1
How does corruption influence the Foreign Direct Investment flows
from the European Union to Africa.
Bachelor Thesis
Author: Maud Visser, 10512624
Programme: Faculty of Economics and Business Supervisor: Alex Clymo
2 This document is written by Student Maud Visser who declares to take full responsibility for the con-tents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Abstract
Foreign Direct Investments (FDI’s) play an important role with respect to the development of a country. Corruption affects the choice of location and amount of FDI. The literature suggests an ambiguous relation between corruption and FDI, a positive relationship exists as a result of corruption being instrumental to overcome the inefficiencies due to bureaucracy and negative relations when corruption results in extra costs and creation of uncertainty. Africa and the EU have many FDI flows and highly different corruption levels. In this paper the way corruption affects the FDI flow from the EU to Africa is analysed. Overall EU countries invest more in less corrupt African countries. The effect of corruption is negative but insignificant. The negative relation by literature was explained mainly by the fact that EU has laws against corruption and has a significant lower corruption level than Africa.
4 Table of content
1. Introduction 5-7
2. Literature review
2.1 Introduction 7-8
2.2 Investigations that found a positive relation between FDI and corruption 8-10 2.3 Investigations that found a negative relation between FDI and Corruption 10-13
2.4 Contradictory results 13-16
2.4a Geographical differences
2.4b Differences in the incorporated variables 3. Data & Methodology
3.1 The relation between FDI flows and corruption in the EU and Africa 16-19 3.1a Corruption in the EU and Africa
3.1b Trade between EU and Africa
3.2 Main hypotheses 19
3.3 Method 19-22
3.3a Correlations between corruption and FDI 3.3b Regression with fixed effects
3.3c Robustness check
3.4 Data 22-23
4. Results
4.1 Correlations average corruption and average FDI 23-25 4.2 Correlation between average ranked corruption and average ranked FDI 25
4.3 Fixed effects 26
4.4 Robustness check 26-27
5. Discussion and Conclusion 27-28
5 1. Introduction
Foreign Direct Investments (FDI’s) are rapidly growing in Africa. In 2014 the global greenfield FDI market grew about one percent, while Africa received a 65 percent increase in capital investments (Fingar, 2015). As stated by the United Nations (UN, 2002): “Foreign direct investment contributes towards financing sustained economic growth over the long term’’. This rapid growth in Africa is an advantage for the further development of the continent.
FDI can fill at least three development gaps. These are the investing gap: FDI provides capital for investment, the tax revenue gap: FDI increases the tax revenues of the country by performing economic activities and the foreign exchange gap: FDI brings foreign currency in the country by investments and export earnings (Quazi, Vermuri & Soliman, 2010).
In 2008 the investment peak in Africa was 72 billion dollars (Quazi et al., 2010), after that it declined. Possible explanations are among others the global financial crisis, the drop-in commodity prices and the political instability in North Africa.
(Graph 1, source: EY, 2015)
As can be seen from graph 1 (EY, 2015), Western Europe is the main investor in FDI projects in Africa. An Africa attractiveness survey of EY (EY, 2015) shows that corruption is the second largest perceived barrier after an unstable political environment for investments in Africa. Though being the second largest perceived barrier, corruption can act in two ways. At the one hand it acts as an advantage for FDI, because it can reduce the inefficiency of bureaucracy and thus save time to get things done. But on the other hand it can act as a disadvantage by increasing the cost, uncertainty and raising ethical questions around doing business.
6 (Map 1, source: Transparency international[TI], 2012)
Map 1 (TI, 2012), shows an overview of the levels of corruption of the various countries; a score 10 (light yellow) implies that the country is perceived as “very clean” (no corruption) and a score 0 (dark red) implies that the country is perceived as “highly corrupt”. As can be seen from map 1 the African continent is mostly red coloured thus relatively highly corrupt. Even though the continent is facing high levels of corruption, the amount of FDI is high. Compared to Africa, Western Europe as main investor in Africa has a relatively low level of corruption.
Given the fact that the theoretical effects of corruption on FDI are ambiguous, empirical work is needed to assess what the sign of the effect is on how corruption affects the amount of FDI flows
from the European Union to Africa.
To have a better understanding, in this paper an investigation is carried out on the relation between FDI and corruption. Globally, about one third of the firms face corruption while doing business. Regarding Africa these percentages are even higher, in North Africa 53 percent and in sub Saharan Africa 38 percent of the firms deal with corruption while doing business (Molenaar & Kakebeeke, 2016). Moreover, in 2013 100 milliard dollars are invested in development worldwide, in the same year ten times this amount is lost in bribery, which is one of the expressions of corruption (De Boer, 2016).
Corruption is a widely spread and costly problem. It is interesting to investigate whether countries and companies in the EU, who are relatively transparent, still interact in corruption abroad and even more important do those companies consider corruption as an incentive to invest in a specific country, even though it is against the law. In 1997 all OECD countries plus few extra countries, together including 23 of the 28 EU countries (excluded are: Croatia, Cyprus, Lithuania, Malta and Romania), signed ‘’the convention on combating bribery of foreign public officials in international business transactions’’. This convention is the first and only anti-corruption convention which focuses on the supply side of the transaction. The convention requires the country to impose a fine on paying bribes abroad. It makes it even possible for countries to coordinate their efforts to restrict or even
7 end corruption (Organisation for Economic Co-operation and Development [OECD], 2011). And so does the geographical origin and destination of the FDI influence the result of the relation between FDI and Corruption?
The literature review gives a variety of answers, but the main conclusion is that uncertainties and extra costs are the most important reasons for not choosing a corrupt country for investments. Countries who have laws against corruption and are not corrupt themselves see corruption as a more important barrier than more corrupt countries, and some papers suggest that corrupt countries don’t even experience corruption as a barrier. Corruption can be positive for investors when it helps to overcome bureaucratic regulations and speed up processes. EU-countries do on average invest more in less corrupt African countries than in the more corrupt countries. But a statistical significant effect is not found in the present paper.
In this paper first the definitions of corruption and FDI are provided, followed by an overview of the literature about the relation between FDI and corruption. The literature and other important factors create insight in what the expected relation between FDI and corruption is. After that data and other important variables and literature about the effect of corruption in Africa on the inflow from EU countries is presented. A correlation and regression analyses on the data is performed to achieve a better understanding of the distribution and effect of FDI in Africa in relation to the corruption, followed by conclusions and discussion.
2. Literature Review
2.1 Introduction
Several investigations have been performed on how corruption affects Foreign Direct Investments. A few investigators found a positive effect, more investigators found a negative effect.
Foreign Direct investment (FDI), is an investment in another country than the investors country of residence, with at least a control of 10% in the business (United Nations conference on Trade and Development [UNCTAD], 2014).
Corruption can be defined as: the abuse of entrusted power for private gain (TI, 2015). Transparency International is an organisation that fights corruption. Corruption is roughly divided into three types:
● Grand corruption: crimes committed at a higher level within the government that distort the central function and policy of the country, where leaders can earn at the expense of the public interest.
8 ordinary citizens.
● Political corruption: abuse by political decision makers by using resources to keep their power, status and wealth.
Several researchers tried to explain the connection between corruption and FDI’s. The question is what the reason could be for the positive or negative relation and whether there is an empirically significant effect between corruption and FDI´s.
