Information about the Population Estimates

Document Map

Jump to: Availability

Jump to:Design

Jump to:Output Variables

Jump to:Guide To Interpreting Data

Jump to:Customised Output

Jump to:Catalogue & Reference Numbers

Jump to:Commercially Available Tables

Jump to:Other Comments

Jump to:Classification(s) Used

Jump to:Glossary of Terms

Jump to:Contact Details

Availability

Valid From: ..	03/31/1996
To: ..	06/30/2008 12:00:00 AM
Frequency: ..	Quarterly

---

Design
Purpose: The estimated resident population of New Zealand indicates the number of residents and their distribution by sex, age and subnational areas.

Population estimates are used for many purposes in all areas of society and the economy; including planning, policy formation, and monitoring change in health, education, electoral areas, superannuation, territorial authorities, transport; and in making business decisions. Estimates are used extensively as 'bottom lines' in the calculation of rates such as death rates, participation rates in education, per capita consumption of alcohol, etc.

General Information ..Contents
Different Population Measures
Component Methodology
Base Population
National Population Estimates
Subnational Population Estimates
Mäori Ethnic Group Population Estimates
Historical De facto Population Estimates
Timetable
Rounding of Population Estimates

Different Population Measures
There are two main population concepts of interest:

1. all people present in a given area at a given time (the 'de facto population concept')
2. all people who usually live in a given area at a given time (the 'resident population concept').

Both of these populations can be counted or estimated.

Population counts
The New Zealand Census of Population and Dwellings counts people where they are on census night (census night population count). As this may not be where they usually live, respondents are also asked to provide the address of their place of usual residence (census usually resident population count).

A person's usual residence is generally self-identified. Visitors intending to be in New Zealand for less than 12 months are generally considered to usually live overseas.

Population estimates
To provide population data between census dates, Statistics New Zealand estimates the population, using the most recent census data as a base. The estimated resident population is updated regularly for population change due to births, deaths and net migration (arrivals less departures) of residents. The estimate gives the best measure of the population that usually lives in an area, for a limited range of variables (age, sex, ethnicity) and limited geographic areas. Population estimates apply either to the number of people at a particular time ('at') or the average over a period of time ('mean'), usually a year.

Population projections
The size and composition of the population at a future date based on specified assumptions about future fertility, mortality and migration. The projected resident population has as a base the estimated resident population at a given date.

Population measures
Historically, population data comprised counts of all people present in an area at a given time. However, increased geographic mobility means people are more likely to move between New Zealand and overseas, and between areas of New Zealand, for work, study and holidays. There is therefore a need for data on the people who usually live in an area at a given time, as this is more relevant for many planning purposes (eg housing, schools).

The question often arises as to what is the best measure of the population of an area. This partly depends on what use will be made of that population measure, and what information is available to calculate the population measure. In some instances it may be appropriate to exclude/include visitors who are temporarily present in an area and residents who are temporarily absent from an area.

There are three population measures commonly produced by Statistics New Zealand:

1. Census night population count: A count of all people present in a given area on a given census night.
   
   The census night population count of New Zealand includes visitors from overseas who are counted on census night, but excludes New Zealand residents who are temporarily overseas.
   
   For a subnational area the count includes visitors from overseas and elsewhere in New Zealand (people who do not usually live in that area), but excludes residents of that area who are temporarily elsewhere on census night (people who usually live in that area but are absent).
   
2. Census usually resident population count: A count of all people who usually live in a given area, and are present in New Zealand, on a given census night.
   
   The census usually resident population count of New Zealand excludes visitors from overseas and excludes New Zealand residents who are temporarily overseas.
   
   For a subnational area the count excludes visitors from overseas and elsewhere in New Zealand (people who do not usually live in that area), but includes residents of that area who are temporarily elsewhere in New Zealand on census night (people who usually live in that area but are absent).
   
   Residents are counted at the meshblock (smallest geographic area for statistical purposes) of their usual address. If insufficient information about a person's usual address is given, then the geographic information that is available is used to select (impute) a meshblock within that area.
   
3. Estimated resident population: An estimate of all people who usually live in a given area at a given date.
   
   The estimated resident population of New Zealand includes all residents present in New Zealand and counted by the census (census usually resident population count), residents who are temporarily overseas (who are not included in the census), and an adjustment for residents missed or counted more than once by the census (net census undercount). Visitors from overseas are excluded.
   
   For a subnational area the estimate excludes visitors from overseas and elsewhere in New Zealand (people who do not usually live in that area), but includes residents of that area who are temporarily elsewhere on census night (people who usually live in that area but are absent).
   
   The estimated resident population at a given date after census includes births, deaths and net migration (arrivals less departures) of residents during the period between census night and the given date.
   
   National population estimates are produced quarterly (reference dates at 31 March, 30 June, 30 September and 31 December) and subnational population estimates are produced annually (reference date at 30 June).

The following table summarises the inclusions and exclusions of each population measure for a subnational area on census night:

Group	Census night population count	Census usually resident population count	Estimated resident population(1)
Residents who are present	Yes	Yes	Yes
Residents who are temporarily elsewhere in New Zealand	No	Yes	Yes
Residents who are temporarily overseas	No	No	Yes
Visitors from elsewhere in New Zealand	Yes	No	No
Visitors from overseas	Yes	No	No

(1) Usually includes an adjustment for net census undercount of residents. For a given date after census, the estimated resident population is also updated for births, deaths and net migration (arrivals less departures) of residents during the period between census night and the given date.

The following table illustrates the differences between the three population measures in 2001:

Selected area	Census night population count (6 March 2001)		Census usually resident population count (6 March 2001)		Estimated resident population (30 June 2001)
			(000)
Auckland City	380		368		389
Manukau City	284		283		298
North Shore City	185		185		194
Auckland Main Urban Area	1,087		1,075		1,130
Auckland Region	1,174		1,159		1,217
Christchurch City	324		316		327
Waimakariri District	37		37		38
Banks Peninsula District	9		8		8
Christchurch Main Urban Area	342		334		346
Canterbury Region	495		481		497
New Zealand	3,821		3,737		3,880

In a major change from historical practice, all population estimates, population projections and derived demographic indices produced by Statistics New Zealand from 1991 adopt the resident population concept. Previously, these figures were based on the de facto population concept. Statistics New Zealand has adopted the resident population concept to ensure that population estimates reflect more accurately the population that usually lives in each area. This concept is preferred for most central and local government administration and planning purposes. Further information on the resident population concept is provided in the Demographic Trends 1996 article 'The 1996 Post Enumeration Survey (PES) and Post-censal Population Estimates'.

Post-censal population estimates and projections are not directly comparable with census counts. Census counts give a snapshot of the population at that time but make no allowance for New Zealand residents temporarily overseas nor for net census undercount.
Back to top

Component Methodology
All estimates of population undertaken by Statistics New Zealand use the component methodology, where the components of population change (births, deaths and migration) are used to update a base population or a previous estimate of population. Censuses of Population and Dwellings provide the starting point for deriving population estimates. Estimates are labelled 'provisional' if they are based on incomplete data. 'Final' estimates are based on complete data (ie when all data required by the estimation methodology is available).

Population estimates are normally only defined for events that have already occurred. This feature distinguishes them from population projections, which are concerned with what will happen in the future, given specific assumptions about future births, deaths and permanent and long-term migration.

The methodologies described below are, apart from that for de facto population estimates, those that are currently used to estimate population numbers.
Back to top

Base Population
The estimated resident population at 30 June in the year of the latest census forms the base population for deriving post-censal population estimates and population projections. The base population is also used in the revision of the intercensal population estimates for the 5-year period prior to the latest census. The base population is produced by single years of age, sex, ethnicity (Mäori, Pacific Peoples, Asian, European, Middle Eastern/Latin American/African, Other) and area unit.

The base estimated resident population of New Zealand, at 30 June in the year of the latest census is based on the latest census usually resident population counts updated for:

1. net census undercount;
2. the number of residents temporarily overseas on census night;
3. births, deaths and net migration (excess of arrivals over departures) between census night and 30 June;
4. reconciliation with demographic estimates at the youngest ages; and
5. non-response to the census ethnicity question (2001 and 2006 only).

The following analysis looks at the derivation of the base population at 30 June 2006. An outline of the adjustments made following the 1996 and 2001 Censuses can be found at the end of this section.

The impact of the adjustments on the estimated resident population at 30 June 2006 is summarised in the following table. These adjustments are also discussed in more detail below.

Adjustment	National estimates	Subnational estimates	Ethnic estimates	Age estimates
Net census undercount	Yes	Yes	Yes	Yes
Residents temporarily overseas on census night	Yes	Yes	Yes	Yes
Births, deaths and net migration between census night (7 March 2006) and 30 June 2006	Yes	Yes	Yes	Yes
Reconciliation with demographic estimates at ages 0-4 years	Yes	Yes	Yes	Yes
Non-response to the census ethnicity question	No	No	Yes	No

Net Census Undercount
The 2006 Post-enumeration Survey (PES), conducted shortly after the 2006 Census, provided estimates of net census undercount. The age, sex and ethnic differentials in net undercount, as measured by the PES, provide the basis for modelling net undercount for those variables combined. For more information on the PES, refer to the results from A Report on the 2006 Post-enumeration Survey. For more information on the modelled adjustments, refer to Adjustment of Post-censal Population Estimates for Census Undercount.

