Structural breaks and GARCH models of exchange rate volatility: Re‐examination and extension
通过复制和稳健性检验Rapach与Strauss的汇率波动率预测研究，发现纳入结构断点的GARCH模型在所有损失函数和预测区间上均优于忽略不稳定性的模型。
Summary We examine the empirical significance of structural changes concerning generalized autoregressive conditional heteroskedasticity (GARCH) models of exchange rate volatility using out‐of‐sample tests by replicating and carrying out robustness checks on the volatility forecasting study by Rapach and Strauss (Journal of Applied Econometrics, 2008; 23, 65–90). We employ the same econometric models but incorporate recent US dollar daily exchange rates data while also using different software, a relatively recent forecast accuracy test and loss metrics. Our objective is to attain scientific replication in a broad sense. Our analysis verifies and broadly aligns with the results obtained in the original study. In particular, we find strong evidence that the models incorporating structural breaks demonstrate superior performance across all loss functions and forecast horizons compared with those models that ignore instabilities.