Univariate Potential Output Estimations for Hungary

Potential output figures are important ingredients of many macroeconomic models and are routinely applied by policy makers and global agencies. Despite its widespread use, estimation of potential output is at best uncertain and depends heavily on the model. The task of estimating potential output is an even more dubious exercise for countries experiencing huge structural changes, such as transition countries. In this paper we apply univariate methods to estimate and evaluate Hungarian potential output, paying special attention to structural breaks. In addition to statistical evaluation, we also assess the appropriateness of various methods by expertise judgement of the results, since we argue that mechanical adoption of univariate techniques might led to erroneous interpretation of the business cycle. As all methods have strengths and weaknesses, we derive a single measure of potential output by weighting those methods that pass both the statistical and expertise criteria. As standard errors, which might be used for deriving weights, are not available for some of the methods, we base our weights on similar but computable statistics, namely on revisions of the output gap for all dates by recursively estimating the models. Finally, we compare our estimated gaps with the result of the only published Hungarian output gap measure of Darvas-Simon (2000b), which is based on an economic model.