Courtesy of Oxford Academic, an interesting new paper examining why countries join or avoid China’s BRI:

Abstract

Although China’s motives for developing the Belt and Road Initiative (BRI) have been well studied, scholars have yet to comprehensively examine why states seek to join the initiative. We fill this gap by examining how and why states join the BRI. Countries join by signing a Memorandum of Understanding (MOU) with China on cooperation under the BRI framework. These MOUs create few or no obligations for the states who sign them but increase the possibility of reaping future economic benefits. Thus, we argue that most states should join the BRI unless they view the costs of participation as higher. We hypothesize, and find support for, the argument that democracies are less likely to join because they view participating in a Chinese-led initiative as more costly than non-democracies. Our statistical analysis using a new dataset of BRI participants and paired case studies provides quantitative and qualitative support for this argument.

China’s “Belt and Road” is Xi Jinping’s signature foreign policy initiative that has attracted substantial international attention. The initiative is associated with up to $1 trillion of infrastructure investment and has also generated concerns that it may increase China’s influence in the international system or contribute to indebtedness or subpar governance, especially among developing states. Politically, China has attached great importance to the Belt and Road Initiative’s (BRI) success, incorporating it into the charter of the Chinese Communist Party, while the United States has vigorously opposed it.

The BRI’s impact will occur through the relationships that China builds with states who join. Through 2021, 141 countries have formally joined the initiative, or roughly two-thirds of states in the international community. To provide a more complete understanding of the BRI, this article investigates how and why states choose to join China’s initiative. Although the literature on the BRI is now vast, much of it examines China’s motivations or intentions for creating the initiative (e.g., Wuthnow 2017). Scholars have provided in-depth accounts of specific states who join (e.g.,Markey 2020), but no comprehensive analysis of why states join exists. Moreover, widely varying measures of participation have been used, leading to confusion about who has joined, who has not, and the benefits (or lack thereof) of participation.

We seek to make three contributions to the study of the BRI. First, we highlight the existence of a consistent, universal, and officially sanctioned measure—the Memorandum of Understanding (MOU) on Jointly Promoting the Construction of the Silk Road Economic Belt and the 21st Century Maritime Silk Road—that the Chinese government uses to recruit states to join the BRI. This measure captures the political willingness of states to join the BRI but is impartial to economic outcomes, allowing for a complete and unbiased assessment of the BRI’s appeal and impact compared to other indicators.

We review the content of these MOUs to better understand the terms by which states join and find that BRI-MOUs create few or no obligations for participating states. They are legally non-binding, lack concrete commitments, and include provisions for non-renewal. In other words, countries incur few costs by signing these MOUs but gain the possibility of receiving massive future economic benefits, however vague.

The content of these MOUs suggests a clear but counterintuitive argument for when countries should join the BRI: Because the initial costs of formal BRI participation are low and the potential economic benefits are high, countries should join the BRI unless they view the costs of signing an MOU as higher for reasons beyond economic cooperation. In other words, the strongest predictor variables of participation should be those that are correlated with the decision not to formally participate in the BRI. Specifically, we investigate whether democracies and states with close security ties with the United States view the costs of joining the BRI as higher and are thus less likely to join than other states.

Second, utilizing BRI-MOUs as the primary indicator of participation, we construct a universal dataset that includes both BRI participants and non-participants and use it to compare the characteristics of states who have joined against those who have not. In doing so, we identify the determinants of BRI participation and evaluate existing explanations for why states join, such as demand for infrastructure, political affinity with China, and corruption in participating states.

Statistical analysis provides support for our intuitions. There exists no robust relationship between economic or alignment covariates and BRI participation, such as infrastructure needs or political affinity. The analysis reveals a strong and statistically significant relationship between regime type and non-participation: Democracies are much less likely to join than non-democracies. Although states with pre-existing security ties with the United States are also less likely to join, this result lacks statistical significance.

To further examine the impact of democratic governance on states’ decisions to join—or, more likely, not join—the BRI, we conduct two pairs of illustrative case studies in which states are matched on all covariates except our variable of interest, democracy. In comparing the BRI decisions of Turkey and Spain, as well as Fiji and Mauritius, we are able to show how democracy appears to exert an effect on participation and non-participation in China’s BRI. We also examine Italy as a clear outlier to our argument.

Finally, studying state participation in the BRI creates an opportunity to examine whether and how the BRI may increase China’s international influence. Although some scholars emphasize the importance of domestic factors in China’s development of the BRI, such as outsourcing surplus capacity (Ferdinand 2016; King and Du 2018; Ye 2018), many others view the BRI as a new tool of statecraft intended to increase China’s influence (Zhou and Esteban 2018; Benabdallah 2019; Kliman et al. 2019). Some scholars even describe the BRI as a grand strategy (Fallon 2015; Rolland 2017; Zhou and Esteban 2018).

Is the BRI an example of China’s “debt-trap diplomacy” or a modern-day version of the Marshall Plan? Many arguments rest on the experiences of just a few states, raising concerns about their overall representativeness. Finding a consistent measure of participation and testing determinants of participation across all states in the international system is an important first step toward constructing externally valid theories of the BRI’s influence. Additionally, China may be able to exercise influence through the BRI at different stages, including when states first join or later on through subsequent economic interactions under the BRI. Examining the process and terms by which states join the BRI can help clarify whether the basis of formal participation creates commitments that would increase China’s influence over states who join at the outset. In other work, we examine how BRI participation shapes subsequent economic interactions with China.

The rest of the article proceeds as follows. First, we provide a brief overview of the BRI’s origins and development. Second, we deconstruct our dependent variable of interest—formal BRI participation—and highlight the central role of BRI-MOUs. Third, we introduce our argument, which is based on an analysis of the initial costs of signing a BRI-MOU. Fourth, we present our dataset and test the relationship between several covariates and BRI participation using survival analysis. Fifth, we conduct several qualitative case studies to examine our hypothesized causal mechanisms in detail. We conclude with a discussion of our findings and their implications for future research, especially for China’s potential to exercise influence through the BRI.

Origins and Development of the BRI

Launched in 2013, the BRI is an ambitious global connectivity initiative that seeks to address a gap in infrastructure funding in developing economies too wide to be filled by multilateral development banks (OECD 2018). More than US $200 billion has been directly invested through BRI projects, and some estimates place China’s total investment at above $1 trillion by 2027, making the BRI the largest developmental initiative ever established (Dossani, Bouey, and Zhu 2020).

