Working Paper, University of Utah, Department of Economics 2011-03
Carlin and Soskice (2005) advocate a 3-equation model of stabilization policy. One equation is a monetary reaction rule MR derived by assuming that governments have performance objectives, but are constrained by a Phillips curve PC. Central banks attempt to implement these objectives by setting interest rates along an IS curve. They label this the IS-PC-MR model. Observing that governments have more tools than just the interest rate, we drop the IS equation, simplifying their model to 2 equations. Adding a random walk model of the unobserved potential growth, we develop their PC-MR model into a state space specification of the short-run political economy. This is an appropriate econometric method because it incorporates recursive forecasts of unobservable state variables based on contemporaneous information measured with real-time data. Our results are generally consistent with US economic history. One qualification is that governments are more likely to target growth rates than output gaps. Another is that policy affects outcomes after a single lag. This assumption fits the data better than an alternative double-lag timing: one lag for output, plus a second for inflation has been proposed. We also infer that inflation expectations are more likely to be backward rather than forward looking.
new Keynesian stabilization policy targets microfoundations real-time data