Like I’m sure many of you are, I have been getting increasingly concerned about what is happening with rail patronage this year and it seemed like as soon as the RWC finished, the growth just hit a wall. The 12 month rolling patronage total is likely to drop quite a bit over the next few months as we finally work the RWC boost our of our system. A graph of rail patronage is below:
Patronage had been climbing pretty steadily for some time, it then took a sharp jump due to the RWC before hitting a plateau. The worrying thing is that this stalled growth could easily be picked up on by the likes of the government to claim that projects like the CRL are not needed. Yet while the graphs show that patronage has stalled, the last thing you could say is that the trains aren’t busy. I would regularly notice that people couldn’t get on to trains from places like New Lynn in the mornings due to them being so full. I would suggest that many of those people probably changed modes but would happily go back to trains if there was the capacity This got me wondering if perhaps the reason for the stall is not that we have soaked up the demand but that there simply isn’t the services available for people to use when they want to. Out west for example, there hasn’t been a timetable that added new services since September 2010 (and new timetable actually takes some away).
So after thinking about this for a while I decided that I wanted to see if there was any kind of correlation between the number of services that we run and our patronage. I only had a few copies of old timetables but with the help of some readers and Auckland Transport, I managed to get my hands on timetables from as far back as 1998. I started by looking at how many services are/were run each day and from there worked out how many services were run each month taking into account things like extra Friday services, weekends and public holidays. What I haven’t been able to include is the extra services run for events like the RWC or times when trains are running for upgrade works. Due to the erratic nature of comparing month to month I have looked at the 12 month rolling totals and after plugging in all of the numbers for as far back as I have patronage data I got this:
As you can see there does appear to be a pretty strong correlation between the number of services that are run and patronage. It really does suggest that one of the reasons we are seeing a plateau in patronage is very much likely to be at least partly the result of there not really having been any additional services put on for roughly two years. It is also perhaps worth noting that the timetable changes of the last couple number of years have not really added in any new peak services with the main increases being in the shoulder peak. This suggests there are probably quite a few gains to be had by continuing to increase not just peak services but also off peak services.
Of course we can put on more services but the next thing I wondered was what how the additional services that were put on impacted on their utilisation. In effect were we making the trains less efficient on an average passenger basis or were they any trends that we could see. So next I graphed the average number of passengers per service to see what the outcome was, I expected that as we added services to the network that their average utilisation would drop but then recover over time. What the graph shows though while that does happen to a degree, since we started improving our rail network not only have the number of people using the trains increased but so has the average number per service. This is shown below:
I suspect that part of the reason for the average numbers to increase over time is that as more services are put on, especially off peak, the rail network becomes a more viable option for a wider variety of trips. I was also interested to see how each line fared. It is a bit hard to separate out the Southern, Eastern and Onehunga patronage completely so I have kept them together. As expected the the Western line got a lot larger bump due to the RWC but even without that it seems to have been performing better. I wonder if that is due to have more destinations along the line providing better off peak patronage?
So if frequencies are key to boosting patronage what could we expect for the future. Well the RPTP gives us a clue as to what is planned once we have our electric trains rolling and it appears that we will see much higher off peak frequencies than we have now.
Using this data I made a little timetable to try and work out how many services we would actually have. This came out at roughly double what we have now so assuming we see a similar trend and a similar level of service utilisation then it appears we can roughly expect patronage to double to around 22m which was about the figure that was suggested we would reach by 2021 in the CRL business case. I do actually think that the utilisation number will probably increase a bit further though once we get our new electric trains due to the changes in the bus network, them being a more attractive choice and their larger capacity. With an average of say 130 passengers per service, this would see patronage on the rail network up around 27m per year and in desperate need of more services that can only really be provided with the CRL.
Processing...
Great work Matt. I’d be interested in any data you have on weekend trips -are weekend trips and patronage growing or are they flat lining? Also interested to see how much patronage is affected by the rail upgrades that are happening every other weekend at the moment.
Unfortunately I don’t have any info on weekend trips or the impact from network shutdowns so can’t comment on those. I’m pretty sure patronage drops dramatically when rail buses though.
Interestingly I did one years worth of timetables for Wellington just to see what the current utilisation was for them. Overall very similar with them averaging 102 people per service vs Auckland at 108
We have caught the rail bus once from Manurewa to Papatoetoe return and it was horrid. If the whole line was out, I would prefer to drive into town or defer travel to another weekend if possible.
Isn’t the key period of time the peaks though? Even if the average load per service is 100, if the key peak trains are full and standing every day, potential patronage will be crowded off.
Data on passenger numbers per train or data per time period is going to be the key to understanding what is happening. It could be that judicious reallocation of units could ease the crowding of key services until the advent of electric units.
