Development effectiveness in fragile states: spillovers and turnarounds

This papers examines aid effectiveness in fragile states with particularly weak policies and institutions. It explores whether aid can assist policies and institutions to improve in situations where they are particularly weak.

The paper’s main findings include:

• once a country meets the criteria for a "fragile state" (LICUS), it remains in this status for very long: expected duration is 56 years and expected cost is around $80bn
• even with the limited traditional instruments of a development agency - various types of aid - something can be done to enhance the chances that countries will turn themselves around
• what form of aid is appropriate at what time is important
• technical assistance has no discernable effect until after a turnaround has clearly begun and governments actually want and need to use this aid
• once there are clear signs that the government has itself embarked upon a turnaround, rapid technical assistance is a highly effective form of aid, increasing the chances that the incipient reform will progress to a substantial and sustained improvement and reducing the chances of relapse
• around 4% of GDP appears to be the right amount of technical assistance in early reform environments, although this will obviously have to be nuanced by the particular circumstances
• the government itself should be the best judge of its needs. A practical way of implementing the implied approach would be for newly reforming governments to be given a "Technical Assistance Account" that they can draw down at their own initiative for international expertise
• technical assistance should be organised by donor agencies to be highly responsive to changes in development circumstances
• the evaluation of technical assistance needs to take into account that it is intrinsically a high risk investment, analogous to venture capital
• non-technical aid to improve the pre-conditions for a turnaround should focus on education
• the point at which marginal costs equal marginal benefits is when aid is 22% of GDP. As donors are not risk neutral cannot pool their risks except to a limited extent, 22% should better seen as an upper bound rather than as a target
• non-technical aid also has a role once reform is underway: aid should scale up during the middle of the first decade of turnaround. At this stage there is some evidence that it should be substantial - well in excess of 12% of GDP.