Does Conditionality Still Work? China’s Development Assistance and Democracy in Africa - by CWP Alumni Xiaojun Li

Sunday, Jul 9, 2017
by dsuchens

Previous studies have found that the democratizing effect of conditional aid is temporally contingent: the collapse of the Soviet Union as an alternative source of aid enhanced the effectiveness of Western aid conditionality with respect to democratic reforms being adopted in Africa. Does conditionality still work with China’s rise as a major donor since the early 2000s? This article examines this question by leveraging the first Forum on China–Africa Cooperation (FOCAC) as the temporal dividing point and new measures of Chinese aid to Africa, based on expert opinions and media reports. The results show that the democratizing effects of the OECD’s development aid in Sub-Saharan Africa have indeed diminished. Furthermore, results from a synthetic control analysis suggest that major recipients of Chinese economic assistance did not achieve higher levels of political freedom than other comparable countries in the post-FOCAC period. These findings support the thesis that the democratizing effect of aid conditionality works only during a period when recipient countries do not have other alternative sources of aid, allowing donors to more credibly commit to enforcing conditionality.

Proponents of foreign aid argue that one mechanism by which aid can contribute to democratization in recipient countries is through conditionality—i.e., the exercise of leverage by donors who attach conditions of democratic governance to the disbursement of aid (Robinson 1993; Stokke 2013; Stone 2008). Conditionality does not always work, however, when donors have other, competing, priorities (Bearce and Tirone 2010). This was the case during the Cold War when the United States and other donors gave aid to “strengthen corrupt but geopolitically useful autocracies” (Brautigam and Knack 2004, 275). When the Cold War ended, Western donors could refocus on using conditional aid to push for democratic changes. The temporal effect of aid conditionality is supported by empirical evidence; scholars have shown that the relationship between conditional aid and democratic transition in Africa and beyond is contingent upon the historical context (Bermeo 2016; Djankov et al. 2008; Dunning 2004; Kelley 2008; Kersting and Kilby 2014; Knack 2004; Wright 2009).

China’s rise as a major donor to Africa (and developing countries in general) may pose new challenges to the effectiveness of aid conditionality in the new millennium (Qian 2015). Since the first Ministerial Conference of the Forum on China–Africa Cooperation (FOCAC) in 2000, the amount of Chinese development assistance going into Africa, including aid and finance, has been increasing exponentially (Thaler 2012).1 Besides the increased amount, China’s assistance is often considered more attractive by recipient countries because it has few or no political strings attached and is often disbursed much more quickly and efficiently than assistance from Western nations. Consequently, while traditional donors have criticized China’s approach to aid, many African countries embrace the assistance from Beijing, or at least are glad to have more options.

Does conditionality still work with this change in the aid landscape? Building upon earlier work on aid conditionality and the credible commitment problem in donor–recipient relationships, this article argues that the increased availability of Chinese aid will encourage African recipients to resist pressure from Western donors to improve democratic governance. In the meantime, traditional donors may also feel compelled to compete with the new donors, reducing their credibility in enforcing the conditions. Consequently, the positive relationship between Western aid and democracy among the African recipients will dissipate as the amount of Chinese aid increases.

Few studies have tested the effect of Chinese aid cross-nationally due to the paucity of reliable data, relying instead on case studies (Chen and Kinzelbach 2015; Hackenesch 2015). This article circumvents this problem using the first Ministerial Conference of FOCAC as the temporal dividing point, as well as new measures of Chinese aid to Africa, based on expert opinions and media reports. Empirical analysis shows that the democratizing effects of OECD development aid in Sub-Saharan Africa have indeed diminished in the twenty-first century. Furthermore, results from a synthetic control analysis suggest that the major recipients of Chinese economic assistance have registered lower levels of political freedom than other comparable countries in the post-FOCAC period.

This article makes several contributions to the literature on donor intent in aid giving and international democratization. The empirical findings support the thesis that aid conditionality works only during a period when recipient countries do not have alternative sources of aid, allowing donors to more credibly commit to enforcing aid conditionality. The temporal divergence in the effects of aid conditionality also suggests that future studies should consider disaggregating the post-Cold War period when examining the aid–democratization relationship.

The remainder of the article consists of five sections. Section 2 briefly reviews the literature on aid conditionality, focusing on recent scholarship that demonstrates that the effect of aid conditionality is temporally contingent. Section 3 lays out the main argument that the availability and influx of China’s “no-strings-attached” aid to the African continent in the 2000s dampened the efforts of Western donors to make aid conditional upon democratic governance. Section 4 discusses the research design and data used to test the hypotheses derived in the previous section. Section 5 presents the main empirical results. The last section concludes with some discussion on the policy implications of the findings.

2 Donor Intent, Aid Conditionality, and Democracy in Recipient Countries

A large body of literature has explored the relationship between foreign aid and democracy in recipient countries, leading to two contrasting views on the effectiveness of aid in promoting democracy. Skeptics maintain that aid inhibits democratization because aid is similar to other non-taxable incomes, such as oil and remittances. Consequently, it leads to the same problem of the resource curse (see, e.g., Ahmed 2012; Bueno de Mesquita and Smith 2010; Morrison 2009; Smith 2008). Supporters of foreign aid dismiss this pessimistic view by pointing out that aid donors can play a major role in influencing the outcomes of aid. They argue that one mechanism through which aid can contribute to democratization in recipient countries is through conditionality—i.e., the exercise of leverage by donors who attach conditions of democratic governance to the disbursement of aid (Robinson 1993; Stokke 2013; Stone 2008).

Several explanations can be given for conditionality being effective in inducing reform and policy changes. First, conditionality can solve the principal agent problem between the donor and the recipient when a mismatch exists between their preferred policies (Williamson 1983). Second, conditionality can help recipient governments commit to a policy when they face domestic resistance for its implementation (World Bank 2005). Finally, recipient governments may want to use conditionality as a signaling device to demonstrate to potential investors the credibility and predictability of the policy environment and to distinguish themselves from non-reformers (Marchesi and Thomas 1999). In other words, conditionality should prevent the misappropriation of money if the recipient wishes to remain dependent on future transfers (Dreher 2009).

