This paper uses parametric bootstrapped within-groups (WG) and first-difference generalized method of moments (FD-GMM) estimators of a dynamic panel data (DPD) model. The parametric bootstrap of WG and FD-GMM estimators are known to work well for small sample cases wherein the variances of the individual effects and the disturbances are almost of the same magnitude. The bootstrap can be used when sample size is small enough to achieve asymptotic optimality of estimators. These estimation procedures are used to characterize how the Filipino household changes how they prioritize different expenditure items over time and across regions using the Family Income and Expenditure Survey (FIES) data with panels formed at the regional level.