Comparing estimates of hunger prevalence using several divisors for per capita energy consumption

• Main Author: Arcilla, Rechel Guino
• Format: text
• Published: Animo Repository 2007
• Subjects: Energy consumption; Philippines > Economic conditions; Statistics and Probability
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/11994
• Institution: De La Salle University

Per capital energy consumption is being used to monitor the hunger situation in the country. However, choosing the divisor for per capita calculation poses a difficulty. Several competing divisors were considered: size of the household N, size of the household adjusted for elasticity of need N0.7, and total consumption unit TCU. Using the design-based combined ratio estimator rg, which is a ratio of the total number of hungry households to the total number of households, prevalence of hunger estimates were obtained for the different divisors and compared in terms of their bias, mean square error and coefficient of variation (CV).Resulting empirical cumulative distributions of per capita energy consumption differed, with the distribution using N0.7 as divisor significantly shifted to the right. This difference resulted to varying estimates of hunger prevalence: 73.77% for divisor N, 60.95% divisor TCU, and 26.17% for divisor N0.7. Nevertheless, these estimates were accurate with negligible biases relative to their standard errors,.precise having standard errors lower than 0.011, and reliable with CVs at most 3.42%.