The outsize role of cross-border financial centres

BIS Quarterly Review  |  June 2022  | 
13 June 2022

Financial centres that cater predominantly to non-residents account for an outsize share of cross-border financial activity. These so-called cross-border financial centres are typically located in small economies, in contrast to global financial centres located in large economies. Economies of scale and scope benefit global centres, but physical distance works against the tendency of financial activity to concentrate. So do regulation and taxation, which have set cross-border financial centres apart and propelled their rise. At the same time, these centres pose challenges to regulatory consistency across countries and complicate the analysis of capital flows. 1

JEL classification: F21, F36, F38, G15.

Financial centres that specialise in cross-border activity have become an entrenched feature of the global financial system. Until the 1970s, international financial intermediation was concentrated in a few major cities that also served as centres for domestic activity, notably London and New York. Since then, financial centres that cater predominantly to non-residents – henceforth, cross-border financial centres – have loomed large as intermediaries of cross-border financial flows. Small economies that host cross-border financial centres saw their share of global external assets and liabilities rise from around 15% in the late 1980s to 30% in the late 2010s – even as their share of world GDP remained constant at less than 3% (Graph 1, left-hand panel).

What explains the growing importance of cross-border financial centres in international intermediation? Economies of scale and scope favour the concentration of activity in financial centres embedded in larger economies. Yet, G20 economies, which account for around 80% of world GDP, saw their share of global external assets and liabilities fall from more than 65% in the late 1980s to around 55% in the late 2010s (Graph 1, right-hand panel). The opposing trends in the shares of cross-border centres and G20 economies demonstrate the importance of factors that work against the concentration of financial activity. One such factor is physical distance between financial intermediaries and their counterparties. Another is differences in regulation and taxation; the former propelled the early growth of cross-border centres while the latter has played a greater role more recently.

Key takeaways

  • Cross-border financial centres, which cater predominantly to non-residents, have become an entrenched feature of the global financial system.
  • Geography, regulation and taxation work against the natural tendency of financial activity to concentrate in a few global financial centres.
  • The importance of foreign direct investment in cross-border centres, mainly in the form of funds that only pass through, is indicative of substantial activity motivated by tax considerations.

This feature proceeds as follows. It first defines different types of financial centre. It then outlines the growth of activity in cross-border centres, before turning to the factors that influence a country's emergence and persistence as such a centre. The concluding section outlines potential challenges to regulatory consistency and transparency arising from the role of cross-border centres in international intermediation.

A financial centre is a location, usually a city or district, where intermediaries involved in the provision of financial services are concentrated. Insofar as it is open to foreign participants, any financial centre can be considered international. But not all international financial centres are alike. In general terms, a common way to distinguish among them is by the size of international financial business relative to total economic activity. We follow this approach in this feature.2

One type of financial centre serves mainly resident counterparties and consequently has a low share of international business. This type is synonymous with national financial centres, where banks are headquartered and stock exchanges are located. Their international activity largely involves channelling domestic funding to foreign borrowers and foreign funding to domestic borrowers. Istanbul and Mumbai are examples.

Another type of financial centre serves mainly non-resident counterparties and consequently has a very high share of international business. They channel funds from one country to another, often via entities with a minimal physical presence, such as booking offices, special purpose entities (SPEs) and shell companies. They are neither an ultimate source nor final destination for investments and are usually embedded in small economies, as in the case of Bermuda and the Cayman Islands (Lane and Milesi-Ferretti (2011)). We refer to these as cross-border financial centres and explain in the Box how they differ from offshore centres.

A third type combines the functions of national and cross-border financial centres. Centres of this type are typically located in economies that issue a reserve currency and include some of the largest financial centres. Their international business is very large in absolute terms but not necessarily relative to total economic activity. These are often referred to as global financial centres. London and New York are the classic examples (Cassis (2006)).

