We investigate the impact of aviation accidents on stock index returns. We find that the strong and significant negative impact of accidents on NYSE index returns found by Kaplanski and Levy (2010) becomes much weaker when we control for just a few extreme accidents. These unusually high-impact events generally involve acts of terrorism, such as 9/11, or otherwise coincide with major economic events that were unrelated to the accidents. Our results cast doubt on event studies which suggest that investor mood alone can cause stock markets to suffer large losses. We create three new datasets for aviation accidents, from two different sources, which we link to stock market data. To identify extreme accidents and terrorist incidents, we construct several measures of media coverage, based on the New York Times online archive. We find a significant impact of all severe aviation accidents on NYSE index returns, in line with Kaplanski and Levy. However, accidents involving U.S. carriers alone do not have a significant impact. This suggests that news coverage about “close to home” accidents does not in itself influence investor mood. When we extend our analysis to other market indexes (FTSE100, CAC40, NIKKEI225), we find no evidence of an aviation accident effect that is consistent with the mood effect. Information about terrorism, on the other hand, has a very strong impact.