An Industry-Based Prediction Model for Regional Unemployment Rates Based on Officially Available Data

The paper attempts to determine regional unemployment rates for NUTS 2 regions based on data available on a regular basis from Eurostat and ILO. In the first part it is shown how the evolution of regional unemployment rates is determined by national unemployment rates by industry and the regional employment structure by industry, with key relationships between industries being identified. In the second part, a model of predicting regional unemployment rates from national unemployment rates by industry and regional employment structure is proposed, using unbalanced panel regression models. Specific data mining modeling approaches, not currently used in economics research, are employed to produce and validate the predictions. Conclusions point out that reliable predictions of regional unemployment rates can be done using national unemployment rates by industry and regional employment structure, using models that could be built and used by policymakers