Indigenous Australians and the National Disability Insurance Scheme
Appendix 1: Projection methodology for Remote Service Delivery Areas
Population projections: Overview
A variant of the standard Cohort Component Projection model is used for the projection of the Indigenous population of each Remote Service Delivery Area (RSDA). The necessary inputs to the model include hazard rates of births (fertility), deaths (mortality), and mobility (net migration). With estimates of projected fertility, mortality and migration, and base population estimates by five year increments in age for group and sex, the population of each region is projected over a 20year period (2006–2026) using the standard cohortcomponent method. The population aged 5 years and over and the population aged under 5 years are calculated as follows:
where
 • A(r,g,x,y) is the population in region r of sex g aged x in year y
 • A(r,f,x,y) is the female population in region r aged x in year y
 • b(r,x,y) is the fertility rate at age x in year y
 • m(r,g,x,y) is the migration ratio in region r for sex g; that is, the factor by which a cohort changes through migration in its transition from age x in year y to age x+5 in year y+5
 • s(r,g,x,y) is the mortality survival ratio in region r for sex g; that is, the probability that a person aged x in year y will survive to age x+5 in year y+5, and
 • SR(g) is the proportion of births that are of sex g.
The standard model above is altered to enable: (1) the inclusion of Indigenous and nonIndigenous births (see Chapter 1); and (2) the solving of a series of linear equations to estimate the effect of a 10 per cent diagonal migration scenario (see Chapter 1). This alteration to the standard model results in a simultaneous population projection of three populations. The first is the projection of the RSDA population of Indigenous persons only. The second is the projection of the RSDA population of nonIndigenous females only. The final comprises projections of survivors of the Indigenous population projections.
Four sets of scenarios are included in this modelling framework. Combining the migration, fertility and mortality scenarios leads to projection scenarios for each region shown in Table A1.1
Table A1.1 Indigenous population projection scenarios
Scenario ID 
Migration 
Fertility 
Mortality 
1 
0 
Constant 
Constant 
2 
10% diagonal 
Constant 
Constant 
3 
0 
25year convergence 
25year convergence 
4 
10% diagonal 
25year convergence 
25year convergence 
The remaining components of the projection methodology are outlined below.
Components of population change
Creating baseline Estimated Resident Population by sex by age by Indigenous status
As only total population counts are available, it is necessary to disaggregate these counts by age, sex and Indigenous status. Estimation of the age structure is imputed from ABS estimates of the age structure of the broader Indigenous region of which each RSDA is a member (ABS 2008a). A further complication is that Estimated Resident Population (ERP) counts by Indigenous Region (IREG) are upper censored to age 65+. To maintain the heterogeneity in the mortality estimates from the ABS Life Tables (ABS 2009b, 2009c), the upper censoring is split into 5year age groups 65–85+ years. This is estimated using counts of Indigenous/nonIndigenous status by age at the relevant State level (ABS 2008b). ERP estimates are for June 2006.
Births
The projection of births by RSDA requires three separate estimates of fertility: (1) Indigenous births to Indigenous mothers; (2) Indigenous births to nonIndigenous mothers; and (3) nonIndigenous female births to nonIndigenous mothers. The final fertility estimate is necessary for female births only, as the Indigenous projections only require the input of total births from the nonIndigenous population. An initial investigation was undertaken to see whether areaspecific births data could be used to calculate the relevant fertility rates. Doing so resulted in too large a variation to be driven by fertility differences, and hence Statelevel births data was used for each of the three measures of fertility outlined above. To smooth out irregularities in the data, three yearaveraged fertility rates are used based on data for 2005–2007. Two fertility assumptions were used in the projections. The first series keeps fertility rates (Indigenous and nonIndigenous) constant throughout the projection period. The second series results in a convergence of Indigenous and nonIndigenous fertility over a 25year time frame. That is, the convergence occurs outside the projection period. The convergence results in the same fertility level (Total Fertility Rate) and well as fertility probability distribution (Age Specific Fertility Rates).
