Multilevel Analysis of Urban Poverty in Sri Lanka

Urban poverty in Sri Lanka has long been neglected and given less priority in policy interventions and academic contributions following favorable nominal poverty statistics which underscore the need for urgent attention in matters pertaining to urban livelihood. The negligence of the urban sector and the urban poor for many decades has created a severe dearth of academic contributions. Thus the main objective of this paper is to examine poverty covariates in the urban sector to void some of these knowledge gaps which are vital to be addressed at the current development phase in the country. This study utilizes multilevel models to analyze poverty variations in the urban sector taking both individual and contextual levels simultaneously. It was found that about 11 per cent of total error variance of poverty was attributed to the contextual level. While age, gender, family size, education level and employment status significantly contributed at the individual level; female labor force participation, magnitudes of main employment sectors: agriculture, service and industry were significant at the structural level. It was evident that the risk of poverty varied across districts considerably but there was no evidence of significant variations in the patterns of covariates. Findings of this study suggest that poverty alleviation policies should focus on multilateral way. Keywords: Urban poverty; Asset index; Multilevel mixed models; Sri Lank