Periodicity and Stochastic Trends in Economic Time Series

This is an advanced graduate textbook in econometrics. A large proportion of the data studied by econometricians are series of observations of the same variables made over time (time series). This book provides a comprehensive account of how to allow for seasonal fluctuations in these data by using periodic models.

---

This book provides a self-contained account of periodic models for seasonally observed economic time series with stochastic trends. Two key concepts are periodic integration and periodic cointegration. Periodic integration implies that a seasonally varying differencing filter is required to remove a stochastic trend. Period cointegration amounts to allowing cointegration part-term adjustment parameters to vary with the season. The emphasis is on econometric models that explicitly describe seasonal variation and can reasonably be interpreted in terms of economic behaviour. The analysis considers econometric theory, Monte Carlo simulation, and forecasting, and it is illustrated with numerous empirical time series. A key feature of the proposed models is that changing seasonal fluctuations depend on the trend and business cycle fluctuations. In the case of such dependence, it is shown that seasonal adjustment leads to inappropriate results.