Fragile and war-torn countries often exacerbate domestic instability and increase the likelihood of transnational threats, including armed conflicts, illicit activities, and humanitarian crises.Footnote 1 The potential spillover effects of these conditions have been a concern for policymakers around the world, leading to robust international state-building projects aimed at promoting the performance, capacity, and responsiveness of state institutions.Footnote 2
Much of the literature on international state building treats it as a monolithic process whereby the international community imposes reform efforts to strengthen domestic institutions.Footnote 3 However, despite becoming a core component of the “liberal international order,”Footnote 4 the landscape of international state building is increasingly fragmented. A wide variety of actors, including the United Nations (UN), the European Union (EU), the United States (US), and China, engage in state-building processes. Each may seek different objectives and use different implementation strategies. Moreover, while civilians are central to successful state-building efforts, they may interpret this fragmentation in a variety of ways, which may, in turn, affect whether they accept the (re)construction of state institutions by international actors.Footnote 5 Consequently, we ask: how does international state building vary, what preferences do civilians have toward the process, and are civilians unified in their preferences?
We answer these questions in two steps. First, we identify and explicate variation in international state building along three dimensions: the actor involved; the type of institution targeted; and the form of involvement. Second, we leverage an original conjoint experiment fielded in multiple rounds across three locations in postconflict Liberia to examine how civilians’ state-building preferences vary along each of these three dimensions. We find that civilians strongly prefer state-building processes overseen by the UN or the US over those administered by China; they also prefer state-building efforts performed by the Liberian government over those implemented by China. We also find that civilians prefer the (re)construction of security-oriented institutions over non-security-oriented institutions and that they prefer less direct forms of state-building engagement. Importantly, however, these results are not uniform across respondents: preferences vary by sex, socioeconomic status, and experiences of wartime violence. We explore this heterogeneity in greater detail by leveraging additional questions in our surveys.
Conducting an initial foray into understanding civilian preferences around international state building, we make several contributions. First, we provide researchers with a conceptual mapping of three primary dimensions of international state building. While any state-building process undoubtedly includes additional nuance, we maintain that these three features meaningfully capture variation in who engages in (re)building which type of state institution, and how they do so.
Second, we shift the focus of the state-building literature away from domestic elites and international actorsFootnote 6 and toward civilians. Civilians often have little say in state-building processes, yet their preferences are important for long-term peace, security, and stability.Footnote 7 While scholars have examined the conditions under which government institutions become legitimate,Footnote 8 less attention has been paid to civilians’ perceptions of the international state-building process. As in work that has looked at aid-donor characteristics,Footnote 9 we examine the public’s assessment of the various features that compose international state-building efforts. But we also join recent efforts to assess how variation in actor attribution impacts public perceptions of state institutions.Footnote 10
Third, we contribute to the literature on peacekeeping and international interventions by showing that there is local, micro-level support for third-party enforcement of various processes in conflict-affected countries.Footnote 11 Importantly, while our findings may qualify previous work suggesting that domestic actors often resist or undermine international interventions,Footnote 12 they emphasize that such support is not inevitable but depends on the actor, their form of involvement, and the type of targeted institution.
Finally, we also weigh in on discussions of the international engagement of illiberal and emerging powers.Footnote 13 Our findings suggest that, despite China’s increased investment in state-building projects around the world,Footnote 14 many civilians may ultimately prefer other actors, including their own government.
