
Democracy is under threat. According to a report of the Varieties of Democracy Institute, “the level of democracy enjoyed by the average global citizen in 2022 is down to 1986 levels” and “more than 35 years of global advances in democracy have been wiped out in the last decade” (Papada et al. 2023, 9). Freedom House also claims that global democracy has declined for the 17th consecutive year (2023, 1). From 2002 to 2022, while the number of democratizing countries has decreased from 43 to 14, the number of autocratizing countries has increased from 13 to 42 (Papada et al. 2023, 7). The decline of democracy is not only a problem of developing countries. Some advanced democracies are also experiencing democratic backsliding due to the rise of populism and ethno-nationalism as economic inequality deepens and disinformation explodes in the public sphere (Power 2023, 26-28).
Democratic donors have continued to provide democracy aid1 to promote democracy in developing countries. Has such democracy aid promoted democracy? Among a few quantitative studies on this question, while most of them have argued its positive effect (Finkel et al. 2007; Kalyvitis and Vlachaki 2010; and Scott and Steele 2011), Nielson and Nielson (2010) suggest a skeptical view that the seemingly positive effect of democracy aid is driven by selection effects, that is, most democracy aid goes to countries, where donors have higher expectations that changes towards democracy would follow. One of the most recent studies on this matter, a study by Lührmann et al. (2018), also maintains a positive view, while recognizing that the effect of democracy aid varies by regime type. While these studies have greatly shed light on this question, their works still have limitations in several aspects: the coverage of donors, lack of analysis over a recent period, and operationalization of ‘democracy aid.’
In this paper, to find out whether democracy aid promotes democracy, I focus on the donors, not recipients, because foreign aid (including democracy aid) functions as an important part of foreign policy (Carothers 1999, 85-122; Milner and Tingley 2013; Morgenthau 1962). That is, donors do allocate their scarce resources to certain recipient countries for certain purposes. Bermeo (2011) argues that donors’ intent does matter to encourage democracy in recipient countries in her study about the relationship between foreign aid and regime change. Concurring with her, I claim that to achieve the objectives of democracy aid, accurately allocating their resources to the right sub-sectors and implementing proper aid programs are critical as Carothers indicates (1999, 255-280). In addition, although the effectiveness of democracy aid also depends on the will and capacity of recipient countries, from a policy perspective, it is not easy to secure the support and willingness of leaders and elites from recipient countries, especially when they lead authoritarian or hybrid regimes.
This article seeks to answer what kind of democracy aid works to promote democracy, by comparing the effects of democracy aid and its sub-categories on a democracy index, the Varieties of Democracy Index (V-Dem, Coppedge et al. 2022). Using aid data from OECD/DAC database, I cover democracy aid from all donors from 2002 to 2019, during which time the importance of democracy aid increased significantly. In the following, first, I explore the emergence of democracy aid as a major part of official development assistance (ODA) and problematize the varied operationalization of democracy aid concept by different donors and scholars. Then, I introduce my research design and discuss the results of statistical tests. Finally, I suggest policy recommendations for donors.
While there was democracy-related aid even in the Cold War era especially from the United States (for a detailed history, see Carothers 1999), the issue of democracy (and governance more broadly) became an important agenda in the development field at the end of 1980s (Bräutigam 1992, 4). The end of the Cold War motivated Western donor countries to pay more attention to the governance and strengthening of democracy of partner countries because the necessity to keep supporting non-democratic regimes due to competition with the Soviet Union had waned drastically. Indeed, the amount of democracy aid of USAID (called “Democracy and Governance Funds”) increased from \$128 million in 1990 to \$817 million in 2003, measured in 2000 dollars (Finkel et al. 2007, 415). However, as Carothers and De Gramont (2013, 69) indicated, when donors started to include the agenda of governance in development, it was about states’ economic management, “effective financial management and a well-regulated market.” Although the concept of governance in development widened, it mostly entailied economic aspects For example, when Dollar and Pritchett (1998) claimed “sound management,” a term they use as a synonym of “good governance” in the paper, it meant controlled inflation, fiscal balance, open trade regimes, rule of law, quality public bureaucracy, and low corruption (12). Burnside and Dollar (2000, 897) posited that “aid has a positive impact on growth in developing countries with good fiscal, monetary, and trade policies but has little effect in the presence of poor policies,” and in this case the donors’ interest in governance lay mainly in the economic aspect. In practice, as De Alcántara (1998, 107-108) indicates, the World Bank’s support in the name of “good governance” in the 1990s had been dedicated primarily to improving public sector management (mainly ficsal aspects) and privatization. In sum, donors’ understanding of and support for governance in the 1990s was more about technical management of economic neo-liberalization.
