The mechanisms of subnational population growth and decline in
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Natalie Jackson and Lars Brabyn
The mechanisms of
subnational population
growth and decline in
New Zealand
1976-2013
The broader rationale for the Tai Timu
Tangata project is outlined by Jackson
(infra). Essentially, current New Zealand
has a relatively young and rapidly growing
Introduction: population. However, widespread
subnational depopulation between 1996
This article summarises key findings from the strand of and 2013 saw one-third of the nation’s
TAs decline in size; Auckland and 12 TAs
the Tai Timu Tangata. Taihoa e? project that examined the shared 90 percent of growth, while the
mechanisms of subnational population change in New remaining 10 percent of growth was
spread very thinly across 32 TAs. The
Zealand for 143 towns, 132 rural centres and 66 territorial situation has led to some towns being
authority areas (hereafter TAs), for the 37-year period disparagingly labelled as ‘zombie towns’
(NBR 2014), and contrasted against their
1976-2013. Because of space constraints we present the more successful growing counterparts.
With two large cities among those
information as a set of 10 summary observations. For the declining (Dunedin and Invercargill), and
underlying analyses please refer to Jackson, Brabyn and Maré structural ageing known to drive a
reduction in natural increase (the
(2016); Jackson and Cameron (2017), Jackson, Brabyn, Maré, difference between births and deaths), we
wished to better understand why some
Cameron and Pool (forthcoming); and Jackson and Brabyn
areas are growing and others not.
(forthcoming). Specifically, we wished to know whether
parts of subnational New Zealand might
Dr Natalie Jackson was Professor of Demography at Waikato University and Adjunct Professor of be following their international
Demography at Massey University during the period of this research. She is also Director, Natalie
Jackson Demographics Ltd (demographics@nataliejackson.net). Dr Lars Brabyn is Senior Lecturer,
counterparts in declining from what is
School of Geography, University of Waikato (larsb@waikato.ac.nz). proposed as a ‘new’ and increasingly
Page 22 – Policy Quarterly – Volume 13, Supplementary Issue – June 2017
intractable form of population decline Figure 1: Percentage of Territorial Authority Area, Town and Rural Centre
(where net migration loss is accompanied Populations experiencing net migration gain or loss, 1976-2013
by natural decrease), as opposed to the Net Migration Gain
‘old’ form, where natural increase is 90.0
positive but fails to offset net migration
loss (Bucher & Mai 2005, cited in Matanle 80.0
& Rausch 2011: 19-20, 46-47). 70.0
The question was not merely academic.
60.0 51.7
Globally, population growth is theorized 51.5 51.0
Percentage
to end around the end of the present 50.0
century, but much sooner across the
40.0
developed world (Lutz, Sanderson &
Sherbov 2004; Davoudi, Wishardt & 30.0
Strange 2010; Reher 2007, 2011, Lee &
20.0
Reher 2011). Reports from newly
depopulating countries such as Japan are 10.0
largely negative, indicating reduced 0.0
investment in local infrastructure, TAs (N=66) Towns (N=143) Rural Centres (N=132)
widespread abandonment of schools, 1976-81 1981-86 1986-91 1991-96 1996-01 2001-06 2006-13
homes and business, and general social, Net Migration Loss
economic and environmental damage 90.0
78.8 78.8
outside of the main cities; at the same 74.2
80.0
time, opportunities arising from a 71.2
potential ‘depopulation dividend’ 70.0 62.2
(Matanle 2017) need to be engaged with 57.6
60.0
in a timely manner, long before local 50.3 50.3 49.2
Percentage
councils and similar agencies are 50.0
overwhelmed with sustained depopulation 40.0
that they have not been anticipating
(Matanle & Sato 2010; Audirac 2012; 30.0
Martinez-Fernandez, Kubo, Noya & 20.0
Weyman 2012; McMillan 2016).
10.0
Indeed, if subnational New Zealand is
following its international counterparts 0.0
while at the same time being relatively TAs (N=66) Towns (N=143) Rural Centres (N=132)
youthful and growing strongly at national 1976-81 1981-86 1986-91 1991-96 1996-01 2001-06 2006-13
level, we asked if there could be broader Source: Jackson and Brabyn (forthcoming).
Notes: These data were first published in Jackson et al. 2016. Proportions here differ slightly, due to database revisions.
theoretical implications that would
contribute to a theory of depopulation, New Zealand’s coding. The demographic Our 143 towns and 132 rural centres
and thereby support councils and other components of change (births, deaths, conform to Statistics New Zealand’s
agencies to plan for this eventuality in a natural increase) for each area were then ‘urban areas’. This means that our ‘towns’
more positive way (McMillan 2016). retrospectively modelled using TA-level range in size from major (>30,000
Because the data we needed for the fertility and survivorship rates. An people) to minor (1,000 – 9,999 people)
exercise were not available for many estimate of net migration (total and by urban areas, and our ‘rural centres’, from
subnational jurisdictions, or for the age) was then extracted via the 300 to 999 people, in terms of their size in
period required, or on consistent conventional residual method (net change 1976. Under these arrangements,
geographical boundaries, we had to first minus natural increase = net migration). Auckland is divided into four zones,
extract them via statistical means. The The exercise permitted us to develop while data for rural districts is excluded.
methodology for creating this unique births, deaths, natural increase and The latter was an unfortunate necessity,
database is briefly described in the population by age data, both with and reflecting the large number of small units
attached Appendix, with more detail without migration. Although the baseline that would have needed to be analysed;
available in Jackson et al., (2016). Key data and all rates for modelling have been however, where possible we note trends
methodological points are that ‘usually sourced from statistics New Zealand, the for the aggregate rural district
resident population’ data for all census resulting output has been developed for population.
years 1976-2013 were aggregated to 2013 the purposes of this project, and should Methodologically it should also be
geographic boundaries based on Statistics not be seen as official statistics. noted that each data point carries equal
Policy Quarterly – Volume 13, Supplementary Issue – June 2017 – Page 23The mechanisms of subnational population growth and decline in New Zealand 1976-2013
Figure 2: Percentage of TAs, Towns and Rural Centres experiencing net migration and deaths, while mobility theories by
loss, and actually declining, 1976-2013 definition deal with ‘open’ populations
Territorial Authority Areas affected by migration, and are applied sub-
78.8 78.8 nationally, nationally and globally. To avoid
80.0
getting tangled in theoretical constructs
70.0
57.6
(but see but see Pool, infra; Jackson and
60.0
Pool, forthcoming), we proceed here on
50.0 the basis that the New Zealand population
Percentage
60.6
40.0 is in fact both ‘open’ as in mobility theories,
30.0
47.0 and is ageing structurally, as theorised in
demographic transition (following Dyson
20.0 28.8
2011).
