Global banking has changed substantially in the wake of the global financial crisis, as new regulation, stricter supervision, and stronger risk management prompted international banks to scale back foreign activities (Fratzscher and Bremus 2015, Claessens 2017, De Haas and Van Horen 2013, 2017).

A prime example of this retrenchment is the sharp decline in correspondent banking. Driven by stricter anti-money laundering and counter-terrorism financing rules, many global banks have curtailed their correspondent banking services, especially in low-income countries.

Correspondent banks hold deposits from other banks (respondents) and provide them with cross-border payment services, enabling the local banks of exporters and importers to make trade-related payments without direct bilateral account relationships. 

Correspondent banks also offer trade finance services, such as letters of credit (Schmidt-Eisenlohr and Niepmann 2016), which mitigate non-payment or non-shipment risks. Through these activities, these banks enable trade transactions that might otherwise not occur due to financial constraints or a lack of trust between counterparties.

Correspondent banking's cross-border nature makes it vulnerable to abuse by criminals who exploit national differences in legislation, bank secrecy laws, and enforcement to disguise illicit funds. The fight against financial crime has suffered from the weak enforcement of legal frameworks designed to counteract it (CGD 2015). However, after the global financial crisis, the prosecution of such offenses intensified. In 2014, the US Department of Justice signalled that any global transaction threatening the integrity of the US financial system could be tried in a US court. This was widely regarded as a turning point, heralding stricter regulatory enforcement in anti-money laundering and counter-terrorist financing. In response to this shift, banks reassessed the cost of regulatory compliance in correspondent banking. They subsequently undertook severe pruning of their correspondent networks, terminating relationships deemed no longer cost-effective or too risky (Financial Stability Board 2017, Rice et al. 2020).

This retrenchment has led to a significant reduction in the number of correspondent banking relationships worldwide. Figure 1 visualises this de-risking, showing the Gini coefficient of active correspondent banks per country-pair between 2012 and 2022. Until 2014, the coefficient is stable, after which it increases steadily, reflecting growing concentration as more banks withdrew. In a recent paper (Borchert et al. 2024), we investigate the firm-level consequences of this correspondent bank retrenchment.

To identify the impact of the withdrawal of correspondent banks on firm activity, we join three datasets. Information on the loss of correspondent bank relationships comes from two proprietary surveys among respondent banks in emerging Europe: the third Banking Environment and Performance Survey (BEPS III) and a survey we conducted together with EBRD’s Trade Facilitation Programme. We link these bank-level data to information about firms’ main bank as reported by Bureau van Dijk’s Orbis database. We then match this information with additional firm-level data from Orbis. We focus on four emerging European countries – Bosnia & Herzegovina, Croatia, Hungary, and Turkey – which have traditionally relied heavily on correspondent banking services.

Using a staggered difference-in-differences approach, we then systematically compare the export performance of firms with a main bank that loses a correspondent banking relationship (treated firms) to similar firms with a main bank that has not lost any correspondent relationship up to the event year (control firms). We match treated with observationally similar control firms in terms of their pre-event export turnover, total assets, and total factor productivity, and keep all firms on the common support.

We find that a decline in the supply of correspondent banking services negatively affects both the extensive and the intensive export margins. Figure 2 shows that once a firm’s main bank loses a correspondent bank relationship (t=0), it is more likely to stop exporting or to export less. At the extensive margin, the likelihood that a firm continues to export declines by 5.2 percentage points in the short term and by 19.8 percentage points in the medium term (four years after the shock). At the intensive margin, four years after the event, the export revenues of affected firms are 57% lower than those of similar control firms.

Looking at other real outcomes, we find that when correspondent banking relationships are terminated, some affected firms can compensate for the resulting export decline by boosting domestic sales. However, many other firms experience a decrease in total revenues, lay off employees, or go out of business entirely.

We also assess the relevance of these disruptions at the level of affected villages, towns, and cities (‘localities’) to provide evidence on the local equilibrium effects of terminated correspondent relationships. We find that the negative impacts of terminated correspondent relationships on firm-level trade reverberate through the local economy. When comparing localities where many bank branches lost a correspondent bank to localities without such disruptions, local exporters on average exhibit significant export declines on both the extensive and intensive margins.

The negative outcomes of the decline in the supply of correspondent banking services on firms’ export activities are most pronounced among smaller and younger firms. These firms find it especially difficult to switch to other banks when their main bank can no longer offer the necessary services to facilitate international trade transactions. Additionally, firms whose bank had relatively few correspondent relationships to begin with suffer more, as the withdrawal of even a single correspondent bank can leave their main bank with insufficient cross-border payment and trade finance options. This underscores the importance of a diversified network of correspondent banking relationships in mitigating the adverse effects of correspondent bank withdrawals on firms.

We expand our analysis to 17 emerging European markets using bilateral sectoral trade data to support the external validity of our firm-level results and to examine whether firm-level export declines aggregate to the industry level. 

Moreover, at the industry level, we also have data on imports available. We exploit the tightening of the US regulator’s enforcement of financial crime legislation in June 2014 as a structural break that negatively shocked the supply of correspondent services. 

The sector-level results confirm our firm-level evidence: the export growth rate is, on average, 8 percentage points lower in countries with a high withdrawal in correspondent banking than in countries where no or only few correspondent banks left. Similarly, import growth for high-withdrawal countries decrease an additional 24 percentage points relative to low-withdrawal countries.

This column has provided novel evidence on how the recent sharp decline in correspondent banking relationships, triggered by growing regulatory compliance costs, is hindering firms’ ability to engage in international trade. Our findings demonstrate how abrupt increases in regulatory scrutiny can unintentionally disrupt vital correspondent banking networks, resulting in significant and lasting negative impacts on firms’ export opportunities and local economic dynamism.

The impact of broken correspondent relationships may be long-lasting, as rebuilding local knowledge and international trade links takes time. In the meantime, government-backed trade finance schemes can help mitigate the fallout (Kabir et al. 2024). 

In the longer term, efforts could focus on two key areas: first, improving respondent banks' compliance with international financial-crime regulations and risk-management procedures; and second, exploring new digital technologies to facilitate safe and efficient cross-border payments (Panetta 2023).

Relatedly, efforts can be intensified to develop centralized databases that allow correspondent banks to verify the legitimacy of their respondent banks' customers. The wider adoption of standardised legal entity identifiers (LEIs) can also help to improve transparency and facilitate due diligence. 

By addressing these issues, the international banking community can work towards rebuilding trust within the correspondent banking network, ultimately supporting global trade and economic growth.

By Lea Borchert  & Ralph De Haas