Extreme dependence in investor attention and stock returns – consequences for forecasting stock returns and measuring systemic risk

We characterize co-movements in investor attention by modeling multivariate internet search volume data. Using a variety of copula models that can capture both asymmetric and skewed dependence, we find empirical evidence of strong non-linear and asymmetric dependence in the attention investors give to companies. Modeling three years of daily stock returns and search volumes from Google Trends for 29 bank names, we find a striking similarity between the dependence structure inherent in stock returns and the dependence in the corresponding time series of search queries. We then document the existence of significant asymmetric and skewed tail dependence in the joint distribution of stock returns and investor attention. Finally, stock returns and internet search volumes appear to evolve concurrently in real time with neither one leading the other. Our findings have important implications, e.g. for the analysis of banks' interconnectedness based on equity data and the pricing of investor attention in the cross-section of stock returns.     

Diversification and generalized tracking errors for correlated non‐normal returns

Abstract

The probability distribution for the relative return of a portfolio constructed from a subset n of the assets from a benchmark, consisting of N assets whose returns are multivariate normal, is completely characterized by its tracking error. However, if the benchmark asset returns are not multivariate normal then higher moments of the probability distribution for the portfolio's relative return are not related to its tracking error. We discuss the convergence of generalized tracking error measures as the size of the subset of benchmark assets increases. Assuming that the joint probability distribution for the returns of the assets is symmetric under their permutations we show that increasing n makes these generalized tracking errors small (even though n « N). For n » 1 the probability distribution for the portfolio's relative return is approximately symmetric and strongly peaked about the origin. The results of this paper generalize the conclusions of Dynkin et al (Dynkin L, Hyman J and Konstantinovsky V 2002 Sufficient Diversification in Credit Portfolios (Lehman Brothers Fixed Income Research)) to more general underlying asset distributions.     

An encyclopedia for stock markets? Wikipedia searches and stock returns

We present empirical evidence that collective investor behavior can be inferred from large-scale Wikipedia search data for individual-level stocks. Drawing upon Shannon transfer entropy, a model-free measure that considers any kind of statistical dependence between two time series, we quantify the statistical information flow between daily company-specific Wikipedia searches and stock returns for a sample of 447 stocks from 2008 to 2017. The resulting stock-wise measures on information transmission are then used as a signal within a hypothetical trading strategy. The results evidence the predictive power of Wikipedia searches and are in line with the previously documented notion of buying pressure revealed by online investor attention and the trading patterns of retail investors.     

Too close for comfort? Geographic propinquity to political power and stock returns

We show that firm headquarters’ geographic proximity to political power centers (state capitals) is associated with higher abnormal returns. Consistent with the notion that this effect is rooted in social network links, we find it is more pronounced in communities with high levels of sociability and political values’ homophily, and that it dissipates when firms move their headquarters to another state. Finally, in line with the view that investors perceive such networks to be associated with political risk, we find that this effect is particularly strong when there are substantial levels of corruption, dependency on government spending, and politicians’ turnover.     

Predictability and the earnings–returns relation

This paper studies the effects of predictability on the earnings–returns relation for individual firms and for the aggregate. We demonstrate that prices better anticipate earnings growth at the aggregate level than at the firm level, which implies that random-walk models are inappropriate for gauging aggregate earnings expectations. Moreover, we show that the contemporaneous correlation of earnings growth and stock returns decreases with the ability to predict future earnings. Our results may therefore help explain the apparently conflicting recent evidence that the earnings–returns relation is negative at the aggregate level but positive at the firm level.     

The output gap and stock returns: Do cyclical fluctuations predict portfolio returns?

This study examines whether the output gap leads portfolio stock returns. The paper conducts in-sample and out-of-sample forecasting of US stock portfolios formed on the basis of size and value. First, the paper finds cross-sectional portfolios are predictable in-sample by the output gap. Out-of-sample evidence is weaker but still generally supports the finding that the historical average benchmark can be beaten. Secondly and most importantly, we find mixed evidence that the Fama–French factor mimicking portfolios can be forecasted by the output gap. In particular, there is some out-of-sample predictability of the size effect (SMB) suggesting this lags the output gap. However, the output gap, a key business cycle indicator, cannot predict the value effect (HML) either in-sample or out-of-sample. Our results add to the prior literature which finds that the factor mimicking returns are related contemporaneously (Petkova and Zhang, 2005) or lead (Liew and Vassalou, 2000) economic indicators.     

Why does shareholder protection matter for abnormal returns after reported insider purchases and sales?

We use a unique data set of more than 240,000 reported insider transactions across 15 European countries and the USA to analyze the link between country-level shareholder protection and abnormal returns following insider trades. We show that abnormal returns after insider transactions are positively correlated with country-level shareholder protection against expropriation by corporate insiders, which supports the information-content hypothesis. Market reaction to insider purchases increases with shareholder protection because shareholder protection enhances the transparency and trustworthiness of insiders’ actions, and limits possibilities for direct profit diversion, so that more information is eventually reflected in stock prices. For insider sales, shareholder protection decreases their negative information content. We conjecture that this is due to the effect of greater transparency and trustworthiness strengthening the diversification and liquidity reasons for selling in better shareholder protection countries. We find limited support for the rent-extraction hypothesis that conjectures that shareholder protection is associated with insider trading dollar profits.     

Financial constraints and stock returns — Evidence from Australia

Using multiple discriminant analysis, we construct an index that measures firms' external financial constraints in an Australian setting. We form portfolios of firms based on our financial constraints index and find that financially constrained firms earn lower return than their unconstrained counterparts. Moreover, stock returns of financially constrained firms are found to move together, indicating the potential existence of a financial constraints factor. Neither the variation nor the mean return of the constraints factor are well explained by existing asset pricing models, suggesting an independent role for our financial constraints factor in affecting stock returns.     

Discussion of “Investor recognition and stock returns”

Lehavy and Sloan (2008, Review of Accounting Studies) note that prior studies find that earnings and cash flows explain only a small portion of the cross-sectional variation in stock return. This motivates them to investigate empirically the ability of a behavioral model of capital market equilibrium proposed by Merton (1987, Journal of Finance, 42, 483–510) to explain the remaining variation in stock returns. Their primary findings show that security value is, as predicted, increasing in investor recognition of the security and that investor recognition is incremental to and more important than cash flows in explaining the cross-sectional variation of stock returns. While the research question is intriguing and well motivated, a number of methodological limitations may limit the reliability of the findings/interpretations. In this paper, I first evaluate the motivation and potential contribution of the Lehavy and Sloan (2008) study. I then outline methodological limitations underlying the study and offer ways of overcoming them. In the final section, I state my conclusions.     

How good can heuristic-based forecasts be? A comparative performance of econometric and heuristic models for UK and US asset returns

This paper systematically investigates the sources of differential out-of-sample predictive accuracy of heuristic frameworks based on internet search frequencies and a large set of econometric models. The volume of internet searches helps gauge the degree of investors’ time-varying interest in specific assets. We use a wide range of state-of-the-art models, both of linear and nonlinear type (regime-switching predictive regressions, threshold autoregressive, smooth transition autoregressive), extended to capture conditional heteroskedasticity through GARCH models. The predictor variables investigated are those typical of the literature featuring a range of macroeconomic and market leading indicators. Our out-of-sample forecasting exercises are conducted with reference to US, UK, French and German data, both stocks and bonds, and for 1- and 12-months-ahead horizons. We employ several forecast performance metrics and predictive accuracy tests. Internet-search-based models are found to perform better than the average of all of the alternative models. For several country-asset-horizon combinations, particularly for UK bond returns, our heuristic models compare favourably with sophisticated econometric methods. The heuristic models are also shown to perform well in forecasting realized volatility. The baseline results are supported by several extensions and robustness checks, such as using alternative search keywords, controlling for Fama–French and Cochrane–Piazzesi factors, and implementing heuristic-based trading strategies.     

