Bank funding: evolution, stability and the role of foreign offices

The Covid-19 pandemic and the war in Ukraine have furthered a sustained retreat from global banking. The funding sources of local banking systems have shifted from cross-border to local and, within cross-border, from inter-office to unrelated sources. An increased share from local sources could improve the stability of funding, but greater cross-border funding from unrelated sources could be detrimental, as it is more sensitive to global shocks than is inter-office funding. The shift in the funding mix reflects salient differences between the funding models of domestic banking offices, which have been gaining in importance, and foreign banking offices – ie subsidiaries and branches with foreign parents. Furthermore, as individual foreign bank affiliates have withdrawn, the effect has been to increase the nationality concentration in local banking systems, which historically has correlated with higher funding volatility.¹

JEL classification: F34, G15, G21.

The Covid-19 pandemic and the war in Ukraine have given further impetus to a generalised retreat from global banking that began in the aftermath of the Great Financial Crisis (GFC). A retrenchment of foreign bank offices (FBOs) – subsidiaries or branches whose parent is in another country – changes the structure of a local banking system, not least because their funding models differ from those of domestic bank offices (DBOs). In this article, we document changes in the funding sources of local banking systems, trace the effects to the relative evolution of FBOs and DBOs and draw implications for funding stability.

Our analysis uses the BIS international banking statistics. We define a local banking system to comprise all banks located within a given jurisdiction (Graph 1). Thus, such a system includes both FBOs (eg French banks operating in Brazil) and DBOs (eg Brazilian banks operating in Brazil). We group bank funding sources into three broad categories: local funding (from residents), cross-border inter-office funding (from related banks), and other cross-border funding (from unrelated sources, eg banks, non-bank financial entities or non-financial corporates). We cannot include in most of the analysis countries that do not report local positions in the BIS international banking statistics, such as China and the United States.²

There are four key findings. First, the mix of local banking systems' funding sources has been shifting in recent years, from cross-border to local and, within cross-border, from inter-office (ie from related banks) to unrelated sources. Second, this shift has implications for funding stability, as local funding is typically more stable, but unrelated cross-border funding is more sensitive to international shocks than is inter-office funding. In particular, we show that inter-office funding was the more resilient part of cross-border funding during the GFC and in the face of fluctuations in global financial conditions. Third, the shift in funding mix over the past five years is driven primarily by a relative retreat of FBOs, which rely more than DBOs do on funding from cross-border sources. This is consistent with a generalised retreat from global banking business models, as banking groups have started operating more out of their home offices rather than via their foreign affiliates. Fourth, the retreat of individual foreign offices drives an increasing concentration of FBO nationalities in local banking systems, which has historically been connected with more volatility in cross-border funding.

This article builds on two strands of international banking research. The first is the literature on the stability of foreign bank funding and lending (eg Claessens and van Horen (2013), McGuire and von Peter (2016)). We contribute to it by highlighting the importance of inter-office flows for cross-border funding stability and studying sensitivities to global shocks across cross-border funding sources. The second is the literature on banks' international business models (eg Cetorelli and Goldberg (2012), CGFS (2010), McCauley et al (2010)). We bring aspects of these studies up to date by showing how the funding profiles and the relative size of FBOs and DBOs have evolved in recent years.

The rest of the article is organised as follows. The first section documents changes to banks' funding mix and sets out the implications for funding stability. The second section looks under the hood, showing that these trends are driven by the evolution of FBOs and DBOs and their differences in terms of funding models. The third section documents an increase in FBO nationality concentration, which also has implications for funding stability. The final section concludes.

Funding sources: recent shifts and relative stability

The mix of funding sources for local banking systems has been shifting in recent years. Following a post-GFC retrenchment, total liabilities grew by 39%, from $70 trillion in Q1 2015 to $101 trillion³ in Q4 2021 (Graph 2, left-hand panel). However, the share of funds obtained from cross-border sources showed some decline, falling from 24% to 22% over the same period (centre panel). Within the cross-border segment, funding shifted away from inter-office, falling from 31% of cross-border liabilities to 29%. Overall, local banking systems have replaced cross-border funds, especially inter-office funds for banks in advanced economies (AEs), with local funding (right-hand panel). While these shifts seem small in percentage terms, the decline in the cross-border funding share represents a $1.7 trillion reallocation of the global aggregate.

