Summer 2009 Stephen Knack
Windfall revenues from both foreign aid and natural resource exports can weaken governments’ incentives to design and maintain efficient tax systems
A large body of literature shows that states benefiting from oil and other natural resource rents suffer from a "resource curse": they tend to be less democratic, more corrupt, and less interested in building state capacity for collecting taxes. Some studies have extended the resource curse phenomenon to official development assistance, arguing that both aid and resource revenues can be thought of as sovereign rents that generate dysfunctional rent-seeking behavior, and tend to receive less scrutiny from legislators and citizens than other revenues.
Arguably, these revenues can reduce a government’s dependence on taxes from its own citizens, leading to long-run dependence on foreign aid and erosion of the quality of the tax system. Indeed, empirical work has found foreign aid to reduce tax revenues as a share of national income.
A recent study by Stephen Knack tests the hypothesized relationship between windfall revenues and the quality of tax systems as measured by the World Bank’s Efficiency of Revenue Mobilization (ERM) rating for 135 developing countries. This measure scores countries on a scale of 1 to 6 taking into account a range of tax policy and administration characteristics. Knack finds that high levels of aid do tend to undermine the quality of tax systems. An increase in aid’s share of gross national income by about 30 percentage points is associated with between a half-point and a one-point reduction in the ERM rating, depending on the econometric specification used. The effect of aid remains significant under all specifications and is not sensitive to outliers.
Revenues from natural resources are also associated with weaker tax systems: an increase in net fuel exports per capita of about $500 is associated with a half-point reduction in the ERM rating. However, this effect is neither as strong nor as statistically robust as that of aid, and it is sensitive to a small number of outliers. Moreover, the effect varies substantially depending on which measure of natural resource abundance is used.
These findings are inconsistent with existing perspectives on the impact of aid revenues relative to that of natural resource revenues. It has been argued that while aid reduces the need for taxation and thus increases sovereign rents relative to better scrutinized revenues, it comes with various donor-imposed mechanisms of scrutiny that may offset its negative effects by substituting for reduced pressure from citizens.
The study warns that while this inconsistency is noteworthy, it may arise due to several factors. For example, it is possible that these data measure aid rents more accurately than they do resource rents. As a result, estimates of the impact of resource rents on quality of tax systems are less accurate than the corresponding estimates for aid, and natural resource revenues may still in fact be more damaging than the latter.
In addition, the difference in impacts may be an artifact of the sample. The dependent variable, ERM, is produced only for countries eligible for World Bank loans. Some high-income, resource-abundant countries with large resource rents, such as Saudi Arabia and Kuwait, are excluded from the sample. This might bias the results.
The study also cautions that policy responses to dependence on windfall revenues should distinguish between the impact of aid on the quality of tax systems and the net impact of all donor activity, which might include policy advice not linked in any way to aid flows. The latter cannot be estimated in the framework of this paper, as aid flows are easily measurable unlike the policy advice that might accompany them.
Given this, what can be done to mitigate the negative impact of development aid and natural resource revenues on the quality of governance in less-developed countries? The study presents several options.
• Substantially reduce the share of wind-fall revenues going to central governments by bypassing them, for example through distribution of oil revenues and cash grants directly to people, increased delivery of aid through NGOs and private firms or allocation of a larger share of aid to public goods such as agricultural or anti-malarial research.
• Scale up donor efforts at technical assistance through revenue reform projects to encourage the expansion of efficient domestic revenue-raising capacity in the long run, as well as through investments to improve tax administration, encourage compliance and broaden the tax base.
• Support initiatives calling for disclosure of revenues and expenditures associated with natural resource rents and foreign aid. Examples include the Extractive Indus-tries Transparency Initiative at http://www.eitransparency.org, the Publish What You Pay coalition of civil society organizations at http://www.publishwhatyoupay.org and the International Budget Project at http://www. internationalbudget.org.
Stephen Knack. Forthcoming. "Sovereign Rents and Quality of Tax Policy and Administration." Journal of Comparative Economics. Also published as Policy Research Working Paper 4773. World Bank, Washington, DC. Vol. 3 (4) Summer 2009.