effective use of resources. I think one thing NZ Bus should be commended for is the high rate of asset renewals they’ve achieved since taking over from Stagecoach. “It won’t happen overnight but it will happen” may be the appropriate motto here – I’d give NZ Bus another few years and see what the fleet’s like then.

        2. I think the increase in bus patronage in the past couple of years has also taken some by surprise so older buses have had to be kept on the road longer than originally anticipated.

        3. I’m amazed just how many kilometers you can get out of a bus or coach. I was talking to a driver on an Aussie Greyhound about this. The coach is on the road as many as 20 hours a day, stopping only to load and unload and for driver meal breaks. They run through the night. It does 1500-2000km a day for, say, 350 days a year equals 700,000km a year. And the coach was around 15 years old, which equals about 10 million km. It was still in reasonable condition. I tried to find out how they maintain them and how often they replace the engines but he said he didn’t know… he just drove them.

          By contrast, I used to own a Toyota station wagon that was absolutely stuffed at 320,000km when it broke down and I decided it was time to send it to the wreckers.

  8. Well said with this:

    In the long run time-of-use road pricing and parking reforms are probably needed if we are to push the public transport network towards the point where it is financially self-supporting – which would allow us to ignore characters like Joyce altogether. These are, ironically enough, the very externalities that justify subsidies in the first place. There’s a nice symmetry if market prices for roads and parking would result in PT become operationally self-financing.

    The issue will be that all modes should be capital self financing too. The technical means exist to do all of this. However, be wary that it is likely the market based outcomes would mean you have quite different PT from what is currently envisaged, it may be more bus based (free flowing roads do away with the need for expensive bespoke parallel networks), it may result in significant decentralisation in terms of location and time (as all peak trips are much more expensive), and see people living closer to where they work, or less likely to want to work in heavily built up areas unless they are paid enough to justify the high road or PT prices to access these locations.

    However your point “congestion causes too much driving” is wrong and tautological (like saying being overweight causes too much eating). It is the largely flat and non time or location variable marginal price for road use that causes it.

    1. Yes I agree Liberty, capital self-financing is the next step. One thing to remember, though, is that the land on which the road network sits has never been valued. Incidentally this is one of the key reasons why buying back KiwiRail was a good idea: You cannot expect the rail network to make a return on capital values (including land), while the road network does not.

      So moving to a capital recovery situation is actually fairly complex and may well disadvantage the road network more than public transport, because it uses relatively more highly valuable land (just think of all the land used for roads in downtown Auckland). Unless of course you just talking about self-financing for new (rather than past) capital investment?

      Maybe we should break it into three steps: a) operationally self-financing, b) new capital self-financing, and ultimately c) completely self-financing.

      Not sure what you mean by your last point: I don’t remember writing “congestion causes too much driving.” If I did then it’s an unintentional mistake …

        1. There is always the option of covers, although the breadth is significant. If NZTA operated commercially it could consider how it could optimise use of such land.

          Same can be said of Wellington rail yards of course (Auckland’s one have had value partly realised, but the big problem with just about any recovery of transport land in NZ is the Public Works Act requirement to offer back any land taken under it to the estate of the former owners, and for it to remain set aside for Waitangi Tribunal claims).

      1. Stu, the problem with valuing the land under the road network is that because it provides access, the value of neighbouring land is directly related to the presence of the road. Without a road, the value of the land drops dramatically. Railways offer nothing of that, and so the land can be valued quite readily, because the alternative use of a railway corridor is relatively straightforward. You can close the odd road here and there, but Auckland without roads would not function. Property developers build roads, so the value of properties is linked to that, and roads themselves generate considerable revenue.

        However, you can get a return on the value of the road and rail assets themselves, independent of the land. So it is not a problem. The road network has long generated a net return, which is the surplus of fuel tax and RUC over maintenance, which is then typically “reinvested” in new capital for the network. The railway has not done that, as a whole, for decades, although parts of it have.

        The phrase is “Congestion is not the sole externality that causes too much driving” that I was meaning, regards.

        1. Yes you have to have a road of some kind to access individual properties but you could argue that if it wasn’t for the congestion we wouldn’t have roads as wide as we do. With smaller roads we could have more land available for other purposes.

        2. Oh yes, recognise that the value of land is intertwined with the presence of road access. But it’s not impossible to value the land on which roads sit – you could just factor down values to account for the spillover benefits of road access, e.g. assume 50% of the adjacent property value (opportunity for some interesting econometric analysis there).

          And I think the more important analysis needs to happen, as always, at the margin. That means that we think less about selling off Customs Street entirely and more about whether we’d take just one lane and lease it to adjacent land-owners, who would probably value it more than road users.

          So even in situation where land was reallocated at the “margins,” I think we would still see a reasonable down-sizing in the footprint (and hence capacity) of the urban road network. And by moving in incremental steps you could monitor the subsequent impacts on surrounding land values, congestion levels, and driver WTP.

          My gut feeling is that most central city roads would be pared down to one lane in each direction: I.e. minimum capacity that provides some degree of access. Maybe a few more lanes through intersections …

        3. I suggest it entirely depends on how much road users are willing to pay to use and park on the road. It may be that this, in the context of a high urban per km congestion charge, and perhaps high on street parking charges, makes it different.

          The key thing is that decision would need to be made based on all factors. A quiet pedestrian friendly street with good access for deliveries, taxis vs a wide arterial with lots of through traffic paying through the nose, or something in between. The property owners are probably best placed to make that decision. It would justify much more tunnelling in built up areas, although that would depend on local tunnelling costs. Most ignore that in Auckland tunnelling is hienously expensive because of the geology.

        4. Yes agree, but my “gut feeling” is based on experience with drivers willingness to pay, or perhaps lack there-of (not that public transport users are any better 🙂 ).

    2. P.s. Completely agree with your observations on the likely implications of road pricing/parking reforms – buses will, I think. become relatively more attractive because of free-flowing roads.

  9. question – not sure if or how these two things are related. But the SOI of AT states that “Goal 5.5 “Public transport subsidy per passenger kilometre”, existing perfomance is $0.33 and while the goal is “Decrease %”. So the patronage graphs in the post presumably include subsidies of all sorts. Have these been drawn under the assumption that this performance measure for AT occurs? If not, what might happen to these patronage forecasts.

    1. Very related! Just note that these are not forecasts – they are calculations based on historical data. So looking forward AT just need to keep doing what they’re doing for the last 3 years and they will achieve their SOI goal of reducing subsidies per passenger kilometre.

Leave a Reply

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