The single most pressing economic problem nowadays in many countries including Thailand is household debt. The household debt problem affects the whole economy and living standards of the country. To solve such problem, precise information on household debt is crucial for policy makers to make correct decisions. Therefore, an efficient estimation method is required to obtain precise estimates of the household debt information. In this paper, we apply two small area models: the Fay-Herriot model and the hierarchical Bayesian Fay-Herriot model, to estimate the household debt as alternative methods to the existing direct estimates. Results from our study suggest that the estimates obtained from the two models outperform the direct survey estimates.
Keywords
small area estimation, Fay-Herriot model, hierarchical Bayesian Fay-Herriot model, empirical best linear unbiased predictor, household Debt