|
Contents >>
Appendix 3. Estimating overseas migration
Introduction
A3.1. Persons arriving in, or departing from Australia provide information in the form of incoming and outgoing passenger cards. (Reproduced copies of the passenger card are provided in Appendix 1 of the ABS publication Overseas Arrivals and Departures (3401.0)). Incoming persons also provide information in visa applications, apart from people travelling as Australian and New Zealand citizens. These and other information (such as passport and visa data) available to the Department of Immigration and Multicultural Affairs serve as a sources for statistics of overseas arrivals and departures. Passenger cards are covered in greater detail in Chapter 4.
A3.2. Statistics on overseas arrivals and departures relate to the number of movements of travellers rather than the number of persons travelling (ie. the multiple movements of individual persons during a given reference period are all counted). However, the statistics exclude the movements of operational air crew and ships' crew, transit passengers who pass through Australia but are not cleared for entry, and passengers on pleasure cruises commencing and finishing in Australia aboard ships. Similarly, ABS statistics exclude those persons not travelling under standard visa conditions, such as 'boat people' and the 1999 intake of Kosovo refugees. These movements are however documented as part of border control processes.
Estimation
A3.3. Overseas arrival and departure statistics are derived from a combination of full enumeration and sampling. All permanent movements and all movements with a duration of stay of one year or more are fully enumerated and processed. On the other hand, statistics relating to movements with a duration of stay of less than one year are compiled from a sample of these movements and are therefore estimates which may differ from the statistics that would have been obtained if details on all these movements had been processed.
A3.4. Since January 1997, variable sample skips have been used in the random selection of records to be sampled. Separate skips are applied for each country of citizenship and the skip may vary for each processing month. Over a year, about 3.5% of all short-term movements are selected for sampling.
A3.5. Net Overseas Migration is calculated as the sum of three separate components:
| NOM | = | PLA - PLD + CJ | |
|
where | | | |
| NOM | = | Net Overseas Migration | |
| PLA | = | Permanent and Long-term Arrivals | |
| PLD | = | Permanent and Long-term Departures | |
| CJ | = | Category Jumping | |
Category of travel
A3.6. Passengers are classified according to their length of stay (in Australia or overseas) or 'category'. This is primarily determined by the passenger's selection of his or her own category on the passenger card, which is accepted provided it is consistent with other information on the card. There are three main categories: Permanent movement, Long-term movement and Short-term movement.
A3.7. Permanent movement comprises
- Arrivals of settlers (ie. persons who hold migrant visas, regardless of stated intended period of stay, and New Zealand citizens who indicate an intention to settle) and those who are otherwise eligible to settle (eg. overseas born children of Australian citizens), and
- Departures of Australian residents (including former settlers) who on departure state that they are departing permanently.
A3.8. Long-term movement is defined as
- Arrivals of visitors (with the exception of those who hold migrant visas - see above) and the departures of Australian residents who intend to stay, in Australia or abroad respectively, for twelve months or more, and
- Departures of visitors and the return of Australian residents who had stayed, in Australia or abroad respectively, for twelve months or more.
A3.9. Short-term movement is the residual, that is:
- Arrivals and departures of travellers whose actual period of stay is less than twelve months, with the exception of those who, on arrival, hold migrant visas.
A.3.10. However, a significant number of travellers (ie. visitors to Australia on arrival and Australian residents going abroad) state exactly twelve months or one year as their intended period of stay and stay for less than that period. Accordingly, in an attempt to maintain consistency between arrivals and departures, movements of travellers who report their actual or intended period of stay as being one year exactly are randomly allocated to long-term or short-term, in proportion to the number of movements of travellers who report their actual length of stay as up to one month more, or one month less than one year.
Category jumping
A3.11. From July 1976 onwards, there has been an adjustment for category jumping in net overseas migration. Prior to December quarter 1989, adjustments for category jumping were made only to revised population estimates.
A3.12. Category jumping is the term used to describe the net effect of changes in travel intentions from short-term to permanent/long-term or vice versa. It arises when the duration of a person's journey differs from that originally indicated (on the arrival/departure card at the beginning of the journey) in such a way as to affect his/her categorisation. For example, an Australian resident departing for a short-term visit overseas (stating that he/she intends to stay abroad for less than twelve months) in fact stays more than twelve months, thereby changing his/her travel category from short-term to long-term.
A3.13. If no adjustments were made for category jumping, a resident who states on the passenger card that he/she is departing on a short-term trip but never returns to Australia would not be subtracted from Australia's population. However, if this resident returned after an absence of more than a year he/she would be added to the population even though he/she had not been subtracted on departure. Similarly, a visitor who arrives on a short-term trip but who never departs would not be added to the population; and if he/she departs after one year he/she will be subtracted from the population even though he/she had not been added in on arrival.
