Karel de Wit, Erasmus University Rotterdam
Project Description/Abstract
We propose a new modeling approach for univariate time series with component dynamics that correspond to different levels of aggregation. The model is called the block-autoregressive (BAR) model, as it is based on the application of a vector autoregressive model to univariate data that is partitioned into ‘blocks’ of observations. These blocks of observations are transformed to linear combinations by an orthonormal basis change, which unveils linear dynamics that would otherwise be obfuscated. Model estimation and selection are based on least squares and eigenvalue decompositions, which allow us to estimate the latent orthonormal basis of a process in addition to its regression coefficients. The goodness of fit of the estimated model can then be verified with a new chi-squared test. In our paper, we discuss several simulated and empirical examples which demonstrate both the flexibility and the explanatory potential of the BAR model.
Co-Authors
Maria Grith, Erasmus University Rotterdam
Dick van Dijk, Erasmus University Rotterdam