3228.0.55.001 - Population Estimates: Concepts, Sources and Methods, 2009  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 12/06/2009   
   Page tools: Print Print Page Print all pages in this productPrint All  
Contents >> Estimating national and state population >> Intercensal revision of population estimates

INTERCENSAL REVISION OF POPULATION ESTIMATES

2.53 When the Census year (i.e. 30 June) population estimates become available for the states and territories and Australia they can be compared with the alternative estimates for the same date already produced by updating the previous Census year estimate in quarterly steps using births, deaths and migration data. The estimate calculated using the current Census-based estimate is commonly referred to as the 'rebased' estimate, whereas the estimate calculated using the previous Census-based estimate is called the 'unrebased' estimate. The difference between the two estimates for each state and territory and Australia is called the 'intercensal error'.

2.54 Of these two estimates, the new Census-based estimate is customarily adopted as the true estimate.

2.55 The intercensal error can be explained in two different ways;

  • as a residual error from the previous Census (calculated by the unrebased estimate less the rebased estimate), or
  • as an indicator of how much the unrebased estimate over or under-estimated the true estimate (calculated by the rebased estimate less the unrebased estimate).

2.56 All references to intercensal error in this publication use the first interpretation, i.e. a positive number indicates that the unrebased estimate was higher than the rebased estimate and a negative number indicates than it was lower than the rebased estimate.

2.57 To overcome the break in continuity that using the new (rebased) estimate would entail, all intercensal population estimates using the previous Census as their base are then revised. In doing so, it is assumed that the error is accumulated by an equal number each quarter over the intercensal period. For example, in the case of a five-year intercensal period, (i.e. 20 quarters), the population at the end of the first quarter is adjusted up or down by 1/20 of the intercensal error, the second quarter is adjusted by 2/20, the third by 3/20 etc.

2.58 Details of the intercensal error for the states and territories and Australia at 30 June 1991, 1996, 2001 and 2006 are shown in Table 2.2.

2.2 Intercensal Error, State/territory of usual residence - at 30 June

1986-1991
1991-1996
1996-2001
2001-2006

Persons ('000)

New South Wales
2.4
-15.8
-27.0
2.8
Victoria
7.0
-20.9
35.6
-36.2
Queensland
11.1
15.9
5.8
-24.0
South Australia
10.4
4.5
-9.6
-12.8
Western Australia
(a)29.9
-3.3
11.5
-7.3
Tasmania
-6.3
-1.1
-1.8
-0.8
Northern Territory
-6.7
(b)-5.2
-0.2
-2.9
Australian Capital Territory
4.2
-1.2
-4.3
-4.7
Australia(c)
51.9
-27.5
10.6
-85.7

Percent of population

New South Wales
0.04
-0.25
-0.41
0.04
Victoria
0.16
-0.46
0.74
-0.71
Queensland
0.34
0.48
0.16
-0.59
South Australia
0.72
0.31
-0.63
-0.82
Western Australia
(a)1.83
-0.19
0.60
-0.35
Tasmania
-1.35
-0.23
-0.39
-0.16
Northern Territory
-4.05
(b)-2.86
-0.11
-1.40
Australian Capital Territory
1.45
-0.39
-1.36
-1.40
Australia(c)
0.30
-0.15
0.05
-0.41

(a) Approximately 9,000 persons of the intercensal error caused by PES over-adjustment in 1986.
(b) 1,045 persons of the intercensal error for the NT can be attributed to an under-estimate of the 1991 NT estimated resident population.
(c) Includes a small component of intercensal error for Other Territories.


2.59 The accuracy of Australia's post-censal population estimates compares favourably with that of other countries (see Table 2.3).

2.3 International comparison of absolute intercensal discrepancy, Percent of population - 1981 to 2006

Country
1981
1986
1991
1996
2001
2006

Australia
National
0.3
0.3
0.3
0.2
0.1
0.4
State/territory(a)
1.6
1.2
1.2
0.4
0.6
0.7
Canada(b)
National
0.4
0.4
0.1
0.6
0.2
0.3
Province(a)
1.8
1.2
1.2
0.9
0.7
1.0
USA (1980, 1990 and 2000)(b)
National
2.1
-
0.5
-
2.4
-
State(a)
2.5
-
1.5
-
2.6
-
England and Wales
England and Wales(a)
0.2
-
0.4
-
0.4
-
Constituent Counties and Unitary Authorities(a)(c)
1.8
-
2.5
-
1.8
-
New Zealand(d)
National
0.4
0.3
0.6
0.1
(e)0.8
1.1

(a) Average absolute error.
(b) Equates with error of closure which is the difference between population estimates produced prior to a Census and corresponding Census counts. (No account is taken of variations in undercount between censuses)
(c) Unitary authorities, single tier authorities, were introduced between 1995-98. These superseded a number of county districts although many areas retained the two tiered structure.
(d) Intercensal discrepancies from 2001 are based on the resident population concept, while figures for earlier years are based on the de facto population concept and equate with Census error of closure.
(e) Intercensal discrepancy for 2001 is calculated from the published estimates at the time, so does not reflect the revision to the base population at 30 June 1996 which occurred in 2002 following revision to estimated net undercount in the 1996 Census. Using the revised 1996 base population yields a 2001 intercensal discrepancy of 0.4%.