Residents Temporarily Overseas on Census Night
From external migration data it was determined that 63,900 civilians were temporarily overseas at the time of the 2006 Census. From Defence Force information it was determined that 600 military personnel were temporarily overseas at the time of the 2006 Census. External migration data was used to provide the age, sex and ethnic (using country of birth as a proxy) characteristics of residents temporarily overseas. Census information on absentees overseas was used to provide the geographical usual residence of residents temporarily overseas.

'Temporarily overseas' is defined as being out of New Zealand for a period of less than one year. Most of those New Zealand residents who temporarily depart from New Zealand do so for relatively short periods of time. Over 80 percent of residents temporarily overseas return to New Zealand within one month of departure, and about 95 percent return within three months of departure.

Births, Deaths and Net Migration Between Census Night (7 March 2006) and 30 June 2006
There were an estimated 18,900 live births during 8 March 2006 to 30 June 2006. This figure includes a small adjustment for late registration of births. There were an estimated 8,900 deaths during 8 March 2006 to 30 June 2006. For deaths, no adjustment for late registration was required given the pattern of registration from previous years.

The net migration component can be divided into two sub-components – external and internal migration. From external migration data it was determined that there was a net loss of 1,300 people from the New Zealand population during 8 March 2006 to 30 June 2006. The age, sex, ethnicity and geographical usual residence of migrants was based on information from external migration records. Internal migration was assumed to be zero for each area unit because there was no reliable validated evidence of internal migration during this short period.

Reconciliation with Demographic Estimates at Ages 0-4 Years
Demographic estimates of population aged 0-4 years at the national level were derived from births, deaths and external migration data, independently of census. At the national level the demographic estimates are considered more accurate than census-based estimates because birth and death registrations, and external migration data, currently have virtually complete coverage. By comparison, the accuracy of census coverage at individual ages cannot be reliably estimated from post-enumeration surveys because of the limitations of sample size and associated sampling errors. However, beyond ages 0-4 years the demographic estimates become increasingly susceptible to migration category jumping – migrants changing their status because of differences between intended and actual length of stay/absence as reported on the passenger arrival/departure cards. The ratio of demographic estimates to census-based estimates at the national level was used to weight subnational and ethnic population estimates at ages 0-4 years.

Non-Response to the Census Ethnicity Question
At the 2006 Census there were 167,784 residents who did not state an ethnicity. These people have been assigned an ethnic indicator (Yes/No) for each of the broad Mäori, Pacific, Asian, European, Middle Eastern/Latin American/African (MELAA), and Other ethnicities based on the characteristics (age, sex, geographical area) of the 3.86 million residents who did specify an ethnicity/ethnicities.

The imputation adjustment for the various ethnic groups in 2006 was as follows: Mäori (+28,000), Pacific (+18,000), Asian (+17,000), European (+105,000), MELAA (+2,000) and Other (+17,000). The size of the adjustments made to each of these ethnic groups is larger than the corresponding adjustments made for census undercount. The estimated resident population at 30 June 2006 for Mäori was 10 percent higher than the enumerated census usually resident population count. Similarly, the estimated resident populations for the Pacific, Asian, European, MELAA and Other ethnic groups were 13, 14, 7, 11 and 7 percent higher, respectively, than the census usually resident population counts.

Users of ethnic specific population data need to be aware of this adjustment. The 2006-base Mäori ethnic group population estimates incorporate the adjustments for ethnic non-response outlined above, as do the ethnic-specific population projections. Census counts however, are not adjusted for those people not specifying their ethnicity.

Estimated Resident Population at 30 June 2006
The census usually resident population count at the 2001 Census was 4,027,947. The estimated resident population at 30 June 2006 was 4,184,600 – an increase of 156,600. A summary of the adjustments made to the census usually resident population count at the 2006 Census, to arrive at the estimated resident population at 30 June 2006 (ie base population), is given in the table below.

Adjustment	Number
2006 Census usually resident population count	4,027,947
+ net census undercount	+80,100
+ residents temporarily overseas (civilian and military)	+64,500
+ births (8 March to 30 June 2006)	+18,900
- deaths (8 March to 30 June 2006)	-8,900
- net migration (8 March to 30 June 2006)	-1,300
+ reconciliation at ages 0-4 years	+3,400
= Estimated resident population at 30 June 2006	= 4,184,600

Adjustments to the 2001 Base Population

(a) Net Census Undercount : The 2001 Post-enumeration Survey (PES), conducted shortly after the 2001 Census, provided some estimates of net census undercount. The age, sex and ethnic differentials in net undercount, as measured by the PES, provide the basis for modelling net undercount for those variables combined. For more information on the 2001 PES, refer to the results from the 2001 Post-enumeration Survey and A Report on the Post-enumeration Survey (2001).

(b) Residents Temporarily Overseas on Census Night: From external migration data it was determined that 53,300 civilians were temporarily overseas at the time of the 2001 Census. From Defence Force information it was determined that 1,200 military personnel were temporarily overseas at the time of the 2001 Census. (External migration data does not include movements of military personnel on duty). External migration data is used to provide the age, sex and ethnic (using country of birth as a proxy) characteristics of residents temporarily overseas. Census information on absentees overseas is used to provide the geographical usual residence of residents temporarily overseas. For more information refer to New Zealand Residents Temporarily Overseas at the 2001 Census.

(c) Births, Deaths and Net Migration Between Census Night (6 March 2001) and 30 June 2001: There were an estimated 17,400 live births during 7 March 2001 to 30 June 2001. This figure includes an adjustment for late registration of births based on the pattern of registration from previous years. There were an estimated 8,800 deaths during 7 March 2001 to 30 June 2001. For deaths, no adjustment for late registration was required given the pattern of registration from previous years. The net migration component can be divided into two sub-components – external and internal migration. From external migration data it was determined that there was a net loss of 4,500 people from the New Zealand population during 7 March 2001 to 30 June 2001. The age, sex, ethnicity and geographical usual residence of migrants was based on information from external migration records. Internal migration was assumed to be zero for each area unit because there was no reliable validated evidence of internal migration during this short period.

(d) Reconciliation with Demographic Estimates at Ages 0-9 Years: Demographic estimates of population aged 0-9 years at the national level were derived independently of census from births, deaths and external migration data. At the national level the demographic estimates are considered more accurate than census-based estimates because birth and death registrations, and external migration data, currently have virtually complete coverage. By comparison, the accuracy of census coverage at individual ages cannot be reliably estimated from post-enumeration surveys because of the limitations of sample size and associated sampling errors. However, beyond ages 0-9 years the demographic estimates become increasingly susceptible to migration category jumping – migrants changing their status because of differences between intended and actual length of stay/absence as reported on the passenger arrival/departure cards. The ratio of demographic estimates to census-based estimates at the national level was used to weight subnational and ethnic population estimates at ages 0-9 years.

(e) Non-Response to the Census Ethnicity Question: At the 2001 Census there were 150,546 residents who did not state an ethnicity. These people have been assigned an ethnic indicator (Yes/No) for each of the broad Mäori, Pacific, Asian, European and Other ethnicities based on the characteristics (age, sex, geographical area) of the 3.587 million residents who did specify an ethnicity/ethnicities. This was the first time that Statistics New Zealand carried out imputation procedures for ethnic non-response, when deriving the base estimated resident populations. Following the 1996 Census, those people not specifying their ethnicity were excluded from the ethnic specific estimated resident populations at 30 June 1996, and hence were not incorporated into the ethnic specific population projections, or the estimated resident populations for the Mäori ethnic group. The imputation adjustment for the various ethnic groups in 2001 was as follows: Mäori (+26,000), Pacific (+13,000), Asian (+11,000), European (+113,000) and Other (+1,000). The size of the adjustments made to each of these ethnic groups was larger than the corresponding adjustments made for census undercount. The estimated resident population at 30 June 2001 for Mäori was 11 percent higher than the enumerated census usually resident population count. Similarly, the estimated resident populations for the Pacific, Asian and European ethnic groups were 13, 14 and 7 percent higher, respectively, than the census usually resident population counts. Users of ethnic specific population data need to be aware of this change in methodology. The 2001-base Mäori ethnic group population estimates incorporate the adjustments for ethnic non-response outlined above, as do the ethnic-specific population projections. Census counts however, are not adjusted for those people not specifying their ethnicity.

(f) Usual Residence of Students: The 2001 Census 'Help Notes' instructed tertiary students to put their term-time address as their usual address. However, comparison with previous census results suggested some students had not put their term-time address as their usual address. The estimated resident population of nine area units, spread over six territorial authorities, was adjusted upwards by between 100 and 700 people. These were offset by small decreases to other area units. The net adjustment at the territorial authority level was most significant for Dunedin City (+1,400). Christchurch City, Palmerston North City, Hamilton City and Selwyn District were also adjusted upwards by between 150 and 250 people.