The Importance of Civilian Preferences in a Fragmented International State-Building Process
We define international state building as any “externally led social engineering reflected in efforts by the international community and individual states to rebuild states.”Footnote 15 While it may include a variety of processes, international state building includes actions aimed at creating, reforming, or strengthening state institutions,Footnote 16 with the primary goal of establishing a state that is “regarded as legitimate by the people over whom it exercises authority.”Footnote 17 Consequently, it usually takes place in weak and failing states—particularly those that have experienced severe violence or conflict.Footnote 18
Much of the literature understands international state building as emanating from one of two processes.Footnote 19 It may be provided to overcome the resources or enforcement restrictions that conflict-affected countries face,Footnote 20 or it may be imposed on societies as a model for building “legitimate” states under the “liberal international order.”Footnote 21 In both cases, it is treated as a unified, monolithic process.Footnote 22 However, the state-building landscape has become increasingly fragmented, with a wide array of actors involved in various ways in rebuilding a variety of state institutions. For example, alongside the UN and the US—the historically preeminent bilateral state builder—are other bilateral actors and regional state builders. They include Western actors, such as the EU, the Organization for Security and Cooperation in Europe (OSCE), and Sweden, as well as non-Western, illiberal actors, such as China and Russia.Footnote 23
Moreover, international state builders often use a variety of distinct, and at times divergent, state-building efforts.Footnote 24 For example, over the past twenty-five years, a variety of actors have sought to reform and strengthen different state institutions in Tajikistan through various initiatives. China and the US have provided financial resources and equipment to the country’s security (military and police) institutions to combat border insecurity, drug trafficking, and terrorism, while the OSCE has worked closely with Dushanbe to monitor and train officials in Tajikistan’s human rights and anticorruption institutions.Footnote 25
Though international state builders primarily engage with domestic elites,Footnote 26 understanding civilians’ perceptions of the increasingly fragmented international state-building process is important for several reasons. First, civilian preferences regarding this process may affect their overall perception of its outcomes—that is, whether they find state institutions rebuilt by international actors legitimate. While some research finds that international state building has largely failed or had minimal effect,Footnote 27 this literature often relies on measurements of “success” that are distinct from those tied to civilian perceptions. However, civilians do not always reject international state-building processes,Footnote 28 and there is no clear definition of what constitutes “success” for civilians, as they may understand it in a variety of ways.Footnote 29 Thus, we maintain that examining civilian preferences for how international state building occurs is important because civilians who support various components of these processes may find their outcomes more legitimate.
Relatedly, because a core goal of contemporary state building is the (re)establishment of the social contract between the state and its citizens,Footnote 30 failing to consider the latter’s preferences about the “terms” of the contract may undermine such efforts. In fact, the literature on constitution building emphasizes the importance of civilians’ attitudes in promoting public buy-in.Footnote 31 In this way, perceptions of the process of international state building may affect the degree to which the social contract is (re)established in the postconflict period.
Third, perceptions of the international state-building process may impact individuals’ political behavior and countries’ broader political trajectory. Citizens of postconflict Kosovo, for example, have taken to the streets in protest of the UN’s direct management of their country’s state-building process,Footnote 32 as well as the EU mission’s executive authority over various institutional reforms.Footnote 33 In extreme cases, such perceptions may even alter leaders’ election prospects. For example, the decision of Manasseh Sogavare, the former prime minister of the Solomon Islands, to withdraw from re-election in 2024 was partly due to domestic criticism of his close relations with Beijing and China’s focus on rebuilding public infrastructure over healthcare institutions.Footnote 34
Finally, paying attention to civilian preferences toward international state building helps promote “local ownership” of the process.Footnote 35 One of the main criticisms of international state building is that it ignores local needs and preferences.Footnote 36 Only by understanding these preferences can changes be made to the process that better align it with what local populations may want.
Components of the International State-Building Process
We have seen that civilians often hold preferences toward internationally backed state-building processes that are consequentially meaningful. However, we have not yet considered what features of these processes might lead to differing attitudes. Here, we explicate three primary dimensions along which international state building—and, consequently, people’s preferences regarding it—may vary: the actor involved; the target institution; and the form of involvement.
First, international state builders can vary according to type—that is, their status as either a multilateral or bilateral actor. Multilateral actors include international organizations, such as the EU, OSCE, or UN—each of which engages in state-building activities primarily through the missions they deploy to recipient countries.Footnote 37 Additionally, individual states may engage in state-building activities through bilateral programs. Moving beyond mere foreign aid, states engage in bilateral state building when their intervention involves a form of “shared sovereignty.”Footnote 38 This might include, for example, arrangements to increase the capacity of a postconflict country’s police or military force.Footnote 39 While the US is one of the largest bilateral state-building actors, many other countries have also been involved in such activities. For example, Australia has played a central role in postconflict police reform throughout Oceania,Footnote 40 while the UK has engaged in state-building interventions in Sierra Leone.Footnote 41
There is also meaningful variation within bilateral efforts. In particular, while Western states have historically been the preeminent bilateral state builders, non-Western countries, including China and Russia, have increasingly taken on a leading role. For example, over the past two decades, China has increased its state-building support in Africa, Asia, and Latin America through various infrastructure and security assistance programs.Footnote 42 Importantly, while Western states may seek to institutionalize liberal, democratic norms in their state-building programs abroad,Footnote 43 non-Western states like China neither actively export nor actively condemn antidemocratic actions in recipient states.Footnote 44 Their state-building practices may be more narrowly directed at furthering strategic national interests.Footnote 45 Nevertheless, such an orientation represents a clear distinction among bilateral state builders.
