Estimation of a Structural Break Point in Linear Regression Models

This study proposes a point estimator of the break location for one-time structural in linear regression models. If magnitude is small, least-squares date has two modes at ends finite sample period, regardless true location. To solve this problem, I suggest an alternative based on modification objective function. The modified function incorporates estimation uncertainty that varies across potential dates. new consistent and unimodal distribution under small magnitudes. A limit provided in-fill asymptotic framework. Monte Carlo simulation results outperforms estimator. apply method to estimate U.S. real GDP growth UK stock return prediction