Xi Jinping first announced the BRI in a pair of speeches given during trips to Kazakhstan and Indonesia in September and October of 2013, respectively. The “Silk Road Economic Belt” refers to overland links through Central Asia, while the “21st Century Maritime Silk Road” denotes the maritime routes passing through Southeast Asia and the Indian Ocean. Though Xi’s initial speeches did not reveal concrete proposals for the BRI’s implementation, they did project a vision that went beyond a basic infrastructure plan to forge a “community of interests” through “win–win outcomes” and “mutual trust.” Xi’s speeches signaled China’s willingness to engage economically with its neighbors while respecting the development paths “independently chosen” by their people (i.e., refraining from criticism of their political systems) (Xi 2013a, 2013b).

More concrete plans for the BRI emerged in 2015, when the National Development and Reform Commission (NDRC), Ministry of Foreign Affairs (MFA), and Ministry of Commerce (MOFCOM) jointly published “Vision and Actions on Jointly Building Silk Road Economic Belt and 21st Century Maritime Silk Road” (hereon Vision and Actions). This document is the most authoritative, publicly available government document on the BRI to date (“Vision and Actions on Jointly Building Belt and Road” 2015). Vision and Actions largely reaffirmed Xi’s proposal, providing a more detailed explanation of the “Silk Road spirit” and laying out cooperation mechanisms with participant countries. The document also defines the BRI’s geographical scope, which consists of the Silk Road Economic Belt, the Maritime Silk Road, and six major corridors passing through Asia, Europe, and Africa.

Vision and Actions outlines five “cooperation priorities” (hezuo zhongdian) for the BRI, which frame its implementation. The first is “policy coordination” (zhengce goutong), the promotion of macro-economic policy, development strategy, and large-scale project coordination through the establishment of bilateral exchange mechanisms. The second is “facilities connectivity” (sheshi liantong), the establishment of physical infrastructure and cooperation in areas like technical standards, and civil aviation. The third, “unimpeded trade” (maoyi changtong), includes increasing customs cooperation, lowering tariffs, eliminating investment barriers, and encouraging technical cooperation. The fourth is to promote “financial integration” (zijin rongtong) through the expansion of bilateral currency swaps, renminbi bonds, and the establishment of a “regional financial risk early warning system.” Finally, increased “people-to-people bonds” (minxin xiangtong) include the creation of scholarships, the expansion of tourism, the strengthening of medical and epidemic cooperation, and the leveraging of “new media tools” to foster a “harmonious” environment.

Vision and Actions also details mechanisms for implementing the BRI, such as pilot projects, working groups, and implementation roadmaps. Specifically, the report encourages MOUs as a mechanism by which states should participate in the initiative, stating that “we [Chinese diplomats] should encourage the signing of cooperation MOUs or cooperation plans” (hezuo beiwanglu huo hezuo guihua). Figure 1 shows the number of countries that have formally joined the BRI since 2013.

Figure 1.

Formal BRI participation by year.

Open in new tabDownload slide

The next milestone in the BRI’s development was the first Belt and Road Forum (BRF), held in Beijing in May 2017 and attended by twenty-nine foreign heads of state (Darby 2017). Later in 2017, at the 19th Party Congress, the BRI was incorporated into the Chinese Communist Party’s charter. Finally, as evident in figure 1, the rate of BRI participation accelerated in 2018. In 2019, a second BRF was held with a roster of thirty-seven world leaders and delegates from 150 countries.

How Countries Join the BRI: MOUs

This section identifies and justifies MOUs as a consistent, universal indicator for measuring BRI participation. It then reviews the contents of these MOUs and argues that they are designed to lower the initial costs of joining for potential participants.

Why MOUs

In international law, MOUs are non-binding agreements between parties that provide a framework for cooperation and understanding (von Glahn and Taulbee 2015). They typically identify the terms of an agreement and the expectations of the parties involved. Although they are non-binding, because MOUs clearly articulate intentions of the signatories, they can help build trust and foster cooperation, and often serve as a first step toward more formal, binding agreements in the future.

Previous scholarship has used other measures for BRI involvement, such as attendance at BRI forums, participation in the Asian Infrastructure Investment Bank, and MOUs (Nolan and Leutert 2020; Serrano Moreno, Telias, and Urdinez 2021). We use MOUs as the primary measure for participation for three reasons: (1) The Chinese government has highlighted MOUs as the means by which states join the initiative; (2) a completed MOU reflects mutual willingness by both sides to cooperate under the BRI; and (3) most importantly, these MOUs reflect initial political interest but not subsequent economic cooperation.

First, Visions and Actions highlights MOUs as the appropriate policy tool for enlisting states to join the BRI. Because MOUs are officially sanctioned and encouraged by China’s central government, MOUs have been adopted universally across all countries. As listed in the online appendix, China signed BRI-MOUs with 141 countries between 2013, when the initiative was announced, and 2021.¹ In the BRI’s first 2 years, these MOUs were signed by the PRC’s Ministry of Commerce and an equivalent partner agency (e.g., a Ministry of Economy). Since 2016, however, when the NDRC became the lead implementing agency, all MOUs have been signed by representatives of the NDRC or MFA on the Chinese side and by heads of state or chief diplomats in participating states.

Second, because MOUs are signed by both sides, a completed MOU reflects mutual desire to pursue cooperation. Because an MOU must be negotiated and agreed upon between the two parties, completed MOUs are a useful indicator of a state’s participation in the BRI because both sides have explicitly considered and agreed to the terms of the MOU. This helps avoid confusion as to whether or not a country has actually joined the BRI.

Third, MOUs precede substantive cooperation between China and other states, reflecting an intent to cooperate but not cooperation itself. MOUs are frameworks for future cooperation, whereas economic contracts are binding agreements between two parties that set out the terms and conditions of a business relationship. Within the context of the BRI, the initial political decision to cooperate—captured by an MOU—must be analytically distinct from subsequent economic engagement under the name of the BRI.

Subsequent economic cooperation under the BRI is a poor indicator of formal participation for several reasons. First, no complete database of BRI funding exists that can distinguish general economic cooperation with China from BRI-specific cooperation. One database that analysts often use, the China Global Investment Tracker, simply counts all funding a state receives from China after signing an MOU as BRI funding, even if the state received significant amounts of loans or investments before joining. Additionally, given the importance China’s political leaders have attached to the BRI, Chinese actors have an incentive to characterize projects as belonging to the BRI even if the counterparty does not.