10.0 Following are ten key observations
0.0 drawn from the project. These are
1976-81 1981-86 1986-91 1991-96 1996-2001 2001-06 2006-13
necessarily brief, but, as noted, more
Net Migration loss Actually Declined
Towns formally elaborated in the referenced
80.0 papers. We conclude the article with a
70.0 short summary and consideration of
62.2 related policy implications.
60.0
50.3 50.3
50.0
Percentage
Observation 1: The majority of areas
40.0 55.9
experience net migration loss rather than gain
30.0 Figure 1 compares the percentage of TAs,
20.0 35.7 towns, and rural centres experiencing
28.7 net migration gain or loss from 1976-
10.0
0.0
2013. The majority of TAs and rural
centres experienced somewhat greater
1976-81 1981-86 1986-91 1991-96 1996-2001 2001-06 2006-13
Net Migration loss Actually Declined
net migration loss than gain at most
observations. For towns the split was
Rural Centres
80.0 typically close to 50:50, although net
74.2 72.7
70.0
loss was somewhat greater than gain
between 1986 and 1991, and 1996 and
60.0
50.0
2001. Recalling that rural district data
50.0 62.9 were not analysed in detail for this
Percentage
59.8
40.0 project, their aggregate suggests a slightly
47.0 different story, with slightly more years of
30.0
gains than losses, and those gains being
20.0
sustained since 2001.
10.0
0.0 Observation 2: Net migration loss does not
1976-81 1981-86 1986-91 1991-96 1996-2001 2001-06 2006-13
Net Migration loss Actually Declined always result in population decline
Source: Jackson and Brabyn (forthcoming)
Figure 2 shows the percentage of TAs,
towns and rural centres experiencing net
weight, irrespective of size. This was a communities. The exclusion of rural migration loss from 1976 to 2013, and
deliberate choice, in that our research is districts from our detailed analyses the percentage actually declining. The
concerned with the extent to which means that such communities have only difference between the two measures
individual jurisdictions – all of which broad trends on which to deliberate. is accounted for by natural increase
have implications for such things as rates One major theoretical conundrum also (births minus deaths), discussed further
and resources – are affected, rather than needs to be acknowledged. As outlined in below. On average 59 percent of TAs
what proportion of the total New the introduction to this Issue, the project experiencing net migration loss actually
Zealand population is affected (Jackson has been informed by both demographic declined; for towns and rural centres
& Cameron 2017). For example, TA level and mobility transition theories. the proportions were 70 and 84 percent
patterns and trends are of interest to TA Demographic transition theory (Davis respectively. These data indicate that TAs
councils, while those for towns and rural 1945) is a ‘global and national’ level theory and towns have been more able than rural
centres are of interest to local councils, that essentially treats populations as centres to cover their net migration loss
and their respective planners and ‘closed’, that is, changing only due to births with natural increase, and conversely, that
Page 24 – Policy Quarterly – Volume 13, Supplementary Issue – June 2017net migration loss is a stronger predictor Figure 3: Percentage of Territorial Authority Area, Town and Rural Centre
of net decline for rural centres than towns Populations Growing/Declining 1976-2013 by impact of migration on size
and TAs. 100
90
Observation 3: The majority of areas are 25.8
32.6
smaller with, than without, migration 80 39.2
In total, 62 percent of TAs, 66 percent of
towns and 50 percent of rural centres were 70
larger in 2013 than in 1976. However, 17.4
60
Percentage
36.4
because net migration was often negative,
a proportion of those growing across the 50 27.3
period was smaller with migration than
40
they would have been without. By ‘with
migration’ we mean population change due 30 49.2
to both migration and natural increase; by 37.9
33.6
‘without migration’ we mean population 20
change due to natural increase only (see
10
Appendix). Thus, those growing, but
smaller than they would have been without 0
Territorial Authority Areas Towns (N=143) Rural Centres (N=132)
migration, owe much of their growth to
natural increase. We elaborate on these Group 1 (growing, Group 2 (growing, Group 3 (declining, Declined both with and
larger with migration smaller with migration smaller with migration without migration (1)
interactions further below. than without) than without) than without)
The exercise generated three groups of Source: Jackson and Brabyn (forthcoming).
Notes: (1) refers to one rural centre (Waitati) which declined both with and without migration, but slightly less 'with'
interest (Figure 3):
Group 1: Areas that grew between 1976 Figure 4: Percentage of Territorial Authority Area, Town and Rural Centre Populations
experiencing natural increase or natural decrease, 1976-2013, and projected
and 2013 and were larger with for TAs to 2043
migration than without. This
Natural Increase
situation pertained to 26 percent of Projected TAs
100.0
TAs, 39 percent of towns, and 33
percent of rural centres.
80.0
Group 2: Areas that grew between 1976
and 2013, but were smaller in 2013 60.0
with migration than without. These
Percentage
areas experienced periods of net 40.0
migration loss, but it was either
partially or fully offset by natural 20.0
increase. This situation pertained to
36 percent of TAs, 27 percent of 0.0
towns, and 17 percent of rural TAs (N=66) Towns (N=143) Rural Centres (N=132)
centres. In the aggregate it also 1976-81 1981-86 1986-91 1991-96 1996-01 2001-06 2006-13
pertained to rural districts, the total 2013-18 2018-23 2023-28 2028-33 2033-38 2038-43
population of which was larger in Natural Decrease
2013 (513,951) than in 1976 70.0
Projected TAs
(398,436), but smaller than it would
60.0
have been in the absence of migration
(599,218). 50.0
Group 3: Areas that declined in size
40.0
Percentage
between 1976 and 2013, all of which
were also smaller in 2013 with 30.0
migration than without. In most cases 20.0
these areas experienced natural
10.0
increase, but it was completely offset
by net migration loss. This situation 0.0
TAs (N=66) Towns (N=143) Rural Centres (N=132)
pertained to 38 percent of TAs, 44
percent of towns, and 39 percent of 1976-81 1981-86 1986-91 1991-96 1996-01 2001-06 2006-13
2013-18 2018-23 2023-28 2028-33 2033-38 2038-43
rural centres.
Policy Quarterly – Volume 13, Supplementary Issue – June 2017 – Page 25The mechanisms of subnational population growth and decline in New Zealand 1976-2013
Table 1: Combination of components of change that deliver net growth or net decline decrease is as yet barely evident at TA
Outcome Combination of Components level, the projections (medium variant)
GROWTH A: Natural Increase and Net Migration are both positive
indicate that it will be the experience of
the majority (61 percent) of TAs by 2043,
B: Natural Increase offsets Net Migration Loss
and thus of many of the towns and rural
C: Net Migration Gain offsets Natural Decrease
centres which comprise them (see also
DECLINE D: Net Migration Gain fails to offset Natural Decrease Jackson & Cameron 2017, and Cameron
E: Natural Increase fails to offset Net Migration Loss infra).
