In just over two weeks, on Tuesday 5th March, we will have the next census and the forms for it started to be delivered yesterday. It’s the first one since March 2006 as the planned 2011 census was cancelled due to the Christchurch Earthquakes. The data that comes from the census is important and is used by many agencies as well as private organisations to assist with in their planning. That isn’t always a good thing though as the Ministry of Transport for example, often still references the data obtained for the last census to justify their the governments stance on issues, despite some of the massive changes that have occurred with PT in the last few years. When it comes to mode share, cars will still continue to dominate the figures but the big question is how much it will change by. One area I am particularly keen to see is how much mode share and the the number of cars owned has changed in areas that have access to good quality PT.

With that in mind I thought I would look at how some of the transport stats we monitor have changed since the time of the last census.

First PT usage overall. March is traditionally the busiest month for PT and hopefully this year is no exception.  Perhaps a bit more concerning is that throughout 2012 we saw very sluggish growth, or in some cases reduced PT usage. Still, since the last census a lot has happened with our PT system. Since then we have completed the double tracking of the western line, which was only really just under way during the last census, built the northern busway as well as having seen a number of other improvements to various bus services.

PT changes 2006-2012

Next we have vehicle use. Unfortunately there are only a few points where we get consistent information which makes things a bit harder. The main source is traffic volumes over the harbour bridge and 2006 was when they peaked. At the time the bridge averaged almost 169,000 vehicles per day however as of the end of 2012, that was sitting at just over 157,000.

Related to vehicle usage is the price of petrol, something that has gone up substantially in recent years with it now not uncommon to see prices above $2 per litre. Over the just the last week, prices seem to have shot up quite a bit and are now sitting just below their all time highs at $2.20 per litre. I filled up just last week out west for $1.99 although the west and south do tend to have lower fuel prices. It will be interesting to see if that that has an impact on PT use. Back in 2006 petrol was around $1.46 which inflation adjusted works out at around $1.72

Are you a numbers person? if so what numbers from the census are you most looking forward to finding out?

Share this


  1. Re: the statistic I’m most looking forward to, it’d probably be the changes in population density at the meshblock level from 06 to 13.

    One reason to not get our hopes up too high is that the census measures only journey to work. So all those students who are now catching PT to uni will not count as PT – they’ll be recorded as whatever mode they used to get to their part-time job.

    1. Isn’t your ‘work’ as a full time student your study? Or does it specify paid employment?

      if so transport needs for anyone in the voluntary sector, care of relatives, etc, or of course students is intentionally not captured by the Census?

      1. Nope, it only counts paid employment; doesn’t look at travel for study, voluntary work, etc. Various groups have repeatedly raised this issue when the Census content has been reviewed each time; “main means of travel to education” in particular would seem a particularly important one. Stats NZ keep resisting it, citing loss of data series continuity – that’s silly; they’d still be able to maintain the “travel to work” data series, but we’d also have a new “travel to study” series as well. They also say that the data is already captured by the MoT Household Travel Survey – sure, 4800 households nationwide is going to allow us to drill down to local areas with great precision…

      2. Yeah its quite simple Patrick unless you “worked” sometime in the 7 days up to/including the Sunday 3rd of March in some kind of paid employment (for at least 1 hour or more) or Farm work! you don’t get to fill in Questions 33 to 42 about income, where you “work” and what you do, or the mode you used to cover the greatest distance to work on Census day.

        You can see a sample copy of the 2012 individual Census form here:
        See question 41 for the travel question, and Q32-42 for the related ones about work etc.
        Q41 is pretty much as Stu alludes to as to modes.

        So maybe we need to come with a scheme to employ a few of those otherwise ignored tertiary students for 1 hour at least in the lead up to Census day so their PT voice counts?
        Hmm, if we paid some of them $1 an hour to go to their nearest train station for the AM and PM peaks the Tuesday before Census daya, and then travel to the NEXT station up the line, alight and then count the numbers on and off each train they;d be in paid work, and travelling to to their work by train! Then on Census day they could truthfully answer Q’s 32-42).

        At least that way you’d get some accurate stats of PT usage (trains anyway) that could later tie to the census numbers on the same day for a few hundred $ at best.
        And if repeated on Census data would also allow cross checking of the leakage rate on PT usage if you cross checked it with HOP card usage and paper ticket sales on the same day.

