Changes in the EUR/USD exchange rate can be predicted using news sentiment, according to researchers at the University of Zurich.
Matthias Uhl found that by applying the theory of frequency filtering to financial asset pricing theory, a trading strategy can be developed based on price momentum and news sentiment.
A trading strategy was tested from December 2004 and December 2014 based on price momentum and news sentiment in order to compare the two. Over the ten year period, the mean return per year for the momentum strategies was around 4%, with an annual volatility of 10%. The information ratio sets the returns in relation to the volatility and these were around 0.4.
The mean returns per year for the news sentiment strategies based on news sentiment were on average between 5% and 9%. That is double the returns of the strategies based on price momentum. The annual volatility is similar at 9.9%, but the maximum drawdowns were lower than for price momentum. Given similar volatility and almost double annual returns, the information ratios are also roughly twice as high as for price momentum, ranging from 0.5 to 0.9.
This article will shortly appear in the Journal of Behavioral Finance. More information can be found here.