Previous research points to a stability hierarchy among funding sources, particularly in the face of international shocks. Most stable is local funding, especially deposits (Ongena et al (2015)). Within cross-border funding, inter-office funding has been shown to be the most stable (Reinhardt and Riddiough (2015)),⁴ probably as a result of the pursuit of banking group-wide objectives (De Haas and Van Lelyveld (2010)). Cross-border funding from unaffiliated sources is generally the least stable.

Consistent with these stylised facts, we find that inter-office funding was the most stable part of banks' cross-border funding during the GFC. Overall, cross-border funding plunged from its peak in Q1 2008 through Q1 2009, but inter-office cross-border liabilities declined by only 13% in aggregate, while liabilities sourced from unrelated banks and non-banks, declined by 29% and 18% respectively (Graph 3, left-hand panel).^5,6

In line with previous studies, we also find that inter-office funding is less sensitive to global financing conditions, as proxied by the broad dollar index (Graph 3, centre panel).⁷  For a one standard deviation change in the dollar index (a roughly 8% appreciation), cross-border liabilities fall by less than 1.5% for inter-office but by 5.5% for other sources (for an overall decline of around 4%). Net changes (whereby cross-border claims are subtracted from liabilities) suggest that inter-office sources actually compensate for the retrenchment of unrelated entities.

Swings in a local banking system's cross-border funding go hand in hand with similar swings in domestic lending. We obtain this result from an empirical analysis, that considers both gross and net cross-border liabilities. In particular, we find that, when such liabilities decline, so do local claims, by an equal amount for AEs and by 40% of the gross cross-border liability decline for emerging markets (EMEs) (Graph 3, right-hand panel).

Underlying foreign and domestic banking office trends

The shifts in global funding mix indicated above reflect the confluence of three factors: the declining share of foreign banking offices (FBOs) relative to domestic banking offices (DBOs); the differences between the FBO and DBO funding models; and changes in liability structures within these groups, particularly among DBOs. We discuss these factors in turn.

Following the post-GFC retrenchment in international banking, FBOs' liabilities grew from around 2016, but by less than those of DBOs did (Graph 4, left-hand panel). As a result, FBOs' share in local banking system liabilities declined, from 43% in Q1 2015 to 41% in Q4 2021 for those sourced cross-border. In the case of local liabilities, the corresponding decline was from 11% to 9% for local liabilities (centre panel).

FBOs' funding models differ from those of DBOs in several respects. Cross-border positions account for a far higher proportion of FBOs' balance sheets, reflecting these offices' generally greater international focus (Graph 4, right-hand panel).⁸  In addition, FBOs are nodes in global banks' international intermediation networks, and part of their function is to help manage flows of liquidity within global banks. This naturally generates a higher proportion of inter-office positions. FBOs also utilise much more US dollar funding (Box A). These differences across FBOs and DBOs would affect the funding structure of local banking systems, as the share of FBOs declines.

Within the aggregate FBO and DBO groups, the funding mix has been mostly stable over the past five years, with a few notable shifts. For both FBOs and DBOs, the share of funding from local sources remained largely unchanged between Q1 2015 and Q4 2021 (Graph 5, left-hand panel). However, DBOs have reduced the use of inter-office sources, from 26% to 20% of cross-border liabilities over the same period (centre panel).

In sum, the decline in FBOs' liabilities relative to those of DBOs, the differences in these offices' funding structure, and changes in the funding mix within each group have delivered the following shift at the level of local banking systems: from cross-border inter-office to other cross-border and local funding. (Graph 5, right-hand panel). The underlying reshuffling of DBO funding sources is consistent with the generalised retreat of global banking: banking groups have reduced the role of their foreign affiliates to do more business directly from their home offices.