F: Natural Decrease and Net Migration Loss
ZERO GROWTH G: Natural Increase = Net Migration Loss Observation 5: There are complex
H: Natural Decrease = Net Migration Gain interactions between net migration and
Source: Jackson, Cameron and Pool (2015). Notes: By ‘offset’ we mean to ‘completely conceal’ natural increase/decrease
Net migration and natural increase
In sum, migration is not a panacea for hold that the emergence of natural decrease or natural decrease interact in several
growth, while growth per se may not is initially ‘incipient’ (intermittent onset), different ways to generate three different
reflect net migration gain, but rather, be 47 towns and rural centres (17 percent) outcomes: growth, decline, and zero
the result of natural increase offsetting experienced natural decrease across more growth. Our project has identified three
underlying net migration loss. than one five-year period between 1976 combinations that generate growth, three
and 2013, and 22 (8 percent) experienced that result in decline, and two that result
Observation 4: The majority of areas it across five or more periods. In 2013, 19 in zero growth (Jackson, Cameron & Pool
experience natural increase – but this will towns (13 percent) and 20 rural centres 2015). We refer to these as Types A-C
soon change, with profound implications (15 percent) were regularly experiencing growth, Types D-F decline, and Types G
During the period 1976-2013 the natural decrease; for towns this had and H zero growth (Table 1, and Figures
vast majority of TAs, towns and rural increased slightly, and for rural centres, 5 and 6).
centres experienced natural increase, reduced. These findings reflect those For the policy community the typology
with levels in 2013 slightly higher for for the counties of the United States is important because it provides ‘advance
towns than rural centres (Figure 4, left and Europe (Johnson, Field & Poston warnindding end notdg’ of the permanent
panel). In the aggregate, rural districts 2015), and confirm that New Zealand is ending of growth (see also Jackson 2014).
also experienced natural increase across following its older counterparts, but is as Type A growth, where both elements are
each five-year period. Natural decrease yet at a much earlier stage. positive, is more robust than either Type B
thus remains relatively low (right panel). Attempting to extend the analysis or C growth, where one or other element
However, supporting the main tenets of forward, we drew on projections at TA is negative, and is likely to be sustained for
demographic transition theory, which level.1 These data show that while natural much longer. In particular, areas growing
Figure 5: Percentage of Territorial Authority Areas by Cause of Growth and Decline, 1976-2013, and
Projected 2013-2043 (Medium Variant)
Observed Projected
100%
21 21 21 17
30 27
80% 33 39 6
44 41
50 48
11
18 67
21 20
60%
32 26
32 21
24
26 29
40% 24 29 20
17
11 18
20% 61 23 20
45 18 27
42 14
32 29 20
30 24
0%
1976-81
1981-86
1986-91
1991-96
1996-01
2001-06
2006-13
2013-18
2018-23
2023-28
2028-33
2033-38
2038-43
A: Growth from both positive components B: Natural Increase offsets Net Migration Loss C: Net Migration Gain offsets Natural Decrease
D: Net Migration Gain fails to offset Natural Decrease E: Natural Increase fails to offset Net Migration Loss F: Decline from both negative components
Zero Growth
Source: Jackson and Brabyn (forthcoming).
Page 26 – Policy Quarterly – Volume 13, Supplementary Issue – June 2017because of Type B growth, where natural Figure 6: Percentage of Towns and Rural Centres by Cause of Growth and
increase is completely concealing net Decline, 1976-2013
migration loss, are highly vulnerable to Towns (N=143)
100%
decline as natural decrease emerges, and
their growth should not be seen in the 90%
same light as Type A. Type D decline
80% 37 34 37
(where net migration gain fails to offset 38 40 40
44
natural decrease) is more ‘preferable’ to 70%
Type E or F decline, in that the former has
60% 7
a good chance of reverting to population
16 17
growth, at least in the short-term. By 50% 20
15
16
contrast, Type E decline typically indicates 22
40%
that natural increase is becoming very
low, while Type F decline, the ‘new’ form 30%
52
of population decline, foreshadows a self-
20% 36 32
reinforcing and increasingly intractable 29 32
27 27
type of depopulation. In order to ‘extend’ 10%
the period covered in our 1976-2013
0%
analysis to indicate what will unfold for
1976-81 1981-86 1986-91 1991-96 1996-01 2001-06 2006-13
towns and rural centres, we include TA A: Growth from both positive components B: Natural Increase offsets Net Migration Loss
level projections to 2043 (medium C: Net Migration Gain offsets Natural Decrease D: Net Migration Gain fails to offset
variant). E: Natural Increase fails to offset Net Migration Loss Natural Decrease
Zero Growth F: Decline from both negative components
Type A Growth: Across the period 1976-
2013 the majority of TAs and towns Rural Centres (N=132)
grew from ‘Type A’ growth, where 100%
both net migration and natural 90% 20 17
24
increase were positive, although there 30 34
37 35
80%
was an anomalous period, nationally,
12 15
between 1996 and 2001, when this 70% 11
cause of growth was overwhelmed by 8
60% 9
Type E decline (see below). The 11 17
percentage of TAs growing from Type 50%
A growth increased overall across the
40%
period, from 21 to 41 percent (Figure
55 51
5). For towns, Type A growth 30%
43
44
remained relatively stable, at 38 37 34
20% 38
percent of towns in 1976 and 37
percent in 2013; for rural centres, 10%
Type A growth declined across the
0%
period, from 30 to just 17 percent of 1976-81 1981-86 1986-91 1991-96 1996-01 2001-06 2006-13
rural centres (Figure 6). In the A: Growth from both positive components B: Natural Increase offsets Net Migration Loss
aggregate, rural districts also grew C: Net Migration Gain offsets Natural Decrease D: Net Migration Gain fails to offset
E: Natural Increase fails to offset Net Migration Loss Natural Decrease
from Type A growth between 1991
Zero Growth F: Decline from both negative components
and 1996, and 2001-2013. The TA
level projections in Figure 5 show this Source: Jackson et al. (forthcoming)
form of growth increasing until 2018,
then steadily diminishing across the respectively. For rural centres the downwards, and over the next few
period, pertaining to just 17 percent proportions were 8.3 and 15 percent. decades will increasingly leave areas
in 2043. In the aggregate, rural districts experiencing net migration loss,
Type B Growth (where natural increase experienced this type of growth ‘unprotected’. TA projections indicate
completely conceals net migration between 1981 and 1991, and 1991- that this form of growth will diminish
loss) declined across the period for 1996. This type of growth places as the proportion experiencing
TAs and towns but increased for rural affected areas in a somewhat more natural decrease grows, with Type B
centres. For TAs it pertained to 32 vulnerable position than Type A growth accounting for as few as
percent in 1976 and 29 percent in growth, because as indicated above, 6 percent of TAs in 2043.
