This paper documents that the public debt-to-GDP ratio predicts negatively one- to five-year cumulative nominal consumption growth. Moreover, a higher debt-to-GDP ratio is associated with higher yield spreads, controlling for output gap and inflation. I examine these facts in a New Keynesian DSGE model in which growth and inflation are endogenous. In this model, high government debt forecasts low growth and deflation, making bonds attractive assets in high debt states. Furthermore, due to mean-reversions of fundamental processes that drive the economy, longer-term bonds are better hedges than shorter-term ones, resulting in increases in the slope of the term structure at times of high public debt and hence the empirical regularities seen in the data. My paper thus furthers our understanding of what determine bond yields and the impact of quantitative easing.