Report to the Minister of Statistics: Submitting Cabinet paper
The 2018 Census – Submission of Detailed Business Case and
Release of Tagged Contingency Funding to Cabinet Economic
Growth and Infrastructure Committee
Date:
17 June 2014
Priority:
High
Security level: In confidence
File number:
MM1377
Action sought
Action
Deadline
Minister of
Note the contents of this briefing;
18 June
Statistics
Sign the attached Cabinet paper and the
Cab 100 declaration;
Submit the Cabinet paper and
The 2018
10am,
Census Focussed Detailed Business Case to 19 June
the Cabinet Office, for consideration by the
Cabinet Economic Growth and Infrastructure
(EGI) Committee; and
Note that copies of the Business Case will
be made available at the EGI Committee
meeting for your colleagues to refer to, with
the Executive Summary made available as
Appendix 1 to the above Cabinet paper.
Contact details
Name
Position
Telephone
First
contact
Sean
Private Secretary to (04) 817 9633 021 244 0895
Broughton
the Minister of
Statistics
Carol
Deputy
(04) 931 4947 021 946 604
Slappendel
Government
Statistician,
Operations
Recommended action
It is recommended that you:
1.
Note the contents of this report;
2.
Sign the attached Cabinet paper
The 2018 Census – Submission of Detailed
Business Case and Release of Tagged Contingency Funding and the Cab 100
declaration;
3.
Submit the attached Cabinet paper and
The 2018 Census Focussed Detailed
Business Case to the Cabinet Office by 10am 19 June 2014, for consideration by
the Cabinet Economic Growth and Infrastructure (EGI) Committee on Wednesday
25 June 2014; and
4.
Note copies of
The 2018 Census Focussed Detailed Business Case will be made
available at the EGI Committee meeting of Wednesday 25 June 2014 for your
colleagues to refer to, with the Executive Summary made available as Appendix 1
to the above Cabinet paper.
Carol Slappendel
Hon Nicky Wagner
Deputy Government Statistician
Minister of Statistics
Date:
Date:
Report to the Minister of Statistics: Submitting Cabinet paper
The 2018 Census – Submission of Detailed Business Case and
Release of Tagged Contingency Funding to Cabinet Economic
Growth and Infrastructure Committee
Purpose
1.
To seek your approval of, and signature on, the attached Cabinet paper
The 2018
Census – Submission of Detailed Business Case and Release of Tagged
Contingency Funding, for submission, along with
The 2018 Census Focussed
Detailed Business Case, to the Cabinet Economic Growth and Infrastructure (EGI)
Committee meeting of 25 June 2014 and subsequently to Cabinet. The Cabinet
paper is due to the Cabinet Office by 10am on Thursday 19 June 2014.
Background
2.
You have considered previous drafts of the Cabinet paper and Focussed Detailed
Business Case for the 2018 Census and provided feedback to Statistics NZ
officials in your meeting with them on 11 June 2014.
Main changes to the Cabinet paper
3.
Statistics NZ has redrafted the Cabinet paper on the basis of your comments of
11 June 2014, and has consulted with monitoring agencies on these changes.
4.
The Cabinet paper has also been considered by the Officials of the EGI
Committee (OEGI). OEGI provided comments on the Cabinet paper text, which
have been incorporated, particularly paragraph 39 regarding the potential for use
of RealMe in the 2018 Census.
5.
The main changes to the Cabinet paper are:
a reduction in the overall length and general rewording, along with reordering
and the insertion of sub-headings, to improve the flow and tone of the paper
removal of the section on commercial viability, with incorporation of its key
points into a new section on information technology (starts paragraph 36)
insertion of a section on the links between the 2018 Census and Statistics
2020 programmes (paragraph 41)
reduction of the discussion on analysis of variations to the 2018 Census
(paragraph 21), with detail on this work retained within Appendix 1 - the
Executive Summary to the Business Case
6.
OEGI also suggested you may wish to discuss this Cabinet paper with the
Ministers of Finance and Economic Development prior to the EGI Committee
meeting. Copies of the final draft of the Cabinet paper have been provided to your
office to inform your colleagues prior to these discussions.
Office of the Minister of Statistics
Cabinet Economic Growth and Infrastructure Committee
THE 2018 CENSUS – SUBMISSION OF DETAILED BUSINESS CASE AND RELEASE OF
TAGGED CONTINGENCY FUNDING
Proposal
1. This paper provides an overview of the 2018 Census focussed detailed business case,
completed by Statistics New Zealand (NZ) in response to a direction by Cabinet in February
2014, and seeks approval to release funding, currently held in a tagged contingency, to
deliver the next census in 2018.
EXECUTIVE SUMMARY
2. Following submission of a single stage business case in February 2014, Cabinet agreed, in
principle subject to Budget 2014 decisions, that a modernised census be held in March 2018.
Budget 2014 decisions have now been taken, with the first year of funding (2014/15) for the
2018 Census to be released on 1 July 2014 and the remaining funding held in a tagged
contingency to be released following submission to Cabinet of a focussed detailed business
case.
3. The census model needs to change if it is to be sustainable over time. Transforming the
census to a model based solely on administrative data is under investigation for longer term
change. In the short to medium term, modernising the survey based census model will
significantly improve the efficiency and the effectiveness of the census.
4. The key features of a modernised 2018 Census include:
a “digital by default” strategy (with an online completion target of 70%, resulting in
significant reductions in field staff and a model less sensitive to population growth)
using real-time management information to direct field effort to those least likely to
respond (targeting resources at improving response rates for hard-to-reach populations)
efficient re-use of administrative data to support collection and processing of the census
(generating savings and assisting longer term transformation)
revising census content and providing earlier, more flexible census outputs (to better meet
New Zealand’s changing information needs).
5. In preparing the focussed detailed business case requested by Cabinet, the Executive
Summary of which is attached as Appendix 1, Statistics NZ has confirmed that the
modernised 2018 Census is the best available method to deliver changes needed for the
next census. It provides the best value for money for the next census and further censuses
and is commercially viable. The modernised census model is estimated to produce savings
(compared to 2013 costs) of at least 5% over the 2018 and 2023 census cycles, with current
modelling estimating savings as high as 11%. Statistics NZ has strong management
structures in place and planning in train to ensure successful delivery of the 2018 Census.
6. While a revised quantitative risk analysis revealed potential for increased costs, Statistics NZ
has since reduced some cost components and identified options to absorb cost pressures,
should they arise, within the Vote Statistics baseline and is confident it can successfully
deliver the modernised 2018 Census within the total estimate of $118.9M submitted for
Budget 2014. This amount includes a $5M project contingency, held for release at the
discretion of the Ministers of Finance and Statistics.
1
BACKGROUND
7. The New Zealand Census of Population and Dwellings (the census), conducted since 1851
and a legislative requirement under the Statistics Act 1975, produces the authoritative, and in
some cases the only, comprehensive source of population and dwellings statistics including
information on local communities and small populations. The 2018 Census will be the 34th
census conducted in New Zealand.
8. On 19 February 2014 the Cabinet Economic Growth and Infrastructure (EGI) Committee
considered options for the next census [EGI Min (14) 2/11 refers]. Three options were
presented: a ‘status quo’ census in 2018, a ‘modernised’ census in 2018, and a ‘modernised
plus’ census in 2021. A single stage business case was submitted, with a modernised 2018
Census as Statistics New Zealand’s preferred option.
9. The EGI Committee agreed, in principle, subject to Budget 2014 decisions, that a
modernised census be conducted in March 2018, with updated content and increased use of
technology and administrative data. The EGI Committee also agreed to a work programme
for further investigation of longer term options with a final report on census transformation
investigations, sufficient to decide a preferred future direction for the New Zealand Census,
to be presented to Cabinet by 31 October 2015. Cabinet confirmed these decision on 24
February [CAB Min (14) 6/8 refers].
10. Cabinet approved in Budget 2014 [CAB Min (14) 13/8 (31)] a total appropriation of $7.9M
capital and $111.0M operating costs, to deliver the next census in 2018 over 5 years and to
continue the Census Transformation work programme over 18 months. This funding is
allocated as follows:
$13.099M operating and $0.547M capital for 2014/15 in Vote Statistics, for the 2018
Census programme and Census Transformation work, and
the remaining $105.3M funding, from 2015/16 to 2019/20, is allocated to a tagged
contingency to be released on Cabinet’s approval of a focussed detailed business case.
11. In association with Budget 2014, the Minister of Statistics announced, on 19 May 2014, that
a modernised census will be held in 2018.
CONFIRMING THE CASE FOR THE MODERNISED 2018 CENSUS
12. Statistics NZ has developed a focussed detailed business case on the 2018 Census in close
consultation with monitoring agencies. The business case demonstrates that the agreed
option, a modernised 2018 Census:
is supported by a robust case for change - the 'strategic case'
optimises value for money - the 'economic case'
is commercially viable - the 'commercial case'
is financially affordable - the 'financial case', and
is achievable - the 'management case'.
13. This Cabinet paper focuses on the key points of the business case. The Executive Summary
of this business case is attached at Appendix 1.
2
Why change the Census model?
14. High quality, relevant census information is critical to New Zealand, with an estimated net
present value over the next 25 years of close to $1.4 billion1. However, the census model
needs to change if it is to be sustainable over time. Compelling drivers for change include:
rising costs and the potential for efficiencies
the ongoing trend of declining response rates (particularly for small sub-groups and small
areas)2
the increasing availability and sharing of administrative data across government, and
technological advances.
15. Transforming the census to a model based solely on administrative data is under
investigation for longer term change. In the short to medium term, modernising the survey
based census model will significantly improve the efficiency and the effectiveness of census
and provide a step on the way to a census based on administrative data. Work will be
undertaken within the 2018 Census programme to further test the potential of replacing
traditional census collection with administrative data, building on work already underway
comparing the quality of current administrative data with 2013 Census data.
The key features and benefits of the modernised 2018 Census
16. Modernising the census will be achieved through a fundamental shift in the way people are
contacted and respond to completing the census – away from a large force of census
collectors delivering paper forms to every household to the majority of people completing
their census online without the need for a collector to visit.
