Introduction
Between 2008 and 2019, Latin American (LAC) countries borrowed $132 billion from Chinese development finance institutions (DFIs) (Ray and Myers Reference Ray and Myers2023) and $155 billion from the World Bank, including its International Bank for Reconstruction and Development (IBRD) and International Development Association (IDA) windows. Over 20 LAC countries have become signatories of China’s Belt and Road Initiative (Nedopil Reference Nedopil2022). LAC was the first region the United States visited to promote its own infrastructure development program, Build Back Better World (B3W), in 2021, with stops in Colombia, Ecuador, and Panama (Wilkinson Reference Wilkinson2021). The region has become a battleground in ongoing competition between the United States and China in the past decade, and development finance is one tool these great powers can use to exert influence or impose consequences on the region (Urdinez et al. Reference Urdinez, Mouron, Schenoni and de Oliveira2016; Zelicovich and Yamin Reference Zelicovich and Yamin2024). However, interviews with policymakers and cross-national loan data show that LAC borrowers’ interests supplement lenders’ preferences as explanatory factors for China’s development finance.
Using data on 18 LAC countries’ borrowing from 2008–2019 and a case study of Ecuador, this article investigates the conditions under which LAC countries choose to borrow from Chinese DFIs, China Development Bank (CDB) and Export-Import Bank of China (CHEXIM) versus their United States-backed counterpart, the World Bank.Footnote 1 It then qualitatively traces the interactions between these different explanatory factors through interviews with government officials in Ecuador, one of the LAC countries that has borrowed the most from China.
Ecuador is an emblematic case for studying China’s development finance. Like many countries borrowing from China, it has faced a continued need for external finance and turned to China for both project-specific and discretionary lending. Furthermore, its return to bond markets in 2014 after its 2008 default provides an opportunity for a before-and-after analysis (Zanoni et al. Reference Zanoni, Díaz, Paredes, Gaston Andrian and Lorenzo Maldonado2024). Ecuador has financed the construction of seven hydroelectric facilities with Chinese loans, along with other energy, transportation, and infrastructure projects. This includes the troubled flagship project, the Coca-Codo Sinclair hydroelectric dam, completed in 2016 but which has since suffered from degrading materials and environmental damages (Villavicencio Valencia et al. Reference Villavicencio Valencia, Bruno Segovia Mejía and Ramiro Velasco Erazo2022). The two countries established a discretionary loan facility for infrastructure projects in 2010, backed in part by oil export revenues, which has since been renewed in five phases, most recently in 2018 (Ray and Myers Reference Ray and Myers2023). Ecuador and China have also developed close political relations; a particular highlight mentioned by interviewees was President Xi Jinping’s state visit to Ecuador in 2016 (ECU 12, ECU 14).
This article first replicates earlier studies showing how macroeconomic factors such as creditworthiness are key drivers of China’s development finance. It then demonstrates how these conditions interact with political factors and project-specific goals. Statistical results show that leftist political parties are more likely to borrow from China, and interviews suggest that these political considerations are important at a higher level and earlier phase of decision-making. This article finds that countries’ borrowing decisions are not only related to China’s relative power, but also to their foreign policy alignment towards the United States. Interviewees highlighted China’s emergence as an alternative to traditional institutions and described their goal to diversify their financing. LAC countries also strategically borrow from different lenders for distinct types of projects, preferring loans from Chinese DFIs when borrowing for infrastructure projects with larger loan sizes, based in part on Chinese companies’ interest and expertise. Officials in borrowing countries are aware of the tradeoffs between different types of financing and seek to maximize the development benefits from each. Altogether, these results complement prior studies of supply-side factors by providing a more nuanced view of borrowers’ reasons for seeking development finance from China versus the World Bank.
China’s Development Finance in Latin America
This article contributes to an emerging literature on the “demand-side of sovereign debt” (Cormier Reference Cormier2024) which aims to explain how developing countries choose their creditors. Authors have identified political ideology and partisanship (Ballard-Rosa et al. Reference Ballard-Rosa, Mosley and Wellhausen2022; Cormier Reference Cormier2024), interest group preferences (Bunte Reference Bunte2019), and leaders’ transparency preferences (Mosley and Rosendorff Reference Mosley and Peter Rosendorff2023) as key demand-side factors. While acknowledging the continued importance of supply-side factors, the works collectively demonstrate that incorporating demand-side variables improves explanations of lending patterns and debt composition. Aside from Bunte (Reference Bunte2019), most prior research does not specifically consider China’s role as a new creditor. The results of this article support earlier findings that even within limits imposed by external factors, such as lender preferences or macroeconomic conditions, borrowing countries do exercise agency (Humphrey and Michaelowa Reference Humphrey and Michaelowa2013). While this article probes borrowers’ preferences, both “push” and “pull” factors contribute to borrowing outcomes (Li et al. Reference Li, Gallagher, Chen, Yuan and Mauzerall2022).
Existing scholarship suggests three main explanations for China’s development finance in LAC: (1) China’s own preferences and interests, (2) the characteristics of the projects or loans themselves, and (3) considerations about the balance of power between the United States and China. This literature has focused mainly on supply-side factors, with far less written about the demand-side factors driving LAC countries’ choices. Authors have found that profit and commercial concerns (Jenkins Reference Jenkins2018; Li et al. Reference Li, Gallagher, Chen, Yuan and Mauzerall2022; Gallagher Reference Gallagher2021), the One China Policy (Hwang Reference Hwang2021; Telias and Urdinez Reference Telias and Urdínez2022; Gachúz Maya and Urdinez Reference Gachúz, Carlos and Urdinez2022), and reputation building as a “win-win” or “responsible” partner for developing countries (Gallagher Reference Gallagher2021) are key drivers for China’s lending, aid, and broader economic activities in the region. Some authors suggest that China has geopolitical motives for lending to LAC (Ellis Reference Ellis2011; Mendez and Alden Reference Méndez and Alden2021), but others argue that China is simply filling needs the United States has neglected (Urdinez et al. Reference Urdinez, Mouron, Schenoni and de Oliveira2016). While China does lend higher amounts to countries that tend to agree with it on key foreign policy questions (Flores-Macías and Kreps Reference Flores-Macías and Kreps2013; Landry Reference Landry2018), evidence on whether China intentionally supports more authoritarian or less politically stable regimes in the region is mixed (Brand et al. Reference Brand, McEwen-Fial and Muno2015; Gallagher Reference Gallagher2021; Telias and Urdinez Reference Telias and Urdínez2022).
Second, recent scholarship has begun to explore how variations on loan characteristics present borrowing countries with a choice between different types of financing. China’s loans differ from traditional Western financing in terms of their conditionality profile (Kaplan Reference Kaplan2021; Gallagher Reference Gallagher2021), environmental and social governance practices (Gallagher and Yuan Reference Gallagher and Yuan2017; Ray Reference Ray2020), and the types of projects they support (Gallagher Reference Gallagher2021). Survey experiments have shown that elites in developing countries prefer higher value projects, transportation infrastructure, untied aid, grants-based financing, and at least some regulations surrounding transparency, environmental and social governance, and labor safeguards (Blair et al. Reference Blair, Custer and Roessler2022). Bunte (Reference Bunte2019) demonstrates that domestic interest groups’ perceptions of distributional consequences differ between loans from Western institutions, BRICS countries, and private creditors, which in turn influence the government’s policymaking process.
Third, the international relations literature has analyzed countries’ strategies for navigating United States–China economic competition in the region, approaching these “triangular relations” from LAC’s perspective, challenging assumptions that they lack agency (Gallagher Reference Gallagher2016; Fortin et al. Reference Fortin, Heine and Ominami2021). Throughout the twentieth century, LAC countries have oscillated between alignment with and autonomy from the United States (Russell and Tokatlian Reference Russell and Gabriel Tokatlian2013), a dynamic that shapes the region’s political and economic relations with China (Brito Munita and Tagle Montt Reference Munita, Ignacio and Javier Tagle Montt2023). Evidence from Southeast Asia and Africa has shown that countries are successfully leveraging this geostrategic context to pursue domestic development goals, but authors have yet to conduct similar investigations in LAC (Brown and Harman Reference Brown and Harman2013; Hwang Reference Hwang2021; Kuik Reference Kuik2021; Gallagher et al. Reference Gallagher, Bhandary, Narassimhan and Tam Nguyen2021).
Finally, this article offers an important extension of the exit-voice-loyalty literature (Hirschman Reference Hirschman1970; Gehlbach Reference Gehlbach2006). While China has emerged as a new exit option for countries blocked from Western financing (Bunte Reference Bunte2019), many countries that can access more traditional financing continue to borrow from China. Between 2014–2019, after it regained access to bond markets and World Bank loans, 10 of Ecuador’s 19 loans were from China, indicating that even when Western financing was available, Ecuador chose to borrow from China in over half the cases. Evidence from Africa indicates that countries are borrowing from both Western and Chinese institutions, but for distinct purposes (Hwang Reference Hwang2021). The non-exclusive nature of development finance institutions allows states to simultaneously exercise their exit option by borrowing from China and leverage the prospect of closer relations with China to extract preferred finance from Western institutions (Gehlbach Reference Gehlbach2006; Ray Reference Ray2021).
This section has highlighted recent progress towards incorporating borrowing country agency within explanations of China’s development finance patterns. By analyzing domestic politics, international system dynamics, and project-level characteristics, this article contributes a more robust understanding of the demand-side drivers of China’s development finance. The use of qualitative evidence from elite interviews adds further nuance to the picture by delineating the interactions and sequence of these factors.
Theory and Hypotheses
This article analyses the drivers of LAC countries’ choices of foreign development partners between two ideal types: Western United States-backed institutions (World Bank) and Chinese DFIs (CDB and CHEXIM). Based on interview evidence, the level of analysis in this article is the individual project, rather than annual amounts, which more appropriately reflects a state’s decision between sovereign lenders. Policymakers explained that the process for seeking development finance begins with a list of projects, rather than a specific amount of capital (ECU 19, ECU 28). Previous authors have measured annual funding, which more closely approximates lenders’ decisions about the amount they are willing to lend, as opposed to borrowers’ decisions over which projects they seek finance for. In line with previous research, this article does not contend that lenders’ preferences are unimportant, but rather that borrowers’ decisions merit further exploration (Humphrey and Michaelowa Reference Humphrey and Michaelowa2013; Mosley and Rosendorff Reference Mosley and Peter Rosendorff2023; Cormier Reference Cormier2024). The rest of this section hypothesizes the conditions under which countries borrow from China.
Economic Indicators
Existing literature has focused on macroeconomic indicators of developing countries as the primary determinants for China’s overseas development finance in comparison to Western finance. This article thus begins by incorporating sovereign credit risk (Gallagher Reference Gallagher2021; Dollar Reference Dollar2018), trade with China (Landry Reference Landry2018; Dreher and Fuchs Reference Dreher and Fuchs2015), political stability (Dreher and Fuchs Reference Dreher and Fuchs2015; Landry Reference Landry2018; Li et al. Reference Li, Gallagher, Chen, Yuan and Mauzerall2022), and macroeconomic fundamentals (Landry Reference Landry2018; Gallagher Reference Gallagher2021).Footnote 2 GDP per capita is also included as a measure of economic development, which Humphrey and Michaelowa (Reference Humphrey and Michaelowa2013) find affects developing countries’ choices of external creditors.
Together, these factors constitute an Economic Stability vector and offer a baseline understanding of development finance patterns. Politically stable and lower risk countries will be more likely to borrow from the United States-backed World Bank, a reflection of the World Bank’s typical preferences. Countries will be more likely to borrow from Chinese DFIs when they trade more with China, a reflection of China’s preference. These relationships may also reflect borrowers’ preferences, however, such as more transparent leaders seeking Multilateral Development Bank financing over bilateral (Mosley and Rosendorff Reference Mosley and Peter Rosendorff2023).
Domestic and International Political Conditions
This article captures both popular and elite politics via public opinion and party ideology. Research has shown that while it is not the primary determinant of foreign policy, public opinion does exert indirect influence on decision-makers (Foyle Reference Foyle1999; Kertzer and Zeitzoff Reference Kertzer and Zeitzoff2017). Prior work has found that China’s development finance has no significant effect on public opinion of China in LAC (Eichenauer et al. Reference Eichenauer, Fuchs and Brückner2021). This article makes two conceptual distinctions, first by treating opinion as an independent variable, and second by analyzing opinion of the United States, rather than China, as the driving factor for borrowing decisions. This construction is based on its central role in Western-led DFIs as described by interviewees (ECU 2). Therefore, a negative correlation between public opinion of the United States and borrowing from China is consistent with policymakers using public opinion as a heuristic tool for deciding which lender should finance a given project.
More immediately proximate to borrowing decisions is the political ideology of the current government at the time a loan agreement is reached. While leftist LAC governments’ state-led development models imply generally higher levels of borrowing for investment in major projects (Coelho et al. Reference Coelho Farias De Souza2020), this article considers the influence of ideology on decisions between creditors. Prior research on sovereign debt has shown that left and right governments have distinct borrowing preferences (Ballard-Rosa et al. Reference Ballard-Rosa, Mosley and Wellhausen2022; Cormier Reference Cormier2024). Leftist governments in LAC have pursued closer relations with China based on the conjuncture between domestic political economic proposals rejecting the perceived United States-imposed neoliberal agenda and a favorable international environment (Sánchez Reference Sánchez2011; Coelho Farias De Souza et al. Reference Coelho Farias De Souza2020; Aguas and Pampinella Reference Aguas and Pampinella2022; Zapata and Martínez-Hernández Reference Zapata and Adrián Martínez-Hernández2020). I hypothesize that based on this ideological and foreign policy affinity, LAC countries with leftist governments are more likely to borrow from China than the World Bank.
This article also analyses the structural characteristics of the international system, namely the balance of power between the United States and China and LAC countries’ alignments within this context, as predictors for borrowing decisions. The debate about how countries will respond to power transitions and hegemonic challengers is long-standing in the international relations literature, and it largely centers on a decision to bandwagon with the rising challenger or balance against it alongside the existing hegemon (Waltz Reference Waltz1979; Walt Reference Walt1987; Schweller Reference Schweller1994; Kydd Reference Kydd2020). Bandwagoning countries will borrow more often from China as its relative power increases; balancing countries will do the opposite.
This article builds on recent work on non-alignment in LAC and interview evidence suggesting a different conceptualization of how countries position themselves vis-à-vis United States–China competition. In addition to the balancing-bandwagoning dichotomy, a diversification strategy implies that states seek to distance themselves from the United States generally rather than align with China particularly (Russell and Tokatlian Reference Russell and Gabriel Tokatlian2013; Fortin et al. Reference Fortin, Heine and Ominami2021). Research on the drivers of foreign policy alignment with China has found that trade with China is a significant predictor, but the evidence for development finance is mixed (Flores-Macías and Kreps Reference Flores-Macías and Kreps2013; Dreher and Fuchs Reference Dreher and Fuchs2015; Li et al. Reference Li, Gallagher, Chen, Yuan and Mauzerall2022; Landry Reference Landry2018). This article reverses the hypothesized direction of the causal relationship and treats the United States as the referent state. Thus, diversifying countries, namely those increasingly distant from the United States in their foreign policy, are more likely to borrow from China. This Political vector brings together new work from the foreign policy literature to explain a question asked primarily in development economics scholarship.
Project Characteristics
Development finance may support a wide range of projects, from large, expensive infrastructure projects to smaller healthcare interventions in particular communities. Furthermore, lending institutions tend to support certain types of development projects, whether based on expertise or preference. The World Bank has moved away from funding major infrastructure projects in the past few decades, while Chinese DFIs have filled this gap (Gallagher Reference Gallagher2021; Kaplan Reference Kaplan2021). I hypothesize that borrowing countries take advantage of these conditions to strategically select lenders for different types of projects, prioritizing China for infrastructure projects (Hwang Reference Hwang2021).
DFIs also differ on the terms of their lending, including the sizes of the loan, interest rates, maturities, and grace periods. Loans from Chinese banks tend to be larger, less concessional, and have shorter grace periods (Malik et al. Reference Malik, Parks, Russell, Walsh, Solomon, Zhang, Elston and Goodman2021; Morris, Parks et al. Reference Morris, Parks and Gardner2020). While they certainly do not prefer higher interest rates or shorter grace periods, developing countries have nonetheless chosen to borrow from CDB and CHEXIM. Prior research has demonstrated that borrowers make tradeoffs between political objectives and debt terms (Ballard-Rosa et al. Reference Ballard-Rosa, Mosley and Wellhausen2022; Mosley and Rosendorff Reference Mosley and Peter Rosendorff2023; Cormier Reference Cormier2024). This article incorporates interview evidence to shed light on how policymakers in Ecuador evaluate these tradeoffs, and which political and project goals they are willing to prioritize over less favorable lending terms.
This Project vector rounds out the factors this article evaluates as predictors for borrowing countries’ choices between Chinese development finance and the World Bank. The next section outlines the data and methods that will be used to test these hypotheses and articulate the interactions of the various factors.
Research Design
The main empirical objective of this article is to establish the factors that influence a LAC country’s decision to borrow from either the United States-backed World Bank or Chinese DFIs. To accomplish this goal, this article employs a combination of quantitative analysis of logistic regression models and novel qualitative evidence from semi-structured interviews with government officials. The study period begins in 2008, when China issued its first white paper outlining strategic priorities for relations with LAC as a region, elevating it in China’s overall global outlook (People’s Republic of China 2008). It ends in 2019 prior to the global economic disruptions in the wake of the COVID-19 pandemic. This article specifically focuses on sovereign borrowing from official creditors and thus excludes private companies’ borrowing from the same creditors and countries’ borrowing from private creditors, such as commercial banks.
The primary quantitative analysis estimates a logistic regression model for each of the three vectors introduced above with a binary dependent variable of lending institution (assigned 1 for China, 0 for World Bank) across 18 LAC countries. The political and project models retain the economic vector to evaluate their effects on prior findings. All three models are multilevel mixed-effects models consistent with the data structure of observations (projects) nested within higher-level clusters (countries) (Sommet and Morselli Reference Sommet and Morselli2017; Brown Reference Brown2021).Footnote 3
The unit of analysis in this article is the individual loan. This departs from previous studies analyzing total or proportional borrowing from China in a country-year. This departure is important to allow analysis not of how much development financing a country receives, but rather of which lending institution among multiple options provides the finance for a given project. Excepting political ideology and credit ratings, country-year independent variables are lagged one year.
The results from the statistical analysis are then examined more deeply in how they interact with each other using a single case study. Ecuador was selected as a crucial case (Gerring Reference Gerring2017) where supply-side factors were likely to be the strongest predictors for borrowing decisions due to the challenges Ecuador faced from its default on significant portions of its foreign debts in 2008 and political instability in the 1990s and early 2000s.
With the election of President Rafael Correa in 2007, Ecuador’s domestic political economy and foreign policies shifted dramatically (Coelho Farias De Souza et al. Reference Coelho Farias De Souza2020). Ecuador expelled its World Bank representative and ceased all new and ongoing World Bank operations in the country (World Bank 2022, 2024). The resumption of World Bank operations in 2013 resulted from a combination of political will from Ecuador’s government and activism from the World Bank via sub-national transportation infrastructure projects (World Bank 2016, 2024). A 2024 report from the World Bank’s Independent Evaluation Group (IEG) cited interviews with country staff who explained that their reengagement strategy started with sub-national and informal meetings to “meet the government’s desire to limit the World Bank’s visibility” (World Bank 2024). The staff interviewed for the report suggested that while the World Bank was intentionally limiting its financing in Ecuador, it also faced direct political constraints to a closer relationship.
Meanwhile, Ecuador has been among the top ten borrowers from China in the world (Malik et al. Reference Malik, Parks, Russell, Walsh, Solomon, Zhang, Elston and Goodman2021) and after Venezuela and Brazil, has borrowed the most from China in LAC. Across the study period, Ecuador borrowed over $18 billion from China in 24 specific loans. These loans were concentrated in the energy sector, renewables and fossil fuels, but also included loans for budget support (Ray and Myers Reference Ray and Myers2023). Several were backed with future oil shipments to China, a characteristic frequently taken as evidence of China pressuring Ecuador to accept unfavorable loan terms to ensure access to key natural resources for China’s own interests. Ecuador borrowed $2.3 billion across 13 loans from the World Bank during the study period, a significantly lower amount and fewer total loans.
The timing of these loans is instructive. Between 2008–2013, over 80% of Ecuador’s loans were from China. Between 2014–2019, only 53% were from China. Ecuador issued its first new bonds in 2014 and signed its first post-default loan with the World Bank in 2013. Therefore, the second half of the study period addresses a counterfactual of how Ecuador might have borrowed if it had retained access to Western finances; it continued to borrow from both sources. Data collected in interviews with policymakers aim to shed light on why Ecuador chose to borrow from both China and the World Bank.
Interview Data
In total, 19 interviews were conducted with policymakers between January and March 2024 with a response rate of 37%. Interviewees were selected based on their proximity and participation in key development finance and relevant foreign policy decisions during the study period. Current and former officials from the Ministry of Foreign Affairs and Human Mobility (MREMH); Ministry of Production, Foreign Trade, Investment, and Fishing (MPCEIP); Ministry of Defense (MDN); and sector-specific ministries were interviewed. While the Ministry of Economy and Finance (MEF) is a key negotiator for loan terms, none of the current or former officials from the MEF accepted an interview request. To fill this data gap, officials from the implementing ministries were interviewed, as they were proximate to these decisions and interacted directly with MEF officials. Because many of the Chinese-funded development projects were energy infrastructure, these interviews focused on the Ministry of Energy and Mining (MEM) and Coordinating Ministry for Strategic Sectors (MCSE). Finally, members of the National Assembly’s Commission on International Relations and Human Mobility were contacted.
Interviewees were identified through a two-step process, first by reviewing personnel lists published by relevant ministries and second by snowballing from the initial list. Three interviewees were identified through snowballing, and three others were identified via references from academics, leading to 32% of interviewees identified through snowball sampling. Interviews were conducted in Spanish, either in Quito or virtually if the interviewee was working overseas.Footnote 4 Table 1 provides additional details.
Interview Data
Regression Data
The quantitative data encompasses 18 LAC countries’ borrowing between 2008 and 2019.Footnote 5 The binary dependent variable is the lending institution for a given project, assigned 0 for the World Bank and 1 for China. Data for World Bank loans was retrieved directly from its project database. CDB and CHEXIM loans were collected from the Chinese Loans to Latin America and the Caribbean Database (Ray and Myers Reference Ray and Myers2023) and cross-checked with AidData 2.0 (Custer et al. Reference Custer, Dreher, Elston, Fuchs, Ghose, Lin and Malik2021; Dreher et al. Reference Dreher, Fuchs, Strange and Tierney2022). The data includes 518 loans, 443 from the World Bank and 75 from Chinese banks. The average loan size from the World Bank was $342 million and the median was $70 million. For Chinese banks, the average size was $1,668 million with a median of $300 million (Figure 1).
Chinese and World Bank Lending to LAC, 2008–2019.
For the economic model, governance is measured using the Political Stability and Absence of Violence index of the Worldwide Governance Indicators. Inflation rates and GDP per capita were collected from the Global Economic Monitor database. Data on LAC countries’ trade with China was collected from the International Trade Statistics Database, published by UN Comtrade, scaled as a percentage of GDP. Sovereign credit ratings were collected from Trading Economics’ listing of Standard and Poor, Moody’s, Fitch, and Dominion Bond Rating Service and scaled from 0 to 10, based on their grade, with 0 indicating almost no risk and 10 indicating high risk. Where multiple ratings were issued in the same year, the simple average was taken. Where no credit rating was issued, the prior year’s value was carried forward. Summary statistics for all independent variables are available in Table 2.
Summary Statistics for Independent Variables
For domestic politics, public opinion is measured with a question in the Latinobarómetro surveys asking respondents to evaluate the United States on a four-point favorability scale, where higher values indicate more positive opinion (Corporación Latinobarómetro 2020). The year-over-year change is used to capture the trends of public opinion rather than a simple static measure. For the three years when surveys were not conducted—2012, 2014, and 2019—the values were linearly interpolated. Party ideology data was collected from the V-Party database (Lindberg et al. Reference Lindberg, Düpont, Higashijima, Marquardt, Bernhard and Döring2022; Pemstein et al. Reference Pemstein, Marquardt, Tzelgov, Wang, Medzihorsky, Krusell, Miri and von Römer2020). A country-year measure was constructed by combining electoral data on the head of government’s party and a measure of that party’s left-right lean. More negative values indicate a more left-leaning party.
For international politics factors, the relative balance of power is measured using Beckley’s (Reference Beckley2018) approach that conceptualizes power as a net resource (GDP X GDP per capita) and most accurately tracks the rise and fall of great powers throughout history. China’s relative power is measured as China’s proportion of the United States’ and China’s collective net power resources. Data on UN voting alignment was obtained from Bailey et al.’s (Reference Bailey, Strezhnev and Voeten2017) ideal point estimates for all UN General Assembly votes. The alignment trend was estimated by calculating the absolute distance between each LAC country’s ideal point and the United States’ ideal points for each year and taking the year-over-year difference. This captures whether a country is actively moving away from the United States, in other words pursuing a diversification strategy.
Lastly, data on project characteristics was collected from several sources, starting with the lending institutions. The amounts were available for all loans in the dataset. Interest rates, maturities, and grace periods for Chinese loans were collected from AidData 2.0 where available and supplemented with data from the World Bank’s International Debt Statistics. Where available, the annual average values of these variables for all new external debt commitments to China were assigned to loans without a specific interest rate, maturity, or grace period. The maturities and grace periods for World Bank loans were calculated according to the agreement signing dates, first and last repayment dates provided directly from the World Bank.Footnote 6 World Bank interest rates for 2008–2014 were collected from Morris et al. (Reference Morris, Parks and Gardner2020). Interest rates for 2015–2019 were estimated using these authors’ methodology approximating the World Bank’s own procedure. Countries were retroactively classified into the four lending categories based on their historical income levels in relation to the World Bank’s Graduation Discussion Income (GDI) and assigned an interest rate based on the average of fixed- and variable-spread rates for their respective category, which was then added to the contemporaneous LIBOR.Footnote 7 For all lenders, a dummy variable for infrastructure projects was created after qualitatively harmonizing sectors across lenders. Chinese loans were manually coded into one of the 11 World Bank Sectors.Footnote 8 Projects in the following sectors and sub-sectors were then classified as infrastructure: Energy and Extractives (except Mining, and Oil and Gas), Transportation (except Public Administration), ICT Infrastructure, and Water/Sanitation/Waste (individually reviewed projects).
Results
Overall, the results provided in this section demonstrate that incorporating both political factors and pragmatic concerns significantly improve explanations of why LAC countries choose to borrow from Chinese banks versus the United States-backed World Bank. Table 3 presents qualitative results from the semi-structured interviews with policymakers, and Table 4 presents the quantitative results for the mixed-effects logistic regression models. Marginal effects plots for each of the models are included, with Figures 2–4 showing how the predicted probability of a loan being from a Chinese DFI changes over the range of a given independent variable, holding the other variables constant at their means. This section pairs the statistical findings with more detailed interview commentary to illuminate the interactions between the various factors and outline a sequence of decision-making.
Coded Interview Results
Results of Likelihood Models of Lender
*p < .1; **p < .05; ***p < .01.
Marginal Effects of Significant Economic Variables.
Marginal Effects of Significant Political Variables.
Marginal Effects of Significant Project Variables.
The most common reason that interviewees gave for Ecuador’s extensive borrowing from China was a desire to exit from the United States-led global financial system, including the World Bank and IMF. The importance of the project’s sector was nearly as common, with interviewees referencing Ecuador’s need for new energy generation projects and other types of infrastructure, such as transportation. Each of these factors was mentioned by nearly half of the interviewees. As Table 4 shows, countries’ change in UN vote distance from the United States and the infrastructure dummy are both highly significant and positively related to the likelihood that a project will be financed with Chinese loans.
The next most cited factors were also a combination of political and pragmatic considerations: the loan amount and political ideology of the former Rafael Correa government. When interviewees referred to “ideology,” there was overlap in two conceptualizations of the term. First, it often signified the state-led development model requiring large amounts of capital for public investment in major infrastructure projects. Second, it also meant a certain aversion to the market-based global political economy headed by the United States and an affinity towards the pink tide leftist governments in LAC that were challenging this structure. The statistical models are only able to incorporate the former, but interview quotes provided throughout the section illustrate commentary on the latter. Several interviewees mentioned the lack of access or difficulty accessing alternative sources of credit, such as IMF loans or private bonds, as factors in their decision-making, mirroring the supply-side concern of lenders. The sovereign credit risk variable in the statistical analysis is significant across all model specifications, indicating that this remains influential even in the context of other political or pragmatic considerations.
Finally, interviewees discussed several factors that are not readily quantifiable for statistical analysis, including the levels of interest from Chinese companies and Western lenders in Ecuador’s proposals, the types of conditionality that lenders impose, and the expertise that Chinese companies have in certain sectors. A former MPCEIP official explained that the “explicit announcement from Chinese officials of their interest…[in] being a source of public investment” was a key factor in Ecuador’s selection of Chinese loans (ECU 20). Coupled with interest, interviewees cited Chinese companies’ expertise and efficiency as positive factors because the financing typically incorporated direct contracting with Chinese companies to complete the project (ECU 4, ECU 8). In contrast, World Bank loans require a lengthy bidding process to identify companies to execute the project, which are often won by Chinese companies in the end (US GAO 2023).
It is well known that Chinese loans do not come with the same types of conditions as loans from the World Bank; one interviewee explained that one of the advantages of Chinese loans is that “there weren’t policy conditionalities” associated with them (ECU 20). The same interviewee warned, however, that Chinese loans are “not free money” because they do come with commercial conditions such as buying Chinese products or using Chinese technology. Policymakers considered this tradeoff and decided that commercial conditionalities were more favorable for their domestic political dynamics. The rest of this section discusses the results of the regression models in context of additional interview responses.
The economic model (column 1 of Table 4) largely replicates the results of prior studies of the determinants of Chinese development finance, showing that economic stability and credit risk are significant drivers, namely that middle income countries with poorer credit ratings tend to borrow from China. Countries with higher GDP per capita and lower inflation, but poorer governance and creditworthiness, are more likely to borrow from China for a given project. A former MPCEIP official explained that Ecuador’s “access to bond markets was affected by the non-payment of its debts” in 2008, limiting its alternative financing options (ECU 20). This suggests that a convergence of demand- and supply-side considerations, including both deliberate exit from and externally limited access to traditional finance, is responsible for Ecuador’s borrowing patterns. The sovereign credit risk is the only economic factor that retains significance across all three models.
The results of this first model depart from earlier findings in that trade with China is not a significant predictor for borrowing. Additionally, none of the interviewees mentioned trade with China as a factor in their decision-making. Indeed, one official argued that the notion of dependency on China was “incorrect” and instead cited Ecuador’s imports from the United States in areas such as wheat and technology as being more influential in its policy decisions (ECU 5). These findings suggest that trade may be a key predictor for the amounts that China decides to lend but is not a factor for countries in deciding which lender to use. Figure 2 presents these results visually, showing how the probability of a country borrowing from China changes over the range of each independent variable, holding the others constant at their means.
The political model (column 2 of Table 4) demonstrates the importance of domestic and international political factors. The more leftist the political party in power at the time of a loan, the more likely a country is to borrow from China for a given project. Figure 3 illustrates that right-leaning parties have almost a zero percent probability of borrowing from China for a given project. Several interviewees highlighted the overwhelming influence of leftist political ideology for Ecuador’s borrowing decisions, particularly during the ten years of Rafael Correa’s presidency (ECU 1, ECU 3, ECU 8, ECU 5, ECU 19, ECU 16, ECU 20). An MREMH official stated that the ideological component was “stronger” than other factors (ECU 8), while a former MPCEIP official emphasized the “very, very political” nature of these decisions (ECU 20). A member of the National Assembly concluded that “Correa’s socialist government” was the primary reason for Ecuador’s significant debt to China (ECU 16). While Ecuador may be a somewhat extreme case of the influence of political ideology on borrowing decisions, the statistical results support the conclusion that the left-right lean of political parties is relevant across the region. Public opinion is not significantly related to borrowing, consistent with a former government official’s statement that the public does not have much knowledge or influence in foreign policy (ECU 5).
The results of the international political variables provide insights into how geopolitical dynamics and development finance decisions interact. Both China’s relative power and the change in vote distance from the United States are positive and significant. These results indicate that countries are making borrowing decisions based on both the material balance of power and their ideational positioning vis-à-vis the United States. As shown in Figure 3, the greater the movement away from the United States compared to the prior year, the more likely a country is to borrow from China; a one-unit increase (compared to zero change) in vote distance from the United States increases the probability of borrowing from China nearly fivefold.
These statistical results are supported by comments from several interviewees (ECU 1, ECU 7, ECU 19, ECU 20), illustrating that geopolitical considerations are present on the demand-side in addition to the supply side (Flores-Macías and Kreps Reference Flores-Macías and Kreps2013; Landry Reference Landry2018). One former MPCEIP official declared that these decisions were “clearly” more about distancing from the United States, explaining “it was not embracing or pursuing China, but resistance, antipathy towards the US” (ECU 20). Another emphasized that at the beginning, China appeared as an “exit from the international monetary funds” (ECU 7). A former ambassador contextualized these comments by highlighting the association of these traditional lenders with the United States “because of the role the US has in the World Bank and International Monetary Fund” (ECU 2). These results indicate that alignment with China is not the key factor in borrowing decisions, but rather a desire to diversify both borrowing and international relations beyond the “hegemony of the US” (ECU 18, ECU 20).
Finally, the statistical and interview results demonstrate that pragmatic project considerations (column 3 of Table 4) are key to understanding LAC’s borrowing decisions, especially the tradeoffs they’re willing to make. For projects that need higher amounts of financing, countries are more likely to borrow from China. As Figure 4 shows, a $10 billion loan has fairly even chances of coming from either lender, but a loan of $20 billion has a nearly 100% probability of coming from China. A former MPCEIP official explained that borrowing from China is partly because traditional lenders “are never going to give us financing of the same magnitude as China” (ECU 20). Another put it succinctly: “China was a cash machine” that Ecuador could leverage (ECU 27).
Another key driver for LAC states’ borrowing decisions is the project sector; when it involves building infrastructure, countries are significantly more likely to borrow from China. While only 10% of non-infrastructure loans came from China, 37% of infrastructure loans did. Existing literature has demonstrated this relationship can be explained by China’s priorities (Gallagher Reference Gallagher2021; Kaplan Reference Kaplan2021), but interviews provide evidence that LAC countries are strategically allocating their projects. Multiple MREMH officials explained that Ecuador had significant energy deficiencies and sought to meet them through renewable, hydroelectric energy (ECU 3, ECU 4, ECU 8). These priorities were identified in national plans prior to China taking a growing role in project financing (ECU 19, ECU 27). China’s financing was “absolutely key” to accomplishing this goal (ECU 5). A former ambassador emphasized the different advantages of lending countries, where the United States and Europe no longer have the capacity to build projects such as dams, but “China does do it, China is interested” in these projects (ECU 4), again indicating that borrowers see the package deal of financing and contractors as an asset.
Loans with shorter grace periods and higher interest rates are also more likely to have been from Chinese DFIs. This finding clearly reflects China’s preferences as a lender. Interview evidence is illustrative in explaining why countries are willing to accept these less favorable terms: high-level political preferences for loans from China over the World Bank (ECU 5, ECU 18, ECU 20, ECU 28). The negotiation of terms such as interest rates, grace periods, and maturities occurs only after political leaders have identified a preferred lender. In Ecuador, the MEF debt committee led these negotiations, and the political constraints they faced are consistent with prior research demonstrating the limits on technocrats from ideologically driven actors (Cormier Reference Cormier2024).
This sequencing of demand-side factors is key to understanding borrowing outcomes. Interviews indicated that the political factors identified in this article—party ideology and an exit from the United States-led multilaterals—play a greater role at the beginning of the process, while more pragmatic considerations such as the type and terms of financing have more influence later. At an international level, Ecuador first sought an alternative to borrowing from the World Bank and IMF where the United States has the “strongest voice” (ECU 18); it was not specifically China that they were pursuing (ECU 20). Domestically, the “project requirements come from the Ministry of Foreign Affairs,” where delegations on official visits to potential lenders seek financing for different projects (ECU 28). An official from Correa’s governments emphasized the interaction of political considerations and loan amounts, explaining that the political aspects of loan negotiations come “at the beginning when we’re talking about the scale of financing” (ECU 5).
Once political actors and interests had initiated the borrowing process, the MEF debt committee led the negotiations, only calling on the Ecuadorian embassy in China and MREMH when they hit an obstacle. Interviewees largely agreed that when it comes to the specifics of the loan terms, such as the interest rate and repayment scheme, the MEF handles them (ECU 5, ECU 7, ECU 12, ECU 29). A former official described the MEF’s debt committee’s consideration of factors such as the average cost of existing debt, the operational cost of the proposed loan, historical borrowing limits, and available market-based options (ECU 5). One of the advantages interviewees cited of negotiating with China was the lack of political conditionalities (ECU 12, ECU 20). However, commercial conditionalities did present challenges during negotiations. In these instances, “when [the MEF] needed support, they called us” at the MREMH to use diplomatic channels (ECU 12), again suggesting a central role for high-level political actors. An official familiar with the negotiation process stated that contrary to the perception that China will easily lend to any country, accessing credit from Chinese DFIs was “very difficult” (ECU 4). Interviewees cited Chinese DFIs’ insurance requirements (ECU 4), commercial interests (ECU 12), and desire to charge high interest rates “without going beyond what the [borrowing] country could pay” to ensure a return on their investment (ECU 27). On the other hand, Ecuador insisted that a “limit of 20–25%” of any project workforce could come from China, with most jobs reserved for the local labor force, a condition that interviewees perceived as frustrating to China’s negotiators (ECU 3, ECU 12).
Finally, the results demonstrate that in development finance, great powers do not dictate terms to their borrowers. One MREMH official said that observers often conclude that their borrowing from and close relations with China were due to “pressure from China, but it was not like that” (ECU 1). Another MREMH official explained that China has been “very receptive to our requests” for financing, challenging the argument that Chinese loans are pushed onto developing countries (ECU 3). A former ambassador argued that many of these criticisms neglect to consider the global context confronting Ecuador and other developing economies in the wake of the 2008 financial crisis that “slammed the capitalist-Western world” (ECU 4). In that moment, when Ecuador needed financing for its national energy plan, China was able to not only offer capital, but also “technology and companies.” Even in this challenging global context, a member of the National Assembly emphasized that the “US did not abandon us…Correa threw them out” (ECU 16). An official from the Correa government described this strategy as “diversification to achieve our objectives” (ECU 5). Even interviewees who were highly critical of the Correa governments’ decisions to borrow heavily from China acknowledged that this was at least partly a deliberate choice after “fighting with the US, IMF, and World Bank” (ECU 19).
These comments underscore the necessity of combining both supply- and demand-side factors to achieve a full explanation of China’s development finance in LAC. Lenders and borrowers factor in sovereign credit risk, but also consider project-specific needs such as the amount of financing, the cost of the loan, and whether the lender and potential executing companies have the requisite interest and experience. Framing these pragmatic concerns are the domestic and international political environments, where a left-leaning party may be more inclined to borrow from China, especially when officials seek to reduce their dependency on United States-led DFIs.
Conclusion
This article set out to empirically analyze the political and pragmatic determinants of Chinese development finance in LAC. Prior work on this question has emphasized China’s preferences as a lender, with scholars only recently beginning to incorporate LAC states’ agency into the equation. This article bridges literature on the determinants of Chinese development finance, LAC foreign policy, and power transition theories. It shows that domestic politics within borrowing countries, their foreign policy alignment with major powers, and their specific project goals are all significant predictors of their borrowing decisions. Political parties with a more leftist lean are more likely to finance development projects with loans from Chinese banks. Beyond simply bandwagoning with China, countries seek to diversify beyond United States-led financial institutions. Finally, pragmatic considerations such as the amount of available financing and the project sector are very strong predictors for where a state will borrow.
These findings do not negate the role of lenders’ preferences and external limits placed on borrowing countries, but they do demonstrate that borrowers hold and express preferences that are an additional and important part of the story explaining China’s development finance. These demand- and supply-side factors are often complementary, as in the case of Ecuador, where it faced challenges to accessing external market-based finance, but it also sought to intentionally reduce its reliance on what policymakers saw as a United States-led hegemonic financial system. Incorporating supplementary interview evidence from Ecuador study bolsters the theoretical lens of borrower agency that is more complicated to achieve in the statistical analysis, due to these complementary dynamics.
This article has contributed a more nuanced picture of the drivers of China’s development finance, especially the micro-level insights into the decision-making process within borrowing countries. It shows that the story is not a simple case of countries taking whatever they can get, even for countries like Ecuador with constraints on their access to financing. The findings have important implications for understanding how developing countries navigate power transitions, a key question for the prospects of global stability in the coming decades as the United States and China compete for global influence both with and beyond development finance.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/lap.2024.54
Acknowledgments and Funding information
I am grateful to Rebecca Ray, Kevin Gallagher, all the Global China Fellows, and two anonymous reviewers for their insightful feedback. Special thanks to Lauren Mattioli, Catherine Abou-Khalil, Chas Walker, and Si Wu for comments on the early versions. The research for this working article was completed with grants from the Boston University Graduate Research Abroad Fellowship and the Boston University Center for Latin American Studies David Scott Palmer Grant.
Competing interests
The author declares none.
Open access
