Page 368 - Contributed Paper Session (CPS) - Volume 2

CPS1888 Chen C. et al.
are obtained by adding up the three-month GST value of these units. We
assume there is no uncertainty arising from the use of modelling and/or
sampling, as actual data is used.
Direct transformed approach - The direct transformed approach uses one
part of the GST data that is fit for use, to estimate another part of the data
which can meet most our quality criteria, but the reporting frequency is not fit
for the purpose of producing quarterly outputs. This method is applied to the
GST two-monthly and six-monthly filers with simple reporting structures.
• For two monthly filers we always have one month of GST data for the
reference quarter missing, either the first month data or the third month
data, depending on the reporting type of the filer. Modelling in
conjunction with monthly filers is used transform the two-monthly data
to quarterly data, with no other data sources involved. The estimated
quarterly GST value of a two-monthly filer can be calculated as follows
̂
̂ =

Where is a ratio of the total GST value of eligible monthly filers for the three
̂

months of the reference quarter, to their GST value for the two months where
the two-monthly filer files a GST return.

∑ 3 ∑ ℎ
̂
= =3−2 =1 ,,

∑ ∑ ℎ
=−1 =1 ,,

Where
ℎ = ℎ
= ℎ
= ℎ [1~4]
= ℎ ℎ ℎ ℎ ℎ ℎ
[1~12] =

= ℎ = {

= {1, … , }, ℎ ℎ ℎ
ℎ

For the two-monthly data we have implemented a quality indicator in selected
outputs, measuring the variability in the aggregate estimates that occurs in
the data transformation.

• Six monthly filers can file GST returns in any two months of the year that
are six months apart. That means we always have to deal with data filed
for six different months and which contains the reference quarter. Firstly,
we disaggregate six-monthly data to two-monthly data based on the

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