In this paper theories and results described in the various investigations will be considered. First investigations that show a positive effect of corruption are discussed, followed by papers that show negative effects of corruption on investments. Finally an investigation is included which discusses contradictory results.
2.2 Investigations in which positive relations are found between FDI and corruption
Already for a long-time investigator try to find explanations on if and how FDI benefit from corruption. Huntington (1968), Leff and Lui (as cited in Cuervo-Cazurra, 2006), were among the earliest that brought up theories on the advantages of corruption. The most mentioned argument is that corruption could facilitate transactions and speed up procedures that would encounter
otherwise a lot more difficulty and time if it would work out anyway. In this context corruption, could be seen as ‘’speed money’’. Some empirical investigations (partly) confirm the above theory, but the authors gave different explanations for the positive relation.
Egger and Winner (2005) argue that the degree of corruption correlates with determinants of the institutional environments which are difficult to measure. They investigated the short and long run effect. They found a positive relation for both the long and short term. The long-term impact of corruption on FDI seemed to be even more positive than the short term. Their explanation is that the positive effect is caused by the fact that investors are enabled to work around regulations and restrictions. In one of the specifications of the model they included the interaction term between corruption and legal quality, a positive outcome was estimated but the it wasn’t significant so they are careful with drawing conclusions.
Quazi et al. (2014), analysed the effect of FDI inflows to African countries and also found a positive correlation. Their explanation is in line with Egger and Winner, that accepting bribes and other forms of corruption makes it easier for companies to overcome the host countries complex bureaucratic regulations or to navigate between them, considering the host country as the country receiving the FDI. Moreover, some cynics even believe that adopting to local customs, that may involve corrupt practices, is the only way of accomplishing foreign operations in developing countries
9 by developed countries. They also assume that the positive relation is due to the presence of weak regulatory frameworks. As soon as the weak regulations are replaced by stronger ones over the years, the relationship will probably change because the support of the outcomes diminish and may follow a negative pattern. They suggest for further research on that topic to find evidence for their expectations.
Cuervo-Cazurra (2006) only found a partly positive relationship; in this paper the effect between the home (the origin country where the FDI’s come from) and host country corruption was investigated. His theory that businesses dealing with corruption in the home country are not
negatively influenced by host country corruption, was confirmed by his empirical research. Investors from corrupt home countries are used to pay bribes and experience this as just a regular way of doing business, they are familiar with how to deal with bribes. They even consider it as an advantage when the bribes they have to pay in the host country are lower than the ones in their own countries. Also expertise on how to work with bribes is valuable in such cases as there is no official information on this. Therefore it could even be experienced as an advantage over others when you have this expertise.
Even though the outcomes of the studies above have in common that they show a positive impact of corruption on investments, the explanations of the authors for this positive relation are different. They all agree that corruption serves to overcome obstacles and speed up processes. Few argue that the effect of corruption on investments is positive in Africa because of existing weak regulatory frameworks. Others argue that businesses that are used to do business with bribes, consider it as a regular way of doing business.
Autor(s) Countries included Variables used
Quazi Vermuri and Soliman (2014)
Host countries: Africa Home countries: Almost all countries covered (1995-2012)
Dependent: annual FDI inflow as % of GDP FDI inflow to an African country Independent: (1) FDI-1, (2) FDI-2, (3) Corruption, (4) Market Size, (5) Government effectiveness, (6) Infrastructure, (7) Economic openness, (8) Economic Freedom. Cuervo-Cazurra (2006) Home countries: Almost all
countries covered
Host countries: 106 (diverse)
Dependent: FDI inflows
Independent: (1)Host country corruption, (2)Home CC with laws, (3)Home CC with high corruption.
Control: (1) GDP, (2) Population, (3) Distance, (4) landlocked, (5) Island, (6) Common border, (7) Common language, (8) common colony, (9) Ever colonial link, (10) restrictions in trade, (11) restrictions on FDI.
Egger and Winner (2005)
Host: 73 ( mix of developed and less
developed countries (1995-1999)
Dependent variable: Inward FDI stocks
Explanatory variable: (1) Real GDP (2) Secondary school enrolment (3) Real GDP* secondary school enrolment (4) Legal quality (5) Corruption (short run) (6) Corruption (additional long run impact)
Other explanatory variables: (1) Inflation, (2) Trade Openness, (3) Political accountability, (3) External conflicts, (4) Internal Conflicts, (5) Bureaucratic quality. (6) legal quality, (7) Share of literate Adults, (8) Reduction in transboundary
10 emission pressure, (9) Environmental Governance, (10) environmental
sustainability, (11) Memberschip in international environmental agreements, (12) Protected area in percent of total area
Helmy (2013) Host: 21 MENA countries (2003 -2009)
Dependent: total FDi inflows of a country.
Explanatory variable: (1) CPI (2)Per capita income of the country
(3)Openness, (4) secondary school enrolment, (5) international homicide rate, (6) Fiscal freedom, (7) Investment freedom
Table 1 shows key variables of the investigations done, enabling comparison of the key variables with the negative outcomes, thus a positive relation between FDI and corruption.
As can be seen from the above table, three out of the four investigations above found a solely positive relation; those studies show a variation in regression analyses and with respect to included Countries and Variables. The host countries are respectively Africa, the Middle East & North Africa and a diverse sample of countries and as home countries all countries were included that invested in the host countries for both FDI Flows or stock. Thus they don’t make any distinction in home
countries, only Egger and Winner excluded several times one country group in one of their regressions. Also with respect to the papers that found a positive relation the variables used are based on solely the host countries. The corruption perception index from Transparency International is used as the main variable for corruption.
According to Quazi et al. (2014), China is heavily investing in infrastructure, which may imply that countries with a poorly developed infrastructure are probably more attractive for China than for other countries which could cause problems with respect to the sign of the variable when not discriminating within the home countries. They also include FDI-1 and FDI -2, the investments in the
year and two years before. This is considered as important because decision makers may not react instantaneously to changes and foreign investors are mostly risk averse. Furthermore the presence of FDI can influence the decision maker.
Cuervo-Cazurro(2006), was the only of the four investigators who discriminated within the home countries by controlling for home country corruption on existence of laws on corruption and concluded that laws on corruption switch the result from positive to negative.
2.3 Investigations that found a negative relation between FDI and Corruption
Contrary to the above outcomes, other authors found a (partly) negative effect of corruption on FDI. Most obvious reasons are that corruption increases costs of doing business and introduces
uncertainty on the reliability of the market procedures. Also international laws exist which try to decrease the rate of corruption which implies that doing business while dealing with corruption is considered illegal.
11 Qian and Sandoval-Hernandez (2016) studied the effect of corruption distance, which is the difference in corruption between home and host countries, on FDI flows. They found evidence that higher corruption distance has a negative influence on the decision to invest and also on the amount of money invested. A close level of corruption between host and home country seems to be the prominent factor for FDI. This implies that corrupt countries, for example African countries, attract more FDI from developing countries having little distance of corruption.
Wei (2000) considers corruption as an extra tax, which is negatively correlated with FDI. So in case a country would either increase their tax or corruption level, both levels would negatively affect FDI inflows. His analyses show results that confirm the idea that corruption increases the costs of doing business and because of that has a negative effect on FDI inflows.