The impact of a shift away from FBOs on funding stability is a priori unclear. Granted, there is evidence that foreign banks tend to be more vulnerable to international shocks than domestic banks are (McGuire and von Peter (2016)). However, such stylised facts reflect foreign banks' greater use of non-deposit funding, especially cross-border funding, whereas FBOs and DBOs that fund themselves similarly are similarly stable.⁹  In addition, FBOs tend to have more stable cross-border funding because their parents support them in times of stress, probably as part of an overall optimisation for the banking group (Eguren-Martin et al (2022), De Haas and Van Lelyveld (2006, 2010), Barba-Navaretti et al (2010)). A decline in FBOs' share of the local banking system's cross-border liabilities reduces the reach of these inter-office networks. In the light of the discussion in the previous section, this could impair local banking systems' funding stability, all else equal.

Impact of FBO closures

Many global banks have recently announced sales or closures of their foreign affiliates (either subsidiaries or branches) across host countries.¹⁰  This is yet another facet of a general retreat from global banking business models. To investigate this phenomenon in a systematic way, we drill down to the level of individual foreign banking offices – branches and subsidiaries – that contribute the data underlying the BIS banking statistics.¹¹

Exits of individual foreign affiliates – through closure or divestiture – have contributed to the relative retrenchment of FBOs studied above. The number of individual foreign offices has declined over the past five years by about 8% in both advanced and emerging markets and by about 17% in OFCs (Graph 6, left-hand panel, bars).

Because FBOs' and DBOs' funding models differ, exits of individual foreign affiliates are likely to affect the overall funding mix of a local banking system. Consistent with this, we see that over the 2016–21 period, declines in the number of foreign bank affiliates¹² were followed on average by a 1.4% decline in the total cross-border funding of local banking systems. This compares with an average 5% increase after the net entry of foreign bank affiliates. The effect of changes in the number of FBOs on the growth of inter-office funding is more striking (Graph 6, centre panel).

More generally, an individual office's exit reduces the presence of one nationality and opens the door to banks from other – typically, already present – nationalities to step in (either by purchasing the exiting affiliate or expanding to take over market share). This can raise the concentration of nationalities in the local banking system (Graph 6, left-hand panel, dots). Increasing concentration reduces the diversification of ultimate funding sources, which makes overall funding more vulnerable to financial stress in individual home countries and thus more volatile (Peek and Rosengren (2000)). Over the past 10 years, we find a material relationship between the nationality concentration in a local banking system and the volatility of its cross-border liabilities, but not with that of its cross-border claims. Concretely, changes in such concentration explain roughly 25% of the variability in a system's total cross border liabilities (right-hand panel, bars).

Conclusion

This article has highlighted changes in the mix of three funding sources for local banking systems – local, cross-border inter-office, and other cross-border. These sources differ in their stability. Local funding is most stable, in part because its major component is deposits. Inter-office funding provides access to cross-border (and foreign currency) funding through a channel that remains relatively resilient during periods of stress. While the focus above was on international shocks, inter-office networks are also an important source of diversification following domestic shocks (Reinhardt and Riddiough (2015), D'Avino (2015)).

Since 2015, a shift toward local funding, which benefits stability, has been counteracted by a decline in the share of inter-office in cross-border funding, which has the opposite effect. These shifts in a local banking system's funding mix stem from a retrenchment of foreign bank offices, which have ceded funding share to domestic bank offices with weaker cross-border inter-office channels.

The importance of cross-border sources for funding stability is most relevant for economies that cannot readily substitute for such sources locally. For example, foreign currency funds typically need to be sourced cross-border. Thus, the ongoing retreat of foreign banking offices makes central bank swap lines more relevant as a tool to ensure access to foreign currency funding after a shock.¹³

While this article has emphasised the resilience of foreign bank offices' cross-border funding, the activities of such offices can also be destabilising. Global banks with extensive networks of foreign affiliates can become increasingly complex, which increases their riskiness (Aldasoro et al (2022)). Further, at times of stress in their home country, FBOs may be drawn upon to help support the parent. This could be at the expense of the local banking system and may syphon liquidity away from the host economy (Cetorelli and Goldberg (2012)).