Moreover, measuring participation using economic indicators of cooperation (e.g., signed contracts) would overlook those countries that are willing to participate but never receive the economic benefits of cooperation. In other words, measuring BRI participation using foreign aid or loans for construction projects would overstate or misrepresent the initiative’s impact by assuming that either one automatically accompanies BRI participation, which is not true. Using an indicator that reflects political interest but precedes actual economic cooperation avoids this form of selecting on the dependent variable.

By using MOUs as our dependent variable, we privilege a political and economic framing of the BRI over others (e.g., environmental or educational), which may shape our understanding of the initiative and its appeal (Lesch and Loh 2022). Nevertheless, research demonstrates that there are genuine economic (Scobell et al. 2018; Ye 2018; Lai 2020) and political (Wuthnow 2017; Clarke 2018; Zhou and Esteban 2018) motivations underlying the BRI’s creation. Participation has economic (Foo, Lean, and Salim 2020; Hillman 2020; Wang et al. 2020) and political (Rolland 2017; Gong 2019; Sutherland et al. 2020) effects for states as well. Thus, it is reasonable to assess participation through this lens to study which economic and political characteristics most determine state participation.

In sum, MOUs are the official, state-sanctioned measure by which states join the BRI. Unlike other indicators, MOUs capture a state’s political willingness to participate without selecting the dependent variable by using measures of realized economic cooperation. As such, we believe they are the most valid indicator of BRI participation.

Content of BRI-MOUs

This section reviews the characteristics and content of BRI-MOUs and reaches two important conclusions. First, MOUs focus on general principles of cooperation and not specific BRI projects or actions. Essentially, they are best viewed as signals of intent to cooperate rather than evidence of realized cooperation. China can claim to add a country to the BRI, and the partner gains the potential of greater Chinese investment and access to China’s economy. Second, MOUs are designed to be low-cost agreements, thus encouraging more countries to join. They are non-binding and do not legally commit China or signatories to specific actions. Most are also applicable for a limited duration of time (generally 3–5 years) and contain provisions for non-renewal. These characteristics support our overarching argument—that the costs of BRI participation are usually low, at least at the outset when an initial political decision to participate is made.

Our analysis is based on the full text of twenty-six MOUs. Descriptive statistics (in the online appendix) demonstrate that these countries are slightly wealthier, slightly more democratic, and trade slightly more with China than BRI participants in general. However, these documents span several regions, capture varying degrees of financial integration with China, and involve states with different political systems. More importantly, the content and format of MOUs are strikingly similar across documents.

Recycled language is common in agreements between states: Low-capacity governments use copied text to increase their reach and efficiency, while powerful states use it to propagate preferred norms. Bureaucratic culture and standard operating procedures also play a role in replicating language (Allee and Elsig 2019). In China’s case, the similarity between MOUs likely reflects the centralized, top-down nature of foreign policy execution as well as China’s desire to normalize its preferred norms of bilateral interaction in its relationships with other states. Given the similar content and format of MOUs across states and the highly centralized nature of China’s bureaucracy, we have little reason to believe that the MOUs we have collected are not generally representative. In comparing the MOUs obtained from Sudan and New Zealand, for example, we do not find substantial differences that would warrant caution in extending our analysis more generally.

China’s BRI-MOUs follow a common structure, with sections addressing: (1) the principles and objectives of cooperation, (2) areas of cooperation, (3) cooperation mechanisms, (4) the settlement of differences, (5) legal status, and (6) provisions for renewal and termination. The standardized format suggests that they are based on language proposed by the Chinese government, which forms the starting point for negotiation. Few, if any, changes are made if the partner country has no objections.²

1. Principles and objectives: The first section of the MOUs concerns the principles and objectives of BRI cooperation. MOUs generally describe the principles of cooperation as “extensive consultation, joint contribution, and shared benefits” (gongshang, gongjian, and gongxiang). However, there are also points of heterogeneity between MOUs. Some explicitly acknowledge existing commitments that must be respected. For example, Thailand’s MOU references a need to respect “domestic laws and regulations,” while the Philippines’ highlights the importance of “mutual respect for territorial integrity and sovereignty.”³

2. Areas of cooperation: The areas of cooperation in MOUs largely mirror those outlined in Vision and Actions. They are as follows: (i) policy coordination, (ii) facilities connectivity, (iii) trade, (iv) financial integration, and (v) people-to-people ties. Each area is usually discussed in a general language without reference to details. For example, Montenegro’s MOU anticipates “cooperation in infrastructure. . . such as ports, port terminals, highways, railroads, logistics centers and industrial parks.”⁴ Such ambiguity underscores the fact that MOUs reflect an initial interest in cooperation but not cooperation itself.

3. Cooperation mechanisms: A section on the mechanisms of cooperation is always present in MOUs but is often left quite vague. Nevertheless, the mechanisms of cooperation demonstrate an effort to lower the costs for countries who join by making use of existing diplomatic frameworks. First, MOUs emphasize utilizing existing bilateral mechanisms, which reduces the diplomatic costs of participation by avoiding the need to agree on new arrangements. Second, where MOUs do mention the creation of new mechanisms, this language is often quite tentative. For example, Italy’s MOU states that “the Parties may conclude arrangements for collaboration in specific sectorial fields” [emphasis added]. Finally, pilot programs are frequently mentioned, which lower the costs of participation by enabling participants to trial new projects with no long-term commitment.

4. Settlement of differences: Most MOUs reviewed state that any differences over interpretation or implementation will be settled “amicably” using “direct consultations.” For example, Latvia states that “the two Participants will settle differences in the interpretation and implementation of this MOU through friendly consultations.” Although the BRI does have a centralized dispute resolution mechanism—commercial courts were set up in Xi’an and Shenzhen in 2018 to hear disputes—none of the MOUs reviewed contain clauses mentioning these courts or mandating their use. If such clauses exist, they likely do so in the context of commercial contracts negotiated separately from the MOUs (Dahlan 2020). The absence of such clauses substantiates our claim that MOUs are best viewed as signals of intent to cooperate rather than evidence of effectuated cooperation.

5. Legal status: None of the BRI-MOUs surveyed is legally binding. Indeed, in public international law, MOUs are recognized as agreements that, unlike treaties, are informal and non-binding (Kolb 2013), which makes them easier to negotiate. Thus, MOUs are often viewed as “agreement[s] intended to have political (or moral) weight” rather than incur legal obligations (Kolb 2013, 142). Eighteen out of twenty-four MOUs surveyed contain a clause that explicitly emphasizes that the MOU does not constitute a legally binding agreement. Bosnia Herzegovina asserts, for example, that “This Memorandum does not constitute legally binding obligations for the two Participants” and “is only an expression of their common will to jointly advance the Belt and Road Initiative.”