Deaths
State level Life Tables were used to calculate fiveyear survival ratios for both Indigenous and nonIndigenous populations (ABS 2009b, 2009c). For the projections, two scenarios are used: (1) Keeping nonIndigenous survival ratios at 2006 levels over the full projection period; and (2) converging the agespecific survival ratios of Indigenous and nonIndigenous persons over a 25year period. The Indigenous survival ratios are increased linearly to the nonIndigenous rates over this time. That is, the average rate of change is constant over the projection period.
Migration
The final inputs required for the projection are estimates of migration, both in terms of level and age distribution. Two migration scenarios are used in the projections. The first sets net migration to zero for both the Indigenous and nonIndigenous population projections. The second assumption is more complicated, and is set out below.
This assumption results in a 10 per cent diagonal increase in the Indigenous population of each age cohort (excluding births), above that which would have been achieved with zero migration. As an example, Table A1.2 displays two age groups, 5–9 and 10–14 in 2006 and 2011. The ratio B/A represents the survival of the 5–9 cohort to 10–14 years in 2011, assuming no migration. To calculate the migration assumption, B/A is increased by 10 per cent (i.e. B/A + 0.1). An iterative linear search algorithm is then used to estimate the level of migration (net of survival) necessary to increase the 2011 population aged 10–14 by approximately 10 per cent. This approach has the advantage of imputing both an age structure of migration and an absolute level of migration. This is particularly pertinent given the poor data quality at the RSDA level. The estimated migrants are projected as a subset of the population for future years, subject to both population decline through survival and population increase through fertility.
Table A1.2 Example migration table
2006 
2011 

5–9 
A 

10–14 
B 
Summary of results
The projection results are summarised by grouping all ages together in Tables A1.3, A1.4 and A1.5. In Table A1.3, the estimated nonIndigenous population in 2006 is given, as well as a projected population in 2026, calculated by applying the nonIndigenous annualised growth rate of 0.96 per cent per annum from Biddle and Taylor (2009). These are presented alongside the estimated Indigenous population in 2006 as well as the projected Indigenous population in 2026 for each of the four scenarios.
In Table A1.4, these population estimates/projections are used to calculate a projected annualised growth rate between 2006 and 2026. These are presented alongside the projected growth rates between 2006 and 2026 for the Indigenous Region that the community is located in, based on the zero migration scenario in Biddle and Taylor (2009). The main point to note in Table A1.4 is that in the absence of migration, all communities have a projected growth rate that is less than that for the region as a whole. This is driven mainly by the higher levels of births of Indigenous children to nonIndigenous mothers in the less remote parts of the regions which, demographically, dominate the results for the region as a whole. It is important to note, however, that the projected growth rates in the RSDAs are comparable to those for remote and very remote Australia in Biddle and Taylor (2009)—1.74 and 1.63 per cent per annum respectively—and larger than the nonIndigenous growth rate (0.96).
The proportion (percentage) of the population in the area estimated/projected to be Indigenous in 2006 and 2026 is given in Table A1.5.