References

Aldasoro, I and T Ehlers (2018): "The geography of dollar funding of non-US banks", BIS Quarterly Review, December, pp 15–26.

Aldasoro, I, T Ehlers, P McGuire and G von Peter (2020): "Global banks' dollar funding needs and central bank swap lines", BIS Bulletin, no 27.

Aldasoro, I, E Eren and W Huang (2021): "Dollar funding of non-US banks through Covid-19", BIS Quarterly Review, March, pp 31–40.

Aldasoro, I, B Hardy and M Jager (2022): "The Janus face of bank geographic complexity", Journal of Banking and Finance, vol 134(C).

Avdjiev, S, C Koch, H S Shin and W Du (2019): "The dollar, bank leverage and the deviation from covered interest parity", American Economic Review: Insights, vol 1, no 2, pp 193–208.

Barajas, A, A Deghi, C Raddatz, D Seneviratne, P Xie and Y Xu (2020): "Global banks' dollar funding: A source of financial vulnerability", IMF Working Papers, WP/20/113.

Barba-Navaretti, G, G Calzolari, A Franco Pozzolo and M Levi (2010): "Multinational banking in Europe: Financial stability and regulatory implications lessons from the financial crisis", Economic Policy, vol 25, pp 703–53.

Bruno, V and H S Shin (2014): "Assessing macroprudential policies: case of South Korea", Scandinavian Journal of Economics, vol 116, no 1, pp 128–57.

Cetorelli, N and L Goldberg (2012): "Liquidity management of U.S. global banks: Internal capital markets in the great recession", Journal of International Economics, vol 88, no 2, pp 299–311.

Claessens, S and N van Horen (2013): "Impact of foreign banks", Journal of Financial Perspectives, vol 1, pp 1–18.

Claessens, S and N van Horen (2015): "The impact of the global financial crisis on banking globalization", IMF Economic Review, vol 63, pp 868–918.

Committee on the Global Financial System (2010): "Funding patterns and liquidity management of internationally active banks", CGFS Papers, no 39.

Correa, R, L Goldberg and T Rice (2015): "International banking and liquidity risk transmission: evidence from the United States", IMF Economic Review, vol 63, pp 626–43.

D'Avino, C (2015): "Net interoffice accounts of global banks: The role of domestic funding", International Journal of Financial Studies, vol 3, no 3, pp 177–93.

De Haas, R and I Van Lelyveld (2006): "Foreign banks and credit stability in central and eastern Europe: A panel data analysis", Journal of Banking and Finance, vol 30, pp 1927–52.

----- (2010): "Internal capital markets and lending by multinational bank subsidiaries", Journal of Financial Intermediation, vol 19, pp 689–721.

Ehlers, T and P McGuire (2017): "Foreign banks and credit conditions in EMEs", BIS Papers, no 91.

Eguren-Martin, F, M Ossandon Busch and D Reinhardt (2022): "Global banks and synthetic funding: the benefits of foreign relatives", Journal of Money, Credit, and Banking, forthcoming.

Fender, I and P McGuire (2010): "Bank structure, funding risk and the transmission of shocks across countries: concepts and measurement", BIS Quarterly Review, September, pp 63–79.

Financial Stability Board (2022): "US dollar funding and emerging market economy vulnerabilities".

Ivashina, V, D Scharfstein and J Stein (2015): "Dollar funding and the lending behavior of global banks", Quarterly Journal of Economics, vol 130, no 3, pp 1241–81.

McCauley, R, and P McGuire (2014): "Non-US banks' claims on the Federal Reserve", BIS Quarterly Review, March, pp 89–97.

McCauley, R, P McGuire and G von Peter (2010): "The architecture of global banking: from international to multinational?", BIS Quarterly Review, March, pp 25–37.

McGuire, P and G von Peter (2016): "The resilience of banks' international operations", BIS Quarterly Review, March.

