Search

In most macroeconomic models, variations in nominal variables, such as inflation or money growth, are considered to be important determinants of cyclical fluctuations in real activity. The major hypothesis in that respect is the so-called Phillips-curve. In its modern interpretation, it maintains that only unexpected changes in nominal magnitudes produce real effects. Reliable empirical evidence on these effects is therefore crucial for the building of macroeconomic models and the conduct of monetary policy. Time series of output, industrial production or employment however contain growth and seasonal components in addition to cyclical elements. The empirical implementation of business cycle models therefore requires assumptions with respect to growth and seasonal parts as well, in order to isolate cyclical movements and to avoid misspecified equations. It has become general practice to assume that economic growth can be reasonably well approximated by a deterministic linear time trend. Seasonality is either captured through the explicit introduction of dummy variables or the use of seasonally adjusted data. Again, these procedures assume a deterministic seasonal structure. It is generally concluded in this literature, that unanticipated and possibly also expected changes in nominal magnitudes have non-negligible real effects

In most macroeconomic models, variations in nominal variables, such as inflation or money growth, are considered to be important determinants of cyclical fluctuations in real activity. The major hypothesis in that respect is the so-called Phillips-curve. In its modern interpretation, it maintains that only unexpected changes in nominal magnitudes produce real effects. Reliable empirical evidence on these effects is therefore crucial for the building of macroeconomic models and the conduct of monetary policy. Time series of output, industrial production or employment however contain growth and seasonal components in addition to cyclical elements. The empirical implementation of business cycle models therefore requires assumptions with respect to growth and seasonal parts as well, in order to isolate cyclical movements and to avoid misspecified equations. It has become general practice to assume that economic growth can be reasonably well approximated by a deterministic linear time trend. Seasonality is either captured through the explicit introduction of dummy variables or the use of seasonally adjusted data. Again, these procedures assume a deterministic seasonal structure. It is generally concluded in this literature, that unanticipated and possibly also expected changes in nominal magnitudes have non-negligible real effects