1 This publication contains statistics of persons arriving in, and departing from, Australia, together with the major characteristics of travellers. More detailed statistics can be made available on request (see paragraph 31).
SOURCE OF THE STATISTICS
2 Persons arriving in, or departing from, Australia provide information in the form of incoming and outgoing passenger cards. Incoming persons also provide information in visa applications (apart from people travelling as Australian or New Zealand (NZ) citizens). These and other information available to the Department of Immigration and Citizenship (DIAC) serve as a source for statistics of overseas arrivals and departures (OAD).
3 In July 1998, DIAC revised the incoming and outgoing passenger cards and associated procedures as well as computer systems. Following these changes, some questions on the passenger cards were not compulsory and answers to these questions were not checked by Customs officers. The question on marital status was deleted. Data on marital status is now derived from visa applications (only for certain visa classes) and is therefore not available for Australian or NZ citizens. The changes also affect the data for 'previous country of residence' which is imputed for Australian and NZ citizens. For more information see the May 1998 issue of this publication. Since July 1998, there have been additional minor changes to both incoming and outgoing passenger cards.
4 From July 2001, DIAC adopted a new passenger card processing system which involved electronic imaging of passenger cards and intelligent character recognition of the data stored in the images. This process has yielded several improvements to the processing of passenger card data, most notably the detailed information about missing values. There have also been several changes to data quality. Information on these changes appears in the Appendix (Data Quality Issues) linked to the Explanatory Notes of this issue.
5 The statistics in this publication relate to the number of movements of travellers rather than the number of travellers (i.e. multiple movements of individual persons during a given reference period are each counted separately). The statistics exclude the movements of operational air and ships' crew, transit passengers who pass through Australia but are not cleared for entry, passengers on pleasure cruises commencing and finishing in Australia, and unauthorised arrivals.
STATE AND TERRITORY CLASSIFICATION
6 Following the 1992 amendment to the Acts Interpretation Act to include the Indian Ocean Territories of Christmas Island and the Cocos (Keeling) Islands as part of geographic Australia, population estimates commencing with September quarter 1993 include estimates for these two territories. To reflect this change, another category of the state/territory classification has been created, known as Other Territories. Other Territories includes Jervis Bay Territory, Christmas Island and the Cocos (Keeling) Islands. OAD data for Other Territories are not available prior to February 1995.
7 The classification of countries in this publication is the Standard Australian Classification of Countries. For more detailed information refer to the ABS publication Standard Australian Classification of Countries (SACC) (cat. no. 1269.0).
8 The statistics on country of birth, citizenship, residence or main destination have certain limitations because of reporting on passenger cards. For instance, United Kingdom includes England, Scotland, Wales, Northern Ireland, the Channel Islands and the Isle of Man. Similarly Korea includes both the Republic of Korea and the Democratic People's Republic of Korea.
9 A large number of short-term residents departing state Europe as their main destination on the passenger card. These responses are grouped into inadequately described, unless otherwise indicated.
10 OAD 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. All movements with a duration of stay of less than one year are sampled. Statistics relating to these movements are therefore estimates which may differ from statistics which would have been obtained if details of all these movements had been processed. Sample standard errors can be found in the Standard Errors document linked to the Explanatory Notes of this issue.
11 Since January 1997 variable sample skips were used in the selection of records to be sampled. Separate skips were applied for each country of citizenship. Over a year about 3.5% of all short-term movements were selected for sampling. However, for operational reasons variable skips across months were ceased from August 2000. From January 2005, 4.9% of all short-term movements have been selected for sample. The skip values correspond to the lowest skip value (i.e. the highest sample selected) for each country from a sample design including individual month and direction of travel as sub-categories.
12 The statistics in this publication have been rounded to the nearest 100 for short-term movements and to the nearest 10 for permanent and long-term movements. As a result, sums of the components may not add exactly to totals. Analysis featured in the Key Points and Main Features of this publication is based on unrounded data. Calculations made on rounded data may differ to those published.
CORRECTIONS AND IMPUTATIONS
13 The imprecision due to sampling errors should not be confused with errors arising from imperfections in reporting, which may occur in any data collection, whether sampled or not. Every effort is made to minimise such errors, both through careful design of the passenger cards and through checks on the information once it is received. During the edit process some items are corrected where they conflict with other known information. Missing replies to certain items such as age, state and country of stay/residence are also imputed by reference to other related items. Information on non-response rates and data imputation appears in the Appendix (Data Quality Issues) linked to the Explanatory Notes of this issue.
