3133.0 - Demography Working Paper 2003/1 - Estimated Resident Population and Measurement of Category Jumping, 2003  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 06/02/2003   
   Page tools: Print Print Page Print all pages in this productPrint All

INTRODUCTION

1 This working paper outlines a recently detected deficiency in the current measurement of migration category jumping, the actions that will be taken in the short term for estimates of Estimated Resident Population (ERP) and Net Overseas Migration from September Quarter 1997 onwards, and proposed research and development to minimise the impact of this issue on ERP and Net Overseas Migration statistics.


SUMMARY

2 Estimated Resident Population for States, Territories and Australia incorporating revisions to final 2001 Census rebased estimates from September Quarter 1996 to June Quarter 2001 and subsequent quarter preliminary estimates will be released in Australian Demographic Statistics, September Quarter 2002 (ABS cat. no. 3101.0) on 20 March 2003. Total populations for States, Territories and Australia for December Quarter 2000 to September Quarter 2002 will be released in Population, Australian States and Territories - Electronic Publication (ABS cat. no. 3239.0.55.001) on the ABS website on 18 February 2003.

3 The ABS will set category jumping to zero for the final estimates of September Quarter 1997 to June Quarter 2001 due to a deficiency in the current method of estimation. Any error resulting from this action will be included in the final estimate of 1996-2001 Intercensal Discrepancy. Category jumping will also be set to zero for September Quarter 2001 and subsequent quarters for preliminary ERPs in the short term until a new methodology is developed. See table 1 for a list of the ERPs that are influenced by this action.

4 ABS has commenced a project to develop a new methodology for measuring category jumping based on new data on international passenger movements provided by the Department of Immigration and Multicultural and Indigenous Affairs.


CATEGORY JUMPING AND ESTIMATED RESIDENT POPULATION

5 Overseas arrivals and departures are classified according to length of stay (in Australia or overseas), recorded in months and days by travellers on passenger cards. There are three main categories of movement:

  • permanent movements;
  • long-term movements (one year or more); and
  • short-term movements (less than one year).

Overseas arrivals and departures are included in the calculation of ERP where the movement (arrival or departure from Australia) is permanent or long term as stated on the first leg of travel.

6 Category jumping is the term used to describe changes between intended and actual duration of stay of travellers to/from Australia, such that their classification as short-term or as long-term/permanent movers is different at arrival/departure from that after 12 months. Category jumping consists of two components - an Australian resident component and an overseas visitor component. The scenarios below describe some of the circumstances of international travellers that are of interest when measuring category jumping.

Category jumping scenarios.
Scenario 1.
    Passenger A states on arrival in Australia that their intention is to stay less than 12 months (i.e. short term) and are not included in ERP. However, their visit becomes 24 months and they are then classified as category jumpers (from short term to long term) and added to the ERP for the initial quarter of arrival.

Scenario 2.
    Passenger B states that they are going to be staying in Australia for more than 12 months. They are added to the preliminary ERP on this basis. However, they then leave within 6 months and will then be taken out of the ERP for the initial quarter of arrival.

Scenario 3.
    Australian resident C states that they are leaving Australia for less than 12 months. They decide to stay away for more than 12 months, and so an adjustment is made to remove them from ERP for that initial quarter of departure.

Scenario 4.
    Australian Resident D states on departure that they are leaving Australia for more than 12 months. However, they return before the end of the 12 month period. They will have been initially removed from preliminary ERP in the quarter of departure, however, an adjustment is made to add them back into the population for the departure quarter when the data becomes available for their return to Australia.

7 The current method for measuring category jumping uses gross movement data from the incoming and outgoing passenger card data provided by the Department of Immigration and Multicultural and Indigenous Affairs (DIMIA). This information is analysed against those arriving and departing from Australia which provides information on long-term arrivals, short-term arrivals and residents arriving and departing Australia. The methodology is explained in greater detail in Appendix 3 - Estimating Overseas Migration in Demographic Estimates and Projections: Concepts, Sources and Methods (ABS cat. no. 3228.0).


DELAYS AND DEFICIENCIES IN CATEGORY JUMPING ESTIMATION

8 Delays in receipt of passenger card data from the Department of Immigration and Multicultural and Indigenous Affairs have occurred due to processing delays with incoming and outgoing passenger cards from August 2000 onwards. For quarters where completed migration data was not available, the ABS used projections or partial estimates for Net Overseas Migration (including Category Jumping) to calculate ERPs. Complete data on permanent and long term arrivals and departures have now been received and this data will be used in calculating final estimates of ERPs from September Quarter 2000 to March Quarter 2001.

9 With the completion of the migration data time series, a deficiency has been identified in the current measurement of migration category jumping. This problem can be traced to a redesign of the passenger cards and processing systems in July 1998 where a series break has been detected. In particular, the measurement of duration of stay/absence on the second leg of travel (which is important for determining a change in duration of stay/absence) was changed from a question on the passenger card completed by the traveller, to a calculation based on comparison of movements using DIMIA records of previous movements. The absence of a complete series of passenger card data over the past two years has delayed identification of this problem and development of an alternative measurement technique.