Liquidity and expected returns—Evidence from 1926–2008

This paper re-examines the liquidity effect on stock expected returns in the NYSE over the period 1926–2008, the pre-1963 period, for which there is a lack of research, and the post-1963 period. The results from the entire sample of 1926–2008 show that expected returns increase with the stock level illiquidity. However, illiquidity level has explanatory power in the cross-sectional variation of stock expected returns only over the post-1963 period, and is, both economically and statistically, insignificant for the whole sample and the pre-1963 period. These findings are robust after taking into account various characteristics such as size and risk controls. On the other hand, evidence from the entire sample and the pre-1963 sample suggests that the systematic liquidity risk plays a significant role in the cross-sectional variation of stock expected returns. The different result for the pre- and post-1963 is explained by the portfolio shifts occurred during the economic downturns.     

‘Optimal’ probabilistic and directional predictions of financial returns

This paper examines the probability of returns exceeding a threshold, extending earlier work of Christoffersen and Diebold (2006) on volatility dynamics and sign predictability. We find that the choice of the threshold matters and that a zero threshold (leading to sign predictions) does not lead to the largest probability response to changes in volatility dynamics. Under certain conditions there is a threshold that has maximum responsiveness to changes in volatility dynamics that leads to ‘optimal’ probabilistic predictions. We connect the evolution of volatility to probabilistic predictions and show that the volatility ratio is the crucial variable in this context. The overall results strengthen the arguments in favor of accurate volatility measurement and prediction, as volatility dynamics are integrated into the ‘optimal’ threshold. We illustrate our findings using daily and monthly data for the S&P500 index.     

Does the connectedness among fossil energy returns matter for renewable energy stock returns? Fresh insights from the Cross-Quantilogram analysis

The response of renewable energy stock returns to the dynamics of fossil energy markets is a vital concern of low-carbon transitions. There is still sparse literature documenting the directional dependence of renewable energy stock returns on the connectedness among fossil energy returns, even though previous studies have examined the relationship among renewable energy stocks and fossil energy markets. Additionally, the conclusions of prior studies are quite far from reaching a consensus regarding the relationship between the renewable energy stock and the fossil energy markets. To this end, by using the TVP-VAR based connectedness approach and Cross-Quantilogram techniques, this study does the first attempt to unpack the complicated and controversial directional dependence of renewable energy stock returns on the returns and connectedness of fossil energy markets, considering various market conditions and time horizons. The empirical analysis demonstrates that, first, the directional dependence of renewable energy stock returns on fossil energy returns is pronounced during extreme market conditions, whereas they appear to be decoupled from fossil energy returns during normal market conditions. Second, the total connectedness between fossil energy returns transmits a substantial shock to renewable energy stock returns during most market conditions, which is in stark contrast to the information transmission directly originating from fossil energy markets. The performance of renewable energy stock markets improves with stronger fossil energy return connectedness, whereas weaker fossil energy return connectedness hinders it. Additionally, further study reveals that the directional dependence of renewable energy stock returns on the net connectedness of the crude oil market is dominated by negative dependence when the net connectedness of the crude oil market is low, whereas it displays positive dependence when the net connectedness of the crude oil market is high. This directional dependence pattern on the net connectedness of the crude oil market is opposite to that exhibited in the net connectedness of the coal and natural gas markets. Third, in general, the directional dependence of renewable energy stock returns on fossil energy returns is more pronounced in the short term but diminishes over the medium and long terms. Conversely, the directional dependence of renewable energy stock returns on fossil energy return connectedness persists over the medium and long terms. Final, with the outbreak of the Global Financial Crisis during 2007–2008, we notice an abrupt jump in the directional dependence of renewable energy stock returns on fossil energy returns and their connectedness, particularly during extreme market conditions. Our findings provide noteworthy implications for energy transformation, energy security, and climate mitigation.     

A high-low-based omnibus test for symmetry, the Lévy property, and other hypotheses on intraday returns

We consider different models for intraday log-returns: Lévy models, symmetric models, and Lévy processes subjected to independent continuous time-changes. For these models, we show bivariate interchangeability of intraday up- and downside volatility ratios which are built using daily high-low prices. Using conditional inference permutation tests on bivariate interchangeability, we develop an omnibus test for the above-mentioned models. Empirically, we find strong evidence against intraday returns belonging to these model classes, as we reject bivariate interchangeability of the volatility ratios for half of the components of the DJIA, two thirds of the S&P 500 shares and almost all stocks of the German DAX.     

