National Family and Household Projections 2013-base update

Reference period

This release contains updated 2013-base projections of families and households usually living in New Zealand. These supersede the 2013-base projections released in October 2015. The new projections have the estimated families and households at 30 June 2013 as a base, and cover the period 2014–38 at one-year intervals. The projection period is limited to 25 years because of the uncertainty of family and household projections, as discussed in Nature of projections.

See NZ.Stat for more detailed projection results, including individual years.

As with the 2001-base and 2006-base national family and household projections, a 'propensity' method has been used to produce the latest projections. The family and household projections are derived from estimates (for 2014–16) and projections (for 2017–38) of the New Zealand population, by multiplying the population by assumed living arrangement type rates for each age-sex group. The projections of population by living arrangement type are subsequently aggregated to give projections of families (by broad family type) and households (by broad household type). Before the 2001-base projections, we produced household projections using a 'household head' method.

Consistency with other projections

These family and household projections complement National Population Projections: 2016(base)–2068 released on 19 October 2016. However, only the medium projection of the family and household projections and the median projection (50th percentile) of the national population projections are designed to be directly comparable. Other projections or percentiles cannot be directly compared because the projection assumptions may be incompatible.

Changes since the previous 2013-base projections

Review of assumptions

Deriving the projections involves a review of all projection assumptions. We updated the national family and household projections to incorporate the latest demographic information, notably the 2016-base national population projections (released on 19 October 2016), and living arrangement type rates (LATRs) assumptions. Variant A assumes LATRs remain constant at 2013 levels. Variant B assumes LATRs will change linearly between 2013 and 2038, based on an assessment of observed trends between 1986 and 2013 and likely future trends.

See Living arrangement type rates (LATRs)

for more information.

Projection assumptions

We formulated the projection assumptions after analysing short-term and long-term historical trends, recent trends and patterns observed in other countries, government policy, and other relevant information.

Base population

These projections have as a base the estimated resident population (ERP) of New Zealand at 30 June 2013. We derived this population (4.442 million) from the census usually resident population count at 5 March 2013 (4.242 million), with adjustments for:

• net census undercount (+104,000)

• residents temporarily overseas on census night (+82,000)

• births, deaths, and net migration between census night and 30 June 2013 (+9,000)

• reconciliation with demographic estimates at ages 0–9 years (+5,000).

We derived the estimated number of families and households indirectly from the estimated resident population and the estimated living arrangement type rates for each age-sex group. The estimated number of families (1.247 million) and households (1.648 million) are equivalent to the census family count (1.136 million) and census household count (1.550 million), respectively, at 5 March 2013, with adjustments for:

• net census undercount

• families and households temporarily overseas on census night

• change between census night (5 March 2013) and 30 June 2013

• families and households temporarily absent from their usual residence but elsewhere in New Zealand on census night.

Fertility

There are three alternative fertility variants – designated low, medium, and high – which assume that in 2038 the total fertility rate will reach 1.65, 1.85, and 2.05 births per woman, respectively.

The medium fertility variant assumes fertility rates of women aged under 35 years will decline between 2017 and 2036, but increase for women aged 35 years and over. By comparison, the low fertility variant assumes fertility rates of women aged under 39 years will decline between 2017 and 2036, but increase for women aged 39 years and over. The high fertility variant assumes that fertility rates of women aged under 24 years will decline between 2017 and 2036, but increase for women aged 24 years and over.

A sex ratio at birth of 105.5 males per 100 females is assumed, based on the historical annual average of the New Zealand population.

Mortality

There are three alternative mortality variants – designated low, medium, and high – which assume that mortality rates will continue to drop so that the period life expectancy at birth for males will increase to 85.7, 84.8, and 83.8 years, respectively, by 2038. The corresponding life expectancies for females in 2038 will be 88.5, 87.7, and 86.9 years. The life expectancy at birth in 2012–14, from complete period life tables, was 79.5 years for males and 83.2 years for females.

