Do panel data permit the rescue of the Balassa-Samuelson hypothesis for Latin American countries?

Abstract

This paper tests empirically the Balassa-Samuelson (BS) hypothesis (i.e. rapid economic growth is accompanied by real exchange rate appreciation because of differential productivity growth between tradable and non-tradable sectors) using annual data for twenty Latin American countries. The paper applies new panel data unit-root tests proposed by Im et al. (Discussion paper, University of Cambridge, June 1997) and new panel data cointegration techniques suggested by Pedroni (Oxford Bulletin of Economics and Statistics, 1999) and the results are compared with those obtained with conventional time series unit-roots and cointegration tests. The main finding is that whereas a standard time series approach rejects the Balassa-Samuelson hypothesis for 11 countries out of 20, new panel cointegration techniques permit the rescue of this hypothesis for Latin American countries, as well as for Central American and South American groups of countries considered separately.