Measuring PT success

A lot of effort, and money, has gone into improving Auckland’s public transport system over the last decade – mostly in the form of upgrading the rail system, building the Northern Busway and providing the bus companies with increasing levels of public subsidy. I’m the last person in the world to say that this increased spend has been unjustified (with the possible exception of the increase in bus subsidy), but the amount of money and effort that has gone into improving public transport clearly leads to the question of “how do we judge that it’s been worthwhile?”

ARTA, and now Auckland Transport, focus a lot on patronage increases. Rail patronage has increased dramatically since Britomart station opened in 2003 – from around 2.5 million trips a year to 9 million trips a year. Bus patronage hasn’t grown much though in the same time period – as shown in the graph below from ARTA’s September monthly business report (are we ever going to see patronage reports from Auckland Transport?)

The orange lines show monthly bus patronage. While it obviously bounces around on a month by month basis, the longer term trends for bus patronage in Auckland since 2002 don’t make for the best reading.

To show this a bit more clearly, I’ve taken the statistics from the graph above and popped them into an Excel Spreadsheet. I’ve then compared how each month between 2002 and 2010 compared with the ‘average patronage’ for that month over the nine years. Where patronage is greater, the month is highlighted green, less is red and on average is yellow:

The two shaded months (May 2005 and October 2009) indicate months when there was industrial action (at least I presume that’s what caused the unbelievably low figure back in 05) that disrupted the figures. What’s immediately obvious about the statistics is the obvious question of “what the heck happened in 2005 to 2007?” The other five full years looked at (2002-2004 and 2008-2009) all had total bus patronage of above 45 million trips – and I’m sure 2010 easily got above 45 million too eventually – but those three middle years all had patronage of merely over 42 million.

So patronage levels tell us something when it comes to determining how successfully our PT system is working. It’s pretty obvious that something went terribly wrong in those three middle years: perhaps it was low international student numbers, perhaps it was rail patronage cannibalising bus patronage, perhaps it was because there were few, if any, improvements to the bus system during this time.

However, there’s something else to throw into the mix when it comes to analysing patronage, and that’s the fact that Auckland’s population continues to grow. In fact, between 2002 and 2010, Auckland’s population grew from 1.254 million to 1.459 million: an extra 205,000 people. So you would expect patronage to increase in line with population growth to be an indicator that the system is ‘standing still’, with an increase in per capita boardings being the true test of whether public transport patronage is actually increasing.

Dividing the annual number of bus trips in Auckland over the 2002-2010 time period by the population of the Auckland region shows some interesting results:
So from 36.5 bus trips per capita per year back in 2002, we’ve actually decreased to a figure of 32.5. Now obviously some of those trips would have gone from the bus onto the train – as rail patronage has increased significantly – and I’m sure the 2010 figures (which remember don’t include October to December) would show a much higher per capita figure than any year since 2004, but overall I think this provides an interesting insight into answering questions about what has happened to our bus network in recent years. (I will do something similar for total PT patronage in a future blog post).

So there are clearly some flaws in completely focusing on total patronage, without having regard to population change. But simply focusing on ‘boardings’ also has its limitations. The benefits, to reducing congestion, to the environment and so forth, of a long trip being made by public transport instead of by car, is surely greater than that of a short trip. Auckland’s roads, in particular, benefit a heck of a lot more from someone taking a peak hour trip from Papakura to the CBD on the train than they do from someone taking an off-peak trip for a couple of kilometres to visit a friend. This is reflected in the NZTA economic evaluation manual when it comes to assessing the economic benefits of PT projects, but otherwise doesn’t seem to be reported much.

One exception to the general lack of focus on public transport trip length were figures presented by ARTA to the ARC last year (on the very last page of that document) – which indicated a huge increase in the amount of passenger kilometres travelled on the PT network.

An improved rail system, plus improvements to the bus system that have focus on longer distance trips (like the Northern Busway services and improved frequencies to growing areas of Botany and Flat Bush) have contributed to making the average public transport trip in Auckland a lot longer over the last couple of years. This is likely to have had a much greater impact on reducing congestion than if trip lengths had stayed the same.

