| |
||||
![]() |
||||
Explanatory Note
Issues and Clarifications on the Model-Based Estimation Methodology
for the Annual Food and Poverty Thresholds
April 2007
![]()
(Posted 16 April 2007)
These Notes were prepared to address the various questions raised regarding the 2006 and 2007 model-based thresholds. It includes a compilation of the major issues and limitations of the model-based estimation methodology, together with the official clarifications of NSCB.
I. On the model used to estimate thresholds without using the actual prices
A number of competing models were generated in modeling the effective annual growth rate of the food threshold. From among the list, the final model was selected based on the statistical properties implied by the model such as consistency of statistical significance and stability of parameter estimates across different time periods. For the purpose of forecasting thresholds, emphasis is given to the accuracy of forecasts rather than the interpretability of the structure of the final model. Hence, the model structure is not supported by a known theoretical framework. This limitation is to be addressed through improvement of the underlying methodology, to be done periodically (i.e., per NSCB Resolution No. 5, Series of 2007, Approving the Model-based Estimation Methodology for the Annual Food and Poverty Thresholds, the review of the methodology will be conducted in December 2007 and May 2010).
II. On the year-on-year growth rate of the model-based estimates of thresholds
Overall, this methodology of model-based estimation of food thresholds is inclined to produce long-run estimate(s) of threshold inflation (since the variable being modeled is the effective annual food threshold growth rate). Consequently, when two time periods are being compared, there is higher probability of minimal downward bias with respect to the realized value, when the realized value is brought about by erratic price increases among basic food commodities.
In evaluating the soundness of the model-based estimates of food and poverty thresholds, ideally, one can use the information from the CPI corresponding to the bottom 30 percent of the entire income distribution. This can be done by comparing the growth rates of the former with the inflation implied from the latter. However, due to unavailability of data on CPI for the bottom 30 percent, the NSCB settled on using CPI for all income groups for comparative purposes. Hence, the growth rates that can be derived from the model-based thresholds may not be directly comparable with the available inflation computed from CPI for the entire income distribution.
III. On the release of model-based estimates of food and poverty thresholds
As approved by the NSCB Executive Board, model-based thresholds are to be generated for both FIES and non-FIES years. After the release of preliminary model-based estimates in December preceding each reference year, continuous improvement on the model-based estimation methodology is to be undertaken as more data become available. With this updated model, model-based estimates of thresholds are to be revised for non-FIES years. On the other hand, for FIES years, preliminary thresholds will be revised using the menu-based methodology using actual prices from the National Statistics Office and the Bureau of Agricultural Statistics. Further, these menu-based thresholds will be used in the computation of other poverty statistics, which include poverty and subsistence incidences (during FIES years), and other measures of poverty such as income and poverty gaps, and severity of poverty index.
IV. Statistical advocacy
Towards appreciation of model-based estimation of food and poverty thresholds among data users, the NSCB will conduct statistical advocacy and public information activities that present the limitations of the methodology, in partnership with other government agencies, such as the Bangko Sentral ng Pilipinas (BSP), among others. Two advocacy forums on the model-based thresholds are tentatively scheduled in July and August 2007, as rider activities from the statistical advocacy forums regularly conducted by the BSP.
For inquiries, please contact Ms. Jessamyn O. Encarnacion or Mr. Art Martinez, Jr. at telephone number +632-8965390 or through e-mail addresses jo.encarnacion@nscb.gov.ph and am.martinez@nscb.gov.ph.