Where Cuervo-Cazurra (2006) found a positive relation between FDI and host country corruption when the home country was corrupt as well, he found a negative relationship between FDI’s from non-corrupt countries and corruption of the host country. Countries who signed laws against bribery face a higher cost and risk of engaging in bribery abroad which discourages them further from doing business.
Apart from the host and home country corruption differences, Cuervo-Cazurra (2008) argued that it is not only the level of corruption but also the type of corruption that affects the level of incoming FDI. He points out that corruption affects FDI negatively in developed countries, because of the increased costs and uncertainty. On the other hand, in transition economies this effect was (partly) compensated by the benefits of bypassing inadequate regulations by accepting
corruption/paying bribes. He investigated this by dividing corruption in two categories: arbitrary (uncertain corruption) and pervasive (widely present corruption). Pervasive corruption negatively influences FDI in developed as well as in transition economies because of the extra costs. Arbitrary corruption is having significant a less negative influence on transition economies probably because operating in a transition economy already comes with a lot of uncertainty. This indicates that companies prefer to deal with ‘’an unknown evil’’ above higher costs.
The risk taken or the uncertainty, is also an argument that has been described in other investigations for example mentioned by Shleifer and Vishy (as cited in Cuervo-Cazurra, 2008). Paying bribes or making ‘under table agreements’ doesn’t mean that the person you paid the bribe or made the agreement with will adhere to the promise. Going to court in such a case is not possible because paying bribes is considered an illegal practice. Uncertainty about sticking to a promise is one thing but paying bribes also gives an incentive to create more bribes, which creates more uncertainty (De Soto, 1989).
All the above described papers mentioned that mostly the increased cost and increased uncertainty caused by corruption are the main deterrents why companies see corruption as a barrier
12 for investing in a specific country. Habib and Zurawicki (2001) mentioned that costs and uncertainty are not the only reasons. They provided apart from the economic explanation an ethical argument. They conclude that corruption is seen as morally wrong and immorality discourages companies from engaging in bribery. The more economic reasons Habib and Zurawicki (2001) provided are that companies may also try to avoid corruption because it is difficult to manage, risky and costly. Conclusively, corruption discourages FDI’s according to the above authors (summarized in table 2) by increasing costs, creating uncertainty and because of ethical considerations.
Author Countries included Variables Wei (2000) Home countries: 8 industrial countries
Host country: 33 diverse countries
Dependent: Bilateral stocks of FDI
independent: (1) Tax-rate, (2) corruption, (3) tax credit, (4) political stability, (5) GDP, (6) Population, (7) distance,
(8) linguistic tie, (9) OECD, (10) Wage Cuervo-Cazurra;
(2008)
+/- 90 diverse countries Dependent: Ln FDI inflows Log of FDI inflows into the country Independent: (1) Host country corruption, (2) Host country pervasive corruption, (3) Host country arbitrary corruption, (4) GDP, (5) Population, (6) GNI per capita, (7) FDI limitations, (8) Inflation, (9) Oil producer, (10) Transition economy, (11) Home country GDP, (12) Home country corruption, (13) Distance, (14) Landlocked, (15) Island, (16) Common border, (17) Common language, (18) Colonial link
Cuervo-Cazurra (2006) 183 home countries 106 host countries
Dependent: FDI inflows
Independent: (1) Host country corruption, (2) Home CC with (3) laws, (4) Home CC with high corruption
Control: (1) GDP, (2) Population, (3) Distance, (4) landlocked, (5) Island, (6) Common border, (7) Common language, (8) common colony, (9) Ever colonial link, (10) restrictions in trade, (11) restrictions on FDI. Habib Zurawacki
(2001)
111 countries (whole spectrum) 1994–1998
Dependent: Foreign direct investment (logFDI) for the years Independent: (1) Population, (2) GDPGRW, (3) GDP per capita, (4) Inflation, (5) Corruption (6) Openness (7) Political risk Habib and Zurawicki
(2002)
Large number of host countries offering a broad perspective on FDI and corruption (1996-1998)
home countries:: Germany, Italy, Japan, Korea, Spain, UK and the USA
Dependent: Bilateral FDI flows
independent: (1) Population. (2) GDP growth,
(3) log GDP/capita, (4) Unemployment, (5) trade/gdp, (6) Science & technology, (7) cultural distance, (8) Distance, (9) Economic ties, (10) CPI, (11) Political stability, (12) TI chapters, (13) abs difference in CPI Asiedu (2006) Cover 22 countries in SSA over the
period (1984-2000)
Dependent: 100*FDI/GDP
independent: (1) GDP, (2) Natural resources, (3) human capital, Literacy rate, (4) Inflation rate, (5) FDI openness, (6) Corruption, (7) Infrastructure, (8) effectiveness of the rule of law, (9) number of coups, (10) number of riots (11) number of assignations
Hossain (2016)
List of countries (1998-2014) Host countries: several countries from: South and South-East Asian, Latin America and the Caribbean and Africa
Dependent: Total FDI inflows a host country receives at time t divided by the host country’s total population (that is, FDI per capita
Independent: (1)Corruption, (2) GDP/pop, (3) Growth rate GDP, (4) POPG, (5) U PopG, (6) Trade openness
13 (7) Foreign companies in host country, (8) Risk, (9) SCH (2nd educ) & Literacy, (10) Inflation, (11) bureaucracy, (12) Democracy, (13) Law Teixeira & Guimarae
(2015)
96 countries 2000-2010 The indicator for FDI inflows (our dependent variable) is a dummy variable, which takes the value 1 when the average (2007–2010) ratio of FDI net inflow in GDP is above the mean and 0 otherwise.
Independent”: (1) Market Dimension, (2) GDP, (3) Geographic location, (4) Accessibility, (5) Natural resources, (6) Education/literacy, (7)R&D, (8) Labour costs, (9)Labour laws flexibility, (10)Quality of the work force, (11) Politics Incentives for FDI attraction, (12) Taxes, (13) Currency (exchange rates), (14) Political situation, (15)Corruption, (16)Bribe Payer’s Index, (17)Global Corruption Barometer, (18)Human Development Index, (19)Economic Freedom Index, (20) International Country Risk, (21)Guide indicator, (22)Business International Index
Quazi (2014) Host countries: Cambodia, China, Indonesia, South Korea, Laos, Malaysia, Philippines, Thailand, Vietnam, Bangladesh, Bhutan, India, Maladies, Nepal, Pakistan and Sri Lanka. (1995-2011)
Dependent: Annual FDI inflow
Independent: (1) Corruption, (2) FDI t-1, (3) economic freedom, (4) rate of return, (5) infrastructure, (6) reginal difference (7) market size, (8) human capital, (9) political stability
Qian, X., & Sandoval-Hernandez, J. (2016).
54 country pairs (18 industrial, 27 developing and transition economies) 1997-2007
Dependent: The Logarithm of FDI flow to a host country in current USD Independent: (1) corruption distance, (2) corruption, (3) Corruption distance, (4) real GDP, (5) Unemployment,
(6) distance, (7) Common language, (8) legal (common legal system), (9) R&D distance, (10) trade openness, (11) natural resources, (12) Political risk, (13) Number of days and procedures, (14) regulation cost, (15) Time trends
Table 2 shows key variables of the performed investigations that show negative relationships between corruption and FDI.