International financial business has many aspects, from origination, underwriting and trading to legal, accounting and other corporate support services. For the purposes of analysing capital flows and international interconnectedness, the most relevant aspect is the extent to which this business results in the accumulation of external positions. These positions comprise a country's outstanding claims on and liabilities to non-residents in the form of portfolio investment, direct investment, other financial investment (consisting mainly of bank loans and deposits), and the market value of derivatives.3

We measure cross-border intermediation as the minimum of external financial assets and liabilities. This captures the extent to which a country acts as a conduit for financing between non-residents, as opposed to a source or destination for investments. The minimum underscores a country's role in cross-border financial intermediation and disregards its role as an external creditor or debtor.

To gauge whether this measure of cross-border intermediation is high or low relative to economic activity, we scale by GDP and identify cross-border centres as outliers in the distribution of this ratio across countries. GDP is readily available for almost all economies, including many dependent territories, and is highly correlated with less readily available measures of financial activity, such as total financial assets. The outliers are detected annually, using a method that is less sensitive than alternatives to the shape of the distribution.4

Cross-border financial intermediation in 2020 is shown for a sample of over 200 economies in Graph 2, both in US dollars (y axis) and as a ratio to GDP (x axis). Cross-border financial centres appear on the right side of the graph (red dots). The British Virgin Islands (VG) and Cayman Islands (KY) stand out, with cross-border intermediation a thousand times greater than their GDP, followed by Bermuda (BM) and Luxembourg (LU). A few economies, such as the Netherlands (NL), Hong Kong SAR (HK) and Singapore (SG), are examples of cases in the boundary zone between cross-border and global centres. Several other economies were classified as cross-border centres in the past but had fallen out of the group by 2020 (orange dots).

Global financial centres appear at the top of Graph 2. They feature the highest US dollar values of cross-border financial intermediation; in addition, their ratios to GDP are higher than the cross-country median but lower than those of cross-border centres. The largest global centres are London (GB) and New York (US). Shanghai (CN) is not yet in the league of global centres because the large US dollar value of its cross-border intermediation is small relative to the size of the Chinese economy.

The group of countries classified as cross-border financial centres is reasonably stable over time, but not static. Over the 1995–2020 period the number of cross-border centres in any given year varied between 12 and 18, with a total of 23 different economies belonging to this group at some point in time (Table 1). Many had already emerged as cross-border centres in the 1970s and have remained prominent since then, including the Cayman Islands (KY), British Virgin Islands (VG) and Luxembourg (LU). A few have faded in importance, notably Bahrain (BH), Panama (PA) and Vanuatu (VU) in the 1990s. Others have grown in recent years, such as the Netherlands (NL) and Mauritius (MU).

Since the 1970s, greater financial integration has caused external assets and liabilities to grow faster than GDP worldwide and especially so in cross-border centres. Across all economies, the ratio of cross-border intermediation to GDP increased sevenfold between the early 1980s and 2020, from 0.3 to 2.1 (Graph 3, left-hand panel). Over the same period, this ratio increased about fifteenfold for cross-border centres, from 1.5 to 23. Accordingly, the share of these centres in global financial activity rose fast – unlike their share in world GDP (Graph 1, left-hand panel).

The rapid growth of cross-border centres, as well as the financial integration of emerging market economies, has contributed to a more diverse geographic distribution of external positions (Lane and Milesi-Ferretti (2018), Broner et al (2020)). The concentration of external liabilities has been roughly stable, partly due to the growth of US external liabilities as a share of the global total. At the same time, the concentration of external assets across all countries fell by about half between the 1970s and 2010s (Graph 3, right-hand panel).

What explains the growing prominence of cross-border financial centres in the global financial system? Network effects as well as economies of scale and scope work in favour of concentrating financial activity in a few large economies, particularly global financial centres. In these centres, the presence of a diverse set of market participants lowers the costs of matching borrowers and savers with different preferences, thus attracting still more intermediaries and customers. Similarly, a pool of financial experts attracts more experts, thus promoting specialisation and agglomeration effects.

However, the coexistence of many cross-border financial centres alongside a few global ones demonstrates that forces against centralisation are at play too. Information frictions increase with geographical and cultural distance, which supports activity in financial centres closer to customers. Advances in transportation and communication have reduced these frictions but have not eliminated them. High rents and congestion in cities put further limits on concentration, especially of activities that do not benefit from network effects, such as business registrations. Moreover, national borders matter: laws, regulations and taxes are country-specific, and can be tailored to influence the geographic distribution of financial activity.

