Periodic autoregressive conditional duration

We propose an autoregressive conditional duration (ACD) model with periodic time-varying parameters and multiplicative error form. name this (PACD). First, we study the stability properties moment structures of it. Second, estimate parameters, using (profile two-stage) Gamma quasi-maximum likelihood estimates (QMLEs), asymptotic which are examined under general regularity conditions. Our estimation method encompasses exponential QMLE, as a particular case. The proposed methodology is illustrated simulated data two empirical applications on forecasting Bitcoin trading volume realized volatility. found that PACD produces better in-sample out-of-sample forecasts than standard ACD.