17. The key features of the modernised 2018 Census, which ensure the delivery of the desired
effectiveness and efficiencies in accordance with revised investment objectives and critical
success factors (see Appendix 1), include:
mailing out census internet codes and directing people to internet response by
default, with a target of 70% online completion, which will enable a significant
reduction in census collectors, deliver significant savings (up to $10M, compared to
2013 costs) and provide a census model less sensitive to population growth and more
efficient over time
improving the public’s experience of the census, making responding to the census
online very easy and straightforward, consistent with the Better Public Services
programme Result Area 10 outcome of making it easy to transact with government in
a digital environment
using an administratively sourced address list of most New Zealand households to
mailout internet codes, efficiently re-using existing information, in line with the
Government ICT Strategy
improving data quality through targeted follow-up of non-responding households,
directing resources to areas with low response rates, using web-based and real-time
management information
1 Bakker, Carl.
Valuing the Census: A report prepared for Statistics New Zealand which quantifies the benefits to New Zealand
from the use of census and population information. (Wellington: Taylor Duignan Barry, April 2013)
2 At a national level, 1996 Census achieved net coverage of 98.4% and a response rate of 95.6%. 2006 Census achieved net
coverage of 98% and a response rate of 94.8%. The 2013 Census followed this downward trend, achieving net coverage of 97.6%
and a response rate of 92.9%. These national rates may be higher or significantly lower for sub-populations and small population
areas.
3
extensive use of new media channels and communications to help the public
understand the changes in census collection to ensure high online uptake, overall
response rates and target hard-to-reach groups
census systems using or adapting Statistics 2020 platforms3 (generating around $7M
savings), which will be available for re-use in the census following 2018, creating
further efficiencies and consistency with Statistics NZ strategic direction
increased use of administrative data in census processing, improving efficiency and
data quality (generating around $1.5M savings)
different statistical methodologies in response to the new model, which will then be
available for re-use for the following censuses, reducing costs over time
revised census content to better meet New Zealand’s changing information needs
providing earlier, more flexible and more accessible census information products
(outputs), to better meet customer needs
managing change and risk through a comprehensive test driven development
strategy throughout the 2018 Census programme (more extensive than previously,
reflecting the degree of change in the 2018 Census).
Value for Money Assessment
18. A value for money assessment, comparing the modernised 2018 Census with the Status
Quo option (a no-change 2018 option presented to Cabinet in February 2014), confirms that
the modernised 2018 Census delivers the best value for money, outperforming the Status
Quo comparator in both monetary and non-monetary terms.
19. The modernised 2018 Census delivers benefits well in excess of its direct costs. It further
increases value for money by reducing costs over two census cycles, producing currently
estimated savings of at least 5%4 (possibly as high as 11%) over the 2018 and 2023 census
cycles. This will meet or surpass the 5% efficiency target set by the investment objectives.
Over the next two censuses, the modernised census is expected to produce $4.10 of
benefits to the national economy for every $1 spent.
20. The modernised 2018 Census will also ensure essential data quality is maintained and
deliver on non-monetary benefits such as improving the public’s experience of the census,
making it easier for people to interact with the census in a digital environment, embracing
technological change and ensuring greater efficiency in citizens’ transactions with
government (in line with relevant Government strategies such as Better Public Services
(BPS) and the ICT Strategy).
21. In the preparation of this business case, Statistics NZ explored a range of variations to the
modernised 2018 Census model, for example whether New Zealand could move to a fully
online census more quickly. The analysis concluded that, on the basis of current evidence,
the modernisation currently planned for the 2018 Census is at the maximum level that can
sensibly be achieved at this time, and that the modernised 2018 Census model is the optimal
current mix of variable settings.
3 Statistics NZ is now three years into its 10 year programme of transformation, “Statistics 2020 Te Kāpehu Whetū: Achieving the
Statistical System of the Future”, aimed at delivering innovation and change to ensure business sustainability and increased value
from fit-for-purpose statistics. There are direct synergies between this strategic direction and the direction of modernising and
transforming the census.
4 Compared to a 2013 model, inflated to 2018 and 2023 dollars.
4
2018 Census: The Financial Case
Costs
22. In March 2014, Cabinet was advised that the 2018 Census would cost up to $111.0M in
operating, and $7.9m in capital costs, a total of $118.9M. These estimates included an
allowance for contingency.
23. Through the detailed business case process, Statistics NZ has revisited these estimated
costs and the efficiencies that will arise from modernising the census. This further work has
included a revised Quantitative Risk Assessment (QRA), and more detailed planning of the
required investment in conjunction with the Statistics 2020 transformation programme.
24. The revised QRA revealed a potential further $6.4M cost to deliver the 2018 Census,
bringing the total potential cost to $125.3M at the 85th percentile5. The main cause of the
cost increase was a further $3.0M cost for personnel. A further $0.7M capital cost was also
identified.
25. Statistics NZ has since reduced some project costs and identified options to absorb some
cost pressures within the Vote Statistics baseline. A thorough review of costs found that the
overhead costs of running the census could reduce by $3.0M. Overhead costs comprise
variable costs incurred by Statistics NZ’s normal business units to support the census (such
as recruitment costs).
26. The $3.0M reduction in overhead costs lowers the cost of delivering the 2018 Census to
around $122.3M at the 85th percentile. The cost at the 50th percentile (the most likely cost)
is around $117.3M. These estimates indicate that a $5.0M contingency is prudent, being the
difference between the most likely cost ($117.3M) and the cost at the 85th percentile of risk.
Funding and affordability
27. Based on the March 2014 estimates, Budget 2014 appropriated the first year of census
funding ($13.1M operating and $0.55M capital), with the balance being held in a tagged
contingency.
28. This provides a total of $118.9M of funding, leaving a $3.4M gap to allow for the 85th
percentile of risk. Statistics NZ proposes to meet these costs, if they arise, from within the
Vote Statistics baseline. The Statistics NZ Four Year Plans will be used to monitor the
capacity of the Vote to absorb any pressures and identify any trade-offs that may be
necessary. Routine project reporting will also be used to monitor cost pressures.
29. Additionally $5.0M is reserved as contingency within the $118.9M, for release at the
discretion of Joint Ministers (Finance and Statistics). In practice, Statistics NZ will first seek
to absorb any census cost pressures within its baseline, while revealing any trade-offs,
before seeking to drawdown the $5.0M contingency if necessary.
Statistics NZ well placed to manage the successful delivery of the 2018 Census
30. Statistics NZ has a long history of successful census delivery and has undertaken extensive
planning, research and piloting of the changes needed for 2018.
Online development well piloted
31. The 2018 Census is part of a planned, incremental and disciplined modernisation of an
existing business process. Online completion of the census has been used in New Zealand
5 The 50th percentile represents the QRA’s assessment of the most likely cost of the 2018 Census. The 85th percentile represents
the QRA’s assessment of the amount by which the programme could be reasonably expected to exceed costs, given known risks.
The difference between the 50th percentile and the 85th percentile is generally accepted as the amount of contingency which
should be reserved in financial planning.
5
since 2006, building from 7% online completion in 2006 to a national online completion rate
of 34% in 2013 and of 65% in the Oamaru district in 2013, where the mailout model for 2018
was piloted.
2018 model in line with international developments
32. The 2018 Census model is in line with international census developments, particularly in
Canada and Australia, both of whom are running censuses in 2016. Statistics NZ is working
in close collaboration with the statistical agencies in these countries, including participating in
peer review and adapting methodologies, where possible.
Programme planning well advanced
33. Programme planning, recruitment, statistical process planning and census design are well
underway for the 2018 Census. Programme governance has been established, with
membership of the 2018 Census Programme Board confirmed, the Senior Responsible
Owner (SRO) for the programme identified (Deputy Government Statistician, Operations)
and governance arrangements for the 2018 Census programme agreed, alongside and
intersecting with the Statistics 2020 transformation programme. The proposed programme
structure has been designed to support and drive the large streams of work need to deliver
the significant efficiency gains required to be successful and mitigate programme risks.
34. The 2018 Census programme will not be fully established until July 2014, when funding
becomes available and key staff are appointed. The 2018 Census has 15 extensive planning
deliverables, many of which are currently in development. Comprehensive planning will be
completed in the first six months of the programme. Significant progress has been made in
programme organisation and identifying the required strategies and plans that will guide the
programme.
Design allows for testing throughout
35. Adaptation will occur throughout the 2018 Census programme, using an enhanced test
driven development model, to ensure that what is actually delivered (collection and
processes) will be the optimal mix (e.g. of address checking, collector visits, mail-outs) to
meet the key performance indicators and investment objectives and deliver quality
information, within budget. This may deliver modernisation gains in excess of those already
expected and accounted for.
Information technology changes carefully planned
36. Investment in information technology (IT) is a relatively small component of the overall spend
on the 2018 Census, at around 18%6. This investment does not represent a single IT
system. The largest IT investment involves an enterprise collection platform that will be
initially developed for the 2018 Census and is intended to be further developed for Statistics
NZ data collections as a whole. Proof of Concept work has recently been completed which
gives confidence in the preferred delivery mechanism. The 2018 Census will also re-use and
enhance existing production IT capability for processing and dissemination.
37. 2018 Census IT systems will be made up of modularised components which allow flexibility
in terms of when and how they are sourced. This allows individual components to be pre-
tested, does not rely on one solution or supplier, and minimises costs by allowing existing
system components to be used. Integration between the modules is key and integration
testing has been built in.
38. The 2018 Census programme structure and planning includes specific roles and rigorous
early testing to manage and reduce integration risk, building on experience gained in the
2013 Census where Statistics NZ was one of the first government agencies to move to using
6 A total of around $22M operating and capital expenditure, hardware and software, both outsourced and in-house.
6
link to page 9
Infrastructure as a Service (IaaS). This new service was used for hosting the 2013 online
census and was achieved within an 8 month development to implementation timeframe. The
move to IaaS required the successful integration of systems and processes from 3
organisations.
39. Statistics NZ is intending to develop an easy-to-use internet response system for the 2018
Census. Internet codes will be provided to households that will link individuals to their
census night location, which is fundamental to the census. Statistics NZ has investigated the
use of RealMe for the 2018 Census. However, as there is no need to identify or have an
ongoing relationship with individuals, RealMe would not be useful in this context. The use of
RealMe in the 2018 Census would create additional complexity and risk creating
unnecessary confusion which could jeopardise census collection and data quality. The
potential for future RealMe usage for future census purposes is being explored in the Census
Transformation programme of work, and is also being investigated for other Statistics NZ
surveys.