The conceptualization of governance has widened in the 2000s. The argument of ‘sound institution and policies’ has been more explicitly expanded to the political sphere in the Monterrey Consensus, the outcome of the United Nations International Conference on Financing Development in Monterrey, Mexico, in 2002. It announced that “[S]ound policies and good governance at all levels are necessary to ensure ODA effectiveness” (United Nations 2002, 9).2 The 9/11 attack on the United States (U.S.) in 2001 and increasing interest about fragile countries thereafter raised issues of governance and democracy significantly. The 2002 National Security Strategy (NSS) (White House 2002) emphasized “development by opening societies and building the infrastructure of democracy.” Later, through the 2006 NSS, the U.S. Government elevated ‘development’ as one of three pillars for U.S. national security next to defense and diplomacy (White House 2006). The militarization of aid (Kilby 2012) and the rising interest in democracy and governance in the 2000s pushed many donors to pour huge quantities of resources into political aid such as ‘democracy aid’ or ‘aid for governance’ (Easterly 2003; Scott and Steele 2011). Interestingly, in the OECD/DAC statistics, the category of ‘Government and Civil Society’ was newly created for reporting aid activities related to governance and democracy in 2002. Donors’ increased interest in institution building during that time and the rise of rights-based approach (Carothers and De Gramont 2013: 98-104) promoted more directly political objectives as a development goal to strengthen democratic institutions. According to OECD/DAC (Development Assistance Committee) Statistics, the total amount of ODA for democracy and governance from donors to developing countries has increased from \$4.1 billion in 2002 to 16.3 billion in 2019 (OECD 2022).3 As Figure 1 shows, the proportion of ‘Democracy Aid’ in total ODA has increased from 6.7% in 2002 to 10.0% in 2019 although there has been some fluctuation during the period. This trend attests that the importance of ‘Democracy Aid’ has increased.
The effect of democracy aid, however, has been relatively neglected in academic research about foreign aid, considering the accelerating securitization of foreign aid and increased volume. First of all, while there are many studies that question the effect of aggregated foreign aid on governance and/or democracy at the national level, the results are not consistent: while some argue that foreign aid positively affects governance and/or democracy by using cross- national quantitative methods (Brown 2005; Crawford 2001; Goldsmith 2001; Dunning 2004; Wright 2009), others refute them by suggesting different statistical results (Bräutigam and Knack 2004; Knack 2001 and 2004; Rajan and Subramanian 2007; Busse and Gröning 2009).
Although there are a few prominent pieces of scholarship, which directly question the effect of democracy aid on democracy (See Table 1 in page 98), there are several issues. First, Finkel et al. (2007), Scott and Steele (2001; 2005), Scott and Carter (2020) focus only on U.S. aid. While the U.S. is one of main donors in democracy aid, it only accounts for around 25% of all democracy aid from 2002-2019.4 Therefore, their articles cannot show the effect of global democracy aid on democracy. Secondly, considering the accelerating securitization of foreign aid since 9/11, the stall of the third wave of democratization, and the decline of democracy, it is necessary to examine the recent performance and outcomes. Therefore, it is necessary to examine the recent effect of democracy aid on the global level. Finally, the operationalization of democracy aid is problematic.
