Statistical matching of income from FIES to CPH: The Central Visayas case study

• Main Authors: Duazo, Kathleen Joy L., Tolentino, Pauline Anne L.
• Format: text
• Language: English
• Published: Animo Repository 2016
• Subjects: Income > Philippines > Central Visayas; Statistical matching; Central Visayas (Philippines) > Census; Statistics and Probability
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/14912
• Institution: De La Salle University

Census of Population and Housing (CPH) and Family Income and Expenditure Survey (FIES) are both nation-wide surveys that give information regarding the country. While both surveys are individually limited in information, they complement one another for the purpose of providing policy-makers more detailed integrated information. A framework that offers an approach is statistical matching. In this paper, multiple parametric and unconstrained non-parametric methods were examined with respect to the goal of matching CPH with total household income based on observations in FIES. Using conditional mean matching and draws based on conditional predictive distributions under the parametric framework, results showed that using log-transformed data at global-level matching is better in both sensibility and consistency. On the other hand, results from the distance hot deck method under the nonparametric framework showed that the provincial-level is preferred due to considerations in sensibility of the results across different wealth classes and consistency of the generated distributions to those expected from FIES, particularly for the provinces of Siquijor and Negros Oriental. Overall, matching income to CPH from FIES at the global-level under the linear model parametric approach is most for its results being sensible, when evaluated across different wealth percentiles, and being consistent, when compared to corresponding income percentiles estimated from FIES.