Measuring stock market contagion: Local or common currency returns?

Empirical research on contagion between international stock markets generally focuses on index returns converted into US dollars. This paper argues that it would be more appropriate to use returns denominated in countries' local currencies, as only these returns accurately reflect price fluctuations in national stock markets. Returns converted into a common currency also reflect fluctuations in the exchange rate, which is shown to bias the outcomes of a contagion test.