Working Paper 57 - Exorcism of the Ghost: An Alternative Growth Model for Measuring the Financing Gap

In 1997, William Easterly authored an intriguing paper titled “The Ghost of Financing Gap:How the Harrod-Domar Model Still Haunts Development Economics”. His investigation showsthat the Harrod-Domar model fails drastically in explaining the performance of investment andgrowth. He reveals that the Harrod-Domar model, which supposedly died 40 years ago, is stillused by leading international and regional organizations. Indeed, more than 90 percent of thecountry economists in the World Bank still use the model to make growth and resource requirementsprojections, and to provide advice on economic policy.We argue in this paper that in addition to its well known shortcomings, the Harrod-Domar modelis susceptible to the misinterpretation that all capital inflows (mainly foreign assistance in the caseof Africa) required to fill the financing gap would be invested. This is because it focuses on thesaving-investment relationship in a planned sense (ex ante). In practice, once the foreign exchangeto bridge the ‘planned’ resource gap is obtained, there is no guarantee that all of it would financeinvestment. The wrong presumption that all foreign exchange resources to fill the domestic resourcegap would be invested gave rise to the aid-financed investment paradigm.Recent research has announced the failure of the aid-financed investment in Africa on the empiricalgrounds that a dollar of ODA did not lead to a dollar of investment. We argue that such a conclusionis misleading because it was based on a faulty hypothesis. In many African countries, foreign aidto bridge the resource gap, particularly under program aid might be predestined, by design, toconsumption and not investment. The presumption that all foreign assistance is intended forinvestment might have partly contributed to the negative assessment of aid effectiveness, and hence,to the diminished public support in donor countries for aid programs.This paper introduces the extended version of the Balance of Payments Constraint growth modeldue to Thirlwall and Hussain (1982), as an alternative approach to replace the Harrod-Domarmodel in development practice. Unlike the Harrod-Domar model, the Thirlwall-Hussain model isnot liable to the presumption that all aid would be invested. It clearly shows that the requiredcapital inflows to fill the domestic resource gap would finance imports and that, at least, part ofthese imports would be for consumption and not for investment. Also, the Thirlwall-Hussain modelis capable of measuring, separately, the effects of import volume, export volume, the terms oftrade, domestic inflation and the exchange rate on the amount of foreign exchange required toachieve a target growth rate.The paper illustrates the use of the Thirlwall-Hussain model in measuring the financing gap using asample of 24 African countries. The illustration is conducted against the objective of reducing thenumber of the poor by 4 percent per annum as a requirement for achieving the internationaldevelopment goal of reducing poverty by 50 percent by the year 2015.The results show that given the terms of trade, export volume and initial imbalance in the currentaccount, virtually all the countries in the sample would require large inflows of foreign resources toachieve this target. If export performance is not improved relative to imports, the financing gapwould increase overtime, snowballing to very large amounts. For the 24 countries in the sample,the financing gap is estimated at a total of US $ 44.8 billion (or 19.5 percent of projected GDP)per annum in the first five planning years alone. The cumulative amount of resources for all the 24countries over the first five planning years reaches about US $ 224 billion.However the results obtained should be considered as illustrative of the use of the Thirlwall-Hussain model in measuring the financing gap rather than indicative of the amount of foreignassistance required. This is mainly because the model is applied using generalized assumptionsabout future growth in the model’s variables. Intimate knowledge of individual country circumstanceswould enable the development practitioner to come up with more informed predictions.Unaffected by the quality of data is the theoretical finding of the Thirlwall-Hussain model. Byshifting the focus from the saving-investment relationship to the export-import relationship, themodel provides new pointers for development strategies and for measuring the effectiveness ofdevelopment assistance. It demonstrates that foreign aid can contribute to higher growth ratebecause it finances the excess of imports over exports. However, the model implies that such adependency on aid to support higher growth rates will continue unless the production structureand the pattern of trade in the recipient country are changed to increase export expansion relativeto imports.Thus, if the ultimate goal of foreign assistance is to help poor countries graduate to a self-sustaininggrowth path, the model implies two broad pointers for measuring the long-term developmenteffectiveness of foreign aid. The effectiveness of development assistance should be measuredeither in terms of its ability to promote export relative to imports in the recipient country and/or interms of its ability to create the conditions that will attract private capital. However, even in thislatter case, the model implies that development effectiveness of private capital must be measuredby their contribution to expanding export earnings relative to imports.The basic lesson for development practice is that when foreign exchange is the binding constraint,which is the case of most African countries, overall rates of return on aggregate investment oughtto be measured in terms of foreign exchange earnings. Development practitioners should, thereforeconcentrate less on monitoring the ‘Harrod-Domar variables’ such as saving and investment ratiosand more on innovative variables that reflect the foreign exchange productivity of investment.