High-frequency market reactions to unscheduled stock-specific news : an empirical analysis of the intraday market dynamics at the Oslo stock exchange
MetadataShow full item record
- Master Thesis 
We use pre-processed news data combined with high-frequency stock data from the Oslo Stock Exchange to test the following hypotheses: (1) the sentiment of news articles can predict the direction of intraday abnormal returns; and (2) intraday volatility and trading activity increase around the arrival of news articles. First, we find that abnormal returns are significantly negative at 17 basis points 90 minutes after a negative news release. In contrast, we cannot establish a significant relationship between abnormal returns and news with a positive or neutral sentiment. Second, by using a high-frequency vector autoregressive model, we find that: (1) volatility increase on average 0.47 standard deviations ten minutes before a news arrival; and (2) money value traded increase by 0.48 and 0.47 standard deviations five and ten minutes before news arrivals. Thus, our results suggest that negative news articles affect the abnormal returns more than positive news articles and that unscheduled news affects the intraday volatility and trading activity at the Oslo Stock Exchange.