Empirical economists often filter data prior to analysis to remove features that are a nuisance from the point of view of their theoretical models. Examples include trends and seasonals. This article describes how data filters work and the rationale that lies behind them. It focuses on the Baxter–King and Hodrick–Prescott filters, which are popular for measuring business cycles.
KeywordsARIMA models Band-pass filters Baxter–King filter Business cycle measurement Cramer’s representation theorem Data filters Deterministic linear trends Gaussian log likelihood Generalized method of moments Granger causation High-pass filters Hodrick–Prescott filter Impulse response function Rational-expectations business-cycle models Seasonal adjustment Seasonal fluctuations Shocks Spurious cycle problem Stochastic general equilibrium models Stochastic trends Trend reversion Vector autoregressions
JEL ClassificationsC2 C4
- Benati, L. 2001. Band-pass filtering, cointegration, and business cycle analysis. Working Paper No. 142, Bank of England.Google Scholar
- Blanchard, O., and D. Quah. 1989. The dynamic effects of aggregate demand and supply disturbances. American Economic Review 79: 655–673.Google Scholar
- Fok, D., P. Franses, and R. Paap. 2006. Comparing seasonal adjustment methods. In Palgrave handbook of econometrics: Volume1; Econometric theory, ed. T. Mills and K. Patterson. Basingstoke: Palgrave Macmillan.Google Scholar
- Granger, C., and P. Newbold. 1986. Forecasting economic time series. New York: Academic Press.Google Scholar