Second, actors may target different types of institutions in the state-building process. On the one hand, traditional state-building models suggest that expanding the administrative state is crucial, as a necessary first step in both building state capacity and increasing state legitimacy.Footnote 46 Indeed, in many historic state-building processes, administrative state institutions took precedence before other state institutions emerged, including security.Footnote 47
On the other hand, contemporary state-building projects often take place in weak, failed, or conflict-affected states. Consequently, while it may include the provision of various goods and services, a sequencing of state-building activities necessarily unfolds, with certain institutions taking a priority over others. As Krasner and Risse and Lake demonstrate, security provision often becomes a first-order concern within today’s postconflict state-building processes, as states with increased security capacities are largely understood to be less prone to armed conflict and criminal violence.Footnote 48 Thus, while any state-building effort may incorporate aspects of both, a meaningful distinction can be made here between security-oriented and non-security-oriented institutions. The former are mainly institutions like the military and police; the latter comprise a wider variety of administrative institutions, such as postal services, social services, and regulatory bodies.
Third, international state building may vary with respect to actors’ form of involvement.Footnote 49 At one end of the spectrum, an actor may be directly or heavily involved with a given state-building reform. This often arises when the actor has a degree of independent or executive oversight in the process. UN transitional administrations in Cambodia, East Timor, and Kosovo, for example, held managerial, decision-making authority in each country’s initial, postconflict, state-building period. Thus, they were responsible for designing and implementing a variety of institutional reforms in each country with comparatively less input from domestic actors.Footnote 50
At the other end of the spectrum, an actor may be indirectly involved in a state-building initiative, taking a more “hands-off” approach.Footnote 51 In these instances, for example, state builders may “monitor” or provide financial or technical assistance. For example, China has regularly provided Tajikistan’s security institutions with testing equipment and simulation training.Footnote 52 In such contexts, while state-building efforts are undoubtedly taking place, the international actors have less of a direct influence over the specific contours of the process.
Measuring Civilian Preferences for International State Building
Case Selection: State Building in Liberia
We leverage insights from postconflict Liberia to examine how variation along each of these three dimensions influences civilians’ preferences for postconflict state building. Liberia represents a “crucial” case for our analyses because it has undergone an extensive state-building process, with a variety of international actors involved in rebuilding a variety of institutions. Although Liberia has received assistance from many different international actors (including Canada, the EU, and Sweden), the UN, US, and China represent the main international statebuilders.
The UN has been deeply involved in Liberia’s state-building process since the end of the country’s second civil war, in 2003. With the adoption of the Accra Peace Agreement, the UN Security Council authorized the establishment of a multidimensional peacekeeping force, the UN Mission in Liberia (UNMIL).Footnote 53 The mission was involved in a variety of state-building tasks, including reformation of the judiciary and democratic institutions, but it was primarily involved in reforming the country’s security sector. For example, during its fourteen-year deployment period (2003–2018), UNMIL contributed nearly USD 5 billion and thousands of personnel toward reforming the country’s police force.Footnote 54
In contrast, the US has a long history of state building in Liberia. In 1822, the American Colonization Society established the Colony of Liberia as a colony of freed slaves on the African coast.Footnote 55 Fondly referring to Liberia as the US’s fifty-first state, many Liberians consider the US one of their country’s strongest allies.Footnote 56 Indeed, this “special relationship” has solidified over time through the consistent engagement of the US in security activities in the country and its commitment to (re)building Liberia’s security sector. For example, the US helped establish Liberia’s armed forces throughout the twentieth century. Since its civil wars, Liberia has been one of the main recipients on the African continent of security-oriented state-building assistance from the US.Footnote 57
China is the newest of these three state-building actors. Initially gaining a foothold in the country via personnel contributions to UNMIL, China now invests consistently in Liberia. Its state-building efforts focus on public infrastructure. For example, it has invested heavily in the (re)construction of roads, bridges, and ministerial and sports complexes in both the capital and rural areas. Moreover, China has also contributed to Liberia’s education system, by (re)constructing school and university buildings,Footnote 58 and to its security sector.Footnote 59
Though undoubtedly a crucial case, Liberia also represents a “difficult” test for our analyses. Studies have suggested that Liberia represents a generally successful case of postconflict state building, and such processes are often treated as homogeneous and unified.Footnote 60 Thus it may be more difficult to identify any meaningful variation in civilian preferences along each of the three dimensions. Consequently, if we find any significant differences in Liberians’ attitudes along each dimension of the state-building process, then we may expect variation to be observed in other postconflict countries, where internationally backed state building is often more divisive.