2013; for towns, 22 and 17.5 percent natural increase is trending
Policy Quarterly – Volume 13, Supplementary Issue – June 2017 – Page 27The mechanisms of subnational population growth and decline in New Zealand 1976-2013
Figure 7: Percentage of Territorial Authority Area, Town and Rural Centre Populations population change for either towns or
by impact of migration on structural ageing (as indicated by percentage rural centres. Instead, the long-
aged 65+ years, with and without migration)
theorized arrival of zero growth as the
100.0 ‘end point’ of demographic transition
has been shown in our project – as it
90.0
has internationally – to be a
80.0 transitional stage through which most
70.0 areas pass on their journey from
66.4
growth to decline, and in a few
Percentage
60.0 78.8
84.7 temporary cases, vice versa.
50.0 In sum, the TA level projections
40.0 provide advance warning of the ongoing
impact of structural ageing, with both
30.0
Type A and B growth steadily diminishing
20.0 and Type C growth (net migration gain
33.6
10.0 21.2
15.3
offsetting natural decrease) ultimately
taking over as the main cause of growth.
0.0
Territorial Authority Areas (N-66) Towns (N=143) Rural Centres (N=132) Type D, E, and F decline each grow, with
Younger/same with Migration Older with Migration Type F decline (both components
Source: Jackson and Brabyn (forthcoming). negative) taking over as the main cause of
decline around the mid-2030s, and of
Type C Growth (where net migration gain to 29 percent; and for towns and rural population change per se in 2043 (Jackson
offsets natural decrease) has as yet centres, increased from 27 to 32 et al., 2016).
been the experience of relatively few percent, and 47 to 60 percent,
areas, because natural decrease itself respectively. Aggregate data for rural Observation 6: The vast majority of TAs,
is as yet not widespread. However, TA districts suggests they experienced towns and rural centres are older as a result
level projections indicate that Type C this combination between 1976 and of migration
growth will now become more 1981 only. TA projections indicate Across the period 1976-2013, migration
prominent, pertaining to around 20 this form of decline will remain a caused the majority of TA, town and
percent by 2043, and becoming the common experience for the towns rural centre populations to have older
largest driver of growth and the and rural centres which comprise age structures than would have been the
second-largest driver of population them, although will diminish over case in the absence of migration (Figure
change. Type C growth may be a more time as it gives way to Type F decline 7). That is, migration either removed
robust driver of growth than Type B, (see below). This is the shift from the young people and/or added older people,
at least in the short-term, because ‘old’ form of depopulation referred to causing the proportion aged 65+ years
while migration remains positive it by Bucher and Mai (2005) to the ‘new’ to be greater than it would have been
has a good chance of offsetting form, where both elements are without migration.2
underlying natural decrease. negative. This is known as ‘age-selective’
Type D Decline (where net migration gain Type F Decline (where both elements are migration (see below), which has also
fails to offset natural decrease) is as negative) has so far played a minor been shown internationally to accelerate
yet similarly uncommon, not yet role in New Zealand’s subnational structural ageing as much as the
experienced at TA level and population change, albeit more conventional driver, low fertility rates
pertaining to less than 3 percent of evident for rural centres than towns. (Johnson et al., 2015).
towns and 4 percent of rural centres At TA level the combination has not The greatest ‘juvenescent’ (youth-
in 2013. TA projections indicate that yet been experienced. However, imparting) impact was for towns, only 15
this form of decline will become projections indicate that it will percent of which had populations younger
noticeable from around 2028 and become notable at TA level from the with, than without, migration. The same
pertain to around 6 percent of TAs in early 2030s, become the major driver situation pertained to one-fifth of TAs
2043. of decline, and be the main cause of and one-third of rural centres. This means
Type E Decline (where natural increase population change per se, by 2043 that, by contrast, 79, 85, and 66 percent of
fails to offset net migration loss), has, – when it will account for around 27 TAs, towns, and rural centres respectively
by contrast, been very common across percent of TAs. were older with, than without, migration.
New Zealand’s TAs, towns and rural Zero Growth: No cases of zero growth Notably this was not the case for New
centres, although diminishing for the were observed at TA level during 1976 Zealand’s rural districts, which, in the
former and increasing for the latter. to 2013, although a few came very aggregate, had an almost identical
For TAs, the proportion fell from 45 close. It was also a minor cause of percentage aged 65 and older in 2013 both
Page 28 – Policy Quarterly – Volume 13, Supplementary Issue – June 2017with and without migration (13.0 and Figure 8: Percentage of Territorial Authority Area, Town and Rural Centre Populations
13.1 percent respectively). by extent of impact of migration on structural ageing (as indicated by reduction
or increase in percentage aged 65+ years, with and without migration)
Only a minority of TAs, towns and
100.0
rural centres have thus become more
youthful as the result of migration, a 1.5
finding which, at least in terms of 80.0 10.5
international migration, runs counter to 21.2
many governmental pronouncements on 60.0 13.6
35.7
Percentage
the issue (see McDonald & Kippen 1999
17.4
and Kippen & McDonald 2000 for a 40.0
36.4
refutation of this perception for Australia; 17.4
and Jackson & Cameron 2017 for New 20.0 29.4
Zealand). 13.6 11.4
4.2
Moreover, many jurisdictions are 0.0
-4.9 -9.1
decidedly older as a result of migration -10.6 -3.5
-4.5 -4.2 -7.6
(Figure 8). At TA level, 36 percent had -20.0 -4.5
-12.1
populations that were between 20 and 50
percent older because of migration, and -40.0
31 percent, between 50 and 100 per cent Territorial Authority Areas (N-66) Towns (N=143) Rural Centres (N=132)
older. By contrast, just one was more than 50% younger
50 percent younger (Wellington City). 100% older
Almost 30 percent of New Zealand towns Source: Jackson and Brabyn (forthcoming).
had populations between 20 and 50
percent older as the result of migration; percent of rural centres (and 15 percent of gains of those of parenting age and thus
36 percent, between 50 and 100 percent TAs) had greater than 20 percent of their also their children – and reinforced by loss
older; and 10 percent, more than 100 respective populations aged 65+ years. at the oldest ages. By contrast, the towns
percent older. Rural centres had larger As a result, towns are on the one hand experiencing the greatest ageing effect
proportions than towns that were both currently less likely than rural centres from migration all see net losses at 15-19
more than 100 percent older (13.6 to be experiencing natural decrease (as or 20-24 years, and make their gains at the
percent) and more than 50 percent noted above, in 2013, 13 and 15 percent older ages, particularly ‘early retiree’ age.
younger (4 percent compared with 3 respectively), but on the other, seeing a The reverse effect from age 75 is also very
percent for towns), as the result of more rapid shift to that situation. As both clear, reflecting the classic move many
migration. Rural centres also had greater Pool and Brabyn (this issue) propose, older people make ‘back’ towards health
proportions in each of the ‘younger with tertiary education and jobs are attracting services and/or family.
