RDP 2007-02: Current Account Deficits: The Australian Debate 4. Optimality and Sustainability: An Empirical Assessment

The intertemporal approach to the current account forms the foundation of Pitchford's (1989b, 1989c, 1990) view of the current account. Using the methodology developed by Campbell (1987) and Campbell and Shiller (1987), several studies test whether Australian current account data support the intertemporal model, with mixed results. Milbourne and Otto (1992), using quarterly data, reject the intertemporal model while, in contrast, Cashin and McDermott (1998) and Otto (2003), using annual data, and McDermott (1999), using quarterly data, find supportive evidence, but only after 1975, 1980 and 1991 respectively. Bergin and Sheffrin (2000) extend the intertemporal model to account for external shocks by allowing the interest rate and exchange rate to vary. They find that this improves the fit of the model by better capturing volatility, thereby providing support for the intertemporal model.

Following these studies, this section of the paper examines optimality through the lens of the intertemporal approach to the current balance, but with two innovations. First, account is taken of the effect of the capital market opening and financial market deregulation with the advantage of a longer sample of data postdating these changes. Prior to this, net foreign debt may have been less than optimal (due to consumption and/or investment being too low), and credit constraints may have prevented optimal consumption smoothing in the face of shocks to income. The second innovation is to account for the fact that shocks to the Australian net cash flow (output minus investment and government expenditure) may be correlated with shocks in the rest of the world and, as a result, have a limited effect on the current account (Glick and Rogoff 1995). That is, global shocks should lead to changes in the world interest rate rather than current account balances.

The full details of the model and estimation approach, along with detailed results, are reported in Appendix A. In summary, there is tentative evidence in support of the intertemporal model. In particular, the current account balance appears to adjust in a way that is consistent with consumption smoothing in the face of temporary shocks to output, government expenditure and investment. This is true, however, only in the period after financial liberalisation in the early 1980s, in line with the removal of capital controls and the easing of credit constraints. There is also evidence of consumption tilting, whereby Australian residents appear more impatient than the world as a whole. This has contributed to a persistent current account deficit in the order of 4½ per cent of GDP since the mid 1980s.

It is worth considering what might justify a persistent degree of impatience and the resulting long history of current account deficits. In the case of Australia, the desire to build up the capital stock (both private and public) while still maintaining a relatively high level of consumption would seem a natural consequence of a relatively undeveloped, ‘new’ country with considerable natural wealth. This is particularly true of one which benefits from a relatively steady flow of immigrants and institutional features conducive to sustaining a relatively prosperous and stable lifestyle.

While the estimates presented in Appendix A suggest that the extent of this impatience appears relatively modest, it is not possible to test the solvency condition – that is, whether or not the intertemporal budget constraint has been satisfied. Indeed, as Milesi-Ferretti and Razin (1996) note, in practice it is difficult to determine whether a country running persistently large current account deficits is solvent at any given time. The more feasible test is to examine the sustainability of the situation – that is, to determine the level of trade surplus (and hence also the current account balance) required to stabilise the level of net foreign liabilities (relative to GDP) given plausible assumptions about output growth and the costs of servicing net foreign liabilities. A number of studies have undertaken this type of exercise for Australia. For example, Gruen and Sayegh (2005) find that an average goods and services trade surplus around ½ to ¾ per cent of GDP can sustain foreign liabilities at a ratio of 60 per cent (which compares to the actual deficit on the trade account of 1½ per cent of GDP, on average, since 1980). Alternatively, if the trend current account balance (of about 4½ per cent of GDP since 1984) were to be sustained, net foreign liabilities would eventually stabilise around 86 per cent of GDP (assuming average growth of nominal GDP of 5½ per cent per annum).

The limitations of such calculations, however, is that they do not consider what sort of changes would be needed to bring about the turnaround in the trade balance (and the associated reduction in the current account), nor exactly when these changes need to occur. Again, this reflects the difference between solvency and sustainability, with the latter being an assessment of what constitutes a stable equilibrium, while the former allows for the possibility that there may be even higher, and potentially sustainable, levels of foreign indebtedness that are welfare enhancing.