Do Survey Expectations of Stock Returns Reflect Risk-Adjustments?

Abstract

Motivated by the observation that survey expectations of stock returns are inconsistent with rational return expectations under real-world probabilities, we investigate whether alternative expectations hypotheses entertained in the literature on asset pricing are consistent with the survey evidence. We empirically test (1) the notion that survey forecasts constitute rational but risk-neutral forecasts of future returns, and (2) the notion that survey forecasts are ambiguity averse/robust forecasts of future returns. We find that these alternative hypotheses are also strongly rejected by the data, albeit for different reasons. Hypothesis ( 1) is rejected because survey return forecasts are not in line with risk-free interest rates and because survey expected excess returns are predictable. Hypothesis ( 2) is rejected because agents are not always pessimistic about future returns, instead often displaying overly optimistic return expectations. We speculate as to what kind of expectations theories might be consistent with the available survey evidence.

The term structure of expectations

“…The main takeaway is that our survey-based measure of short rate expectations has on average performed substantially better in predicting short rates than the forward curve, and the performance gap widens with the forecast horizon. This finding is consistent with the notion that forecasters do not simply report risk-neutral expectations extracted from the forward curve when being surveyed about short rates, a conclusion also shared by Adam et al (2021). At the same time both measures have come closer in the past 20 or so years, reflecting the overall decline of term premiums.…”

Five Facts about Beliefs and Portfolios

“…We also contribute to a macro-finance literature that has debated the tradeoffs between behavioral and rational modeling approaches, with survey evidence providing an important input (see Cochrane, 2011Cochrane, , 2017Greenwood and Shleifer, 2014;Adam, Matveev and Nagel, 2018). Among the many proposed equilibrium models, the most relevant for our work are those that directly incorporate survey evidence (e.g., Barberis et al, 2015;Adam, Marcet and Beutel, 2017;Bhandari, Borovička and Ho, 2016) and those that feature heterogeneous belief (e.g., Scheinkman and Xiong, 2003;Geanakoplos, 2010;Caballero and Simsek, 2017;Martin and Papadimitriou, 2019).…”

Necessary Evidence for a Risk Factor's Relevance

“…As pointed out by Adam et al (2020), survey evidence "allows researchers to consider alternatives to the rational-expectations assumption in an empirically disciplined way." Amromin and Sharpe (2013) explores how investors use current economic conditions to inform their stock-market expectations.…”