ABSTRACT
In Thirlwall’s (1979) seminal paper the ideawas advanced and empirically verified that for a large groupof countries the rate of growthof output is balance of payments (BP) constrained because this sets the limit to the growth of demand to which supply can adapt (Thirlwall and Hussain, 1982). The main essence of ‘Thirlwall’s Law’ is that in an open economy, expenditure cannot grow faster than income growth without creating a current account deficit on the BP. A current account deficit cannot be sustained through an indefinite inflow of capital, because deficits above a certain percentage of Gross Domestic Product (GDP) trigger negative signals to the international community that force countries to adjust (McCombie and Thirlwall, 1997a). For a given rate of growth of exports, the brunt of the adjustment falls on a reduction in income growth to restore BP equilibriumbecause relative price changes do not act as an efficient BP adjustmentmechanism. An open economy’s real economic growth rate is therefore determined by export growth for a given income elasticity of the demand for imports. The rate of growth of exports, in turn, is mainly a function of ‘world’ income or ‘world’ demand. The main objective of this chapter is to apply a ‘generalised’ version of the
BP growth model by testing for long-run relationships between the output growth rates of OECD countries and two neighbouring regions; South Africa (SA) and the rest of the Southern African Development Community (RSADC).1 In this context, the chapter attempts to make the following contributions to the existing theoretical and empirical literature. First, Thirlwall’s BP constrained growth model is a specific case involving a
bilateral trade relationship between one country and the ‘rest of the world’. In this chapter, the specific case is generalised into a multilateral trade relation between an individual country (SA) and blocks of countries (OECD and RSADC). One of the main findings of the chapter is that the policy implications of the ‘generalised’ BP growth model present a different perspective compared to the ‘specific’ BP model. The policy suggestions are particularly relevant to neighbouring regions
that participate in a mutual economic cooperation scheme, and specifically, where one of the regions (SA) dominates the other (RSADC) in terms of economic size.2
The ‘generalised’ BP growth model provides a supplement to McCombie’s (1993 andChapter 5) theoretical version inwhich the long-run growth rate of an advanced country (or blocks of countries) depends on the growth rate of another through the BP constraint. Second, since all the output growth rate variables in this chapter are stationary
{I(0)}, the econometric methodology employed departs from standard cointegration techniques such as the Johansen procedure which tests whether non-stationary variables {I(1)} cointegrate to form an I(0) process (see Johansen and Juselius, 1990). The methodology draws on recent advances in time-series econometric techniques such as Pesaran et al.’s (2001) bounds testing procedure and the autoregressive distributed lag (ARDL) approach to estimating the long-run (Pesaran and Shin, 1999). The bounds testing procedure is particularly relevant to this application as it not only provides critical values for a set of purely I(1) variables but also for a set of purely I(0) variables. The ARDL approach, on the other hand, yields consistent estimates of the long-run coefficients that are asymptotically normal even when all the underlying regressors are I(0). As a supplement to the long-run growth rate equations, simulations based on impulse responses and forecast error variance decompositions are used to illustrate the mutual interdependence of the two neighbouring regions through the BP constraint. The rest of the chapter is organised as follows. The second section derives the
‘generalised’ version of theBPmodel. The third section discusses themain features of the data. The fourth section presents the econometric methodology and the fifth section the empirical results. The last section concludes with policy implications.
Following Thirlwall (1979, 1999), SA’s current account of the BP, measured in its own domestic currency, may be written as:
