Using a spree estimator to estimate the number of unemployed people across subregions

Abstract

The methodology of small area estimation (SAE) plays an important role in the field of modern information gathering, which aims to cut survey costs while lowering the responent burden. SAE methods have an advantage over clasical methods since they enable reliable estimates at lower levels of spatial aggregation and with more domains, where the representative approach displays too much variability. This means that small area estimation can be used to handle cases with few or no observations for a given domain in the sample. However, cell total estimates for lower levels of spatial aggregation or subpopulations tend to differ from estimates calculated by means of higher levels of representation, which is possible due to their adequate sample size. One way of coping with this incompatibility is by applying a SPREE estimator. This is used to adjust the values in the cells of an estimated contingency table to the totals obtained by means of the representative method. Internal cells can initially be filled with data from previous censuses, or current administrative registers. The method seems to be particularly useful for estimating the parameters of the labour market, since the methodology used in the Labour Force Survey can only yield data at the level of a province. The users of statistical data, however, expect information which is more geographically disaggregated. Considering the above, the aim of the present paper is to demonstrate the potential of the SPREE estimator for estimating the number of unemployed at the level of subregions in the Wielkopolska province using data from the unemployment register and the Labour Force Survey.