The 10 investigations in table 2 include a wide variety of home and host countries. Some focus more on the developing countries as host country and others took just a wide range of host countries. There are two other interesting differences in the above investigations. Some investigators specifically focussed on the relation between corruption and FDI, where others focussed on the drives for FDI and included corruption as just one of the variables. Five of them incorporated a form of home country corruption in their investigation. Also common languages and former investment are widely used variables.
2.4 Contradictory results
As mentioned earlier, contradictory results could be partly explained by the way the authors designed their investigation. But probably the main reason why contradictory results are found, is that the question what the effect of corruption is on FDI, just doesn’t have one specific outcome. In this paper further explanation of this will be divided into two parts: geographical differences and
14 differences in variables used, as both influence the outcomes. Furthermore the most informative investigation will be discussed separately.
2.4a Geographical differences
As already stated in the research of Cuervo-Cazurra (2006), differences are explained by variations in home country corruption and especially between the home countries who signed the 1997 OECD convention against bribery and the countries who did not. Already some investigators made a difference in geography by choosing one or a related group of host and home countries. Nwaogu and Ryan (2014) divided the regions based on different geographical characteristics and economic growth paths, and conclude that different regions may not attract the same type of FDI. They also stated that motives from home countries for investing differ across regions. For example the argument that China is investing in infrastructure which makes it a specific attractiveness parameter for them (Quazi et al., 2014).
The outcome that seems interesting is that the investigations on Africa and MENA countries have a positive outcome. Thus the underlying assumption of Quazi et al. (2014), that the positive relation between FDI and corruption in Africa, is existing because of the weak regulatory
environment in Africa is worth investing.
2.4b Differences in the incorporated variables
The two main variables used in the studies are corruption and FDI, but corruption is depending on many other variables, which could affect the results.
Corruption has a many-sided background and no unambiguous meaning and thus is measured in many different ways. Most indexes have a wide selection method, but it can be questioned whether it makes sense to measure corruption just in one way. There are differences in appearance, like bribery or money laundering, differences in how to face it with active/passive and differences in where it appears. Cuervo-Cazurra (2008) split up corruption into known and unknown corruption where he concluded that unknown corruption had less impact in already corrupt home countries than known corruption. Also the difference in active and passive corruption could be interesting to investigate. By active participating in corruption it is meant that you know what you are doing and pay to get your thing done, by passive participating, it is not that clear that you are involved but it is just a way to adjust to the system which is known to be not right but doesn’t feel that wrong as actively participating. But the many variations in how corruption appears could be
15 valuable as well. Some investigations already used a bribery index. But splitting up corruption in several measures could provide a more detailed and reliable outcome on the research. Apart from the way corruption is measured now and the index is used, it could be interesting to investigate whether some kind of a turning point can be identified where companies decide to not further accept corruption or in which cases they won’t even start with accepting corruption. It may be that companies don’t mind being involved in bribery/corruption until a certain extend. This is probably a level which is not traceable by legal authorities, until where the benefits outweigh the costs or in which case the amount of money paid for corruption doesn’t feel morally wrong. Knowing this ‘’turning point’’ would improve the value of the research by giving insight until what level corruption acts as a deterrent for FDI. Investigations on this matter are not available within the existing
literature and investigations in this direction could help to give a better insight and view in the ongoing debate.
Also the choice of control variables within the investigations may explain the differences in outcomes. Nwaogu and Ryan (2014) investigated the influence of the spatial interdependence on FDI’s. The so called ‘Third country effects’, the differences in the incorporated variables between neighbouring countries, are important. Leaving them out in the regression can result in biases, inconsistencies and inefficient parameter estimates (Anselin, as cited in Nwaogy and Ryan, 2014). Those third country effects could also be crucial for the measurement of the effects of corruption on FDI. When a country is corrupt but all the surrounding countries are more corrupt, it may receive more FDI’s because of a relatively favourable level of corruption.
The number of explanatory variables used is also different. In the three investigation with solely positive outcomes less explanatory variables, which were solely based on the host country, have been used than in the investigation of solely a negative outcome. Another one is the variables of colonial history and home country corruption. When looking for example at Africa, having a colonial history plays an important role in being the choice for FDI, leaving this out of the control variables could cause biases. Also home country corruption is not included in the papers that show positive results, while in the investigations with a negative outcome this appears to cause an important switch in results.
With respect to the above conclusions, the outcomes of the investigation by Cuervo-Cazurra (2006) seems one of the more informative ones by giving insight and recognition that the outcome depends on the way it is tested. Home country corruption seems to be important factor, which is fully in line with the underlying theory, especially the existence of laws in the home country. As countries with laws against corruption are mostly less corrupt, they face higher costs as well as a higher level of uncertainty by not having expertise in corruption. The investigations could be further improved by specifying geographical differences. For example the reason the PR of China invests in
16 Africa is different from why the EU invest in Africa, which could still cause inadequate answers. Quazi et al. (2014) can be considered as more relevant for this research on differences for EU and Africa, as they used parameters which are interesting for especially Africa.
3. Data and Methodology
This research will focus on evaluating existing literature and data on corruption on the basis of geographical differences. The literature will focus on the corruption and FDI in and between the EU and Africa and data will provide and overall insight of the de situation and the expected effect of corruption on FDI. Because one of the theories is that the choice of the home and host country do influence the results of investigations, two regions are specified in this investigation. As home country all the European Union countries together are defined; as host country all individual African countries. First the corruption in the EU and Africa will be further outlined, then the FDI relation between the two continents will be discussed, thereafter the research method and results will be presented.
3.1 The relation between FDI and corruption in the EU and Africa
3.1a Corruption in the EU and Africa
As seen from map 1, which was presented in the introduction, the overall corruption in the European Union is pretty low. Corruption appears everywhere thus also in the European Union. Many
initiatives are taken in the recent years to fight corruption. Apart from the anti-corruption measures within the EU laws, there are also some laws who fight against involving in corruption by European companies outside the EU. As already mentioned before, the OECD convention against bribery is one of the conventions which most EU-countries ratified. But there are more rules against corruption, for example The Council of Europe criminal law convention on corruption. This law acts as an instrument against the co-ordinated criminalisation for a broad scale of corruption practices (for example active and passive bribery in almost every sector, money-laundering, trading influence etc.). Countries are required to provide effective and dissuasive sanctions and measures. Legal entities are subject to eventual sanctions (Council of Europe portal, 2017).
As the perceived corruption map in the introduction shows, Africa faces high levels of corruption. Corruption practices mostly occur in secrecy and have many facets, which makes it
17 difficult to give an unambiguous measure of corruption. According to the United Nations Economic Commission for Africa (2016), weak governance institutions are one of the main causes and determinants of corruption. Those weak institutions are a threat for the prospect of positive outcomes from structural transformation processes. Also, Cross-border financial flows are seen as both a cause and a determinant of corruption.