(g) Centenarians: International experience suggests age misreporting is relatively significant among centenarians (age 100+ years). The single year of age distribution of counts from the 2001 Census showed a 'spike' at age 100 years. From closer examination of census records, it was determined that about 100 centenarians (out of 400) should have their age adjusted. These were offset by small increases to other ages.


Adjustments to the 1996 Base Population
Only two adjustments were made in deriving the original base resident population after the 1996 Census; an adjustment for net census undercount (originally 45,000), and an adjustment for residents temporarily overseas at the time of the 1996 Census (39,700). For more information see Adjustment of Post-censal Population Estimates for Census Undercount and New Zealand Residents Temporarily Overseas at the 1996 Census. Following the 2001 Post-enumeration Survey (PES), the 1996 PES results were revised using the same methodology as that used for the 2001 PES, thus aligning the methodologies of adjustment at 1996 and 2001 (refer to the results in A Report on the Post-enumeration Survey (2001)). Consequently the adjustment for net census undercount was revised from 45,000 to 60,000.


Back to top

National Population Estimates
The estimated resident population is produced using the component methodology and 'date of event' method. Date of event method means that demographic events (births, deaths and migration) used in producing population estimates are classified by their date of occurrence. For demographic accounting, each demographic event is classified by the year and quarter of birth of the person to whom the demographic event applies.

The estimated resident population uses census data (census usually resident population count) as a starting point. Consequently estimates of the resident population based on one census cover one intercensal period (and slightly beyond) until more recent census data becomes available. When the more recent census data becomes available, this is incorporated into the new base for the next 5-year period, etc. After each Census of Population and Dwellings, Statistics New Zealand routinely revises all its national and subnational population estimates, using data from the latest two censuses.

Updating Estimated Resident Populations:
National population estimates are produced quarterly. To produce population estimates for the current quarter (ie the quarter under consideration, or the reference quarter), previous estimates (or the base population) are updated using births, deaths and external migration data for the current quarter. Birth and death registrations are captured by the Department of Internal Affairs. Statistics New Zealand receives this vitals data each month. Live births to women resident in New Zealand are classified by the sex of the child and by the year and quarter of birth. Deaths of New Zealand residents are classified by the sex, and by the year and quarter of death of the deceased. To meet production timetables the numbers of births and deaths that occurred in the most recent quarter are estimated. Estimation is necessary as not all births and deaths that occur in a quarter are registered by the end of the quarter. Migration data is not estimated, as all required data is available from Statistics New Zealand's external migration databases.

Provisional and final estimates are produced for the resident population each quarter. Provisional estimates are produced for the Hot Off The Press and Media releases. This is the first public release of population estimates for a quarter and they are released within six weeks following the end of the reference quarter. The provisional estimation for quarterly deaths has a 95 percent confidence interval of ±250 (or about 3.5 percent, based on an average of 7,000 deaths per quarter). The final estimates are derived when an extra quarter of detailed vitals data is available, about 5 months after the end of the reference quarter.

The estimate of deaths that occurred in the reference quarter is taken as those deaths that were registered either in the reference quarter or in the subsequent quarter, and that occurred in the reference quarter. This final estimation for deaths is very accurate as very few deaths are registered outside one quarter of their occurrence (about 1 death in a 1,000 is not). The final estimation for births that occurred in a quarter is less accurate, as births are not registered as closely to their occurrence as are deaths (around 5 percent of births are not registered either in the quarter of occurrence or in the subsequent quarter). The final estimation for quarterly births has a 95 percent confidence interval of ±500 (or about 3.5 percent, based on an average of 14,000 births per quarter).

The procedures used to estimate quarterly births and deaths take account of the seasonal variation in these events and the time lag between these events and their registration. The second estimation of deaths for a quarter gives some insight into how this can be done, where the number of deaths that occurred in the current quarter is estimated as the product of registered deaths in the current quarter times the ratio of actual deaths to death registrations for the same quarter of the previous year. Other estimations for births and deaths use similar but more complex procedures.

No adjustments are made for category jumping between the short-term and permanent and long-term categories in external migration. The extent of category jumping, however, is monitored. Category jumping estimates are imprecise and don't become available until more than a year after the end of any period of interest. Category jumping between the 1991 and 1996 Censuses was estimated as a net loss of about 9,000 residents from the resident population. (No adjustments were made for this loss when resident estimates were taken back to 1991, as an age by sex breakdown of the loss could not be accurately determined.) Category jumping in external migration added an estimated net 18,000 to the resident population between the 1996 and 2001 Censuses. In the intercensal period between the 2001 and 2006 Censuses, category jumping added an estimated 47,000 to the resident population.

A similar methodology was used to derive resident estimates going back from the 1996 Census to 1991 as was used going forward from 1996. Resident estimates prior to 30 June 1996 used vital registrations data up to 30 June 1997. No estimation of births and deaths (by date of occurrence) was necessary as births and deaths that occurred prior to the 1996 Census were almost totally registered by 30 June 1997.

Estimates of the resident population of New Zealand (based on the 1996 Census) are available quarterly, on an 'at' basis from 31 March 1991 onwards. They are also available quarterly on a 'mean year ended' basis from 31 December 1991 onwards. Estimates are available for the resident population classified by sex and single year of age (0-89 years and 90+) or by 5-year age group (0-4, 5-9, ..., 85-89, 90+).
Back to top

Subnational Population Estimates

Population estimates for regions, territorial authority areas and urban areas

The estimated resident populations for these subnational areas are produced annually at 30 June. They are constrained to agree with the corresponding national estimated resident population. The estimation of subnational populations uses the same component methodology as for estimates of the national population, but with the addition of an extra component – internal migration between the subnational areas of New Zealand. External migration for each territorial authority (ie movements between the TA and outside New Zealand) is estimated from external migration data. Internal migration is estimated since there is no direct source of annual internal migration data.

The formula used for estimating the resident population in each territorial authority is:
Population (current year) = population (previous year) + births - deaths + net external migration + net internal migration.

All components of the formula (on the right hand side of the equation) are estimates of the number of events that occurred during the current year ended 30 June.

Estimates of births and deaths occurrences for each territorial authority in the current year are derived from past experiences of ratios of registrations to occurrences.

External migration data on travellers arriving in or departing from New Zealand on a permanent and long-term basis are classified by territorial authority. All overseas travellers are required to fill in arrival/departure cards when entering or departing the country. Due to a significant non-response to the territorial authority question on the arrival/departure cards, external migration for territorial authorities are estimated using a simple proration methodology, based on country of birth classification, to incorporate the non-response component.

A measure of internal migration (movement by residents within the country) at subnational level is available only for the 5-year intercensal period as a whole. During the intercensal years, internal migration is estimated using a range of symptomatic data sources. The section below 'The use of administrative data in subnational population estimates' describes the symptomatic data that forms an input to estimating the internal migration component. The symptomatic data is supplemented by information from territorial authorities. A letter from Statistics New Zealand is forwarded to each territorial authority requesting any information that is likely to affect population growth in their area.

Subnational estimates are produced annually, 'at 30 June', for regions, territorial authorities and urban areas. Estimates of the total population of these subnational areas become available in mid-October of the same year. Estimates of the population of these subnational areas by five-year age group (0-4, 5-9,10-14, ..., 80-84, 85+) and sex generally become available in December each year, once area unit population estimates have been produced.


Population estimates for area units

Area units are non-administrative areas intermediate in size between meshblocks and territorial authorities.

The estimated resident populations of area units are produced annually at 30 June, and generally become available in December each year. They are constrained to agree with the estimated resident population of associated territorial authority areas. In 2007, as part of a government initiative to make more information freely available, Statistics New Zealand began to release area unit population estimates on its website (via Table Builder).

Statistics NZ produces area unit population estimates using a component methodology, where the components of population change (births, deaths and migration) are used to update a base population. The formula used for estimating the resident population in each area unit is:
Population (current year) = population (previous year) + births - deaths + net migration (external and internal migration combined).

Occurrences of births and deaths are estimated using information from birth and death registrations (available at the area unit level). The estimation procedure also takes into account the ratio of occurrences to registrations from previous periods.

In contrast, information relating to net migration is not readily available at the area unit level. For this reason the net migration component requires a special estimation methodology.

Initially, the net migration component for each area unit is estimated using a model which takes into account the relationship between net migration and building consents during the last intercensal period (2002–2006), building consents during the period of interest, and occupancy rates at the time of the 2006 Census. The initial estimate of net migration for each area unit is then adjusted so as to meet the net migration constraint for associated subnational areas, and to incorporate any salient information received by Statistics NZ. For example, while producing area unit population estimates at 30 June 2007, adjustments were made to the net migration component of several area units, so as to reflect the impact of new prisons and retirement homes that had been constructed in those area units.


Subnational population estimates by five-year age group and sex

Subnational population estimates by five-year age group (0-4, 5-9, 10-14, ...., 80-84, 85+) and sex generally become available in December each year once area unit population estimates have been produced.

