The robust permanent income model revisited

Abstract: The robust permanent income model discussed in a number of works, see e.g. Hansen et al. (1999, 2002), is reformulated as a linear quadratic tracking problem with a time-varying intercept following a ‘Return to Normality’ model. The results in Tucci (2005), which implicitely assumed desired paths for the states and controls equal to zero, are generalized to the case where the objective function depends upon arbitrary desired paths. By comparing robust control with the optimal control for a linear system with time-varying parameters, in this more general case, it is confirmed that the decision maker applying the former is indeed assuming a very special kind of model misspecification