Another measure that’s often used to calculate the “success” of public transport is ‘modeshare’, which effectively calculates the percentage of people who use public transport to get to work/university/school (for those 15 and over). Modeshare’s disadvantage is that it pretty much completely ignores off-peak patronage – people using PT to go to the movies on the weekend or pensioners going shopping during the week. So while modeshare may be useful in telling us something about PT’s ability to reduce peak-time congestion (ie. if we increased Auckland’s 7% modeshare to something like Vancouver’s 16%, the roads are likely to end up significantly less congested), it doesn’t tell us much else. An example of this is the fact that Christchurch has a lower PT modeshare than Auckland (5% compared to 7% I think), but because Christchurch’s PT system works well for off-peak trips, they actually have more per-capita annual boardings than Auckland does.

A single, region-wide, modeshare figure also ignores the difference between trips made along congested corridors or accessing places like the CBD with short trips to work from one suburb to another that may not generate much congestion at all. So while Auckland’s 7% overall modeshare sounds incredibly low, if you look at what happens in important places things are quite different. Around half of all commuters into Auckland’s CBD each morning use public transport and around a third of people crossing the Harbour Bridge a peak times do so on the bus. In other words, if it weren’t for public transport we’d need twice as many road lanes into the CBD (and twice as many carparks in the central city), plus another few lanes on the Harbour Bridge.

While place specific measures of peak-time modeshare may give us more useful information about how PT is helping the city function, it does focus a lot on a narrow (if very valid) view of the purpose of public transport as a congestion reducing tool (or perhaps more realistically, a measure for shifting more people at a set level of congestion). We may want to have other measures, like what amount of CO2 emissions or other air pollutants are being reduced (if any, given the state of many of our buses) through the use of public transport. Or how many people are able to participate in society without having to own and drive a car. Or how many families are able to get by with one car, instead of two or more – and the huge economic benefit that provides them.

I think that fundamentally, the best tool to use to measure the success of public transport is dependent on what you want your system to do. What makes a good public transport system? Is it one that is attractive for people to use largely at peak times, but not so good at other times? Is it one that gets longer trips off the road? Is it one that reduces auto dependency – by that do we mean the perceived necessity to drive to work or the perceived necessity for a household to own more than one car?

The ‘burden that PT eases’ test is one that I think is useful to focus on. This would include the burden of congestion, the burden of CO2 emissions, the burden of social exclusion if you can’t drive for physical or financial reasons, the burden of having to own more cars than you want (or can really afford), the burden for businesses having to spend a lot of money on providing parking, the burden of urban sprawl that auto-dependency encourages and so forth. What this all really means is that measuring PT success is actually a lot more complicated than simply reporting a patronage figure every month. It’s also a lot more complicated than ensuring there’s a bus stop within a few hundred metres of every house in the city – as if that bus comes only once a day it’s hardly of much use.

I think perhaps the biggest burden Auckland suffers due to its below-par public transport system is the perceived need the general population has to own a lot of cars. There is almost one car in Auckland for every adult – a very high number by international standards. Now while I’m sure some people enjoy owning lots of cars, and some people might need a particular work car because of the nature of their job, I don’t doubt there’s a lot of the population out there that owns more cars than they would, if truly given the choice. So perhaps the best measurement tool for PT’s success in Auckland could be reducing our dependence on owning so many cars, and measuring this in a way that’s independent on income.

After all, 75% of Manhattan residents don’t own a car – and that’s just about the richest little bit of the entire United States.

7 comments to Measuring PT success

  • Ross Clark

    Per-capita tripmaking is a good measure of how effective a system is. If you go back to the Mike Lee paper you posted, the chart on p10 makes for fascinating reading:

    Why did I find it so interesting?

    * Auckland and Christchurch were on a par, Christchurch slightly better if anything.
    * Portland was better again – but only just! depsite its $US 2bn light rail system.
    * The Wellington region was much better again still (Wellington City proper manages about a hundred trips per person per year on its bus system, refer: In fact, it was better than its ‘big brothers’ of Perth, Vancouver, Toronto, and Brisbane.

    So, two question for open debate:

    (a) why does Wellington do better? (and these numbers don’t include the railway network); and,
    (b) does this shed any light on Auckland’s situation?

  • Matt L

    2005 was when we started to see capacity increases on the trains with the introduction of the SA trains and was when we also really started to see stations being upgraded, it was also the time that train use started really picking up. My guess is that between 2005 and 2007 the decline of bus passengers was due to them moving to trains however by 2008 most of those that could easily use a train instead of bus had done so and as such the growth of bus usage since then is a result of genuine increases in bus use.