In 2008, Statistics New Zealand adopted a new method for producing subnational population estimates by age group and sex, which involved a new method for estimating subnational net migration.

At the territorial authority level, estimates of birth occurrences (by sex) and death occurrences (by age and sex) are derived from birth and death registrations. Estimates of net migration for each territorial authority (by age and sex) are initially derived using the latest territorial authority population projections, and national age-sex net migration constraints for the June year of interest. Where Statistics New Zealand holds information about a significant local development at the territorial authority level (eg, a newly completed prison), the initial age-sex net migration estimates are adjusted so as to take such developments into account.

At the area unit level, estimates of birth occurrences (by sex) and death occurrences (by age and sex) are derived from birth and death registrations. Estimates of net migration for each area unit (by age and sex) are initially derived by apportioning net migration estimates for each territorial authority across associated area units, while taking into account the age-sex distribution and net migration constraint for each area unit. Where Statistics New Zealand holds information about a significant local development at the area unit level (eg, a newly completed prison, student hostel, apartment complex or retirement village), the initial age-sex net migration estimates are adjusted so as to take such developments into account.

Statistics New Zealand continues to explore new methods that could improve the quality of subnational population estimates. At present, work is being undertaken to explore alternative methods for estimating net migration at the area unit level.



The use of administrative data in subnational population estimates

Background
This section aims to provide users of subnational population estimates with some information on the administrative data sources currently used, and an evaluation of these data sources against specified criteria. It is important to note that all data referred to here is used in aggregated form only and the confidentiality of individual data is protected.

For the production of subnational estimates, population change is obtained by estimating natural increase (excess of births over deaths occurrences) and net migration. The migration component is computed at the territorial authority level and consists of estimated net external migration and estimated net internal migration. Net external migration is derived from responses to the New Zealand address question on the arrival/departure cards of permanent and long-term migrants and includes an adjustment for non-response. Estimation of the internal migration component takes account of internal migration during the most recent intercensal period, and is also guided by a range of symptomatic indicators or administrative data sources (eg building consents) and information provided by local authorities.

For the production of subnational population estimates prior to 2004, estimation of internal migration was limited mainly to two administrative data sources being indicators of population change, namely residential building permits and employment data. However, due to increasing levels of mobility of residents between territorial authorities, it has increasingly become important to use a wider range of administrative data sources, in order to maintain the quality of subnational population estimates.

Evaluation criteria
The following criteria were used to assess the usefulness of administrative data sources as symptomatic indicators of internal migration:

• population and coverage
• data content
• geographic level
• timeliness
• historic availability
• consistency
• electronic capture

These criteria were applied to administrative data sources already available to Statistics NZ. As other administrative data becomes available, it will also be evaluated against the same criteria. All of the data sources listed below are used when estimating internal migration.

The assessment of the data sources is followed by some brief comments on limitations of the data as indicators of internal migration.

Administrative data sources used in subnational population estimates from 30 June 2004 onwards
1. Residential Building Consents Data

(Source: Statistics NZ)

Population and coverage	Counts of new residential building consents issued during the year ended June.
Data content	Counts of new residential building consents by territorial authority and area unit.
Geographic level	Residential building consents are available for TAs, area units and meshblocks.
Timeliness	Building consents database is updated on a monthly basis.
Historic availability	Time series for June years are available from 1990 onwards.
Consistency	No data definitional changes are expected.
Electronic capture	Part of Statistics NZ building consents database.

Building consents data covers new residential building well but covers demolitions of existing building poorly. Local building consents may reflect replacement of housing stock rather than additional stock. The classification of residential building consents does not indicate whether the building consent is intended for an "all year" residence or a holiday home that only will be occupied for parts of the year.

2. Electoral Enrolment Data

(Source: Electoral Enrolment Centre, Wellington)

Population and coverage	All New Zealand residents who are eligible to vote. Hence the electoral enrolment data includes NZ residents aged 18 years and over who are registered on the electoral roll. The electoral roll does not include "new migrants" (ie those immigrants who have arrived in NZ in the past year and are not citizens of New Zealand, Cook, Niue or Tokelau).
Data content	Summarised counts by age group (18-19, 20-24, 25-29,...,70-74, 75+ years) from the active electoral roll for general electoral districts and territorial authorities at 30 June.
Geographic level	Electoral districts and territorial authorities.
Timeliness	The electoral enrolment database is updated monthly. Datasets of electoral counts, at 1 July, are supplied at the end of July by the Electoral Enrolment Centre.
Historic availability	The Electoral Enrolment Centre does not maintain a historic time series of electoral counts for electoral districts and territorial authorities. Statistics NZ has commenced an annual time series of electoral counts from 30 June 2003 onwards. A time series of enrolment rates by age group has been established from 30 June 2003.
Consistency	No data definitional changes are expected.
Electronic capture	Electoral counts are supplied as an Excel file.

Electoral enrolment data will be affected by registration campaigns, by-elections and electoral cycles. Enrolment rates vary by TA and also by the age group of those enrolling.

3. IRD Data

(Source: LEED (Linked Employer-Employee Data project), Statistics NZ)

LEED brings together, or links, datasets held by the Inland Revenue Department (IRD) as administrative tax records with data held by Statistics NZ. The IRD administrative data is obtained from the IRD's Data Warehouse, and is derived from Employer Monthly Schedule (EMS) payroll returns.

Population and coverage	The EMS captures all people engaged in paid employment in the month covered by the return. It also includes people engaged in self-employment. People receiving social security benefits are required to hand in an annual income tax return and are therefore included in the IRD database. As a result, the IRD database includes some children. The usual resident population concept that forms the basis for the Statistics NZ population estimates series is not included as an indicator in the IRD database.
Data content	The dataset supplied to the Population Statistics Unit contains counts of individuals who have been registered with the IRD. The snapshot at 30 June includes the following variables: date of birth, sex, TA address of living at 30 June in the current year and TA address of living at 30 June of the previous year. The counts are limited to those with client status code recorded as 'Active' in both years.
Geographic level	The dataset forms a subset of the IRD data where individuals' addresses are possible to geocode to a valid TA code.
Timeliness	A dataset of counts at 30 June is supplied to Population Statistics Unit in August from 2006 onwards.
Historic availability	A time series has been established of counts by age, sex and TA on an annual basis from 30 June 2001 onwards. This enables a time series of counts of annual net movements by age and sex between TAs from 30 June 2002 onwards.
Consistency	No data definitional changes are expected.
Electronic capture	The dataset is supplied in Excel format.

Continued improvements in the maintenance of the IRD administrative system to reflect the population of interest are necessary to make the IRD data series a valuable input to the estimation of internal migration.

4. School Rolls Data

(Source: Data Management Unit, Ministry of Education, Wellington)

Population and coverage	The Data Management Unit of the Ministry of Education collects snapshot data of all New Zealand schools at 1 July each year. Numbers of regular students by age, sex, year of schooling, ethnicity, type of school, type of student and nature of attendance (full-time or part-time) are collected.
Data content	Population Statistics Unit, Statistics NZ, is supplied with a dataset of school roll counts by year of schooling, sex and TA.
Geographic level	TA
Timeliness	A dataset of the most recent update of school rolls is supplied two-three months following the requirement of schools to provide roll information at 1 July.
Historic availability	A time series of school roll counts by year of schooling, sex and by territorial authority from 30 June 1996 onwards is available.
Consistency	No data definitional changes are expected.
Electronic capture	School rolls are supplied as an Excel file.

Changes in the school attended, school zoning and other factors limit the usefulness of school rolls as indicators of population change. The relationships between local rolls and local residential populations are dynamic.

Other administrative data sources
Time series, at 30 June of each year, are maintained of numbers of university students by university location; of offenders by prison location; and of armed forces by living location.

Back to top

Mäori Ethnic Group Population Estimates
The definition of Mäori used for population estimates has changed over the years.

Before 1986
Mäori population estimates were based on people of half or more Mäori blood.

1986-1991
The term 'sole Mäori' was used between the 1986 Census and the 1991 Census to describe those census respondents who stated New Zealand Mäori as their sole ethnic group. These estimates were of the de facto population. Mäori births and deaths (defined as being of half or more Mäori blood) were used to update the base population at census. Mäori net migration was assumed to follow the same pattern as applied over the period 1982-1986 (ethnicity information was not collected on external migration cards after 1986).

1991 Onwards
Mäori population estimates refer to people belonging to the Mäori ethnic group. These estimates are based on the census usually resident population counts of the Mäori ethnic group at the 2001 and 2006 censuses, updated for:

1. net census undercount;
2. the number of residents temporarily overseas on census night;
3. births, deaths and net migration (excess of arrivals over departures) between census night and 30 June;
4. reconciliation with demographic estimates at the youngest ages; and
5. non-response to the census ethnicity question.

