Page 367 - Contributed Paper Session (CPS) - Volume 2

CPS1888 Chen C. et al.
2.1 Data assessment
Various aspects of GST data are examined within this model, such as
reporting structures, timeliness, conceptual alignment, and reporting
frequency. These are shown as decision boxes in Figure 2.
Conceptual alignment - Administrative data has not been collected for
statistical purposes, therefore its alignment to the conceptual and definitional
requirements of our outputs needs to be assessed. We take the annualised
GST values and compare it with our annual financial data and found that in
most case the GST data can be used in the production of quarterly outputs.
We also developed a GST outlier strategy to effectively remove sales of large
capital items, which may have material impact on levels and movements in
some industries.
Timelines –To use administrative data in our quarterly outputs we must
have timely data. A significant finding in our assessment is the fact that at least
85 to 95 percent of GST data by value is available within current quarterly
production timeframes. Wherever the GST data is unavailable, we developed
a GST imputation strategy to obtain fit-for-purpose data.
Reporting structures - Administrative data is usually collected from legal
units and this is also the case with GST. Legal units that are most likely to be
statistically fit for use are those that correspond to statistical units with activity
predominately in a single industry. As the unit structures become more
complex, e.g. legal units linked by ownership to provide (referred to as GST
groups), the use of GST data becomes more problematic. With the assessment
model, we found that over 95% businesses by count have a simple reporting
structure and the GST data can be used directly.
Reporting frequency – Businesses in New Zealand can file GST data on a
monthly, two-monthly (referred to as TA if file in odd months and as TB if file
in even months) and six-monthly basis depending on turnover size. For
quarterly outputs, a monthly reporting frequency is ideal as data can be easily
added up to a quarterly value. Businesses with a reporting frequency of more
than a month require some form of transformation to enable the publication
of quarterly data.

2.2 Transformation methodology
Depending on the data assessment and its fitness for use, the most
appropriate transformation methodology is established for the various ‘parts’
of the business population.
Direct unmodified approach - The direct unmodified approach uses the
GST data as it is received (with imputation and capital adjustment in place).
This is applied to GST monthly filers with a simple structure, and small GST
monthly groups where the reporting unit or a single legal unit within the group
accounted for greater than 80 percent of the group value. The sales measures

356 | I S I W S C 2 0 1 9