The most common types of corruption practices in Africa are bribery, notably through concealment and tax evasion, and accounting irregularities (as stated by Burke and Cooper, in the African governance report, 2009). Collusion between suppliers and public officials within the international supply chain, bribery in international transactions, money laundering and customs tariff avoidance are the most common types of corruption in cross-border acts. Africa undertakes action against corruption by a signing a convention in which they made agreements on what the African countries should do to prevent and criminalise corruption (African Union, 2003).
Even though laws exist in the EU and in Africa, companies and governments still fail to use, install or impose them. Scandals still occur (TI, 2015). The widespread corruption in Africa can, per the African union, only be tackled by African leadership companies. But European countries and companies play an important role too. They must ensure that they not get involved and participate in corruption by for example paying bribes. The existing laws help to prevent this. It is even suggested that fighting corruption should be a central part of development strategies.
3.1b Trade between EU and Africa
Why do we or don’t we invest in Africa? As already mentioned unstable political environment and corruption are the major perceived barriers For FDI’s toward Africa. Other threats are weak security, poor basic infrastructure, lack of highly skilled labour, inconsistency and lack of transparency in regulatory policy, unattractive tax policies and financial incentives. As can be seen in table 3, natural resources and economic growth are the main factors for choosing Africa as FDI destination.
Technology, media and telecommunications (TMT), financial services, Consumer Products and Retail (CPR) and Real estate, Hospitality and Construction (RHC) are the sectors where most investments go to (EY, 2015).
18 (Table 3, attractiveness of African parameters, source: EY (2015))
Europe and Africa share a history of colonialism. Colonialism can lead to a common native language and a relatively familiar political and social environment. Research confirmed that there is a
significant and positive relation between colonial ties and bilateral FDI’s (Tong, 2005).
Some trade agreements exist between several groups of African countries and the EU. An old one, Lomé (called after and signed in Lomé, the capital of Togo) is signed in 1975. This agreement covers subjects such as security of the relations, non-reciprocal preferences, mutual interest and interdependence, respect for sovereignty and equality between partners. After this first step the agreement was extended in 1990 when subjects as human rights, good governance and democracy were included (European Union [EU], 2005).
Furthermore new agreements are still being negotiated. The past five years the EU and its former colonies, under which countries within Africa, are working on development negotiations and far-reaching trade agreements the so called ‘’Economic Partnership Agreements’’. Mostly these agreements improve the opportunities for trade by decreasing or stopping quotas or duties. These agreements also decrease uncertainty and improve sustainable development, which may improve the FDI flows towards Africa.
But the agreements are also criticised. Many NGO, policy makers and business
representatives argue that the agreements are too broad and importunate for Africa. Some even argue that such agreements could harm the African development (Draper, 2007).
This research only includes data until 2012, so new trade agreements don’t affect the outcome, but will be summarized in this paper, because these trade agreements should be incorporated in new research in the ongoing debate on corruption and FDI (European Commission, 2005).
19 East Africa, were signed. The grouped African countries mostly share few identical export goods. In 2016 the EU signed a trade agreement with the Southern African development community and still two major negotiations are ongoing which are between the EU and Central Africa, and the EU and East and Southern Africa.
In combination with the investigations in the literature review and the above mentioned studies a combination of valuable variables for the investigation of the relation between corruption and FDI is provided. Apart from the variables already mentioned in table 1 and 2, The following variables should be considered when trying to find out the relation between FDI and corruption in the EU and Africa: Natural resources, economic growth, domestic markets, large low cost labour force, demographics and political and social environment. Those parameters influence decision making for FDI in Africa (EY, 2015). Also some barriers, which are not already mentioned, such as weak security, poor basic infrastructure, unattractive tax policies and financial incentives could be included (EY, 2015).
Also trade agreements, laws against corruption/home country corruption and colonial history, which influence language and familiar political and social environment should be considered to be included, except when they are applicable on every country in the regression.
3.2 Main hypothesis
The hypothesis to be tested is: corruption negatively influences FDI inflow in Africa from The EU, where b1 is the corruption coefficient.
h0: b1=0 h1: b1<0
3.3 Method
First four correlation tables are composed. Those will give more insight in the distribution of EU FDI’s in Africa with respect to the corruption in Africa. Second a regression between FDI and corruption with respect to fixed effects is performed. Lastly a robustness check by a regression with control variables is performed. All the analyses are performed in Excel.
20
3.3a Correlations between corruption and FDI
When the correlation between the averages of FDI and Corruption is analysed, the average
corruption is based on the available data on corruption between 2001 and 2012. The average FDI’s inflows are taken from the available data on FDI between 2001 and 2012.
Average corruption Average FDI
Mean 2.92 314.62 Standard Error 0.13 120.67 Median 2.79 18.21 Standard Deviation 0.96 870.19 Sample Variance 0.93 757239.25 Range 4.69 4835.92 Minimum 1.21 -165.50 Maximum 5.90 4670.42 Sum 152.05 16360.05 Count 52.00 52.00
(Table 4, descriptive statistics on the total average corruption and total average FDI)
In table 4 the descriptive statistics are shown. Because of the outliers in differences in FDI inflows, as seen in this table, it is difficult to find clear results. Therefore FDI inflows are split into an average FDI inflow from above |100| and an average inflow beneath |100| (in millions of US dollars). The descriptive statistics are displayed in Table 5 and 6.
Average
Corruption Average FDI
Average
Corruption Average FDI
Mean 3.04 799.78 Mean 2.85 11.39
Standard Error 0.18 285.37 Standard Error 0.18 4.99
Median 2.90 265.16 Median 2.65 4.29 Standard Deviation 0.83 1276.19 Standard Deviation 1.04 28.24 Sample Variance 0.68 ######## Sample Variance 1.09 797.43 Range 2.90 4835.92 Range 4.69 140.42 Minimum 1.98 -165.50 Minimum 1.21 -55.17 Maximum 4.88 4670.42 Maximum 5.90 85.25 Sum 60.85 15995.53 Sum 91.20 364.52 Count 20.00 20.00 Count 32.00 32.00
(Table 5, descriptive statistics on the total average corruption (>100) (Table 6, descriptive statistics on the total average and total average FDI) corruption (<100) and total average FDI)
Also a correlation between ranked average FDI and ranked average corruption is calculated. The FDI inflows are ranked from 1-52, whereby 1 is lowest FDI receiver and number 52 receives the most FDI’s. The Corruption is ranked from 1-52, whereby 1 is most corrupt and 52 is least corrupt. This way the ranked popularity and corruption distribution is shown and can be compared.
21
3.3b Regression with fixed effects
A regression analysis is performed between the FDI inflow and the corruption index. Fixed effects are incorporated in the regression. Fixed effects regression is used as a method for controlling for missing variables in panel data where the omitted variables vary across entities (states) but do not change over time. So in this case the fixed country effects could consist of factors such as size of the country. Incomplete FDI data are left out, and for the few incomplete corruption indexes the first follow up score was used to fill the gaps. Only 15 countries are tested in the regression, all the countries who didn’t had at least 10 years of FDI date are left out and after that the countries with the highest average FDI inflow are used. The following regression is performed:
FDI Inflowst,c = b0 + b1 Corruption Perception Index t,c + b2 C +……+b15 C
where C is the dummy variable for every specific African Country, and the little t and c are
respectively year and country. Including every dummy variable leads to biases; one of the dummies was excluded, which is Uganda in this regression. So Uganda will be the reference dummy that is shown in the intercept.