6. Provisions for renewal and termination: MOUs are typically in effect for 3–5 years and are automatically renewed, unless one of the parties submits written notice communicating otherwise. Time periods can be individually negotiated. El Salvador and Costa Rica signed nearly identical MOUs, for example, but El Salvador signed an MOU with a 5-year term while Costa Rica opted for a shorter, 3-year term.⁵ MOUs can also be amended.

Termination is also possible but requires written notice. For example, the Philippines’ MOU notes that it will “remain in effect for a period of four (4) years, unless a Participant officially notifies, through diplomatic channels, to terminate this MOU at least three (3) months before its expiration date.” To our knowledge, only one MOU—with Romania—has been terminated in this manner.⁶

Our Argument and Alternative Explanations

Below, we present our argument and the three alternative explanations against which we test it. Our argument takes the conclusions from the previous section as a starting point: The costs associated with signing MOUs are low, and the potential benefits are high.

We view BRI-MOUs as an example of “non-serious bargaining,” which reflects cooperation that is largely symbolic and not substantive in nature. According to James Fearon, non-serious bargaining occurs when the shadow of the future is short and enforcement is impossible (Fearon 1998). The characteristics of BRI-MOUs reflect these conditions: The shadow of the future is short (MOUs are effective for 3–5 years and contain provisions for non-renewal), and enforcement is not possible (MOUs are not legally binding). The review of MOUs in the previous section revealed minimal, largely cosmetic differences between agreements, which further substantiates the notion that states do not approach bargaining seriously.

Nonetheless, the future benefits of joining the BRI, however vague, appear quite high. Vast sums of money—as much as 1 trillion USD (Chatzky, McBride, and Berman 2023)—have been associated with the BRI. Although a state lends its support to China’s narrative by signing an MOU, the main attraction appears to be potential future benefits, such as increased Chinese investment or access to China’s market.

Because MOUs themselves generate few costs but are accompanied by substantial potential gains, we argue that most states should choose to join the BRI. States that choose not to join will be those that view the costs of signing a BRI-MOU with China as much higher, such as lending their support to a Chinese-led initiative. Put differently, the determinants of non-participation will be just as important—if not more important—than the determinants of participation.

We contend that states will view the costs of joining the BRI as higher under two conditions. The first is when a state has a democratic political system. Democracies will view the costs as greater because they are more likely to be opposed either to China’s preferred principles of cooperation or to endorsing cooperation with an increasingly influential autocratic state that is often viewed as seeking to reshape the norms of the international system.

This argument reflects the informational and normative, as opposed to institutional, logics that underlie arguments about the influence of democracy on a state’s international behavior. That is, democracy affects a state’s actions abroad through the public’s perceptions of threat and morality: democratic publics view non-democratic states as more threatening and untrustworthy than democratic ones (Maoz and Russett 1993; Tomz and Weeks 2013). Even democratic leaders who may be more inclined to consider joining the BRI must contend with a public or elite that might be opposed to such a move.

The second condition is whether a state is a close security partner of the United States, which should raise the costs of signing an MOU for two reasons. The first involves a material interest in maintaining US support. Allies, recipients of US arms sales, or security partners may be discouraged from maintaining even non-security relationships with Washington’s strategic adversaries (Campbell and Sullivan 2019). Because the United States has been openly skeptical of the BRI since its inception, its security partners are more likely to view joining as risking their ties with Washington. When Italy signed its BRI-MOU in 2019, for example, a White House spokesman called the BRI a “vanity project” and asserted that Italy should avoid it (Reuters 2019). Similarly, in 2019, Vice President Pence publicly thanked Iceland for not joining the BRI and linked this decision with the health of the bilateral relationship, stating that “the U.S. is grateful for the stand Iceland took rejecting China’s Belt and Road financial investment” (Pence 2019). More recently, President Biden publicly called the BRI an initiative of “debt and destruction” (Biden 2023).

A second reason why US security partners may be less likely to join involves the tradeoff between security and autonomy. States that rely on others for their security are often thought to cede a degree of foreign policy autonomy. Lake argues that the structure of the international system is best explained as a social contract of authority between the United States and smaller states who subordinate their autonomy in exchange for security (Lake 2007). Morrow makes a similar argument about alliance relationships (Morrow 1991). The tradeoff between security and autonomy predicts that states benefitting from US security ties will subordinate other objectives (such as joining the BRI) to preserve these relationships.

To summarize, we argue that the costs of a state’s initial decision to participate in the BRI are low and that joining the BRI promises high potential economic gains. All things being equal, we expect that countries will choose to participate in the BRI unless they view the costs of joining as much higher, such as lending political support to a Chinese-led initiative. We propose two variables—democracy and security ties with the United States—that may raise the costs of participation and expect these variables to be negatively correlated with BRI participation. Although this argument is straightforward and intuitive, it has nonetheless been overlooked in favor of less generalizable, state-specific explanations for participation in the BRI.

We test our argument against three extant alternative arguments for why states join the BRI. The first stresses economic need (Rolland 2017; Ferchen et al. 2018; Dollar 2019). Even though benefits of joining the BRI are diffuse and uncertain, much of the discussion around the BRI focuses on investment in infrastructure. Thus, countries with low-quality infrastructure should be more likely to join the BRI to access the funds that China has declared it intends to invest. More generally, poorer countries should be more likely to join than wealthier ones. Arguments about China’s “debt-trap diplomacy” often rely on the assumption that countries join for economic reasons.

A second set of alternative arguments stresses a partner state’s political or economic affinity with China. Countries who already have close ties with China should be more likely to join the BRI to sustain their cooperation, as China has attached great importance to the success of the BRI (Dreher et al. 2018). Thus, states with close political affinity, such as similar voting records in the UN, or states that already trade substantially with China should be more willing to join the BRI.

A third set of alternative arguments highlights that corruption in domestic political institutions may increase the probability that a state decides to join the BRI. Specifically, some discussion of the BRI and of China’s overseas investment points to how national leaders may be more likely to pursue investment from China for personal gain. Thus, countries with higher levels of corruption should be more likely to join the BRI (Fulton 2020).

The Survival Analysis Model and Data

In this section, we discuss why survival analysis is the most appropriate statistical technique for hypothesis testing, explain how we use such analysis in the context of the BRI, and describe our new dataset.

