Text mining and the information content of Bank of Canada communications

This paper uses Latent Semantic Analysis to extract information from Bank of Canada communication statements and investigates what type of information affects returns and volatility in short-term as well as long-term interest rate markets over the 2002-2008 period. Discussions about geopolitical risk and other external shocks, major domestic shocks (SARS and BSE), the balance of risks to the economic projection, and various forward looking statements are found to significantly affect market returns and volatility, especially for short-term markets. This effect is over and above that from the information contained in any policy interest rate surprise.