Institutional prerequisites

A prerequisite for becoming an international financial centre is a stable, transparent legal environment that instils confidence in the predictability of contractual terms and the effective resolution of disputes. Dubai's decision to establish its financial centre as a separate administrative area, with its own courts as well as commercial and civil laws based on English common law, illustrates the importance of the legal framework. Lebanon is another example, in that Beirut's standing as a financial centre was undercut starting in the 1970s by civil wars and the weakening of domestic institutions.

Another prerequisite is the unrestricted use of foreign currencies. This is particularly important for financial centres in small economies, which lack any sizeable activity in their own domestic currency. To achieve scale, they transact in the US dollar and other reserve currencies. Some adopt a reserve currency outright as their domestic currency. Financial counterparties readily transact in reserve currencies, and often do so outside the reserve-issuing country. For example, non-banks outside the United States borrow US dollars mainly offshore, from banks outside the United States (Graph 4, left-hand panel). A substantial share of offshore lending in US dollars, yen, sterling and Swiss francs – lending by banks outside the currency area – takes place from cross-border centres (red bars).

Cross-border financial centres are also typically open to foreign financial institutions and their expertise. In many, foreign banks account for the lion's share of the banking system's cross-border positions (Graph 4, right-hand panel). Foreign-owned institutions bring skills, technology, networks and reputation that locally owned institutions in small economies have difficulty matching. They also bring access to funding and liquidity, which help cross-border financial centres to operate at scale.

Geographic aspects

Geography naturally entails some dispersion of activity across financial centres. The physical reality of time zones requires several financial centres that allow for trading, clearing and settlement around the clock. New York, London and Tokyo span 24 hours with their market opening hours. But even small time differences can be inconvenient for financial institutions and their customers. Furthermore, information frictions increase with physical and cultural distance (Mian (2006)). This creates opportunities for multiple financial centres even within a given time zone, as in the case of Hong Kong SAR and Singapore.

The advantage of being closer to customers helps explain why cross-border financial centres tend to have more of a regional focus than global centres do. In particular, many centres cater to a large neighbouring economy. To gauge the geographic reach of a financial centre, we use the BIS locational banking statistics to examine bilateral links between banks and their counterparties abroad (Graph 5). Most cross-border centres book a smaller percentage of cross-border bank positions outside their region (33%) than the average country (38%), let alone global financial centres like the United States (74%) and Japan (83%). Several cross-border centres transact almost exclusively within their region, including Cyprus (CY) and Gibraltar (GI). The British Virgin Islands (VG) is a notable exception, owing to its extensive links with Asia in particular.

Some cross-border financial centres play a role intermediating between regions. The Cayman Islands (KY) – located in the Americas – source 65% of their bank-related liabilities from other regions, notably from Japan and other countries in Asia; yet the bulk of their cross-border claims are on borrowers in the Americas, pushing the interregional share below 40% on the asset side (Graph 5). Similarly, about half of Malta's (MT) liabilities are owed to creditors outside Europe, but 80% of its assets remain invested within the region.

Tax and regulatory considerations

Cross-border centres that lack the economies of scale and scope enjoyed by financial centres in large economies often compete for financial business by offering lower taxes or lighter regulations. Centres that compete on these grounds face a highly elastic demand for their services; small differences can have a significant impact on their market share (Sinha and Srivastava (2013)). As a result, much of the business booked in such centres involves entities that have low setup costs and a minimal physical presence, so that the business can be relocated with ease (Dixon (2001)).

The first cross-border centres had their origins in regulatory arbitrage. Differences in the regulatory treatment of banks' domestic and foreign funding create the equivalent of a tax wedge, which enables affiliates abroad to offer higher interest rates to depositors and lower ones to borrowers. The most relevant regulations include ceilings on deposit rates, reserve requirements and deposit insurance premiums (McCauley et al (2021)).