Research Design
We measure civilians’ preferences for each dimension of international state building using a set of conjoint experiments embedded in two rounds of representative surveys implemented in three locations across Liberia: Monrovia, Bong Mines, and Gbarnga (Table 1). We selected these locations because of their variation across several factors that may condition civilians’ experience of and position on postconflict state building: community type, exposure to wartime violence, and relative exposure to postconflict state-building processes.
Table 1. Relative empirical variation in the wartime and postconflict state-building experiences of three locations in Liberia
Note: Wartime violence is measured by battle-related deaths, from Gleditsch et al. Reference Gleditsch, Peter Wallensteen, Sollenberg and Strand2002.
The first round of the survey was conducted in January 2020. The second round, conducted in January 2022, replicates the conjoint experiment (with a slight modification, as described later) and was implemented to assess whether findings from the first round are robust across location and time (for example, before and after the COVID-19 pandemic declaration). Table 2 describes the two rounds.
Table 2. Survey locations, dates, and respondent totals
Survey Recruitment and Implementation
We sampled between 500 and 1,700 adult respondents in each location.Footnote 61 To obtain representative samples, we used a two-stage sampling procedure. Within each city/community, enumerator areas (EAs) were established according to common geographic boundaries. We adopted a stratified sampling procedure to select the number of respondents in each EA proportional to the total population in each location, according to Liberia’s most recent census. From there, enumerators engaged in a random-walk procedure to select a dwelling for respondent selection: after starting in each EA’s community market or central plaza, the enumerator randomly selected a street to walk down, alternating on each street between selecting the third house on the right and the third house on the left. Once a dwelling was identified, the enumerator selected a random adult to interview using a Kish table. Following each interview, the enumerator continued to the end of the street and repeated the random selection process.
The surveys were administered face to face in the first language of the respondent and by enumerators of the same sex as the respondent. Each survey interview took about fifty minutes to complete, beginning with introduction and consent and ending with a debriefing. Each survey included additional questions on baseline characteristics, including income level, sex, and exposure to wartime violence.
Research Ethics
Approval from an institutional review board (or the local equivalent) is the minimum requirement to meet our institutions’ standards for research with human participants.Footnote 62 However, the sensitivity of the survey and the nature of the face-to-face interviews warrant additional ethical considerations.
First, we understand that postconflict state-building processes may be highly contentious; having enumerators ask individuals throughout a community about the topic might raise suspicion and fears, making them—and willing participants—a potential target. On top of the social dimension are the physical risks inherent in survey enumeration. While the risks that enumerators and respondents face cannot be completely eradicated, it is our duty to eliminate the ones we can and minimize the others. We sought to do so in three ways. First, where possible, enumerators were instructed to notify and receive approval from local leaders before conducting interviews. Second, each enumerator was paired with a member of the opposite sex in a single EA; this not only allowed the team to ensure that the enumerator and their respondent were of the same sex but also guaranteed that no enumerator was traveling alone. Third, we provided insurance for our research team so that we could provide medical care in the case of injury during the survey.
We also understand that unforeseen circumstances can put individual enumerators, or more generally the team and project, in danger. Consequently, we worked with the local survey team to construct a general safety plan outlining various conditions under which the project would be paused or terminated. In addition, as a team, we checked in with each other daily to ensure that each enumerator was safe and secure. Finally, seeking to minimize the gap between the ethical standards to which academic institutions adhere in the Global North and those in place in other regions of the world, we also obtained approval from the University of Liberia’s Institutional Review Board.
