US Stock return predictability with high dimensional models

This article investigates whether the HML, SMB along with the long-term reversal and the momentum factors exhibit both in-sample and out-of-sample forecasting ability for the US stock returns. Our findings suggest that these factors contain significantly more information for future stock market returns than the typically employed financial variables. We also go one step further and test whether these variables can proxy for the aforementioned factors and find that the default spread and to a lesser extent the term spread contain important information for the evolution of the factors examined. Finally, we show that appropriate decompositions of the factors in their size and value components can enhance predictability.

Health care costs represent a large and growing component of business and consumer expenditures in the US. Medical inflation represents these costs, and it differs from aggregate inflation and other market factors with respect to its rate of growth, statistical properties and the extent to which it can be hedged by households and firms. Using multiple model specifications for the 25-year period from 1985 to 2009, we find medical inflation is robustly priced in the cross-section of US stock returns. It commands a risk premium of between 31 and 51 basis points per annum per unit change in beta. Medical inflation is also unique in that it represents the only inflationary component that robustly explains the cross-section of stock returns in this manner and is not subsumed by other common factors in the literature. These results quantify the health care industry’s unique and significant role in the US economy and stock market, further rationalizing the substantial attention this industry receives.

ABSTRACTIn this article, we document the asymmetric role that the US stock market plays in the international predictability of excess stock returns during recession and expansion periods. Most of the positive evidence accrues during the periods of recessions in the United States. During the expansions, there is only a limited evidence supporting the importance of lagged US returns in predictability of stock returns in 10 industrialized countries.

This paper investigates which US stock price index is strongly influenced by the Japanese stock markets. Our empirical tests as to the time-varying correlations derived from a multivariate generalized autoregressive conditional heteroskedasticity (MGARCH) model reveal the following evidence. First, (1) we clarify that all US stock returns of the Dow Jones, S&P 500, and NASDAQ Composite indices are statistically significantly positively influenced by the returns of TOPIX and the Nikkei 225. Moreover, we also find that (2) the changes of the NASDAQ-S&P 500 (NS) ratio are statistically significantly positively influenced by the returns of TOPIX and the Nikkei 225 while the changes of the Dow Jones-S&P 500 (DS) ratio and the Dow Jones-NASDAQ (DN) ratio are statistically significantly negatively influenced by the returns of TOPIX and the Nikkei 225. These relations between the returns of TOPIX and the Nikkei 225 and the US stock price index ratios are clearly shown in all three sample periods employed in this paper. Hence our results suggest that, in our analyzing periods, (3) the NASDAQ index is most strongly positively connected with the Japanese stock markets.

We consider the problem of neural network training in a time‐varying context. Machine learning algorithms have excelled in problems that do not change over time. However, problems encountered in financial markets are often time varying. We propose the online early stopping algorithm and show that a neural network trained using this algorithm can track a function changing with unknown dynamics. We compare the proposed algorithm to current approaches on predicting monthly US stock returns and show its superiority. We also show that prominent factors (such as the size and momentum effects) and industry indicators exhibit time‐varying predictive power on stock returns. We find that during market distress, industry indicators experience an increase in importance at the expense of firm level features. This indicates that industries play a role in explaining stock returns during periods of heightened risk.

A recent study by Rapach, Strauss, and Zhou (Journal of Finance, 2013, 68(4), 1633–1662) shows that US stock returns can provide predictive content for international stock returns. We extend their work from a volatility perspective. We propose a model, namely a heterogeneous volatility spillover–generalized autoregressive conditional heteroskedasticity model, to investigate volatility spillover. The model specification is parsimonious and can be used to analyze the time variation property of the spillover effect. Our in‐sample evidence shows the existence of strong volatility spillover from the US to five major stock markets and indicates that the spillover was stronger during business cycle recessions in the USA. Out‐of‐sample results show that accounting for spillover information from the USA can significantly improve the forecasting accuracy of international stock price volatility.

