The impact of public information on financial markets

Information is the global currency of today and nowhere is it more true than in financial markets. Specialized agencies like Reuters, Bloomberg and Dow Jones invest large sums to develop the fastest, most accurate and most reliable systems of delivering thousands of news every day to investors around the world. The sheer volume of news has long exceeded the capacity of human beings to process it manually but advances in computer linguistics and text processing open up unique possibilities for a systematic and automated analysis. Based on these advances, this dissertation deals with the impact of news announcements on the prices of stocks mentioned in the news, on the prices of other stocks, and on the broader market. Separately, it deals with the question of whether quantifying the search for information can offer insights about investor uncertainty. The first contribution of this dissertation is to show that negative news is, broadly speaking, more informative. This is true both with respect to the announcing company as well as other companies in the market. Greater impact of negative news seems related to the fact that periods of bad financial news coincide with periods of increased interest in the economy. Using internet searches as a proxy for information search provides clear evidence of this relationship. The second main contribution is to give insight into the activities of news agencies, which seem more successful at filtering out interesting company announcements and presenting their content in a more transparent way, than in generating novel stories on their own. The final contribution is to present a parsimonious method for aggregating the tone of news across companies in order to obtain market-wide indicators of information flow. A measure of news disagreement emerges as a particularly successful predictor of future returns and volatility.