Schätzung des Potentialoutput und des Output Gap: Evaluation verschiedener Verfahren in der Anwendung auf die Schweiz

Potential output, the level of economic activity consistent with a stable inflation rate, and the output gap, measured as the deviation of GDP from potential output, are key concepts in macroeconomics. However, potential output is not directly observable but must be estimated – either on basis of a production function or by filtering methods. This paper tries to assess the pros and cons of the two approaches. Univariate filters decompose GDP in trend and cycle according to purely statistical criteria. Hence, it is somewhat questionable whether the extracted trend conforms to the economic notion of potential output. Moreover, substantial revisions must be expected at cyclical turning points. Better suited for addressing policy issues is the production function approach, as it derives potential output from the inputs in the production process (labor, capital, technical progress).This structural method has several drawbacks as well. It is more laborious and involves various assumptions that are to some extent a matter of judgement. The multivariate state-space model presented in this paper takes a middle course. It remains in the realm of filtering methods but links the unobserved output gap with cyclical indicators like the unemployment rate, the investment rate and the inflation rate. By including this information, the estimated trend is subject to notably smaller revisions, and the interpretation as potential output stands on more solid ground.