ON EXPORT-LED GROWTH IN AUSTRALIA AND CANADA: COINTEGRATION, CAUSALITY AND STRUCTURAL STABILITY*

This paper uses cointegration and vector-error correction (VEC) modelling to investigate whether the export-led growth hypothesis applies to Australian and Canadian exports and labour productivity. The Johansen (1988) procedure is used to test whether manufacturing and aggregate measures of exports and labour productivity are cointegrated. Exports and labour productivity are found to be cointegrated, suggesting the export-led growth hypothesis holds for the small, open Australian and Canadian economies, although the estimated cointegrating vectors suggest the relationship is of small order. The tests suggested in Hansesn (1992), for parameter stability in relationships with I(1) variables, are estimated to examine if the parameters estimated from the Johansen procedure are unstable and whether the null of cointegration can be rejected in favour of structural change. No evidence is found of parameter instability in either the Australian or Canadian long-run relationships at conventional significance levels. Estimation of VEC models suggests that exports are seen to cause productivity growth, although the quantitative impacts are again small. The reverse causality is rejected for both countries, except for the Canadian manufacturing sector for which there is a small, significant positive effect of labour productivity on manufactured exports.