Structure and Modeling of the Conjoint Experiment
As noted, we use a choice-based conjoint experiment embedded in each of the surveys.Footnote 63 Although additional care must be taken when interpreting their results,Footnote 64 conjoint experiments are useful over other survey experiments because they allow us not only to directly control for exposure to each independent variable of interest but also to test several causal relationships present across the hypotheses simultaneously.Footnote 65 Moreover, these factors more accurately represent real-world decision-making processes than other survey experiments,Footnote 66 while also allowing individuals to respond to sensitive aspects of each topic more discretely, thus more accurately capturing their true preferences.Footnote 67
The attributes in the conjoint experiment correspond to the primary dimensions of international state building, and each feature therein represents the operationalization of the corresponding dimension. The first dimension is the actor involved—here the UN, the US, or China. We choose this set for two reasons. First, we include a variety of actors both to examine how such variation may shape state-building preferences and to minimize any biases we may find if we rely on a single actor. Second, we include these three actors because of their prevalence in Liberia. For a baseline, we use the recipient state: the Liberian government.
The second dimension is the type of institution. As noted, our primary distinction is between security-oriented and non-security-oriented institutions. For the former, we specify a “special forces” team: the Armed Forces of Liberia Special Forces Team. This operationalization represents an institution aimed at promoting the physical security of civilians as it demonstrates a clear strengthening of the country’s armed forces. For the latter, we use either the Liberian Postal Service (round 1) or the Liberian Land Authority (round 2). These are examples of two common state-building institutions: postal services and land-granting authorities.Footnote 68 Consulting with the local research team and country experts, we opted for the post office in the first round because of its role in the bureaucratic expansion of states.Footnote 69 In the second round, we chose the land authority because of its role in representing the legal expansion of the state. Given the importance of land issues in Liberia,Footnote 70 the land authority represents a potentially more “desirable” non-security-oriented institution and thus a harder test for any distinctions between security and nonsecurity institutions. Nevertheless, each institution represents fundamental services that a contemporary state is meant to provide and therefore is central to postconflict state-building processes.Footnote 71
Crucially, we ensured that it was plausible that each international actor engaged with each institution through each form of involvement. For example, while China has been primarily involved in expanding non-security-oriented institutions in Liberia, it has also helped reform the country’s security sector through cooperative training and technology provision.Footnote 72 At the same time, while the UN and US have been heavily involved in security-sector reform, they have also been involved historically in building the capacity of non-security-oriented state administrations.Footnote 73
The third dimension is the form of involvement in the state-building process. To capture more direct forms of state-building involvement, we use the phrase “has rebuilt.” To capture indirect, limited participation, we use the phrase “has monitored the reconstruction of.” While this dimension of international state building undoubtedly operates on a spectrum, we rely on these operationalizations for three reasons: to maximize the clarity of distinction; to match measurement decisions in recent studies;Footnote 74 and to contextualize the experiment following our consultation with the local research team. The overall design of the conjoint experiment is presented in Table 3.
Table 3. Outline of the conjoint experiment
The experimental portion of the survey began with the enumerator reading a short introductory statement on the task to be completed:
I would now like to provide you with a few statements of recent developments in Liberia. For each set of two statements, please identify which you find more favorable.
Enumerators then read two randomly assigned combinations of the conjoint experiment to each respondent and asked which he or she preferred. This selection process (“task”) was completed three or four times by each respondent, with a new combination provided each time.Footnote 75 For example, one possible combination of statements was:Footnote 76
[1] The Liberian Government has rebuilt the Armed Forces of Liberia Special Forces Team.
[2] The United States has monitored the reconstruction of the Liberian Postal Service.
Following Leeper, Hobolt, and Tilley, we present the results derived from the marginal mean estimands for each of the attribute levels.Footnote 77 Because the conjoint experiment was forced-choice with two profiles (state-building events) in each task, values above 0.5 indicate an attribute level that increases the favorability of a profile over another, and values below 0.5 represent an attribute level that decreases it.
We expect each attribute to have an additive effect on each respondent’s preference. Thus, in a standard regression framework, the probability that respondent i selected a certain state-building event S can be written as
Pr(S)i = actori + institutioni + involvementi + ϵi
Because participants were asked to complete the conjoint experiment multiple times, we include robust standard errors clustered by respondent.
Results and Discussion
Figure 1 presents the primary results of the conjoint experiment, while Figure 2 presents the heterogeneous effects across three key demographic characteristics: sex, socioeconomic status, and experience of wartime violence. In this section, we highlight and expand on several key findings from the analyses.Footnote 78
Figure 1. Marginal means of attribute levels
Note: Horizontal bars represent 95% confidence intervals; standard errors are clustered by respondent.