migration’ groups. young people to the larger cities, while Although not studied in detail,
lifestyle and amenity factors are attracting aggregate data for rural districts suggests
Observation 7: Towns are more likely than older people from both the rural centres an age migration profile that falls between
rural centres to have more than 20 percent and the larger towns to the smaller towns. the two extremes. On the one hand the
aged 65+ years data show consistent net loss at 15-19 and
The finding of greater proportions of rural Observation 8: Age-selective migration is 20-24 years, but on the other, in most
centres than towns being younger as a accelerating structural population ageing in years, net gains at the main parental and
consequence of net migration is one of the most areas child ages, offset by net loss at age 65 and
more surprising findings of this project. Figure 9 shows the age distribution of above. Similar to Rolleston and
We had theorised that rural centres would migration for the five towns experiencing Arrowtown, this combination leads to a
be more affected by the loss of young the greatest juvenescent impact of relatively low percentage aged 65+ years,
adults and have greater levels ‘ageing- migration (left figures) and the five and further indicates that the structural
in-place’, and this combination would towns experiencing the greatest ageing ageing of rural districts is primarily due to
have caused those areas to age faster. The impact (right figures). The respective left- ageing-in-place, albeit in some cases the
explanation is that while the populations and right-hand skews clearly illustrate ageing of past migrants, but not direct in-
of rural centres are, on average, older than the difference in impact. Queenstown, migration at older ages.
those of towns (in 2013, 17 and 14 percent Rolleston, Wellington, Arrowtown and As also proposed, these migration age
aged 65+ years respectively), there is a Central Auckland all gain a disproportion profiles are typically altered by natural
greater proportion of towns with more of migrants at the younger ages, albeit increase, both the net difference between
than 20 percent aged 65 years (Jackson et Rolleston and Arrowtown experience births and deaths, and change in the size
al., forthcoming, Tables 3 and 4).3 In 2013, minor net loss at 20-24 and 15-19 years of individual cohorts as they age. That is,
41 percent of towns compared with 29 respectively, but this is offset by substantial when larger cohorts take the place of
Policy Quarterly – Volume 13, Supplementary Issue – June 2017 – Page 29Age Distribution (%) Age Distribution (%) Age Distribution (%) Age Distribution (%) Age Distribution (%)
-5.0
0.0
5.0
10.0
15.0
20.0
25.0
-4.0
-2.0
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
-400
-200
0
200
400
600
800
1000
-20.0
-10.0
0.0
10.0
20.0
30.0
40.0
50.0
60.0
-300.0
-200.0
-100.0
0.0
100.0
200.0
300.0
400.0
Rolleston
0-4
Arrowtown
0-4 0-4
0-4
Queenstown
0-4
5-9 5-9 5-9
5-9
Wellington Zone
5-9
10-14 10-14 10-14
10-14 10-14
15-19 15-19 15-19
Central Auckland Zone
15-19 15-19
20-24 20-24 20-24
20-24 20-24
Notes: Different scales on Y-axis
25-29 25-29 25-29
25-29 25-29
30-34 30-34 30-34
30-34 30-34
35-39 35-39 35-39
35-39 35-39
40-44 40-44 40-44 40-44
40-44
45-49 45-49 45-49 45-49
45-49
50-54 50-54 50-54 50-54
50-54
55-59 55-59 55-59 55-59
55-59
60-64 60-64 60-64 60-64 60-64
65-69 65-69 65-69 65-69 65-69
70-74 70-74 70-74 70-74
Towns experiencing the greatest youthful impact from migration
70-74
75-79 75-79 75-79 75-79 75-79
80-84 80-84 80-84 80-84 80-84
Page 30 – Policy Quarterly – Volume 13, Supplementary Issue – June 2017
85-89 85-89 85-89 85-89 85-89
90-94 90-94 90-94 90-94 90-94
95+ 95+ 95+ 95+ 95+
Age Distribution (%) Age Distribution (%) Age Distribution (%) Age Distribution (%) Age Distribution (%)
Tairua
-15.0
-10.0
-5.0
0.0
5.0
10.0
15.0
20.0
-10.0
-5.0
0.0
5.0
10.0
15.0
20.0
25.0
-10.0
-5.0
0.0
5.0
10.0
15.0
20.0
25.0
30.0
-20.0
-15.0
-10.0
-5.0
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
Waimate
-300.0
-250.0
-200.0
-150.0
-100.0
-50.0
0.0
50.0
100.0
150.0
200.0
0-4 0-4 0-4
0-4 0-4
Whangamata
5-9 5-9 5-9
5-9 5-9
10-14 10-14 10-14
Mangawhai Heads
10-14 10-14
Katikati Community
15-19 15-19 15-19
15-19 15-19
20-24 20-24 20-24
20-24
The mechanisms of subnational population growth and decline in New Zealand 1976-2013
20-24
25-29 25-29 25-29
25-29 25-29
30-34 30-34 30-34
30-34 30-34
35-39 35-39 35-39
35-39 35-39
40-44 40-44 40-44
40-44 40-44
45-49 45-49 45-49
45-49 45-49
impact and the five towns experiencing the greatest ageing impact from net migration, 1976-2013
50-54 50-54 50-54 50-54
50-54
55-59 55-59 55-59 55-59
55-59
60-64 60-64 60-64 60-64
60-64
65-69 65-69 65-69 65-69
65-69
70-74 70-74 70-74 70-74 70-74
Towns experiencing the greatest ageing impact from migration
75-79 75-79 75-79 75-79 75-79
80-84 80-84 80-84 80-84 80-84
Figure 9: Net migration age distribution (percentage of migrants at each age) for the five towns experiencing the greatest youthful
85-89 85-89 85-89 85-89 85-89
90-94 90-94 90-94 90-94 90-94
95+ 95+ 95+ 95+
95+smaller ones, and vice-versa, the net percent, accelerating since 2001 (+29 towns the relationship is negative (the
change for those age groups may be less or percent). Rural centres are thus more higher the natural increase, the lower
greater than indicated by migration alone. vulnerable than towns to the loss of the net change, and vice-versa), but it is
However for the most part, it is the natural increase via decreased births, not statistically significant after 1996. For
concentration of migration at either while towns are becoming more rural centres the relationship is significant
younger or older ages, and particularly the vulnerable to it from increased deaths. at only three observations, and it is
level of net out-migration around age 15- Mirroring the trends at 15-44 years is positive (the higher the natural increase,
24 years, and/or in-migration at 50+ the equally strong but negative correlation the higher the net gain).
years, that determines the ‘speed’ of between the percentage aged 65+ years By contrast, the correlation between
structural ageing. and natural increase7, indicating that the net growth and net migration gain is both
higher the percentage aged 65+ years, the very strong and positive, and never below
Observation 9: The proportion of women lower the natural increase. Although to +0.92 (p=0.01 for all observations, for
aged 15-44 is a stronger driver of natural some extent auto-correlated with the both towns and rural centres). Thus, areas
increase or decrease than the total fertility proportion of women aged 15-44 years, with higher net migration gain have
rate this relationship was similarly found for higher growth – and vice-versa.
