Differential effects of unconventional monetary policy

Did the Eurosystem's quantitative easing from 2015 to 2018 have differential effects regarding the bank lending volume to different institutional sectors, industry sectors, or types of loans? To investigate this question, this paper employs linked microdata of the German banking system. These allow for computing the volume of bond redemptions at bank level as a measure of banks' exposure to QE. Because when a bond matures, the bank is faced with the decision of whether to reinvest the proceeds into bonds or whether to rebalance into another asset such as loans. When the central bank squeezes bond yields through large-scale purchases, banks with more redemptions have a stronger incentive to rebalance. However, a fixed effects model reveals no significant difference between banks with a high exposure compared to the control group regarding their overall loan growth. Neither can any of the mentioned differential effects be observed. While these findings are at odds with some of the previous empirical literature, they are in line with theories that argue that lending is purely demand-led and any central bank action geared towards the supply side of the loan market merely constitutes 'pushing the string'.