Figure 2. Marginal means of attribute levels by subgroup
Note: Horizontal bars represent 95% confidence intervals; standard errors are clustered by respondent.
Results for Actor
Regarding the state-building actor, we find that civilians prefer the US, the UN, and the Liberian government over China.Footnote 79 The estimates for the former, ranging from 0.52 (Liberian government) to 0.546 (US), are positively and significantly associated with support for a given state-building process. This is particularly true for the US, with an estimand that is statistically distinct from that of the Liberian government. In contrast, China’s involvement is negatively and significantly associated with a given state-building process (an estimate of 0.434).
These results both align with and diverge from earlier scholarship. They substantiate recent scholarship by highlighting that civilians rarely hold neutral opinions on international actorsFootnote 80 and reinforce other studies that find that donors’ characteristics shape public support.Footnote 81 Our results are also consistent with studies that find no meaningful increase in support for China resulting from its assistance but a meaningful increase in support for the US under such conditions.Footnote 82 The distinct impact of the US may be driven by its long-standing and close history with Liberia.Footnote 83 This would imply that individuals who have closer ties to the US, such as those of a higher socioeconomic status,Footnote 84 should favor the US as a state builder. Indeed, wealthier individuals are more likely to support US-backed state-building processes (Figure 2).
At the same time, we find the results somewhat surprising. Research has demonstrated a strong favorability toward all international donors over domestic governments when it comes to aid assistance.Footnote 85 These findings also implicitly contrast with past work showing the various benefits that come with Chinese state-building and development assistance.Footnote 86 Relatedly, the high visibility of Chinese state-building projects in the country should imply more favorability, as scholarship finds that exposure to projects—especially completed projects—consistently improves donor perceptions.Footnote 87 Indeed, Blair and Roessler find that Liberians’ exposure to Chinese aid and investment improves their perceptions of Chinese donors and investors.Footnote 88 However, our findings demonstrate that respondents held relatively clear opinions of each international state builder included in the experiment (see Figure A2 in the appendix). This suggests that our results are not driven purely by respondents’ disproportionately low exposure to China’s previous activities in the country.
We find that variations in sex and socioeconomic class do not meaningfully predict opposition to China’s involvement in the state-building process (Figure 2).Footnote 89 Consequently, we explore three other potential explanations for these findings. First, it may be the case that people simply hold less positive views on China’s efforts in Liberia. Figure 3 suggests that this may be the case. In particular, additional survey results (round 2) demonstrate that respondents find China to be less helpful for Liberia than the UN and the US. We also find that respondents are disproportionately more likely to trust the UN and the US, as compared to the Liberian government (see Figure A4 in the appendix).
Figure 3. Liberians’ views on international state builders
Note: The bar plot shows the distribution of respondents’ answers to the question, “How helpful or unhelpful is [X] for Liberia?”
Second, the negative impact of China on respondents’ state-building preferences may be driven by normative considerations. Previous work has demonstrated that civilians are more likely to support donors who are more democratic and transparentFootnote 90 because transparency and accountability promote trust.Footnote 91 This means that Liberians with more authoritarian views may be more willing to support Chinese state building. Testing this proposition directly would introduce post-treatment bias, but we see some evidence for it by relying on other questions in the survey. Those who score higher on the standard authoritarian scale are more likely to trust China (see Table A17 in the appendix), suggesting that they might also prefer its state-building activities over other actors.
Finally, recent studies have also found that Chinese assistance, in particular, may stoke division—increasing (real or perceived) levels of corruption, political patronage, or intergroup conflict among recipient populations,Footnote 92 as well as the protest rate of local populations.Footnote 93 Victims of wartime violence may be particularly sensitive to these concerns and therefore opposed to developments that may further create political and social cleavages. Indeed, we find that those who experienced particularly severe forms of wartime violence are even less likely to prefer Chinese-backed state-building processes (Figure 2). This trend is supported by regression analyses that find those with knowledge of armed groups are less likely to trust China (Table A17 in the appendix).
Results for Type of Institution
Turning to the type of institution, we find that Liberians have a clear preference for state-building activities that involve the security-oriented institution (Armed Forces of Liberia Special Forces Team) over the non-security-oriented institutions (postal service/land authority). With a value of the former at 0.546 (standard error of 0.003), the estimand is eight percentage points—and twenty-four standard deviations—greater than that for the latter (0.466). Moreover, the estimands operate in different directions: the impact of the Armed Forces of Liberia Special Forces Team is positive and statistically significant, while the impact of the postal service/land authority is negative and statistically significant.
