Research

We investigate the performances of two methods of complete subset averaging—complete subset linear averaging (CSLA) and complete subset quantile averaging (CSQA)—on the problem of corporate bond return prediction. We find that the two methods are overwhelmingly better than univariate linear regression and simple forecast combination. Meanwhile, CSQA is better than CSLA in most cases. For practical implementation, we also provide discussions on the selection of the hyperparameter k when applying these complete subset averaging methods.

We employ an ensemble learning approach, “stacking”, to refine and combine a variety of linear and nonlinear individual stock return prediction models. In an application of forecasting U.S. market excess return, stacking with a simple structure can outperform the traditional historical mean benchmark, Mallows model averaging, simple combination forecast, complete subset regression, combination elastic net forecast, and several other models in terms of both in- and out-of-sample performance measures on a consistent basis. More importantly, we find that the out-of-sample gains of stacking are especially evident during extreme downside market movements. Overall, stacking can generate substantive improvements in market excess return predictability.

This paper investigates how the COVID-19 pandemic affects the connectedness network of stock market volatility in 19 economies around the world. Our method builds on the Diebold-Yilmaz volatility network model to construct the volatility spillover index, and uses lag sparse group LASSO to accommodate the high-dimensional system. We find that the outbreak of the COVID-19 pandemic strengthens the overall volatility connectedness, and the global connectedness level remains high throughout 2020. In particular, connections across different continents have become stronger during this period. However, China is shown to be disconnected from the global volatility connectedness network until late November 2020. We find evidence that China is not the main source of volatility spillover during the COVID-19 pandemic.

A major research task in asset pricing is predicting the value of future stock returns. We examine a seemingly simpler and yet less explored task—predicting the direction of future stock returns. Theoretically, a simple model in Christoffersen and Diebold (2006) suggests that direction prediction is closely related to the widely recognized volatility clustering effect. Empirically, we compare direction and value prediction of the S&P500 excess returns via a variety of models, including linear, regularized linear, machine learning, and combination methods. The general findings indicate that for a typical investor, direction prediction generates greater economic gains than value prediction does because of the information embedded in volatility and higher moments of returns. From the models’ perspectives, while machine learning and combination methods can enhance both direction and value prediction, the enhancement is more significant in the realm of direction prediction, and the predictive power stems from the volatility mechanism. Additional analyses suggest that value and direction prediction share similar stock market risk factors.

We propose a simple, linear-regression-based method for prediction of the time series of stock returns. The method achieves out-of-sample performances comparable to machine learning methods while having ignorable computational costs. The key component of the method is to integrate a straightforward cross-market factor screening into the iterated combination method proposed by Lin, Wu, and Zhou (2018). Our empirical results on the U.S. stock market show that the method outperforms many state-of-the-art machine learning methods in certain periods. The method also exhibits greater utility gain and investment profits in most periods after considering transaction costs.

This paper provides evidence that the U.S. Securities Acts of 1933 and 1934 improved market stability by imposing mandatory information disclosure. Despite being the world’s first national level financial regulation laws, the seminal Acts’ effects have remained controversial. We argue that improvement of accounting report and mandatory disclosure of insiders’ holdings help reduce the volatility of security returns. We provide evidence that reductions in NYSE stocks’ idiosyncratic volatility and market volatility are causally related to enforcement of the Acts. Overall, our evidence shows one positive effect of the initial introduction of mandatory disclosure by the 1933/34 Acts.

Using predictors from corporate bond, Treasury, and stock markets, we find that corporate bond returns are predictable based on an ensemble learning approach, “Stacking”. Combining various linear and nonlinear models together, we show that Stacking generates significant predictive power in terms of both statistical measure and economic gain across bond ratings and maturities. Long term government bond return, leverage ratio, default return spread, and stock liquidity are the most critical factors in corporate bond return prediction. We also find that a simpler meta model performs best.