INTERPRETATION OF FRACTAL DIMENSIONS IN POVERTY ESTIMATION

Abstract

The paper analyzed the poverty incidence data among families in the Philippines across the different provinces for three periods: 2006, 2009, and 2012, using a surrogate measure of data roughness through fractal dimensions. Fractal dimensions as statistical quantities are robust measures which are not easily influenced by extreme observations unlike other variance-based poverty measures. Results showed that poverty incidence is most pronounced and severe in 2009 and significantly least severe in 2012. This did not coincide with the classical analysis of the National Statistical Coordination Board which averred no significant improvement in the poverty situation in the Philippines using the same data report. Moreover, multifractal spectrum analysis revealed that some provinces which used to be members of the 20 poorest provinces were delisted in 2012. Attribution of the improvement in the poverty situation are discussed in the paper considering some of the poverty alleviation projects in the Arroyo administration and the current Aquino administration’s poverty alleviation thrusts.