Ellen, welcome to the discussion on second residence/address.

Thank you for your input. Your suggestion will give us further clarity on the usefulness of data generated. How the question will be framed depends on user needs. What you have suggested will guide us on what variables are required to meet that need.

Again, thank you.

I have already researched unoccupied dwellings from 2013 and made some maps of them. See my website for a presentation that looked at the distribution. http://www.ollivier.co.nz/download/census/index.shtm
My conclusion is that the majority of the 10.6% are indeed in coastal communities at the beach, but where it is hard to let them out due to a lack of local jobs. Note that the counts are not available at the meshblock level.

Related to the second dwelling is the fact that all dwelling statistics are based on 'occupied dwellings' or 90% of homes. Should every series have a count for 'not measured' so that the total counts include the unoccupied dwellings? Otherwise there is a significant uncertainty that has not been quantified.
It is not evenly spread. Some communities have 50-80% unoccupied dwellings so the counts for those areas need to be discounted.

It is important to look up the definition of a 'dwelling'. It matches the valuation roll definition of a 'unit' . Correspondents may be surprised to see that each motel unit is a 'unit' because it has bedroom, bathroom, kitchen. This makes some areas with lots of accommodation look empty in March when the census is held.

Since the second address will not be visited by the census collector, how will the address be validated? I fear a lot of indecipherable addresses and false positives. Second homes will often not have a delivery address so will be missing from NZ Post. Even the council does not give these places an address, only a rating and legal description.

Kim, welcome to the discussion on second residence and for sharing your findings on dwellings.

At the moment, we are still looking at how the second residence/address question(s) will be structured and it depends on inputs like what you have mentioned. This is important for us to be aware of user needs such that the questions can be correctly framed to meet that need.

Thank you.

I'm writing from point of view of and IDI (Integrated Data Infrastructure) data analyst and assuming that 2018 Census data will be linked to IDI. For IDI data analysis, the extra address information (both the second residence and 12 months previous) will help fill major gaps in the patchwork of address information in the IDI currently. Having second residence will help avoid misleading 'non-matches' if Census and admin data are compared only on the basis of the first address given in the Census.

Welcome to the discussion Charles, and thank you for sharing.
IDI is treated separately from Census. Any linking of the two sources must undergo a privacy impact statement and should be cleared with the Privacy Commissioner and the Government Statistician. SNZ does not link Census and IDI.
For anonymized addresses in IDI to be relevant and updated, it must be addressed from the administrative data source, for instance IRD or MSD.

Hi Cecile

As an IDI user I'm aware that the Census isn't currently linked to the IDI. But I understood that a decision on whether to link the 2013 Census was pending
(i.e. aiming to have the kind of clearances you mention confirmed or denied over the next couple of months).

CS

Addresses in IDI are really such a challenge. We can do more if addresses are updated - especially those collected by IRD.

Thank you Charles for your interest.

If it is possible to get good information through census on all of the following I think adding this question to the census is a win-win.
•regional population fluctuations,
•family connections across households,
•New Zealand’s housing stock, and
•second home ownership rates.

I am sure there would be challenges with framing a question(s) but the benefits seem worth trying for. I am particularly interested in how second residence information might help with understanding of different types of families' living arrangements - especially where grandparents are raising kids, step families and extended families.

Lois, welcome to the online discussion on second residence.
We will be collating all inputs to second residence from the online discussion, roadshows and formal submissions. If second residence will be included in the 2018 Census, the number of questions will be limited due to respondent burden. Some of the information needs you've stated will be derived from several attributes. Rest assured that your inputs will be included.

The discussion so far seems to be resident-focussed - obviously a key target for Census2018. I wonder if we might discuss the topic from a residence-focus - in other words the buildings themselves. Someone owns every residence. If it's important to know (for an occupied residence) how many bedrooms, bathrooms etc. there are, Census2013 made no effort to gather that information for unoccupied residences. Is that to change for 2018? I have other questions but will leave it at that for now.

Welcome to the discussion Michael and thank you for your comments.
Second residence is not final yet to be included in the census. We are consolidating comments/suggestions as to its need and weighing the importance of each one. Your comments included.
There is another sub-topic which is about housing which is considering variables like dwelling counts, number of rooms etc.
The Household Economic Surveycounts bedrooms, kitchens, dining rooms, lounges or living rooms, and rumpus rooms. It doesn't count bathrooms, showers, spa rooms, laundries, halls, garages and pantries. For customised data you can e-mail info@stats.govt.nz or call toll free 0508 525 525.
I hope this helps.