2.60 There are two areas which contribute to the intercensal error:
  • errors in the Census-based estimates of the population at the current or previous Census date (a one-off error at the time of compiling the estimates); and/or
  • errors in the estimates of any of the components of population change since the previous Census (an error which compounds over the intercensal quarters).

2.61 After the intercensal error at the state/territory level is adjusted for revisions using Census data on interstate migration, the remaining unattributable portion is referred to as the intercensal discrepancy. The intercensal discrepancy acts as a balancing item, that when combined with births, deaths and migration equals the difference between the two 30 June Census year population estimates.

2.62 The final intercensal error is the preliminary intercensal error adjusted for finalised estimates of births and deaths which occurred between 2001 and 2006. As there is a lag before finalisation of these components of population change, final intercensal error is calculated considerably later than the preliminary levels.

2.63 The final intercensal discrepancy is equivalent to the final intercensal error but with the interstate migration revisions introduced to account for intercensal error. They are therefore identical at the Australian, but not state and territory level.

2.4 Definitions of Intercensal Error and Intercensal Discrepancy

Definition Unrebased estimate Rebased estimate

Preliminary intercensal error Preliminary Preliminary
Final intercensal error Revised Final
Final intercensal discrepancy Final(a) Final

(a) Revisions to interstate migration on the basis of the current Census.


2.64 An assessment of the accuracy of Census date population estimates and the components of population change is given in Chapter 9 - Data sources.


Intercensal revision 2001-06

2.65 For the 2001-06 state and territory intercensal estimates, the 2006 Census allowed an assessment of how much of the intercensal error was due to inaccuracies in estimates of interstate migration. The 2006 Census data which allowed this assessment were from the two Census questions on usual residence 1 year ago and 5 years ago. Estimates of interstate migration based on Census data were used to revise the interstate migration component of population updates for the intercensal period, before the intercensal discrepancy was calculated.

(a) Revisions of interstate migration

2.66 Comparisons of the 2006 Census counts of 2001-06 interstate migration (adjusted for undercount and to include an estimate for ages 0 to 4) with the Medicare-based estimates of interstate migration (see Chapter 7 - Estimating interstate migration) for the 2001-06 period are shown in Table 2.5.

2.5 Net Interstate Migration, Comparison between Census-based and Medicare-based estimates - 2001 to 2006

Interstate migration 2001-06
Census(a)
Medicare(b)
Difference
Final intercensal error
30 June 2006(c)

New South Wales
-125 465
-136 330
10 865
2 767
Victoria
-13 121
-2 197
-10 924
-36 229
Queensland
147 890
164 362
-16 472
-24 034
South Australia
-10 192
-12 639
2 447
-12 780
Western Australia
3 993
-1 399
5 392
-7 259
Tasmania
4 113
3 105
1 008
-793
Northern Territory
-4 514
-8 474
3 960
-2 948
Australian Capital Territory
-2 375
-6 428
4 053
-4 673
Australia(d)
-
-
-
-85 715

- nil or rounded to zero (including null cells)
(a) Census counts have been adjusted for net undercount.
(b) Medicare-based estimates accumulated quarterly.
(c) A positive number indicates that unrebased ERP as at 30 June 2006 was higher than the rebased ERP. A negative number indicates it was lower than than the rebased ERP.
(d) Includes a small component of intercensal error for Other Territories.


2.67 The approach used for revising population estimates for the intercensal period assumes that intercensal error is primarily due to errors in estimating interstate migration (as it is the most difficult component to estimate - unlike births, deaths and overseas migration, there is no direct measure), to the extent that this is supported by the Census-based migration data.

2.68 The Census is considered the best measure of interstate migration at the time of the Census, so assuming it does not increase the intercensal error, this is the preferred base for migration estimates. Putting this into practice, if the difference between Census-based and Medicare-based migration estimates is the opposite sign as a state or territory's intercensal error, then that migration difference is interpreted as accounting for the intercensal error.

2.69 The 2001-06 revised net interstate migration levels were derived by:

Step 1 - Adjust interstate migration levels

2.70 Step 1a - For states and territories where the difference and intercensal error have the same sign, set migration levels to Medicare-based migration estimate.

