The Treasury

Global Navigation

Personal tools


An Analysis of Benefit Flows in New Zealand Using a Social Accounting Framework WP 13/01

Appendix B. The Benefit Dynamics Dataset

The BDD dataset is derived from the SWIFTT benefit payments system. Historical data generated by taking snapshots of SWIFTT data at given points in time are held on the Information Analysis Platform (IAP). In their raw form, the historical data do not readily permit analysis of benefit dynamics. The variables of interest are distributed across a number of different structures and complex matching routines are required to link them with spells on benefit. In addition, the structures are characterised by a considerable amount of “noise”: multiple records must be examined to establish a chain of events and, in some cases, retrospective action means that these cannot be established by working through records chronologically. The BDD links information recorded in SWIFTT for the same individual over time, strips away the noise in the historical data, and organises the data in a form amenable to longitudinal analysis. In the December 2003 update, the data set was enhanced to incorporate selected variables from the job seeker register system SOLO and the Unified Client View Phase Two (UCVII) system, other administrative data systems held by MSD.

The BDD set is stored in the secure environment of the IAP and access is restricted to a small number of authorised analysts. The data set contains no name and address information that could be used to identify individual benefit recipients.

The dataset covers the period from 1 January 1993 to the date of the most recent update. From October 2010, the data are updated at the end of each month (previously updates were quarterly). These updates are each created with a lag of at least two to three months to allow retrospective action to 'bed down' records for the end of the period.

Because of retrospective action, successive versions of the BDD will give a different benefit history for small numbers of individuals, and analyses repeated on a later version of the data set may produce slightly different results. In order to minimise interruption to projects using a given version of the data set, the most recent update and the one that preceded it are both available through the IAP. The year and month of the update is indicated by the four digits at the end of the name of each file - 1006 indicates the update to 30 June 2010 for example.

For the period up to 1 June 1996, the data set is based on consolidated files of historical SWIFTT data. The consolidated files reflect a snapshot of the situation of each individual in SWIFTT at monthly intervals only. For variables which do not have a change date recorded (such as the partnership status indicator APORT and the district office variable DIST), this leads to some imprecision in the timing of events.

Page top