14 Errors of this kind differ from discrepancies arising from the fact that certain information reflects the travellers' intentions at the time the passenger cards were completed. These intentions are, of course, subject to change. Particularly affected is the distinction between permanent and temporary movement and in the latter case, length of intended stay, country in which most time will be spent and main reason for journey.
SEASONAL ADJUSTMENT AND TREND ESTIMATES
15 Seasonally adjusted and trend estimates of short-term overseas movements are shown in tables 1, 2, 3, 4, 7 and 8.
16 Seasonally adjusted estimates are derived by estimating and removing systematic calendar related effects from the original series. In the short-term visitor arrival and short-term resident departure series, these calendar related effects are known as seasonal (e.g. increased travel in December due to the Christmas holiday period) and trading day influences (arising from the varying length of each month and the varying number of Sundays, Mondays, Tuesdays, etc. in each month). Each influence is estimated by separate seasonal and trading day factors which, when combined, are referred to as the combined adjustment factors.
17 From November 2004 ABS has introduced an improved method for removing trading day effects from seasonally adjusted estimates. Corrections for trading day effects are now applied as prior corrections to the original estimates, rather than being applied within the seasonal adjustment process. This is now consistent with the treatment of any corrections for large extremes, changes in level, changes in seasonal pattern, Easter, and other effects. This change in methodology will result in revisions to seasonally adjusted and trend estimates. From July 2003 concurrent seasonal adjustment methodology has been used to derive the combined adjustment factors. This means that data from the current month are used in estimating seasonal and trading day factors for the current and previous months. Concurrent seasonal adjustment replaces the forward factor methodology used since seasonal adjustment of short-term visitor arrivals began in 1969 and short-term resident departures in 1976.
18 Concurrent adjustment can result in revisions each month to the seasonally adjusted estimates for earlier periods. However, in most instances, the only noticeable revisions will be to the combined adjustment factors for the current month, the previous month and the same month a year ago. Although there is no specific information paper on concurrent adjustment to short-term visitor arrivals or resident departures, more detail on the method in general can be found in the Information Paper: Introduction of Concurrent Seasonal Adjustment into the Retail Trade Series (cat. no. 8514.0).
19 More recently, the ABS implemented improved methods of producing seasonally adjusted estimates, focused on the application of Autoregressive Integrated Moving Average (ARIMA) modelling techniques. The revision properties of the seasonally adjusted and trend estimates can be improved by the use of ARIMA modelling. ARIMA modelling relies on the characteristics of the series being analysed to project future period data. The projected values are temporary, intermediate values, that are only used internally to improve the estimation of the seasonal factors. The projected data do not affect the original estimates and are discarded at the end of the seasonal adjustment process. The OAD collection uses ARIMA modelling where appropriate for individual time series. The ARIMA model is assessed as part of the annual reanalysis and following the 2006-07 annual reanalysis 96% of time series use an ARIMA model. For more information on the details of ARIMA modelling see 'Feature article: Use of ARIMA modelling to reduce revisions' in the October 2004 issue of Australian Economic Indicators (cat. no. 1350.0).
20 Seasonal adjustment procedures do not aim to remove the irregular or non-seasonal influences which may be present in any particular month, such as the effect of major sporting and cultural events, changes in airfares and the fluctuation of the Australian dollar relative to other currencies. Irregular influences that are highly volatile can make it difficult to interpret the movement of the series even after adjustment for seasonal variation. Trend estimates take these irregular influences into account.
21 The trend estimates of short-term overseas visitor arrivals and short-term Australian resident departures are derived by applying a 13-term Henderson-weighted moving average to all months of the respective seasonally adjusted series except the first and last six months. Trend series are created for the last six months by applying surrogates of the Henderson weighted moving average to the seasonally adjusted series.
22 While this technique enables smoothed data for the latest period to be produced, it does result in revisions to the smoothed series, principally of recent months, as additional observations become available. There may also be revisions as a result of the re-estimation of the seasonal factors. For further information, see A Guide to Interpreting Time Series - Monitoring Trends (cat. no. 1349.0).
23 Occasionally situations occur that necessitate breaks being applied to the trend series. These breaks are necessary because of a change in the underlying level of the original series. While the breaks apply to an individual country (e.g. Indonesia) a consequence is that breaks are also applied to the regional total series (e.g. Total South-East Asia) and the total series. Breaks currently included in the trend series are as follows:
- October 2002: Short-term resident departures - Indonesia, Total South-East Asia and Total - Trend series breaks due to the decrease in movements resulting from the Bali bombing of 12 October 2002.
