Table of Contents
2. Literature Review
3. The Research Design
4. Limitations and Concluding Remarks
5. Publication bibliography
7. Statutory Declaration
More than 60 years of development assistance in Sub-Saharan Africa seem to have yielded little success. Africa today, still viewed by many as a backward continent, lacks economic diversification and development. It is home to more than 400 million people living in extreme poverty, or an income below 1.90 Dollar a day (Olinto, Uematsu 2017). With an absolute poverty rate of 43 percent, Sub-Saharan Africa maintains the global lead, accounting for roughly a third of the world’s poor (Beegle et al. 2016). Despite receiving the most aid out of any world region, macroeconomic impacts fail to materialize (OECD 2017). Western public opinion often points the finger at a so-called corrupt nature of African people, and African leaders in particular. The argument is that aid and other sources of income are used to line their corrupt pockets, rather than to progressively developing their country. For example, the current President of the United States, Donald Trump, stated that “[e]very penny” that goes to Africa “will be stolen” (Trump 2013) and is basically wasted. However, looking at estimated corruptions numbers, one could argue that these allegations are not just mere xenophobia. Only four Sub-Saharan African countries score over the Corruption Perception Index (CPI) midpoint of 50, while the vast majority lies in the bottom third of the global ranking (Transparency International 2017a). Concerning Africa’s poverty situation and its developmental state, it appears that aid effectiveness has been severely hindered by the occurrence of corruption on the continent. However, one strain of literature goes even further, suggesting that aid is actually a contributing cause for corruption (Alesina, Weder 2002; Asongu 2012; Brautigam, Knack 2004; Djankov et al. 2008; Knack 2001; Rajan, Subramanian 2007; Svensson 2000). The increased monetary inflow enables the recipient government to rent seek these funds. On the other hand, another strain of literature claims the contrary, arguing that aid will be used to improve institutions, and therefore accountability capabilities, while also closing the income gap for public employees, and thus limiting small- scale corruption (Charron 2011; Dalgaard, Olsson 2008; Dunning 2004; Mohamed et al. 2015; Okada, Samreth 2012; Tavares 2003; van Rijckeghem, Weder 2001). Both sides confidently support their views with empirical studies.
In this paper, we will explore the question of whether foreign development aid has an enhancive or reductive effect on corruption levels, and then conclude which of the theoretical camps is correct in the specific context of Sub-Saharan Africa. First, we will discuss the relationship between aid and corruption by reviewing the most relevant studies in literature, both on a general level and in African context. In addition to empirical findings, theoretical mechanisms for each strand shall be revealed. Next, we will establish our own quantitative breakdown, based on the newest available data. After the definition of important terms and the presentation of our created indices, we will run a Pearson correlation analysis using 48 Sub-Saharan African countries with data from 2002 to 2014. Furthermore, we will test the findings for robustness, using a multivariate ordinary least squares (OLS) regression. Results will be interpreted and applied to our discussion. We will deliberate potential shortcomings, before the final section will summarize and conclude.
2. Literature Review
In order to make sense of the following review of studies on the aid-corruption nexus, we must first find comprehensive and practical definitions for these central terms. Development aid, sometimes referred to as international or foreign aid, entails the international transfer of goods (such as food or military equipment), capital (financial resources such as grants or concessional credits), or services (such as technical advice or training) from one country or international organization to the benefit of another country (Williams 2015). Official Development Assistance (ODA) represents the most typical form of foreign aid, and will be the primary form of aid referenced in this paper. The Organization for Economic Cooperation and Development (OECD) defines ODA as “government aid designed to promote the economic development and welfare of developing countries” and excludes any military loans and credits (OECD 2018). Aid is delivered via two methods: bilaterally, between a donor and recipient country, or multilaterally, between a development agency (for example the United Nations, World Bank, or other non-governmental organization) and recipient country. Furthermore, only aid delivered to states on the OECD’s developing countries list is defined as ODA (OECD 2018).
Leading corruption authority Transparency International (2017b) defines corruption in general terms first, as “the abuse of entrusted power for private gain." Then, more specifically, it distinguishes between various types, identifiable by the government sector involved and total monetary loss. Grand corruption occurs at high levels of government, whereby leaders benefit despite the public's detriment, usually by manipulating government policies or harming central political functioning. Petty corruption involves low- and mid- level government abuse of power in everyday public domains. It occurs between ordinary citizens and public officials, typically in the access of public goods or services such as in hospitals, schools, and police departments (Transparency International 2017b). Tavares (2003) argues that “corruption is likely to arise in situations where resources are transferred with substantial discretion without accountability to the decision maker [...]. Foreign aid is, by definition, such an instance: it provides goods or finance at below market prices to governments or population groups. Aid disbursements are typically handed free to local authorities that then distribute them, with considerable discretion, among their fellow citizens (Tavares 2003, p. 100).”