    Considering how much investment is currently being put into rail I think we do need to look at the PT system as a whole as there was always going to be some level of mode shift however looking at the overall PT usage there was obviously a drop in PT usage from 2003 and it took until 2008 to get back to the same level.

    I have all the monthly figures for train use from mid 2007 if you want them. Lastly I hope we will see patronage figures from AT but it isn’t looking promising so far.

  • George D

    The price has increased, while the user experience has remained pretty bad. Apologies for repeating a little from the other thread:

    “The current cash fare for a single stage is $1.80…”
    I haven’t caught a bus in Auckland in a long, long time. I remember $1.20 sections in the early 2000s. The bus fare structure even then was steep.

    I think bus prices should go down. Melbourne, for example, has a very simple two stage system. Up to two and a half hours (a 7.31am ticket is valid til 10am)transport on all modes in the large zone one will cost you $2.96 using their Myki card, and daily $4.90. Compare this to Auckland, where a daily trip the same distance will cost you about $12-13 dollars on the bus (and take over an hour each way)

    If you want to look at patronage on buses, you have to look at simple ticket prices.

    I don’t think flat fares are the answer to everything – I’d like to be able to jump on and jump off cheaply – but the current system is quite broken. When you add in the other reasons why people might avoid buses (slow, uncomfortable, bad/reckless drivers, bouncing around giving a headache, lower quality of patron, poor routes) you can see why patronage has dropped radically in per-capita terms. If there is _any_ alternative to the bus, I’ll use it. It also means that bus accessable parts of Auckland are underutilised. As Auckland is now a single entity, and seeks to develop into a 2 million plus city, there’s a lot of opportunity to bring in people and parts that are currently neglected. High quality reasonable cost pt solutions (RTNs) are definitely needed.

  • It would be interesting to see what kinds of services were making money and when.
    I think the key to getting public transport subsidy down lies in:

    1. Getting pubic transport as a “general way to get around” rather than singly focused on commuter trips
    2. All day use, particularly in the off peak and weekends (because the commuter market is already largely taken by PT anyway)
    3. Cross town and radial trips (these trips’s share of the travel market is mostly still the domain of car)

    Public Transport is “successful” when you can live in a city (works, shop, play, dine, recreate) without a car and not feel like your mobility choices have been limited.

  • I agree these are pretty key. Auckland has a huge bus fleet, heaps of drivers and a whole wack of support infrastructure… but they really only get significant return on it for about two hours one way on weekday mornings, and two hours back the other way on weekday afternoons.

    As you suggest, in Auckland about 50% of CBD weekday commuters already take public transport and those services tend to be full… there is little opportunity to grow that further. The real efficiency is if we can get those same buses full in the counter-peak, off peak and weekends too.

    One thing for reducing the subsidy would be to gross contract all routes, I wonder if Mr Joyce realises that.

  • I don’t know if car ownership is a good measure, simply because people will own a car just for the option value and “civilised” value. The option value is the ability to take long weekend drives, emergency trips, holidays, outings or carry loads- and many people would have a car or drive it to the park and ride or use public transport for commuting. They don’t want to get rid of the car, they just want the public transport option when they want to.

    The “civilised value” is because people see cars as a necessity, just like people see having a roof on your house or a television in your living room as a necessity. So you are always going to get a high value if you measure that, no matter what, and I doubt that will change, so its not useful as a metric I think.

    mode share is good, but I would suggest using the off peak patronage or sunday patronage as a test for how good a PT system is. You could also put in the fare box ratio and peak to base ratio (busiest hour vs lunchtime “baseline” patronage and transfer ratio along with trips/population.

    The reason why I think these are good measures is because they show if the public transport network is being used for all day trips, not just “there-is-no-sane-alternative” access to and from work. The transfer ratio is a direct measure of integration with other modes.

    • Nick R

      It’s not so much owning a car, but having access to one which provides the “civilised value”. For example in a houshold of two people you don’t need two cars to take long weekend drives.

      So I think a good measure would be a metric of the number of cars per houshold vs. the number of people in the houshold, or perhaps simply the number of car trips made and/or kilometres travelled per year per person.

Leave a Reply