Testing for parameter changes in linear state space models
Open Access
- 31 May 2021
- journal article
- research article
- Published by Wiley in Applied Stochastic Models in Business and Industry
- Vol. 37 (6) , 1060-1079
Abstract
Linear state space models (LSSMs) provide a very general framework for multiple time series analysis. We propose a novel statistical procedure for testing validity of a LSSM which is focused on the detection of changes in parameters of the given LSSM. We derive the moments as well as the asymptotic distribution of the test statistic, and investigate the test size and the test power for changes in means, variances, and autoregressive coefficients. In the empirical application we test the validity of LSSMs applied to daily realized volatility time series.Keywords
This publication has 38 references indexed in Scilit:
- Forecasting intraday volatility in the US equity market. Multiplicative component GARCHJournal of Financial Econometrics, 2011
- Efficient estimation of a multivariate multiplicative volatility modelJournal of Econometrics, 2010
- Seasonality with Trend and Cycle Interactions in Unobserved Components ModelsJournal of the Royal Statistical Society Series C: Applied Statistics, 2009
- A unified approach to nonlinearity, structural change, and outliersJournal of Econometrics, 2007
- Forecasting Using Principal Components From a Large Number of PredictorsJournal of the American Statistical Association, 2002
- The distribution of realized stock return volatilityJournal of Financial Economics, 2001
- Diagnostics for Time Series AnalysisJournal of Time Series Analysis, 1999
- Exact Initial Kalman Filtering and Smoothing for Nonstationary Time Series ModelsJournal of the American Statistical Association, 1997
- Evidence on Structural Instability in Macroeconomic Time Series RelationsJournal of Business & Economic Statistics, 1996
- Long memory relationships and the aggregation of dynamic modelsJournal of Econometrics, 1980