Yes and no. There have been ~4tph arriving at Britomart from the western line at least as far back as early 2005 and that is still the case today. Sure the trains have gotten longer and that has helped and I don’t doubt that it is the peak usage that holds the numbers to the level they are now but it is not the peak alone that has been driving the growth. As I’m not working at the moment I have been catching trains during the day more and noticing that most carriages tend to have 30 or more people onboard which means that for most trains we are talking 100+ passengers in the middle of the day. Would as many people use the trains if they had to wait an hour or more for a service during the middle of the day?
Matt, your proper measure here is not services, but seat kilometres, i.e. services times train capacity times kilometres travelled, or you could simplify it to services times capacity, which would be simpler but a reasonable proxy for this execrise
Seat km is a worthwhile measurement for looking at the overall system utilisation but that isn’t something people think about when trying to decide if they will use the train or not. If we are trying to attract new passengers to the network then they aren’t going to know how many seats are available, they will look at the timetable to see if the times that trains run suit their needs.
I also don’t know just how many seats there are on each service so it would simply be impossible for me to work out with the data available. I did consider looking at pax per route km but that would have taken a lot longer to work out.
Matt, it’s also a measure of system capacity, in the second graph, sheduled services and patronage pretty much ran in parallel except for a leap in 08/09
so it’s worth asking if this shift was due to more total capacity in the system, i.e. carriages added to existing trains, or if there was another explanation, I appreciate that you might be working with limited data available to you
Bear in mind that the peaks only last four hours a day, five days a week. That’s twenty hours out of about one hundred and twenty of service. There is a lot more patronage growth to be had from all-day every-day services.
And enormous potential for growth on weekend Western Line services, something that AT stubbornly fails to recognise, citing electrification work as their excuse. The frustrating thing is that electrification of the Western Line could have been substantially finished by now, but it seems to have petered out over the past couple of months, with work left bizarrely half-completed. People won’t even think about catching trains when there’s an hourly service because it limits flexibility; and they certainly get annoyed when they board a 2 car ADL to find it packed to the gills as is often the case, particularly on Saturdays.
I’m not so sure about the plateau. I think a straight line fit through the 12-month rolling data suggests that the latest data is roughly in line with a constant passenger increase, and that the ‘plateau’ remains the result of the RWC data. If the 12-month rolling was the same six months from now then I’d be more concerned.
We haven’t had Septembers numbers yet to so the RWC figures are still in the most recent data
Axio you are right. But, and I think Matt’s work shows this, it is likely to stay flat all next year and well into 2014 because there simply won’t be the services to carry new demand until the EMUs are up and running.
Patrick, certainly agree with you. There are simply not the services to carry any new demand between now and 2014 and that coupled with the RWC Opening Night stuff-up, the poor customer service Veolia persist in providing, the lack of service functionality in the new ticket machines, the continuing HOP card fiasco (will only get worse once its launched on 28 Oct), a lack of properly gated stations, ongoing fare evasion and the continuing inability of AT to properly engage with its customer base, will see patronage effectively ‘flat’ until the EMUs come on stream in 2014.
Although I hate the cliche, it is definitely a case of “perception is reality”. As a seasoned marketing manager, I just cannot understand why AT continues to effectively feed this perception – their marketing, PR and overall customer communication are unbelievably inadequate. Until AT’s marketing and PR teams are cleaned out and people brought in who really know how to reach customers, keep them engaged and keep them loyal, the next few years are going to be very depressing ones.
Rob, to present the an alternate point of view, AT has done everything it can (new trains, new HOP, new Bus Integration/Contracts) but these are seriously major pieces of work that take years. The ball is rolling but time is the problem. We want them now. It would be a mistake to advertise and get more customers now when things are already bad and more passengers would only make things worse. A major advertising push will surely come when the new EMUs arrive and there is a major increase in capacity for passenger growth.
It does worry me that AT don’t seem to put much effort into keeping statistics on how full their train services are, but on the other hand theres nothing they can do until the EMUs arrive anyway — all the trains are in use.
Wellington’s weekday evening off-peak patronage was effectively killed when they thinned the ½-hourly service (last train midnight) to hourly (last train 11pm) back in 2004. The excuse at the time was staff-shortages, but the hourly evening service has struggled on since then with seriously dwindling numbers. Whoever plans these things can’t get their heads around the fact that an hourly suburban service is a real turn-off to customers. And add to that the impediment of Wellington Railway Station not being in the part of town where most evening things happen. I fear for the future of the evening service altogether the way things are going.
Yeah I guess one of the things is that the service needs to be frequent enough that people can rely on it. A service once an hour in the evenings means that you have a long wait if you miss it so just in case you will use a different method, usually driving.
If you have old Wellington timetables or even better the number of services that are run monthly I would be happy to see how Wellington compares with the same analysis.
This post is gold. It shows an elasticity greater than 1 which is shows a severe under-serving.
Can’t catch a service which doesn’t exist.
Possibly not – remember there are many other changes that have occurred in the same period as services have been increased, e.g. ongoing station upgrades, increasing fuel prices, and also general population growth.