Conventionally, the shift from natural the United States and European counties The analysis confirms that net
increase to natural decrease is associated by Johnson et al., (2015: 665-666). The migration is the stronger ‘predictor’ of net
with a total fertility rate (TFR)4 of less relationship is somewhat stronger for change; however, the foregoing analyses
than 2.1 births per woman, for around a
generation (c. 25 years). New Zealand’s
TFR is still around 2 births per woman, What has been missing from the debate
although at TA level it ranges from 1.5
for Queenstown-Lakes District to 3.3 is an understanding of the widespread
for Opotiki District (Statistics New
Zealand 2016). However, reflecting the
nature of decline across the country, and
age-selective migration patterns above, in particular, an understanding that it is
we found a much stronger relationship5
between natural increase and the not limited to small rural towns.
percentage of women aged 15-44 years6,
than with the TFR (Jackson et al.,
forthcoming). Although we have reliable
subnational TFR data for only three
observations, our findings concur with
Johnson et al., (2015: 667-669), whose New Zealand’s towns than rural centres, demand that a caveat be added: only when
study on the counties of the United States reflecting the finding that many towns migration is examined as a discrete
and Europe showed that the higher/lower have relatively old age structures, with variable.
the proportion of women at the main proportions at older ages presumably Moreover, another important finding
childbearing ages, the more/less likely it increasing the number of deaths. While a from the correlation analysis is that net
is for them to sustain the natural increase greater proportion of towns than rural migration and natural increase are
required to offset migration loss. centres have been experiencing natural negatively correlated. That is, the higher
Also important is our finding that the increase, the proportion is falling at a the net migration gain, the lower the
national percentage of women aged 15-44 faster rate than for rural centres. natural increase, and vice-versa. For towns
years (for New Zealand) peaked in 1991 at the correlation is statistically significant
46.6 percent and has since fallen to 39.8 Observation 10: Net migration explains most (p=0.01) at all but one observation.
percent, the latter almost identical to that of the variance in net change, but natural Although more modest for rural centres,
for the USA and Europe. Unless fertility increase determines whether the outcome is and statistically significant at just three
rates rise substantially, diminishing positive or negative observations (p=0.01), the negative
proportions of women at reproductive We found that net migration explains correlation is a salutary finding, as it
age mean diminishing birth numbers. The around 95 percent of the variance in the appears that migration does not
proportion of towns with lower than the net population change of towns and rural necessarily mean ‘more births’, as is often
national proportion of women in these centres across the period 1976-2013. This believed. Rather, it may reflect recent
childbearing age groups increased from is because natural increase is somewhat trends in the composition of migrants,
77 percent in 1976 to 87 percent in 2013 more even across these jurisdictions, and from primarily families, to primarily non-
(+14 percent). For rural centres the is mostly positive. At the same time, we family workers (students and others on
proportion (lower than the national found that natural increase has a relatively temporary visas and/or early retirees
average) has increased from 70 to 90 weak relationship with net change. For moving for lifestyle reasons). That is, not
Policy Quarterly – Volume 13, Supplementary Issue – June 2017 – Page 31The mechanisms of subnational population growth and decline in New Zealand 1976-2013
only is migration not a panacea for glaringly evident over the next few and projections at TA level indicate that it
growth, but it may further hasten the end decades, as natural increase gives way to will become increasingly prevalent from
of natural increase, a finding also noted ageing-driven natural decrease. As Brabyn 2028. By 2043 it is projected to affect
for the United States and Europe by (infra) argues, the contribution of natural around 27 percent of TAs, and to be the
Johnson et al., (2015) and for Europe over increase to population change has thus far largest cause of depopulation and the
a very long period by Murphy (2016: 239). been relatively even across New Zealand, single largest cause of population change
with the result that migration accounts per se. Type A growth will by then have
Summary and discussion for 95 percent of variation. However, as diminished to around 17 percent of TAs.
Over the past few years, population change we show in this article, the actual level of Answering our second research question,
in New Zealand has attracted much media growth or decline is heavily dependent on the various combinations delivering
and political attention. Areas experiencing whether natural increase augments or population growth and decline also do more
rapid growth have been contrasted with offsets that net migration gain or loss. than simply affect population size; they also
so-called ‘zombie towns’ (NBR 2014), We have shown here that there are have a profound effect on the age structure
the latter usually singling out a few rural several combinations of natural increase of each population. Net loss at younger
towns that have experienced precipitous and net migration, three of which deliver ages typically makes age structures older,
decline. What has been missing from growth, three decline, and two, zero and net gain, younger. Net in-migration at
the debate is an understanding of the growth. The distinction between the older ages further accelerates – or, where
negative, slows – structural ageing. This
... the question as to whether the
‘age-selective’ migration has resulted in
just 15 percent of New Zealand towns
ending of population growth and onset having populations in 2013 that were
younger as the result of migration across
of depopulation will be ‘good’ or ‘bad’ the period 1976-2013; conversely, 85
has as yet been subjected to very percent of towns were older. The latter
situation also pertains to two-thirds of
little serious research – or theoretical rural centres and four-fifth of TAs. Only a
minority of our study areas have thus
development ... become more youthful as the result of
migration, a finding which, at least in
terms of international migration, runs
counter to many pronouncements on the
widespread nature of decline across growth and decline types is important for topic. Moreover, in many cases areas are
the country, and in particular, an policy purposes. Type A growth, where substantially older (or, in fewer cases,
understanding that it is not limited to both elements are positive, is somewhat substantially younger).
