摘要：The analysis of Kirchgässner and Wolters (2009, KW henceforth) corrob- orates the findings of Baltensperger, Jordan and Savioz (2001), 1 who find compelling evidence for M3 as information variable in the Swiss monetary policy process. Using data covering the period from 1983 to 2008, KW extend the analysis and estimate money demand equations for the monetary aggregates M1, M2 and M3. They obtain economically sensible estimates, in particular a unit income elasticity for M1 and M3. For all money aggregates, they report a nega- tive sensitivity to the 3-month LIBOR, which confirms the prerequisite of con- trollability for a money aggregate to be used in policy analysis. Money demand is estimated to be stable over the sample period. In particular the interest semi- elasticity of M1 and M2 demand is estimated to be robustly negative. This is less so for M3 demand, see below.