40. The Government Chief Information Officer via the GCIO ICT Assurance Team will oversee
assurance of the 2018 Census (as per the All-of-Government ICT Assurance Framework).
This will be achieved by the GCIO approval of programme’s Assurance Plan, and ongoing
oversight of the execution of the Assurance Plan. GCIO are working with Statistics NZ to
finalise and approve this plan by 31 July 2014.
2018 Census and Statistics 2020 programmes working together
41. Statistics NZ has carefully planned the way in which the Statistics 2020 and 2018 Census
programmes will support each other to achieve transformation. Strategies to assist this
include:
Shared IT architecture vision and operating model between the two programmes – the
IT architecture requirements of the two programmes are very similar, with the more
straight forward 2018 Census developments providing a basis, and impetus, for the
more complex developments for Statistics 2020
Shared governance - including the SRO for the 2018 Census also being the SRO for
the Enterprise Collection Platform Project (see also para
graph 33)
Planning for the intersections of both programmes integral to both designs
Close oversight of both programmes by the Statistics NZ Enterprise Programme
Office, which has organisational responsibilities for managing change across the
organisation, including through extensive regular monitoring and programme reporting
Quality assurance is integral to programme design
42. Quality Assurance for the 2018 Census will include:
Up to 5 independent quality assurance (IQA) reviews and at least one further Gateway
review (A Gateway review and IQA were conducted in May 2014).
Oversight by central agencies through major project monitoring of 2018 Census and
Statistics 2020.
Testing and evaluation – a comprehensive programme of testing across the end-to-end
business cycle will be undertaken, covering resources, processes and systems.
Dress rehearsal – a major run-through of 2018 Census end-to-end systems and
processes will be undertaken in March 2017.
Post enumeration evaluation – independent quality assurance will be conducted
following the 2018 Census to measure the accuracy of census coverage (undercount
and overcount). A process of post enumeration quality assurance has been conducted
on New Zealand censuses since 1996.
7
Risk management strategies well advanced
43. Managing risks will be of particular importance in the 2018 Census, as census systems,
processes and methods are modernised. Risks will be actively monitored and managed
through the life of the programme.
44. A detailed risk register is well advanced. Further development will occur as senior census
staff are appointed and as planning deliverables are finalised towards the end of 2014.
Seven key risks of the 2018 Census have been identified. Significant work has already been
undertaken to identify the main mitigation strategies for these risks (see Appendix 2).
Statistics NZ is confident these risks can be effectively monitored, mitigated and managed.
Key Decision for Cabinet
45. This Cabinet paper seeks the release of the remaining $105.3M funding for the 2018 Census
from 2015/16 to 2019/20, and for 2015/16 census transformation work, from the 2018
Census tagged contingency to Vote Statistics.
Financial Implications
46. A sum of $111M operating and $7.9M capital funding is sought for continued census
transformation and the 2018 Census. As voted in Budget 2014, the first year of funding,
$13.1M operating and $0.55M capital will be released from 1 July 2014. The following
changes to census appropriations for the following years 2015/16 to 2019/20, representing
the balance of the funding requirement for the 2018 Census and 2015/16 census
transformation work, are proposed:
$m – increase/(decrease)
Vote Statistics
2015/16
2019/20 &
Minister for Statistics
to 2018/19
outyears
Departmental Output Expense:
2018 Census of Population and Dwellings
97.988
-
(funded by revenue Crown)
Statistics New Zealand
Capital Injection
7.359
-
Total Operating
97.988
-
Total Capital
7.359
-
47. The 2018 Census programme will run over a 6 year cycle. The proposed funding
arrangement is for a multi-year appropriation (MYA) over five years. This MYA is for the total
census budget. In the final year of the five year appropriation, Statistics NZ will ask for a new
single year appropriation to be established. It will then carry forward funding for the sixth
year into the new one year appropriation. The indicative spending profile for the new multi-
year appropriation, described above, is as follows:
$m – increase/(decrease)
Indicative annual spending
2014/15
2015/16
2016/17
2017/18
2018/19
profile
Operating
-
13.150
18.747
48.540
17.551
Capital
-
4.097
3.181
0.080
-
8
Human Rights, Treaty of Waitangi and Privacy Considerations
48. The proposals in this paper are consistent with the New Zealand Bill of Rights Act 1990, the
Human Rights Act 1993, and the Privacy Act 1993. The proposals outlined in this paper
have no specific gender implications.
Legislative and regulatory implications
49. The year of the next census, 2018, is in accordance with Section 23 of the Statistics Act
1975, which requires that the census be held in 2013 and every five years thereafter. Section
23 also requires that the day on which the census shall be taken will be appointed by the
Governor-General by Proclamation, with subsequent notice given in the New Zealand
Gazette. It is conventional practice that the census date is decided and made public early in
the census cycle. For example, the Proclamation for the planned March 2011 Census was
made in June 2009.
Consultation
50. Consultation with a wide range of government agencies was undertaken previously on the
options for the next census and census transformation, as outlined in the February report to
Cabinet. A number of those agencies specifically supported the modernised 2018 Census
option, whilst others did not express a preference.
51. Consultation on this paper has been undertaken with the following agencies and
departments: the Treasury, State Services Commission, and the Department of Internal
Affairs (the Government Chief Information Officer). The Department of Prime Minister and
Cabinet has been informed.
Comment from the Government Chief Information Officer (GCIO): ICT Assurance
52. Ongoing monitoring of dependency risks: We note that a significant portion of the technology
changes required to modernise the 2018 census reside in the Statistics 2020 programme,
with the Census 2018 programme intending to pick up and further develop technology
platforms delivered or upgraded in Statistics 2020. We note the dependency risk that this
approach introduces, and will monitor this risk through targeted assurance activities in the
wider programme Assurance Plan.
53. Assurance to date: Assurance on the programme to date has been comprehensive, with
Gateway and IQA reviews on the business case phase being timely and appropriate.
54. Finalise Assurance Plan: The overall Assurance Plan for the programme going forward is not
yet finalised. GCIO ICT Assurance are working with programme team and look forward to
finalising this by the end of July 2014.
Recommendations
55. It is recommended that the Cabinet Economic Growth and Infrastructure (EGI) Committee:
a)
note that in February 2014 Cabinet, inter alia:
i.
agreed in principle, subject to Budget decisions, that a modernised New Zealand
Census of Population and Dwellings be conducted in March 2018 (Option 2), with
updated content and increased use of technology and administrative data;
ii.
directed Statistics New Zealand to complete a focussed detailed business case on
the preferred option, reporting on the commercial, financial and management
aspects of the 2018 Census, and submit this to EGI by 30 June 2014; and
iii.
agreed in principle that, subject to Budget decisions, the first year of funding for the
2018 Census ($13.1M operating and $0.55M capital costs) be released from 1 July
2014;
9
link to page 12 link to page 13
iv.
noted that Cabinet agreement to release the balance of funding will be sought
through the submission of the focussed detailed business case referred to in ii
above [CAB Min (14) 6/8 refers];
b)
note that Budget decisions confirmed the above agreements [CAB Min (14) 13/8 (31)
refers], with funding for 2014/15 voted in Budget 2014 and funding for the years 2015/16
and beyond held in a 2018 Census tagged contingency;
c)
note that Statistics New Zealand has completed a focussed detailed business case, in
consultation with monitoring agencies and in accordance with the Cabinet direction
outlined in recommendation
a) above, the Executive Summary of which is attached to
this Cabinet paper with the full focussed detailed business case available at this
Committee meeting;
d)
note that the further work undertaken in the preparation of the focussed detailed
business case confirms the modernised 2018 Census is the best available method to
deliver changes needed for the next census; provides methodology to address falling
response rates; provides the best value for money for the next census and further
censuses; is commercially viable; and has strong management structures in place and
planning in train to ensure successful delivery;
e)
note that, while a revised quantitative risk analysis revealed potential for increased
costs, Statistics New Zealand has since reduced some cost components and identified
options to absorb cost pressures within the Vote Statistics baseline and is confident it
can successfully deliver the modernised 2018 Census within the total estimate of
$118.9M submitted for Budget 2014;
f)
note the Minister of Statistics will, through the annual Statistics New Zealand Four Year
Plan, review forecast cost pressures and Statistics New Zealand’s ability to manage
2018 Census financial risk, should it emerge;
g)
approve the following change to appropriations, with a corresponding impact on the
operating balance and debt:
$m – increase/(decrease)
Vote Statistics
2015/16
2019/20 &
Minister for Statistics
to 2018/19
outyears
Departmental Output Expense:
2018 Census of Population and Dwellings
97.988
-
(funded by revenue Crown)
Statistics New Zealand
Capital Injection
7.359
-
Total Operating
97.988
-
Total Capital
7.359
-
h)
note that the indicative spending profile for the new multi-year appropriation, described
in recommendation
g) above, is as follows:
$m – increase/(decrease)
Indicative annual spending
2014/15
2015/16
2016/17
2017/18
2018/19
profile
Operating
-
13.150
18.747
48.540
17.551
Capital
-
4.097
3.181
0.080
-
i)
agree that the proposed changes to appropriations for 2014/15 above be included in the
2014/15 Supplementary Estimates and that, in the interim, the increases be met from
Imprest Supply;
10
link to page 13 link to page 13 link to page 13
j)
agree that the expenses and capital expenditure incurred under recommendatio
n g)
above will be charged against the 2018 Census tagged contingency;
h)
note that, as a result of the decisions in recommendatio
n g) above, the 2018 Census
tagged contingency is fully draw down and effectively expires;
i)
note that the total operating increase in recommendatio
n g) above includes a project
contingency of $5.000 million; and
j)
authorise the Ministers of Finance and Statistics to release the contingency funding on
submission of revised cost information by Statistics New Zealand, in consultation with
the Treasury and the Government Chief Information Officer, provided all other means of
managing any 2018 Census cost pressures within Vote Statistics have been exhausted.
Nicky Wagner
Minister of Statistics
Date:
11
Appendix 1
2018 Census Focussed Detailed Business Case
Executive Summary
On 19 February 2014 the Cabinet Economic Growth and Infrastructure Committee (EGI)
considered options for the next census [EGI Min (14) 2/11 refers]. Three options were presented:
a ‘status quo’ census in 2018, a ‘modernised’ census in 2018, and a ‘modernised plus’ census in
2021. A single stage business case was submitted, with a modernised 2018 Census as Statistics
New Zealand’s preferred option.
