Technology- and investment-led growth effects of economic integration: a panel cointegration analysis for the EU-15 (1960–2000)

Abstract

Using a panel cointegration approach we test for technology- and investment-led growth effects of economic integration for the EU-15 Member States over the period 1960 to 2000. Integration is measured by an index that is mainly based on tariff reductions and accounts for both GATT-liberalization and European integration. We find that integration has induced sizeable level effects on GDP per capita of some 44%, with both technology-led and investment-led effects playing an important role. While integration-induced efficiency increases materialize within a few years, integration-induced effects on the equilibrium stock of capital require a long time to work themselves out.

Acknowledgements

I am grateful to Joerg Breitung for a number of clarifying comments on the two-step estimator and for providing the critical values for the cointegration tests in Table 1. I also wish to thank the editors for their valuable remarks. The usual disclaimer applies.

Notes

¹ Since we did not find significant and robust results for human capital, we omit human capital from the production function from the beginning to simplify the exposition.

² The single unit root tests do not suggest that the present paper is subject to this problem (but again, the low power of the unit root tests has to be borne in mind).

³ Note that the Larson et al. (2001) test allows for a heterogeneous cointegrating relationship. Under the null of no integration (r = 0) this makes no difference (since there is no cointegrating relation to be equated across countries under the null). But also for r > 0 the gain in power from imposing homogeneity will vanish asymptotically.

⁴ The lagged level of INT was used since it yielded a better fit. This is plausible, since the effects of integration do not materialize immediately.