These preferences align with research suggesting that the reconstruction of security-oriented institutions should be prioritized to promote lasting peace.Footnote 94 This distinction appears within public opinion in other contexts, as recent work has also found that civilians place a premium on security in the postconflict period.Footnote 95 Indeed, the government of Liberia has prioritized funding the security sector, with funding increasing each year since 2019.Footnote 96
However, the literature and our findings raise important questions regarding the conditions under which civilians in postconflict countries may prioritize non-security-oriented institutions. Here, the heterogeneous effects suggest that certain populations are less likely to prefer security institutions. First, women are less likely to prefer the Armed Forces of Liberia Special Forces Team to the nonsecurity institutions. This is somewhat unsurprising, given that women across the world are less likely than men to prefer securitized policies.Footnote 97 Women may also be less willing to prefer the rebuilding of special forces because these types of groups are often involved in sexual and gender-based violence.Footnote 98
Women may also prefer the (re)construction of administrative institutions over security institutions for other reasons. Increasing administrative capacity (for example, via post offices that issue identification cards) makes women more legible to the state,Footnote 99 and land institutions help civilians acquire land titles, which are crucial for women’s economic empowerment.Footnote 100 Thus, some of the variation by sex could be driven by a preference for improvements in the administrative capacity of states as an avenue for their empowerment.
Second, we find that higher-income individuals prefer security-oriented institutions. This is consistent with findings that higher-income individuals prefer stronger policing and security measures.Footnote 101 In contrast, lower-income civilians may be more likely to benefit from an expanded administrative state and see less benefit from expanded security forces, perhaps because they are disproportionately likely to be targeted by such institutions.Footnote 102
Finally, we find meaningful variation in respondents’ state-building preferences based on their wartime-victimization status. Those who experienced a mild degree of community-level wartime violence are less likely to prefer the security-oriented institution, as compared to those who did not experience any form of violence (Figure 2). Although this may seem puzzling, recent work has found that those who experience a minimal level of wartime violence may be particularly supportive of welfare-oriented institutions, as such experiences may increase prosocial behavior.Footnote 103 Nevertheless, the relationship may be parabolic because those who are on the extreme ends of the wartime-victimization scale—that is, “none” and “severe”—favor security-oriented institutions at similar rates.
Ultimately, our findings along this second dimension of state building suggest that civilians in Liberia have a strong preference for the (re)construction of security-oriented institutions. However, not all civilians share this preference, with important differences across sex, socioeconomic status, and wartime victimization.
Results for Form of Involvement
Turning to the third dimension of international state building, we find that civilians prefer that state builders engage in more limited, indirect forms of state building rather than direct forms. The estimand for the former (here, “monitored the reconstruction of”) is 0.512 (Figure 1). This is not only positive and statistically significant but also distinct from the estimand for the latter (here, “has rebuilt”): 0.487.
As with the other two dimensions, these results align with previous work on the effectiveness of international state-building efforts. Research has shown that state-building efforts that bypass domestic actors to impose reforms often fail because they lack input legitimacy.Footnote 104 This also appears to be the case in situations of delegated authority: international state builders that engage in negotiated forms of influence—instead of “complete takeovers”Footnote 105—allow domestic oversight, inclusion, and thus ownership over the process.Footnote 106 In this way, the results suggest that civilians prefer forms of state building that implicitly necessitate a degree of domestic involvement in the process.
At the same time, these significant results we observe in the aggregate may be primarily driven by wartime victimization. The primary difference we observe is for individuals without any community-level wartime victimization (Figure 2). While these individuals may hold particularly strong attitudes toward each form of involvement, victims of wartime violence may place less emphasis on how state building takes place. This may be determined by priorities similar to those of low-income individuals—that is, a focus on what reforms are made, rather than how. Alternatively, it may be driven by a desire to ensure that such processes are successfully executed. Recent work suggests that more direct forms of participation are better at ensuring that reforms are implemented,Footnote 107 and victims of wartime violence may also prefer that decisive action be taken in periods of postconflict transition.Footnote 108 Thus, they may be relatively less willing to prioritize a limited form of actor involvement in the state-building process.