The EGI Committee agreed, in principle, subject to Budget 2014 decisions, that a modernised
census be conducted in March 2018, with updated content and increased use of technology and
administrative data. Cabinet confirmed the decision on 24 February [CAB Min (14) 6/8 refers].
Budget 2014 decisions have now been taken [CAB Min (14) 13/8(31]. The intention to conduct a
modernised census in 2018 was publicly announced by the Minister of Statistics on 19 May 2014.
The single stage business case for the modernised 2018 Census was submitted for consideration
in Budget 2014. Cabinet approved in Budget 2014 [CAB Min (14) 13/8(31)] a total appropriation
of $7.9M capital and $111.0M operating costs over 5 years, to deliver the next census in 2018,
and to continue investigations, over the next 18 months, into the use of administrative data to
produce census information (Census Transformation). This funding is allocated as follows:
a)
$13.099M operating and $0.547M capital for 2014/15, for the 2018 Census programme
and Census Transformation work, and
b)
the remaining funding, from 2015/16 to 2019/20, is allocated to a tagged contingency
to be released on Cabinet’s approval of this focussed detailed business case, to be
submitted by 30 June 20147.
Cabinet directed Statistics NZ to report, in a focussed detailed business case, on the commercial,
financial and management aspects of the 2018 Census programme which provides additional
information on:
• the benefits and risks
• detailed scope and costs
• detailed management and assurance plans and
•
processes for a revised quantitative risk analysis and for accessing contingency
funding [EGI Min (14) 2/11].
This focussed detailed business case provides the additional information requested by Cabinet
and, along with the associated Cabinet paper, seeks the release of the remaining funding for the
2018 Census, from 2015/16 to 2019/20, from the tagged contingency to Vote Statistics.
The Strategic Case
This focussed detailed business case confirms the case for census change. The traditional
model of census collection needs to change if it is to be sustainable over time. The main drivers
for change continue to be:
7 Monitoring agencies reviewed the risk rating of the programme, from medium to high, in late 2013. Had the high risk rating
been given initially, a two-stage business case would have been required. Given the work that had been completed on the single
stage business case, it was agreed that Statistics NZ supplement the single stage business case with this focussed detailed
business case on the modernised 2018 Census.
12
Rising costs and potential efficiencies – it is getting harder to run a traditional census in an
efficient manner, population growth contributes to rising costs, digital options are more
efficient and provide opportunities to maximise self-response and concentrate personnel
on the hard-to-count non-respondents.
Response rates – declining response rates8, impacting on the quality of data especially for
small sub-populations and small areas, are a trend for New Zealand and internationally.
Administrative data – increasing availability and sharing of administrative data across
government, efficiencies from re-use of data, consistent with Government strategic
directions.
Technological advances – census systems are currently based on outdated technology
(over 10 years old) and set up for paper-based survey delivery and collection; the public
expect easier and faster modern methods; and there are opportunities from technological
and methodological advances in combining and analysing large datasets and in census
collection modes.
Transforming the census to a model based solely on administrative data is under investigation for
longer term change, through the Census Transformation programme of work which will report
back to Cabinet by 30 October 2015. In the short to medium term, modernising the survey based
census model will significantly improve the efficiency and the effectiveness of census and provide
an important step towards a future census model based on administrative data.
The investment objectives and critical success factors for the 2018 Census have been revised in
the preparation of this focussed detailed business case as follows:
Investment Objectives for the modernised 2018 Census
a) To undertake a Census of Population and Dwellings in 2018 that meets statutory
requirements.
b) To at least maintain the quality of census information compared with the 2013 Census.
c) To improve, by 20%, the timeliness of census information products released to users
following a 2018 Census compared to that released following the 2013 Census.
d) To reduce, over two full census cycles, the average cost of the census by 5.0%, starting in
2014 and using the 2013 Census adjusted for annual inflation at 2.0%.
Critical Success Factors for the modernised 2018 Census
a) Strategic alignment:
How well does this option support Statistics NZ's strategic priorities and other strategic
programmes such as Statistics 2020, and the Government's Better Public Service Result
Area 10 and ICT Strategy.
8 At a national level, 1996 Census achieved net coverage of 98.4% and a response rate of 95.6%. 2006 Census achieved net
coverage of 98% and a response rate of 94.8%. The 2013 Census followed this downward trend, achieving net coverage of 97.6%
and a response rate of 92.9%. These national rates may be higher or significantly lower for sub-populations and small population
areas.
13
b) Value for money:
How well does this option optimise value for money by balancing costs and efficiencies
against other goals of modernising census processes and delivery, and meeting statistical fit-
for-purpose quality by?
i.
maintaining the coverage rate achieved by the 2013 Census
ii.
improving the respondents experience of participating in the census process
iii.
meeting customers’ information needs
iv.
reducing sub-national and sub-group variation in response rates compared with
the 2013 Census
v.
decreasing the need to impute for missing data for key variables, sub populations
and small areas.
c) Commercial partnerships:
How well placed is the community of contractors for supply and service to meet the capacity
and capability requirements for this option, including ability to deliver to specifications, ability
to modify requirements in response to the census testing regime (test driven design), and
optimise value for money over the term of the contract?
d) Risk and feasibility:
How well placed is Statistics NZ's capacity to achieve this option within the time-frame as
specified by the Statistics Act 1975?
Modernising the census will be achieved through a fundamental shift in the way people are
contacted and respond to completing the census – away from a large force of census collectors
delivering paper forms to every household to the majority of people completing their census
online without the need for a collector to visit. Efficiency and quality gains can be made by
greater use of modern technology and more sophisticated statistical methodologies throughout
the census process.
The key features of a modernised 2018 Census, which will ensure the achievement of these
revised objectives and critical success factors include:
a “digital by default” strategy, with an online completion target of 70% (resulting in
significant reductions in field staff and a model less sensitive to population growth;
and in accordance with Government strategic directions)
using real-time management information to direct field effort to those least likely to
respond (targeting resources at improving response rates for hard-to-reach
populations)
extensive use of new media channels and communications (to explain the changes in
census collection, maximise online uptake and overall response rates, and target
messages to hard-to-reach populations)
efficient building, or re-use, of Statistics NZ corporate platforms to support census
collection and processing (achieving savings, ensuring strategic alignment and
ensuring systems are maintained to enable re-use in future censuses)
efficient re-use of administrative data to support collection and processing of the
census (generating savings and assisting longer term transformation)
revising census content and providing earlier, more flexible census outputs (to better
meet New Zealand’s changing information needs).
14
The Economic Case
In the preparation of this focussed detailed business case, Statistics NZ considered:
the potential for variations on the design of the modernised 2018 Census to provide faster
modernisation within the same funding, or to produce further efficiency savings; and
the value for money proposition for the modernised 2018 Census, including the sensitivity
of the predicted benefits to the main programme risks.
Potential for variations
Statistics NZ explored the potential for variations on the modernised 2018 Census option, as
agreed by Cabinet in February 2014, to test the potential to move faster, where possible, towards
modernisation within the same funding.
Generating further savings
Options were first explored to generate savings which could be redirected towards faster
modernisation. Given the critical role of collecting census information to ensure a high response
rate and meet coverage targets (i.e. to count everyone), Statistics NZ considered potential areas
where funding could be reduced without significantly compromising census collection, and
therefore data quality.
It concluded that the only options for savings, without affecting critical collection activities, were
to: a) limit census content redevelopment; and/or b) reduce the amount of census information
analysed and released by Statistics NZ.
Census questions (content) have changed only minimally since 2001. No content change in
2018 would mean questions remain static for longer than 20 years (until potentially the following
census in 2023). The questions of the 2018 Census must meet the information needs of 2018
and beyond, not the information needs of the past. Given the investment needed to update
census content for 2018 is relatively small, at around $600,000, and given the significant impact
outdated content would have on the realisation of census benefits, it was concluded that not
updating content could not be justified.
Another potential area considered for savings was reducing census outputs, releasing information
on only a minimum set of variables and charging for any more detailed information. Analysis
concluded that reducing access to census information would significantly erode the value to New
Zealand of the census, far in excess of the relatively small financial saving. Another issue with
reducing census outputs is that it would be very difficult to defend gathering information from
everyone in New Zealand which is then not made publically available, and/or charging the public
to access standard information breakdowns on information they are legally required to provide to
Statistics NZ. Census information underpins trust in government and the principles of open
democracy and failing to output all census information could undermine that trust. Also charging
central, regional or local government agencies for standard information breakdowns for use in, for
example policy analysis and resource allocation models, would simply shift costs to other Votes
or to regional and local councils.
Moving faster towards modernisation
Two opportunities to increase modernisation in 2018 were investigated: a) greater use of
administrative data in the 2018 Census; and b) moving to a fully online census more quickly.
Greater use of administrative data in 2018 is limited by the current quality of administrative data
across government. The focus at this point needs to be on improving the quality of administrative
data, as is being carried out in the Census Transformation project. Analysis concluded that,
15
given the constraints of current administrative data, current plans to achieve efficiencies through
using administrative data in census processing are at the maximum level that can sensibly be
achieved at this time.
Statistics NZ also considered whether it would be possible to move more quickly to a fully online
model in 2018. The target for completion of census online in 2018 is currently 70%, reflecting the
results of the 2013 Oamaru trial which achieved 65% online uptake, analysis of the national
online uptake (34%) in 2013 and international experience. This target may be adjusted, following
field testing in the development phases of the 2018 Census and reviewing the Australian and
Canadian experiences in their 2016 censuses. Therefore greater efficiencies in the field
operation may be achieved than are currently expected. However, planning based on a higher
than 70% target at this point would not be supported by available evidence, and would risk
significantly over-estimating savings and under-estimating costs.
Conclusion
In considering variations, it is important to bear in mind the adaptive nature of the 2018 Census,
where an enhanced test driven development model will be used to ensure an optimal build and
active management of development risks. An increased testing programme for the 2018 Census
will ensure that what is actually delivered (collection and processes) will be the optimal mix (e.g.
of numbers of collectors, of real time management information needed to direct field effort, of
communications design to achieve online targets etc.) to meet objectives and deliver the greatest
possible efficiencies while maintaining data quality.
