International tourism, on which we focus in this paper, is recognized to contribute to long-run growth through a whole list of diverse channels. This belief that tourism can promote, if not, plainly, cause long-run growth is known in the literature as the Tourism-Led Growth Hypothesis (TLGH). Our case study of Brazil can also be taken as a particular test for such hypothesis. In our twofold empirical exercise, two different econometric methodologies are applied to two distinct data sets showing, among other things, that results are independent of either data or methodology. On the one hand, annual data from 1965 to 2007 for Brazil as a whole are used for a cointegration analysis to look for the existence of a long-run relationship among variables of economic growth, international tourism earnings and the real exchange rate. The relationship among the variables that we find, can be considered as weakly exogenous, but the test does not support Granger-causality. On the other hand, high quality data for the 27 Brazilian states though for a shorter period, from 1990-2005, allows for the use of the dynamic panel data model proposed by Arellano and Bond (1991). We show that the long-run elasticities between real per capita GDP with respect to tourism receipts and the real rate of exchange are 0.13 and 0.30, respectively. Finally, we compare our results with similar studies also investigating the TLGH showing a relationship between the value of the elasticity of per capita GDP with respect to tourism and the levels of development of tourism in each particular country.