This article introduces a new class of volatility models known as Periodic log-Generalized Autoregressive Conditional Heteroscedastic (P − log GARCH) models, which incorporate periodic variations in the coefficients. These parameter variations are particularly relevant when incorporating seasonality into economic decision-making theory. The P − log GARCH formulation exhibits desirable properties, such as unconstrained positivity of parameters, absence of extreme value clustering, and a volatility trend. Specifically, the proposed model, without a trend, demonstrates periodic stationarity and is well-suited for data characterized by robust seasonal volatility. We investigate the probabilistic structure of this model and establish necessary and sufficient conditions for the existence of stationary solutions in a periodic sense. Additionally, we examine the strong consistency and asymptotic normality of the generalized quasi-maximum likelihood estimator (GQMLE) under mild assumptions. To assess the performance of our model, we conduct a Monte Carlo study to examine the finite-sample properties of the GQMLE. Finally, we present empirical evidence by applying the P − log GARCH model to analyze the exchange rates of the Algerian Dinar against the U.S. dollar and the Euro, thereby demonstrating its practical utility.
MSC Classification: 62G20 , 62M10