Determinants of Measurement Errors in Immigration Flow Statistics: A Case of Nine Central & Eastern European Countries

This study analyses the determinants of measurement errors in immigration flow statistics in nine Central and Eastern European Countries (CEECs) between 2007 and 2019. This research is based on immigration flow statistics reported to Eurostat and a synthetic dataset of immigration flows generated as a part of this study. Based on these, a within-between random effects (REWB) model is employed to analyse the effects of public sector corruption, government effectiveness and research and development (R&D) in the government sector on the level of undercount and accuracy in available migration statistics. Additionally, an auxiliary analysis is conducted using a random effect model to identify if administrative register maintenance costs, dynamic migration processes and financial allocations can be attributed to missing disaggregated immigration flow data. This study identifies that there exists a non-linear relationship between public sector corruption and the measurement errors. Initially, the measurement errors drop with rising levels of corruption and increase beyond a specific threshold, partially supporting the perception of public sector corruption as an act of ‘greasing the wheels’. While public sector corruption is shown to decrease the measurement error, the directionality is attributed to overall low levels of corruption amongst the CEECs and the fact that the inflection point lies in the tail end of our sample. An additional conflicting finding of the study is that R&D expenditure seems to increase the level of inaccuracy and that government effectiveness has no effect on the measurement errors in the given context. However, the analysis of missingness reveals that increases government sector R&D expenditure does reduce the probability of missing data, though, the results hold only if countries with no missing data are excluded from the sample.