Date of Award

8-2021

Degree Name

Doctor of Philosophy

Department

Economics

First Advisor

Dr. Matthew L. Higgins

Second Advisor

Dr. Eskander Alvi

Third Advisor

Dr.Devrim Yaman,

Keywords

Behavioral finance, investor sentiment, bullish sentiment, stock returns, economic surprises, bearish sentiment

Abstract

In the first chapter titled “The Asymmetric Effect of Sentiment on U.S. Equity Returns”, we test the asymmetric impact of investor sentiment, proxied by the Baker-Wurgler (2007) investor sentiment index, on expected stock returns in the U.S. We regress sentiment on market and economy-wide fundamentals, use the residuals as a measure of excess sentiment and estimate long-horizon return regressions using positive and negative components of excess sentiment as predictors. We hypothesize that excessive optimism leads investors to make significant portfolio changes whereas excessive pessimism makes investors more cautious about investing, due to loss aversion. Primary results confirm our hypothesis with a significant positive sentiment coefficient and an insignificant negative sentiment coefficient. Our results hold for an alternative investor sentiment measure, multiple stock market indexes and stock portfolios based on book-to-market ratio, size, operational efficiency, and level of investment. Long-horizon regressions are plagued by two econometric problems: overlapping observations and persistent predictors. We correct for these issues by providing Hodrick (1992) standard errors.

In the second chapter titled “The Asymmetric Effect of Sentiment on Global Equity Returns.”, we test if excess investor sentiment has an asymmetric impact on expected stock returns in thirteen industrialized countries, using long-horizon regression. We regress consumer confidence, a proxy for investor sentiment, on economic indicators and use residuals as a measure of excess sentiment for each country. We regress expected stock returns on positive and negative components of excess sentiment for 6,12,24 and 36 months horizon and correct for econometric problems associated with long-horizon regression by providing Hodrick (1992) standard errors. We find evidence of a statistically significant difference in the effect of bullish and bearish sentiment on stock returns for most countries in the sample. Primary results hold for portfolios based on book-to-market ratio, earnings-price ratio, and dividend yield.

In the third chapter titled “Do Economic Surprises Affect Stock Returns? The Role of Sentiment.”, we test whether the effect of macroeconomic surprises on stock returns is impacted by investor sentiment, proxied by the Federal Reserve Bank of San Francisco’s daily sentiment index. We employ an event study methodology with separate regressions for six real economic indicators: GDP, industrial production, unemployment, retail sales, durable goods, and continuing jobless claims. We regress the daily stock returns for release dates of macroeconomic indicators on macroeconomic surprises. We test if positive and negative sentiment affects the portfolio choices of investors in response to unexpected macroeconomic news. We find consistent results with significant coefficients for pessimistic investors, as they make portfolio changes in response to news, and insignificant coefficients for optimistic investors, as they ignore news about real economic activity. We conclude that loss averse investors take a cautious approach to investing when they are bearish about overall stock market, unlike when they are bullish about stock market. Primary results hold for multiple stock market indexes, different stock portfolios and an alternative categorization of investor sentiment as low, high, and medium sentiment.

Access Setting

Dissertation-Open Access

Share

COinS