Buffered Autoregressive Models With Conditional Heteroscedasticity: An Application to Exchange Rates
提出BAR-GARCH模型，能同时捕捉条件均值和方差的缓冲现象，用于汇率数据，比传统AR-GARCH和门限AR-GARCH模型更优。
This article introduces a new model called the buffered autoregressive model with generalized autoregressive conditional heteroscedasticity (BAR-GARCH). The proposed model, as an extension of the BAR model in Li et al. (2015), can capture the buffering phenomena of time series in both the conditional mean and variance. Thus, it provides us a new way to study the nonlinearity of time series. Compared with the existing AR-GARCH and threshold AR-GARCH models, an application to several exchange rates highlights the importance of the BAR-GARCH model.