|Page tools: Print Page Print All
FINAL NOM ESTIMATES
6.19 Travellers arriving in Australia mark their incoming passenger cards with an indication of their intended duration of stay as:
6.20 Travellers departing Australia are required to indicate whether they are:
6.21 Australian residents departing are required to indicate their intended length of stay overseas, so as to distinguish between long-term resident departures (LTRD) and short-term resident departures (STRD).
6.22 The categories of long-term resident return (LTRR) and short-term resident return (STRR) are based on actual (continuous) duration of stay overseas. Incoming and outgoing passenger cards are matched and this information is supplied to the ABS by DIAC. Similarly the categories of long-term visitor departure (LTVD) and short-term visitor departure (STVD), are based on actual (continuous) duration of stay in Australia. A record of departure is matched with their last arrival to calculate duration of stay. ABS then converts this duration into short-term (under 12 months) and long-term (12 months or more).
6.23 While each traveller may make a number of movements during a reference quarter, for the purpose of measuring NOM under the improved methods, they are assigned one and only one category of travel for the reference quarter. See Step 2 below for a description of how this is determined.
Implausible travel sequences
6.24 There are circumstances where implausible travel sequences appear in the data. For example, a traveller is recorded as having two sequential arrivals in Australia without a departure in between, or conversely, two departures from Australia without an arrival. In these instances the implausible travel sequences are repaired using a logical imputation. The repair of implausible sequences is necessary in order to derive an estimate of duration of stay or absence, since time spent in Australia is derived by summing up the duration between each arrival and departure pair.
Final estimates calculation steps:
6.25 Final NOM estimates are calculated in the following sequential order.
Step 1: Derive person-level data on overseas movements
6.26 Data from the Overseas Arrivals and Departures (OAD) collection, along with additional data obtained from the Travel and Immigration Processing System (TRIPS) which is sourced from DIAC.
6.27 Each OAD file contains information derived from incoming and outgoing passenger cards, and is matched for each movement to selected TRIPS data items through a unique personal identifier. OAD files do not include records where a passenger card has not been matched to TRIPS (e.g. if the traveller records incorrect passport details on their passenger card, and the mismatch between this and the TRIPS record has not yet been resolved). Records from TRIPS relating to these unmatched movements are extracted and added to the quarterly OAD files for NOM estimates.
6.28 The quarterly files used for NOM estimates also contain a number of logical edits and imputations, mainly at the movement level. Person-level data are constructed from these movement data by matching movements using personal identifiers.
Step 2: Determine the category of travel for each overseas traveller
6.29 Travellers are assigned to one, and only one, category of travel during a reference quarter. The algorithm used to assign a category of travel to each traveller is based on:
Step 3: Derive ERP flag at start of reference quarter
6.30 An 'ERP flag' is used to indicate whether a traveller is 'IN' or 'OUT' of the ERP at the start of the reference quarter. During the start-up period for the improved NOM methods, this flag has been determined by looking at the individual's previous travel history.
6.31 For subsequent periods, a traveller's ERP flag at the start of a reference quarter will be carried forward from their flag at the end of the previous quarter. If the traveller has no flag in any of the previous six quarters, their ERP flag at the start of the reference quarter will be imputed based on the direction of their first movement within the quarter i.e. 'OUT' of ERP if they are arriving in Australia; 'IN' ERP if they are departing from Australia.
Step 4: Derive movement history and required data items for each overseas traveller
6.32 Movement histories are constructed for each traveller based on their overseas arrivals and departures during the 16 months following an overseas movement that takes place in the reference quarter. In addition, a number of key demographic data items are derived for each traveller, including their financial year of birth (used to calculate age at 30 June), sex, country of birth and country of citizenship. In most cases these derived items are available from the source OAD and/or TRIPS datasets. In cases where required demographic data items are missing, logical imputations similar to those applied in existing OAD and ERP systems are used.
Step 5: Identify implausible movement sequences, and impute 'missing' movements
6.33 Implausible movement sequences appear in the data for a number of reasons, but are mainly due to non-matches between travel information (i.e. visa applications or passport information) and existing TRIPS information. Non-matches can occur when a traveller's personal details change (e.g. marital status, family name), or when their travel documentation is updated (e.g. new passport and country of passport, perhaps combined with changes to personal details). When a non-match occurs, the DIAC processing systems assign a new personal identification number to the movement before referring it to a resolution process. While most non-matches are resolved at a later date, administrative data provided to the ABS for the OAD or from TRIPS may not include all revisions relating to this process.