Taking account of this range of factors, it was concluded that the variations explored did not offer
any viable alternatives to the modernised 2018 Census option agreed by Cabinet, given that:
the level of savings which could be generated, from no change to census content and
the reduction of standard information outputs, would have a disproportionately
negative effect on realising the benefits from, and achieving the investment objectives
and critical success factors of, the 2018 Census;
no substantive additional modernisation gains could be relied on for the 2018 Census;
and
the test driven development model will ensure the best mix of various factors to
deliver the optimal build on the day and may deliver modernisation gains in excess of
those already expected and accounted for.
Value for money
Given the lack of viable variations on the modernised 2018 Census option, detailed economic
assessment was undertaken comparing the modernised 2018 Census with the Status Quo option
(a no-change 2018 option rejected by Cabinet in February 2014). This value-for-money
assessment included examining:
Whole-of-life costs and net present value across two census cycles (2018 & 2023)
The ratio of present-value costs to present-value monetary benefits
Multi-criteria analysis of non-monetary benefits.
This economic assessment demonstrated that the modernised 2018 Census represents the best
value for money as compared with the Status Quo option, as the additional investment required
to deliver a modernised 2018 Census is outweighed by the significantly greater benefits from
modernisation.
The analysis of monetary costs and benefits demonstrated that the modernised 2018 Census will
produce residual value which will be realised in the 2023 Census, and provide significant savings
16
over two census cycles. The modernised 2018 Census is currently estimated to produce
savings, comparative to 2013, of at least 5% (possibly as high as 11%) over the 2018 and 2023
census cycles, meeting or surpassing the 5% efficiency target set by the investment objectives.
Over the next two censuses the modernised model will produce $4.10 of benefits to the national
economy for every $1 spent.
In comparison, a status quo (no change) census would be poor value for money, with increasing
costs over two census cycles. Failure to address declining response rates and to update census
content means a status quo census would produce a significant decrease in overall census data
quality and fail to meet New Zealand’s information needs. A status quo census would not align
with relevant Government strategies (such as the ICT Strategy and BPS), would fail to embrace
technological change and would not improve the public’s experience of the census.
A revised quantitative risk analysis, undertaken in May 2014, identified the main drivers of risk for
the 2018 Census programme as difficulty in predicting:
high non-response areas
self-response rates (those who will respond without a visit from a collector)
integration of census with corporate technology architecture.
Statistics NZ has developed effective strategies to mitigate these risks and minimise their
potential impacts. The sensitivity analysis undertaken indicates that, providing strategies to
mitigate and minimise these risks are implemented successfully, there should be no significant
reduction in overall monetary and non-monetary benefits. The Net Present Value (part of value
for money assessment) of the modernised 2018 Census is so high that monetary costs would
have to increase by 200% or monetary benefits decrease by 70% (or a combination of both) for it
to become a negative value. This is a highly unlikely scenario.
The economic analysis of the modernised 2018 Census concludes that this option will deliver the
greatest value for money, particularly when balanced against other objectives for modernising the
census, and may exceed the savings target set by the investment objectives for the 2018
Census.
The Commercial Case
Statistics NZ will embark on a programme of up to twenty separate procurements to enable the
successful delivery of the 2018 Census operation, with a total procurement value of around
$23M. The commercial case of the focussed detailed business case details how these
procurements will be approached and managed, in accordance with Government’s procurement
principles, rules, guidance and policies; to ensure procured goods and services will deliver a
modernised 2018 Census which meets investment objectives and critical success factors, and
realises benefits.
In many cases, the goods and services procured for the modernised 2018 Census will be similar
to those procured for the 2013 Census. Statistics NZ is confident the market can meet the
procurement needs of the 2018 Census in a timely and cost-effective manner.
The largest single procurement of information technology (over $1M) will be the development of
the enterprise collection platform that will be initially developed for the 2018 Census and is
intended to be further developed for Statistics NZ data collections. Market analysis to date has
confirmed the market could most likely provide a solution that meets the core requirements for all
Statistics NZ collection activities including the 2018 Census. Another potential solution is to use
components developed by the Australian Bureau of Statistics (ABS) for their 2016 Census.
Communications with the ABS have confirmed their system, developed in-house, is very similar
to what is required for the New Zealand 2018 Census and ABS are willing to share what they
have developed with Statistics NZ.
17
Careful consideration has been given to the location of the Enterprise Collection Platform project
within Statistics NZ programme governance structures, with the Enterprise Collection Platform
Project Board now reporting to the 2018 Census Programme Board and having the same Senior
Responsible Owner (Deputy Government Statistician, Operations). The Government Chief
Information Officer will provide ongoing ICT assurance in the planning and monitoring of the 2018
Census (as per the All-of-Government ICT Assurance Framework).
The Financial Case
The Next Census single stage business case and Cabinet paper in February 2014 sought
agreement for funding of $111M in operating and $7.9 in capital for the 2018 Census, and to
continue census transformation. Cabinet agreed [EGI Min (14) 2/11] and Budget 2014 decisions
confirmed [CAB Min (14) 13/8 (31)] that the first year of funding, ($13.1M) operating and $0.55M
capital) be released from 1 July 2014, with the remaining funding released through the submission
of a focused detailed business case in June 2014. As a result of further work in the preparation of
this business case, Statistics NZ confirms funding requirements are consistent with the single stage
business case submitted for Budget 2014 (that is, $118.9M in total, being $7.9M capital expenditure
and $111.0M of operating expenditure). However the composition of those costs has changed.
Financial viability over time
The Financial Case demonstrates that the modernised 2018 Census continues to represent the
most financially affordable option for the next Census. As noted in the Economic Case, the
changes proposed in the modernised 2018 Census may realise an estimated 5%, or potentially as
high as 11%, saving across two census cycles (2018 and 2023), meeting or exceeding the 5%
efficiency target set by the revised investment objectives. However, investment is needed in the
modernised 2018 Census in order to realise these savings over time.
Savings are expected to be achieved on the basis of $24M investment in the 2018 Census,
including for new systems and processes, greater testing and integration costs, and additional
communications and marketing to inform the public of census changes. This investment is offset
by up to $19M of efficiencies realised in the 2018 Census, for example from fewer field staff,
reductions in processing costs through re-use of administrative data, and savings from adapting
Statistics NZ corporate platforms for census use.
Further efficiencies are expected over time. The 2018 Census will invest in systems, processes
and methodologies which will be available for re-use in any subsequent survey-based census,
resulting in a reduction of future costs. Cost reductions of around $13M, in comparison to the 2018
Census, have been estimated for the census beyond 2018. In addition, the development of the
new collection model also avoids substantial cost increases driven by population growth now and
in the future, as the previous paper-based manual collection model is much more sensitive to
population growth than the new “digital by default” model.
Costs reviewed and confirmed
Since February 2014, Statistics NZ has completed further work to review both the costs of the 2018
Census and the efficiencies that will be achieved in modernising the census. This work has included
a revised quantitative risk analysis (QRA) and more detailed planning of the required investment in
conjunction with Statistics 2020.
The revised QRA modelled a most likely estimated total 2018 Census cost (including contingency)
at $125.3M, an increase of $6.4M from the $118.9M submitted for Budget 2014. This QRA reflects
the significant degree of uncertainty inherent in this stage of census preparations for a new census
model. The QRA identified the most significant uncertainties affecting the QRA results as:
18
the adaption of the Statistics 2020 corporate IT platforms for census use
the collections self-response rate of 70%
the average non-response follow up time estimate.
Statistics NZ has undertaken more detailed planning which has included developing plans and
mitigation strategies for the management of the areas of most significant uncertainty identified by
the QRA. For example, since the revised QRA was conducted, a new programme governance
structure has been developed to ensure close connection between Statistics 2020 Enterprise
Collection Platform development and the 2018 Census. Further planning has also included
reviewing the assumptions used in the cost estimates for personnel and overheads.
In line with the QRA risks, detailed planning has highlighted some areas where efficiency targets
may be challenging, particularly in achieving the full savings from reductions in personnel. Current
estimates now suggest around $3M of additional personnel effort may be required.
Statistics NZ has also reviewed the calculation of overheads for the programme, removing the fixed
organisational overhead which had been allocated under a standard FTE formula. This review of
overheads has resulted in a $3M reduction in cost which Statistics NZ proposes to reallocate within
the 2018 Census funding.
Whilst recognising the risks identified by the QRA, Statistics NZ remains confident it can
successfully deliver the modernised 2018 Census within the Budget 2014 appropriation of $118.9M
and is not seeking additional funding. This decision is based on the following rationale:
Approximately one third of the financial risk reflected in the revised QRA is related to
capital and its associated depreciation flows which, if they materialise, Statistics NZ is
confident it can manage within current cash reserves
Previous experience in delivering censuses, including the 2013 Census which is
forecast to be on budget with limited use of contingency
The application of rigorous financial governance will be used to track the approved
census budget and ensure planned efficiencies are delivered, including further
revised QRAs and Gateway reviews, and oversight by Central Agencies
Regular cost monitoring and reporting will occur, including reporting in detail on the
cost drivers of key activities and use of rigorous project and programme disciplines
Statistics NZ willingness to first look to manage cost pressures within Vote, via trade-
offs or, if possible, the delivery of further efficiencies.
Statistics NZ does, however, seek a change in the composition of the funding, reducing the
contingency component from $16.4M to $5.0M, with a corresponding increase in operating funding.
This re-composition reflects the certainty of increased costs factored into the revised QRA, which
were previously considered uncertain and accounted for in the original $16.4M contingency. A
contingency of $5M is consistent with the revised QRA advice on the contingency amount required.
If cost pressures manifest in manageable amounts across the life of the programme, Statistics NZ
is confident these could be absorbed within baseline. However, if cost pressures manifest
materially in one financial year, a request to joint Ministers to draw down on the contingency, as
per usual practice, would need to be considered in consultation with the Treasury.
19
Managing the Modernised 2018 Census
Statistics NZ is well placed to successfully manage the delivery of the modernised 2018 Census
on time and within budget. The management case of the focussed detailed business case details
the arrangements that are planned and in progress to ensure that the 2018 Census is
successfully delivered, meeting its objectives, managing its risks and fully realising its benefits.
