Estimating poverty and vulnerability to monetary and non-monetary poverty: the case of Vietnam

Drawing on three-wave panel data from the Vietnam Housing Living Standard Surveys (VHLSS) 2010, 2012, and 2014 and employing a fuzzy method, this paper estimates chronic and transient poverty across multiple dimensions (income, education, health, housing, basic services, durable assets, economic status) in Vietnam. Using standard deviation as a measure of risk, this study further defines vulnerability as a probability for becoming poor and estimates vulnerability to poverty from the stochastic variation of expected deprivation within a defined interval. We further apply the method of multilevel analysis to assess the deprivation of households and distinguish vulnerability as influenced by idiosyncratic (household-specific-level) and covariate (province-level) shocks. It is observed that while the number of chronic poor in all dimensions is quite low, the proportion of chronic poor in the housing dimension is the highest (around 5% over the applicable years nationwide). Regional variation in non-monetary dimensions of poverty is substantial and clearly distinct from monetary poverty. We show that there are more multidimensionally poor households that are vulnerable to idiosyncratic shocks than to covariate shocks, and the proportion of vulnerable households (to covariate shocks) in the housing dimension is significantly greater than that in other dimensions. Almost all covariates of household and province are significantly different between vulnerable and non-vulnerable groups across the multiple dimensions of poverty other than health. Our findings suggest an urgent need for policy attention on the explicit nature of vulnerability and on the many dimensions of poverty in specific regions, and to look beyond the current official monetary-based approach.