It is not an easy question to define or operationalize democracy aid with respect to democracy. Savun and Tirone (2011, 235) divide ‘democracy aid’ into three categories: (1) state institutions, (2) civil society, and (3) electoral assistance. While such categorization seems sound, reality might be different. Thomas Carothers,5 who has extensively published books and articles about ‘democracy aid,’ stated that:
“Even though this type of international assistance [democracy aid] has a specific goal—to foster and advance democratization— criteria for assessing it are elusive. “Democracy aid” itself is a catchall term for an endeavor that has acquired enough moving parts to make drawing boundaries around it difficult” (Carothers 2015, 59).
As his statement shows, the concept of democracy aid is fuzzy. This fuzziness gets more problematic when scholars operationalize ‘democracy aid,’ because there are several linkages between the practices of donors (such as projects and programs), the reporting of such practices as data (e.g., OECD/DAC), and what scholars count as ‘democracy aid.’
First, there are the practices of donor agencies. For example, according to the study of Scott and Steele (2005, 448), NED’s grants were:
“directed toward (a) promoting and supporting worker rights and political participation (26.3 per cent); (b) building and supporting civic participation and education (25.3 per cent); (c) promoting human rights (10.9 per cent) and market reforms (10.8 per cent); (d) developing building political institutions such as parliaments and political parties (9.9 per cent); (e) developing the institutions and activities of a free press (8.9 per cent); (f) elections (6.0 per cent); and (g) basic conflict resolution in societies suffering from such instability (2.0 per cent).”
However, many scholars, including myself, might not agree with an argument that suggests support to market reforms and conflict resolution is part of democracy aid. For another example, the USAID Strategy on Democracy, Human Rights, and Governance indicates that the “strategy provides a framework to support the establishment and consolidation of inclusive and accountable democracies to advance freedom, dignity, and development” (2013, 4). However, the first paragraph of its main homepage for “Democracy, Human Rights, and Governance” states that:
“Democratic governance and human rights are critical components of sustainable development and lasting peace. Countries that have ineffective government institutions, rampant corruption and weak rule of law have a 30-to-45 percent higher risk of civil war and higher risk of extreme criminal violence than other developing countries” (USAID 2023).
Some would say that ineffectiveness and corruption is not directly related to democracy, considering the existence of many ineffective and corrupt but democratic countries in Asia and Africa. These cases show that it may be problematic to count all projects of certain institutions or divisions as cases of democracy aid.
Secondly, what to include as democracy aid can be problematic. As Table 1 in page 98 shows the tree of OECD/DAC statistical reporting consists of many sub-categories. For example, in their article, Kalyvitis and Vlachaki (2010) counted all “150” aid cases as democracy aid. However, in this case, the data included conflict-related activities such as “Removal of land mines and explosive remnants of war (15250)” as a part of democracy aid, which many scholars would not agree with. The sub-categories of “1511*” are more generally related to governance, not directly to democracy. Lührmann et al. (2018) count only four subcategories (15150, 15151, 15153, and 15160) as democracy aid. However, from the perspective of egalitarian democracy or participatory democracy ‘Women’s equality organisations and institutions’ (15170) should be part of democracy aid. And for other scholars, ‘legal and judicial development’ (15130) may be critical to protect human rights and democratic principles as a part of rule of law. In addition, those components related to ‘good governance’ such as public finance management or anti-corruption, may improve democracy, although it is hard to say they are an essential part of democracy. Another significant issue in the study of democracy aid is that authors do not clarify how they operationalize democracy aid. As seen in Table 2 in page 99, many authors do not disclose what components are counted as ‘democracy aid’ in their studies. That is, the current practices of operationalizing democracy aid are quite arbitrary.
In brief, previous research about the relationship between democracy and democracy aid has three critical issues: (1) the operationalization of democracy aid is fuzzy and arbitrary due to the inclusion of other types of aid not directly related to democracy; (2) some of them only deal with U.S aid. which accounts for only 25% of democracy aid; (3) and they do not cover recent periods when democracy aid has become more important.
