The slope of the yield curve and the term spread have important implications for macroeconomic outcomes being amongst the key variables in predicting output growth (Estrella and Hardouvelis (1991)). We first document empirical evidence showing the relevance of financial business profitability in explaining real output growth and the linkages between spreads and expected profitability. We then develop a DSGE model with endogenous term spreads derived from banks’ risk assessment of potential liquidity shortages affecting their balance sheets. In our model, fluctuations of the future profitability of banks portfolios affect the ability of banks to cover for any liquidity shortage and hence affect the premium banks require to carry maturity risk. During a boom, profitability is increasing and hence spreads are low, while during a recession profitability is decreasing and hence spreads are high, in accordance with the data. We show that allowing banks to sell long term assets to the Central Bank after a liquidity shock leads to a sharp decrease in long term rates and term spreads. Furthermore, such interventions have significant impact on long term investment, decreasing the amplitude of output responses after a liquidity shock. The short term rate does not need to be decreased as much and inflation turns out to be much higher than if no QE interventions were done.