Testing the demand-led and BOP-constrained growth model: a system approach

Abstract

In a series of articles of the journal Applied Economics there was an extensive debate on the Balance-of-Payments (BOP)-constrained growth model and its testing. This led to a single-equation-based test for the necessary condition of the BOP-constrained growth model that the BOP constraint binds. Since then many extensions and applications of the basic BOP-constrained growth model have appeared in the related literature, but the empirical model evaluation methodology has not been developed to further test the sufficient condition that causality runs from foreign demand or exports determined by it to GDP rather than the other way round. This study utilizes the standard vector error correction model (VECM) to formalize a system approach to empirical testing of both the necessary and sufficient conditions.

Notes

¹ Mostly Equation 5b is considered since otherwise statistical inference is problematic (see McCombie, 1989). The reason is that calculation of confidence intervals for the π⁻¹ε empirical analogue leaning on the estimates of π and ε from Equations 1 and 2 becomes cumbersome. The problem could be solved using the bootstrap though.

² In the sense that no over-identifying restrictions are imposed.

³ The BOP-constrained growth model defined above implies three cointegrating vectors and, for simplicity, it is assumed that this holds hereafter. However, due to other reasons, there could be more cointegrating vectors and they should be left in β. This does not invalidate the Post-Keynesian BOP-constraint growth theory, just points to that the theory presents only a partial explanation of the actual process.

⁴ Although, at the first glance, testing this hypothesis for a small country might seem redundant (but not that of exogeneity of exports discussed in the next section), it is not that clear what a small economy is and when it starts affecting the foreign demand. For instance, foreign demand might be not an exogenous one, if small economies traded mostly with one another.

⁵ In addition, the silent assumptions of the model on exogeneity–for instance, that foreign income affects exports and that the domestic GDP affects imports, but not vice versa–could be tested analogically.