The 2018 Census programme has been divided into 5 phases over 6 financial years. End of
phase reviews, in conjunction with the testing programme, provide points for review, learning and
adaption.
While programme planning, recruitment, statistical process planning and census design are well
underway, the 2018 Census Programme will not be fully established until July 2014, when
funding becomes available and key staff are appointed. The 2018 Census has 15 extensive
planning deliverables, many of which are currently in development. Significant progress has
been made in programme organisation and identifying the required strategies and plans that will
guide the programme. Comprehensive planning will be completed in the first six months of the
programme.
Programme governance has been established, with membership of the 2018 Census Programme
Board confirmed, the Senior Responsible Owner (SRO) for the programme identified (Deputy
Government Statistician, Operations) and governance arrangements for the 2018 Census
programme agreed, alongside and intersecting with the Statistics 2020 transformation
programme.
The proposed programme structure has been designed to support and drive the large streams of
work need to deliver the significant efficiency gains required to be successful. Recruitment is
underway for key senior staff.
Managing change within the 2018 Census
The 2018 Census General Manager and the SRO will have key roles in leading, driving and
managing change and ensuring new thinking develops alongside new methods which are in turn,
fully integrated into system and process development. The proposed census structure includes
a Manager, Programme Design, Test and Office role, which provides a specialist project
executive role which will support the programme management of the entire programme and be
responsible for overseeing effective integration of change across the programme.
Key strategies for managing 2018 Census change include:
a rigorous test driven development programme
emphasis on a think-do learn through all phases
capturing, documenting and sharing lessons learnt throughout the 2018 Census lifecycle
reporting monthly, at phase ends, and as required
detailed re-planning at least yearly with plans being re-baselined and approved annually by
Statistics NZ’s Executive Leadership Team (ELT).
The 2018 Census team will work closely with the Statistics NZ Enterprise Programme Office,
which has organisational responsibilities for managing change, through governance and regular
reporting structures. The inclusion on the 2018 Census Programme Board of a representative
20
from the Statistics NZ Enterprise Programme Office as well as three ELT members will ensure a
close connection with organisational transformation.
Quality Assurance Planning
Work on the Quality Assurance and Management Strategy and Plan is underway. Quality
Assurance for the 2018 Census will include:
Up to 5 independent quality assurance (IQA) reviews and at least one further Gateway
review (A Gateway review and IQA were conducted in May 2014)
Oversight by central agencies through major project monitoring of 2018 Census and
Statistics 2020
Testing and evaluation – a comprehensive programme of testing across the end-to-end
business cycle will be undertaken, covering people, processes and systems.
Dress rehearsal – a major run-through of 2018 Census end-to-end systems and
processes will be undertaken in March 2017.
Post enumeration evaluation – independent quality assurance will be conducted
following the 2018 Census to measure the accuracy of census coverage (undercount
and overcount).
Retrospectives – six-weekly internal reviews of project and programme progress, using
agile methods, will be undertaken to ensure that census work is on track and changes
are made to improve areas of weakness.
Closure reports from 2013 Census – project and programme closure reports from the
2013 census will feed into programme initiation and detailed programme and project
planning for the 2018 Census.
International collaboration – ongoing collaboration with other statistical agencies,
particularly the International Census Forum, the Australian Bureau of Statistics,
Statistics Canada and the national statistical agencies in the UK, will provide
opportunities for peer review and to learn from their experience in preparing and
undertaking censuses in 2016.
Benefits Management Arrangements
A benefits mapping exercise has been undertaken which identifies the top three benefits of
conducting the modernised 2018 Census as:
a) All levels of government have the information required to make decisions of national and
policy importance
b) Statistics NZ continues to deliver information of national, regional and local significance
throughout census change
c) The next census is delivered by a sustainable lower cost operating model.
Careful management of benefits will be an integral part of the management of the 2018 Census.
A benefits realisation plan is in development and will be implemented and monitored throughout
the life of the 2018 Census programme, with responsibility for the oversight of benefits
management resting with the Project Executive, Programme Design, Test and Office.
Risk Management Arrangements
Managing risks will be of particular importance in the 2018 Census, as census systems,
processes and methods are modernised. Risks will be actively monitored and managed through
the life of the programme.
21
A detailed risk register is well advanced. Further development will occur as senior census staff
are appointed and as planning deliverables are finalised towards the end of 2014. Seven key
risks of the 2018 Census have been identified. Significant work has already been undertaken to
identify the main mitigation strategies for these risks. Statistics NZ is confident these risks can be
effectively monitored, mitigated and managed.
22
Appendix 2
2018 Census Programme Risks Summary
2018 Census Programme Risks Summary
#
Risk Description
Existing Controls in Place
Treatment/ Risk Mitigation Actions
1 Business processes and methods - If 2018 Census business • Annual replanning and baselining practices - overseen • Implementing an effective test driven development programme
processes and methods are not sufficiently robust or tested, by the Census Programme Board and Enterprise
• Using the Canadian Census delivery and fol ow-up model as a basis
delivery to investment objectives and critical success factors Programme Office (EPO)
f
o
r
2
0
1
8
C
e
n
s
u
s
planning and testing
wil be compromised. Realisation of benefits wil in turn be
• Regular programme risk reviews and reporting
•
E
a r l
y
pl
a
n
ni
n
g i
s
underway – for example work breakdown
impacted.
• Regular programme improvement focused
structures (refer paragraph XX 194) are wel advanced
retrospective reviews • Using international expertise and experience e.g. ongoing
• Programme management support and oversight from involvement in the International Census Forum (ICF) and conducting
Statistics NZ EPO an International Review of New Zealand’s 2018 Census preparations
• Agreed thresholds of tolerance with immediate
(using a panel of invited ICF) members
escalation to central agencies where significant
• Architecting for change - ensuring our architecture wil accomodate
variance from agreement l
a
t
e
c
h
a
n
g
e
s
t
o business processes.
• Oversight by central agencies of major projects - 2018
Census and Statistics 2020
2 Statistics NZ's quantum of change - If the quantum of
• Integrated oversight by the EPO of al Statistics NZ
• Sharing governance between the 2018 Census, Statistics 2020
change occurring across the 2018 Census programme and change programmes and projects, including the 2018
Transformation and the Census Transformation project wil reduce silo
across the organisation (e.g. Statistics 2020, organisational Census
effects
re-grouping) exceeds capability then delivery of the 2018
• Oversight of 2018 Census budget matters by ELT and • Monitoring and reporting of risks is done weekly by teams and
Census wil be compromised.
the Chief Financial Officer (i.e. outside the structure for monthly by projects and the 2018 Census programme
the 2018 Census programme)
• RIsks and issues are escalated to the Executive Leadership Team as
appropriate and within agreed tolerances.
• Using a think-do-learn cycle involving a range of testing strategies,
rapid review, reporting and refining practice.
• Keeping key stakeholders informed and managing expectations
involved in implementing new methods wil help people understand how
our learning is driving down uncertainty.
3 Programme costs and savings - If programme costs are
• Oversight of 2018 Census budget matters by ELT and • Expected savings are spread across the 2018 Census programme
underestimated then expected savings wil not be realised
the Chief Financial Officer (i.e. outside the structure for and so are not reliant on only one aspect of efficiencies
and the census budget wil be impacted.
the 2018 Census programme) • Gateway reviews as scheduled
• Additional programme and project support and
• Contingency review scheduled early in the 2018 Census programme
oversight, including close integration with the EPO • Contingency plans include organisational response (e.g. examining
• Monthly reporting of finances to Census Programme
organisational budgets for re-prioritisation of funding outside the 2018
Board to ensure close monitoring of budgets
C
e
n
s
u
s
o
u
t
p
u t
c
l
a
s
s
e
s
)
a
n
d
r
e
v
i
e
w
of
s
c
o
p
e f
or the 2018 Census
• Annual re-planning, review and re-baselining cycles to
ensure fiscal best practice
• Oversight by central agencies of major projects - 2018
Census and Statistics 2020
4 Architecture - If 2018 Census systems are not built in time
• Shared governance between the 2018 Census and the • Implementing an effective test driven development programme,
and their architecture is not integrated then delivery of the
Statistics 2020 Transformation Programme, on which
including proof of concepts work and integrated testing early in the
2018 Census wil be compromised. This wil compromise
much Census architecture is dependent
process
benfit realisation by impacting the quality of data produced
• Programme management support and oversight from • Creating a new Testing Project Manager role
from the census. This wil impact national decision making
Statistics NZ EPO
• Starting early development – some architecture is already wel
and policy development.
advanced
• Architecting for alternatives - ensuring our architecture wil al ow
alternative sources of capability to be readily substituted in
contingency situations.
• Planning contingencies - includes exploring alternative sources of
relevant architecture (e.g. Australian Bureau of Statistics systems)
and having in place clear decision points for implementing
contingency plans.
5 Self-response rate - If self-response rate is below the
• Regular programme risk reviews and reporting
•
P
r
o
m
o ti
n
g
s
el f - r
e
sponse by providing response choice (e.g.
required level to achieve efficiencies then additional
• Regular programme improvement focused
respondents can request a paper form)
investment is required to meet planned targets. Low self-
retrospective reviews • Self response is encouraged by using a wave approach to non-
response rates may also compromise realisation of benefits. • Programme management support and oversight from
response fol ow-up (mail out of internet codes and other information is
Statistics NZ EPO
fol owed up, where necessary, with reminder letters, provision of
• Agreed thresholds of tolerance with immediate
requested forms and support)
escalation to central agencies where significant
• Flexibility wil be built into field management designs
variance from agreement
•
T
h
e
c
ol l
e
c t i
o
n solution wil be designed to facilitate real time field
• Oversight by central agencies of major projects - 2018 management.
Census and Statistics 2020
• A rigorous and comprehensive testing programme is planned. There
are clear mechanisms for esacaltion and reporting of any issues
arising in the field. The aim here is for early warning and rapid
response to what we are learning through ongoing testing in the field.
• Daily MIS reporting through operational phases wil provide ongoing
business intel igence which wil enable targeted, real time management
of field responses to issues
• Self-response is supported by increased end-to-end public
communications
• Testing the impact of communications approaches on self-response
rates is integrated into the 2018 Census testing programme
• Creating a new Communications and Marketing Manager
6 Communications strategies - If planned communications
• Regular programme risk reviews and reporting
• I
n
c r
e
a s i
n
g c
o
m
m
unications investment compared with 2013 Census
strategies do not adequately compensate for new methods in • Regular programme improvement focused
• Creating a new Communications and Marketing Manager
the 2018 Census, response targets wil be impacted and
retrospective reviews • Communications planning wil begin early in the census cycle
compromise realisation of benefits.