Starting with the 2001 base, estimates from 30 June 1991–2001 were produced by updating for natural increase (excess of births over deaths) and an assumed annual population loss of 800 over the period 1991–2001 due to the combined effect of migration (there is no external migration data available for the Mäori ethnic group) and inter-ethnic mobility (people changing ethnic identification). The net annual outflows assumed over the period 1991–2001 followed the pattern of arrival by Mäori in Australia, as measured by the 2001 Australian Census. The population estimates of the Mäori ethnic group for 1996 have not been reconciled with the Mäori population enumerated by the 1996 Census because of different ethnic questions asked in the 1996 and 2001 censuses.

Population estimates for the Mäori ethnic group from 30 June 2001–2006 have been revised using results from the 2001 and 2006 censuses. Over this period, an annual population loss of 3,150 due to the combined effect of migration inter-ethnic mobility, was assumed. The revised estimates for 2001–2006 supersede previously published estimates.

The estimated resident populations from 2006 were obtained by updating the base population at 30 June 2006 for births, deaths and the combined effect of migration and inter-ethnic mobility during the ensuing period. Mäori ethnic group estimates from 30 June 2006 assume net migration flows as a proportion (20 percent) of the net permanent and long-term migration of New Zealand citizens to Australia, with adjustments for inter-ethnic mobility and Mäori migrating to countries other than Australia. The assumed net migration flows are based on data obtained from censuses in other countries including Australia, the United Kingdom and the United States of America.

The estimated resident population of the Mäori ethnic group is derived quarterly, starting from 1991. Estimates are available on the Statistics New Zealand website at 30 June of each year and for mean years ended December, classified by sex and single year of age (0-89 years and 90+), 5-year age group (0-4, 5-9, ..., 85-89, 90+), and selected age groups (eg 0-17, 15-64, 18+).

Back to top

Historical De facto Population Estimates

National De facto Population Estimates

Estimates of the de facto population of New Zealand are available for the period 1936-1995.

Type	Frequency	Period
At	31 December	1936-1995
Mean year ended	31December	1937-1995
At	31 March, 30 June, 30 Sep, 31 Dec	1976-1995
Mean year ended	31 March, 30 June, 30 Sep, 31 Dec	1977-1995

The estimates are available for the total, Mäori and non-Mäori populations by sex and single year of age (0-89 years and 90+) or by 5-year age group (0-4, 5-9, ..., 85-89, 90+).

Subnational De facto Population Estimates

A reorganisation of local government during the 1988–1992 period created 16 regional council areas and 74 territorial authority areas. Annual estimates of the de facto population of regional council areas, territorial authority areas, and main and secondary urban areas, are available for the period 1986–1995.

Area	Type	Frequency	Period
Regional council	At	31 March	1986-1995
Territorial authority	At	31 March	1986-1995
Urban (main and secondary)	At	31 March	1986-1995

No de facto population estimates after 1995 have been produced. Following the 1996 Census, national and subnational population estimates and projections have only been produced for the resident population concept. The base population for de facto population estimates is the census night population count shown by the most recent Census of Population and Dwellings. The base population was updated using registered births, registered deaths and total external migration. De facto population estimates do not include an adjustment of the base population for census undercount, as the first Post-enumeration Survey measuring census undercount was conducted after the 1996 Census.
Back to top


Timetable
National population estimates
Provisional estimates of the resident population are available within 6 weeks after the end of the reference quarter. The provisional estimates are released via the Hot Off The Press and Media releases. Final estimates are available about 5 months after the end of the reference quarter.

Subnational population estimates
Subnational estimates are produced annually, 'at 30 June', for regions, territorial authorities and urban areas. Estimates of the total population of these subnational areas become available in mid-October of the same year. Estimates of the population of these subnational areas and of other user-defined areas, by five-year age group (0-4, 5-9,10-14, ..., 80-84, 85+) and sex, generally become available in December each year once area unit population estimates have been produced.



Back to top

Rounding of Population Estimates
Estimates are rounded to reflect the accuracy of the processes by which they are derived. Owing to rounding, individual figures may not sum to stated totals.

At the national level, estimates of the total resident population are rounded to the nearest 100; and estimates for subgroups defined by age are rounded to the nearest 10.

Subnational estimates are rounded according to the following rules:
(i) Populations of less than 10,000 are rounded to the nearest 10;
(ii) Populations of 10,000 or greater, but less than 20,000 are rounded to the nearest 50;
(iii) Populations of 20,000 or greater are rounded to the nearest 100.
Back to top

---

Output Variables

Output variables are pieces of individual information that can be extracted from the survey/output data. Often output variables can be cross tabulated with other output variables, for example sales by industry classification. The list below contains all current and past output variables which have been released for this survey/output.

Variable Name	Description
National Estimated Resident Population (quarterly, at or mean year ended)	Population by Single Year of Age (0-89, and 90+) and Sex Population by 5-Year Age Group (0-4, 5-9, 10-14, ..., 85-89, 90+) and Sex Population Change Population Age Distribution (percent) Median Age (half of the population is older than this age) Natural Increase (births minus deaths) Net permanent and long-term migration
Subnational Estimated Resident Population (annually, at 30 June)	Population by territorial authority, regional council, urban area and area unit by 5-Year Age Group (0-4, 5-9, 10-14, ..., 80-84, 85+) and Sex
National Estimated Resident Population of Mäori Ethnic Group (annually, at 31 December)	Population by Single Year of Age (0-89, and 90+) and Sex Population by 5-Year Age Group (0-4, 5-9, 10-14, ..., 85-89, 90+) and Sex Population Change
	Population Age Distribution (percent)
	Median Age (half of the population is older than this age)

Changes in Output Variables over time

---

Guide to Interpreting Data

Summary of Changes to Survey/Output ..

1988	Local Government Amendment Act No 3 provides for the constitution of 14 Regional councils. The regional council areas cover every territorial authority area in New Zealand with the exception of Chatham Islands County. These replaced 22 Local Government Regions.
1 Nov 1989	Local government reorganisation creates 74 territorial authority areas. These replaced 213 Local Authorities.
1991	Census labour force question changed so that the census labour force will relate more closely to the definition used in the Household Labour Force Survey (HLFS). Classification of urban areas revised into a three-part classification consisting of main, secondary and minor urban areas differentiated by population size.
1 July 1992	The number of regions increased to 16 following boundary reorganisation in the northern South Island.
Sept 1995	Changes to the ethnic question on the birth and death registration forms means that the ethnic classification is no longer according to degree of blood, but rather, according to self-identification. This also means that a greater range of ethnic groups can be identified and multi-ethnic responses can be generated. See Births and Deaths Profiles.
1996	First post-enumeration survey held in New Zealand following the 1996 Census of Population and Dwellings.
1996	The population concept used for population estimates and projections changed from de facto to resident. The de facto population includes all persons in New Zealand at a given time. It includes overseas visitors, but excludes residents temporarily overseas and net census undercount. The estimated resident population includes the estimated net census undercount and the estimated number of residents temporarily absent overseas, but does not include overseas visitors. Resident population estimates do not show the quarterly fluctuations that de facto estimates are subject to because of seasonal changes in the numbers of overseas visitors temporarily in New Zealand, and changes in the numbers of New Zealand residents temporarily out of New Zealand. Differences between mean estimates show little or no seasonal variation. Over the period 1991-95 revised mean year ended estimates of the de facto population were about 2% smaller than the corresponding estimates of the resident population. (The revised de facto estimates used information from the 1991 and 1996 Censuses while the resident estimates only used information from the 1996 Census.) The reference date for subnational population estimates and projections shifted from 31 March to 30 June.
1999	A revision of the age by sex estimates of the resident population was carried out in September 1999. The revision affected the age by sex estimates from 1991 onwards. The estimated totals for males, females and the total population remain unchanged. The revision was necessary because of data quality issues relating to the enumerated population at younger ages. Censuses in all countries tend to under-enumerate the population at all ages, and, in particular, the number of young children. Under-enumeration in the 1996 Census was confirmed by the first Post-enumeration Survey (PES) undertaken in New Zealand, which was carried out shortly after the 1996 Census. When post-censal estimates of the resident population were first produced following the 1996 Census, statistical adjustments were included for the estimated under-enumeration in the 1996 Census and for the estimated number of NZ residents temporarily overseas at census. The size of the PES limited the precision of the statistical adjustments for census undercount that were made. The precision of the adjustment for census undercount is particularly important for ages 0-5 years as for these ages estimates based only on births, deaths and external migration can provide alternative estimates to estimates built on census data. A demographic approach (ie using births, deaths and net permanent and long-term migration) has been used to provide a set of estimates for younger ages.
1999	In October 1999, estimates for 1998 and 1999 already published were revised to correct for a shortfall in birth registrations during 1998 (as indicated by birth notifications received from hospitals).
1 Jan 2001	Chatham Islands District became Chatham Islands Territory.
2001	Second Post-enumeration Survey carried out in New Zealand following the 2001 Census of Population and Dwellings.
2001	The ethnic question in the 2001 Census of Population and Dwellings changed from that asked in 1996. (The 2001 question was very similar to that asked in 1991.) As a result, the ethnic data from the 2001 Census is not comparable with that from the 1996 Census. This change means that 1996-base population estimates for the Mäori ethnic group are not comparable with the soon to be released 2001-base population estimates. Consequently, estimates for the Mäori ethnic group will be revised back to 1991, using results from the 2001 Census.
2002	The base populations for the Mäori, Pacific, Asian, and European ethnic groups, at 30 June 2001 incorporate an adjustment for those people not specifying their ethnicity. This is the first time the base estimated resident populations have included such an adjustment. A revision of the estimated resident populations for the period 1991-2001 was carried out in September 2002. The revision was necessary for two reasons. Firstly, the 2001 Census and the 2001 Post-enumeration Survey (PES) allowed a new base population for population estimates to be defined at 30 June 2001. Secondly, a revision of population estimates based on the 1996 Census was necessary, as analysis of the 2001 PES results led to revised estimates of undercoverage in the 1996 Census, from 1.2% to 1.6% in total. About 16,000 residents were added to the 1996-base population because of the revised undercoverage estimates. Population estimates for the period 1991-1996 use the revised 1996-base population.
2004	On 1 March Tauranga District became Tauranga City.
2006	March, Banks Peninsula District amalgamated with Christchurch City.
2007	A revision of the estimated resident populations for the period 2002–2006 was carried out in August 2007. The revision was necessary because the 2006 Census and the 2006 Post-enumeration Survey (PES) allowed a new base population for population estimates to be defined at 30 June 2006.
2008	In December, Statistics New Zealand adopted a new method for producing subnational population estimates by age group and sex. This new method was used to revise subnational population estimates, by age group and sex, at 30 June 2007. Estimates of the total population of each subnational area, at 30 June 2007, were not affected by this revision.