In this research, I statistically examine the effectiveness of democracy aid on democracy, using data from OECD/DAC (from 2002 to 2019) and from V-Dem6 (varieties of democracy) from 2000 to 2021. The core and main difference of this research from other previous studies is that I compare the effects of several categories of democracy aid on five different conceptualizations of democracy.
As mentioned earlier, as the operationalization of democracy aid varies by scholars, the conceptualization of democracy also varies by scholars. Therefore, the democracy index itself should be multi-faceted. Although the Polity Project is the most widely used democracy index, as Coppedge (2012) and Munck and Verkuilen (2002) discuss, it might not the best index for democracy because it excessively focuses on the electoral components of democracy. Unlike other democracy indices such as the Polity Project (Center for Systemic Peace 2021) and Freedom House Rating (Freedom House 2022), the V-Dem Index conceptualizes democracy through five different lenses: polyarchy (electoral democracy), participatory democracy, deliberative democracy, liberal democracy, and egalitarian democracy. While the V-Dem Index is relatively new, its usefulness and superior performance over other democracy indices due to its “underlying definition and measurement scale,” the coverage of non-missing observations, and “the availability of disaggregate data,” is well indicated (Boese 2019, 95). Therefore, I use the five different indices of V-Dem as indicators of my dependent variable for my research.
Regarding the data related to aid, used as independent variables, I extract from OECD/DAC (2022) over the period from 20027 to 2019.8 Because the list of aid recipient countries of OECD/DAC has been regularly updated (OECD 2022), if countries’ data points have been less than 50% in the target years (from 2002 to 2019), I exclude them from my analysis. Finally, I cover 128 developing countries. For control variables, from the literature of democracy, democratization and foreign aid, I include economic development (GNI per capita, Burkhart 2000; Lipset 1959), urbanization (the portion of population who lives in urban areas, Barro 1999; Lipset 1959), trade openness (the percentage of trade over GDP, Hadenius 1992), and aid dependency (the percentage of ODA over GNI, Bräutigam and Knack 2004; Brown 2005; Knack 2001; Goldsmith 2001). The data for control variables is from World Bank (2022) and calculation with other data from OECD/DAC (2022).
I use a Simultaneous Equation Model for this research due to an endogeneity problem between democracy aid and democracy. That is, democracy aid may go to countries, where donors have higher expectations that changes towards democracy will follow. In practice, donors tend to provide aid to the countries, which can show better results to justify their uses of the budget to their own taxpayers. This endogeneity between aid and democracy (and other links) is an important problem to control as several studies about democracy aid indicate (Finkel et al. 2007; Scott and Steele 2011). To control the endogeneity issue, I use a Simultaneous Equation Model (SEM).9
In addition, it is more reasonable to assume that democracy aid has a lagged effect on democracy, not in the same year of aid disbursement or the next year. Therefore, I use the two-year lagged effect of democracy aid and control variables on the changes of democracy.
Finally, when I use democracy aid as an independent variable, I test both democracy aid per capita and at the national gross level. In general, it is common to use the aid amount per capita in the studies of foreign assistance because the actual amount offered to each citizen of recipient countries needs to be calculated. However, in the case of democracy aid, it is quite tricky to use aid per capita because many projects or programs of donor agencies are about establishing institutions. That is, if a policy is adopted through democracy aid, while it is applied to all the citizens, it is hard to say the cost of establishing the policy proportionately increases by the size of national population. However, in the case of electoral support, because the size of population matters, using aid per capita makes more sense. Therefore, I test both for democracy aid per capita and at the national gross level and try to interpret the relevant results.10 Table 3 is the summary of my model.
Table 4 is the result of statistical tests, using democracy aid per capita. As seen in the table, most components of democracy aid are significantly correlated to the increase of different conceptualizations of democracy. However, aid for ‘Elections’ is not statistically significant with four conceptualizations of democracy, except for ‘electoral democracy.’
