By: Arshad Ali Bhatti, IIIE
This paper addresses the issue of outliers in finance-growth literature and provides a robust sensitivity analysis of some past studies and an updated data set. We employ the robust regression methods of median quantile regression and least trimmed squares. It shows that the findings of past studies are sensitive to outlier observations. Further, we find that the positive effect of financial development on growth disappears and even becomes negative once we use our extended data set of 86 countries over the period 1997-2006. This last finding is consistent with Rousseau and Wachtel (2011). Moreover, we investigate whether our understanding of the finance-growth relationship can further be improved by introducing a measure of R&D into the standard analysis. We note that our measure of R&D has a strong positive effect on growth and may proxy the role of an omitted variable which is highly correlated with economic growth.http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2432519
