Monetary policy at the zero lower bound: Information in the Federal Reserve’s balance sheet

I examine the impact of the actual purchases of Treasury securities by the Federal Reserve on Treasury yields. Using structural stability tests I find significant breaks in the relation between these variables. I find that in the zero lower bound period following the first phase of quantitative easing, May 2010 to December 2015, the actual purchases of Treasury securities by the Federal Reserve are positively related to changes in Treasury yields. This effect is driven primarily by the positive relation of the Treasury purchases with the bond risk premium, but they are also positively related to the expected inflation rate and the real rate of interest. The evidence is consistent with the liquidity channel hypothesis as put forward by Krishnamurthy and Vissing-Jorgensen (2011), since the Federal Reserve’s Treasury purchases also strongly predict a lower corporate yield spread. Using a macro-finance term structure model I provide counterfactual estimates of the Treasury yields in the zero lower bound period. The counterfactual 10-year yield and the term premium are considerably smaller during QE2 but close to the actual time series for most of the rest of the lower bound period.