• Programme management support and oversight from • The effects of communications strategies wil be tested and plans
Statistics NZ EPO
refined as a result of testing
• Agreed thresholds of tolerance with immediate
• Lessons learned from the 2013 Census (an evaluation of the 2013
escalation to central agencies where significant
communications package is underway) and from overseas, wil be
variance from agreement i
n
t
e
g r
a t
e
d
i
n t
o
2018 communications planning
• Oversight by central agencies of major projects - 2018 • Addditional focus on reactive communications to manage known and
Census and Statistics 2020
emerging col ection issues
7 National address list - If the national address list has a
• Regular programme risk reviews and reporting
•
G
ai
n
a
n
u
n
d
e r
s t
a
nding of the quality of the available address
significant number of missing addresses (or duplicate
• Regular programme improvement focused
information early in the 2018 Census programme using 2013 Census
addresses), which results in mailout targets not being met
retrospective reviews address information compared to the NZ Post address database.
this wil require higher levels of col ector visits to respondents • Programme management support and oversight from
• Current settings and costings al ow for 100% address canvassing
(i.e door-knocking) and therefore higher costs than
Statistics NZ EPO
(physical checking of addresses against the national address list)
estimated.
• Agreed thresholds of tolerance with immediate
prior to the first mailout to dwel ings
escalation to central agencies where significant
• Enhanced management information systems are being built to
variance from agreement
p r
o
vi
d
e
r
e
al t i
m
e information, back and forward from the field, to
• Oversight by central agencies of major projects - 2018 enable quick response (mitigation actions) to significant issues.
Census and Statistics 2020
23
2018 CENSUS STRATEGY
Version 1.0 endorsed by Census Programme Board 22 October 2014 and adjusted for
Statistics New Zealand’s Strategic Refresh 14 January 2015
1. Purpose
This document sets out the strategic direction for the 2018 New Zealand Census of
Population and Dwellings. The strategy wil focus our efforts and guide the work of the
programme and the output of the 2018 Census in order to meet the needs of our customers.
2. Organisational context
Data is a national strategic asset. Statistics New Zealand's purpose is to empower decisions
by adding value to New Zealand's most important data and the Census of Population and
Dwellings has a key role to play in this.
To help improve outcomes for all New Zealanders, our goal is to:
double the value of the data provided by Statistics New Zealand to New Zealand
by 2018 with a ten-fold increase in the value of data provided by Statistics New
Zealand to New Zealand by 2030.
In order to meet this goal Statistics New Zealand has defined a number of strategic roles for
the department:
provider -provide independent and trusted data
enabler - enable New Zealand decision makers
innovator - innovate to drive value for customers
steward - steward data now and in the future
The 2018 Census has a part to play in each of these roles and this is reflected in the
strategic goals for the census.
3. Aim of the census
The New Zealand Census of Population and Dwellings (the census), conducted since 1851
and a legislative requirement under the Statistics Act 1975, produces authoritative and, in
some cases, the only comprehensive source of population and dwelling statistics, including
information on small communities and small populations. The 2018 Census will be the 34th
census conducted in New Zealand.
The aim of a census in New Zealand is to produce the best possible count of the population
and dwellings and deliver the count, along with other related statistical information, to
customers through a useful and accessible range of products and services.
4. Background
High-quality, relevant census information is critical to New Zealand, with an estimated net
present value over the next 25 years of close to $1.4 billion1. The census model needs to
change if it is to be sustainable over time. Compelling drivers for change include:
rising costs and the potential for efficiencies
1 Bakker, Carl.
Valuing the Census: A report prepared for Statistics New Zealand which quantifies the benefits to New
Zealand from the use of census and population information. (Wellington: Taylor Duignan Barry, April 2013)
the ongoing trend of declining response rates (particularly for small sub-groups and
small areas)2
the increasing availability and sharing of administrative data across government, and
technological advances.
As part of its Census Transformation Strategy, Statistics New Zealand is investigating the
production of census information increasingly from administrative sources, rather than the
current “full enumeration” survey model. Internationally, countries with administrative
censuses produce census information at much lower cost, and information can be produced
more frequently.
In February 2014, Statistics New Zealand confirmed that a modernised census in 2018 is stil
the best available method to deliver changes needed for the next census. While existing
administrative data cannot, at present, replace current census statistics, there is potential to
use existing administrative data within the current census model. Transforming the census to
a model based solely on administrative data remains under investigation for longer-term
change, through the Census Transformation Project.
In the short-to-medium term, modernising the survey-based census model will significantly
improve the efficiency and effectiveness of the census and provide a step on the way to a
census based on administrative data. Work will also be undertaken within the 2018 Census
programme to further test the potential of replacing traditional census collection with
administrative data, building on work already underway comparing the quality of current
administrative data with 2013 Census data.
In June 2014, Cabinet approved funding for a modernised 2018 Census based on a detailed
business case with the following investment objectives:
To undertake a Census of Population and Dwellings in 2018 that meets statutory
requirements
To at least maintain the quality of census information compared with the 2013
Census
To improve, by 20%, the timeliness of census information products released to
customers following a 2018 Census, compared with the 2013 Census
To reduce, by 5% over two full census cycles, the average cost of the census,
starting in 2014 and using the 2013 Census adjusted for annual inflation at 2%.
5. The 2018 Census Strategy
A key part of the modernisation that underpins the 2018 Census strategy is to radically alter
the mix of modes used in the current collection model, by introducing mail-out, not delivering
paper forms to every dwelling and making the internet the primary response mode. At the
same time we must maintain the accuracy of basic counts achieved in recent censuses.
The enhanced multi-modal census design will modernise census operations, support
improved interaction between the public and government, and reduce the real cost of census
data collection.
Communication strategies to inform respondents – all people in New Zealand on census day
– will need to expand, not only to call people to complete their census forms, but also to
inform people about the changed collection model. For example, many people wil not see a
2 At a national level, 1996 Census achieved net coverage of 98.4% and a response rate of 95.6%. 2006 Census achieved net
coverage of 98% and a response rate of 94.8%. The 2013 Census followed this downward trend, achieving net coverage of
97.6% and a response rate of 92.9%. These national rates may be higher or significantly lower for sub-populations and mall
population areas.
2
census collector on their doorstep before census day and they will not receive paper forms
by default.
Given the magnitude of the planned change and the risk involved to the quality of census
counts, a thorough development and testing programme is required.
6. Strategic goals
The key strategic goals for the 2018 Census are:
1. Improve data quality while modernising
2. Reduce the cost of collection operations
3. Contribute to organisational capability
4. Increase use of administrative data
5. Make digital engagement easy
6. Adopt test-driven development
7. Deliver customer-driven products and services
The following pages provide further information on these seven goals and highlight the
strategies and actions to achieve them. Note that the goals are not in priority order and that
some actions apply to several goals.
3
Strategic Goal 1: Improve data quality while modernising
Inherent risks to information quality come with change to our business model for the census.
The step-change for the 2018 Census means that risks to data quality are much greater than
for the 2013 Census, which was much the same as the 2006 Census. While modernising,
we must protect the accuracy of basic counts, as they are the key reason for taking a
census.
There are six elements of statistical quality: accuracy, relevance, coherence, timeliness,
accessibility and interpretability. Al wil have quality targets for 2018.
In 2018 we aim to maintain high coverage and response rates and address sub-group and
subnational variation in responses rates, while modernising. Specific emphasis will be on
improving Māori response rates, after their decline in the 2013 Census and recognising the
importance of census information for Māori.
Emerging data needs will be considered for new topics in the 2018 Census, but decisions on
content wil reflect the role the census plays in an integrated official statistics system.
Objectives and strategies
Actions
Ensure accuracy of national counts and reduce variation in subnational response rates Design, test and implement
Develop modern collection processes to enable real-time
more modern and targeted
management of field resources
collection approaches
Undertake thorough development and testing of changes
Target hard-to-reach populations
Develop new communication
Review the statistical quality management strategy
approaches to support the new
collection process
Research other models to better understand where non-
response pockets wil occur
Undertake a post-census coverage check, e.g. Post-
enumeration Survey
Increase public communications to compensate for less
doorstep communication
Develop public communication strategy to encourage self-
response and support non-response follow-up
Diversify communication strategies to reflect population
changes
Ensure an integrated communications strategy for the
census, under the Statistics New Zealand brand
Improve Māori national and subnational response rates Design, test and implement
Ensure understanding of Māori responsiveness across the
specific Māori collection
Census Branch in preparation for engagement
strategies
Implement a specific strategy to improve Māori response
rates
Develop Māori communication
Focus on Māori when making modernisation changes
strategies that support the new
collection process and
Focus on Māori when diversifying communications strategies
and integrating with the Statistics NZ brand and
encourage high Māori
communications strategy
response
Maintain relevant, coherent and fit-for-purpose census information with lower processing
cost
4
Improve relevance of census
Make necessary content changes for emerging key data
content and maintain quality of
needs
foremost and defining variables Make content-related changes driven by new methodologies
Make necessary content changes for foremost variables
Implement the relevant
Manage customer expectations through targeted content
changes to drive the ‘make
consultation processes
digital engagement easy’ goal
Include automatic edits, auto-fil ing and prompts for the
Review statistical quality
online form
management strategy
Improve edit and imputation approaches
Use IDI data to:
Use administrative data to
o improve work variable quality
reduce cost of processing
o replace work and income-related responses
supplementary variables
Strategic Goal 2: Reduce the cost of collection operations
The cost of distributing and collecting census forms is substantial, with col ection operations
consuming over half the census budget. Unless major changes are made to the “full
enumeration” census model, the cost of census-taking wil keep climbing with more dwellings
and inflation.
The new collection model uses a property data frame to mail out internet access codes to
households, followed by waves of reminder letters, which may include paper forms. Field
staff wil focus on delivering forms in those areas not covered by the mail-out, and on the
extensive targeted non-response follow-up activity. Re-engineering of the collection model
will increase data-processing efficiency by achieving better integration of the collection and
data-processing phases.