small rural towns. Both Dunedin and more robust than Type B growth, where Again, these effects have policy
Invercargill cities, for example, are smaller natural increase conceals net migration implications, as areas gaining young
now than they were in 1976, despite their loss, and the two should not be conflated. adults (especially from internal migration)
annual gains of many of the nation’s As natural decrease unfolds, areas are at the same time playing a role in the
tertiary education students. Also missing currently growing from Type B growth ageing of the towns and rural centres they
from the commentary is that in many will become increasingly vulnerable to have left. Akin to ‘watering the neighbour’s
cases, growing towns and rural centres overall decline. Type E decline, on the garden’ (Attané & Guilmoto 2007), future
are growing only, or largely, because of other hand, where natural increase fails to policy may need to consider having
natural increase – the difference between conceal net migration loss (the widely younger/growing areas compensate their
births and deaths; this component experienced, ‘old’ type of decline), has the ageing/declining counterparts. This
is regularly rendered invisible. Even potential to revert to Type B growth, if proposition is strongly supported by
Auckland’s growth has been primarily the migration in those areas turns positive. correlations that show the relative size of
result of natural increase, accounting for Type E decline is also less perverse than the population of women aged 15-44
58 percent of the region’s growth over the the ‘new’ type of depopulation, Type F years is a critical factor in sustaining
past 25 years (Jackson 2016). decline, where there is both net migration natural increase. Second only to the strong
The widespread perception of loss and natural decrease. Answering our positive correlation between net migration
migration as the primary driver of growth main research question, we found that this and net change, the correlation for women
and decline has diverted attention away new, ‘dual’ form of depopulation is as yet aged 15-44 years and natural increase
from the all-important population affecting a relatively small number of New shows that the higher the former, the
replenishing role of natural increase (see Zealand’s towns and rural centres (13-15 higher the latter. Conversely, the lower the
also Cochrane & Pool, infra, on the critical percent), and is not yet evident at TA level. percentage of women at these ages, the
role of Mäori), a deficit that will become However, it is emerging intermittently, lower the natural increase. This very
Page 32 – Policy Quarterly – Volume 13, Supplementary Issue – June 2017strong correlation, which concurs with Zealand’s relatively high birth rates and age considerably (Atoh 2000; Lutz,
that found for the United States and currently very high per capita levels of net Sanderson & Sherbov 2004; Reher 2007,
Europe (Johnson et al., 2016), has also international migration may suggest to 2011; World Bank 2009; Haartsen &
strengthened over time, and more so for some that the ageing-driven ending of Venhorst 2009; Audirac 2010; Lee & Reher
rural centres, where the loss of women population growth is ‘over the horizon’, 2011; Matanle and Rausch 2011;
aged 15-44 years is greater than for towns. and thus may be paid less attention in the Martinez-Fernandez et al., 2012, among
Moreover, the correlation between short term. We emphasise that this is not many others). In many cases there is the
women aged 15-44 years and natural the case at the subnational level (see also potential for a ‘depopulation dividend’, if
increase is also substantially stronger than Matanle 2017, which compares New the situation is accepted and engaged with
that between natural increase and the Zealand with Japan). Although the decline positively and in a timely manner
total fertility rate, for both towns and experienced across the period 1976-2013 (Matanle 2017). Accordingly, it is time for
rural centres – although we have only was disproportionately the result of net our policy makers to begin revisiting our
three observations on which to base this migration loss, that loss and its age- policies and the (growth) principles on
proposition. While also positive, selective nature has greatly accelerated the which most are based (United Nations
indicating that the higher/lower the TFR, structural ageing of affected areas. These 2000:4).
the higher/lower the natural increase, the areas will struggle to return to long-term
1 These and all projections used in this article are based on
relatively weak correlation values indicate population growth, as the increased Statistics New Zealand (2015e). They thus predate the
that there is as yet very little relationship proportions at older ages and decreased latest ‘2013-base - 2043 Update’ (2017) projections which
became available after the article was written. Comparison
between the two. Accordingly, we can at proportions at reproductive age afford indicates that population growth will be slightly higher in the
first decade of the projection, but thereafter the outcomes are
least tentatively conclude that age- little chance of resurrecting the natural almost identical to those presented here.
selective migration is the major driver of increase they once enjoyed. As natural 2 It should be recalled that both ‘with’ and ‘without’ migration
includes natural increase.
New Zealand’s current shift to natural increase diminishes, more and more 3 In 2013, the percentage of the usually resident population
aged 65+ years ranged, for towns, from 3 percent (Waiouru,
decrease, rather than the conventional migrants will be required simply to largely a military camp area) to 36 percent (Tairua, a North
driver, low fertility. maintain each population at the same Island east coast beach town), and for rural centres from 0.5
percent (Burnham Military Camp) to 44 percent (Pauanui,
This finding is further supported in size. This may happen for popular also a North Island east coast beach town, immediately
another, even more surprising finding, retirement towns, but not most towns. adjacent to Tairua).
4 The Total Fertility Rate is a ‘synthetic’ measure of the average
which shows that net migration and Ageing and declining areas thus require number of children a woman would have across her lifetime
if she were to experience all of the age-specific fertility rates
natural increase are negatively correlated. regionally-specific migration policies that occurring in that particular year.
Although the correlations are relatively give primacy to local, rather than national, 5 In order to summarise the patterns and trends we examined
the relative strength of key relationships, by applying the
weak, they are consistently negative and demographics. Pearson Correlation Coefficient ‘r’ to various combinations
of the data. The Pearson Correlation Coefficient measures
indicate that the higher the net migration, Finally, the question as to whether the the linear strength of the relationship between two arrays of
the lower the natural increase, and vice- ending of population growth and onset of data, with +1.00 meaning that each item moved in exactly
the same direction at the same rate of change (whether
versa – a finding also reported for the USA depopulation will be ‘good’ or ‘bad’ has as positively or negatively), and -1.00 meaning that each item
and Europe (Johnson et al., 2015; Murphy yet been subjected to very little serious moved in the opposite direction.
6 p(0.01) for all observations for both towns and rural centres.
2016). Migration may thus not be the research – or theoretical development (see By contrast the TFR was significant at either 0.01 or 0.05
for two of the three observations, for both towns and rural
‘bringer of babies’ that many believe. Matanle, infra). The end of population centres.
These findings are important from a growth is likely to have both positive and 7 p(0.01) at all observations for both towns and rural centres.
8 Description: The baseline data was created by Dave Maré
migration policy perspective, not least negative outcomes – positive, for example, (Motu Research) under microdata access agreement with
Statistics New Zealand, MAA2003/18. dave.mare@motu.
because they have implications for the in terms of fewer people to consume org.nz. The tables contain counts of the 1976, 1981,
ideal composition of international resources and damage the environment, 1986, 1991, 1996, 2001, 2006 and 2013 usual resident
population by age and sex, grouped by 2013 geographic
migrants; that is, perhaps low growth/ and potentially the development of a area boundaries (Territorial Authority and Urban Area). The
declining areas need more who are likely greater sense of community. However, its Urban Area classification has been extended to identify rural
centres (ua13=501) separately (using 2013 Area Unit
to have families, and fewer merely to negative outcomes will include the loss of codes).