Table A1.3 Estimated/projected nonIndigenous and Indigenous population in 2006 and 2026 by RSDA
NonIndigenous 
Indigenous 

Region Name 
2006 
2026 
2006 
2026S1 
2026S2 
2026S3 
2026S4 
Amata 
29 
35 
341 
462 
663 
455 
657 
Angurugu 
33 
40 
1 013 
1 372 
1 972 
1 356 
1 958 
Ardyaloon 
31 
38 
243 
328 
470 
324 
467 
Aurukun 
70 
85 
1 059 
1 451 
2 081 
1 423 
2 051 
Beagle Bay 
24 
29 
238 
320 
459 
317 
456 
Coen 
60 
73 
239 
330 
473 
324 
466 
Doomadgee 
61 
74 
1 102 
1 546 
2 211 
1 507 
2 168 
Fitzroy Crossing 
444 
537 
733 
1 028 
1 467 
1 014 
1 454 
Galiwinku 
147 
178 
2 158 
2 930 
4 208 
2 895 
4 178 
Gapuwiyak 
52 
63 
1 208 
1 637 
2 352 
1 618 
2 336 
Gunbalanya 
88 
107 
1 141 
1 584 
2 271 
1 556 
2 243 
Halls Creek 
261 
316 
1 092 
1 515 
2 168 
1 492 
2 145 
Hermannsburg 
68 
82 
938 
1 261 
1 811 
1 246 
1 798 
Hope Vale 
45 
54 
797 
1 092 
1 565 
1 071 
1 543 
Lajamanu 
105 
127 
735 
1 010 
1 446 
991 
1 427 
Maningrida 
176 
213 
2 600 
3 610 
5 174 
3 545 
5 111 
Milingimbi 
49 
59 
1 086 
1 506 
2 159 
1 479 
2 132 
Mimili 
36 
44 
289 
393 
563 
387 
558 
Mornington Island 
94 
114 
1 028 
1 446 
2 067 
1 409 
2 026 
Mossman Gorge 
0 
0 
165 
233 
333 
227 
327 
Nguiu 
85 
103 
1 463 
2 031 
2 911 
1 994 
2 875 
Ngukurr 
73 
88 
1 055 
1 446 
2 072 
1 418 
2 043 
Numbulwar 
64 
77 
713 
968 
1 391 
957 
1 381 
Umbakumba 
21 
25 
434 
589 
845 
582 
839 
Wadeye 
146 
177 
2 074 
2 880 
4 128 
2 828 
4 077 
Walgett 
1 002 
1 213 
1 220 
1 748 
2 488 
1 696 
2 429 
Wilcania 
154 
186 
453 
637 
910 
618 
888 
Yirrkala 
212 
257 
1 472 
2 005 
2 877 
1 981 
2 857 
Yuendumu 
92 
111 
701 
946 
1 358 
935 
1 348 
Source: Authors’ own calculation; Biddle & Taylor 2009
Table A1.4 Projected annual Australian Indigenous growth rates between 2006 and 2026 by RSDA and IREG
Remote Service Delivery Area 
Indigenous Region 

Region name 
Scenario 1 
Scenario 2 
Scenario 3 
Scenario 4 
Name 
Growth rate 
Amata 
1.53 
3.38 
1.45 
3.33 
Port Augusta 
1.75 
Angurugu 
1.53 
3.39 
1.47 
3.35 
Nhulunbuy 
1.69 
Ardyaloon 
1.51 
3.35 
1.45 
3.32 
Broome 
1.82 
Aurukun 
1.59 
3.43 
1.49 
3.36 
Cape York 
1.79 
Beagle Bay 
1.50 
3.35 
1.44 
3.32 
Broome 
1.82 
Coen 
1.63 
3.46 
1.53 
3.39 
Cape York 
1.79 
Doomadgee 
1.71 
3.54 
1.58 
3.44 
Mt Isa 
1.93 
Fitzroy Crossing 
1.71 
3.53 
1.63 
3.48 
Derby 
1.89 
Galiwinku 
1.54 
3.39 
1.48 
3.36 
Nhulunbuy 
1.69 
Gapuwiyak 
1.53 
3.39 
1.47 
3.35 