Change to the type and extent of census communications will inform the public of the
changed processes.
Extensive development and testing wil successfully transform the current business model.
The new model, introduced in 2018, wil be extended in the next census.
Objectives and strategies
Actions
Reduce collection cost Change forms delivery process Document the new business process model and develop an
to maximise self-response
end-to-end statistical design
Work with the Property Data Frame Project to develop the
frame
Draw on other countries’ expertise
Use canvassing to improve address list quality
Redesign field operations and
Redevelop and test methodologies to support the new
processes to reduce the
collection model
number of field staff required
Produce real-time management information
Increase use of digital technologies for all field processes
Integrate systems, vendors and business processes
Use addressing quality to identify mail-out target areas
Plan for when and how we use paper forms
Use a ‘wave’ approach to follow up the initial mail out with
reminder letters and paper forms
Redevelop field structure, field scheme, collector workload
model and pay methodology
Employ field staff with skil s to implement targeted non-
response follow-up procedures
Use enterprise col ection
Build case management functionality with the Enterprise
systems
Collection Platform team
5
Increase integration between
Build a processing system adapted for census with the
collection and processing
Household Processing Platform team
systems
Reduce processing effort by promoting online response and
more coding by respondents
Speed up sending fieldbook information to the processing
centre
6
Strategic Goal 3: Contribute to organisational capability
The census is a large operational activity, progressively integrating with corporate initiatives
under Stats 2020. The 2018 Census wil collaborate in and contribute to modernisation
across a range of Statistics New Zealand projects including the household processing
platform, the property data frame, the Enterprise Collection Platform as well as
dissemination tools and services.
The census will integrate with corporate and all-of-government Information Technology
initiatives.
Objectives and strategies
Actions
Improve corporate efficiency by integrating with other corporate business processes Collaborate in a range of Stats
Build collection case management capability with the
2020 projects
Enterprise Collection Platform project team
Work with the Property Data Frame Project to develop the
frame
Specify census requirements for the Household Processing
Platform
Develop customer-driven census products integrated with
other corporate products
Enable Statistics New Zealand to benefit from census-
product expertise
Use corporate and all-of-Government procurement
processes
Grow our people Ensure census staff maximise
All census staff have development plans, incorporating
their skil s and experience
Statistics New Zealand requirements
Census staff contribute to corporate projects
Use corporate and all-of-government Information Technology Integrate census, corporate
Use the Geospatial Management System to develop
and government IT processes
electoral and automatic rebasing systems
and platforms
Use corporate dissemination platforms
Use all-of-government IT services, especially IT
Partner with other agencies to
infrastructure
deliver and use shared
Include Census in the corporate IT Disaster Recovery
services
solution
Make participation in the census easy, in line with Better
Public Services Result Area 10
Develop and implement a procurement strategy
7
Strategic Goal 4: Increase use of administrative data
Statistics New Zealand will use more administrative data in the 2018 Census. An
administratively sourced address list of most New Zealand households wil be used to mail
out internet access codes, efficiently reusing existing information, in line with the
Government ICT Strategy. Administrative data may supplement responses from some
groups (e.g., overseas visitors) and will help with coding some variables. The Census
Transformation Project wil continue to explore the wider use of administrative data in the
census.
This goal is consistent with Statistics New Zealand’s strategy of “administrative data first”.
Statistics New Zealand intends to change the balance of administrative data and direct
surveys. The department’s aim is that by 2020 administrative data wil be the first source of
information, supplemented where necessary by direct collection.
Objectives and strategies
Actions
Achieve collection efficiencies by using administrative data Use administrative sources of
Build a property data frame with the Geospatial team for
address data in the new
mailing out internet codes to most households
census collection process
Ensure legislative changes needed to use administrative
data are made in time
Investigate using administrative Use management information to measure coverage and
data to replace census forms
response rates in real time
altogether
Investigate data sources for replacing census forms
altogether for some non-private dwelling types and unit non-
response
Use administrative data to improve processing efficiency and data quality Use administrative data in the
Replace manual coding of workplace address and industry
coding process
with administrative data
Experiment with administrative data for imputation
Use many administrative data sources for evaluation during
field operations and traditional data evaluation
Use administrative data to reduce respondent burden in post-2018 censuses Use the 2018 Census to test
If feasible, substitute administrative data for responses in
replacing some questions with
2018 and beyond
administrative data in the next
If feasible, include an embedded test for the next census in
census
the 2018 Census
Use the 2018 Census as a benchmark for assessing alternative administrative models Collaborate with the long-term
Make data from the 2018 Census available for the Census
census transformation strategy
Transformation Project
Design the 2018 Census processing system to integrate with
administrative sources
Integrate census data with data from other collections Use the 2018 Census to test
If feasible, substitute administrative data for census data in
data integration and
the census after 2018
substitution for the next census
Strategic Goal 5: Make digital engagement easy
The internet is the first place many people go to for information. New technologies and
applications bring people and content together and allow the repurposing and reuse of data
and information in ways that until recently were either impossible or not considered.
For the 2018 Census online response will be the primary response mode, compared with
mail-back or collector pick-up of paper forms. The online response rate is expected to
exceed 70%, with up to 20% of forms mailed back and 10-15% picked up by collectors.
Statistics New Zealand wil promote the online option to maximise uptake as well as
enhancing the collection systems for a more environmentally sustainable business model.
Statistics New Zealand wil introduce online consultation to increase customer participation.
This wil precede targeted consultation on content and outputs. Communications across all
census phases wil be driven by developments in online social media.
Objectives and strategies
Actions
Make it easy for customers and data suppliers to interact digitally with Statistics New
Zealand Use a customer and supplier-
Understand issues and design from supplier and customer
centric design and delivery
points of view
approach
Use marketing techniques and clear communication
messages
Develop processes and systems that make it easy for the
respondent to participate in the census
Prototyping and usability testing
Make online the primary response mode by transforming collection and communications Design the questionnaire for
Improve data quality with the online form
the online form and mobile
Enable respondents to fill in their forms anywhere, anytime
devices
Enable the ‘hard to reach’ to use the online option, while
removing cost and improving access
Ensure paper forms capture
essential information without
constraining the online form
Use digital communication
Maximise use of social media to promote the census and
channels to promote and
improve response rates
support the new col ection
Target communications to specific subgroups and regions
model
with intelligent use of real-time data
Make online the dominant mode for respondent help and communication Encourage respondents to use
Redesign online respondent help
online help for collection and
Introduce additional online help and support mechanisms
dissemination activities
Develop smarter online forms Improve online-form usability
Review the questionnaire development process
and emphasise “on-line first”
Improve the way online help is stored, presented and
selected
Cater for more types of computer devices
Investigate the benefits of a resumable online form
Investigate using multiple languages for online census forms
Build predictive text features into the online form
Investigate the use of code files and address list information
to automate some responses
Introduce more online edits, respondent prompts and auto-
fil ing capability
Integrate the address list with online form responses
Increase public participation and awareness of census statistics Use online technology for
Use online consultation processes
consultation, communication
Increase the use of social media for communications
and dissemination activities
Increase the amount of census information available online
Make all output and dissemination activities available online
9
10
Strategic Goal 6: Adopt test-driven development
Extensive testing will ensure the 2018 Census meets important information needs. A
comprehensive testing programme wil verify the end-to-end impact of modernising the 2018
Census.
An enhanced test-driven development model wil be used to manage the level of change in a
modernised census. This will ensure that an optimal mix meets key performance indicators
and investment objectives, and delivers quality information, within budget.
Objectives and strategies
Actions
Use the testing programme to reduce risk and ensure quality Integrate the testing
Create an integrated test strategy and plan
programme with the develop-
Use international experience to develop best-practice testing
test-learn-revise cycle,
methodologies
emphasising reduction of risk
Conduct a series of tests, including a major field test in 2015-
associated with change
16 and a Dress Rehearsal in 2017
Plan early and ensure sufficient
Use an iterative process of testing early and often, refining
the test plan as findings emerge
budget for development and
testing
Use the 2018 Census to support the transformation testing
programme
Use evidence from the testing programme for timely and
robust decision-making
Set up a sufficiently resourced testing team early
Bring forward procurement to ensure suppliers participate in
the major tests
Implement a comprehensive risk-management strategy
Use testing as a sound platform for future census transformation Use the 2018 Census
If feasible, conduct embedded tests during the 2018 Census
Programme to inform the
Incorporate Census Transformation test plans into 2018
Census Transformation Project
Census test plans
Include the Census Transformation Project team as a
reviewer of key 2018 Census documents
11
Strategic Goal 7: Deliver customer-driven products and services
The Census of Population and Dwellings is part of an integrated national social and
population statistics system. It provides benchmark population counts at national and local
levels on a regular basis, to satisfy legislative requirements. It has a unique role as an
important source of demographic, social and economic data for small geographical areas
and sub-populations.
Customers’ expectations of timely delivery of census data are increasing – they are no
longer willing to wait the usual 7-9 months from census day. Statistics New Zealand will
need to modernise systems, processes, and products and services to disseminate census
data sooner. This wil involve using corporate developments such as the dissemination
platform. It wil mean looking internationally to other agencies for products, services and
confidentiality methods that can be used to produce the 2018 Census outputs.
Customers expect to access data in different formats from different platforms, wherever they
are. To support these changing priorities, the systems and processes used to disseminate
census data wil have to change.
Objectives and strategies
Actions
Use modern dissemination methods to improve customer access to data Ensure confidentiality methods
Focus on customer information needs when developing
meet the needs of a range of
products and services
customer-driven products and
Extend the range of customer-defined products
services
Change confidentiality methods to support an updated
product mix, including on-the-fly confidentiality
Release census information to customers earlier than we did for the 2013 Census Use corporate dissemination
Change data col ection, processing and evaluation to deliver
platforms by default, with
final census data earlier
census-specific tools only
Test output production systems and processes in time for
when needed
early operational use
Plan dual output production phases (early and status quo) of
all products and services, while managing the associated risk
Introduce mobile and interactive dissemination Develop mobile-optimised
Build or procure products optimised for the mobile customer
products and dissemination
Regular use of infographics, mapping and other
technology to widen support of
visualisations, to support dissemination communications
embedded census data
objectives
Introduce tools to give customers more control over data
12
Document Outline