Investor Attention and the Response of US Stock Sectors to the COVID-19 Crisis

Predictable variation and profitable trading of US equities: a trading simulation using neural networks

In this study, we aim to test the pricing power of market liquidity in the cross-section of US stock returns. We examine three liquidity measures: Pástor and Stambaugh (2003)’s liquidity factor, Bali et al. (2014)’s liquidity shocks, and Dreshsler, Savov, and Schanbl (2017)’s money market liquidity premium. With a large set of test assets and the time-series regression approach of Fama and French (2015), we find that aggregate liquidity is not priced in the cross-sections of stock returns. That is, adding the liquidity factor to common asset-pricing models does not improve the performance of models significantly. Therefore, our results call for more research on the impact of aggregate liquidity on the stock market.

Fama and French (1992) show that size and book-to-price dominate CAPM beta and other variables such as the price-earnings ratio and dividend yield in explaining the cross-section of US stock returns. Comparable evidence for the UK points to a book-to-price effect, but not a size effect (Chan and Chui, 1996; Strong and Xu, 1997). In this paper, our first contribution is to show that a measure of research and development (RD) helps explain cross-sectional variation in UK stock returns. Our cross-sectional results on the association between stock returns and RD are consistent with recent US evidence reported by Lev and Sougiannis (1996, 1999) and Chan, Lakonishok and Sougiannis (2001).Fama and French (1993, 1995, 1996) also show that a three-factor model captures a high proportion of the time series variation in portfolio returns, again for the US. Our second contribution is to show, for the UK, that a modification to the three-factor model to take account of RD activity can significantly enhance the explanatory power of the three-factor model. We show that, as a practical matter, estimated risk premia based on the modified three-factor model can differ considerably from risk premia estimated using the CAPM or the three-factor model. In particular, risk premia for industries in which few firms undertake RD activities tend to be over-estimated.

ABSTRACTIn this paper, we investigate the impact of oil prices on both aggregate and industry US real stock returns over the period 1973–2017. The empirical analysis contributes to the related literature introducing a state-dependent oil price (high and low) and the local projections approach. Our main finding is that, depending on the nature of the shock and industry, the negative effects of oil price shocks become exacerbated -and the positive effects get moderated- if oil prices are already high.

Fuzzy Inductive Reasoning and Nonlinear Dependence in Security Returns: Results from an Artificial Stock Market Environment

Recent financial research has provided evidence on the predictability of asset returns. In this paper we consider the results contained in Pesaran and Timmerman (1995), which provided evidence on predictability of excess returns in the US stock market over the sample 1959–1992. We show that the extension of the sample to the nineties weakens considerably the statistical and economic significance of the predictability of stock returns based on earlier data. We propose an extension of their framework, based on the explicit consideration of model uncertainty under rich parameterizations for the predictive models. We propose a novel methodology to deal with model uncertainty based on ‘thick’ modelling, i.e. on considering a multiplicity of predictive models rather than a single predictive model. We show that portfolio allocations based on a thick modelling strategy systematically outperform thin modelling. Copyright © 2005 John Wiley & Sons, Ltd.

The Comovement of US and UK Stock Markets

This dissertation provides empirical evidences on media-based investor emotions in predicting stock return, conditional volatility, and stock and bond return comovements. We first studied the interaction between US media content and the US stock market returns and volatility. We utilize propriety investor sentiment measures developed by Thompson Reuters MarketPsych. We select four measures of investor sentiment that reflect both pessimism and optimism of small investors. Our objective is two-fold. First, we examine the ability of these sentiment measures to predict market returns. For this purpose, we use dynamic Vector Auto-Regressive models. Second, we are interested in exploring the effects of these sentiment measures on the market returns and volatility. For this purpose, we utilize a Threshold-GARCH model. Next, we investigated the effect of investor emotions (fear, gloom, joy and optimism) in financial futures markets by using Thompson Reuters MarketPsych Indices. The purpose of this study is three fold. First, we investigate the extent of usefulness of informational content of our sentiment measures in predicting stock futures and treasures futures returns using daily data for different measures of emotional sentiments. Second, we investigate whether emotion sentiments affect financial futures returns and volatilities. Third, we explore the role of emotion sentiment factors in volatility transmission in financial futures markets. To the best of our knowledge, this is the first study that extensively explores the role of investors’ sentiment in the most liquid contracts (S&P 500 futures and 10-year Treasury notes) in futures markets.