        Anyone know if AT planning something like this for Census day – be a golden chance to really confirm PT numbers. Good use of my rates $ as far as I can see.

        Lastly, like Stu, now I have bought a new (electric) Bike, I’ll be cycling on the 5th to my work (come hell or high water, hills or dales), to aloow me to put another tick in the Bicycle column
        – it will be the second time ever I’ve ticked that box in the Census forms I’ve filled out the last time was 19 (cough)-ty 1. Man the times really are a changin’ back.
        Oh, and yeah I was told when I bought mine that Mayor Brown bought a pair of electric bicycles as well recently. I immediately asked what colour they were and no, they weren’t “Brown” to match his surname.

        And of course, I’ll do the Census online like last time – its very fast to do it that way and you know it will save some poor bugger having to read your hand writing, code it up and then punch your census form. And hopefully will provide some “fast” stats soon after Census night as well from the online respondents – maybe enough to shake up the local Government election later this year as to PT usage? Even if the results aren’t official they may provide some facts for to the naysayers for a change.

      3. I tired to get a co-ordinated request across the major councils for a “journey to study” in the 2006 Census, Stats are averse to adding questions and complication to the Census and I had proposed a very simple addition to workplace definitions that would minimise additional data collection and complexity

        but I think what killed it was certain councils south of Auckland didn’t understand this issue and asked for masses of additional data

        at the time, comparing the JTW across the bridge with the (then) ARC cordon count on Fanshawe St (taking the Ponsonby off-ramp into account) suggested that in very broad terms (apples and kumquats) about a third of passengers in the AM peak could have been students

          1. Can’t see it flying. There’s a legal compulsion to complete the Census that makes it an almost 100% (for all the publicity given to the handful of people who object, even if there were 4,500 who filled in nonsense or didn’t complete that’d still be only 0.1% non-compliance) dataset. A government of any stripe would, quite reasonably, be very hesitant to make a subset of the population subjected to legal compulsion to complete a second survey that’s not also controlled by central government for administration and/or content, and without that compulsion it would be of very questionable value.

          2. that’s a very interesting question actually – and definitely one to consider. Especially if you integrated with existing SNZ census infrastructure you could start to build up a more detailed picture of what’s happening in urban areas, which do tend to change more rapidly than the rural areas. Maybe we need a census of major urban areas timed to happen approximately halfway between the national census?

          3. Yeah Stu that is what I was thinking. An alternative would be that as surveys increasingly get taken online, the official census could be moved to every four years. Combined with the potential move of the parliamentary term to 4 yearly, it being done in the middle of each government cycle could be useful ensuring that every election electorate boundaries are appropriately adjusted as well as giving governments, or potential governments better information as to ongoing trends.

        1. I use Stats and census data all the time, and to be honest I see very little value in changing census questions, or having an “Auckland-only” one. Five-yearly is pretty good in my opinion, and data comparability is an important consideration. That “travel to work” question has been in place for decades. As for adding additional questions, well I’m sure there are plenty of different groups who’ve made hundreds of suggestions for new questions. But there’s a fine line between wanting to get more information, and not wanting to piss people off from their census forms being too long, and possible drops in the quality of completed forms.

          I imagine that the data on student PT travel is already fairly well-known by AT, since the vast majority of students are likely to have availed themselves of the 40% discount. That won’t cover all their travel – from memory, you can’t get the discount on all types of ticket, so students might often have two cards, one with the student discount and one with only the standard discount – but it’s a good start. People could make an OIA request if they were interested?

          1. That AT data doesn’t capture cycling and walking. The census numbers are used as absolute fact to shape and defend policy. This looks like a huge distortion in the information that drives public spending in a crucial area.

            An additional couple of questions to capture the growing education sector seems like a net benefit against the cost of what? “not wanting to piss people off”.

          2. I’m not arguing against the value of the question Patrick, it does have value. But I’m sure we could come up with dozens or hundreds of other interesting questions, many of which have nothing to do with transport. E.g. a lot of people were keen to get a question in about sexuality, which was denied. They can’t put every good question in, it’s just not possible, and there’s no particular reason why our particular interest would be prioritised above others.