9 Disclaimer: Access to the data was provided by Statistics
‘work’, especially in rural centres. many economies of scale that have been New Zealand under conditions designed to give effect to the
Similarly, where areas are gaining enjoyed in the growth phase, and plausibly security and confidentiality provisions of the Statistics Act
1975. The results presented in these tables are the work of
older people, either through migration or the loss of value of many assets, such as the authors, not Statistics New Zealand. They are not ‘official
statistics’.
ageing-in-place, or alternatively are losing housing. In New Zealand it will have
them to other areas, there are social policy major implications for the way in which
implications in terms of the type of rate revenue has historically been gathered Acknowledgment
resources and services that are needed; by local government councils (Jackson We wish to thank Dr Anne Pomeroy and
one-size-fits-all policies are to be avoided. 2004; Jackson & Cameron 2017). Dr Etienne Nels, both of the University
The same point pertains to areas gaining Ultimately the ending of growth and of Otago, for their most helpful and
or losing families and children; their onset of depopulation is expected to be insightful comments on an earlier version
needs are very different. the dominant situation across most of this article. Any remaining shortfalls
In terms of policy development countries of the developed world by mid- are our own.
regarding structural ageing per se, New century; prior to that our populations will
Policy Quarterly – Volume 13, Supplementary Issue – June 2017 – Page 33The mechanisms of subnational population growth and decline in New Zealand 1976-2013 References Atoh, M (2000) ‘The coming of the hyper-aged and depopulating society Lutz, W, W Sanderson & S Sherbov (eds) (2004) The End of World and population policies: the case of Japan’, Conference Proceedings Population Growth in the 21st Century: New challenges for human New York: Department of Economic and Social Affairs capital formation and sustainable development London & Sterling VA: Attané, I & CZ Guilmoto (2007) Watering the Neighbours Garden: The International Institute for Applied Systems Analysis/Earthscan growing female demographic deficit in Asia. Paris: CICRED Martinez-Fernandez, C, N Kubo, A Noya & T Weyman (2012) Audirac, I (2012) ‘Declining suburbs in Europe and Latin America’, Demographic change and local development: shrinkage, regeneration International Journal of Urban and Regional Research, pp.226-244 and social dynamics, OECD/LEED Working Paper Series. Bucher, H & R Mai (2005) Depopulation and its Consequences for Matanle, P (2017) ‘Towards an Asia-Pacific depopulation dividend in the Regions of Europe. Report Prepared for the Council of Europe, the 21st Century. Regional growth and shrinkage in Japan and New Directorate General III–Social Cohesion DG3/CAHP10(2005) 7 final Zealand’, The Asia-Pacific Journal 15(6), p.5 Cameron, M (2017) ‘’The relative (un)certainty of subnational population Matanle, P & A Rausch with the Shrinking Regions Group (2011) Japan’s decline’ Policy Quarterly, Supplementary Issue, pp.55-60 Shrinking Regions in the 21st Century: Contemporary responses to Davis, K (1945) ‘The world demographic transition’, The Annals of the depopulation and socioeconomic decline Amherst, NY: Cambria Press American Academy of Political and Social Science 237, pp.1-11 Matanle P & Y Sato (2010) ‘Coming soon to a city near you! Learning Davoudi, S, M Wishardt & I Strange (2010) ‘The ageing of Europe: to live ‘beyond growth’ in Japan’s shrinking regions’, Social Science demographic scenarios of Europe’s futures’, Futures 42, pp.794-803 Japan Journal 13(2), pp.187-210 Dyson, T (2011) ‘The role of the demographic transition in the process of McDonald, P & R Kippen (1999) The impact of immigration on the ageing urbanization’, Population and Development Review 37, pp.34-54 of Australia’s population Canberra: Department of Immigration and Haartsen, T & V Venhorst (2009) ‘Planning for decline: anticipating on Multicultural Affairs population decline in the Netherlands’ Econom. en Social Geography McMillan, R (2016) ‘The shrinkage pathway: managing regional 101(2), pp.218-227 depopulation’, in P Spoonley (ed) Rebooting the Regions: How low or Jackson, NO (2014) ‘Sub-national depopulation in search of a theory – zero growth needn’t mean the end of prosperity pp.213-237, Albany, towards a diagnostic framework’, New Zealand Population Review 40, NZ: Massey University Press pp.3-39 Murphy, M (2016) ‘The impact of migration on long-term European Jackson, NO (2016) ‘Irresistible forces: facing up to demographic change’, population trends, 1850 to present’, Population and Development In P Spoonley (ed) Rebooting the Regions. Why low or zero growth Review 42(2), pp.225-244 needn’t mean the end of prosperity pp.47-77, Auckland: Massey National Business Review [NBR] (2014, 12 July) ‘NZ has zombie towns University Press that need to close — economist’, (retrieved 26 March 2017 from: Jackson, NO & L Brabyn (forthcoming) ‘Population ageing and age- https://www.nbr.co.nz/article/nz-has-zombie-towns-need-close- selective migration. Subnational population trends in New Zealand’. %E2%80%94-economist-ns-159124) Draft available from authors on request Reher DS (2007) ‘Towards long-term population decline: A discussion of Jackson, NO, L Brabyn & D Maré (2016) ‘New Zealand’s towns and rural relevant issues’, European Journal of Population 23, pp.189-207 centres 1976-2013 – experimental components of growth’ Working Reher DS (2011) ‘Economic and social Implications of the demographic Paper 7. National Institute of Demographic and Economic Analysis, transition’, Population and Development Review 37, pp.11-33 University of Waikato, Hamilton, New Zealand Rowland, DT (2003) Demographic Methods and Concepts Oxford: Oxford Jackson, NO, L Brabyn, DC Maré, MC Cameron & I Pool (forthcoming, University Press. under review) ‘From ageing-driven growth towards the ending of Statistics New Zealand (2015a) Life Table functions 2005-07 and 2012- growth. Subnational population trends in New Zealand’ 14 by territorial authority area, age and sex Jackson, NO & MC Cameron (2017) ‘The unavoidable nature of population Statistics New Zealand (2015b) Subnational population by age and sex, ageing and the ageing-driven end of growth - an update for New June 1996, 2001, 2006, 2013 Zealand’ Population Ageing: (tba), 1-26. DOI: 10.1007/s12062-017- Statistics New Zealand (2015c) Live births per 1,000 women by age, 9180-8 Mäori and Total, Table DFM019AA Jackson, NO, MC Cameron & I Pool (2015) The mechanisms of Statistics New Zealand (2015d) Number of births, Mäori and Total, by subnational depopulation – towards a typology Presentation to ‘Our territorial authority area, Table VSB011AA People, Our Places’ Population Association of New Zealand Biennial Statistics New Zealand (2015e). Subnational population projections by age Conference, Hamilton June 29-30 and sex, 2013(base)–2043 Jackson, NO & I Pool (forthcoming) A Theory of Replacement Capacity Statistics New Zealand (2016) Customised database. Estimated birth Johnson, KM, LM Field & DL Poston Jnr (2015) ‘More deaths than occurrence at 30 June 1997-2013 by age of mother and territorial births: subnational natural decrease in Europe and the United States’, authority area of usual residence based on 2013 boundaries. Population and Development Review 41(4) pp.651-680 United Nations (2000) Replacement migration. Is it a solution to Kippen, R & P McDonald (2000) ‘Australia’s population in 2000: the way declining or ageing populations? UN Secretariat, Department of we are and the ways we might have been’ People and Place 8(3), Economic and Social Affairs pp.10-17 World Bank (2009) Reshaping Economic Geography chapter 3, Lee, RD & DS Reher (2011) ‘Introduction: the landscape of the Washington DC: World Bank demographic transition and its aftermath’, Population and Development Review 37(1), 1-7 Page 34 – Policy Quarterly – Volume 13, Supplementary Issue – June 2017
You can also read