Usage and Limitations of the Data ..Use of population estimates data
Estimated resident populations (total, Mäori ethnic group, or subgroups defined by age and sex) aid the decisions of government, business and community organisations. They are also widely used in analysing changes in the demographic, economic and social structure of New Zealand, eg in the calculation of mortality rates, participation rates in education, per capita consumption of goods within New Zealand, etc.

For the calculation of rates when the numerators or "top lines" relate to data covering calendar years, mean year ended 31 December population estimates are appropriate as denominators or "bottom lines". For analysing data collected over a short time period (eg a month), estimates of the population at the end of the closest quarter are appropriate as the denominator or "bottom line". The estimated resident population for the total and Mäori ethnic group populations are available for all quarters, ie at 31 March, 30 June, 30 September, and 31 December, from March 1991 onwards.

Subnational population estimates are used in many administrative, statistical and research applications. They are used by Government, regional planning authorities, local bodies, private organisations and individuals for determining future housing and other consumer needs, community amenities and social services at the subnational level. Subnational estimates of population also provide information on the supply of labour for employment and the growth of markets. Subnational estimates are available at 30 June; these estimates can be used as a proxy for mean year ended 31 December population estimates.

Limitations of population estimates data

Limitations of national estimates
The accuracy of population estimates depends on the accuracy of the population base and on the accuracy of the components (births, deaths and migration) used to update population estimates. Some users of population estimates are more interested in accurately monitoring population change rather than in monitoring the overall accuracy of population estimates. Here the accuracy of the base population is discussed separately from issues of accuracy in the components of population change.

The base estimated resident population is derived by adjusting the census population for net census undercount as enumerated by the Post-enumeration Survey (PES) and estimates of residents temporarily overseas at census date (based on external migration and census data). The uncertainty in the net undercount estimated from the 1996 PES had a survey error (twice the standard error) of 7,000. Thus, at 95% confidence, because of uncertainty in the PES alone, the resident population of New Zealand at 5/3/96 could be up to 7,000 higher or lower than the estimated resident population at 5/3/96. The uncertainty in the net undercount estimated from the 2001 PES had a survey error of 10,000.

Residents temporarily overseas (RTOs) at census are estimated using the length of absence data from the migration arrival and departure cards. At the 1996 Census an estimated 40,000 (survey error 1,100) New Zealanders were temporarily overseas ie just over 1% of the NZ resident population. Migration data is the only comprehensive source of data on RTOs; only 16,700 or 42 percent of RTOs were captured by the 1996 Census as absentees overseas (note that whole households may be away overseas at census). At the 2001 Census an estimated 54,500 New Zealanders were temporarily overseas. This estimate was based on matching migration arrival and departure records; it is not possible to assign a measure of accuracy to these estimates.

After the most recent census, national population estimates are updated quarterly using registrations of births and deaths, and permanent and long-term migration, captured by the migration cards. It is necessary to estimate the number of births and deaths that occurred during the most recent quarter as not all of the births and deaths that occurred in the current quarter will be registered by the end of the quarter.

Births and deaths that occurred in the most recent quarter are estimated from registered births and registered deaths, with estimation errors (twice the standard error) of 520 for provisional estimated births and 240 for estimated provisional deaths. Estimation errors for final estimated births and final estimated deaths are smaller, about 200 for births and 0 for deaths (but with a slight negative bias of about 8 deaths missed per quarter). With births and deaths running about 55,000 and 27,000 respectively per year the final annual estimation error for births amounts to about 400 or 0.7%, while for deaths the annual negative bias of 32 is only about 0.1% of annual deaths (one part in a thousand).

Registrations of births data and deaths data in New Zealand can be regarded as complete. Virtually no births or deaths escape the registration process for any length of time. A highly conservative estimate might be that records are 99.5 percent complete. This would equate to approximately 280 births and 140 deaths being missed on average each year. However we have no evidence of any such under-reporting.

Permanent and long-term migration is the least reliable component of population change. In the preparation of international travel and migration statistics, the passenger type (short-term overseas visitor, short-term New Zealand resident traveller or permanent and long-term migrant) is primarily determined by the passenger's response, on the arrival or departure card, to the questions on where they live and their length of stay or absence. If the person's intention changes later during the trip, this passenger type may no longer be correct. Statistics New Zealand continues to monitor the prevalence of this 'category jumping' in international travel and migration statistics. The monitoring carried out to date indicates that, in most years, the net effect of category jumping is negligible.

Population estimates for the period 1991-1996 are based on the 1996 Census; note that as no Post-enumeration Survey was held after the 1991 Census it is not possible to derive a base resident population for 1991. Revised estimates of the resident population for 1996-2001 incorporate the intercensal discrepancy between 1996-base population estimates taken forward to 2001, and the new base population at 2001. Revised estimates of the resident population for 2002–2006 incorporate the intercensal discrepancy between 2001-base population estimates taken forward to 2006, and the new base population at 2006. Intercensal discrepancy is the difference between two population estimates that arises from errors in the census-base estimates and/or components of change. The intercensal discrepancy is introduced smoothly into the revised population estimates, with the size of the adjustment for the intercensal discrepancy depending on the relative closeness of the date of the estimate to the beginning and end of the intercensal period.


Limitations of subnational estimates
Statistics New Zealand produces subnational population estimates using a component methodology, where the components of population change (births, deaths and migration) are used to update a base population.

Subnational estimates are constrained to the national estimates and, therefore, the quality of subnational estimates is affected by the quality of the components of national estimates. There are also some additional uncertainties that arise when estimating the components of population change at the subnational level, particularly while estimating net migration (external and internal migration combined) for subnational areas. Due to these factors, as the time period between the reference date of the base population and the date of the estimate increases, the uncertainty associated with subnational population estimates also increases.

It is also important to note that population estimates for small geographic areas (e.g., area units) are more uncertain than estimates released at broader geographic levels. This added uncertainty arises because:

• information on external and internal migration is not readily available at the area unit level
• localised developments (eg greenfield subdivisions) can have a significant impact on the net migration patterns and population structure of area units, but this impact can be difficult to discern.

Due to estimation of the internal migration component, regional differentials in census undercount and regional variations in the propensity of residents to be temporarily overseas at census, it is difficult to quantify the accuracy of subnational estimates. However, some insight can be gained by considering intercensal discrepancy.

The intercensal discrepancy for the last intercensal period is the difference between 2006-base and 2001-base population estimates at 30 June 2006. At the national level there was an intercensal discrepancy of 45,100, with the 2006-base estimate of New Zealand's resident population (4,184,600) being higher than the 2001-base estimate (4,139,500). At the regional level, intercensal discrepancy (in absolute terms) did not exceed 5 percent for any of New Zealand's 16 regions. At the territorial authority level, intercensal discrepancy (in absolute terms) exceeded 5 percent for 14 out of the 73 territorial authority areas.

Statistics NZ continues to search for ways to improve its subnational estimation techniques. At present, work is being undertaken to explore alternative methods for estimating net migration at the area unit level..

Related Data Sources ..

Sampling Errors ..

Non-sampling errors

Caveats on Release ..Mäori Ethnic Group Estimates, 1991-2001 (2001-base)
Limitations are inherent in the data used to estimate the resident population of the Mäori ethnic group for the period 1991-2001. All population estimates for the Mäori ethnic group are supplied only as a guide for research and other analytical purposes. Mäori ethnic group population estimates based on the 2001 Census of Population and Dwellings supersede all other Mäori ethnic group population estimates derived or published previously.

