Abstract
Employing an advanced bootstrap VAR model with a fixed rolling window, we investigate the causal nexus between insurance activities and banking credit in China. Parameter stability tests show that none of the traditional VAR models have stable parameters, and hence the full sample results are unreliable. Rolling window results indicate the time-varying causal nexus between them in various subsamples in China and two key structural change points – the years 2002 and 2008.
Notes
1 According to these findings in previous works, this article uses residual based modified-LR test on the basis of bootstrap technique, as suggested by Shukur and Mantalos (Citation1997, Citation2000), Mantalos and Shukur (Citation1998), Mantalos (Citation2000), and Hacker and Hatemi-J (Citation2006), to investigate the short-run causal relationship between real insurance density and real banking credit.
2 Using the Johansen cointegration tests we find strong evidence that insurance activities and banking credit are cointegrated. The estimation results are not reported in this article, and available from the authors upon request.