A0912
Title: Extract investor sentiment from price disparity via model-based neural networks
Authors: Hao Ma - Queen Mary University of London (United Kingdom) [presenting]
Abstract: The aim is to show how to identify and estimate investor sentiment. By exploiting the price disparity of dual-listed stocks, it is shown how to identify stock-specific market-wide investor sentiment based on the noise trader theory. Structural estimation is then conducted via deep learning to estimate the Chinese investor sentiment as a general function of firm-level characteristics. This novel model-free sentiment indicator extends the understanding of what and how characteristics drive sentiment dynamics. The framework is further used to extract the sentiment component of each stock in the Chinese stock market and test a wide range of behavioral theories.