2.71 For example, the intercensal error for New South Wales was not explained at all by the Census-based migration data, so the Medicare-based migration estimate was not altered. The intercensal error of New South Wales was 2,767 (i.e. the 2001-based 30 June 2006 population estimate was 2,767 too high), but Census-based migration indicated that, in net terms, fewer people departed the state. As such, if New South Wales' migration estimate was adjusted according to Census-based migration, its intercensal error would increase.

2.72 Step 1b - For states and territories where the difference and intercensal error have opposite signs and intercensal error is larger (further from zero) than the difference set migration levels to Census-based migration estimate.

2.73 For example, the Australian Capital Territory's intercensal error of -4,673 could be entirely attributed to interstate migration being 4,053 lower in Medicare-based migration than Census-based migration. The Census-based migration estimate of -2,375 is used, which reduces the intercensal error from -4,673 to -620.

2.74 Step 1c - For states and territories where the difference and intercensal error have opposite sign and intercensal error is smaller (closer to zero) than the difference set migration levels to Medicare-based migration estimate minus the intercensal error.

2.75 For example, Tasmania's final intercensal error of -793 could be entirely attributed to interstate migration being 1,008 lower in Medicare-based migration than Census-based migration. As an adjustment of only 793 is required to explain the intercensal error, that was the amount the Medicare-based estimates were altered by (giving 3,898 in Table 2.6), rather than the Medicare-based migration difference of 3,105.

Step 2 - Scale interstate migration to sum to zero

2.76 Finally, these new interstate migration levels are scaled to sum to zero according to the state or territory's proportion of 2001-06 Medicare-based interstate arrivals and departures.

2.6 Revisions to 2001-06 Net Interstate Migration

Interstate migration 2001-06
Intercensal discrepancy 30 June 2006
After adjustment to minimise intercensal discrepancy
After zero-sum adjustment (final)
Medicare movements(a)
Preliminary
Final

New South Wales
-136 330
-140 564
1 024 676
2 037
-1 467
Victoria
-2 197
-5 086
699 243
-39 685
-39 118
Queensland
164 362
160 495
935 834
-29 906
-27 901
South Australia
-10 192
-11 346
279 275
-12 001
-11 487
Western Australia
3 993
2 715
309 291
-3 405
-3 145
Tasmania
3 898
3 329
137 759
-561
-569
Northern Territory
-5 526
-6 196
162 178
-660
-670
Australian Capital Territory
-2 375
-3 169
192 098
-1 460
-1 414
Australia(b)
15 468
-
3 743 429
-85 587
-85 715

- nil or rounded to zero (including null cells)
(a) The number of interstate movements registered by Medicare between 30 June 2001 and 30 June 2006.
(b) Includes a small component of intercensal discrepancy for Other Territories.


2.77 The 2001-06 revised interstate migration estimates were then divided into two parts:

(1) Net movement for 2005-06. This used data from the Census question on usual address one year ago (adjusted for undercount and to include an estimate for age zero).

(2) Net movement for 2001-05. These data were obtained by subtracting the figures for 2005-06 from the totals for 2001-06.

2.78 Quarterly interstate arrival and departure estimates for 2005-06 and for the four years 2001-05 were calculated by spreading the total net gain/loss for each state/territory over the 4 quarters for 2005-06 and the 16 quarters for 2001-05 proportionally according to the original quarterly Medicare-based estimates.

(b) Distribution of the intercensal discrepancy

2.79 The 2001-06 intercensal discrepancy was distributed evenly over the intercensal period for each state/territory, sex and birth cohort as follows:

2.80 For cohorts born prior to the 30 June 2001, i.e. the start of the intercensal period,

Equation: eq2_4

where:
      Equation: eq2_5ais the intercensal discrepancy for quarter q
      ID is the total intercensal discrepancy

2.81 For cohorts born between the two censuses, the intercensal discrepancy was split over those quarters including and following the financial year of birth, that is,

Equation: eq2_5

where:
      n = the number of financial years that the cohort has been represented in the population including the financial year of birth of the cohort. For example the discrepancy for the cohort born in the 2005-06 financial year can only be distributed over the 2005-06 period, while the discrepancy for those born in 2004-05 can only be distributed over the two financial years (i.e. 8 quarters) 2004-05 and 2005-06.

2.82 These formulae were also used for distributing the intercensal discrepancy over the quarterly population estimates by marital status and country of birth.

(c) Final revised intercensal population estimates

2.83 The revised estimates of the population for the intercensal period were calculated by the quarterly adjustment of the population, starting with the 30 June 2001 (Census-based) population, for subsequent births, deaths, overseas migration, interstate migration and intercensal discrepancy. This was the same procedure as outlined in paragraph 2.17 for post-censal estimation, except that finalised components of change were available and the intercensal discrepancy component was also included.




Previous PageNext Page