Population estimates are traditionally derived by the component method, ie the components of population change, viz. births, deaths and migration, are used to update the base population to the current date. The quality of the components is critical to producing accurate estimates. The greater the time between the date of the population base and the date of the estimates produced, the more critical the quality of the components becomes, as the effect of data imperfections tend to be cumulative. Ideally, each component of change would be based on current data. This is not possible in deriving estimates for the Mäori population as there is no current data on external migration by ethnicity.

Census Base

The base population for the latest set of Mäori population estimates is the census usually resident population count of the Mäori ethnic group, as enumerated by the 2001 Census of Population and Dwellings, adjusted for: non response to the ethnicity question; net census undercount (as measured by the 2001 Post-enumeration Survey); and residents temporarily overseas. Working from this base, estimates back to 1991, and forward from 2001 were produced. The population estimates for Mäori for 1991 are in broad agreement with the Mäori population enumerated by the 1991 Census, taking account of changes in the census question on ethnicity, unknown under-enumeration of Mäori in the 1991 Census (no Post-enumeration Survey was conducted in 1991), and changes in ethnic affiliation over the period 1991-2001.

Because of changes to the question used to collect data on ethnicity between the 1996 and 2001 Censuses, no attempt has been made to reconcile the Mäori population estimates based on the 2001 Census with results from the 1996 Census.

Migration

Information on ethnicity is not collected on the arrival/departure cards filled in by travellers entering and leaving New Zealand. Thus there is no data available on external migration for the Mäori population – possibly the most volatile component of any population estimates model.

Mäori ethnic group estimates for 1991-2001 were derived assuming a total net outflow of 8,000 over the period (an average annual net loss of 800). The timing of the net annual losses was determined from data on Mäori in Australia, recorded by the 2001 Australian Census. Mäori ethnic group estimates after 2001 were derived assuming an annual net loss of 800. The net outflow, coupled with other known components of population change, largely explains the structural changes which occurred in the Mäori population between 1991 and 2001. Any inaccuracy in the migration component, or in its assumed age profile, may accumulate during the estimation period. Estimates of the Mäori ethnic group population for young adult ages, where migration is common, may be particularly affected.

Births and Deaths

From 1 September 1995 new ethnicity questions were introduced to the birth and death registration forms. The new questions are based on the concept of self-identification. Mäori ethnic group births for the period December quarter 1995 onwards are based on births data compiled using the new question. For the period from the December quarter 1990 to the September quarter 1995, Mäori ethnic group births were initially taken to be those where there was any indication of Mäori ethnicity (based on ethnicity of the parents, and other indicators of ethnicity, as no data was collected on the ethnicity of the newborn; and included proration for non-response). Subsequently these numbers were assessed relative to the estimates by age for the Mäori ethnic group base population.

Mäori ethnic group deaths for the period December quarter 1995 onwards are based on deaths data compiled using the new question, adjusted for date of event methodology. Mäori ethnic group deaths for the period 1991-1995 are derived from death rates that were used for the 1996-based Mäori population projections. It was not considered appropriate to simply prorate deaths up, as the assignation "Mäori" on the earlier death registration forms is likely to depend on age at death. Another factor considered was the scale of the necessary adjustments. (The largest number of "Mäori" deaths over this period ran at about two-thirds of the current Mäori ethnic group deaths).

Inter-ethnic Mobility

Population estimates for the Mäori ethnic group do not include any adjustments for inter-ethnic mobility. Net inter-ethnic mobility is believed to have been negative for some Mäori age groups over the ten-year period 1991-2001. Despite the fact that no adjustment has been made for inter-ethnic mobility, the assumed annual net migration loss (800) may have been different had inter-ethnic mobility gains and losses been known.

The comparability of ethnic response in different data collections - matching 2001 birth and death registrations to Census 2001

The ethnic question used in the 1996 Census was used until mid 2005 to collect ethnic data from birth and death registrations. The ethnic question used in the 2001 Census differed from the 1996 Census ethnic question. Results from the 2001 Census indicated that the change in ethnic question in 2001 may have led to a response shift. Therefore it was important to assess how well the 2001 Census data on ethnicity agreed with corresponding vitals registration data. The key objective of the exercise was to compare ethnic data collected from vitals registration with ethnic data collected from the 2001 Census. The accuracy of ethnic demographic indices and population estimates is dependent on the consistency of ethnic details collected from the different data sources.

Birth and death registrations were matched to data from the 2001 Census. Births that occurred between 1/1/2001 and 6/3/2001 and were registered by 30/6/2002 were matched to the 2001 Census. Birth parents for births that occurred between 1/1/2001 and 30/6/2001, and were registered by 30/6/2002, were also matched to the 2001 Census. Deaths occurring on or after 7/3/2001 which had been registered by 30/6/2002 were matched. For both births and deaths, registration records were matched on sex, date of birth (day of month, month and year), and usually resident meshblock. A deterministic match rule was used whereby a link was declared if all the matching variables agreed exactly on the two records being compared (a vitals registration record and a Census 2001 individual record).

Instances where a vitals record was matched to two or more census records were excluded from any comparisons made. This was done as, without detailed checking of registrations records matched more than once to census records, we had no way of knowing which match was most likely to be a true match. Excluding vitals records matched more than once to census helped keep down the occurrence of false matches in the file of matched records to be used for comparisons of ethnic response. The proportion of false matches remaining in vitals records matched only once to census was estimated to be about 0.5 percent (1 in 200 records).

Whether a match is achieved will be affected by the quality and coverage of the matched data sources. Census data is not available for all New Zealand residents, because of undercount and because some residents were out of the country temporarily at the time of the census. Match success will also be affected by the quality of the matching variables in the two data sources, in particular, missing values and errors (due to the respondent or processing). Match rates above 85% were seen as unlikely to be achieved.

Prior to comparing the ethnic determinations on the vitals records matched only once, the variables affecting the chances of a match were examined. Achieving a match (or not) was found to be affected by the proximity of the vitals event to the date of the 2001 Census (6/3/01), age of the deceased (or age of the mother/father in the case of births) and ethnicity. Additionally, for matching involving births data (for babies, mothers, fathers) match success was also affected by age, sex of the parent, whether the parents were married or not, and other variables. The variables affecting match success were used to derive record weights to take account of the success of matching when comparing counts of ethnicity in the two data sources of the matched data. All results discussed below are based on weighted comparisons. Comparisons of ethnicity collected from the two contexts (vitals and census) focussed on the five main ethnic indicators, European, Mäori, Pacific, Asian and Other.

There was no strong evidence for deriving and using "adjustment ratios" to correct for differing ethnic responses when using mortality data for 2001-02 and 2001 Census data. Evidence to date supports the view that, for the three ethnic indicators European, Mäori and Pacific, the net effect of ethnic mobility was generally minor. Death registrations data on the three main ethnic indicators for the period after the 2001 Census and up to 30 June 2002, are broadly comparable to 2001 Census data. Despite this broad comparability, relatively small differences were apparent by sex and 10-year age group, for European and for Mäori. For Mäori, the differences are relatively small when compared with the adjustment ratios derived by Ajwani et al for the period 1980-1999 (Mäori differences could be examined as Ajwani et al used prioritised ethnic group with Mäori having the highest priority, and our study used Mäori ethnic group). There was insufficient matched data for the Asian and Other ethnic groups to enable sound conclusions to be drawn.

There was no strong evidence for using "adjustment ratios" to correct for differing ethnic responses when using data on newborn, mothers and fathers, for 2001 birth registrations and 2001 Census data. The data examined supports the view that for the five main ethnic indicators the net effect of ethnic mobility was generally minor. Adjustment ratios for newborn of European, Mäori, Pacific, Asian or Other ethnicity were all within 3.1% of 1.00. Analyses taking account of mother's age or father's age were consistent with the summary level analysis.

---

Customised Output

National Estimated Resident Population:
Estimated resident population by single year of age and sex;
Estimated resident population by five year age group and sex.
'At' the end of quarters, from 31 March 1991 onwards;
'Mean year ended', for years ended at the end of quarters, from 31 December 1991 onwards.

Subnational Estimated Resident Population:
Estimated resident population by regional council, territorial authority and urban areas;
by five year age group and sex.
'At 30 June' for 1996 onwards.