Survival analysis is a form of regression used to estimate the effects of one or more predictor variables on the duration of time before some event occurs to a unit. It is more appropriate than standard Ordinary Least Squares (OLS) regression with a binary indicator of BRI membership because it considers not only if something happens, but also when it happens (Jones and Box-Steffensmeier 2004). A unit is said to “survive” so long as the event has not yet happened. Political scientists have used survival analysis to explain various outcomes, including norm proliferation and treaty ascension and ratification (Simmons, Lloyd, and Stewart 2018; Hashimoto 2020). Specifically, we use a Cox regression model with time-varying covariates. The quantity of interest is the hazard ratio, which compares the risk of an event occurring to units with different values of a covariate of interest. For instance, if the hazard ratio for a binary democracy variable (i.e., 1 if the country is a democracy) is less than 1, it means ceteris paribus, democracies are less likely to join the BRI than non-democracies.

Our panel dataset includes 193 countries with observations for eight covariates between 2013 and 2019. We include countries that recognize Taiwan and those that switched recognition during this period. Our results are robust to dropping all Taiwan-recognizing country years, the results of which can be found in the online appendix. We use BRI-MOUs to measure participation in the initiative.

To test our argument, our dataset includes economic, domestic, and alignment variables. First, we include four economic variables. World Bank data on Gross Domestic Product (GDP) and GDP per capita, both in constant 2010 US dollars, are used to test the effect of overall market size and level of economic development. A World Bank measure of the quality of infrastructure from the Logistics Performance Index is used to determine whether BRI participation is driven by a need for increased infrastructure investment. This data take the form of a continuous variable assigned values between 1 (least developed) and 5 (most developed) and is reported by the World Bank every 2 years (The World Bank 2020a). Finally, UN Comtrade data are used to calculate total trade with China to test the claim that pre-existing economic interaction with China is positively correlated with BRI participation (The United Nations 2020).

We include two domestic-level variables to test the effect of a state’s political structure or private interests. First, polyarchy scores from the Varieties of Democracy dataset are used to measure the level of democracy (Coppedge et al. 2020). This variable takes a continuous value between 0 and 1, with larger values indicating a greater degree of openness in the state’s domestic political system. Second, the World Bank Worldwide Governance Indicators corruption index is used to measure the level of corruption (The World Bank 2020b). This is a continuous value ranging from roughly ?2.5 and 2.5 (with lower scores indicating a greater degree of corruption) that we normalize to values between 0 and 1.

Two final variables test whether BRI participation may be the result of international political alignment with China or the United States. A state’s ideal point distance from China in UN General Assembly voting patterns, adapted from UN General Assembly Voting Data, measures affinity to China (Voeten, Strezhnev, and Bailey 2009). A greater value indicates a greater distance from China’s voting preferences. A binary dummy variable indicating whether the partner state houses a US military base is used to measure security ties with the United States (Vine 2019). These sites have more than 10 acres of land, represent a military investment of USD 10 million or greater, and house more than 200 personnel.

Our model can be informally summarized as follows:

BRIPARTICIPATION=GDP+GDPPC+CNTRADE+UNDISTANCE+CORRUPTION+INFRASTRUCTURE+POLYARCHY+U.S.BASE.

Results

The analysis indicates that two factors affect the likelihood of a state joining the BRI: (1) level of democracy and (2) pre-existing trade with China. Market size, income level, UN alignment, corruption, the existence of US bases, and level of infrastructure development do not influence participation.

Table 1 reports the hazard ratios for each coefficient. Standard errors are clustered at the country level to account for serial correlation across time periods for a country. A hazard ratio larger than 1 indicates that an increase in the covariate increases the chance of joining the BRI. A ratio less than 1 denotes a decreased risk of signing on, with all other covariates held constant. For example, the hazard ratio for the GDP coefficient in   table 1 is equal to 0.867, meaning states with higher GDPs are less likely to join the BRI.

Table 1.

Open in new tab

Hazard ratios

Hazard ratios	MOU
GDP	0.867
	(0.101)
GDPPC	1.193
	(0.119)
Logged trade/China	1.155*
	(0.086)
UN distance from China	1.091
	(0.215)
Corruption	0.546
	(0.955)
Infrastructure	0.788
	(0.358)
Polyarchy	0.190***
	(0.526)
US bases	0.607
	(0.339)
Observations	840
R-squared	0.042
Maximum R-squared	0.692
Wald test	29.470*** (df = 8)

SEs clustered by country and for coefficients, not hazard ratio.

Note: ***p < 0.01; **p < 0.05; *p < 0.1.

Recall that our argument emphasizes variables that inhibit states from joining, rather than those that encourage them. Specifically, we hypothesize that democracy and US security ties are the two variables that might increase the costs of joining. As such, these two variables should be negatively correlated with joining—i.e., have hazard ratios less than 1.

Our empirical results for these predictions are substantively correct. Democracy and the presence of a US base decrease the likelihood of joining the BRI. Democracy reaches statistical significance, while US bases do not but have a hazard ratio in the expected direction. All else equal, a unit increase in the polyarchy score (becoming more democratic) reduces the odds that a state will join in that year by 81 percent.

Apart from democracy, only existing trade with China reaches statistical significance—but only at the 0.1 level—and is positively correlated with BRI participation, which supports the hypothesis previously offered by scholars that countries with existing economic ties to China are more likely to join. Under this account, participating states seek to consolidate their existing economic ties with China.

The remaining variables do not yield interpretable results. In particular, UN alignment and infrastructure do not reach significance. This undermines the hypotheses that close international political alignment with China or infrastructure needs motivate countries to join the BRI.

We conduct four robustness checks. First, we extended our analysis to 2021, running the same model but extending the infrastructure variable’s values to 2021. Our substantive results remain the same—only the level of democracy and trade with China reach statistical significance and remain correctly signed.

Second, one assumption of survival analysis is that observations are not “tied,” or one state’s decision to join the BRI does not impact another’s. One might argue that this assumption is violated because many states joined the BRI during the 2017 BRF and the 2018 meeting of the Forum on China–Africa Cooperation (FOCAC), suggesting the presence of network effects. However, network effects are unlikely to compromise our results. First, events like the BRF and FOCAC do not appear to exert a causal effect on joining the BRI. Countries that joined during these events would have joined anyway, and countries that did not join would not have changed their decision simply because they attended the event. MOUs take time to negotiate, and it is unlikely that a country that did not plan on joining the BRI would have joined during a BRF because it witnessed other countries doing so at the event. Indeed, thirty-nine countries sent delegates to the 2017 BRF but did not join that year. Similarly, during FOCAC, which was attended by fifty-three African states, only twenty-eight joined the meeting (Ministry of Foreign Affairs 2017, 2018).

