In assessing the debt-reduction/aid trade-off, it should be noted that the effects of debt reduction and of aid are somewhat different, and will depend on the nature of the aid which is lost. The use of aid is to a great extent within the control of the donors, and much of it will be ear-marked to specific investment projects. The cost of losing this aid will depend to a considerable extent on the nature of the investment concerned, as well as on its financial terms.
By contrast, debt reduction (beyond the level at which debt would otherwise have been paid) releases resources which can be allocated to any purpose (subject to whatever conditionality may be applied). Even if each $1 of debt reduction is off-set by a $1 reduction in aid, this may be beneficial if resources for recurrent spending are a major constraint, or if the aid lost would have supported a relatively low-priority project.
The timing of benefits is an important consideration to take into account. The effect of a reduction in aid applies essentially to the year in which the aid would have been received (although the future income on the investment which it would have financed is lost). In the case of debt-reduction, the effect is a continuing stream of benefits: other things being equal, the debt will be lower (and the debt overhang reduced) in every future year. Similarly, if the debt is reduced below the payable level, the saving on debt servicing will also continue for an extended period.
The strength of the trade-off between aid and debt-reduction in terms of their effects on investment and growth is very difficult to assess, and the results are critically dependent on the specification used. Based on the (rather simplistic) approaches taken in Woodward (1996b), for example:
the relationship of new investment to the stock of debt and net inflows of official finance suggests that $1 1of aid is equivalent in its effect to $20-25 of (nominal) debt reduction; but
the relationship of the growth rate of investment to changes in the debt stock and in net inflows of official financing suggests that $1of aid is equivalent to about $3 of (nominal) debt reduction58.
However, three important factors should be noted in considering these results. Firstly, while the above figures are based on changes in the nominal (face) value of debt, the cost of debt reduction is equivalent to the change in its present value (as defined earlier). Accordingly, a $1 reduction in nominal debt costs substantially less than $1 - perhaps in the order of $0.7559. Secondly, as noted above, the effect of $1 of aid is limited to the year in which it is received, and should therefore be fully captured in the above estimates, while debt reduction generates a continuing stream of benefits, of which the above estimates include only a single year. The real trade-off which we need to consider is therefore between the one-year effect of aid, and the present value of the whole stream of benefits from debt reduction.
Taking account of these factors, and assuming that the stream of benefits from debt reduction continues indefinitely, the effect of $1 of debt reduction is roughly equal to that of $1 of additional aid based on the first relationship, and about 7-8 times as much based on the second. This would appear to suggest that, even if there were a one-to-one trade-off between debt reduction for the HIPCS and aid for low-income countries, the net effect would at worst be neutral and could be strongly positive.
As discussed previously, it would be relatively easy to ensure that the trade-off with aid was less than one-for-one - and at least possible to eliminate it entirely. This would make the effect unambiguously positive. Prioritising the sources which only have an impact on aid flows well into the future (eg IDA reflows) would also tend to increase the net benefits. Even with a one-to-one trade-off with immediate aid disbursements, the effect could be made more positive by diverting aid flows which have a below-average development impact, eg., those on less favourable financial terms, tied to exports from the donor country, or for low-quality or low-priority projects.
Clearly, this conclusion must be treated as tentative, because of the very basic analytical methods used. However, there is some reason to expect that the net benefits may be under-stated: the estimates above exclude investment in human capital (health and education) and a range of other factors which may contribute substantially to growth (eg., maintenance of infrastructure, administrative capacity, political stability, etc.); and these are likely to benefit more from debt-reduction than from aid, as they are primarily related to recurrent spending rather than to investment.