This regression provides more insight in till what extend corruption affects FDI, while
controlling for country fixed effect. Probably the regression turns out to be not significant because of the control variables that are not included in the country fixed effect, but a better insight of the distribution is showed. With the R-square and adjusted R square it can be seen how much effect they may have on each other and till what extend FDI relies on all the other variables.
3.3c Robustness check
Another regression analyses with control variables is performed by controlling for specific factors which are important for FDI between the EU and Africa. Because of the fact that only Africa and EU are used, variables such as trade agreements and home country corruption/laws against corruption are less important as the countries incorporated are rather comparable with respect to these factors and therefore these factors are left out. The following regression is preformed:
FDIt,c=b0 + b1CPI t,c + b2 FDI-1t,c + b3 ECONOMIC OPENNESSt,c+ b4 ECONOMIC GROWTHt,c+b5 MARKET SIZEt,c + b6 INFRASTRUCTUREt,c + b7 NATURAL RESOURSES + b8 INFLATIONt,c+ b9 TAXt,c
The same countries as in the above regression with fixed effects are used, with a time range from 2003 until 2012. Unknown data is covered the same way as done with respect to corruption by using
22 the first following up result; in case this wasn’t available the last known result is used. Congo is missing data for infrastructure and Liberia is missing data for natural resources for the whole-time period. Thus those two countries where excluded in the regression. Due to multicollinearity between infrastructure and corruption (correlation coefficient above 0.7), the control variable infrastructure is removed. Now Congo is put back in the regression again and no multicollinearity came up1. Because
every African country used has a colonial history with one or more EU countries, the control variable colonial history is not included in the regression, which in the end resulted in the final regression:
FDIt,c=b0 + b1CPI t,c + b2 FDI-1t,c + b3 ECONOMIC OPENNESSt,c+ b4 ECONOMIC GROWTHt,c+b5 MARKET SIZEt,c + b6 NATURAL RESOURSES+ b7 INFLATIONt,c+ b8 TAXt,c
Table 7 represents the descriptive statistics for the performed regression.
FDI inflow Perceived corruption FDI -1 Economic Openness Economic
Growth Market Size
Natural
resources Inflation Tax Mean 1244.32 3.10 1062.82 78.58 5.39 3439.93 37.37 9.89 46.78529 Standard Error 173.61 0.08 145.94 2.55 0.38 235.56 3.44 1.15 1.108142 Median 378 2.90 338.50 71.30 5.15 2659.33 12.72 7.97 44.95 Standard Deviation 2024,8 0.97 1701.98 29.76 4.40 2747.11 40.14 13.42 12.92305 Sample Variance 4099327.35 0.94 2896724.71 885.93 19.32 7546625.72 1611.55 180.03 167.0051 Range 10738 4.30 9807 121.40 37.75 9630.94 98.24 124.45 52.7 Minimum -1104 1.40 -1104 35.46 -4.01 399.69 0.00 -20.63 24.2 Maximum 9634 5.70 8703 156.86 33.74 10030.63 98.24 103.82 76.9 Sum 169228 422.20 144544 10687.11 732.97 467830.60 5082.12 1344.68 6362.8 Count 136 136 136 136 136 136 136 136 136
(Table 7, descriptive statistics)
3.4 Data
In this paragraph the variables incorporated in the descriptive statistics in Table 4,5 and 6 are described.
Foreign direct investment (FDI): The data on FDI is obtained from the United Nations conference on
trade and development (2015) and the FDI flows are on a net basis. The UNCTAD (2015) defines FDI as follows. When an investor makes an investment outside his home economy with the perspective to acquire lasting interest in the enterprise, this refers to an FDI. Further the ‘direct investor’ should
1 For correlation table see appendix table 5
23 have an effective voice in the management of the enterprise which is associated with at least 10% of the equity ownership. This implicates that the single foreign investor owns at least 10 percent of the (ordinary) shares or voting power of the enterprise. Exceptions are cases in which it can be proven that 10 percent ownership does not give the investor an effective voice in the management. The voting power does make the differences between an FDI and a foreign portfolio investment. After identifying the direct investment enterprise, the capital flow between the entity and the enterprise operating outside the home economy of the investor should be defined to know which one can be classified as FDI. Only the capital invested directly by the investor or though other enterprises related to the investor, are classified as FDI. Examples are equity capital, the provision of short-term and long-term intra-company loans (between parent and affiliate enterprises) and the reinvestment of earning.
The Corruption Perception Index (the CPI) from Transparency International is used as a measure for corruption. The CPI is a combination of 12 data sources which all rank corruption. The reason that it is called a perception index is due to the difficulty of empirical analyses while not having access to all the data. For making up the data, scandals, investigation or prosecutions are used. Countries are ranked with a score 1-10, where 1 means a high corruption score and 10 means no corruption. (TI, 2015)
Control variables included are the following: FDI-1: The FDI flow of the previous year (Quazi, 2014) is used. Natural resources, these are incorporated because they are an incentive for investing in Africa (EY, 2015) and are measured in Fuel exports as % of merchandise exports (Teixeira & Guimarães, 2015). Economic growth is the growth rate of GDP (Hossain). Inflation is the annual inflation rate (Cuervo-cazurra, 2008). Infrastructure is defined as the quality of port infrastructure (Quazi, 2014). Tax policies: are measured as the total tax rate (EY, 2015). Market size is GDP per capita (Quazi, 2014). Economic openness is reflected by the share of total volume of trade (export plus import) in GDP (Quazi, 2014). All data is obtained from the Word Development Indicators (World bank, 2017) apart from FDI -1, which is obtained from UNTAC.
4.0 Results
4.1 Correlation tables average corruption and average FDI
Graph 2 shows the distribution of the EU FDI’s in Africa with respect to their perceived corruption; the higher the corruption index, the less corrupt the country is.
24 The slope is 0.2142 which shows that the correlation is positive, which implies a negative relation
between FDI and Corruption. Graph 2 shows that on average the less corrupt countries receive more FDI from the EU. This gives more insight of the distribution on EU FDI in Africa, but with a significance of 0.127 no relation and especially no causal relation is defined.
Graph 3 & 4 show the distribution of the countries with respectively a yearly average under $100.000.000 dollar or a yearly average above $100.000.000 dollar.
2 See appendix table 1 for regression output
0 1 2 3 4 5 6 7 -1000 0 1000 2000 3000 4000 5000 Aver age FD I in Mili ons of US$
Graph 2: Average FDI and Average Corruption
0 1 2 3 4 5 6 7 -80 -60 -40 -20 0 20 40 60 80 100
Average Corruption, (1=very corrupt and 10=very transparant)
Aver age FD I in Mili ons of US$
Graph 3:Average FDI (<100) and Average
Corruption
25 In the graphs the correlation statistic is respectively 0.2198 and 0.38633, again a positive relation,
which means a negative relation between the distribution of FDI in relation to Corruption.
4.2 Correlation between average ranked corruption and average ranked FDI
Graph 5 shows the correlation between average ranked corruption and the ranked average FDI, with number 1 as lowest FDI inflow and 52 highest FDI inflow. It can be seen from this graph that the distribution is kind of randomly divided, both high as well as low corrupt countries do receive a variety of FDI inflows.
