Interest Rate Forecasting with Principal Component Analysis Based on Long-Run Covariance Matrix

Principal component analysis (PCA) is one of the most important methods in analyzing and forecasting the term structure of interest rates. However, there are strong indications that it is not adequate to estimate interest rate factors by traditional PCA when there is time dependence and measurement errors. To correct these problems, it is recommended to use the long-run covariance matrix to estimate the principal components, extracting the correct covariance structure present in these processes. In this work, we show that out-of-sample forecasts for the term structure of interest rates constructed with the PCA using long-run covariance matrices appear to be more accurate compared to predictions based on static covariance matrices.