All Times ET
Keywords: Shrinkage, Quantile regression, Instrumental variable, Portfolio
This paper uses shrinkage covariance by investigating factor models. We improve the model from several perspectives. Firstly, we use quantile regression to deal with tail effects in data distribution. Secondly, state variables such as the GDP states are common in stock performance analysis. We use instrument variable (IV) regression to deal with discrete variables. Last but not least, shrinkage is a way to handle high-dimensional data. We present an example using Shanghai stock index comprising of 50 stocks. The results show that a quantile shrinkage model with IV has smallest standard deviation.