Starting in the early 1990s, donor countries increasingly conditioned their aid with democratic governance, which they viewed as being indispensable for economic growth. The US Agency for International Development (USAID), for instance, spent $637 million in 1999 on democracy assistance, with over one-fifth going to Sub-Saharan Africa (Carothers 1999). The World Bank and other multilateral donors also committed substantial resources to programs aimed at strengthening accountability and the rule of law in recipient countries (Stone 2008; World Bank 2005). In contrast to donors’ enthusiastic support for aid conditionality, the empirical evidence seems to be mixed (Crawford 1997; Goldsmith 2001; Heckelman 2010; Kilby 2009; Knack 2004; Öhler et al. 2012).2

Some scholars suggest that the poor results of conditionality may be attributed to structural impediments in the recipient countries (Brown 2005). But the donors are not inculpable either. When there are competing priorities, donors may find it difficult to credibly commit to enforcing the conditions. In a widely cited paper, Thad Dunning argued that the key impediment to using aid for democracy promotion during the Cold War years was the lack of credibility of the United States and its allies’ threats to remove aid when they needed to compete with the Soviet Union for influence and clients (Dunning 2004, 411). Recognizing the geostrategic dilemma faced by Western donors who continued to provide aid to recalcitrant African states to prevent losing clients to the Soviet Union, leaders of recipient countries were able to resist aid conditionality and obtain economic assistance from the West without undergoing domestic reforms. The collapse of the Soviet Union, however, renewed the ability of the West, the only remaining donors, to withhold development assistance if the recipient states failed to carry out democratization reforms. Aid conditionality thus became more effective after the end of the Cold War.

A number of authors have echoed this line of reasoning in exploring the aid–democratization relationship in Africa and beyond. Brautigam and Knack (2004, 275), for example, noted that “the end of the Cold War allows the United States and other donors to target aid more selectively, rather than using aid to strengthen corrupt but geopolitically useful autocracies.” Similarly, Kelley (2008, 229) argued that the end of the Cold War “freed Western countries to push for democratic changes [as] democracy increasingly came to be seen as strengthening rather than undermining security interests.” Kersting and Kilby (2014) concurred that aid may stymie moves toward democracy when donors “cannot credibly commit to enforcing aid conditions due to overriding geopolitical priorities.”

The shift in the politics of aid after the Cold War has received empirical support as well. For instance, Dunning (2004) demonstrated that the democratizing effect of foreign aid in the Sub-Saharan African countries was present only during the post-Cold War period. Similarly, using a dataset that included US foreign assistance program information for 165 countries from 1990 to 2003, Finkel et al. (2007) showed that democracy assistance had a significant impact on democracy building in the recipient countries. Most recently, Bermeo (2016) found that the negative relationship between aid and the likelihood of democratic change was limited to the Cold War period for a group of 129 developing countries from 1973 to 2010.

In sum, the extant theoretical and empirical works on the aid–democratization nexus suggest that aid conditionality is temporally contingent. Aid did not become effective in improving democracy in the recipient countries until after the end of the Cold War, when donors could more credibly enforce conditionality. In the following section, I will argue that the rise of China as a major donor in Africa marks another important temporal shift in the effectiveness of aid conditionality.

3 China’s Development Assistance and Democracy in Africa

China’s engagement with Africa dates back nearly 60 years. The first instance of China’s aid to Africa occurred in November 1956, when Beijing provided Egypt with 20 million Swiss Francs in cash. In 1960, China helped Guinea build its match and cigarette plants, which was China’s first development project in Sub-Saharan Africa. By the mid-1980s, China’s assistance had helped pave the way to diplomatic recognition with over 40 African countries. An oft-cited example of Chinese aid to Africa during this period was the Tanzania–Zambia Railway, completed in 1975 under the auspices of a zero-interest loan of RMB 980 million from China.3

Not until the turn of the twenty-first century, however, did China formulate its official policy on aid to Africa, intensifying its relationship with the entire continent. On October 10, 2000, the first Ministerial Conference of the FOCAC was held in Beijing. The FOCAC is an official forum between China and the African states to “further strengthen the friendly cooperation between China and Africa under the new circumstances, jointly meet the challenge of economic globalization and promote common development (FOCAC 2013).” Ministers from China and over 40 African countries, including the presidents of Algeria, Tanzania, Togo, and Zambia, attended the conference, which adopted the Beijing Declaration of FOCAC and the Program for China–Africa Cooperation in Economic and Social Development.

China’s aid and development finance to Africa has grown exponentially since the conclusion of the first FOCAC and has been regarded by many as “the most significant development on the continent since the end of the Cold War” (Taylor 2012). Following the first FOCAC conference, China reduced or cancelled RMB 10.9 billion worth of debts for 31 heavily indebted or least-developed countries in Africa. In subsequent FOCAC conferences, held every three years, China doled out larger and more ambitious assistance programs to Africa. During the 2006 FOCAC ministerial conference, for instance, then-President Hu Jintao pledged to double the amount of aid by 2009, provide $5 billion in preferential loans in the next 3 years, and establish a $5 billion China–Africa Development Fund (Zafar 2007). By 2012, China had provided more than $10 billion worth of “concessional loans” to Africa (Chen 2012). This amount was doubled to a commitment of $20 billion from 2013 to 2015, which was announced by President Xi Jinping during his first trip to the continent in March 2013 (People’s Daily 2013).

China’s rise as a major donor in Africa once again makes aid conditionality less effective for three additional reasons. First and foremost, Chinese aid requires no commitment from the recipient countries to undergo reforms in governance.4 China’s official discourse claims that its aid is granted in adherence to the principles of “mutual respect of sovereignty and territory integrity,” “no interference in internal affairs,” and “treating each other equally and safeguarding common interests (People’s Daily 2011)” Chinese scholars at the Africa–China–USA Trilateral Dialogue in 2005 opined that since “there is no consensus on a definition of good governance, China does not pre-condition its assistance on the existence of democracy and places more emphasis on a balance among reform, stability, and development.” They further commented that “the international community should not push too hard for democracy” and “should be more patient about letting [the African countries themselves] do the job (Council on Foreign Relations 2007).”

