NSCB - 4th Quarter 2004 National Accounts - Technical Notes on the Estimates of the Philippine System of National Accounts (PSNA) Series 2004-Q4

NSCB Technical Notes on the Estimates of the
Philippine System of National Accounts (PSNA) Series 2004-Q4
(posted 24 February 2005)

Technical notes on the PSNA are regularly published in the NSCB website for a better appreciation of our national accounts. Reference should also be made to the earlier technical notes for the current national accounting practices adopted in the Philippines.

Seasonally Adjusted National Accounts-

Along with the regular quarterly publication of the Philippine National Accounts, the NSCB publishes the Seasonally Adjusted National Accounts series. The following notes deal with the what, why and how of seasonal adjustment.

Seasonal adjustment is the process of removing seasonal effects from a time series (in this case the quarterly national accounts) in order to show certain non-seasonal characteristics of the economy or sector/s of the economy. An example of seasonal effect is the Christmas holiday period wherein most consumers do a lot of shopping compared to the other months or quarters of the year. After this holiday season, sales in the following months, January and February, often decline. We cannot of course readily conclude that with the drop in sales from December to January, the manufacturing or trade sector is on a decline. This is a routine seasonal pattern observed every year.

Seasonal patterns conceal other characteristics of the data that are of interest to analysts. For example, in Agriculture, Fishery and Forestry (AFF) the seasonally adjusted gross value added (GVA) posted a decline of 3.8 percent from third quarter to the fourth quarter of 2004 while the unadjusted series for the same period showed an increase of 35.0 percent. In this case, to the untrained eye, it can lead to confusion or difficulty in assessing the general direction of the AFF time series. To those who are knowledgeable about seasonal adjustment, the seasonally adjusted series is easier to compare and shows those characteristics that seasonal movements tend to hide.

The procedure of seasonal adjustment involves breaking down a series into trend, cycle, seasonal fluctuations, and irregular components:

Trend - is the upward or downward movement of the time series over a period of time;

Cycle – is the almost periodic variation in a series over several years arising out of economic reasons;

Seasonal Fluctuations – are effects that are reasonably stable in terms of annual timing, direction, and magnitude. Possible causes include natural factors (the seasonal weather), administrative measures (starting and ending dates of the school year), and socio-cultural-religious traditions (such as Christmas). Effects associated with the dates of moving holidays like Easter and Ramadan are not seasonal since they occur in different calendar months depending on the date of the holiday;

Irregular Component - anything not included in the trend, cycle, or seasonal fluctuations. Its values are unpredictable as regards timing, impact, and duration. It can arise from natural disasters, unexpected weather conditions, strikes, etc.

At present the NSCB uses the software X-11 ARIMA/2000 (Auto Regressive Integrated Moving Average), developed by Statistics Canada, to derive the Seasonally Adjusted National Accounts Series. The software is based on an iterative estimation procedure wherein the model to be tested by the software is identified by the user. Several diagnostic tests are applied to the models to determine which model best fits the series. Note that the X-11 ARIMA decomposes a series into three components, the trend-cycle, the seasonal factors and the irregular series.

Prior to 2002, NSCB utilizes the concurrent adjustment approach wherein the estimates of seasonal factors are fine- tuned as new estimates become available each quarter. This results to revisions at each reference quarter and can confuse the users. Hence, starting 2002, NSCB utilized the forward factor approach, which relies on the annual analysis of the latest available estimates to determine the factors that will be applied for the coming quarters. NSCB found minimal differences between these two approaches.

Currently, the NSCB through the Technical Committee on Seasonal Adjustment of Philippine Time Series is evaluating other approaches and softwares to further refine the existing method.

Revisions as of January 2005

Based on the Revision Policy approved by the NSCB Executive Board and consistent with the international practice on the revision of National Accounts, the following are the revisions in our previous estimates, based on the revisions and updates made by the data sources themselves, including those made by NSO on Exports, Imports, MISSI, QSPBI, Building Permits, WPI, CPI and RPI; BSP on the BOP; BAS on Agriculture and Fishery, the Forest Management Bureau on Forestry; DOE for Mining and Quarrying; NPC for EGW; as well as DBM and DOF for Construction.

For the third quarter, the following revisions were made:

On the production side,

EGW upward from 2.0% to 3.7%
TCS upward from 8.5% to 11.4%
Trade upward from 7.5% to 7.8%
ODRE upward by 6.7% to 6.9% and
Private services upward from 6.5% to 6.8%
AFF was revised downward from 7.9% to 7.3%
Mining and quarrying downward from –4.8% to –9.4%
Manufacturing downward from 4.7% to 4.2%
Finance downward from 9.9% to 9.0%

On the expenditure side,

PCE was revised slightly downward from 5.6% to 5.5%
Construction downward from 5.8% to 4.5%
Breeding Stocks downward from 5.2% to 5.0%
Exports downward from 16.5% to 15.6% and
NFIA upward from 3.3% to 5.8%

As a result of all the revisions for the third quarter, GDP growth for Q3 2004 was maintained at 6.3% while GNP growth was revised upwards from 6.1% to 6.2%. A more complete tabulation of our revisions is included in our publication.

For inquiries please contact Dir. Raymundo J. Talento at (632) 895-2481 or
e-mail him at rj.talento@nscb.gov.ph.