(Graph 5, X-as Average corruption, Y-as average ranked FDI)
3 See appendix table 2 And 3 for regression output
0 1 2 3 4 5 6 -1000 0 1000 2000 3000 4000 5000
Average Corruption, (1=very corrupt and 10=very transparant)
Aver age FD I in Mili ons of US$
Graph 4:Average FDI(>100) and Average
Corruption
0 10 20 30 40 50 60 0 10 20 30 40 50 60Average ranked Corruption (1=Most Corrupt country and 52 is least corrupt country) Aver age ranked FD I, 1 r eceiv ed least FD I and 52 recei ve d m ost FD I)
Graph 5: average corruption & ranked average
FDI
26 4.3 Fixed effects
As all of the above correlations don’t show any significant results, a regression with fixed effects is performed where time fixed effects and country fixed effects are included. For the regression with fixed effect the following regression is performed:
FDIt,c=b0 + b1CPI t,c + b2 ALGERIA + b3 ANGOLA + b4 REPUBLIK OF CONGO +b5 EGYPT + b6 GABON + b7 KENYA + b8 LIBERIA + b9 MADAGASCAR+ b10 MAURITIUS + b11 MOROCCO + b12 NIGERIA + b13 SOUTH AFRICA+ b14
TANZANIA + b15 TUNESIA
Where CPI is the corruption perception index and b2 – b15 are the dummy variables for each country
except for Uganda. T stands for time and C is for the 16 different countries.
The outcomes in table 8 show a positive effect which suggests that the more transparent/less corrupt the country the more FDI it receives, but the effect is not significant.
The R square and the adjusted R square, which are 0,55 and 0,51, suggest that a little over 50% of the FDI flows can be declared by fixed country effects and corruption.
Regression Statistics Multiple R 0.745262 R Square 0.555415 Adjusted R Square 0.513473 Standard Error 1295.368 Observations 175
Coefficients Standard Error t Stat P-value Intercept 165.1186 788.3797255 0.209440422 0.834372547 Corruption 31.43357 280.4258993 0.112092248 0.91089165 (Table 8, excel regression output, regression with fixed effects)4
4.4 Robustness check
A can be seen in table 9 the perceived corruption is positive, which is in line with the hypotheses but the p-value is 0.094 (>0.05) thus not significant. It cannot be concluded based on this investigation that a relation between corruption and FDI exist because the H0 hypothesis cannot be rejected. The
signs for economic openness, economic growth and market size are negative, which is deviating from the literature. But all three have a P-value above 0.05 and are thus not significant. FDI-1 and natural resources are positive and significant and tax and inflation are negative and not significant.
4 See appendix table 4 for complete regression output
27 SUMMARY OUTPUT Regression Statistics Multiple R 0.819 R Square 0.671 Adjusted R Square 0.650 Standard Error 1197.760 Observations 136.000
Coefficients Standard Error t Stat P-value Intercept 584.742 731.372 0.800 0.425 Percieved corruption 338.010 200.300 1.688 0.094 FDI -1 0.898 0.067 13.364 0.000 Economic Openness -7.797 4.323 -1.804 0.074 Economic Growth -26.490 25.961 -1.020 0.309 Market Size -0.077 0.066 -1.161 0.248 natural resources 10.469 4.569 2.291 0.024 Inflation -3.671 8.286 -0.443 0.659 Tax -14.510 9.933 -1.461 0.147
(table 9, output excel, regression with control variables)5
5.0 Conclusion & Discussion
The main conclusion from the literature in this paper is that there is not just one answer to the question on what and how corruption affects FDI’s; corruption shows positive as well as negative relations with FDI’s. There are theoretical reasons and proves for both perspectives.
Uncertainty, extra cost and ethics are the reasons of a negative influence, the faster and easier processes when accepting corruption are the main reasons for a positive relationship.
When considering how corruption affects FDI flows from EU to Africa, which was the main question in this research, it showed that though they have a kind of contrary corruption levels, Western Europe is still the main investor in Africa. Probably the weak institutions in Africa lead to positive effects of corruption on FDI as where the low corruption and laws in the EU as home country seem to be the most important negative factor for FDI in relation to the corruption.
The correlation tables between (ranked) average FDI’s and the (ranked) average corruption, gave an overview of the distribution of FDI in African countries based on the corruption level. A positive slope was found, which implicates a negative relation. From this correlation it can be seen that in general European Union companies invest more in less corrupt countries, but no statistical significant relation is found. The regression with fixed effects is insignificant and only shows that
28 corruption and fixed country effects determine FDI for a little over 50 percent. The regression with control variables, still shows a positive but insignificant effect between perceived corruption and FDI. Lack of corresponding outcomes in the literature about the overall question how corruption and FDI affect each other and thus also within the specific research on EU in relation to Africa, are probably mainly due to the difficulty of assessing corruption. It is hard to measure and no real marks exist. Because of the lack of availability and transparency in the information, we will be never sure about the values and the reliability of the values. Splitting corruption in various sub variables, for example bribery and non-existence of agreements, may provide a better insight and probably a more measurable thus useful result. In this case situations can occur where bribery has a negative effect, but corruption in another way positively affects FDI. And also the way how corruption appears would affect the outcome.
The lack of significance of the results in this investigation can have several reasons. The FDI data were incomplete which caused problems with the outcome. Nowadays more and more is registered in data bases which makes the incomplete data less of a trouble in future investigations. Also for the regression with fixed effects the use of only corruption indexes, FDI data and fixed effect could cause some trouble, even though a lot of explanatory variables are partly compensated by the choice of home and host countries. They probably play a too important role, that leaving them out of the regression could cause some biases. The regression with control variables, has ambiguous
outcomes with can be caused by missing variables, a too small range of host countries and the lack of other control variables. The possibility of reverse causality, which is not tested here, could also affect both of the above results.
Thus the overall conclusion of this paper is that no significant relation is found between FDI and corruption, but more insight is gained in the differences and complications which come up during studying the relation of corruption and FDI. Choices of host and home countries and explanatory variables all give contradictory results, because relationships are very specific and an overall conclusion probably doesn’t exist.
Furthermore it is valuable to investigate the effect of corruption on FDI with respect to geographical differences. Using countries with the same regulatory system and comparable
corruption and host countries with similar attractiveness to FDI could also be helpful in determining the right variables for the specific comparisons. Also the existence of trade agreements between countries are interesting to incorporate in further investigations. Those could ensure more certainty and help to decrease the negative effects of corruption. Further suggestion is to investigate the existence and height of an eventual ‘’turning point’’, at what level of corruption does it become (more) negative.
29 References
African Union. (2003). African Union convention on Preventing and combating corruption. Retrieved from: http://www.eods.eu/library/AU_Convention%20on%20Combating%20
Corruption_2003_EN.pdf
Asiedu, E. (2006). Foreign direct investment in Africa: The role of natural resources, market size, government policy, institutions and political instability. The World Economy, 29(1), 63-77. De Boer, M. (2013, 13 December). Ontwikkelingslanden zien bijna $1000mrd aan crimineel geld wegstromen. Het financieel dagblad. Accessed at https://fd.nl/
Cuervo-Cazurra, A. (2006). Who cares about corruption?. Journal of International Business
Studies, 37(6), 807-822.
