Economic policy uncertainty and the information content of earnings announcements: Empirical evidence from the US in 1985-2018

Abstract

This thesis studies the impact of economic policy uncertainty on the information content of earnings announcements. There is a large body of literature regarding both economic policy uncertainty and the information content of earnings announcements, but the relationship between the two phenomena has not yet been studied. This study gives an overview of the topic with empirical evidence, paving the road for future research.

Economic policy uncertainty (EPU) index by Baker et al. (2013) has been used as a measure of economic policy uncertainty. This recent measure has been found to capture many aspects of economic policy uncertainty, and therefore it has been widely used by academics in many fields. It has been found to have an adverse impact on many areas of economic, including but not limited to decreased GDP growth, M&A activity and capital investments, and increased stock price volatility and equity risk premium.

This thesis continues the research on EPU index by investigating its impact on the information content of earnings announcement in US stock listed companies between 19852018. This is done by observing the trading volume of stocks in the earnings announcement window, that includes the daily firm/trading volume observations that are within 16 days of the closest earnings announcement. With this dataset of 17,015,045 observations, two analyses are done. Firstly, a visual illustration is provided by replicating Beaver (1968) methodology but incorporating high/low EPU groups to show the difference. In addition, OSL regressions are run for the whole period and sub-periods of approximately 11 years, to explain the trading volume by the release of the earnings announcement, a proxy for economic policy uncertainty, and the interaction term of the two.

It is found that economic policy uncertainty decreases trading volume during the earnings announcement window. In addition, the magnitude of this effect increases during the release day of an earnings announcement, and one day after that. This shows that the information shock caused by earnings announcement decreases during high economic policy uncertainty, implying a decrease in the information content of earnings announcements. The results regarding the changes in this phenomenon across time are conflicting, and no clear conclusions can be drawn from those.

This thesis provides a starting point for research on the area but is also prone to limitations. The most important one being the lack of control variables, that are suggested to be incorporated as a part of future research. In addition, a qualitative approach could be taken to explain why the information content decreases during the times of high economic policy uncertainty.