            It seems unnecessary to change a national census for the sake of asking a question which only applies to the fairly small percentage of New Zealanders who are students in Auckland, when such information could be quite easily collected by survey. I know MIT carries out such surveys, and chances are many of the other institutions do too. And another requirement of the census is that it’s supposed to collect information which cannot otherwise be easily collected, so questions which apply only to a small subset of the population would normally be covered through surveys instead. I would argue that this is more efficient, especially when it’s a fairly localised group like students.

          3. I’m not suggesting limiting anything to Auckland, there are students all over the country and others who move for purposes other than paid employment. And their needs ought to be considered too. Regardless of other issues that may or may not be included it is clear that this is both an important issue with an easy one question fix. Information is great but incomplete information is distorting.

    2. When I filled out the census last time I am pretty sure that I wrote my occupation as “Tertiary Student” and I was one of those that caught public transport to work.

        1. Or I made a mistake back at the time I filled it out? I can’t remember what I filled out back in 2006, but I think I filled out tertiary student (hence why I said I was “pretty sure”). But still, what a lovely, understanding and constructive response from a contributor to this blog…

          1. Ah if it is meant in humour then no problem. Acerbic humour doesn’t come across so well on the net, so you’ll have to excuse me misconstruing that one!

    3. I agree but even so I think there are a decent increase in the proportion of people who now use PT to get to work compared to 6 years ago. I know that in 2006 both my wife and I were using a car, she will be using the train and I would be doing the same if I was working.

  2. I’m interested in seeing cycle and walk mode share (obviously)! Although I do think that a one day count is not that reliable because it could either be really sunny (which will skew numbers up) or rainy (which will skew them down). i’m also very interested in seeing the analysis of the distance people are traveling to work and if that has changed much or increased in Auckland over the last 7 years. My guess is it probably hasn’t really for most people.

      1. But you can write it in, along with any other religion you like.

        Would that mean you could legitimately put in “Better Public Transport” as your religion if you wanted?

      1. The change in religion is one I’m looking forward to having a look at. There’s likely to be a relative increase in religions practiced in Asia, and an increase in the no-religion category.

      2. Well, Stats said the same thing about people filling in “New Zealander” as their ethnicity, but changed their minds once it was ~10% of the responses.

        I think the “Jedi” joke’s over though, and I doubt as many as 10% of people will give it as religion.

  3. Petrol price is a very important issue because we are currently in a bit of a fool’s paradise with regard to exposure to the international price of crude; our record high local currency is sheltering us from the changes this century to the traded price. This means that we have a double vulnerability to price rise. It can come from either further crude rises or a drop in the NZD towards its historical average, or of course both at once.

    And if both were to happen together the change could be very sudden and, for most people, shocking. Auckland has very little structural resilience to this threat having under invested in alternatives to private vehicle dependency for decades now.

    We all need to hope that sudden rises in pump price do not happen any time in the next few years as none of the Transit systems are at all ready to deal a sudden jump in demand. Not rail till the new trains are all operating, bus until the new network is bedded in and not the whole system until the fares plan is operating smoothly.

    And we know from previous sudden petrol price rises that people suddenly try to find alternatives. But most don’t consider these until they find they have to. Unfortunately the choices made for us all by our leaders have not provided much in the way of satisfactory options for much of the population.

    1. This is a very important point Patrick, and one which most NZers are probably blissfully unaware of. Although the bliss part could run out.

  4. On a side note, I see that Stats are currently going to be running the next census in 2018 (5 years after this one), and presumably every 5 years thereafter that.

    I think though that this move will then create a series break from the previous “01” and “06” census years, meaning directly comparing census data from this and subsequent one with earlier ones much more dififcult. The alternative could be to treat the 2013 census (delayed from 2011) as a one off, and do the next one in 2016 putting the censuses back on track on the 01/06 cycle.

    For the planners and users of Census data amongst you, what is your preference?
    – two disjointed sets of census figures from 2013 and later
    – or continuing on the the 01/06 cycle from 2016 with just an “aberration” for the 2013 census to manage?

    Interested to know your thoughts on which way you’d prefer how the census data series went.

    I fear that introducing a “break” in a series will require a significant reversioning on old data to allow alignment from pre-2013 census with later ones, along the line of changing to Decimal currency.
    But since in this case, the change is not really going to be a permanent one would it not be better to use the 2013 data and the 2006 data to synthesise a pseudo 2011 data set which allows a continuous series, with only the 2011 data set being the interpolated one. Of course, might need an act of parliament to do this, but we have time (and 1 election) before then, and in any case its never been a problem to fudge the figures for this government, so why not fudge the Census figures retrospectively for 2011 as well?

