A re-examination of the Balassa-Samuelson hypothesis using recent panel data unit-root and cointegration tests: Evidence from MENA countries
Résumé
The main goal of this paper is to reassess the Balassa-Samuelson hypothesis using annual data for 16 Middle East and North African (MENA) countries. We carry out recent panel data unit-root tests proposed by Im et al. (1997) and recent panel data cointegration techniques developed by Pedroni (1999) and we compare the results with those obtained with conventional time series unit-roots and cointegration tests. Our main finding is that whereas the standard time series approach rejects the Balassa-Samuelson hypothesis for 11 countries out of 16, new panel cointegration techniques rescue this hypothesis for MENA countries. © African Development Bank 2003.