An alleged Canadian income trust announcement leak of 23 November 2005 provides a remarkable example of the sensitivity of event study analysis to the treatment of cross-sectional dependence in returns. Whereas a leak should chiefly have affected the returns on other securities, not income...