Considering that the vast majority of aid is distributed through channels of government budget, aid appears to be highly vulnerable to corruption, especially on the grand scale (Alesina, Dollar 2000).
There exist an abundance of studies examining the effect of aid on corruption. Although, often the nexus is simply a secondary, partial test included in the measure of aid effectiveness on quality of governance institutions (e.g. Svensson 2000; Knack 2001; Brautigam, Knack 2004; Rajan, Subramanian 2007; Djankov et al. 2008; Dalgaard, Olsson 2008; Asongu 2013). This is because there is a consensus in literature that poor quality of governance and institutions is linked to higher levels of corruption, as weak institutions lack checks and balances and facilitate low degrees of government accountability (Mohamed et al. 2015). For our purposes we hence assume that if foreign aid is to reduce corruption, it is also to improve governance and institutional quality, and vice versa.
As one of the forerunners in the discussion of the aid-corruption relationship, Svensson (2000) introduced a game-theoretic rent seeking model supported by empirical evidence from a panel study of 66 countries in the period of 1980 to 1994. He found that in fragmented environments with competition between social groups (measured through ethnolinguistic fractionalization), a higher influx of foreign aid is connected to higher levels of corruption due to increased governmental revenues and incentives for rent seeking activity. In this case, even the mere prospect of aid may raise rent dissipation and inefficient public spending. When concerning Africa, a highly ethnically diverse region, Svensson’s findings would therefore link aid to an increase in corruption.
Knack (2001) analyzed the effect of aid on the quality of governance, indicated by bureaucratic quality, rule of law, and corruption. He examined 80 countries, covering between 1975 to 1995. His empirical findings disagree with Svensson, as his study shows that with greater ethnic homogeneity, aid worsens the quality of governance. However, both agree that aid is rather a rent seeking tool than a promoter of capabilities for qualitative institutional progression. Knack’s findings display that higher levels of aid are linked to a decrease in quality of governance. He claims that aid gives elites the ability to bypass important institutional reforms and keep inefficient economic policies in place, consequently weakening government accountability for its actions. Additionally, without much coordination of aid allocation, conflicts over aid funds can arise, which, as argued by Svensson (2000), may nurture corruption levels. Knack noticed that these detrimental factors to governance quality are not only distinctive of financial flows but also non-financial aid such as technical assistance.
Rajan and Subramanian (2007) support these views in their study of aid effect on governance quality. They argue that a country’s industrial and manufacturing sector is an indicator for good governance, as the complex transactions of manufacturing depend on reliable institutions. Manufacturing is hence a symbol of rule of law, and the ability to enforce contracts and control corruption. Their evidence reports that an increase in one standard deviation of aid lowers industrial growth by magnifying 2.8 percent points. Rajan and Subramanian (2007) infer that aid has a damaging effect on manufacturing and hence governance institutions, including corruption control. As with Knack (2001), effects are also true for mere technical assistance and non-financial aid.
By looking at data from a cross section of 28 Sub-Saharan African countries in the period of 1982 to 1997, Brautigam and Knack (2004) studied the influence of aid on the quality of governance in a specific African context. In concordance with the previously discussed studies, they also found robust statistical proof of aid weakening institutional quality, hence backing the idea of aid fostering corruption. Additionally, they found aid as a reducer of a country’s tax share of GDP. They provide theoretical considerations to make sense of their findings. Aid flows diminish taxation efforts and incentivize fiscal indiscipline, as government revenues do not depend on tax revenues. Therefore, funds are also alleviated from the long-term development of such tax administrations. Brautigam and Knack also believe patronage and corrupt rent seeking behavior to be major contributors to their findings.
Similarly, Djankov et al. (2008) found damaging effects of aid on institutional quality due to rent seeking activity in a 108 country panel study, covering 1960-1999. Interestingly, they argue that aid resembles a “curse” for the recipient country, just as resources do:
“Natural resources and foreign aid share a common characteristic: they can be appropriated by corrupt politicians without having to resort to unpopular, and normally less profitable, measures like taxation44 (Djankov et al. 2008, p. 171).
Comparing the severity of the consequences, the authors found that aid related rent seeking worsens governance institutions even more than rents from natural resources (in their case, oil).
It seems there is little incentive for political elites to deviate from aid rent seeking behavior when it entails no significant repercussions or sanctions for them. In their 63 country study, covering 1981-1995, Alesina and Weder (2002) found that countries with higher corruption levels actually tend to receive more aid inflow. Corrupt governments are not discriminated against, as factors like colonial history, political alliance, and other economic conditions play a larger role in aid allocations.
