“…In contrast, over a third apply a multivariate analysis which supports the hypotheses of export-led growth (Riezman et al, 1996;Islam, 1998), growth-led exports (Arnade and Vasavada, 1995), bi-directional causality (Anoruo and Ramchander, 2000;Awokuse, 2005a) and increased output growth leading to reduced exports (Kónya, 2004). Additionally, when undertaking bivariate analysis, studies reported bi-directional causality (Chow, 1987;Bahmani-Oskooee et al, 1991;Bahmani-Oskooee and Alse, 1993;Rahman and Mustafa, 1997;Ekanayake, 1999), non-causality (Dodaro, 1993;Dutt and Ghosh, 1996;Riezman et al, 1996;Islam, 1998) and that increased output growth leads to a decrease in export growth (Jung and Marshall, 1985;Kónya, 2004). Furthermore, several studies investigate the Granger causality test using the error correction model with bi-directional causality detected (Bahmani-Oskooee and Alse, 1993;Rahman and Mustafa, 1997;Ekanayake, 1999;Anoruo and Ramchander, 2000;Awokuse, 2005a), while evidence of non-causality (Dutt and Ghosh, 1996) and growth-led exports (Arnade and Vasavada, 1995) Toda and Yamamoto (1995) to reveal a negative link running from GDP to exports.…”