This “temporal correlation” of events will mean that this one-variable comparison overstates the actual impacts of service changes. But the elasticity is likely to be fairly close to one nonetheless, and far more elastic than one would expect given that typical service elasticities are between 0.4-0.6. So yes, Auckland is relatively under-served when it comes to rail frequencies. Bring on the EMUS!
Can I add my own question of whether southern line patronage may be affected at least in some part by a lack of park and ride facility in a reasonably built up area and that patronage may be limited, particularly on week days, to those within walking distance of the stations on this line.
Perhaps people prefer cars.
Yes, that must be it. Also, trains are a COMMUNIST PLOT.
I hadn’t thought of it, but perhaps you’re right.
I prefer trains and I am (according to my more libertarian minded friends) also a communist. QED.
Photo proof: http://www.flickr.com/photos/robotman/1078903192/
I honestly stand in awe of people that mine data and graph it as you have done here! It reveals a lot and I’m grateful for that! Thank you.
We might see a statistical bump from HOP (assuming anyone can figure out where to buy the d*** things) purely because they’ll be counting passengers more accurately.
I’m just extremely happy the RPTP proposes 10 minute interpeak rail frequencies during the week. Gotta support that strongly in my submission.
I agree with all the comments about 1 hour between trains, and don’t get me started on the last train leaving Britomart for Onehunga around 9 – so seeing a show and catching a train back home is just not an option, which is ridiculous – but I think with the current timetables and population around the stations you can expect usage to be static.
But as Auckland’s population is expected to grow quite a bit over the coming years, it is not only are 30 minutes trains, which run later that is required but long term it as Gordon Price said – building more places for people to live near stations.
If twice or three times the number of lived near a few of the main stations – that will be a lot more people on the trains, this is what is required to really make train usage increase – and of course make it easier for these new Aucklanders to get around in the city.
Chicken and egg isn’t it Adam, get the train service running properly and living near the stations will increase in popularity. Of course it would also help enormously if the Council’s own regulations especially around parking minimums and density also encouraged growth at these hubs; a lá Vancouver as you say.
Gordon Bennet – perhaps, but the analogy between cars and a PT network that has been neglected for 50 years by the government (despite public pressure and many studies by experts saying that shouldn’t happen), is imagine if your road was blocked and only opened every half hour and you then had a 2 minute window to get you car out of the street or it would close again.
How often would you use your car? Yes, some people would use a car regardless, but if we got the modal share in Auckland down to 50% that would be massive victory and improve the city out of sight. That would still mean 50% of people could travel in their cars.
Frequency, frequency, frequency. That is what quality PT is all about all over the world.
And would actually ‘solve’ congestion, unlike any road building. And it wouldn’t need to be anywhere near a 50 % mode shift, even 5% would make a huge and positive change in the driving experience.
Patrick – Reynolds – sure, I agree 100%. I always just like to make the point that it is not like pro-PT/cycling people are saying noone should drive their cars. Just that LESS people should drive their cars. I dont believe that 85% of Aucklanders want to drive their car to work. I am sure maybe a third of those people would like, at least sometimes, to have another option.
I just think that message isnt getting out. I think if pro-road people knew more about the modal share aims for PT/cycling it would reduce the anxiety.
To paraphrase Abraham Lincoln: “You can [have] all the people [driving] some of the time, and some of the people [driving] all the time, but you cannot [have] all [or 85% of] the people [driving] all the time.” I am sure Abe would have agreed.
For some reason people can’t seem to grasp the fact that people drive, not because its what everyone wants but because that is the one mode we have invested in making things easier for. People are logical and will do what is quickest/easiest/cheapest/most convenient. The government use the argument that we should invest in what people use today but that is a by-product of what we invested in yesterday, we should be investing in what we want people to use tomorrow.
“we should be investing in what we want people to use tomorrow”
Well said. Cars.
Gordon, it was kind of funny the first time around but now you’re just being annoying. Don’t troll again please.
I agree that building more housing near sttions near the southern line would help for rail patronage but then, of course, the value of those dwelling is likely to rise because of their proximity to the rail system and may become unaffordable to the average punter. Park and ride is also fraught with difficulties including finding the land sites and who would use them, PT users or local business employees as tends to happen at the Ellerslie station but at least they may give some option for people (like myself) who live a ways from the rail line.
We are certainly setting up Park and ride for the western line and on for the Northern expressway, so why not for the southern line?
I live in Kingsland (and have since July) and used to love taking the train to the city. Then I realised I can take the bus and it’s much faster, doesn’t have to do the ridiculous reversing at the Grafton station, and actually takes me to where I want to go (near Auckland Uni). I really think the planned connection will make a huge difference, and I hope the government doesn’t start pulling money just because it’s stagnating at the moment. Focusing on a car-only culture in a growing city seems pretty idiotic to me personally.
When the CRL is built it’ll only take 5 minutes from Kingsland to the heart of the CBD. How cool will that be?
It’ll be awesome! Won’t be for many years yet, unfortunately.