A0496
Title: Fluctuation monitoring for structural change
Authors: Wen-Jen Tsay - Soochow University (Taiwan)
Hsin-Chieh Wong - National Taipei University (Taiwan) [presenting]
Abstract: A fluctuation-based procedure is developed to monitor out-of-sample stability in ordinary least squares (OLS)-based linear regression models. To this end, feasible boundaries are proposed for the fluctuation test that asymptotically control the test size, significance level, and maturity. The theoretical analysis yields closed-form expressions for the Laplace transform under fixed maturity and the optimal boundary under random maturity. Simulation studies demonstrate that the proposed procedures attain the nominal significance levels across various scenarios, supporting their effectiveness in detecting structural change.
