Forecasting the changes in daily stock prices in Shanghai Stock Exchange using Neural Network and Ordinary Least Squares Regression

The rise of robo-advisor wealth management services, which constitute a key element of fintech revolution, unveils the question whether they can dominate human-based advice, namely how to address the client’s behavioral biases in an automated way. One approach to it would be the application of machine learning tools during client profiling. However, trained neural network is often considered as a black box, which may raise concerns from the customers and regulators in terms of model validity, transparency, and related risks. In order to address these issues and shed more light on how neurons work, especially to figure out how they perform computation at intermediate layers, this paper visualizes and estimates the neurons’ sensitivity to different input parameters. Before it, the comprehensive review of the most popular optimization algorithms is presented and based on them respective data set is generated to train convolutional neural network. It was found that selected hidden units to some extent are not only specializing in the reaction to such features as, for example, risk, return or risk-aversion level but also they are learning more complex concepts like Sharpe ratio. These findings should help to understand robo-advisor mechanics deeper, which finally will provide more room to improve and significantly innovate the automated wealth management process and make it more transparent.

The authors study the Fama and French three-factor (FF-3F) model in relation to a developing market. To this end, they consider Chinese stock markets over the period 1995–2008, which is to say, over a period when these markets are recognized as “developing” markets influenced by speculative activity. The authors find that the model appears to be working as a form of “principal component analysis for the determinants of stock price formation with book-to-market (B/M) as the “variable of choice” on account of that it captures the earnings-to-price (E/P), cash-flow-to-price (C/P) and sales-to-price (S/P) variables while remaining largely uncorrelated with firm size (whereas E/P, C/P and S/P are themselves positively correlated with firm size). The variables, however, are unrelated to risk as represented by market exposure, volatility, or leverage.

This investigation studies the impact of mutual fund herding on the returns achieved by contrarian strategy from 1990 to 2015 in the Chinese stock market. The relationship between the profit gained by the contrarian strategy and the macroeconomic environment is also examined. First, the returns of the contrarian strategy in China’s stock market are found to be significant. Second, most loser stocks with a high degree of mutual fund herding outperform loser stocks with a low degree of mutual fund herding, revealing that the profitability of an investment portfolio depends on the degree of mutual fund herding. Third, investors should buy loser stocks with a high degree of herding and sell winner stocks with a low degree of herding during a two-year formation period, over which zero-cost contrarian strategies yield the significantly highest return. Finally, the payoff of contrarian strategies is positively related to the herding effect and negatively related to macroeconomic variables.