Can NN-algorithms and macroeconomic data improve OLS industry returns forecasts?∗

Using an extensive range of macroeconomic indicators and a number of two-stage models mixing OLS and a non-parametric approach known as the nearest neighbour algorithm, the authors analyse the potential for improving forecasts of US industry returns over those built by OLS on industry-specific variables only. Basic performance is measured by the average cross-sectional correlation over time between the 55 forecasted returns and the realized returns across industries. Since investors and asset managers typically want a steady performance over time, the volatility of this cross-sectional correlation is further taken into account in an adaptation of the Sharpe Ratio. Strong evidence is found in favour of certain macroeconomic factors as dominant industry return predictors, and some two-stage models based either purely on OLS or a mix between OLS and the non-linear model can lift both cross-sectional forecasting correlation and Sharpe Ratio. However, viewed overall in relation to the benchmark OLS model, performance is not consistently improved by any particular model.     

The cross-section of equity returns and assets’ fundamental cash-flow risk

The decomposition of consumption beta into a component driven by assets’ cash-flow news and one related to assets’ discount-rate news reveals that macroeconomic risks embodied in cash flows largely account for the cross-sectional dynamics of average stock returns. Empirically, we find that differences in expected excess returns between low book-to-market and high book-to-market portfolios are associated with differences in their cash-flow betas, and thus reflect macroeconomic, especially consumption-related risks. This result holds true for a broad set of consumption-based asset pricing models. In addition, the results indicate that the risk premium on equity markets is primarily driven by the exposure of assets’ cash-flow components to the cyclical variability of durable consumption goods.     

Pricing volatility of stock returns with volatile and persistent components

This paper introduces a two-component volatility model based on first moments of both components to describe the dynamics of speculative return volatility. The two components capture the volatile and the persistent part of volatility, respectively. The model is applied to 10 Asia-Pacific stock markets. Their in-mean effects on returns are tested. The empirical results show that the persistent component is much more important for the volatility dynamic process than is the volatile component. However, the volatile component is found to be a significant pricing factor of asset returns for most markets. A positive or risk-premium effect exists between the return and the volatile component, yet the persistent component is not significantly priced for the return dynamic process.

Do industries matter in explaining stock returns and asset-pricing anomalies?

Industry returns cannot be explained fully by well-known asset pricing models. This study reveals that common factors extracted from industry returns carry significant risk premiums that go beyond the explanatory power of size, book-to-market (BM) ratios, and momentum. In particular, this study shows that (1) the small-firm effect is significant only for firms whose market capitalization is below their industry average; (2) the BM effect is an intra-industry phenomenon; (3) a one-year momentum effect is significant only for firms whose BM ratio is smaller than the industry average and limited to non-January months; and (4) there is seasonality in all effects that cannot be explained by risk-based asset-pricing models. Neither rational nor behavioral theories alone can explain industry returns, and it is perhaps too hasty to attribute asset pricing anomalies to a single driving force.     

Factor state–space models for high-dimensional realized covariance matrices of asset returns

We propose a dynamic factor state–space model for high-dimensional covariance matrices of asset returns. It makes use of observed risk factors and assumes that the latent integrated joint covariance matrix of the assets and the factors is observed through their realized covariance matrix with a Wishart measurement density. For the latent integrated covariance matrix of the assets we impose a strict factor structure allowing for dynamic variation in the covariance matrices of the factors and the residual components as well as in the factor loadings. This factor structure translates into a factorization of the Wishart measurement density which facilitates statistical inference based on simple Bayesian MCMC procedures making the approach scalable w.r.t. the number of assets. An empirical application to realized covariance matrices for 60 NYSE traded stocks using the Fama–French factors and sector-specific factors represented by Exchange Traded Funds (ETFs) shows that the model performs very well in- and out of sample.     

China’s anti-corruption campaign and stock returns of luxury goods firms

Financial Markets and Portfolio Management - This paper uses the unexpected launch of the Chinese anti-corruption campaign in 2012 to study the stock market effects of political risk. Political...