Tavares (2003) criticizes Alesina and Weder’s (2002) results as flawed by a reversed causation in their regression. The size of the coefficient is biased because, even if aid would lead to less corruption, more aid is given to more corrupt states. Tavares then conducts his own analysis which controls for factors like oil resources, political rights, colonial history, cultural and geographic proximity to the donor, and ethnic fractionalization. His findings robustly demonstrate aid as a reducer of corruption. An increase of aid of one percent of GDP roughly lowers corruption by 0.2 (on a scale ranging 0 to 10). Tavares provides two explanative mechanisms. First, in the conditionality effect, donors may attach the inflow of aid to certain rules and conditions of good governance, limiting recipient governments’ corrupt behavior. In the second mechanism, the liquidity effect, aid helps overcome public revenue shortages and improve public official salaries. This liquidity effect in turn reduces the incentives for small scale corruption. Van Rijckeghem and Weder (2001) also argue in agreement with Tavares. They deduce empirical evidence for the negative relationship between fair civil service wages and corruption. Bribes and solicitation are reduced, as aid is invested in training and wages of crucial public officials, such as police, judges, and tax collectors.
Some literature does not evaluate the aid-corruption relationship as a strictly black or white issue, but finds that it can be negative or positive depending on certain conditions. For instance, Dalgaard and Olsson (2008) use a large N cross-sectional analysis to study the effects of rents of natural resources and aid on institutional quality, for example, with rule of law and corruption. They conclude that while natural resources like oil and minerals enhance corruption, aid can decrease it. However, the effect is strongly non-linear, meaning it is valid for lower but not higher influx of aid.
Dunning (2004) argues that there is a firm distinction in aid effectiveness between the Cold War and post Cold War time periods. During the Cold War, aid allocation was strongly influenced by geostrategic interests, diminishing what Tavares (2003) called the conditionality effect, weakening the credibility of threats of conditioning. Those threats to withdraw aid if conditions were not fulfilled became more credible in the post Cold War period, improving aid effectiveness. Whether aid decreases or increases corruption may thus depend on specific geopolitical circumstances.
Likewise, Charron (2011) argues for more nuanced distinctions in aid effectiveness in his study. Expanding on Dunning's (2004) results, he claims that post Cold War improvements are based on the Anti-Corruption Movement (ACM) beginning in the 1990s. The movement was the agenda focus on corruption for many major international organizations. The ACM included: the 1996 signing of the “Combating Bribery of Foreign Public Officials in International Business Transactions” convention by all countries in the Organization for Economic Co-operation and Development (OECD), the cooperation of the World Bank and Transparency International beginning in 1997, the launch of the International Monetary Fund (IMF) anti-corruption campaign of 1997, the World Trade Organization (WTO) joining the Working Group on Transparency in Government Procurement in 1996, and the creation of the United Nations (UN) Management Development and Governance Division in 1995 (which dealt with government accountability and transparency). Additionally, the European Union and the United States followed this precedent and engaged in corruption combating campaigns by the turn of the century. Charron (2011) argues that ratifying ACM norms displays that multilateral aid from international organizations have lesser ties to geostrategic and political agendas. By obligating recipient governments to act on efforts to combat corruption, multilateral aid from international organizations is thus more effective in the production of good governance. Charron’s study, with panel data from 82 countries between 1986 and 2006, hence disaggregates aid into bilateral and multilateral aid and a pre- versus post ACM time period, namely the year 1997. The results show that multilateral aid after 1997 reduces levels of corruption. Conversely, multilateral aid before 1997 does not present a clear relationship. Bilateral aid, on the other hand, was either statistically insignificant or even had a positive effect on corruption.
Another interesting study applying post Cold War data is that of Okada and Samreth (2012). Their study further nuanced the aid-corruption link, applying a sample of 120 countries between 1995 and 2009. They first differentiated between multilateral and bilateral aid, and then furthermore, bilateral aid from four major donors. Additionally, by applying a Quantile Regression method (QR), they identified the nexus at different stages of corruption levels and not solely at its mean. Their estimations indicate that aid in general has a decreasing influence on corruption. Effects were most pronounced when multilateral organizations allocate aid to countries who were already low on the corruption interval to begin with. This is because those countries have relatively improved institutional governance quality and better capabilities to control corruption. Influx of aid can then be distributed more efficiently. Bilateral aid from donors with colonial histories like the United States of America (USA), United Kingdom (UK), and France appeared to be rather insignificant, while aid from Japan had corruption reducing effects.