Furthermore, even though these events could affect the timing of the signing, they are unlikely to have a significant effect on our results because time periods are measured in years. For instance, the 2018 FOCAC would only impact our results if the effect was strong enough such that it delayed or accelerated signing by over a year, which is not impossible, but unlikely. Similarly, MOUs signed independently—but in the same year—from the BRFs are sometimes included in the BRF’s list of deliverables. Nonetheless, to guard against network effects, we run a model with a variable that captures the proportion of countries that joined in the previous year in a country’s region and income level, as classified by the World Bank. Our findings remain robust.

Third, we use strategic partnerships with China instead of the UNGA ideal point as an alternative measure for political affinity with China, which does not change our results. Finally, instead of US bases, we use both the UN ideal point distance from the United States and US arms sales as alternative measures of the country’s US security ties. Neither measure changes our results. We include all these robustness checks in the online appendix.

In the subsequent analysis, we focus on democracy, since it is the only hypothesized variable of interest that reaches statistical significance. To provide a more intuitive account, we generate expected durations to event (Kropko and Harden 2020). This method maps the expected sequence of BRI signing, based on each unit’s covariates, to the time scale of the dataset. This enables us to illustrate the change in the expected survival period associated with an increase in each covariate. The advantage of this method is that it gives a sense of scale, which hazard ratios do not. For instance, a doubling in risk can mean going from one to two days or from 1 to 2 years. In substantive terms, we report how much shorter or longer a country is expected to not join the BRI when it transitions from the 10th to the 90th percentile of polyarchy score (i.e., becomes more democratic).

Figure 2 shows that a democratic state is associated with a 0.78-year delay in joining the BRI compared to an autocratic counterpart. We include the expected effect of hosting a US base to show that it goes in the expected direction, but note that it never reaches statistical significance. Similarly, we include the expected effect of trade with China, which shows that going from the 10th to the 90th quantile (increasing) shortens the time to join. In short, becoming more democratic makes a country less likely to join the BRI. For completeness, 90 percent confidence intervals (CIs) were computed by clustered bootstrap at the country level.

Figure 2.

Effect on expected durations (90 percent CI).

Open in new tabDownload slide

We also report adjusted survival curves based on our estimated Cox regression model. These are computed based on the fitted model, rather than a description of the data. Specifically, it computes a predicted survival curve for each subject based on its covariates and takes a weighted average of the curves to get an overall estimate for each subpopulation.

As figure 3 shows, being democratic increases survival probability, meaning that countries are less likely to join. A binary democracy indicator based on a 0.5 polyarchy cut-off was used to generate the adjusted curve for democracy.⁷

Figure 3.

Adjusted survival curve: democracy (95 percent CI).

Open in new tabDownload slide

Illustrative Case Studies

To further examine and illustrate the impact of democracy on joining the BRI, we conduct two pairs of case studies using a “most similar” case selection strategy. In each pair, one state joined the BRI and the other did not. We matched states across all covariates except our variable of interest—democracy—using the caseMatch tool (Nielsen 2016). Although our statistical model suggests hosting US bases does not affect BRI signing, we did theorize that US security ties would have an effect. As such, we include a second pair where neither country hosts US bases to guard against overdetermination. We also examine an outlier, Italy, which joined the BRI despite being a democracy.

Turkey and Spain

A comparison of Turkey and Spain illustrates how democracy exerts an effect on a state’s international behavior through informational and normative mechanisms. In other words, democratic publics view non-democratic states as more threatening and untrustworthy, which then constrains the actions of their leaders when it comes to participation in the BRI. This dynamic is absent in the case of Turkey, which became increasingly undemocratic between 2013 and 2019 (its polyarchy score plummeted from 0.542 to 0.292), but clearly evident in the case of Spain (which had a consistent score of 0.876) (Coppedge et al. 2020). Both countries were matched on covariates other than polyarchy, including the presence of US bases.

Turkey

The case of Turkey, which signed a BRI-MOU in November 2015, shows that as the country became more autocratic, its leaders have become less concerned with domestic costs incurred from its close association with China.

The Turkish Justice and Development Party’s (AKP) democratic backsliding began after an alleged coup attempt in 2007. After this, Turkish President Tayyip Erdogan increasingly consolidated power, detained critics, and placed the media under AKP’s control (Cook 2016). The Gezi Park protests in 2013 were a response to this creeping authoritarianism, but failed to prevent Erdogan’s consolidation of power.

Under Erdogan, Turkey actively sought alternatives to EU financing and new sources of economic growth, including China. Turkey cooperated closely with Chinese businesses even before joining the BRI. For example, a high-speed rail link between Istanbul and Ankara was completed in 2014. Months before signing the BRI-MOU, in September 2015, Chinese firms invested $920 million in the Kumport Terminal near Istanbul (Sweet 2014; Atli 2017). After joining the BRI, Erdogan argued that the BRI is the best substitute for Western alignment, stating that it “isn’t the EU at all costs,” and described the Chinese-led Shanghai Cooperation Organization as an alternative to NATO (Staff 2016; Özsu and Binark 2019).

The Turkish government had great latitude in its approach toward China because it was less constrained by public sentiment. In 2017, after joining the BRI, Foreign Minister Cavusoglu asserted that Turkey would not allow “any activities targeting or opposing China. . . [and would] take measures to eliminate any media reports targeting China” (Reuters 2017b). Turkey has also been increasingly silent on the treatment of Uyghurs in China’s Xinjiang province, despite cultural ties between Turks and Uyghurs and a diaspora population of 45,000 Uyghurs in Turkey. In 2009, Erdogan asserted that China was committing genocide in Xinjiang, but by 2019, Turkey refused to endorse a joint statement condemning China’s actions in Xinjiang (Reuters 2009; Westcott 2019). One expert argued this was “caused by Turkey’s new ruling ideology, its anti-Western perception of the world, and its affinity with non-Western powers that include both Russia and China” (Kashgarian 2020). China also adapted to Turkish demands on using local contractors in exchange for Turkish moderation in Xinjiang (Ergenç and Göçer 2023, 15).

Thus, China appeared to Turkey to be a more attractive partner than the West at a time when its democratic institutions weakened considerably.

Spain

In contrast to Turkey, Spain has not signed a BRI-MOU. The Spanish public’s wariness of China, coupled with negative memories of the authoritarian Franco regime, has limited the Spanish government’s ability to associate too closely with China despite potential economic benefits.