Hi @michaellittlewood
Thank you for your suggestion to consider a residence focus for this topic.
Most of the anticipated user needs for the topic outlined in the Preliminary view of 2018 Census content were around residents (for example understanding split families and fluctuating populations). However, we would be very interested to hear more on how second dwelling information would be used. Would a count of second address/ residence and the type of residence (e.g. a holiday home, work address etc.) be of use? or would you require more in-depth information similar to that collected on the dwelling form (e.g. number of bedrooms)? How would this inform research, policies or decision making? As Cecil (facilitator) mentions, the user needs we identify through this engagement and consultation will help determine what is in scope for 2018 Census. At present we cannot fully identify what will change for 2018 Census.

Hi everyone!
Attached are the 2011 Census releases of ONS related to 'second addresses' and their final recommended questions reflected in their census questionnaire:

• Number of people with second addresses in local authorities in England and Wales
• What does the 2011 census tell us about work-related second addresses in England and Wales

• Dependent children usually resident in England and Wales with a parental second address

• Second address in England and Wales by type

Following is ONS' final recommended questions on second address_refer to pages 22-23

The Australia Bureau of Statistics (ABS) is also considering second residence as new content for their 2016 Census. In Australia:

• There is growing user interest in data about second residence, particularly in relation to: people who regularly stay in a second residence due to work commitments; children in shared custody arrangements who have an alternate usual residence from the one they stayed in on Census night; and students who regularly stay in a second residence due to educational requirements.

• Collection of information on second residence would enable improved assessment of the numbers of temporary residents located in communities and would be used for the compilation of service populations. This, in turn, would facilitate more accurate planning for services in these locations.

• This topic may also incorporate second residence ownership, which would enhance statistics on wealth and home ownership, and potentially be used to improve measures of socioeconomic status.

This is the starting point for us, but we are flexible on what questions to include once we have listed, prioritised and finalised user needs on second addresses. We are keen to hear more from you specifically on user needs.

Thank you.

Hi Cecile,

This information would be very useful for Civil Defence and Emergency Management (CDEM). When evacuations are required knowing how many people require evacuation in a particular area is critical. During the holiday off-season usually resident data is sufficient to plan an evacuation. However, during public holidays and the summer holiday season it is impossible to know how many people may require evacuation from holiday hot-spots such as Coromandel and Taupo. A lot of New Zealander's stay in their or their friend's/familie's holiday homes at these times. Knowing how many dwellings are holiday homes/baches in a location would certainly help planning for holiday evacuations.

Thanks
Derek

@derekphyn, good morning and thank you for your participation in this discussion thread.
This is good to know that civil defence can use the second address information in planning for holiday evacuations. We will try our best to recommend more detailed question(s) so that location of these addresses will be targeted and can provide relevant information in times of emergency situations.
Rest assured that your input will be included.

Definition of 'second address/residence'

Hi everyone!
I would like to know your thoughts on the proposed definition of 'second address' which is patterned after ONS' 2011 Census. SNZ's proposed definition is 30 days or more a year to identify second address.

Thank you.

In response to questions from the Facilitator Tina....

1.) what is the most useful determination of a second address eg 30+ nights per year?
- I define a second address is another dwelling owned by an individual. However, not sure if you open a can of worms with this one by also needing to ask about third address, fourth address and so on....

2.) What information people would most like to know about the characteristics and/or use of these dwellings (second address).
- Primary use (Personal holiday home, rented holiday home, rental, hostel/shelter, Business, Vacant, other); maximum likely number of occupants at any one time; average number of nights per year the 2nd address has occupants; times of the year the 2nd address is most commonly occupied (selection of more than one allowed - All year, Spring, Summer, Autumn, Winter, summer holidays, long weekends, public holidays, weekends, weekdays)

Why does it need to be in another local authority? That doesn't seem to make much sense to me, particularly in this day and age of local government amalgamation. All of Auckland is one local authority, so you immediately exclude information on someone living in the North Shore owning a second property in Franklin - even though that's the same as someone in Wellington's Miramar owning a holiday home in Waikanae.

Agree. Limiting it by location seems unusual.

@gregstephens - Welcome to the discussion thread and thank you for your comments. I'll take on board your concerns regarding limiting it to local authority which @derekphyn concurred. I agree, that with the amalgamation of local councils we might not be able to address user needs like what @derekphyn would like to address during emergency evacuation.

