Association for Public Policy Analysis & Management

Many macroeconomic and financial time series are subject to structural breaks. One of the concerns of economists and statisticians is how to estimate the breakpoints and coefficients under structural breaks. In this paper, we propose the combined estimator of a restricted estimator in which we can have any restriction on the coefficients and unrestricted estimator in which we estimate the coefficients within each regimes separately by using the observations on that specific regimes. We derive the optimal weight by minimizing the exact weighted Mean Square Forecast Error (MSFE). The superiority of this paper over other papers in the same context is that, here we find the combination weight by minimizing MSFE(β,W), with any given forms for W. We compare the in sample and out of sample forecast, and compare the results of our proposed estimator with some of the existing ones. Monte Carlo simulations proof the superiority of our estimator.