Abstract: Existing studies on asset return predictability focus on aggregate performance. We examine the oft-overlooked grouped heterogeneity in return predictability across different assets and macroeconomic regimes. A novel tree-based asset clustering methodology is introduced to partition the panel of asset-return observations according to return predictability, using high-dimensional asset characteristics and aggregate time-series predictors. When implemented on U.S. equities over the past five decades, we find that some characteristics-managed (dollar trading volumes, unexpected earnings, earnings-to-price, and cashflow-to-price) and/or macro-based (dividend yield and default yield) clusters are more predictable, resulting in a heterogeneous predictive model with outperformance. Finally, less predictable clusters generally exhibit lower risk-adjusted investment performance, revealing an important empirical link between return predictability and trading profitability.
Discussant: David Rapach, Federal Reserve Bank of Atlanta
Abstract: We study probability forecasts in the context of cross-sectional asset pricing with a large number of firm characteristics. We evaluate the performance of eight distinct models: Ordinary Least Squares (OLS), Logistic Regression (Logit), Partial Least Squares (PLS), and a set of Neural Networks (NN) with one to five hidden layers. Empirically, we find that a simple probability forecast model such as Logit can surprisingly perform as well as a sophisticated NN probability forecast model, all of which deliver long-short portfolios whose Sharpe ratios are comparable to those of the widely used return forecasts. Moreover, we show that combining probability forecasts with return forecasts yields superior portfolio performance (yielding a Sharpe ratio of 1.76) versus using each type of forecast individually, suggesting that probability forecasts provide valuable information beyond return forecasts for our understanding of the cross-section of stock returns. We also demonstrate probability forecasts provide incremental value for cash flow and fundamental prediction. The profitability factor constructed using probability forecasts generates larger returns and higher Sharpe ratios compared to the traditional expected profitability factor. Overall, our findings highlight the additional insights and value of modeling the probability instead of just the expected returns alone. By shifting the focus from traditional expected return forecasts only to consider also probability-based predictions, our study suggests a new paradigm for asset pricing and portfolio management.
Discussant: Allaudeen Hameed, National University of Singapore
Abstract: This paper demonstrates that economic narratives significantly price the cross-section of stocks. Using a vast dataset of more than 150k digital media sources since 2013, roughly 350 narratives are quantified, and corresponding narrative-mimicking, long-short portfolios are constructed using stock return narrative betas. Narrative-mimicking portfolios of recently trending narratives outperform those of descending attention by about 7% annually, controlling for standard risk factors. The cross-sectional narrative-beta-pricing is independent of past return and is neither significantly impacted by narrative coverage at the stock level nor earnings announcements. The results suggest that while investors respond to short-run narrative shocks as measured by narrative betas, they under-react to long-run narrative trends, manifesting narrative momentum returns.
Discussant: Kuntara Pukthuanthong, University of Missouri
Abstract: We find that managers strategically shift targets in their communications with investors and markets. Using the complete history of the earnings conference call transcripts by U.S. corporations from 2006 – 2020, we employ natural language processing techniques to analyze conference calls and find that managers choose and re-choose targets to ensure they clear their endogenously chosen hurdle. For instance, if they have seen same-store sales growth for 16 consecutive quarters, they will mention this intensively. However, when they encounter a quarter without same-store sales growth, they shift the conversation to another metric, such as cost savings or strategic R&D. When managers change the target, this predicts significant negative returns and realizations for the firm in question. In particular, in the quarter following a moved target, firms underperform by up to 99 basis points per month (t-stat = 4.38) in value-weighted monthly abnormal return (alpha). Moreover, we find that the effects are significantly stronger with more complex targets, non-financial targets, and the most persistent targets.
