This paper examines systematic risk (betas) of Australian government debt securities for the period 1979–2004 and makes three contributions to academic research and practical debate. First, the empirical work provides direct evidence on the systematic risk of government debt, and provides a benchmark for estimating the systematic risk of corporate debt which is relevant for cost of capital estimation and for optimal portfolio selection by asset managers such as superannuation funds. Second, analysis of reasons for non-zero (and time varying) betas for fixed income securities aids understanding of the primary sources of systematic risk. Third, the results cast light on the appropriate choice of maturity of risk free interest rate for use in the Capital Asset Pricing Model and have implications for the current applicability of historical estimates of the market risk premium. Debt betas are found to be, on average, significantly positive and (as expected) closely related, cross sectionally, to duration. They are, however, subject to significant time series variation, and over the past few years the pre-existing positive correlation between bond and stock returns appears to have vanished.