China’s policy of non-interference and a “no-strings-attached” approach to aid is appealing to African leaders. In a 2009 survey of African officials, ranking from junior military officers to former presidents (from countries such as Angola, Mozambique, South Africa, Namibia, Cape Verde, and Zambia), 63 of the 67 interviewees expressed positive views about China’s aid in Africa (Horta 2013). During a recent visit to Guyana by a Chinese delegation led by Zhang Gaoli, then Party Secretary of the Tianjin Municipal Committee,5 Raphael Trotman, speaker of the Guyana National Assembly, made the following comment that nicely sums up the attitude of African elites toward Chinese aid:

We believe that China has been a very good friend from afar and what is unique about China is that its involvement in Guyana has never been one that sought to interfere with our internal political structure. Other countries give aid with conditions—whether they be on governance, on trafficking in persons or a raft of legislation on social issues. China has never given with conditionalities coming with them, and that is something we appreciate (Guyana Times International 2012).

Second, Chinese aid is usually disbursed very quickly and efficiently. A study by the Economic Strategy Institute credits the “sheer competence and speed with which China is able to negotiate and execute its development programs” as “an important element of its appeal” (Olson and Prestowitz 2011). In addition, the existing multilateral aid platform has become increasingly dysfunctional and uncoordinated, leading to wide frustration among recipient countries, which have had to deal with the highly bureaucratic and burdensome systems used for aid delivery by the traditional donors (Woods 2008). This is perhaps best summarized by the following comment made by Abdoulaye Wade, President of Senegal:

I have found that a contract that would take five years to discuss, negotiate and sign with the World Bank takes three months when we have dealt with Chinese authorities. I am a firm believer in good governance and the rule of law. But when bureaucracy and senseless red tape impede our ability to act—and when poverty persists while international functionaries drag their feet—African leaders have an obligation to opt for swifter solutions. I achieved more in my one hour meeting with President Hu Jintao in an executive suite at my hotel in Berlin during the recent G8 meeting in Heiligendamm than I did during the entire, orchestrated meeting of world leaders at the summit (Wade 2008).

Finally, skeptical of and even disillusioned by the Western recipe for economic development inscribed in the Washington consensus, some countries now look to China’s model as a successful example to emulate (Chen and Kinzelbach 2015; Peerenboom 2008). Even though both policymakers and scholars have emphasized that the China model may need to be adapted to fit local contexts, a number of African countries have embraced the “Beijing consensus” (Ramo 2004). For instance, a number of Special Economic Zones, which were key components of China’s reform and opening up policies, as well as engines of fast economic growth, have already been replicated in countries such as Mauritius, Nigeria, Tanzania, and Zambia, usually accompanied by loans and investments from China (Davies 2008).

To summarize, African countries have come to realize that they need options, not conditions, when it comes to aid. China’s aid and development financing provides such options and is thus welcomed with open arms by many African countries.6 Nevertheless, this may be bad news for traditional donors who have thus far been successful in tying aid to democratic reforms and other conditions. Indeed, Chinese aid projects have come out victorious in competitions with the World Bank on a number of occasions (Naim 2007). This causal argument suggests that the availability of Chinese aid would encourage African recipients to resist pressure from Western donors to improve democratic governance. In the meantime, traditional donors may also feel compelled to compete with the new donors, reducing their credibility in enforcing the conditions. This leads to the following two hypotheses:

Hypothesis 1: As China’s aid and development financing to Africa expands, the positive relationship between Western aid and the level of democracy in the African recipients will dissipate, all else being equal.

Hypothesis 2: The more aid a country receives from China, the less likely it is that its level of democracy will increase, all else being equal.

4 Research Design and Data

I tested the two hypotheses regarding the effect of China’s aid to Africa based on Dunning’s (2004) research design, which was used to explore the impact of aid on democracy in a pooled cross-sectional time-series model of Sub-Saharan Africa between 1975 and 1997. The dependent variable is the Freedom House index of political freedom, measured on a seven-point scale in half-point increments, with higher numbers indicating greater political freedom. Aid, the key independent variable, is the ratio of the OECD’s official development assistance (ODA) to the gross national product (GNP) in the recipient country. Additional controls in the model include the gross domestic product (GDP) per capita, a dummy variable for whether or not a country has a legal tradition based on English common law, and a measure of racial, linguistic, and religious fractionalization. In addition, the model includes a dummy variable indicating whether or not the Soviet Union considered a Sub-Saharan African state as a “revolutionary democracy” or “socialist-oriented” during the 1970s and 1980s. To account for potential endogeneity, aid is instrumented with population size and a dummy variable for French colonies.7 Furthermore, to account for the fact that the positive effects of aid on democracy are temporally contingent, Dunning (2004) divided the sample observations into two periods (1975–1986 and 1987–1997) and estimated the same model for each period.

Since I am interested in whether or not the rise of China as a major donor in Africa dampens the effect of Western aid on democracy, I make a couple of modifications to Dunning’s model, in accordance with the two hypotheses above. First, I extend the dataset to 2008 and split the post-Cold War sample into two periods, 1987–2000 and 2001–2008, using 2001 as the dividing point, after the first Ministerial Conference of the FOCAC. A potential concern is that 2001 may also have been the year in which the West, particularly the United States, began curtailing its aid programs in Africa, with more focus on national security priorities in other regions, following the 11 September 2001 terrorist attack. In other words, the diminishing effect of Western aid on democracy could be due to the downsizing of ODA rather than the new availability of Chinese aid. Nevertheless, this is unlikely, as development aid is usually committed years ahead and disbursed in multiple installments over time. Recent research also suggests that since the early 2000s, the European Union and United States have made support for democratic reforms one of the top priorities in their aid packages to African recipients (Hackenesch 2015).

This concern can be further dispelled by the data. Figure 1 plots the mean ratios of ODA to GNP for all 48 Sub-Saharan African countries from 1975 to 2008. We can see that aid from traditional donors to Africa increased steadily for the first two decades of this period, from 8.08% in 1975 to 19.92% in 1994. Aid dropped sharply in the next five years but then climbed again, starting in 2000. In fact, compared to the amount in 2001, the West has given more development aid in every subsequent year except 2006.

Fig. 1

OECD ODA as a percentage of GNP, 1975–2008. The series is the average over all countries.

