Should we track migrant households when collecting household panel data? Household relocation, economic mobility, and attrition biases in the rural Philippines

Based on household panel data that tracked migrant households (with an additional survey cost of 17 percent), this article describes behavior of household relocation and quantifies the extent of attrition biases in estimating the determinants of percapita household consumption and of its growth rate. Many households relocate for non-economic reasons, and to rural destinations, while the small number of urban migrants improved their wellbeing faster than did others. Such heterogeneity among migrants may be a reason behind the negligible attrition biases caused by the omission of migrants, in the inference on the average behavioral coefficients among the original population.