Spain has adopted a cautious approach toward the BRI, with officials citing concerns over transparency, economic viability, and other partners in the EU—possibly a veiled reference to Greece and Italy—becoming too close to China (Ortega 2019). In 2019, Spain’s chief diplomat, Josep Borrell, explained that Spain had decided to take a “constructive” approach to the BRI, which he asserted had “positive potential” if certain EU principles were respected (Cai 2019). Although political figures have often adopted careful language, Spanish firms have expressed enthusiasm for the BRI. In 2017, COSCO acquired a majority stake in Noatum Port Holdings, and in 2018, $17.6 billion worth of deals were signed (Reuters 2017a; China Global Television Network 2018).

These contradictions illustrate how the Spanish government has sought to balance the potential economic benefits of BRI participation with public wariness about China. Although the Spanish government does not prioritize human rights concerns in its relations with China beyond supporting EU statements critical of Chinese practices, the Spanish public is more critical (Esteban and Otero-Iglesias 2018). Polls, for example, suggest that Spanish citizens’ opinions of Chinese influence have turned negative at a faster rate than in other EU countries (Real Instituto Elcano 2017), which likely further constrained Spanish leaders from joining the BRI.

The Spanish public are especially wary of China’s political system and the increasing concentration of power under Xi due to Spain’s legacy of fascist rule. As one report on Spanish perceptions of China notes, “the majority of Spaniards are against anything that reminds them of the Franco regime—one-party rule, media and internet censorship, and restrictions on political rights all fall into this category” (Esteban and Otero-Iglesias 2018).

Spain’s historical experience highlights the normative and informational effects of democracy: Although business focuses on the economic opportunities provided by China’s rapid development, the public tends to be more reluctant to increase ties, creating costs for formal participation in the BRI.

Fiji and Mauritius

A comparison of Fiji and Mauritius also illustrates how democracy exerts an effect on international behavior. Fiji became increasingly undemocratic from 2013 to 2019, with its polyarchy score dropping from 0.465 to 0.146, while Mauritius had a relatively constant score averaging 0.816 (Coppedge et al. 2020). Again, Fiji and Mauritius were matched on covariates other than democracy, but, unlike Spain and Turkey, neither hosts a US military base (obviating this as a potential causal factor).

Fiji

Fiji signed a BRI-MOU in November 2018. Fiji’s decision to join the BRI was enabled by political upheaval and multiple coups in the early 2000s that alienated Western states and forced Fiji’s government to adopt a more independent foreign policy, regardless of domestic concerns regarding Fiji’s close association with China.

Fiji has repeatedly sought to diversify its economic and political relations over the past three decades. After coups in 1987, 2000, and 2006, the country first initiated, and then re-initiated, its “Look North” policy—a reference to countries Asia and China in particular—as economic sanctions from the West drove it to search for new partners (Szadziewski 2020b). Prime Minister Bainimarama stated that the policy reflected an “intention to expand relations with non-traditional partners” in a way that is “essential to [Fijian] national development” (Bainimarama 2011).

Due to these efforts, Fiji enjoyed close commercial ties with China prior to joining the BRI. The Wangguo Friendship Plaza in Suva, which began construction in 2016, was set to become the tallest building in the country. Yet issues with China-funded projects began even before Fiji joined the BRI, and the Wangguo Plaza was delayed and remains incomplete (Szadziewski 2020a).

Fiji’s state-friendly media provided little coverage of the failures, however, and continued to show enthusiasm for the BRI, highlighting the natural alignment between Fiji’s and China’s economic initiatives (Szadziewski 2020a). As two scholars note, “Prior to 2015, Look North narratives dominated media and official discussion of Sino-Fijian relations; however, by 2017, the swing toward the BRI framework appears absolute, with Look North mentions barely registering in Chinese and Fijian sources” (Smith and Wesley-Smith 2021, 302–3).

China’s increasing presence in Fiji has generated political opposition, though it remains unclear if this will influence overall policy. In the 2022 elections, Sitiveni Rabuka, Fiji’s main opposition leader, became prime minister and replaced Bainimarama, who led the 2006 coup and had served as prime minister in 2007. As early as 2018, Rabuka asserted that he was uncomfortable with China’s involvement in Fiji’s affairs (Smith and Wesley-Smith 2021, 388). More recently, Rabuka canceled a police training and exchange agreement with China, and deepened defense cooperation with Australia, noting that “our system of democracy and justice systems are different so we will go back to those [Australia and New Zealand] that have similar systems with us (Al Jazeera 2023).”

Thus, greater democratic contestation is consistent with the role of public opinion in shaping attitudes toward China, even though Fiji joined the BRI during a period when it was undemocratic.

Mauritius

Mauritius, a democracy, has not joined the BRI despite close economic relations with Beijing—including being the first African country to sign an FTA with China. Mauritius’ non-participation likely stems not from domestic suspicion of China, per se, but from its close relationship with India, another democracy that has publicly opposed the BRI. Interestingly, although Mauritius’ experience does not provide support for our hypothesis that American security ties would inhibit countries from joining the BRI, it does suggest that security relationships with other states can increase the costs of BRI participation.

China and Mauritius have had close relations for several decades. Bilateral relations have included a series of cooperation agreements that led to the creation of the $820 million Jinfei Economic Trade and Cooperation Zone in 2008, China’s first overseas special economic zone and the largest injection of foreign capital into the country at that point (Chang 2013). As noted above, Mauritius also signed the first FTA between China and an African nation in 2019, a feat one Mauritian journalist hailed as a “hallmark of the deepening partnership between both countries” (Arden 2021).

Why has Mauritius not joined the BRI? One potential explanation is its relationship with India, a historically important partner and a democracy publicly opposed to the BRI. Two-thirds of Mauritius’ population are Indo-Mauritian (“Mauritius” 2022). India has also tended to view the Indian Ocean as part of its sphere of influence. In 1983, for example, India seriously considered intervening in Mauritius to prevent a pro-Soviet coup (Panda 2015). More recently, India has pursued an aggressive diplomatic strategy aimed at preserving its influence in the Indian Ocean. In Mauritius, this has included high-level state visits, cooperation on defense and security issues, and development assistance (Hall 2019). Mauritius’ non-participation in the BRI could possibly be a signal of continuing alignment with India and India’s concerns about China.

Nevertheless, public opinion could have played a role in the decision to not formally join the BRI. Mauritian journalists have written freely about the potential debt trap and geostrategic implications of the BRI, with some calling it a “toxic campaign” (Prosper 2021). A 2020 public survey revealed that although the public generally welcomes Chinese investments, a plurality believes that Mauritius has “borrowed too much” from China (Afrobarometer 2021). That said, in a recent wave of the Afrobarometer survey in March 2022, a majority of respondents (69 percent) thought that Chinese political and economic influence was either somewhat or very positive (Afrobarometer 2022).

The Outlier: Italy

Italy, a wealthy democracy and the only G-7 country to join the BRI, represents an outlier for our argument because it signed an MOU in March 2019. As it also decided to withdraw from the BRI in July 2023, tracing the mechanisms behind its decision to join and then leave allows us to examine our hypotheses about costs (Armellini 2023).

Prime Minister Giuseppe Conte, who led a populist government dominated by his Five Star Movement, drove the decision to join China’s BRI. Conte sought to signal Italy’s discontent with the EU and seek potential alternative economic partners (Pugliese, Ghiretti, and Insisa 2022). In anonymous interviews, Italian ministers from his government remarked that signing was a low-cost way of mobilizing the government’s Eurosceptic base—effective “political marketing” that they intentionally “played up”—and offered the potential for increased economic engagement with China (Pugliese, Ghiretti, and Insisa 2022, 1048–50). The MOU’s vague nature and Italy’s ability to frame the MOU’s content within existing EU–China frameworks—such as the EU–China Connectivity Platform—further lowered the costs of signing.

Yet consistent with our argument, public wariness toward China motivated Italian leaders to emphasize the non-committal nature of the MOU. Before the signing, Conte sought to defuse concerns by stating that it “won’t mean that the next day we will be forced to do anything. It will allow us to enter into this project and have a dialogue” (Balmer 2019).

As time progressed, Italy distanced itself from the BRI and signaled its intention to withdraw in July 2023 (Armellini 2023) for two reasons. First, promised economic benefits failed to materialize. Italy’s defense minister, for example, noted that the BRI failed to boost exports to China, and described the decision to join as “improvised and atrocious” (Armellini 2023). Prime Minister Giorgia Meloni remarked that France and Germany had closer trade relationships with China than Italy despite not being BRI members (Sorgi 2023). Second, these unrealized economic benefits were coupled with domestic and international political costs. Domestic pushback from Italian legislators on China’s policies on Xinjiang and Hong Kong coincided with a shift in the government’s position away from China toward the EU (Meacci 2021). Internationally, other EU members have criticized Italy’s membership, and some have argued that Italy’s signature led to its isolation by the EU during the Comprehensive Agreement on Investment negotiations with China (Meacci 2021). Both are consistent with our hypothesis that democracies incur costs from a closer association with an authoritarian state.

Public opinion toward China also worsened after Italy joined the BRI, with 38.5 percent of respondents reporting in 2020 that their view of China declined in the preceding 3 years (Gallelli et al. 2021). Reflecting this shift in public opinion, when Mario Draghi became prime minister in 2021, his government publicly characterized China as “an autocracy that does not adhere to multilateral rules and does not share the same vision of the world that democracies have” and claimed that he would assess the BRI-MOU “carefully” (Reuters 2021).

The current Meloni’s government decision to withdraw is thus consistent with trends from previous administrations, and the preceding discussion shows that our hypothesized mechanism—costs incurred when a democracy associates itself with an authoritarian state—did indeed contribute to this decision.

Conclusion

Although the BRI has attracted substantial academic interest, scholars have neglected to interrogate comprehensively one of the most fundamental questions about the initiative: Why do states join? In asking how and why states join the BRI, we shift the analytical focus from China to partner states and provide the first comprehensive study of BRI participation.

Our research reveals important characteristics of BRI participation. First, there is a clear mechanism by which states formally join the BRI: the signature of an MOU on cooperation under the BRI with the Chinese government. Second, an examination of these MOUs reveals that they are—by design—not costly for states to sign. Nevertheless, by signing them, states position themselves to receive potentially huge economic benefits. This suggests a counterintuitive argument: States should choose to join the BRI unless they believe joining entails other costs that decrease the value or appeal of participation. We hypothesize that two factors are most likely to inhibit participation by raising the costs of participation: democracy and security ties with the United States. Third, quantitative and qualitative tests of this argument highlight democratic governance as the strongest predictor of non-participation. We also find that pre-existing trade ties with China are positively correlated with participation, but only at the 0.1 level of statistical significance.

Our results call into question common arguments about the BRI, particularly regarding its potential influence. First, joining the BRI is not particularly costly and thus should not be viewed as a strong signal of affinity with China. The terms by which states join the BRI indicate that the MOUs lack mechanisms or provisions that would increase Chinese influence. Because the costs of participation are by design low, non-participation represents a stronger signal because it incurs a cost in the form of foregone potential economic gains. China has been unable (with a few exceptions) to use the lure of BRI membership to drive a wedge through the advanced industrialized democracies. Moreover, even when China has been successful, democratic publics have remained skeptical of the initiative and, in the case of Italy, appear to have soured on it quite quickly.

Second, and counterintuitively—given the BRI’s branding as an economic connectivity initiative—our results demonstrate that economic need (as measured by low GDP and poor infrastructure) does not predict which states choose to participate. Instead, overall, states who trade more with China are more likely to join. In other words, BRI participation represents a rebranding or consolidation of existing relationships more than it does an expansion of China’s economic reach. Moreover, China’s messaging can be misleading, exaggerating the BRI’s impact. For example, in October 2021, Xinhua reported that “China’s trade with countries along the Belt and Road rose 23.4 percent year on year” (Xinhua 2021), implying that such trade was caused by BRI participation even though we find that higher levels of trade are associated with joining.

Third, joining the BRI does not mean that promised economic benefits will necessarily materialize. Once a state has signed a BRI-MOU, China has achieved its political goal of increasing acceptance and support of the initiative. Moreover, because enforcement of MOUs is impossible, China may have incentives to renege on promises of economic cooperation after MOUs have been signed (Fearon 1998, 285). As Italy suggests, enthusiasm about participation in the BRI declined substantially when economic benefits never materialized, contributing to the government’s decision to leave the initiative.

Superficially, the BRI appears to be an all-encompassing and ever-expanding initiative, reflecting China’s increasing economic power. Upon closer examination, however, the initiative has encountered obstacles that have limited its reach. Democratic governments, in particular, have refrained from joining. A full assessment of the BRI’s actual impact, however, will need to consider the experiences of states that do decide to participate. Whether the BRI has translated into greater Chinese influence and power depends substantially on whether the promised economic benefits of participation materialize in these states and whether China can use these benefits to increase its influence. The low cost of signing BRI-MOUs cuts two ways: It makes it easy for states to join the initiative, but it also makes it easy for them to leave.