We quantify Brazilian banks' systemic risk between January 2010 and January 2019 using Comovement Value at Risk (CoVaR) as a risk measure. We model the conditional depedence beteween Brazilian banks and the representative index of the Brazilian financial system (BFIindex) using copulas. With this risk measure, we assess the impact of several macroeconomic variables on systemic risk in a dynamic linear regression model. Results indicate that systemic risk increased during the global financial crisis of 2008, and that volatility and stock market returns are key determinants of higher or lower systemic risk. Other important macroeconomic variables are the slope of the “zero coupon” yield curve and the 12-month treasury bill rate of change. These results have implications for financial institutions' capital regulation.