The statistical results, based on democracy aid at the national gross level, show a different picture from the previous one. As seen in this table, overall democracy aid (DAC code 151), governance-related assistance (from DAC Code 15110 to 15113), and aid to ‘Legal and Judicial Development (15130)’ are positively correlated to the increase of democracy in most types of democracy. However, aid to ‘Legislature and Political Parties (15152),’ ‘Media and Free Flow of Information (15153),’ and ‘Human Rights (15160)’ are negatively correlated to democracy indices in statistically significant ways. In addition, support for ‘Democratic Participation and Civil Society (15150),’ ‘Election (15151),’ and ‘Women’s Equality (15170)’ is mostly statistically insignificant.
From the above results, several findings can be discussed. First, electoral assistance from donors may not be conducive to the promotion of democracy. Rather, the existence of donors’ support to certain elections sometimes unintentionally justifies flawed or rigged elections in competitive authoritarian countries by the fact that donors supported the election process and participated in the monitoring. For example, while the electoral assistance of United Nations Development Program (UNDP) and ten other donors, introduced the preparation of the 2011 presidential election in the Democratic Republic of Congo (DRC) as “a success story,” (UNDP 2010) the election was somewhat fraudulent as the Carter Center criticized that the results of this election lacked credibility (Carter Center 2011). Though donors expressed their concerns about the legitimacy and credibility of the election results (World Bank 2013), they could not but accept the consequences of the election.12 In addition, Levitsky and Ziblatt suggest that “since the end of the Cold War, most democratic breakdowns have been caused not by generals and soldiers but by elected governments themselves … democratic backsliding today begins at the ballot box” (2018, 9). That is, although elections are a critical component of democracy, aid to elections, not “free and fair,” might not promote democracy.
Secondly, components, which seem more technically related to ‘good governance’ such as ‘Public Sector Policy,’ ‘Public Finance Management,’ ‘Decentralization,’13 and ‘Anti-corruption’ consistently improve democracy per capita and at the national gross level. Aid to ‘Legal and judicial Development’ also appears to promote democracy.14 However, aid to norm-based and institutional components such as ‘Human Rights’ and ‘Legislature and Political Parties’ may not be conducive to the promotion of democracy at the national level. As Bräutigam indicates (1992) institutions are difficult to transfer. Especially, when certain institutions are more norm-based and exogenous to developing countries, they may not be functioning as intended before the norms become internalized. Especially, considering that many developing countries are ruled by authoritarian leaders, who may not be interested in the proper functioning of democratic institutions, aid to such institutions may not be very conducive to democracy.
Finally, it is beyond the scope of this research, to explain why “Media and Free Flow of Information” is negatively correlated to all conceptualizations of democracy at the national gross level. However, I hope to offer one probable reason, namely that the media is often dominated by the state and authoritarian leaders in authoritarian developing countries, and therefore, it may be functioning to impede the development of democracy.15 Of course, this needs further research to provide a more persuasive explanation.
In this article, I problematize the following issues in existing studies of democracy aid: 1) the operationalization of ‘democracy aid’ is fuzzy and arbitrary; 2) they do not cover the recent period in which the volume and importance of ‘democracy aid’ has dramatically increased; and 3) many of these studies only focus on aid provided by the United States. Then, I explore the trajectories of how governance and democracy have emerged as important agendas in the aid community and discuss the complexity of the ‘democracy aid’ concept. Then, I compare the effects of several categories of democracy aid on the Varieties of Democracy Index, using OECD/DAC aid data from all donors since 2002. I find that, while overall ‘democracy aid’ appears to promote democracy, the positive effect is associated with aid targeted towards governance-related issues. Conversely, aid provided to support election and democratic institutions, which involve the embedding of norms, are not yielding the desired results
These results may hint at an important policy implication namely that democracy aid to ensure ‘good governance’ may be more effective to promoting democracy. Many developing and aid-recipient countries are not mature democracies, rather, most of them are authoritarian or hybrid regimes. According to Teorell, who assessed vast hypotheses about the determinants of democratization, argues that socioeconomic modernization and economic freedom impede the downturns of democracy (2010, 145). Another recent study also finds a similar result that “good short-term economic performance reduces the magnitude of democratic downturns” (Coppedge et al. 2022, 25). Reflecting on these observations, it would be better for donors to focus on improving governance and building effective governments which may facilitate socioeconomic modernization and economic development rather than more directly focusing on democratic institutions.16 Such institutions may not function as intended when recipient countries are not equipped with the willingness and capacity to run them. Carothers once warned that “[r]ule-of-law aid cannot substitute for the internal [recipients’] will to reform” (1998, 98) when such aid was once considered a panacea by Western donors. His statement still holds. Meanwhile, democracy aid to building effective governments may help recipients develop capacities to run their countries and democratic institutions in the future.