In sum, we find that civilians have strong preferences for more indirect, limited forms of international state building. However, as with the other dimensions, there can be meaningful heterogeneity across demographic subgroups, with victims of community-level wartime violence less likely to hold polarizing positions on the form of actor involvement.
Conclusion
Postconflict countries often receive state-building assistance from international actors, who engage in various institutional-reform initiatives through different modes of engagement. It is crucial to understand civilian preferences in this fragmented international state-building environment because these preferences not only impact state-building outcomes but also are central to local ownership processes.Footnote 109
We explicate three primary dimensions along which international state building may vary—the actor, the targeted institution, and the form of involvement—and we leverage a series of original surveys fielded across Liberia to assess how variation along each dimension impacts civilians’ state-building preferences. The findings provide a handful of meaningful insights.
First, Liberians’ preferences do indeed vary along each of these three core dimensions, but they also hold significantly consistent preferences for each dimension. Specifically, respondents strongly prefer indirect state-building processes, targeted at security-oriented institutions and overseen by the US, the UN, and (to a lesser extent) the Liberian government, rather than China. Regarding the latter, though our work is exploratory, we find suggestive evidence of consistently negative attitudes toward Chinese-backed state-building processes. Our supplementary analyses suggest that this latter finding may be driven by both individuals’ subjective perceptions of the country and their objective experiences of wartime violence.
Second, Liberian civilians strongly prefer security-oriented state-building processes over non-security-oriented processes. However, this polarization is weaker among women, lower-income individuals, and those with less exposure to wartime victimization.
Third, civilians generally prefer indirect, less involved forms of international state building over more direct, hands-on forms, though individuals without any community-level exposure to wartime victimization may be driving these results.
Taken together, these results offer researchers, policymakers, and practitioners meaningful insights into civilians’ preferences for international state building. They suggest that wartime victimization influences individuals’ preferences regarding postconflict state-building processes. This has important implications for long-term state-building efforts, as it suggests that wartime experiences continue to shape individuals’ political preferences for decades afterward. Such legacies should therefore be prioritized when considering how to engage in state building.
Our findings also support a soft, cautious optimism for internationally backed state-building efforts. While an emerging body of research has, importantly, drawn attention to the challenges that state building faces,Footnote 110 we find that, by and large, civilians in Liberia do desire international state-building assistance but also that such support is substantially conditioned on which features unfold in the process. This both reinforces and qualifies such scholarship. With this in mind, the question becomes not whether internationally backed state building is productive but rather how it may be. Consequently, these results from Liberia, as well as our explication of each dimension of the state-building process, serve to move these conversations forward.
Nevertheless, this study is not without limitations. Our results may be influenced by how we operationalize each dimension of state building and by country-specific features. While the UN, US, and China are leading state builders, other international actors, including Australia, the EU, Sweden, and the UK, are increasingly involved in the postconflict state-building process;Footnote 111 state building may also target a wider variety of institutions, including the judiciary and election institutions. Though our focus provides important theoretical and empirical insights, future research should explore how additional variation along each dimension may shape civilian preferences. This may be done, for example, through quantitative replications of our experiment using different institutions and actors elsewhere or through qualitative case studies.
Moreover, our analyses treat civilians’ preferences as static. While this is useful for an initial conceptual and empirical exploration, state building is a dynamic process, and civilians’ preferences may change over time, especially as international actors become embedded in such contexts.Footnote 112 Future studies should therefore examine how and when such preferences change to provide a more complete picture of the relationship between public opinion and and internationally backed state-building processes.
Acknowledgments
We thank participants of the Building State Legitimacy in Post-Civil War States workshop at the 2018 Peace Science Society (International) annual meeting, Cornell University’s Spring 2021 Comparative Politics workshop, and workshops at Washington University in St. Louis and Duke University for their feedback on earlier versions of this manuscript. We also thank Franklin Zheng for his research assistance. Finally, we thank the anonymous reviewers and editors of International Organization for their thoughtful feedback, suggestions, and questions.
Funding
This research was supported by grants from the UK’s Economic and Social Research Council (ES/S009590/1) and the Folke Bernadotte Academy (DNR 17-00425).
Data Availability Statement
Replication files for this research note may be found at <https://doi.org/10.7910/DVN/IES0FO>.
Supplementary Material
Supplementary material for this research note is available at <https://doi.org/10.1017/S0020818325100829>.
Open access