10 New data on passenger movements has been provided by DIMIA which enable the matching of consecutive international movements and improved measurement of travel sequences. Preliminary exploratory analysis indicates that some long term visitors to Australia who are normally included in the Australian population, would be removed from the population estimates through category jumping when they are, in fact residing in Australia but going overseas for a short period, and are not added back into the population on their return. In addition, of those who subsequently return to Australia, whilst half state they are intending to stay long-term, almost all leave again on a similar basis within less than 12 months. The combine effect of these short departures from Australia would be that the person may never be counted in the population regardless of the overall length of residence in Australia unless they remain in Australia for a continuous period of more than 12 months. This is mostly caused by overseas students (and possibly temporary business migrants, and New Zealand citizens) who reside in Australia for most of the year but often travel overseas for short periods during their overall period of residence in Australia.

Impact of subsequent overseas travel by long term visitors hypothesis
Hypothesis
    A student (or any long term visitor) entering Australia long term (12 months or more) on a long term temporary visa is added to the population (ERP). If they depart within 12 months, for any reason and for any period, they are removed from the population via a category jumping adjustment. That is, even if they visit their home country but once a year for holidays, birthdays, festivals etc., they are removed from the population. It is not clear what happens when they return to Australia - some are added back to ERP as new long term visitors (which may result in duplicate counting of a long term arrival within a year), whilst others are treated as "residents returning from a short trip overseas", and in which case they are not added to ERP (even though, given their long term status, they conceptually should be included).

11 Given this deficiency and the unavailability of an alternative measure in the short term, the ABS intends to set category jumping to zero from September Quarter 1997 to June Quarter 2001 for forthcoming final ERP releases. Any error resulting from this action will be included in the final estimate of 1996-2001 Intercensal Discrepancy. For subsequent quarters (September 2001 onwards), category jumping will be set to zero for preliminary estimates.

TABLE 1 ESTIMATED RESIDENT POPULATION, CATEGORY JUMPING CALCULATIONS

Category jumping calculations

Quarter publishedPreliminary ERPFinal ERP

Sept 1997 – Jun 2000ModelZero
Sept 2000Zero(a)Zero
Dec 2000Zero(a)Zero
Mar 2001ModelZero
Jun 2001ModelZero
Sept 2001Zero(a)New model
Dec 2001Zero(a)New model
Mar 2002Model(b)New model
Jun 2002Zero(a)New model
Sept 2002 – Sept 2003ZeroNew model

(a) Set to zero due to processing delays with passenger cards.
(b) Will be revised to zero pending development of an improved measure.

12 ABS has commenced research to develop potential new category jumping estimation methods based on matching individual traveller movements over time. A State and Territory dimension will be an important design feature criteria for the new method. This will allow either modelling of traveller behaviour or directly measure travellers who satisfy the criteria in hypothesis described after paragraph 10 above.

13 The growing difference between long term arrivals and long term departures has been observed over recent years. This may be symptomatic of the change in duration of stay/absence measurement with increasing numbers of long term visitors on arrival being classified as short term visitors on departure (one of the groups to be measured by category jumping). Recent changes to Australia's Migration Program have led to an increase in long term visitors obtaining permanent residency visas on-shore (e.g. former students and business migrants). Whilst this phenomena is not relevant to the measurement of ERP, ABS is negotiating with DIMIA regarding potential data on on-shore visa grants that may be available for statistical analysis and reporting by State and Territory. Demography Working Paper 2000/4 - Category Jumping : Trends, Demographic Impact and Measurement Issues also looked into concerns with the volatility of category jumping trends, their impact and measurement.


RELEASE OF FINAL 2001 CENSUS BASED ERPs, 30 SEPTEMBER 1996 TO 30 JUNE 2001

14 Much of the change brought about by the decision to set category jumping to zero will be offset by a rebase of the ERP based on the 2001 Census. The 2001 Census rebased ERPs includes adjustments to take into account of:
  • final Census counts on a usual residence basis;
  • final data on Census net undercount including findings reported in Demography Working Paper 2002/2 - Estimated Resident Population and Effects of Census Systems Created Records;
  • final data on Residents Temporarily Overseas on Census night;
  • final births data (on a date of occurrence basis) for the intercensal period;
  • final deaths data (on a date of occurrence basis) for the intercensal period;
  • final interstate migration data for the intercensal period incorporating the 2001 Census interstate migration results; and,
  • final overseas migration data for the intercensal period.

15 Estimated Resident Population for States, Territories and Australia incorporating revisions to final 2001 Census rebased estimates from September Quarter 1996 to June Quarter 2001 and subsequent quarter preliminary estimates (using the new 30 June 2001 base) will be released in Australian Demographic Statistics, September Quarter 2002 (ABS cat. no. 3101.0) on 20 March 2003. Total populations for States, Territories and Australia for December Quarter 2000 to September Quarter 2002 will be released in Population, Australian States and Territories - Electronic Publication (ABS cat. no. 3239.0.55.001) on the ABS website on 18 February 2003.

16 For further information about this paper, contact Peter Meadows on (02) 6252 6639 or p.meadows@abs.gov.au.

Demography Section
Australian Bureau of Statistics
6 February 2002