COLLOQUIUM ON THE STATISTICAL SCIENCES

THE SCHOOL OF STATISTICS

University of the Philippines

invites all students and faculty to a colloquium on

“Assessing Inflation Vulnerability:
An Early Warning System for Inflation in the Philippines”

to be given by

Mr. Christopher John F. Cruz
Master of Statistics (Candidate), School of Statistics

Thursday, 21 March 2013, 4:00 p.m.
UPSS Conference Room
 

Abstract

With the adoption of the Bangko Sentral ng Pilipinas (BSP) of the Inflation Targeting (IT) framework in 2002, average inflation went down in the past decade from historical average but the BSP’s inflation targets were breached 7 times since 2002. Against this context, this paper attempts to develop early warning system (EWS) models for predicting the occurrence of high inflation or inflation crisis in the Philippines. The models could complement the BSP’s existing toolkit (which include a suite of forecasting models) to assess the inflation environment as well as the risks to the inflation outlook. This paper employs two techniques, the policy-based technique (which considers the BSP’s inflation target), and the data-based technique (via Markov-switching models) to identify episodes of high and low inflation. Using the outcomes of the regime classification, logistic regression models are then estimated with the objective of quantifying the possibility of the occurrence of high inflation episodes. Empirical results show that the EWS models have some potential as a complementary tool in the BSP’s monetary policy formulation based on their in-sample and out-of sample forecasting performance.

Key Phrases: Inflation Targeting, Markov Switching Models, Early Warning System

COLLOQUIUM ON THE STATISTICAL SCIENCES

THE SCHOOL OF STATISTICS

University of the Philippines

invites all students and faculty to a colloquium on

“Robust Determinants of Poor Households in the Philippines”

to be given by

Dr. Dennis S. Mapa

Associate Professor and Director for Research

and

Prof. Manuel Leonard F. Albis

Assistant Professor

Thursday, 07 March 2013, 4:00 p.m.

UPSS Auditorium

Abstract

Identifying the poor households is an essential component of the poverty targeting strategies of the Philippine government through its Pantawid Pamilyang Pilipino Program (4Ps), or better known as the Conditional Cash Transfer (CCT) program. The poverty targeting process is a system for identifying who and where the poor households are. The system guarantees the generation and establishment of a socio-economic database of poor households. At the center of this poverty targeting process is the methodology known as the Proxy Means Test (PMT). Essentially, the PMT model is a statistical method used to predict the income of a household based on observable characteristics that correlate with, but are easier to measure than, income. In lieu of the household's actual income (per capita), its predicted income (per capita) from the PMT model is used to compare with the official poverty threshold. The households classified as poor using the PMT model are then used as one of the eligibility criteria for the CCT program (and other government poverty reduction programs). This paper looks at the Bayesian Averaging of the Classical Estimates (BACE) as a procedure in identifying the robust determinants of poor households in the country through the PMT model. Out of the 123 variables used in the models to determine the poor households, the BACE approach is able to identify 30 variables that are considered as “robust” or “strong” determinants.

Key Phrases: Bayesian Averaging of the Classical Estimates (BACE), Poverty Targeting, Conditional Cash Transfer (CCT)