Return direction forecasting: a conditional autoregressive shape model with beta density

Abstract This paper derives a new decomposition of stock returns using price extremes and proposes conditional autoregressive shape (CARS) model with beta density to predict the direction returns. The CARS is continuously valued, which makes it different from binary classification models. An empirical study performed on US market, results show that predicting power not only statistically significant but also economically valuable. We compare probit model, demonstrate proposed outperforms for return forecasting. provides framework