Level-k DSGE and Monetary Policy

This paper develops a new framework of level-k DSGE for monetary policy analysis. Incomplete markets are introduced to guarantee the eductive stability of the equilibrium. k=1.265 is identified using growth and inflation expectations from the Michigan Survey of Consumers, capturing the missing indirect channels and weakened direct channels in households’ forecast rules, as well as the wedge between expectations and reality. With more empirical support, the model is applied in four different issues related to monetary policy. First, the real effects of monetary shocks are accumulative when reasoning levels are low and households’ planning horizons are short. Second, inflation targeting confuses households with the dynamics of GDP, and hence weakens demand stabilization. Third, in liquidity traps, recovery can be fast, slow or even impossible, depending on how deep the recession is. Forth, the initial effect of monetary shocks in far future is dampened by level-k, but the cumulative effects across time can be large. When k → +∞, the level-k DSGE reduces to a basic New Keynesian model as in Gaĺı (2015). ∗Department of Economics, University of Pennsylvania, Email: [email protected]. I am indebted to Jośe-Vı́ctor Ŕıos-Rull for continuous guidance, Dirk Krueger and Harold Cole for support throughout this project. I wish to thank Enrique G. Mendoza, Guillermo L. Ordoñez and Jesús Fernández-Villaverde for very helpful comments, and all participants at the Penn macro lunch and workshop, especially Xiang Fang and Jianhong Xin for discussion. I also wish to thank Bruce Preston, Raffaele Rossi, Joseph Stiglitz, Petra E. Todd, Yong Wang and Eric Young for helpful discussion during my job search. All errors are my own.