@derekphyn thank you for your prompt response and suggestions on questionnaire structure to achieve user needs for second residence.

Still on definition of 'second address/residence'

There are immigrants (permanent residents and citizens) who own two residences, one in New Zealand and another in their country of origin, and those Kiwis who own a second residence in other countries. They travel overseas and stay in their second address for 30 days or more - especially during holiday season. Would they qualify to say 'yes' in this particular question of having a second residence?

Let me know your thoughts on this matter.

Thank you.

Technically yes. But I don’t think we would be interested in knowing that for most things. Only if primary and second residence were both in NZ.

I think that this thread does not separate out the two proposals in the Preliminary 2018 document between a person's second residence and the second dwelling itself.

I was looking from the point of view of the second dwelling which would be useful to have a form filled in for it even if there was no one present on census night. I would not expect the second residence to be linked to the 'first residence' in any way. Maybe a main purpose for every residence would be useful [main/remote_work/holiday/boat].

It was interesting to look at the UK census where the question was asked. Only 1.5% of dwellings were a 'second dwelling'. On our unoccupied count of 10.6% we would expect it to be more than 10% in New Zealand.

The second address for a person was being asked for quite different reasons, mainly to avoid double counting where people split their domicile. I would expect these 'second addresses' to be almost all occupied. I am not sure how these second address counts would be presented. Any examples mocked up?

@kimollivier Welcome back to the discussion.

At this stage, the questionnaire structure is not clear yet and will only be finalised once we consolidate all submissions. However based on the user needs for this topic, potential questions might include:

Whether the individual/household has a second residence based on the finalised definition (yes/no). If yes:

• purpose of that second address (work, study etc);
• location (exact address);
• occupied or unoccupied and when (e.g. winter spring summer or fall) it is occupied/unoccupied

Answers to these questions can be cross-tabulated with other attributes to derive a meaningful statistics/information for an evidence-based decision making.

However, we are also cognizant of the impact on “respondent burden” as we ask more questions. We need to look at ways to reduce it and meet users’ information needs as well. Thus, the prioritisation will be a big challenge for us.

You suggested that only those dwellings that are occupied will be provided with census information. Not so, even unoccupied dwellings will receive census information. The reason for this is that the unoccupied dwelling will have some chance of being occupied during census night.

For the 2018 Census, we will be introducing address canvassing to update our addresses and map database. The address list will create a census dwelling frame that uniquely identifies every dwelling at the address, whether occupied or unoccupied. This will be used to provide census information to respondents, support accurate data collection, and monitor progress in real time. The frame will be updated right through census day.

In previous census counts if the dwelling was unoccupied the dwelling was noted but excluded from all the other variables. Only 'occupied dwellings' are included. My proposal is to get the absentee owner to fill in a dwelling form, if possible, so that the unoccupied dwellings are included in the main tables. If a respondent marks that they have a second dwelling, then could they be asked to fill in a second dwelling form if it is unoccupied on the day? I would have thought that every dwelling needs to have the purpose question, not just secondary ones because the owner might be in the secondary dwelling.

@kimollivier thank you for your continued interest in the discussion because it challenges us to think and visualise the pros and cons.

The census questionnaire will potentially include the second address questions regardless whether the respondent has a second address or not. All of us will have the same census questions to be answered on census night. If the owner happens to be in the second residence on census night, then it would be assumed that they would complete this question with their other address (or, as it might be, their primary).

What data are you specifically looking for on the unoccupied dwellings?

I am interested in everything that is in the dwelling form!
The example that I used was heating fuel. I was looking at the association of air quality with wood fuel use in small towns. I realised that a hotspot analysis was flawed because of the high number of missing records.

I am miffed that 10.6% of the data for dwellings in NZ is not only missing but is also not accounted in the main tables because the table totals are all based on 'occupied' not 'total'.

I would have expected for "statistical rigour" that the tables would include all dwellings and the unoccupied missing values would have a catchall in every table for 'unknown'.
For some areas with lots of absentee owners the percentage can rise to as high as 86% such as Kawau Island.

From other discussions, the justification is that the census is focussed on people, not infrastructure. But for dwellings, this introduces a bias. The presence of people may not be relevant to the analysis.

Currently, we have no information on the facilities offered by about 10.6% of the dwelling stock. Part of that issue is to do with the second residence question and part to do with gaps in the answers on 'housing' where there is another problem - the missing data on ownership. Part of the answer is to ask someone in an occupied dwelling if, on Census night, there is another dwelling which they own but that they know will be unoccupied on Census night. If so, they should complete a second dwelling form that gives the basic information about purpose, facilities etc.