(See the "Design" section of this Profile for information on other estimates available)

---

Catalogue & Reference Numbers

INFOS:

Demography Population Estimates DPE Demography Age Estimates DAE

---

---

Other Comments

---

Classification(s) used

Classifications	Classification Versions	Classification Type
AGE - Age - New Zealand Standard Classification	V1.0	NZ Standard
AGEGPS - Age - 5 Year Groupings	V1.0	Standard
SEX - Sex - New Zealand Standard Classification	V1.0	NZ Standard
BIRTHMON - Date of Birth - Month Classification	V2.0	Standard
BIRTHYR - Date of Birth - Year Classification	V11.0	Standard
ETHNIC05 - Ethnicity New Zealand Standard Classification 2005	V1.0	Standard
DWELTYPE05 - Standard Classification of Occupied Dwelling Type 2005	V1.0	Standard
MAORIDES - Mäori Descent - Standard Classification	V1.0	Standard
USRES - Usual Residence Classification	V1.0	NZ Standard
AU06 - Area Unit 2006	V1.0	NZ Standard
TA06 - Territorial Authority 2006	V2.0	NZ Standard
CON07 - Regional Council Constituency 2007	V1.0	NZ Standard
REGC06 - Regional Council 2006	V1.0	NZ Standard
WARD07 - Ward 2007	V1.0	NZ Standard
MCON07 - Maori Constituency 2007	V1.0	NZ Standard
CB07 - Community Board 2007	V1.0	NZ Standard
SUB07 - Subdivision 2007	V1.0	NZ Standard
UA06 - Urban Area 2006	V1.0	NZ Standard

---

Glossary of Terms

Term	Description
Area Unit	Area units are aggregations of meshblocks with unique names. They are non-administrative areas intermediate between meshblocks and territorial authorities. Area units must either define or aggregate to define urban areas, rural centres, statistical areas, territorial authorities and regional councils. Each area unit must be a single geographic entity with a unique name. Area units of main or secondary urban areas generally coincide with suburbs or parts thereof. Area units within urban areas normally contain 3,000-5,000 population.
Base Population	The starting population, usually distributed by age and sex, from which population estimates and projections are derived. The base population for all 2006-base national and subnational population estimates and projections is the estimated resident population at 30 June 2006.
Census Night Population Count	A count of all people present in a given area on a given census night. The census night population count of New Zealand includes visitors from overseas who are counted on census night, but excludes New Zealand residents who are temporarily overseas. For a subnational area the count includes visitors from overseas and elsewhere in New Zealand (people who do not usually live in that area), but excludes residents of that area who are temporarily elsewhere on census night (people who usually live in that area but are absent). For example: The census night population count of New Zealand at 7 March 2006 was 4,143,279. The census night population count of the Auckland region includes all people present and counted in Auckland region on census night. It includes residents present, visitors from elsewhere in New Zealand and visitors from overseas. It excludes people who usually live in Auckland region but were not in Auckland region on census night.
Census Usually Resident Population Count	A count of all people who usually live in a given area, and are present in New Zealand, on a given census night. The census usually resident population count of New Zealand excludes visitors from overseas and excludes New Zealand residents who are temporarily overseas. For a subnational area the count excludes visitors from overseas and elsewhere in New Zealand (people who do not usually live in that area), but includes residents of that area who are temporarily elsewhere in New Zealand on census night (people who usually live in that area but are absent). Residents are counted at the meshblock (smallest geographic area for statistical purposes) of their usual address. If insufficient information about a person's usual address is given, then the geographic information that is available is used to select (impute) a meshblock within that area. For example: The census usually resident population count of New Zealand at 7 March 2006 was 4,027,947. A person who usually lives in Christchurch city but was in Wellington city on a given census night will be included in the census usually resident population count of Christchurch city, and the census night population count of Wellington city. They will be excluded from the census night population count of Christchurch city and the census usually resident population count of Wellington city.
Cohort	A group of people sharing a common demographic experience. For example, the 1900 birth cohort refers to the people who were born in the year 1900.
Cohort Component Method	A method of estimating or projecting the population by updating the size of each age-sex group in the base population for deaths and migration within each age-sex group during the period between the base date and a given date. New birth cohorts result from births between the base date and the given date. In deriving population projections, the base population is projected forward by calculating the effect of deaths and migration within each age-sex group according to specified mortality and migration assumptions. New birth cohorts are generated by applying specified fertility assumptions to the female population of childbearing age. In deriving population estimates, actual events on births, deaths and migration are generally used instead of assumptions.
De facto Population Concept	A statistical basis for a population in terms of those present in a given area at a given time. The census night population count is a census measure, and the estimated de facto population is a demographic measure, of the de facto population concept.
Estimated Resident Population	An estimate of all people who usually live in a given area at a given date. The estimated resident population of New Zealand includes all residents present in New Zealand and counted by the census (census usually resident population count), residents who are temporarily overseas (who are not included in the census), and an adjustment for residents missed or counted more than once by the census (net census undercount). Visitors from overseas are excluded. For a subnational area the estimate excludes visitors from overseas and elsewhere in New Zealand (people who do not usually live in that area), but includes residents of that area who are temporarily elsewhere on census night (people who usually live in that area but are absent). The estimated resident population at a given date after census includes births, deaths and net migration (arrivals less departures) of residents during the period between census night and the given date. National population estimates are produced quarterly (reference dates at 31 March, 30 June, 30 September and 31 December) and subnational population estimates are produced annually (reference date at 30 June). For example, the estimated resident population of New Zealand at 30 June 2006 was 4,184,600
Ethnicity	An ethnic group is made up of people who have some or all of the following characteristics: a common proper name one or more elements of common culture which need not be specified, but may include religion, customs or language unique community of interests, feelings and actions a shared sense of common origins or ancestry, and a common geographic origin. This definition is based on 'The Ethnic Origins of Nations' (A Smith, 1986). Ethnicity is self-perceived and people can have more than one ethnic group. People can identify with an ethnicity even though they may not be descended from ancestors with that ethnicity. Conversely, people may choose to not identify with an ethnicity even though they are descended from ancestors with that ethnicity. Ethnicity is not the same as birth place. In the case of births and deaths, ethnicity is identified by the person completing the registration form. For births this is usually the parent(s), while for deaths this is most likely to be the funeral director (on the advice of a family member).
Inter Ethnic Mobility (ethnic category jumping)	People changing their ethnic identification over time. Inter-ethnic mobility may occur when different people respond to the ethnicity question for one person. For example, the ethnicity of babies and young children is usually identified by their parents. However, in a later census when these children are old enough to complete their own forms, they will decide for themselves which ethnicity they identify with. This may differ from the ethnicity identified by their parents. Inter-ethnic mobility can also occur when different ethnicities are reported between different collections (eg birth registration form, death registration form, census form) for one person.
Labour Force	People aged 15 years and over who regularly work for one or more hours per week for financial gain, or work without pay in a family business, or are unemployed and actively seeking part-time or full-time work. The full-time labour force comprises persons working 30 hours or more per week and unemployed persons seeking full-time work. The part-time labour force comprises persons working 1 to 29 hours per week and unemployed persons seeking part-time work.
Net Census Undercount	The difference between undercount and overcount. It is usually expressed as a percentage of what should have been the complete count rather than as a percentage of what was counted. The base population for population estimates and projections includes an adjustment for net undercount in the most recent Census of Population and Dwellings.
Regional Council (Region)	The Local Government Commission established regional councils in 1989. Regional councils cover every territorial authority in New Zealand with the exception of the Chatham Islands Territory. The geographical boundaries of regions conform as far as possible to one or more water catchments. In determining regions, consideration was also given to regional communities of interest, natural resource management, land use planning, and environmental matters. There are 16 regions in New Zealand. Twelve of these regions are administered by regional councils while the Gisborne, Tasman, Nelson and Marlborough Regions are administered by their respective district or city councils. The boundaries of territorial authorities are generally the same as regional council boundaries, although there are eight instances where territorial boundaries straddle regional boundaries
Resident Population Concept	A statistical basis for a population in terms of those who usually live in a given area at a given time. The census usually resident population count is a census measure, and the estimated resident population is a demographic measure, of the resident population concept.
Resident Temporarily Overseas	A person who usually lives in New Zealand but who is overseas for a period of less than 12 months. The base population for population estimates and projections includes an adjustment for residents temporarily overseas at the time of the most recent Census of Population and Dwellings.
Subnational	Geographical units of a country (eg area units, territorial authorities, regional councils, urban areas) whose boundaries are defined for administrative, legal or statistical purposes.
Urban Area	Non-administrative areas with urban characteristics and a high to moderate concentration of population. The classification of urban areas was revised for the 1991 Census of Population and Dwellings into three parts – main, secondary and minor urban areas: Main urban areas – centres with populations of 30,000 or more. There are currently 16 main urban areas (12 in the North Island and four in the South). Auckland, Wellington, Hamilton, Napier-Hastings are further subdivided into zones. Secondary urban areas – centres with populations between 10,000 and 29,999. There are currently 14 secondary urban areas. Minor urban areas – centres with populations of 1,000 or more not already classified as urban (that is, not falling within a main or secondary urban area). There are currently 99 minor urban areas and, together with the above two categories, they constitute the urban population of New Zealand. Urban areas are currently defined on the basis of the 1996 Census usually resident population count. As a result, Greymouth is still classified as a secondary urban area, even though the 2001 Census usually resident population count, and the estimated resident populations at 30 June 2001–2005, fall below 10,000.

---

Contact Details

Customer Service E-mail Contact:	info@stats.govt.nz
Survey Enquiries E-mail Contact:	demography@stats.govt.nz

Liability

Statistics New Zealand gives no warranty that the information or data supplied contains no errors. However, all care and diligence has been used in processing, analysing and extracting the information. Statistics New Zealand shall not be liable for any loss or damage suffered by the customer consequent upon the use directly, or indirectly, of the information supplied in this product.