Source: OECD Statistics and WDI

The second change I make is to include a measure of the amount of Chinese aid received by each country, which would allow us to test Hypothesis 2. Unfortunately, China provides no official statistics on its aid disbursement, and some debate still persists as to what actually counts as aid.8 Many scholars have attempted to come up with reasonable estimates of China’s aid to Africa by combining multiple sources, such as the Ministry of Commerce (MOFCOM), the China Export–Import Bank (Eximbank), and various statistical yearbooks (e.g., Brautigam 20082009; Lum et al. 2009). Nevertheless, there are still more questions than answers. To get around the lack of aid data for individual countries, this article takes a different approach by constructing a measure based on expert opinions. In November 2010, I conducted a small survey with the staff and officials working in relevant departments in the MOFCOM and Eximbank.9 In the questionnaire, each respondent was asked to first write down two African countries that, to the best of their knowledge, received the most and least amount of aid from China between 2001 and 2008. They were then asked to place each of the 48 countries in Sub-Saharan Africa on a seven-point scale based on the amount of Chinese aid they had obtained in reference to those two other countries.10 A total of 13 valid questionnaires were collected. As a condition of obtaining their responses, I constructed an aggregate measure of Chinese aid by averaging the scores for each country and designating as major recipient the countries that had an average score above six. This yielded a list of 11 countries: Angola, Benin, Botswana, Cameroon, the Democratic Republic of Congo, Ethiopia, Liberia, Nigeria, Sudan, Zambia, and Zimbabwe. Not surprisingly, most of these countries are frequently in headline news reports on China’s involvement in Africa.

To examine the external validity of this variable constructed from expert opinions, I looked at a number of measures on cultural, security, and economic aspects of China’s engagement with the 48 African countries from 2001 to 2008. Cultural engagement was measured in terms of whether or not a country hosted the Confucius Institute, established to promote Chinese language and facilitate cultural exchanges. Security engagement was measured by two features: whether or not China had a permanent defense attaché office and whether or not it transferred small arms to a country. I used four items to measure economic engagement: whether or not the country received concessional loans from China, the amount of bilateral trade, the amount of economic assistance (contracts, labor exchange, and consulting projects), and foreign direct investment (FDI) from China.

Table 1 compares the two groups—major recipients of China’s aid and other countries—for these seven measures. We can see that countries identified by the survey respondents as receiving large amounts of aid from China also: (1) were more likely to have played host to China’s Confucius Institutes and defense attaché offices; (2) had received small arms transfers, concessional loans, and economic assistance from China; and (3) had traded more with China. The difference in FDI inflows between the two groups of countries, however, is indistinguishable. Because most of these indicators should presumably be highly correlated with aid, these results suggest that the measures of major recipients, based on the expert survey, are fairly reliable.Table 1

Comparison of major recipients of Chinese aid and others


Major recipients (N = 11)

Other countries (N = 37)

Test statistics

Confucius institute




Defense attaché




Arms transfer




Concessional loans








Economic assistance




FDI flow




Oil producers








The test statistics for the percentages are based on the Chi square test. The test statistics for the numerical values are based on the two-sample t test

Source: Hanban Confucius Institutes around the World,” Hanban:, Shinn (2008), Hubbard (2008), China Statistical Yearbook, various years; Author’s survey, Beijing, November 2010

*** p < 0.01, ** p < 0.05, * p < 0.1

In addition to the survey-based measure, I took advantage of a newly available dataset compiled by a team of researchers at AidData, the Chinese Official Finance to Africa Dataset (COFAD). Published in April 2013, the COFAD draws from media reports on Chinese-backed projects, including information on 1673 projects in 51 African countries and on $75 billion in commitments of official finance between 2000 and 2011 (Strange et al. 2013). While its methodology and coverage have some issues, this dataset offers an alternative, quantifiable, country-level measure of Chinese economic assistance in Africa.11

The COFAD tracks three types of financial flows from China: official finance, unofficial finance, and military aid. Official finance includes ODA-like, other official flow (OOF)-like, vaguely official, and official investment. Unofficial finance includes NGO aid, corporate aid, joint ventures, and FDI, with and without state involvement. Military aid includes funds earmarked for non-developmental or non-humanitarian purposes. For the purpose of this study, I use the data on official finance, as it is the most similar to the OECD’s ODA. Figure 2 presents the data from 2000 to 2011. As the figure shows, a lot of variation occurred among the countries. The ten African states that received the largest amounts of official finance from China during this period, in descending order, are Ghana, Nigeria, Ethiopia, Democratic Republic of Congo, Sudan, Zimbabwe, Angola, Liberia, Republic of Congo, and Mozambique. Each of these countries obtained between $4.2 billion and $14.3 billion from China. In contrast, Somalia was only able to get $2.85 million, or two thousandths the amount received by Ghana. It is reassuring that seven of the top ten recipients based on the COFAD also belong to the list of major recipients compiled from expert opinions.

Fig. 2

China’s official assistance to Africa by country, 2000–2011. The map plots the total amount of official assistance each African state received from China between 2000 and 2011.

Source: AidData

For the rest of the variables, I rely on Dunning’s (2004) original dataset and extend it to 2008. Nevertheless, careful examination of his data reveals two measurement problems. First, population and GDP per capita are time invariant. In other words, the same values for these two variables are used for each country over all 23 years in his model. Second, during the process of expanding the dataset to 2008, I found that the key independent variable, the ratio of ODA to GNP, cannot be matched one-to-one with the same variable from the World Bank’s World Development Indicators (WDI). The correlation between the two is 0.951. While the numbers for most country-years are similar, the numbers for three countries (Angola, Ethiopia, and Senegal) are off by 1 year. That is, the ODA/GNP in 1990, for example, is recorded for 1991.

I fix these problems by updating the dataset with the correct measures. Data on population (measured in million people), GDP per capita (measured in PPP constant 2005 international dollar), and GNP (measured in constant local currency unit) for the period between 1975 and 2008 are drawn from the WDI. Data on ODA disbursement (measured in current dollar) are obtained from the OECD’s statistics library. I choose to compute the ratio of ODA to GNP rather than using the data directly from the WDI because the latter has less coverage. Despite some minor discrepancies, the correlation of the two measures is 0.99.