Cuervo-Cazurra, A. (2008). Better the devil you don't know: Types of corruption and FDI in transition economies. Journal of International Management, 14(1), 12-27.
Council of Europe portal, Details of Treaty No. 173. (2017). Criminal law convention on corruption. Retrieved from: https://www.coe.int/en/web/conventions/full-list/-/conventions/treaty/173 Last accessed 12/01/2017.
Draper, P. (2007). EU-Africa trade relations: the political economy of Economic Partnership
Agreements. European Centre for International Political Economy.
Egger, P., & Winner, H. (2005). Evidence on corruption as an incentive for foreign direct investment. European journal of political economy, 21(4), 932-952.
European Union, The cotonou Agreement. (2005). From Lomé I to IV bis. Retrieved from: http://ec.europa.eu/development/body/cotonou/lome_history_en.htm Last accessed 14/01/2017.
European Commission (2005). Trade policies. Retrieved from:
http://ec.europa.eu/trade/policy/countries-and-regions/regions/. Last accessed 13/01/2017. EY, Ey’s attractiveness survey Africa 2015 making choices. (2015). Report of EY attractiveness survey
Africa Making choices. Retrieved from http://www.ey.com/Publication/vwLUAssets/EY-
africa-attractiveness-survey-2015-making-choices/%24FILE/EY-africa-attractiveness-survey- 2015-making-choices.pdf
Fingar, C. (2015, 19 may). Foreign direct investment in Africa surges. The financial times. Accessed at https://www.ft.com/
Habib, M., & Zurawicki, L. (2001). Country-level investments and the effect of corruption—some empirical evidence. International Business Review, 10(6), 687-700.
Habib, M., & Zurawicki, L. (2002). Corruption and foreign direct investment. Journal of international
business studies, 33(2), 291-307.
Helmy, H. E. (2013). The impact of corruption on FDI: is MENA an exception?. International Review of
30 How taxing is corruption on international investors?. Review of economics and
statistics, 82(1), 1-11.
Hossain, S. (2016). Foreign direct investment (FDI) and corruption: Is it a major hindrance for encouraging inward FDI?. African Journal of Business Management, 10(10), 256-269. Huntington, S. P. (2006). Political order in changing societies. Yale University Press.
Leff, N. H. (1964). Economic development through bureaucratic corruption. American behavioral
scientist, 8(3), 8-14.
Molenaar, C., & Kakebeeke, P. (2016, 10 may). 1 op de 3 bedrijven ervaart corruptie. Het financieel
dagblad. Accessed at https://fd.nl/
Nwaogu, U. G., & Ryan, M. (2014). Spatial Interdependence in US Outward FDI into Africa, Latin America and the Caribbean. The World Economy, 37(9), 1267-1289.
OECD. (2011). Convention on combating bribery of foreign public official in international business transactions and Related Documents. Retrieved from:
http://www.oecd.org/corruption/oecdantibriberyconvention.htm
Qian, X., & Sandoval-Hernandez, J. (2016). Corruption Distance and Foreign Direct Investment. Emerging Markets Finance and Trade, 52(2), 400-419.
Quazi, R. M. (2014). Corruption and Foreign Direct Investment in East Asia and South Asia: An Econometric Study. International Journal of Economics and Financial Issues, 4(2), 231. Quazi, R., Vemuri, V., & Soliman, M. (2014). Impact of Corruption on Foreign Direct Investment in Africa. International Business Research, 7(4), 1.
Shleifer, A., & Vishny, R. W. (1994). Politicians and firms. The Quarterly Journal of Economics, 995-
1025.
Teixeira, A. A., & Guimarães, L. (2015). Corruption and FDI: Does the Use of Distinct Proxies for Corruption Matter?. Journal of African Business, 16(1-2), 159-179.
Transparency International. (2015), Corruption perception index [Data file]. Retrieved from http://www.transparency.org/ Last accessed 29/12/2016.
Transparency international. (2015), What is corruption [definitions]. Retrieved from http://www.transparency.org/what-is-corruption/. Last accessed 13/01/2017.
UNCTAD, United Nations Conference on Trade and Development. (2015). Bilateral FDI Statistics [data file]. Retrieved from: http://unctad.org/en/Pages/DIAE/FDI%20Statistics/FDI-Statistics- Bilateral.aspx. Last accessed 23/12/2016.
United Nations. (March, 2002). Report of the international conference on financing for development.
Retrieved from: http://www.un.org/esa/ffd/monterrey/MonterreyConsensus.pdf
United Nations economic commission for Africa, African governance report IV (2016). Measuring
31 http://www.uneca.org/sites/default/files/PublicationFiles/agr4_eng_fin_web.pdf
Wei, S. J. (2000). How taxing is corruption on international investors?. Review of economics and
statistics, 82(1), 1-11.
World Bank, 2017. World Development Indicators [data file]. Retrieved from: http://www.worldbank.org/data/. Last accessed 24/1/2017.
32 Appendix Appendix table 1 SUMMARY OUTPUT Regression Statistics Multiple R 0.214314 R Square 0.045931 Adjusted R Square 0.026849 Standard Error 858.4334 Observations 52 ANOVA df SS MS F Significance F Regression 1 1773807 1773806.667 2.407094115 0.127094 Residual 50 36845395 736907.8988 Total 51 38619202 Coefficients Standard
Error t Stat P-value Lower 95%
Upper 95% Lower 95.0% Upper 95.0% Intercept -251.965 384.1004 -0.655987174 0.514839863 -1023.45 519.5234 -1023.45 519.5234 X Variable 1 193.7617 124.8882 1.551481265 0.12709367 -57.0836 444.6071 -57.0836 444.6071
(Appendix table 1, Regression output of the correlation between Perceived corruption and FDI)
Appendix table 2 SUMMARY OUTPUT
Regression Statistics Multiple R 0.219843385 R Square 0.048331114 Adjusted R Square 0.016608818 Standard Error 28.00335655 Observations 32 ANOVA df SS MS F Significance F Regression 1 1194.765 1194.765 1.523569 0.226662 Residual 30 23525.64 784.188 Total 31 24720.4
Coefficients Standard Error t Stat P-value Lower 95%
Upper
95% Lower 95.0% Upper 95.0% Intercept -5.553229279 14.59303 -0.38054 0.706226 -35.3562 24.24972 -35.3562 24.24972 Average Corruption 5.945247097 4.81658 1.234329 0.226662 -3.89152 15.78202 -3.89152 15.78202
33 Appendix table 3 SUMMARY OUTPUT Regression Statistics Multiple R 0.386304 R Square 0.149231 Adjusted R Square 0.099186 Standard Error 1234.151 Observations 19 ANOVA df SS MS F Significance F Regression 1 4541848 4541848 2.98192 0.102324 Residual 17 25893189 1523129 Total 18 30435037
Coefficients Standard Error t Stat P-value Lower 95%
Upper 95% Lower 95.0% Upper 95.0% Intercept -969.628 1083.512 -0.89489 0.38334 -3255.64 1316.382 -3255.64 1316.382 3.625 599.6336 347.2466 1.726824 0.102324 -132.993 1332.26 -132.993 1332.26