Ambiguity, high-frequency data, and asset prices

Abstract

Many theoretical studies have long argued that ambiguity is associated with stock price, but only a limited number of studies can provide empirical evidence about the form and magnitude of the hypothesized relationship. This thesis consists of two separate empirical studies aiming to fill this gap by using a recent methodology proposed by Brenner and Izhakian (2018) as a measure of ambiguity. In the first study, we quantify ambiguity and ambiguity attitudes for Turkish market and test their impact on the stock prices. Our results show that there is an ambiguity premium different than the risk premium in Turkish market. We also find that Turkish investors are ambiguity neutral in lateral markets but tend to avoid ambiguity as the probability of favorable returns moves away from the historical average. In the second study, we examine the effect of ambiguity exposure on the cross-section of returns in the US equity market. Results show that there is a U-shaped association between ambiguity exposure and expected return. This pattern is more pronounced for smaller stocks. Besides, U-shaped pattern is left-skewed on average; meaning that expected return increases as absolute ambiguity exposure increases but the absolute effect of the positive exposure is larger than the absolute effect of negative exposure. U-shaped trend is essentially same when controlling for widely known cross sectional predictors, other ambiguity proxies, different holding and formation periods, and different states. Overall, our results provide substantial empirical evidence that indicates that ambiguity is priced in stock returns.