Corruption & DBS: Corruption and effect of DBS

Corruption moderates effect of budget support on growth and poverty

Corruption affects which aid modality is more effective in promoting growth and alleviating poverty. This section reviews the available studies and evaluations bearing on this issue. The focus is on the relative impact of general budget support versus that of project support, though sector budget support is discussed in the final subsection.

In summary, available studies indicate that:

• Level of corruption matters: Budget support not advisable in highly corrupt countries
  
• Type of corruption may matter: Political corruption may make budget support unsuitable
  
• Distribution of corruption matters: Targeting aid towards islands of integrity

Fiduciary risk assessments (link to definition) in relation to budget support, should therefore include all these elements in partner countries where corruption may be a problem (here's how).

Level of corruption matters: Budget support not advisable in highly corrupt countries

Aid promotes growth where institutions are good. This is the result of a study by World Bank economists Burnside and Dollar. Though their conclusion has not gone unchallenged, it remains a much cited one in the development debate. The Burnside and Dollar study does not, however, estimate the impact on growth of different aid modalities.

Budget support works better where institutions are good, project support where institutions are bad. In a study by two IMF economists, the analysis of Burnside and Dollar is rerun, with total aid being divided into budget and project support. The result is that budget support has a greater impact on growth in countries with good institutions, whereas project support is more beneficial for growth in countries with bad institutions. According to this study, whether to use budget or project support is a question of the institutional environment. The measure of institutions used, however, is a macroeconomic one, and includes inflation, budget surplus, and openness. The results therefore do not directly tell us which aid modality works best in a corrupt versus non-corrupt environment.

However, corrupt countries have less productive and less pro-poor public spending. Several studies show that corrupt countries spend less on education and health, and more on the military. As human capital formation is conducive to growth, whereas military spending may be detrimental thereto, this may partly explain why corrupt countries have lower rates of growth. Similarly, spending on education and health is directly relevant for poverty reduction, and public resources therefore have less of an impact on poverty in corrupt countries. The implication for budget support is that if aid is less than perfectly fungible, budget support has a lower impact on growth and poverty reduction in corrupt countries.

Type of corruption may matter: Political corruption may make budget support unsuitable

Petty corruption decreases the effectiveness of any type of aid. Where petty corruption is widespread, it functions as a tax on a large range of activities, including activities financed by aid. Though budget and project support fund different activities, it is hard to say whether there is some systematic difference between the activities funded, which makes one form of aid more susceptible to petty corruption. Without additional information, is it therefore hard to assess whether petty corruption favours budget or project support.

Political corruption may be more detrimental to budget support, than to project support. If the people with the power of the public purse are corrupt, topping up that purse will lead to funds being diverted into private hands. And budget support amounts to topping up the public purse. With project support, funds are tied to a specific activity, and cannot always be as directly appropriated by a corrupt political elite. This does not mean that projects are immune to corruption in this sense, but there is at least a potential for safeguarding funds when providing project support.

That political corruption is adverse to budget support, is also suggested by USAID evaluations of budget support to Tanzania, Mozambique and Malawi. While the evaluations argue that budget support to Tanzania and Mozambique has worked reasonably well, budget support to Malawi has not. Since Malawi has approximately the same score as the other two countries on the Transparency International Corruption Perceptions Index, corruption as such does not seem to be the reason for the different effect of budget support.

The Malawi evaluation suggests, however, that a high degree of political corruption is one cause of the bad performance of budget support in Malawi. More evidence is needed to determine whether this is in fact what separates Malawi from Tanzania and Mozambique in this respect. Nevertheless, that political corruption has an adverse effect on budget support, can at least be used as a working hypothesis.

Distribution of corruption matters: Targeting aid towards islands of integrity

In some cases, there may be relatively "clean" areas in an otherwise corrupt set of public institutions. Different ministries, departments, and public institutions, face different incentives for corrupt activities.

By the nature of their activities, some ministries are more subject to public scrutiny than others. For instance, health or education may entail more transparent activities than defence. Some ministries and institutions also serve politically potent stakeholders, and would suffer greater consequences of misconduct (tapping the wages of powerful union members is one example). When incentives differ across the political-bureaucratic system, and the system has certain structural properties, different parts of the system can develop quite different views and practices in terms of corruption.

There is thus a possibility that islands of integrity may develop in an otherwise corrupt institutional setting.

Where islands of integrity exist, sector budget support can be channelled towards them. To use sector budget support in this way, one must map the distribution or variation of corruption across the political system. Again, this requires that funds are less than perfectly fungible, so that they cannot be completely appropriated by other parts of the political system through reallocation. The idea of targeting sector budget support in this way, is related to the more general notion that if certain parts of the public financial management system of a partner country function, aid should be channelled through them.

A word of caution should be expressed on islands of integrity and sector budget support. Firstly, studies of corruption in hierarchies show that where top officials are corrupt, lower level officials are also corrupt. This suggests that if the top ranks of government are corrupt, clean areas in the other and lower parts of government are unlikely to be found. In certain cases, looking for islands of integrity would therefore be a waste of time and resources. Moreover, even if islands of integrity do exist, one should be careful to act in a way that supports their existence. Budget support alters the incentives of public officials, potentially increasing the benefits of corruption, and may therefore alter the behaviour of supposedly clean segments of the political system.