Ongena, S, J-L Peydro and N van Horen (2015): "Shocks abroad, pain at home? Bank-firm level evidence of the international transmission of financial shocks", IMF Economic Review, vol 63, pp 698–750.

Peek, J and E Rosengren (2000): "Collateral damage: Effects of the Japanese bank crisis on real activity in the United States", American Economic Review, vol 90, no 1, pp 30–45.

Reinhardt, D and S Riddiough (2015): "The two faces of cross-border banking flows", IMF Economic Review, vol 63, pp 751–91.

¹ We thank Claudio Borio, Stijn Claessens, Patrick McGuire, Hyun Song Shin, Nikola Tarashev and Goetz von Peter for useful comments. Mert Onen, Swapan-Kumar Pradhan and Jhuvesh Sobrun provided excellent research assistance. All errors are our own. The views expressed in this article are those of the authors and not necessarily those of the Bank for International Settlements.

² We also exclude offshore financial centres (OFCs), where positions are largely unconnected with the jurisdiction's funding needs.

³ These figures are based on a constant sample of reporting countries, and so do not match the global aggregates published on the BIS website, which include all available data.

⁴ This need not always be the case. During the Covid-19 pandemic, inter-office positions jumped considerably, along with other positions (Aldasoro et al (2021)).

⁵ Local liabilities were likely the most stable funding source, but low data reporting over this period prevents a direct comparison. Local claims in local currency (obtained from the BIS consolidated banking statistics), typically funded by local liabilities and shown to be relatively stable through the GFC (McGuire and von Peter (2016)), declined by 14% over the same period.

⁶ Following the GFC, foreign banks in the United States saw a reversal in their inter-office positions, switching from being a net creditor to their parents to a net borrower (Bruno and Shin (2014)). This development reflects attractive interest rates on reserves at the Federal Reserve, which were due to quantitative easing in the United States, leading banks to keep more funds in their Federal Reserve accounts (McCauley and McGuire (2014)). When banks located in the United States are excluded, inter-office liabilities still see a much smaller decline than those to unrelated sources.

⁷ An increase in the broad dollar index is indicative of a tightening of global financial conditions and is associated with a contraction in international banking (Avdjiev et al (2019)).

⁸ FBOs vary considerably in their funding composition, ranging from those that obtain nearly all their funding from cross-border sources to those that obtain less than a quarter (see also Fender and McGuire (2010)). This reflects differences among parent banks' international business models (McCauley et al (2010)). Nevertheless, the majority of FBOs have significant cross-border positions in their balance sheets.

⁹ For instance, Ongena et al (2015) and Claessens and van Horen (2013) show that foreign banks funded locally behave in a similar way to domestic banking offices (DBOs) funded locally. Moreover, internationally funded FBOs do not generally respond to funding shocks more negatively than DBOs with a similar reliance on foreign funding (Ongena et al (2015)). The type of credit and how it is funded matters more than the ownership of the bank (Ehlers and McGuire (2017)), but foreign ownership is notable as it typically indicates a different funding mix.

¹⁰ For instance, since 2021 (in addition to planned exits from Russia): Citi announced its intention to exit retail business in at least 15 emerging market economies; Commerzbank said it would leave Hong Kong SAR, Hungary and Luxembourg; HSBC sold assets in France and Greece and announced sales in Saudi Arabia, Turkey and the United States; Standard Chartered announced exits from at least seven countries in Africa and two in the Middle East; and a few non-US G-SIBs have announced, planned or consummated divestitures of large US subsidiaries

¹¹ The BIS records on a yearly basis the number of banks (head offices, subsidiaries and branch networks) that constitute the reporting population of the BIS locational statistics.

¹² A decline in foreign affiliates is flagged when both the number of banking affiliates and the number of bank nationalities decline. This conservative measure accounts for the possibility that the number of affiliates fall because of a consolidation rather than a true exit. The trend is similar if we use only a decline in the number of affiliates. An entry is flagged when the number of affiliates increases.

¹³ Access to insured FX retail deposits can similarly cushion banks and stabilise their lending supply (Ivashina et al (2015)).