Real-Time Parameterized Expectations And The Effects Of Government Spending

In this paper, we explore the effects of government spending in the real business cycle model where agents use a learning mechanism to form expectations. In contrast to most of the learning literature, we study learning behaviour in the original non-linear model. Following the learning interpretation of the parameterized expect- ations method, agents forecast rules are approximations of the conditional expectations appearing in the Euler equation. We show that variation in agents beliefs about the coefficients of these rules, generates time variation in the transmission of government spending shocks to the economy. Hence, our modelling approach provides an endogenous mechanism for time-varying government spending multipliers in the standard real business cycle model.