A3.14. Chart A3.1 identifies the type of category jumping that may logically occur. People who are in shaded areas in the chart are considered to have jumped a category.
|
A3.1. Category Jumping in Overseas Movement
| |
|
| |
|
| |
A3.15. The types of category jumping which cause difficulties in population estimates are:
- Visitors or residents who were categorised as short-term on arrival (visitors) or on departure (residents) respectively but were not in the same category on completion of their journeys, and
- Visitors or residents who are not in the short-term category on arrival (visitors) or on departure (residents) but were in that category on completion of their journeys.
These are represented by the medium shaded yellow boxes in Chart A3.1 above.
A3.16. Category jumping between the long-term category and the permanent category does not affect population estimates although this may affect the usefulness of the statistics on these movements for other purposes.
A3.17. In estimating category jumping for each quarter, there are two components to be calculated:
- Overseas visitor component - this is derived by calculating the difference between the number of short-term visitors who enter Australia in a reference quarter, and the number of visitors who depart Australia over the year following the reference quarter and state that they arrived in the reference quarter.
- Australian resident component - this is derived by calculating the difference between the number of Australian residents departing Australia in a reference quarter for a period less than 12 months, and the number of Australian residents who return to Australia over the following year and state that they departed Australia in the reference quarter.
This method implicitly takes account of permanent and long-term movements as well as short-term movements.
A3.18. Category jumping is then derived by subtracting the Australian resident component (this component generally reduces the population estimate) from the overseas visitor component (this component generally increases the population estimate). Table A3.2 shows the relative size of each of these components since 1986 and their effect on category jumping.
| | | | | |
|
A3.2. Components of Category Jumping, Years Ended 31 December 1986 to 1998 | |
| | Residents | Visitors | Total | |
| | |
| 1986 | 24,483 | 30,907 | 6,424 | |
| 1987 | 35,362 | 47,608 | 12,246 | |
| 1988 | 28,635 | 52,331 | 23,696 | |
| 1989 | 45,043 | 56,512 | 11,469 | |
| 1990 | 45,256 | 45,807 | 551 | |
| 1991 | 44,580 | 32,726 | -11,854 | |
| 1992 | 84,494 | 59,058 | -25,436 | |
| 1993 | 51,303 | 28,617 | -22,686 | |
| 1994 | 41,663 | 16,932 | -24,731 | |
| 1995 | 18,359 | 20,650 | 2,291 | |
| 1996 | 10,459 | 4,833 | -5,626 | |
| 1997 | 16,668 | 23,985 | -7,317 | |
| 1998 | 18,878 | 26,070 | 7,192 | |
| | |
| | | | | |
A3.19. Preliminary estimates of category jumping have proven difficult to estimate to the desired level of accuracy since the ABS began adjusting population estimates for category jumping in the September quarter 1976. This is because it is necessary to estimate from only 2 quarters of data (ie. the reference quarter and the following quarter) the number of residents returning and the number of overseas visitors departing in the four quarters following the reference quarter of their 'first leg movement'. To illustrate this problem, Table A3.3 shows the pattern of return from 1992 to 1997 of residents who departed in the March quarter:
|
A3.3. Australian Residents Returning from Overseas Who Departed in the March Quarter, 1992 to 1997 | |
| | | Quarter of Return (per cent) | |
Residents
departing
(No.)
| |
| Year | Estimate | Same Quarter | Q+1 | Q+2 | Q+3 | Q+4 | Remainder
(per cent) | |
| | |
| 1992 | Quarter | 66.48 | 20.86 | 3.62 | 2.64 | 1.50 | 4.90 | 454,170 | |
| | Cumulative | | 87.34 | 93.60 | 95.10 | 95.10 | | | |
| 1993 | Quarter | 68.37 | 21.89 | 3.20 | 2.42 | 1.52 | 2.61 | 460,291 | |
| | Cumulative | | 90.26 | 93.45 | 95.87 | 97.39 | | | |
| 1994 | Quarter | 66.29 | 24.18 | 3.39 | 2.18 | 1.27 | 2.68 | 492,142 | |
| | Cumulative | | 90.47 | 93.86 | 96.04 | 97.32 | | | |
| 1995 | Quarter | 68.77 | 23.08 | 3.62 | 2.43 | 1.31 | 0.79 | 512,528 | |
| | Cumulative | | 91.85 | 95.47 | 97.90 | 99.21 | | | |
| 1996 | Quarter | 65.89 | 23.50 | 3.27 | 2.20 | 1.43 | 3.71 | 578,898 | |
| | Cumulative | | 89.39 | 92.66 | 94.86 | 96.29 | | | |
| 1997 | Quarter | 66.47 | 23.83 | 3.41 | 2.21 | 1.23 | 2.86 | 625,077 | |
| | Cumulative | | 90.30 | 93.71 | 95.91 | 97.14 | | | |
| | |
A3.20. For example, when making a preliminary estimate of category jumping to be used in compiling the population estimate for the March quarter 1998 (published September 1998), only March and June quarter 1998 overseas arrivals data are available. This means that an estimate has to be made of the number of Australian residents who will return in the September and December quarters 1998, and the March quarter 1999 (the same estimation has to be made for overseas visitors).
