The short lived EUROLINK model. Centralized and Decentralized approaches to Multi-country models at the European Commission

In 1979, Paolo Ranuzzi, a senior economist within the Commission’s Directorate General for Economic and Financial Affairs (DG II), led the development of ‘EUROLINK’, a short-term macroeconometric multi-country model of the European Economic Community (EEC). EUROLINK was used by the Commission for a few years, for producing macroeconomic forecasts, the EEC budget, and various studies about the dynamics of the EEC economy. However, after two years of intense criticisms, the use and development of EUROLINK was abandoned in 1983. This article documents the history of the short-lived EUROLINK model. We highlight the reasons pushing the DG II to develop such a multi-country model, then the reasons bringing to its abandonment. We argue that, compared to single-country macroeconometric models, multi-country models were considered more suitable to analyse, theoretically and quantitatively, interdependencies across national economies and spillover effects of national policies. To address these issues, EUROLINK combined, via original bilateral trade equations, four heterogeneous large-scale macroeconometric models of European countries, developed by national modelling teams. We characterise this methodology as the “decentralised approach.” Thanks to original archives, we show how this approach was deemed overly complicated and costly by the DG II. After EUROLINK, DG II economists shifted to a different modelling approach (the “centralised approach”), in which the multi-country model (COMPACT, then QUEST) combines identical models of national economies, all entirely built by the DG II team. This approach was considered as one preserving DG II’s “intellectual command” over the modelling activities.