5 Findings

Table 2 reports the effects of Western aid conditionality in the three periods.12 Similar to Dunning (2004), for each period, I run the models with the contemporaneous, one-year, and five-year lags of ODA, which are instrumented with population and French colony to account for potential endogeneity issue.13Table 2

The conditioning effects of aid in three time periods





Model 1a

Model 1b

Model 1c

Model 2a

Model 2b

Model 2c

Model 3a

Model 3b

Model 3c


−0.0156 (0.0246)


0.0221* (0.0136)


−0.00612 (0.0187)




−0.0238 (0.0281)


0.0218* (0.0136)


−0.00627 (0.0181)




−0.0415 (0.0341)


0.0327** (0.0157)


−0.00593 (0.0189)

Cold War

−0.629*** (0.160)

−0.611*** (0.164)

−0.419** (0.168)



0.406*** (0.121)

0.387*** (0.124)

0.291*** (0.110)

0.718*** (0.124)

0.712*** (0.121)

0.722*** (0.115)

0.226 (0.175)

0.206 (0.159)

0.206 (0.136)


0.994*** (0.155)

1.000*** (0.162)

1.093*** (0.166)

0.373*** (0.127)

0.383*** (0.127)

0.347*** (0.130)

0.549*** (0.191)

0.586*** (0.176)

0.643*** (0.168)


−0.00237 (0.00395)

−0.00371 (0.00430)

−0.00736 (0.00449)

0.00234 (0.00251)

0.00251 (0.00258)

0.00309 (0.00302)

−0.0087*** (0.00310)

−0.0087*** (0.00287)

−0.00778*** (0.00271)


−0.332 (1.310)

−0.0406 (1.383)

0.876 (1.280)

−2.631** (1.187)

−2.604** (1.172)

−2.837** (1.185)

2.627 (1.614)

2.764* (1.470)

2.656** (1.303)

Wald Fstatistic










Sargan statistic






























Standard errors from the instrumental variable regressions are reported in parentheses. The instruments for ODA are all independent variables plus a dummy variable indicating whether or not the country was a French colony and a variable for population

*** p < 0.01, ** p < 0.05, * p < 0.1

Focusing on Models 1a–1c, we can see that the results are very similar to those reported in Dunning’s Table 2.14 For the Cold War period (1975–1986), the coefficient estimates on ODA/GNP are negative but statistically insignificant in all three models.15 The sign and statistical significance of the Soviet client dummy variable is also consistent with Dunning’s finding: both statistically significant, but negative in the earlier period and positive in the latter period. These results confirm that the impact of aid diverges in the two periods.

During the pre-FOCAC years (Models 2a–2c), the coefficient estimates on ODA/GNP are still statistically significant and positive at the 0.1 or 0.05 level in all three models, although the magnitude of the effects is considerably smaller (30–50%). By contrast, the coefficients on ODA/GNP in the post-FOCAC years (Models 3a–3c) are all negative and statistically insignificant. This provides support for the first hypothesis that the conditional effect of Western ODA on a recipient country’s democracy has diminished with the rise of China as an alternative source of aid.16

Does the size of Chinese aid matter? I explore this question first by adding the country-level variables for Chinese aid to the model. Because both the survey-based measures and the COFAD measures began in 2000, I am only able to estimate the model for the post-FOCAC period. The results are reported in Table 3. The first thing to notice here is that the ODA variables are statistically insignificant, confirming the finding that aid conditionality no longer worked during this period. In Models 4a–4c, the coefficients on the major recipient dummy variables are statistically significant and negative: the Freedom House scores of countries that received the lion’s share of China’s economic assistance are on average 0.8 points lower. This is quite substantial considering that the average score for these countries in 2000 was 3.14. In Models 5a–5c, I replace the major recipient dummy variables with the official assistance from China. This time, the coefficient estimates are still negative but indistinguishable from zero, probably due to the large amount of missing values in the COFAD dataset. Overall, these results offer partial support for the second hypothesis that countries receiving more aid from China are less likely to see their Freedom House score increase.Table 3

The effects of Chinese aid on democracy in Africa


Model 4a

Model 4b

Model 4c

Model 5a

Model 5b

Model 5c








−0.0138 (0.0188)


0.0127 (0.0205)




−0.0150 (0.0183)


0.0101 (0.0195)




−0.0199 (0.0193)


0.0284 (0.0257)

Major recipient

−0.788*** (0.232)

−0.792*** (0.217)

−0.802*** (0.205)




−0.0319 (0.0602)

−0.0186 (0.0534)

−0.0611 (0.0576)


0.226 (0.173)

0.203 (0.157)

0.187 (0.136)

0.481** (0.209)

0.421** (0.184)

0.414*** (0.155)


0.552*** (0.188)

0.590*** (0.174)

0.630*** (0.167)

0.957*** (0.234)

0.952*** (0.21)

1.000*** (0.205)


−0.00600* (0.00314)

−0.00599** (0.00290)

−0.00523* (0.00275)

−0.00499 (0.00383)

−0.00651* (0.00353)

−0.00481 (0.00359)


2.734* (1.594)

2.902** (1.457)

2.994** (1.304)

0.103 (1.869)

0.719 (1.674)

0.478 (1.514)

Wald Fstatistic







Sargan statistic





















Standard errors from the instrumental variable regressions are reported in parentheses. The instruments for ODA are all independent variables plus a dummy variable indicating whether or not the country was a French colony and a variable for population

*** p < 0.01, ** p < 0.05, * p < 0.1

It is possible that the ineffectiveness of aid conditionality is driven by domestic conditions in the recipient states, as suggested by recent case studies in Angola and Ethiopia (Hackenesch 2015). To further assess the causal effect of having access to aid from China, I employ a synthetic control analysis to compare changes in the Freedom House scores in the major recipients of Chinese aid over time with changes in other comparable countries. The synthetic control method is a data-driven procedure that constructs control units based on a “convex combination of comparison units that approximates the characteristics of the unit that is exposed to the intervention” (Abadie et al. 2010). In the present case, the intervention is the first FOCOC in 2000 and the treatment unit is the major recipient identified by the expert opinions. Because the treatment is applied to multiple states, I combine the units by taking the average and treating them as a single unit.17 I use the panel data between 1986 and 2008 for the synthetic control analysis because data before 1986 contain a fair amount of missing values. In other words, there are 14 years of pre-treatment data.