Migration

There are three alternative migration variants – designated low, medium, and high – which assume long-run annual net migration of 10,000, 15,000, and 20,000, respectively. The medium migration variant assumes net migration decreases from 60,000 in 2017 to 24,000 in 2021, and to 15,000 in 2022 and beyond. The low and high migration variants are 5,000 lower and higher, respectively, than the medium variant for each year. The assumed medium long-run annual net migration of 15,000 reflects the average annual gain of 10–20,000 migrants since the late 1980s.

Living arrangement type rates (LATRs)

There are two alternative LATR variants – designated A and B. Variant A assumes that LATRs will remain constant at 2013 levels. Variant B assumes that LATRs will change linearly between 2013 and 2038, based on an assessment of observed trends between 1986 and 2013 and likely future trends, by sex and single-year of age.

Variant B is the preferred variant, because it has been formulated to produce demographically plausible results. For variant B, the main changes in LATRs assumed between 2013 and 2038 are:

• Partner in couple-without-children family: Increasing rates for males aged over 70 years. Increasing rates for females aged up to 42 years, and 63 years and older. This reflects lower fertility rates with fewer couples having children, couples having children at older ages, and a slight convergence of male life expectancy to female life expectancy with more couples having both partners living to older ages.

• Partner/parent in two-parent family: Increasing rates for females at most ages, especially at ages 30–70 years. This reflects lower fertility rates with fewer couples having children and delayed childbearing.

• Child in two-parent family: Increasing rates at most ages, especially at ages 0–30 years. This reflects decreased rates of single parenting from separation, divorce, and childbearing outside couple relationships, as well as children living with parents at older ages.

• Parent in one-parent family: Increasing rates for males. Decreasing rates for females, which reflects decreased rates of single parenting and lower fertility rates.

• Child in one-parent family: Decreasing rates at younger ages, especially at ages 0–20 years. This reflects increased rates of re-partnering of parent.

• Person in other multi-person household: Decreasing rates for females, this reflects increasing rates of partnerships.

• Person in one-person household: Increasing rates especially at ages 30–70 years for males. The proportion of females years living alone is assumed to drop, given a slight convergence of male life expectancy to female life expectancy.

• Person in non-private dwelling: Decreasing rates at older ages, associated with increasing life expectancy and well-being and declines in morbidity rates, enabling people to live independently at home longer.

• Average number of families per family household is assumed to increase from 1.050 to 1.075, reflecting increasing co-residence of multi-generational families.

• Average number of people per other multi-person household is assumed to increase from 2.637 to 2.773, reflecting group-living in response to housing costs.

For variants A and B, the following factors remain constant at the 2013 levels:

• the proportion of two-parent families with dependent children is assumed to remain constant at 0.817

• the proportion of one-parent families with dependent children is assumed to remain constant at 0.720.

Which projection should I use?

We have produced six alternative projections by combining three population projections with two variants of living arrangement type rates.

The three population projections are:

• Low which assumes low fertility, high mortality, and low migration

• Medium which assumes medium fertility, medium mortality, and medium migration

• High which assumes high fertility, low mortality, and high migration.

The two living arrangement type rates are:

• A which assumes LATRs will remain constant at 2013 levels

• B which assumes LATRs will change linearly between 2013 and 2038.

At the time of release, we consider the Medium B projection the most suitable for assessing future family and household changes. Moreover, only Medium B has been formulated to produce demographically plausible results, by assessing both observed trends between 1986 and 2013 and likely future trends to 2038. Other projections may project significantly different numbers of male and female partners in couple-without-children and/or two-parent families, because the living arrangement type rate A variants are formulated solely from observed historical rates.

The other projections allow users to assess the impact on the number of families and households resulting from different population and/or living arrangement type scenarios. For example, Low B, Medium B, and High B can be used for assessing the effect of different population outcomes combined with variant B living arrangement type rates; and series Medium A and Medium B illustrate the effect of different living arrangement type assumptions combined with the mid-range population scenario.