    PS. I am sure that Stats could clear the decks of the 2013 census in time for 2016, especially now at least 1/3rd of it will be entered online by Census night.
    [and I suspect a good deal higher may well be the case mean in even less post census data cleanup/data entry].

    1. Hmm … the change in timing is not that much of an issue really, because you just adjust your calculations for the slightly longer time between 06 and 13.

  5. it will be interesting to see if the reduction in kilometrees driven is matched by a reduction in cars owned per household, if so, it’s likely a localised effect in urban centres

  6. Currently working part-time (after hours) delivering the census forms.

    Have to say, has really given me some great insights into my local community, just having a 2 minute chat with everyone on their doorstep!

    The pay works out pretty marginal (thinking will be between $15-20 an hour by the time I’m done), but very pleasant work, real eye opener to who and what is in my neighbourhood!

    1. “The pay works out pretty marginal (between $15-20 an hour)” – Believe me, there’s a load of people around here who would be very glad to earn that!

    2. So Adam,
      Do you get a flat fee for the entire census, or is it based on the time you spend on it?

      If a flat fee – if you are lucky you could “offload:” your 200 or so household census area pretty quickly if everyone was home this last weekend, then have nothing to do until Census day right?
      And then if you’re really lucky everyone is home when you come to collect them so its a simple process done quickly.

      So it could be a real easy job, but if no one is home and you have trouble tracking them down, could be a frustrating time wasting exercise right?

      I was thinking that as I saw the Census person making their way around Stonefields – about half the houses where she was were unoccupied (with “SOLD” stickers on the windows), and the other houses there all appeared to have people at home on Saturday so it would be a pretty quick and easy job round there I expect.

      1. Hey Greg, Flat fee, and I’m calculating works out to $15-$20 per hour if work hard.

        Not really possible to do all in a weekend, I spent 2 8-hour days this weekend, and delivered 150 houses, out of my 350 (average sized area for suburbia)

        If everyone does it on the internet, I wont have to pick up a single form, last time around 30% did internet, so should be a fair bit higher this time (last time was ’06), as almost everyone is on the net now.

        Am very happy with the area I got, high density resi, low’ish economic demo, so lots of people home, glad not in stonefields, or in the city (would be chaos trying to do census at a hotel!)

        Bob, yep, that didnt come out right, $15-20 isn’t to be sniffed at, and I’m grateful for the extra cash, and the experiences 🙂

      2. “If a flat fee – if you are lucky you could “offload:” your 200 or so household census area pretty quickly if everyone was home this last weekend, then have nothing to do until Census day right?
        And then if you’re really lucky everyone is home when you come to collect them so its a simple process done quickly.”

        If only! – last weekend I visited 56 dwellings and had no reply from 21 of them. They will need a 2nd and possibly third visit.
        Today, I visited 45 and had 16 no replies (this was between 5.30 and 7.30) – not so cushy really.
        Each collector has about 300 dwellings not the 200 you mentioned. But I have to say that everyone who was home has been very accommodating – not one refusal (so far).

  7. Through historical accident I have never been in NZ on census day and have never filled out a census form! Come March 5th I shall once again miss out, flying out of Auckland that very morning! Statistically I don’t exist, and haven’t done since I was a child 🙁

      1. Because the forms are to be filled out by “persons spending the night in this dwelling on March 5th”, one problem with the ‘snapshot’ approach of the census. On March 5th I’ll be spending the night in the cheapest flight to Singapore money can buy.

        My flatmate will have to complete the brown dwelling for without me on it.

        I suppose I could lie and fill it out as per a usual weekday, but that is technically against the law.

        1. It also creates problems for Stats NZ when they’re trying to estimate NZ’s total population, which they try to bring up to speed correcting for census undercount – people who just got missed, or who were overseas on census night (including yourself). There are actually pretty good records of how many NZers are overseas at any one time, through the departure forms you fill out, and Stats will use those. They also do a Post Enumeration Survey after the census to try and figure out how many people they missed.

          You’re correct that it’s best not to fill out the form.

Leave a Reply