Volume 3, No. 1, April 2004
A Quasi-Bayesian Analysis of Structural Breaks: China's Output and Productivity Series


Xiao-Ming Li*
Department of Commerce, Massey University-Albany, New Zealand


Abstract


A quasi-Bayesian model selection approach is employed to detect the number and dates of structural changes in China's GDP and labour productivity data. It is shown that the predictive likelihood information criterion is valid only among models with well-behaved residuals.


Key words : structural change; predictive likelihood; GDP; labour productivity; C hina
JEL classification : C22; E30

Back