Estimating the money market microstructure with negative and zero interest rates

The Furfine [The microstructure of the federal funds market. Financ. Markets Inst. Instrum., 1999, 8(5), 24-44] algorithm is a key tool for researchers in monetary economics and central banking because it allows identification of the interbank money market microstructure. Furfine's method has been used in numerous studies of money markets but little work has examined how valid its application is at zero and negative interest rates. This work seeks to address this gap. We first extend the standard algorithm to include negative rates and show the bias introduced by the zero interest rate. Secondly, we propose a econometric procedure based on market regularities and the economic likelihood of interbank bilateral relationships to correct the bias. Thirdly, the methodology is applied to TARGET2 data. The impacts of recent monetary policy decisions on key interest rates are studied. We find significantly biased figures when the market microstructure is not correctly estimated.