Category jumping estimation methods
A3.21. Since the commencement of adjusting population estimates for category jumping, a number of methods of producing a preliminary estimate of category jumping have been investigated:
- A regression model based on the relationship between the intended length of stay stated by short-term movers at the commencement of their journey, and the actual length of stay as stated on completion of their journey. This method was discontinued when it was realised that superior estimates could be made by assuming that there was no category jumping.
- Assume preliminary category jumping to be zero. This method was used until September quarter 1989.
- The application of category jumping rates by country of birth, to quarterly estimates of visitors arriving and residents departing. These rates looked relatively stable over time, and produced good estimates for a couple of years before a sudden rise in 1993 in the number of Australian residents departing with the intention of returning within 12 months but staying longer.
A3.22. The method used was to apply category jumping rates by country of birth to estimates of visitors arriving and residents departing. This method was developed after recognising that there were definite differences in the propensity to jump categories between various birthplace groups. Table A3.4 shows the major birthplace groups which influence category jumping numbers, split by the two components, Australian residents and overseas visitors from 1990 to 1997.
|
A3.4. Major Birthplace Groups Contributing to Category Jumping, 1990-91 to 1996-97 | |
| | | | | Period | | | | |
| | | |
| Country of birth | 1990-91 | 1991-92 | 1992-93 | 1993-94 | 1994-95 | 1995-96 | 1996-97 | |
| Australian Residents | |
| Australia | 18,380 | 11,014 | 3,087 | 12,862 | -2,490 | -1,652 | -9,891 | |
| New Zealand | 9,064 | 3,172 | 2,032 | 3,413 | 388 | 874 | 2,021 | |
| United Kingdom and Ireland | 13,835 | 3,420 | 14,906 | 7,356 | -1,539 | 953 | -1,038 | |
| Former Yugoslav Republics | -2,890 | 1,185 | -1,167 | 1,375 | -1,557 | 696 | 514 | |
| Malaysia and Brunei | 3,804 | 4,103 | 3,394 | 432 | 1,359 | 565 | 187 | |
| Viet Nam | 206 | 303 | 1,933 | -431 | 3,422 | -3,131 | -661 | |
| China | 2,603 | 4,828 | 625 | 1,112 | 2,876 | 2,957 | 2,730 | |
| Hong Kong and Macao | 2,875 | 9,214 | 7,041 | 3,477 | 3,901 | 8,233 | 17,336 | |
| Japan | 2,973 | 4,626 | 4,206 | 1,059 | 2,481 | -417 | 86 | |
| India | 178 | 1,650 | 2,446 | 2,124 | -881 | 1,822 | -2,022 | |
| Total | 65,388 | 72,753 | 76,751 | 41,178 | 24,395 | 19,820 | 23,985 | |
| Overseas Visitors | |
| Australia | 3,572 | 6,301 | 3,042 | 8,559 | 7,226 | 4,323 | 1,967 | |
| Fiji | 735 | 896 | 763 | -1 | 250 | 445 | 260 | |
| New Zealand | 9,995 | 4,601 | 7,044 | 10,613 | 6,424 | 7,250 | 5,728 | |
| Papua New Guinea | 2,072 | 710 | 660 | 437 | -115 | 128 | -431 | |
| United Kingdom and Ireland | 5,267 | 5,808 | 4,280 | 4,023 | 6,755 | 6,818 | 16,233 | |
| Philippines | 1,890 | 1,905 | 1,795 | 595 | -4 | 1,991 | 4,369 | |
| Singapore | 427 | 403 | -1,419 | -4,703 | -2,138 | -2,656 | -5,107 | |
| China | 2,262 | 3,592 | 3,563 | 4,603 | 8,910 | 11,296 | 10,024 | |
| Hong Kong and Macao | -288 | 2,673 | 3,118 | -1,519 | -8,770 | -9,024 | -9,296 | |
| United States of America | -628 | -140 | -714 | -1,441 | -6,113 | -1,692 | 407 | |
| Total | 57,063 | 51,445 | 44,122 | 20,346 | 11,478 | 14,296 | 16,668 | |
| | |
| | | | | | | | | |
A3.23. The current method involves delaying the compilation of quarterly population estimates until an extra quarter of overseas migration data becomes available. As Table A3.3 shows, if the compilation of the quarterly population estimate can be delayed until an extra quarter of migration data becomes available, the accuracy of preliminary category jumping estimates can be improved significantly, as more than 90 per cent of Australian residents return within the same or following quarter of departure, whereas less than 70 per cent return in the same quarter as that in which they departed. Estimates of the percentage of (intended) short-term Australian residents who will not return within 6 months, and the percentage of (intended) short-term overseas visitors who will stay more than 12 months are made, based on the experience of recent years for the quarters not yet available. The advantage of this method over previous ones used is that it can take account of sudden changes in category jumping patterns.
A3.24. Having estimated the quarterly category jumping adjustment for Australia as a whole, it is distributed by State, age and sex in proportion to the number of permanent and long-term arrivals for the same quarter.
|