Does a lot help a lot? Forecasting stock returns with pooling strategies in a data‐rich environment

Abstract

A variety of recent studies provide a skeptical view on the predictability of stock returns. Empirical evidence shows that most prediction models suffer from a loss of information, model uncertainty, and structural instability by relying on lowdimensional information sets. In this study, we evaluate the predictive ability of various lately refined forecasting strategies, which handle these issues by incorporating information from many potential predictor variables simultaneously. We investigate whether forecasting strategies that (i) combine information and (ii) combine individual forecasts are useful to predict US stock returns, that is, the market excess return, size, value, and the momentum premium. Our results show that methods combining information have remarkable in-sample predictive ability. However, the out-of-sample performance suffers from highly volatile forecast errors. Forecast combinations face a better bias-efficiency trade-off, yielding a consistently superior forecast performance for the market excess return and the size premium even after the 1970s.

Forecasting asset returns with network‐based metrics: A statistical and economic analysis

“…A similar conclusion is also drawn by Pettenuzzo et al (2014). The more recent literature proves OOS predictability by relying on approaches that include (i) (sophisticated) pooling of forecasts or predictors (Baetje, 2017;Gargano et al 2017;Hull & Qiao, 2017;Rapach et al 2010), (ii) harnessing technical analysis in predictive regressions (Neely et al 2014), (iii) machine learning techniques (Cavalcante et al 2016;Rapach et al 2019;Zhong & Enke, 2017), 2 and (iv) regime-based forecasting (Hammerschmid & Lohre, 2018).…”

Analysis of the relevance of sentiment data for the prediction of excess returns in a multiasset framework

“…The world of finance and economics is multifactorial. Including other data such as commodity prices (Black et al, 2014), the geographic location of companies (Boubaker et al, 2019), business cycles (Liu et al, 2021), information on equity block trades (Kurek, 2014(Kurek, , 2016, and other derived economic data (Baetje, 2018;Cenesizoglu et al, 2019;Rahman et al, 2021) would help predictions. Furthermore, it soon became apparent that including extrafinancial data in order to quantify certain intangibles (such as how the public feels about a stock, or how environmentally sound a company's activity is) has its place in statistical models.…”

The global latent factor and international index futures returns predictability

“…This paper investigates whether the global latent factor estimated by the three‐pass regression filter ( TPRF ) provides statistically and economically significant information for the development of a long‐short strategy for international index futures returns. Voluminous literature explores the aggregate stock market return predictability (e.g., Baetje, 2018; Lo & MacKinlay, 1990; Rapach et al, 2013). In particular, Rapach et al (2013) find that lagged US returns can be used to forecast returns in numerous non‐US industrialized nations but that lagged non‐US returns have a limited forecasting power as compared with US returns.…”