The outcome variable once again is the level of political freedom, measured by Freedom House scores. The predictors are ODA and population, which are averaged over the pre-treatment period and augmented by their lagged terms every two years back.18 The results of the synthetic control analysis are presented in Fig. 3, which plots the Freedom House scores for the major recipients and their synthetic counterpart from 1986 to 2008. Notice that the two trajectories track each other closely for the pre-FOCAC period. This suggests that the synthetic control group provides a sensible approximation of the level of political freedom that would have been observed in the major recipient countries in the absence of a large influx of Chinese aid.

Fig. 3

Trends in Freedom house scores for major recipients vs. synthetic control group

Immediately after the first FOCAC, however, the two lines begin to diverge. While the Freedom House score in the synthetic control group continues to rise, the score for the major recipients drops slightly in 2001, crawls back to the highest pre-treatment level, and then remains flat for three years before falling again in 2007. The discrepancy between the two lines suggests that OECD aid failed to improve democracy in major recipients of China’s unconditional aid, and that the gap increased over time. The magnitude of the estimated impact of being a major recipient is substantial: for the entire post-treatment period, the Freedom House score would have risen by an average of 0.55 points on a seven-point scale, an increase of 16.6%. These results provide stronger support for Hypothesis 2.

6 Conclusion

This study revisits an important and enduring question about the use of foreign aid to promote democracy. Building on earlier studies that focused on the temporal effect of aid conditionality, I discovered that the relationship between aid and democracy in Sub-Saharan Africa over the past three decades has been conditioned by two turns of events: the end of the Cold War in the late 1980s and China’s expanded engagement with Africa in the twenty-first century. The empirical evidence supports the thesis that aid conditionality works only during periods when African countries do not have alternative sources of aid, making the threat to pull the plug on aid more credible.

Proponents of aid conditionality in the West may find these results discouraging. Nevertheless, competition between China and the West is not, by definition, a bad thing for Africa. Recently, during discussions between John Huntsman, former US ambassador to China, and officials from African embassies, Kenyan Ambassador to China Julius Ole Sunkuli remarked that many African officials believe that “competition between donors has had positive consequences for African development” by giving the African countries “options after several decades of a largely Western development model. (Huntsman 2010)” South African Minister Plenipotentiary Dave Malcolmson made a similar comment:

China’s emergence in Africa as a counterbalance to US and European donors has been very positive for Africa by creating competition and giving African countries options. I recall that after the 2006 FOCAC summit, when China announced its commitments to Africa to much international media fanfare, traditional donors changed their attitude. They recognized that they had to measure up to China and “came calling”. The EU subsequently proposed infrastructure projects (after having de facto given up supporting these types of projects) and the World Bank began to support more agriculture projects (Huntsman 2010).

Finally, as China’s aid and development programs continue to expand, they may evolve as well. Risks of defaults in countries that received billions of dollars from China have reportedly prompted Beijing to reconsider its development finance strategies, shifting to a more institutional and multilateral approach to spread its risk (Kynge and Wildau 2015). Evidence for this comes from the establishment of the Asian Infrastructure Investment Bank (AIIB). While the United States has lamented its allies’ choice to embrace this new institution, which is widely regarded as a competitor to the World Bank, the participation of these traditional donors as founding members of the AIIB may lead to more convergence of China’s aid program with the Western model—at which point aid conditionality may once again become effective.



In this paper, I use the words “aid,” “assistance,” and “finance” interchangeably.



See also Kersting and Kilby (2014) for a comprehensive review of the empirical works on this topic.



For a detailed review of the history of China’s aid to Africa, see Shinn and Eisenman (2012).



The only ostensible condition for China’s aid is for the recipients to adhere to the one-China principle and cut their ties with the Taiwanese government. Many African nations see affirmation of Beijing’s sovereignty over Taiwan as a necessary part of their diplomacy with China, an easy concession that costs them nothing but can pay dividends. Some even manage to extract more aid by switching recognition from Taiwan to China or vice versa. The most notable case is the Central African Republic, which has switched sides for a record six times.



Zhang was recently appointed Executive Vice-Premier and a member of the Politburo Standing Committee of the Chinese Community Party.



The other BRIC countries (i.e., Brazil, Russia, and India) as well as South Africa are also increasing their lending to Africa, often unconditionally. China’s aid, however, is particularly attractive given its expansiveness and efficiency as well as China’s economic success over the past three decades.



Former French colonies are more likely to receive more aid, as France is Africa’s largest donor. So are countries with larger populations. For more discussion on these instruments, see Goldsmith (2001, 140–141) and Dunning (2004, 414).



China’s own policy adds to this confusion by encouraging its agencies and commercial entities to combine foreign aid with trade and investment promotion, service contracts, and labor cooperation (Sun 2014).



Departments under MOFCOM that handle and implement foreign aid policies include the Department of Aid to Foreign Countries, the Bureau for International Economic Cooperation, and the Tendering Board for Foreign Assistance Projects.



The order of the countries was randomized in each questionnaire to reduce bias.



For recent works that use this dataset, see Dreher and Fuchs (2015) and Dreher et al. (2014).



The results are robust to alternative specifications and measures of democracy. More details can be found in the online Appendix at



I conducted Durbin–Wu–Hausman tests to test whether 2SLS is needed. I found that the coefficients of the residual from the OLS regression of aid on all other variables are statistically significant, suggesting that OLS may be inconsistent, contrary to the findings reported by Dunning (2004, 414). Furthermore, in most of the models, the instruments successfully passed the Cragg–Donald Wald test and the Sargan overidentification test.



Initially, I was unable to replicate Dunning’s (2004) exact results in Table III with his original data. However, after experimenting with the model specifications, I managed to replicate the exact results by using the natural logarithm of GDP per capita. Interestingly, in his own replication of Goldsmith’s study, Dunning was puzzled by a “slight [but persistent] discrepancy on the coefficient of GDP per capita” (Dunning 2004, 415). This discrepancy can be resolved if we log transform GDP per capita with a base of 10.



The sample size in this period is much smaller because the WDI does not have data on GDP per capita prior to 1980. Nevertheless, replacing this variable with one calculated from Gleditsch’s (2002) expanded population and GDP data (1800–2000) yields very similar results. For ease of comparison with the other models I therefore report the results from using the WDI data.



