Stock market beta and returns on monetary policy announcement days - Evidence from the cross-section of Finnish stock returns
No Thumbnail Available
School of Business | Master's thesis
AbstractPrior literature has documented large average excess returns on prescheduled macroeconomic news announcement days, hence implying that there exists an announcement day premium. A novel study by Savor and Wilson (2014) reports that on announcement days stock market beta is strongly related to returns, however, similar relation is not found on other trading days. In this thesis, I follow Savor and Wilson (2014) and test whether the announcement day premium exists on the Finnish stock market and whether beta can predict cross-sectional return variation on days when news about monetary policy is scheduled to be announced. The central banks whose monetary policy decisions are of interest are the European Central Bank, the Federal Reserve, Riksbank and the Bank of England. In addition, I will test whether differences in investor sentiment can explain the different performance patterns on the subset of trading days. The data set extends from January 1999 until June 2017 and comprises of firms listed on NASDAQ OMX Helsinki All Share. The source of data is Thomson Reuters Datastream database. The announcement days of the central banks are retrieved from their webpages. To test my hypotheses, I follow Savor and Wilson (2014) and use the Fama-MacBeth cross-sectional regression, which I run separately on announcement days and other trading days. Furthermore, I estimate a single pooled regression to directly test whether the equity risk premium differs between the trading days. Prescheduled monetary policy announcement days have been periods of higher average excess returns than other trading days on the Finnish stock market. Converted into monthly return, stock market excess return has been 3,8% on announcement days compared to only 0,3% return on other trading days. Furthermore, I show that stock market beta is significantly related to announcement day returns and this relation holds for individual stocks and for beta-sorted equally-weighted portfolios. Based on Savor and Wilson (2014), the results imply that beta is an important measure of systematic risk: macroeconomic news announcements provide important information to investors about the state of the economy and thus, they require premium to hold higher-beta assets. Specifically, I show that the risk premia earned on the announcement days of the Federal Reserve and Riksbank are significantly higher than on other trading days. Similar relation is not found on announcement days of the European Central Bank or the Bank of England. Lastly, I show that in addition to systematic risk, also investor sentiment varies with the two types of trading days. Thus, both economic fundamentals and investor sentiment affect the returns on the subset of trading days.
Thesis advisorTorstila, Sami
capital asset pricing model, announcement days, macroeconomic news, cross-section of returns