To find out what kind of democracy aid effectively works in certain conditions of recipient countries is another important question. In addition, given that there are various modalities of democracy aid and that effectiveness of actual aid is also dependent on the implementation of various actors and processes in the aid industry (De Haan 2009), this paper lacks important discussions about them. I will leave them as follow-up research agenda for now.
1 While quite a number of scholarly works have been produced, the concept of ‘democracy aid (or assistance)’ is a fuzzy one. For a detailed discussion, see the Section of ‘What is Democracy Aid.’
2 “Good governance is essential for sustainable development. Sound economic policies, solid democratic institutions responsive to the needs of the people and improved infrastructure are the basis for sustained economic growth, poverty eradication and employment creation. Freedom, peace and security, domestic stability, respect for human rights, including the right to development, and the rule of law, gender equality, market-oriented policies, and an overall commitment to just and democratic societies” (United Nations 2002, 7).
3 Specifically, I count the aid included in the category of “Government & Civil Society - general Total” (classification number ‘I.5.a’ or DAC code 151) of OECD/DAC statistics.
4 Calculated from OECD/DAC (2022).
5 According to Google Scholar Search, the term ‘democracy assistance (or aid)’ has been first used in academia by him in the late 1990s.
6 I use V-dem Dataset Version 12, published in March 2022, for this study.
7 OECD/DAC has statistics classified for government-related aid from 2002.
8 While ODA data from 2020 to 2021 is also available as of January 2023, because of the dramatic changes in the two years due to the COVID-19 pandemic, I exclude the data from this research.
9 The form of endogeneity of explanatory variables in the research about relationships between democracy aid and democracy is simultaneity (Scott and Steel 2011, 56-60). SEMs help to deal with the simultaneity issue when one or more explanatory variables are jointly determined with the dependent variables (Wooldridge 2014, 530 and 548-551).
10 While Finkel et al. (2007) mostly uses USAID’s democracy aid at the national level, it also tests democracy aid per capita.
11 Because several CRS categories have been created recently, they do not have enough data points to statistically test in the panel data.
12 There is an interesting study about the relationship between democracy aid and electoral violence. Von Borzyskowski argues that international election support may “contribute to unrest and instability” while it usually helps reduce electoral violence (2019, 15). Given that electoral violence and/or intimidation occur in 58% of elections in Africa (Burchard 2015), where many developing and aid-recipient countries exist, electoral assistance may contribute to the easing of electoral violence, but not to democracy itself.
13 Norris argues that decentralization is “to strengthen democratic participation, representation and accountability, as well as improving government efficiency and effectiveness” (2008, 184). According to the statistical result, his claim works.
14 The aid to this sector generally purports to foster the rule of law. While there are discussions about the relation between the rule of law and democracy, it is generally agreeable that the rule of law is respected in mature democracies. For the details see Maravall and Przeworski 2013 and O’Donnell 2014.
15 Think of the role of the government-controlled Radio-Télévision Libre des Mille Collines (RTLM) in the 1994 Rwandan genocide to spread ethnic hostility and to involve mass killings. For the details see Kellow and Steeves 1998.
16 In her recent article, Power, the current head of USAID, argues that democracies made the biggest errors “to view individual dignity primarily through the prism of political freedom without being sufficiently attentive to the indignity of corruption, inequality, and a lack of economic opportunity” (2023, 28). Her statement resonates with my argument to a certain extent that support to good governance for overall socioeconomic development may be more conducive to democracy than focusing on narrow forms or values of democracy.
![]() |
![]() |