- December 2003: Short-term resident departures - Indonesia, Total South-East Asia and Total - Trend series breaks due to a return to the trend levels experienced prior to the Bali bombing of 12 October 2002.
- October 2005: Short-term resident departures - Indonesia, Total South-East Asia and Total - Trend series breaks due to the decrease in movements resulting from the Bali bombing of 1 October 2005.
- December 2006: Short-term resident departures - Indonesia, Total South-East Asia and Total - Trend series breaks due to the increase in movements to Indonesia to levels closer to, but still lower than, the movements experienced prior to the Bali bombing of 1 October 2005.
An improved correction method has been implemented in the seasonal adjustment process, to remove the effects of Chinese New Year and Ramadan from the seasonally adjusted estimates.
EFFECT OF NEW SEASONALLY ADJUSTED ESTIMATES ON TREND ESTIMATES
- Chinese New Year - Chinese New Year often falls in February but on some occasions falls in January. The movement of Chinese New Year between the boundary of January and February can cause biased seasonally adjusted and trend estimates. The Chinese New Year proximity adjustment method takes into account the graduated increase in activity in the days leading up to the holiday period followed by a graduated return to the normal activity levels in the days following. The proximity correction is only applied to the series from January 1998, as it was not significant before this date. Further details on this adjustment method can be found in 'Estimating and removing the effects of Chinese New Year and Ramadan to improve the seasonal adjustment process' (ABS, Australian Economic Indicators, cat. no. 1350.0, November 2005 issue).
- Ramadan - Ramadan is the ninth month of the Islamic calendar and starts eleven days earlier each year in the Western calendar, so adjustments for this effect apply to different months over the years. The adjustment was made after the detection of a significant influence on travel for Malaysia and Indonesia associated with the end of the Islamic month of Ramadan. The estimates for Total South-East Asia were also corrected as a consequence of the Malaysia and Indonesia series corrections. Other time series did not have a significant Ramadan effect and were not corrected.
Under concurrent seasonal adjustment, the most recent seasonally adjusted and trend estimates are likely to be revised when original estimates for subsequent months become available. The trend revision is a combined result of the revision of the seasonally adjusted estimates and the revision derived from the use of asymmetric moving averages as future data become available. ABS research shows that about 75% of the total revision to the trend estimate at the current end of the time series is due to the use of different asymmetric moving averages when the original estimate for the next time period becomes available. To assess the reliability of the trend estimate at the current end, the 'what-if' charts present trend estimates under two different scenarios for the next time period. The charts show only the impact due to the changes of the asymmetric moving averages and do not include the unknown impact of revision to seasonal factor estimates that would arise when the original estimate for the next time period becomes available.
For a detailed discussion and analysis of OAD time series estimates, see the ABS Demography Working Paper 2004/2 - Interpretation and Use of Overseas Arrivals and Departures Estimates
(cat. no. 3106.0.55.002), available on the ABS website.
27 For further information on the seasonal adjustment process contact the Assistant Director, Time Series Analysis on Canberra (02) 6252 6345 or by email at <firstname.lastname@example.org>.
28 Users of these statistics may also wish to refer to the following ABS products:
Related statistics are also published by DIAC, the Department of Resources, Energy and Tourism and Tourism Research Australia.
30 The ABS also issues a daily Release Advice on the website which details the products to be released in the week ahead.
ADDITIONAL STATISTICS AVAILABLE
31 As well as the statistics included in this and related publications, other relevant unpublished data are available for the following variables:
Country of birth
Age (date of birth)
Marital status (not available for Australian or New Zealand citizens)
Category of travel
Previous/future country of residence
State of intended address/lived
Intended/actual length of stay
Main reason for journey
Country of residence
State or territory of intended address on arrival
State or territory in which most time spent on departure
Occupation (not available for short-term movements)
Intended/actual length of stay overseas
Country spent/intend to spend most time abroad
State or territory of intended address/state or territory lived
Country of embarkation/disembarkation
Airport/Port of arrival/departure
Intention to live in Australia for next 12 months (not available for short-term movements)
This publication draws extensively on information provided by DIAC. This continued cooperation is very much appreciated; without it, the statistics published would not be available. Information received by the ABS is treated in strict confidence as required by the Census and Statistics Act 1905