One alternative explanation for the diminishing effect of aid conditionality is that the post-FOCAC period may coincide with democratic reversal in the recipient countries. This is unlikely to be the case, however, as most such reversals occurred around the late 1990s. Another alternative explanation could be that African states had already transitioned to democracy by the turn of the century, thus nullifying the need to condition aid on democratic reforms. A cursory examination of the data, however, indicates that this is not the case. The average Freedom House score for all of the 48 African states in the 2000s is 3.7, which is near the lower end of the “not free” category.



According to Abadie et al. (2011, 3), in cases where there are multiple units exposed to the intervention, one can “aggregate the data from the regions exposed to the intervention”.



More details on these analyses are available in the online Appendix at



I am grateful to Songying Fang, Andreas Fuchs, Scott Kastner, Kirsten Rodine-Hardy, Ka Zeng, Linting Zhang, and seminar participants at Stanford University and the University of British Columbia for helpful comments. Earlier drafts of this paper were presented in 2013 at the annual conventions of the International Studies Association and the American Political Science Association.

Supplementary material

41111_2017_50_MOESM1_ESM.pdf (174 kb)

Supplementary material 1 (PDF 174 kb)


  1. Abadie, Alberto, Alexis Diamond, and Jens Hainmueller. 2010. Synthetic control methods for comparative case studies: estimating the effect of California’s tobacco control program. Journal of the American Statistical Association 105 (490): 493–505.CrossRefGoogle Scholar

  2. Abadie, Alberto, Alexis Diamond, and Jens Hainmueller. 2011. Synth: an R package for synthetic control methods in comparative case studies. Journal of Statistical Software 42 (13): 1–17.CrossRefGoogle Scholar

  3. Ahmed, Faisal Z. 2012. The perils of unearned income: aid, remittances, and government survival. American Political Science Review 106 (1): 146–165.CrossRefGoogle Scholar

  4. Bearce, David H., and Daniel C. Tirone. 2010. Foreign aid effectiveness and the strategic goals of donor governments. Journal of Politics 72 (3): 837–851.CrossRefGoogle Scholar

  5. Bermeo, Sarah B. 2016. Aid is not oil: donor preferences, heterogeneous aid, and the aid–democratization relationship. International Organization 70 (1): 1–32.CrossRefGoogle Scholar

  6. Brown, Stephen. 2005. Foreign aid and democracy promotion: lessons from Africa. The European Journal of Development Research 7 (2): 179–198.CrossRefGoogle Scholar

  7. Bueno de Mesquita, Bruce, and Alastair Smith. 2010. Leader survival, revolutions, and the nature of government finance. American Journal of Political Science 54 (4): 936–950.CrossRefGoogle Scholar

  8. Brautigam, Deborah A. 2008. China’s foreign aid in africa: what do we know? In China into Africa: trade, aid and influence, ed. Robert I. Rotberg, 197–216. Baltimore: Brookings Institution Press.Google Scholar

  9. Brautigam, Debora A. 2009. The dragon’s gift: the real story of China in Africa. New York: Oxford University Press.Google Scholar

  10. Brautigam, Debora A., and Stephen Knack. 2004. Foreign aid, institutions, and governance in sub-Saharan Africa. Economic Development and Cultural Change 52 (2): 255–285.CrossRefGoogle Scholar

  11. Carothers, Thomas. 1999. Aiding democracy abroad: the learning curve. Washington, D.C.: Carnegie Endowment for International Peace.Google Scholar

  12. Chen, Dingding, and Katrin Kinzelbach. 2015. Democracy promotion and China: blocker or bystander? Democratization 22 (3): 400–418.CrossRefGoogle Scholar

  13. Chen, Deming. 2012. Speech at the Fifth Ministerial Meeting of FOCAC, July 19, 2012. MOFCOM. Accessed 31 March 2013.

  14. Council on Foreign Relations. 2007. Africa-China-US Trilateral Dialogue Summary Report. Accessed 13 August 2013.

  15. Crawford, Gordon. 1997. Foreign aid and political conditionality: issues of effectiveness and consistency. Democratization 4 (3): 69–108.CrossRefGoogle Scholar

  16. Davies, Martyn J. 2008. Special economic zones: China’s developmental model comes to Africa. In China into Africa: trade, aid and influence, ed. Robert I. Rotberg, 137–154. Baltimore: Brookings Institution Press.Google Scholar

  17. Djankov, Simeon, Jose G. Montalvo, and Marta Reynal-Querol. 2008. The curse of aid. Journal of Economic Growth 13 (3): 169–194.CrossRefGoogle Scholar

  18. Dreher, Axel. 2009. IMF conditionality: theory and evidence. Public Choice 141 (1–2): 233–267.CrossRefGoogle Scholar

  19. Dreher, Axel, and Andreas Fuchs. 2015. Rogue aid? The determinants of China’s aid allocation. Canadian Journal of Economics 48 (3): 988–1023.CrossRefGoogle Scholar

  20. Dreher, Axel, Andreas Fuchs, Roland Hodler, Bradley Parks, Paul A. Raschky, and Michael J. Tierney. 2014. Aid on demand: African leaders and the geography of China’s foreign assistance. AidData Working Paper 3.Google Scholar

  21. Dunning, Thad. 2004. Conditioning the effects of aid: cold war politics, donor credibility, and democracy in Africa. International Organization 58 (2): 409–423.CrossRefGoogle Scholar

  22. Finkel, Steven E., Aníbal Pérez-Liñán, and Mitchell A. Seligson. 2007. The effects of US foreign assistance on democracy building, 1990–2003. World Politics 59 (3): 404–439.CrossRefGoogle Scholar

  23. FOCAC. 2013. FOCAC ABC. Accessed 31 March 2013.

  24. Gleditsch, Kristian Skrede. 2002. Expanded trade and GDP data. Journal of Conflict Resolution 46 (5): 712–724.CrossRefGoogle Scholar

  25. Goldsmith, Arthur A. 2001. Foreign aid and statehood in Africa. International Organization 55 (1): 123–148.CrossRefGoogle Scholar

  26. Guyana Times International. 2012. Trotman hails China’s “no strings attached” approach to aid. Accessed 19 March 2013.