6.34 Some travellers with implausible movement sequences in the data will also have a sequence (before and/or after the implausible sequence) of plausible movements over the 16 month period studied. In these cases, an assumption is made that the proportion of a traveller's time spent in Australia during the implausible movement sequence was consistent with their proportion of time spent in Australia during plausible movement sequences. Using their plausible movement sequences, a ratio of their time spent in Australia to their time spent overseas is calculated and applied to the implausible sequence to impute a 'proxy' arrival or departure movement.
6.35 For a very small number of travellers, there may be no plausible movement sequence in the data during the 16 month period analysed (e.g. a movement history may only show two or more overseas arrivals, or only show two or more overseas departures). It is assumed that these travellers spent 50% of the time between movements in Australia.
Step 6: Determine the total duration of stay for each traveller and calculate ERP flags
6.36 The total duration of stay/absence for each traveller is determined by adding durations of stay/absence as shown by movement histories over the 16 month period following an overseas movement.
6.37 As shown in the following diagram, these durations of stay are used to calculate whether a traveller who is 'IN' or 'OUT' of the ERP before the movement is 'IN' or 'OUT' of the ERP after the movement, regardless of their category of travel.
Step 7: Calculate NOM estimates and final impact on the ERP
6.38 Each traveller moving into the ERP during a reference quarter (i.e. an ERP flag of 'OUT' at the start of the quarter and a flag of 'IN' at the end of the quarter) is added to the total NOM estimate for the quarter. Similarly, each traveller moving out of the ERP is subtracted from the NOM estimate. Travellers whose initial and final ERP flags for the quarter are the same make no contribution to NOM. As in current publication tables, NOM estimates can be broken down according to the demographic characteristics of travellers (e.g. state or territory of usual residence, age or sex).
Step 8: Migration adjustment value
6.39 For final estimation, a 'migration adjustment' (MA) value is derived for each traveller in the reference quarter, as a by-product. The MA value is equal to the difference between the traveller's actual contribution to NOM for the quarter (i.e. -1, 0 or +1) and their initial contribution to NOM based solely on their category of travel (i.e. +1 for a permanent or long-term arrival, -1 for a permanent or long-term departure, and 0 for a short-term arrival or departure). The MA value is not required to determine final NOM estimates for the reference quarter, but is used for estimating preliminary NOM for the corresponding quarter two years later.
State or territory of usual residence
6.40 The distribution of improved NOM estimates across states and territories is based on information as reported by travellers on incoming and outgoing passenger cards. There are two data items (State 1 and State 2) that are derived to determine a traveller's state or territory of residence/stay.
6.41 Derivation of the first data item (State 1) is dependent upon the direction of travel (either arrival or departure) and on the type of traveller (either visitor or a resident). For a visitor arrival the state or territory of residence (State 1) is the state or territory where they intend to stay (as indicated on the incoming passenger card). For a visitor departure it is the state or territory in which the traveller states they spent the most time. For a resident arrival it is the state or territory of their intended address in Australia, and for a resident departure it is the state or territory in which they lived. This information is obtained from incoming and outgoing passenger cards.
6.42 As a short-term visitor to Australia may move state or territory during their time in Australia, the second data item (State 2) is used to code them to the state or territory where they spent the most time as reported on their subsequent outgoing passenger card.
6.43 For example, if a short-term visitor arrives in June 2005 and this is their first arrival in Australia, they are allocated to the state or territory they have indicated on their passenger card as their state of intended stay (State 1). However, it is only possible to finalise their contribution to ERP for a reference quarter through the collection of future data i.e. the 16 months following the reference quarter. Therefore, a traveller history is collected over 16 months after the June 2005 reference quarter. If it is found through the traveller's history that they have stayed 12 out of 16 months, then in the June 2005 reference quarter this traveller would be allocated a State 2 value based on their outward movement subsequent to their category of travel movement (i.e. as stated on their passenger card). This would occur regardless of whether they had made numerous arrivals and departures during the 16 month period.
6.44 If it is found over time that the traveller has not stayed in Australia 12 out of 16 months they keep their State 1 allocation for the reference quarter, and as they have not met the required length of time for residency they will not be counted in final NOM.
6.45 State 2 is also derived for short-term visitor arrivals who did not leave Australia at all during the 16 months follow-up period. State or territory of residence for this group is imputed using the State 1 and State 2 distributions of long-term and short-term visitors who have spent more than 12 months, out of 16, in Australia and have made a subsequent departure movement.