Modelling asset returns in the presence of price limits with Markov-switching mixture of truncated normal GARCH distribution: evidence from China

This article proposes a general framework of a Markov Switching GARCH model with a mixture of truncated Gaussian to model asset returns with price limits in China. Theoretically, while retaining many convenient statistical properties of the Gaussian distribution, the proposed model also assumes a flexible time-varying volatility structure to accommodate the feature of the return data under price restrictions in China, such as the clusters near the bounds (due to the 'bound effect'). Empirically, we apply the model to eight representative stocks from Shanghai and Shenzhen stock markets in China. Lastly, we find that our proposed model dominates the conventional volatility models in terms of Value-at-Risk measures.