  27. Hackenesch, Christine. 2015. Not as bad as it seems: EU and US democracy promotion faces China in Africa. Democratization 22 (3): 419–437.CrossRefGoogle Scholar

  28. Heckelman, Jac C. 2010. Aid and democratization in the transition economies. Kyklos 63 (4): 558–579.CrossRefGoogle Scholar

  29. Horta, Loro. 2013. China in Africa: soft power, hard results. YaleGlobal, 13 November 2009. Accessed 13 March 2013.

  30. Hubbard, Paul. 2008. China’s Concessional Loans. In China into Africa: trade, aid and influence, ed. Robert I. Rotberg, 217–229. Baltimore: Brookings Institution Press.Google Scholar

  31. Huntsman, Jon. 2010. African countries prefer Chinese aid to US-China Ccooperation, The Guardian. Accessed 13 August 2013.

  32. Kelley, Judith. 2008. Assessing the complex evolution of norms: the rise of international election monitoring. International Organization 62 (2): 221–255.CrossRefGoogle Scholar

  33. Kersting, Erasmus, and Christopher Kilby. 2014. Aid and democracy redux. European Economic Review 67: 125–143.CrossRefGoogle Scholar

  34. Knack, Stephen. 2004. Does foreign aid promote democracy? International Studies Quarterly48 (1): 251–266.CrossRefGoogle Scholar

  35. Kilby, Christopher. 2009. The political economy of conditionality: an empirical analysis of world bank loan disbursements. Journal of Development Economics 89 (1): 51–61.CrossRefGoogle Scholar

  36. Kynge, James, and Gabriel Wildau. 2015. China: with friends like this. Financial Times, 17 March 2015. Accessed 26 March 2015.

  37. Lum, Thomas, Hannah Fischer, Julissa Gomez-Granger, and Anne Leland. 2009. China’s foreign aid activities in Africa, Latin America, and Southeast Asia. Washington, DC: Congressional Research Service Report.Google Scholar

  38. Marchesi, Silvia, and Jonathan P. Thomas. 1999. IMF conditionality as a screening device. Economic Journal 109 (454): 111–125.Google Scholar

  39. Morrison, Kevin M. 2009. Oil, nontax revenue, and the redistributional foundations of regime stability. International Organization 63 (1): 107–138.CrossRefGoogle Scholar

  40. Naim, Moisés. (2007). Help not wanted. The New York Times. 15 February 2007.Google Scholar

  41. Öhler, Hannes, Peter Nunnenkamp, and Axel Dreher. 2012. Does conditionality work? A test for an innovative US aid scheme. European Economic Review 56 (1): 138–153.CrossRefGoogle Scholar

  42. Olson, Stephen, and Clyde Prestowitz. 2011. The evolving role of China in international institutions. Report prepared for The U.S.-China Economic and Security Review Commission, Washington, DC: The Economic Strategy Institute.Google Scholar

  43. Peerenboom, Randall. 2008. China modernizes: threat to the West or model for the rest?Oxford: Oxford University Press.CrossRefGoogle Scholar

  44. People’s Daily. 2011. Africa welcomes China’s investment. Accessed 19 March 2013.

  45. People’s Daily. 2013. China to provide 20 billion USD loan credits to Africa in three years. Accessed 31 March 2013.

  46. Qian, Nancy. 2015. Making progress on foreign aid. Annual Review of Economics 7: 277–308.CrossRefGoogle Scholar

  47. Ramo, Joshua Cooper. 2004. The Beijing consensus: notes on the new physics of Chinese power. London: Foreign Affairs Policy Centre.Google Scholar

  48. Robinson, Mark. 1993. Aid, democracy and political conditionality in sub-Saharan Africa. The European Journal of Development Research 5 (1): 85–99.CrossRefGoogle Scholar

  49. Shinn, David H. 2008. Military and security relations: China, Africa, and the rest of the world. In China into Africa: trade, aid and influence, ed. Robert I. Rotberg, 155–196. Baltimore: Brookings Institution Press.Google Scholar

  50. Shinn, David H., and Joshua Eisenman. 2012. China and Africa: a century of engagement. Philadelphia: University of Pennsylvania Press.CrossRefGoogle Scholar

  51. Smith, Alastair. 2008. The perils of unearned income. Journal of Politics 70 (3): 780–793.CrossRefGoogle Scholar

  52. Stokke, Olav. 2013. Aid and political conditionality. London: Routledge.Google Scholar

  53. Stone, Randall W. 2008. The scope of IMF conditionality. International Organization 62 (4): 589–620.CrossRefGoogle Scholar

  54. Strange, Austin M., Bradley Park, Michael J. Tierney, Andreas Fuchs, Axel Dreher, and Vijaya Ramachandran. 2013. China’s development finance to Africa: a media-based approach to data collection. Center for Global Development Working Paper.Google Scholar

  55. Sun, Yun. 2014. China’s aid to Africa: monster or messiah? Brookings East Asia Commentary No. 75.Google Scholar

  56. Taylor, Ian. 2012. The forum on China-Africa cooperation. London: Routledge.Google Scholar

  57. Thaler, Farah. 2012. China’s foreign aid to Africa. The Washington Times. Accessed 13 March 2013.

  58. Wade, Abdoulaye. 2008. Time for the West to practice what it preaches. Financial Times.,dwp_uuid=8735dcb2. Accessed 31 March 2013.

  59. Williamson, John. 1983. IMF conditionality. Washington, DC: Institute for International Economics.Google Scholar

  60. Woods, Ngaire. 2008. Whose aid? Whose influence? China, emerging donors and the silent revolution in development assistance. International Affairs 84 (6): 1205–1221.CrossRefGoogle Scholar

  61. World Bank. 2005. Review of World Bank conditionality. World Bank Operations Policy and Country Services.Google Scholar

  62. Wright, Joseph. 2009. How foreign aid can foster democratization in authoritarian regimes. American Journal of Political Science 53 (3): 552–571.CrossRefGoogle Scholar

  63. Zafar, Ali. 2007. The growing relationship between china and sub-Saharan Africa: macroeconomic, trade, investment, and aid links. The World Bank Research Observer 22 (1): 103–130.CrossRefGoogle Scholar

Chinese Political Science Review

June 2017, Volume 2, Issue 2, pp 201–220

News category: