The Role of Export Credit Agencies in Trade Around the Global Financial Crisis: Evidence from G20 Countries

2.1 Stimulus Packages and Global Trade

Following the 2008 GFC, various countries implemented stimulus packages to mitigate the economic downturn. The rapid escalation of the crisis verified the strong ties of globalization. The scale of this crisis is beyond estimation, and Eichengreen and O’Rourke (2010) compared it to the Great Depression regarding the volume of world trade. Many countries announced massive stimulus measures during the GFC, with global stimulus spending estimated at between 1.7 and 2% of the world’s GDP (Khatiwada, 2009). The United States, China, and Japan accounted for 62% of the total stimulus packages implemented in 2009 (ILO, 2011).

During the crisis, G20 nations, representing over 75% of world GDP and 62% of world trade, intervened in markets and trade to a certain extent by implementing national economic stimulus programs. The total stimulus amount provided by G20 countries in 2009 was approximately $692 billion, equivalent to around 1.4% of their combined GDP. Of this total, 39% was contributed by the United States, 13% by China, and 10% by Japan. The composition and speed of implementing these stimulus packages varied across countries – some focused on tax cuts, while others prioritized increased government spending (Prasad & Sorkin, 2009). Specifically, at the national level, many countries aimed to stimulate their economies through large-scale measures such as infrastructure investments, increased public spending, and strengthened social welfare. In contrast, at the financial sector and corporate levels, efforts included bank recapitalization, provision of guarantees, corporate restructuring, support for SMEs, and subsidies to preserve employment (Lin et al., 2014). The crisis affected international trade through supply and demand channels, with GDP growth in destination countries being a crucial determinant of export and import demands (Behrens et al., 2013). In response, the G20 pledged $250 billion to support trade finance through ECAs and multilateral development banks (G20 London Summit, 2009). According to IMF analyses, corporate profitability in both advanced and emerging economies increased significantly following the fiscal stimulus measures implemented in the aftermath of the GFC (Correa-Caro et al., 2018)

2.2 Financial Challenges and Support for Exporters in Crises

Participation in export markets improves the financial health of firms (Greenaway et al., 2007). However, exporters are more reliant on short-term financial liabilities due to higher variable costs, risks, working capital requirements, and sunk costs compared to non-exporters (Maes et al., 2019; Mansilla-Fernández & Milgram-Baleix, 2023; Melitz, 2003). Consequently, any liquidity constraints fundamentally alter exporters’ behaviors and may limit or even prevent some firms’ ability to export (Chaney, 2016; Manova et al., 2015). This effect was also evident during the GFC. For instance, Maes et al. (2019) suggested that “The strong reliance of exporting firms on short-term (asset-backed) funds to refinance their export activities may serve as an explanation for a trade collapse during credit crunches or in periods of low profitability.” Moreover, Chor and Manova (2012) demonstrated that countries with higher credit costs and tighter credit conditions exported less to the United States during the financial crisis.

During the crisis, surveys by the IMF and other organizations indicated that bank-intermediated trade finance declined in value, though not as sharply as merchandise trade (Asmundson et al., 2011). Despite the recognized importance of exports and the vulnerability of exporters, interest in international trade finance was very limited until the 2008–2009 crisis (Auboin & Engemann, 2012). Trade finance is often characterized as both a facilitator of trade and a shock absorber (Irwin & O’Rourke, 2013; WTO, 2016). Inekwe et al. (2018) found that financial distress among US businesses led to a 14% decline in exports, an 11% reduction in investment, and a 9% decrease in GDP growth.

During periods of financial difficulty for exporters, providing financial support for these companies, which act as catalysts of national economies, becomes crucial. Indeed, export-promotion institutions tend to adjust their support based on local and global economic conditions, offering more aid during recessions and less during strong economic growth (Pýcha, 2022). According to the Organization for Economic Cooperation and Development, government institutions or private companies acting on behalf of governments provide officially supported export credits to national exporters. This support includes direct credits to foreign buyers, refinancing, interest-rate support, and insurance or guarantees for credits from private financial institutions, facilitating competition in international markets (OECD, 2024).

2.3 ECAs: Institutional and Financial Structures and Empirical Studies

No uniform structure characterizes ECAs globally since their operational scopes and services vary considerably. Some ECAs restrict their services to providing insurance or guarantees, while others extend their offerings to include loans. In some cases, they provide both services. These agencies are often categorized as either insurers or Eximbanks, depending on their primary functions. Regarding insurance coverage, some agencies specialize exclusively in export credit insurance or investment insurance, whereas the more substantial agencies typically provide both. Furthermore, while some ECAs focus primarily on insuring against political risks, others concentrate on commercial risks, although it is common for larger agencies to manage both risk categories (Stephens, 1999).

Since 1978, the OECD’s Arrangement on Guidelines for Officially Supported Export Credits has regulated ECAs to promote fair competition in the export credit sector, although it is not legally binding. The Competitiveness Report (Export-Import Bank of the United States, 2019) revealed that only 34% of all export credits adhere to this OECD Consensus globally. Competing ECAs from various countries strive to offer their exporters the most favorable conditions possible, suggesting that competition among ECAs could be a reason for the weak adherence to the Consensus. However, Agarwal and Wang (2018) found no evidence that financing provided by ECAs of competing countries increased competition in the global market for US exporters. They also claimed that their findings were not influenced by whether the competing countries were members of the OECD Consensus or received EXIM support. Moreover, Dawar (2020) argued that the current economic slowdown in export growth and the uncontrollable rise in export credit support programs, especially among developed country ECAs, urgently require increased cooperation.

The legal structure, capital source, and risk management frameworks of ECAs are critical to their long-term sustainability. Global best practices recommend that these institutions be structured as agencies under public law and operate independently without relying on government subsidies (Klasen & Janus, 2023). Indeed, most ECAs are either state-owned or under government control. However, their governance structures vary significantly across countries (Pýcha, 2022). Many ECAs face challenges in balancing their mandate to serve government policies with the pressures of operating like a bank, as shifting toward a profit-driven structure risks overlooking their social mission (Groth, 2019).

While ECAs are typically funded through the state budget or receive capital support from the government, some also possess additional capital mechanisms. However, in some countries, operating ECAs at a loss by offering low premium rates to provide exporters with more favorable financing poses a risk regarding the effective and responsible use of public resources (Klasen & Janus, 2023).

Regulations introduced particularly after the GFC have led commercial banks to adopt a more cautious approach toward international lending, thereby increasing the role of ECAs in export financing. Traditionally acting as “lenders of last resort,” ECAs have become more proactive, expanding their financing programs and stepping into roles once dominated by private-sector lenders (Congressional Research Service, 2019). However, since the OECD Arrangement does not cover non-member countries such as China, Brazil, and India, these nations are able to offer financing below market terms. This creates competitive challenges for ECAs in OECD member countries and raises concerns about the overall effectiveness and fairness of the arrangement (Ilias, 2012).

According to recent studies, many countries have increasingly favored the independent public agency model for their ECAs due to its greater flexibility and customer-oriented approach in terms of institutional structure. However, the financial empowerment frameworks of these agencies vary depending on the country’s economic size and export volume (SERV, 2020). Nevertheless, the number of underperforming ECAs remains significant and cannot be overlooked (Yazdi et al., 2019)

Numerous studies support the positive impact of ECAs on countries’ and firms’ exports. Moser et al. (2008) conducted a study covering the period from 1991 to 2003 in Germany and documented that public export guarantees had a statistically significant positive impact on exports (Blackmon, 2016). Badinger and Url (2013) investigated Austrian exporting firms and found that export credit guarantees significantly influenced trade among these entities. Similarly, Choi and Kim (2021) observed in South Korea that short-term export credit insurance mitigated financial constraints for firms, consequently enhancing exports. This effect was notably more pronounced in exports originating from developing countries or smaller firms. In a noteworthy study from a developing country context, Polat and Yeşilyaprak (2017) analyzed Türkiye’s exports to 212 destinations, revealing that a 1% increase in export credit insurance led to a rise in Türkiye’s exports by between 3 and 17%. These findings emphasize the pivotal role of ECAs in bolstering national export volumes. Moreover, they highlight how ECAs can significantly enhance the competitive position of countries on the global stage by demonstrating their importance in international trade dynamics.

It is difficult to assert that ECA facilities positively impact exports, employment, growth, and global welfare. Indeed, Soh (2014) reported that while ECA-backed export credit support positively impacts exports in the long term, its presence in a country does not correlate with higher economic growth or improved employment rates. In contrast, Egger and Url (2006) found that export credit guarantees provided by Austria’s ECA had a significant, albeit modest, short-term effect on exports. Janda et al. (2010), in their study on the Czech Export Bank, showed that the credit support provided by the bank had a positive but statistically weak effect on exports. Hur and Yoon (2024), in their study covering South Korean firms for the 2006–2015 period, found that while both export credit and insurance increased exports, only credit support positively affected employment growth. Additionally, concerns have been raised that ECA credits may distort fair competition in international trade, support environmentally harmful activities, and potentially contribute to human rights violations (Darbellay, 2021; Dawar, 2020; Kim, 2020; Linder, 2019; Shearing, 2013)

Studies examining the impact of ECAs during crises between the 1990s and the first decade of the 2000s are almost nonexistent. Because in the 1990s, government-backed ECAs largely withdrew their credit provisions as the private sector began to offer short-term guarantees (Blackmon, 2016). However, private banks rapidly reduced short-term export credit facilities during the GFC. As a result, following the decline in international demand, the contraction in trade finance became the second most significant factor behind the slowdown in global trade (Mora & Powers, 2009). Indeed, in the fourth quarter of 2008, medium- and long-term trade finance transactions declined by approximately 40% (Chauffour et al., 2010). Thus, during the crisis, ECAs proved to be a vital component in “greasing the wheels” of international trade (Blackmon, 2016). Some of the ECA-related studies focusing on the GFC period are as follows: Utilizing the 2008 Financial Crisis as a dummy variable, Köksal and Genç (2019) reported that in 22 high-income countries, export credit insurance notably increased exports, with a significant impact observed for medium to long-term insurance policies. During the GFC, Felbermayr and Yalcin (2013) found that such a crisis mitigated the decline in German exports. Auboin and Engemann (2014), utilizing data from 91 countries between 2005 and 2011, found that insured commercial credits from ECAs had a strong positive impact on trade in both crisis and non-crisis periods.

A general literature review shows that although ECA-backed financing typically produces positive outcomes at both national and global levels, it also yields weak results for certain macroeconomic variables in some countries. Several prior studies have explored the role of ECAs during periods surrounding the GFC. However, existing studies are either geographically limited or differ from the present study’s design. For instance, Moser et al. (2008) and Felbermayr and Yalcin (2013) focus exclusively on Germany. Auboin and Engemann (2014), whose timeframe is more comparable, analyze the period from 2005 to 2011 to examine trade dynamics before, during, and after the crisis. However, their approach differs in two ways: they use an unbalanced panel covering 91 countries and adopt an import-oriented perspective. In contrast, the current study focuses on the G20 countries, which collectively account for approximately 85% of global GDP, 75% of international trade, and two-thirds of the world’s population. Moreover, we adopt an export-oriented lens, offering a more targeted and policy-relevant contribution to the literature. This distinction positions our study to fill a clear gap in existing research – namely, the lack of a systematic, comparative analysis of ECA impacts on exports across critical crisis periods within the world’s most influential economies. In addition, the study can also be seen as a reflection of the Keynesian approach to financial markets, which aims to compensate for market failures and stimulate aggregate demand during extraordinary times.

3.1 Data Nature and Source

We briefly describe the dataset before moving on to the econometric analysis – to reveal the direct insurance or lending impact of ECAs on the export of G20 countries during the GFC. The sample dataset contains 17 countries (excluding Saudi Arabia and Indonesia due to missing variables) and spans from 2005:Q2 to 2020:Q2. All variables, except GFC, are transformed into natural logarithmic form to provide reliable and consistent estimates (Table 1). This transformation addresses issues such as heteroscedasticity or non-linearity. Economic and financial data, particularly those related to trade flows, GDP, and exchange rates, often exhibit heteroscedasticity, where the variance of the error terms increases with the magnitude of the variables (Lütkepohl & Xu, 2012). Box et al. (2015) discussed the log transformation as a means to stabilize the residual variance, leading to a more adequate model for forecasting. A potential drawback of using log transformations is that they are undefined for zero or negative values. To mitigate this issue, we carefully examined our dataset and ensured that none of the variables used in the regression analysis contained such problematic values.

3.2 Econometric Model

Currently, one of the cornerstones of empirical trade theory is the gravity equation. Formulated by Jan Tinbergen in 1962, the theory connects the volume of trade between countries to their economic size, geographical distance, and relative trade barriers. Its theoretical foundations were extended by Anderson (1979) and further developed through subsequent contributions by Bergstrand (1985, 1989) and Deardorff (1998). The basic gravity model of Tinbergen (1962) is defined as

where GDP and distance between countries determine the export of i to j, and ϵ _ij refers to the stochastic error term in the model. We measure the distance based on trade routes between the most populous cities of countries, such as New York instead of Washington for the United States, or Istanbul instead of Ankara for Türkiye. Generally, the traditional gravity model considers the distance the airline measures; however, nearly 80% of international trade in goods is carried by sea (UNCTAD, 2022). For this reason, as in the paper of Simdi and Unal (2022), the study uses the distance of the SEARATES platform (SEARATES, 2024) because it provides the total trade route time between cities in hours. The trade route (by taking sea routes and road distances in terms of hours) provides a more precise assessment of the impact of distance on exports. Most traditional gravity models rely on physical linear distance. Consistent with these models, the expected effect of distance remains negative. However, our study aims to enhance awareness regarding the accurate measurement of trade routes by incorporating travel time as a distance variable. The GDP and distance variables of our gravity model are calculated by considering the export shares of partner countries. For example, the total export of Germany in 2020Q4 is $387 billion, and the export share of the top 20 trade partners equals 79%, equaling 100% for each to calculate the “GDP” variable. The same method has been applied for the “Distance” variable (Table 2).

Therefore, the partner country GDP variable for Germany 2020:Q4 is as follows:

= 11.1% × GDP of USA in 2020Q4 + 10.5% × GDP of China in 2020Q4 + 9.4% × GDP of France in 2020Q4 + …

The study considered the same calculation method for the distance variable as used for Germany 2020:Q4:

= 11.1% × Trade distance between Berlin and New York + 10.5% × Trade distance between Berlin and Shanghai + 9.4% × Trade distance between Berlin and Paris + …

Complying with the objective of this study, the basic gravity model is augmented with more variables to increase the explanatory power of the regression and is expressed as follows:

The augmented gravity of the model includes REEXC, GFC, and export credit exposures to explain exports. Changes in exchange rates have been used as an explanatory variable in augmented gravity models (Egger, 2002; Egger & Pfaffermayr, 2003; Martinez-Zarzoso & Nowak-Lehmann, 2003). An increase in REEXC means the domestic currency is appreciating, which is expected to reduce exports.

The study employs Phillips–Perron (PP) and Harris Tzavalis unit root tests to prevent spurious results, confirming that the variables are either stationary at levels (I(0)) or after first differencing (I(1)) (Table A3). Since the sample variables are stationary at I(0) and I(1), we intend to incorporate the ARDL method into the augmented gravity model. Because it is a well-established framework in trade analysis and allows us to integrate key variables relevant to our study – such as ECA support, REER, and the GFC dummy – to better explain export performance during crisis and non-crisis periods. This approach was initially proposed by Charemza and Deadman (1997) and further developed and advocated by Pesaran et al. (1999, 2001) due to its multifaceted advantages, including suitability for small sample sizes, flexibility in accommodating variables integrated of order zero [I(0)] and one [I(1)], and the ability to specify different lag lengths for different variables (Ali et al., 2017; Rahman & Kashem, 2017).

Based on the ARDL framework, this study constructs an unconstrained error correction model that articulates both long-term and short-term dynamics among the sample variables. To justify it, the Kao Residual Cointegration Test is performed, which confirms the presence of a long-term equilibrium relationship (ADF t-statistic = –7.724326, P-value = 0.0000). This supports the validity of estimating both long-run coefficients and the error correction term using the Panel ARDL framework. The ARDL equation is formulated as follows:

where ε i t represents the error term assumed to be white noise, Δ denotes the first-order difference operator, and p is the number of lags. The terms in parentheses represent the long-run equilibrium relationship, and the coefficient ϕ is the error correction term, which captures the speed at which the system returns to equilibrium after a shock. The coefficients β ₁, β ₂, …, β ₈ are the long-run coefficients of the explanatory variables. The coefficients γ i j (for i = 1, …, 9, j = 1, …, p) represent the short-run dynamic effects of changes in the explanatory variables over the lag structure.

3.3 Variables and Model Justification

In this study, the standard gravity model structure is not considered due to specific concerns around the endogeneity of insured trade credits as an explanatory variable (Auboin & Engemann, 2014). This potential source of endogeneity arises from trade credits and trade flows may be determined simultaneously – higher trade volumes could increase demand for trade finance, and higher availability of trade finance could increase trade flows.

Our analysis differs from the standard gravity model approach because of the nature of the sample data. While traditional gravity models usually focus on bilateral trade flows, we considered aggregated short-term insured trade credit data from exporting countries. To account for these data restrictions and better tackle the endogeneity issues stemming from trade finance, this study relies on a modified version of the classical import estimation equation, using trade credit insurance as an additional explanatory variable following Brandi and Schmitz (2015).

In addition, as this study only uses the total amount of export credit instruments provided by ECAs in the exporting country rather than data specific to each bilateral trade flow, the econometric model considers the total exports from the home country as the dependent variable (Auboin & Engemann, 2014). This enables the research to account for the aggregate effects of ECAs providing export finance through support of export performance without breaching data availability restrictions or presenting additional endogeneity concerns related to bilateral trade inconsistencies.

We employed the automatic lag order selection feature to justify the model selection, which selects the best-fitting model based on the Akaike information criterion (AIC). This approach ensures optimal model specification by choosing the lowest AIC score, resulting in the ARDL (4, 4, 4, 4, 4, 4, 4, 4, 4) model. The ARDL framework’s ability to estimate both short-run and long-run relationships simultaneously makes it ideal for examining the dynamic impact of ECAs on exports before, during, and after the GFC. Additionally, the study verified that all variables were appropriately tested to avoid including I(2) variables. The appropriate modification of ARDL model orders is sufficient to tackle serial correlation and regression endogeneity problems (Gemmell et al., 2016; Kanas & Kouretas, 2005). An ARDL model with sufficient lags addresses the endogeneity problem, provided that the regressors are not cointegrated among themselves and the primary focus is on the long-run parameters (Pesaran et al., 1999).

4.1 Results

4.1.3 Model Selection Summary

Figure 1 reports the AIC scores for various specifications of an ARDL model. The x-axis represents unique model configurations, numbered from 1 to 16, while the y-axis shows the corresponding AIC scores, where lower scores indicate a model that effectively balances goodness of fit with simplicity to prevent overfitting. The model with the lowest AIC score (Model 16; ARDL [4, 4, 4, 4, 4, 4, 4, 4, 4]) is the most preferred due to its optimal trade-off between model complexity and fitting accuracy. This is especially relevant due to small to moderate-sized samples of the study, where AIC offers superior predictive accuracy compared to other criteria like BIC, which heavily penalizes models with a larger number of parameters (Burnham & Anderson, 2004). Unlike BIC, which emphasizes parsimony, AIC focuses more on minimizing information loss, making it more suitable for accurately capturing the dynamic relationships among variables. The highest log likelihood score of 2,146.674432 demonstrates the best data fit among the considered models. Given the parameter estimates, this score also indicates that the model has the most significant probability of generating the observed data. Its superior AIC value of −3.115943 indicates an efficient balance between model accuracy and simplicity (Table A1).

Figure 1

Model selection criteria. Source: Authors.

While AIC is prioritized, the model’s BIC (0.089610) and HQ (−1.895749) scores are also relatively low, suggesting that the model is not overfitted and remains robust even when penalizing for complexity.

4.2 Discussion

Our findings reveal multifaceted and significant insights into the impact of insured export credit exposures on trade (export). The medium-term and short-term export credit support have both strong positive and negative effects on trade in the long run, respectively, but not in the short run during the sample period (supported by Soh Young (2014)). However, this effect does not vary during the GFC period. Notably, the total insured ECA had a negative impact in the long run and a positive impact in the short run (Köksal and Genç (2019), Auboin and Engemann (2014), and Egger and Url (2006) found a similar result, but not specific to the crisis). Considering the short-run equation, about 0.25% of the departure from the long-run equilibrium is corrected each period (quarter). Several theoretical and conceptual explanations exist for why short-term ECA instruments may negatively affect firms’ long-term export performance. Crisis-driven ECA support can trigger sudden export surges, but once withdrawn, exports often decline, revealing weak underlying competitiveness. Echoing Friedman’s (1969) “helicopter drop” analogy, indiscriminate ECA credit – especially to inefficient or “zombie” firms – can increase moral hazard and reduce incentives for innovation. Moreover, ECA-backed financing may crowd out market-based alternatives, weakening firms’ exposure to market discipline. In the long run, prolonged or poorly targeted ECA interventions can distort domestic markets and harm global trade efficiency.

During the financial crisis, 11 countries experienced a positive short-term impact on exports. In contrast, Brazil, France, India, Türkiye, and the United Kingdom faced adverse effects, with three of which are developing economies. Chauffour et al. (2010) emphasized that developing countries should only rely on ECAs when specific preconditions, such as financial capacity, institutional capability, and governance, are adequately met. This highlights the difficulties developing nations encounter in effectively utilizing ECAs during times of crisis. Furthermore, Aydemir and Gerni (2011) suggested that ECAs in developing economies may require structural reforms to enhance their effectiveness in managing economic shocks. Therefore, we recommend restructuring ECAs in developing countries to address institutional weaknesses and strengthen their crisis management capacities.

Regarding cross-section short-term analysis, country-wise short-term error correction process reports France’s rapid adjustment capability, followed closely by Japan, the United Kingdom, and Russia, indicating robust error correction mechanisms in response to deviations from long-term equilibrium. These results are essential, as they suggest that these countries are highly responsive to economic shocks, potentially due to well-established financial systems or effective economic policies.

Generally, while policymakers often credit ECAs’ risk-bearing capacity to their guarantee schemes, the actual acceleration to exports is predominantly driven by direct credit programs. Because direct support mechanisms enable exporters to access financing directly and swiftly, especially in high-risk markets or where intermediary banks are reluctant to provide funding, ECAs offer flexible solutions to promote exports. Additionally, direct support helps reduce financial intermediation costs. However, it is crucial to note the drawbacks associated with these schemes, notably their potential to directly increase public debt and their functional overlap with private financial sector activities. Based on the findings of this research, although ECAs offer significant advantages in managing exporter risks, reliance on direct credit schemes should be carefully evaluated due to their fiscal implications and possible redundancy with existing private solutions.

The OECD’s Arrangement on Guidelines for Officially Supported Export Credits also fails to prevent finance-based unfair competition among countries. Establishing more binding and enforceable principles at the global level is critically important.

From the theoretical aspect, the findings of this study contribute to the literature on international trade and economic recovery by explaining how different nations’ export capabilities and financial systems adjust in the face of global economic fluctuations. The implications are significant for policymakers and financial institutions. For countries with slower adjustments, targeted reforms in financial regulations or more aggressive monetary and fiscal interventions may be necessary to enhance responsiveness to global economic shifts. Understanding these dynamics for international businesses and investors could guide strategic decisions about where to allocate resources most effectively during economic downturns, enhancing risk management.