@kimollivier
I will note all your observations especially the analysis on the dwelling form. For some users of the data, they want presence of people. For those like you, the infrastructure is more important. Both user groups are important to us, and we would like to meet that need to maximise the use of census data.

@michaellittlewood welcome back.

Noted your interesting comments about the dwelling itself (the infrastructure), especially on user requirements. These will be used to steer the development of the census questionnaire and subsequently the production of census data outputs.

As I said earlier, there are two groups who are keen on this new census content - those who are interested on the dwelling itself and another on residents of the dwelling. SNZ wants to meet both users' requirements to maximise census data outputs.

@kimollivier hi again! I revisited your last comments. For detailed info requirements, you can actually request customised tables from SNZ. There are three ways you can do this:

• e-mail info@stats.govt.nz;
• call 0508 525 525 toll free in New Zealand; or
• at the bottom of every output release, we have a form whether we are meeting your info requirements or not -

The 'contact us' which is located at the upper right hand of our homepage lists our contact details.

I hope you can use any of these so that we can serve you better on your data requirements. This holds true for census and other releases that might be of interest to you and your work.

Still in relation to your last comment, re air quality and heating fuel, SNZ will be releasing data related to the environment. This might be of interest to you. Its a joint project with MfE,

We have 15 more days to go and this online discussion about the 2018 Census content will close. I have summarised data users' requirements in this thread as follows:

Summary points

Definition: 30 days or more a year to identify second address in New Zealand

Two groups of potential data users:

Group 1: Those who are interested in the residents of the second address (eg emergency evacuation during holiday period, school administration purposes, fluctuations in regional/local population during weekdays and weekends etc). This group would like to know:

• The purpose of the second address (work, study, children living with both parents but in different addresses etc)
• Location (exact address)
• Type of dwelling (e.g. house, townhouse, bach, apartment, moveable dwelling etc)
• Period that it is occupied/unoccupied of the second residence (e.g. weekends, weekdays, school holidays etc)

Group 2: Those who are interested in the dwelling (or the structure, the building itself). This group are those who want to use Census data to plan, develop and build infrastructure (e.g. drainage, sewerage, environment, housing construction and development etc). They want to know about:

• Housing stock – occupied and unoccupied
• Location (exact address)
• Type of dwelling (eg house, townhouse, apartment, moveable dwelling etc)
• Number of storeys/levels (one, two, three...)
• Number of bedrooms in the dwelling
• Number of rooms in the dwelling
• Heating fuel for the dwelling
• (all other questions reflected in the dwelling form)

Let me know your thoughts whether I have captured your data requirements. Thank you all!

I have said this in another forum, and I will repeat it here to provide balance for the summary above.

I cannot understand how knowledge of a second address that is a holiday home will practically assist any planning, be it for emergency or other purposes.

Without knowing how many people actually stay in a holiday home at any point in time, it would potentially be dangerous to make assumptions about the level of service required. I feel it is beyond the scope of the census to expect people to provide details on how many whanau or other guests use a holiday home during any given period.

The quality of information provided is likely to be close to meaningless over the period of five years (who can say when they are going to stay in a holiday home with how many others more than a week or two into the future?).

I believe that the census is not the right tool to provide the information listed above, and given that the number of questions that can be asked is limited I feel that there are other questions that would provide more meaningful information and therefore have more merit.

Thank you @suzannebecher

There is indeed a limit to what we can ask in Census. Most of all we have to be cognizant of the impact on respondent burden.
As I said previously, we have to balance everything - requirements and respondent burden.

Thank you for sharing your thoughts. Highly appreciated.

Formal submission to second residence/address

Public engagement and consultation on the 2018 Census
content will close on 30th June 2015. There’s been some great discussion in this thread so far and just wanted to encourage all of you to make a formal submission through these links before our engagement and consultation closes. All the discussions on loomio will be formally assessed but sending in a formal submission will add weight. Links as follows :

• Formal submission
• Content determination framework

Thank you.

Hi there. Just belatedly adding some thoughts from the Office of the Children's Commissioner. We definitely support the inclusion of a question that can capture the dynamic nature of family life in 21st century New Zealand, particularly the reality that many children spend time in more than one house. This is related to the discussion about recording stepfamilies also. If this second address question is included, thought should be given about how to ensure it captures information about how many children live at more than one address. Thanks.