Aggregation and euro area Phillips curves

Abstract

This paper examines the issue of the impact of aggregation in the empirical analysis of euro area labour markets. A Phillips Curve describing the adjustment of unit labour costs is estimated at the national and aggregate level for the 5 largest euro area countries. Potential sources of aggregation bias are investigated - such as differences in parameter coefficients and a lack of correlation in the independent variables across countries - as well as the potentially offsetting statistical averaging effect. Finally the out-of-sample forecasting performance of both approaches is evaluated. The results point to some limited advantages of analysing wage developments at the national rather than at the area-wide level. The paper concludes that if major advantages in undertaking national analysis do exist, they are likely to arise from the ability to develop country-specific structures for the Phillips Curves and not from aggregation biases that emerge when a common structure is used.