Nhulunbuy 
1.69 
Gunbalanya 
1.66 
3.50 
1.56 
3.44 
Jabiru 
1.79 
Halls Creek 
1.65 
3.49 
1.57 
3.43 
Kununurra 
1.92 
Hermannsburg 
1.49 
3.34 
1.43 
3.30 
Apatula 
1.63 
Hope Vale 
1.59 
3.43 
1.49 
3.36 
Cape York 
1.79 
Lajamanu 
1.61 
3.45 
1.51 
3.38 
Katherine 
1.76 
Maningrida 
1.65 
3.50 
1.56 
3.44 
Jabiru 
1.79 
Milingimbi 
1.65 
3.50 
1.56 
3.43 
Jabiru 
1.79 
Mimili 
1.54 
3.39 
1.46 
3.34 
Port Augusta 
1.75 
Mornington Island 
1.72 
3.55 
1.59 
3.45 
Mt Isa 
1.93 
Mossman Gorge 
1.72 
3.56 
1.59 
3.46 
Cairns 
2.03 
Nguiu 
1.65 
3.50 
1.56 
3.44 
Jabiru 
1.79 
Ngukurr 
1.59 
3.43 
1.49 
3.36 
Katherine 
1.76 
Numbulwar 
1.54 
3.40 
1.48 
3.36 
Nhulunbuy 
1.69 
Umbakumba 
1.53 
3.39 
1.47 
3.35 
Nhulunbuy 
1.69 
Wadeye 
1.65 
3.50 
1.56 
3.44 
Jabiru 
1.79 
Walgett 
1.81 
3.63 
1.66 
3.50 
Bourke 
1.58 
Wilcania 
1.72 
3.55 
1.56 
3.42 
Bourke 
1.58 
Yirrkala 
1.56 
3.41 
1.50 
3.37 
Nhulunbuy 
1.69 
Yuendumu 
1.51 
3.36 
1.45 
3.32 
Apatula 
1.63 
Source: Authors’ own calculation; Biddle & Taylor 2009
Table A1.5 Proportion of Australian population estimated/projected to identify as Indigenous in 2006 and 2026 by RSDA
2006 
2026–S1 
2026–S2 
2026–S3 
2026–S4 

Amata 
92.2 
92.9 
95.0 
92.8 
94.9 
Angurugu 
96.8 
97.2 
98.0 
97.1 
98.0 
Ardyaloon 
88.7 
89.7 
92.6 
89.6 
92.6 
Aurukun 
93.8 
94.5 
96.1 
94.4 
96.0 
Beagle Bay 
90.8 
91.7 
94.0 
91.6 
94.0 
Coen 
80.0 
82.0 
86.7 
81.7 
86.5 
Doomadgee 
94.8 
95.4 
96.8 
95.3 
96.7 
Fitzroy Crossing 
62.3 
65.7 
73.2 
65.3 
73.0 
Galiwinku 
93.6 
94.3 
95.9 
94.2 
95.9 
Gapuwiyak 
95.9 
96.3 
97.4 
96.3 
97.4 
Gunbalanya 
92.8 
93.7 
95.5 
93.6 
95.5 
Halls Creek 
80.7 
82.7 
87.3 
82.5 
87.2 
Hermannsburg 
93.2 
93.9 
95.7 
93.8 
95.6 
Hope Vale 
94.7 
95.2 
96.6 
95.2 
96.6 
Lajamanu 
87.5 
88.8 
91.9 
88.6 
91.8 
Maningrida 
93.7 
94.4 
96.0 
94.3 
96.0 
Milingimbi 
95.7 
96.2 
97.3 
96.1 
97.3 
Mimili 
88.9 
90.0 
92.8 
89.9 
92.8 
Mornington Island 
91.6 
92.7 
94.8 
92.5 
94.7 
Mossman Gorge 
100.0 
100.0 
100.0 
100.0 
100.0 
Nguiu 
94.5 
95.2 
96.6 
95.1 
96.5 
Ngukurr 
93.5 
94.2 
95.9 
94.1 
95.9 
Numbulwar 
91.8 
92.6 
94.7 
92.5 
94.7 
Umbakumba 
95.4 
95.9 
97.1 
95.8 
97.1 
Wadeye 
93.4 
94.2 
95.9 
94.1 
95.8 
Walgett 
54.9 
59.0 
67.2 
58.3 
66.7 
Wilcania 
74.6 
77.4 
83.0 
76.8 
82.7 
Yirrkala 
87.4 
88.7 
91.8 
88.5 
91.8 
Yuendumu 
88.4 
89.5 
92.4 
89.4 
92.4 
Source: Authors’ own calculation; Biddle & Taylor 2009
Previous Next