Applying and interpreting model-based seasonal adjustment. The euro-area industrial production series

The recent economic crisis has altered the dynamics of economic series and, as a consequence, introduced uncertainty in seasonal adjustment of recent years. This problem was discussed in recent workshops at the European Central Bank and at Eurostat in the context of adjustment of the Euro Area Industrial Production (EPI) series. Because a seasonal component is unobserved and undefi ned, it is diffi cult to compare results from different adjustment methods. Within the regARIMA model-based approach, however, a framework for systematic analysis and comparison of results is indeed present. The EPI series is analyzed under the TRAMO-SEATS framework. The purpose of the analysis is not to compare alternative methods, but to show how the results of the model-based analysis can be exploited at the identifi cation, diagnostics, and inference stages of modeling, and in the selection of an appropiate seasonal adjustment (and underlying model). Despite the uncertainty induced by the crisis (and the revisions to the unadjusted data), the automatic procedure, with ramps to capture the spectacular 2008 drop in the series, provides excellent and stable results.