Do Happy People Make Optimistic Investors?
Bar-Ilan University – Graduate School of Business Administration
Hebrew University of Jerusalem – Jerusalem School of Business Administration; Fordham University
Monash University
Monash University – Department of Finance
February 13, 2015
Journal of Financial and Quantitative Analysis (JFQA), Volume 50, (2015), 145-168.
Abstract:
Do happy people predict future risk and return differently from unhappy people, or do individuals rely only on economic facts? We survey investors on their subjective sentiment-creating factors, return and risk expectations, and investment plans. We find that non-economic factors systematically affect return and risk expectations, where the return effect is more profound. Investment plans are also affected by non-economic factors. Sports results and general feelings significantly affect predictions. Sufferers from seasonal affective disorder have lower return expectations in the autumn than in other seasons, supporting the Winter Blues hypothesis.
Do Happy People Make Optimistic Investors? – Introduction
Several market-based empirical studies show that non-economic factors (for example, aggregate investor sentiment) are significantly correlated with stock prices. Investor sentiment is defined in the literature in various ways. The most widely accepted definition is probably that suggested by Baker and Wurgler (2007). They define investor sentiment as “investors’ belief about future cash flows and risk not justified by the facts at hand” (p. 129). We employ this definition in this study, and we analyze the direct link between sentiment and risk and return expectations on the individual level. While we do not employ an aggregate market sentiment empirical study, we use previously published sentiment results as a springboard for establishing some of the hypotheses tested in this paper on the individual investor level.
First, we explore the association between sentiment-creating factors reported by each individual and a “general feeling” (mood) variable. Second, we study the relation between sentiment-creating factors and the individual’s subjective market judgment regarding expected return and volatility (risk). Finally, we explore whether and how sentiment affects actual investment plans. To the best of our knowledge, this is the first study to examine, on the individual level, the direct relation between sentiment-creating factors, subjective estimates of future return and risk, and investment plans, where the analysis uses data from a sample of subjects who actually trade in the stock market. Working with data on individual investors who completed a questionnaire in different months, different days, and different hours of the day, the likelihood that some other common factors induce the observed correlation is relatively small; hence, we show compelling evidence of a causal relation between sentiment and individuals’ return expectations.
The cross-sectional analysis relies on the Longitudinal Internet Studies for the Social Sciences (LISS) panel of CentER data at Tilburg University.1 This panel consists of a sample of approximately 5,000 households representative of the population of the Netherlands. After screening those individuals who hold stocks in their investment portfolios, we submitted questionnaires to approximately 900 individuals in three waves during a time span of one year. Each individual reported next-month and next-year subjective return and risk expectations regarding both the Dutch and U.S. stock markets. Each individual also reported on several sentiment-creating factors, some of which are found in previous studies to be correlated with stock prices at the aggregate level. These factors include general feeling (which has not been tested before and allows us to test for the individual mood), recent results for the individual’s favorite sports team, the individual’s perception of contemporaneous weather, and whether the individual is “a spring person” in general and suffers from seasonal affective disorder (SAD) in particular.
Since the questionnaire was filled out three times during the year, we can test what effect the season, particularly the SAD season, has on sentiment. As the subjects filled out the questionnaire on different days, the reported day can be used to test for the day-of-the-week sentiment effect. Unlike market-based empirical studies, we also have data corresponding to weekend non-trading days, which are generally characterized by above-average positive mood. Finally, each individual reported on plans to buy or sell stocks in the coming month. Thus, each individual simultaneously reports on subjective stock market return and risk expectations, subjective perception of contemporaneous sentiment-creating factors, and investment plans.
See full PDF below.