The liberalisation of capital accounts since the 1980s and the strengthening of financial regulation and supervision worldwide have reduced opportunities for regulatory arbitrage, especially in the banking sector. However, the regulation of non-bank financial institutions (NBFIs) tends to be more fragmented than that of banks and thus continues to offer opportunities for cross-border arbitrage. For instance, capital requirements for captive insurance companies are lower and more flexible in some cross-border centres than they are in larger economies (KPMG (2021)). In the area of digital innovation, a number of economies have banned or restricted cryptocurrency businesses, whereas many cross-border centres have enabled their expansion. NBFIs now account for the largest share of intermediation via cross-border centres, which highlights their potential as a channel for regulatory arbitrage (see below).

Corporate taxes are another area where cross-border centres often differentiate themselves. Several cross-border centres have tax rates of zero on some types of corporate income, including the British Virgin Islands (VG), the Cayman Islands (KY), the Isle of Man (IM), Guernsey (GG) and Jersey (JE). This has the effect of enticing foreign businesses, as these same countries have an unusually large number of company registrations per capita (Graph 6, left-hand panel). While statutory tax rates are not uniformly low in cross-border centres, other features, such as exemptions and double taxation treaties, further enhance the attractiveness of a tax regime.

The size of foreign direct investment (FDI) in many cross-border centres is indicative of how much activity is motivated by tax considerations. In cross-border centres, most FDI takes the form of pass-through funds, whereby companies route investments through one or more financial subsidiaries on their way to a destination (Weyzig (2013)). Even if some pass-through activity reflects the complexity of multinationals' corporate structures, a larger proportion is explained by tax optimisation strategies (Borga and Caliandro (2020)).

In many cross-border centres, FDI assets and liabilities are unusually large and of similar magnitude, confirming the pass-through nature of FDI. Globally, FDI accounts for less than 30% of external liabilities (Graph 7, right-hand panel). By contrast, it accounted for about 40% of cross-border centres' aggregate liabilities in 2020. The proportion was even higher in Cyprus (CY), Malta (MT) and Mauritius (MU) (red bars). These centres have small domestic economies, which suggests that little of this FDI was invested in the production of goods or services. The steady rise in outstanding FDI, more than doubling as a share of the external liabilities of cross-border centres since 1995 and up over tenfold as a share of their GDP, points to the growing importance of tax optimisation strategies (Graph 7, left-hand panel).

Authorities in many cross-border centres now strive to implement international tax and regulatory standards. Many centres have joined international efforts to tackle tax evasion: for example, by participating in the automatic exchange of information under the auspices of the Global Forum on Transparency and Exchange of Information for Tax Purposes. In 2021, 137 countries – including many cross-border centres – agreed to a plan to impose a minimum corporate tax rate of 15% (OECD (2021)). Similarly, in cross-border centres the regulations in place against money laundering and the financing of terrorism mostly comply with the technical requirements set out by the Financial Action Task Force (Graph 6, right-hand panel, y-axis). That said, the effectiveness of these regulations varies across centres (x-axis).

To be sure, not all cross-border financial centres are in full compliance with international standards. In some places financial secrecy still attracts clients wishing to engage in activities that are illegal elsewhere. A series of data leaks in recent years, documented in a database curated by the International Consortium of Investigative Journalists, sheds light on the complex web of corporate structures that can be used to hide wealth or illicit activities (Díaz-Struck et al (2022)).

Business specialisation

Competitive pressures lead many cross-border financial centres to implement similar regulatory and tax regimes. Thus, to achieve a lasting advantage, cross-border financial centres also tend to specialise in certain activities. They build an ecosystem of financial, advisory and legal services that cater to specific demands and differentiate them from other financial centres. Indeed, centres sometimes tailor their regulatory and tax framework to appeal to certain types of activity (Sinha and Srivastava (2013)).

When cross-border financial centres first emerged in the 1970s, they typically specialised in banking. A substantial share of this banking activity consisted of intragroup transactions among affiliates of the same (foreign) banking group. Even business involving unrelated parties, such as lending to non-bank borrowers, was typically originated by banks based elsewhere (McCauley and Seth (1992)). In effect, assets and liabilities were recorded in cross-border centres for financial reporting purposes, but mind and management for making meaningful lending and borrowing decisions were located elsewhere (Dixon (2001)).

Since the mid-1990s, activity in cross-border financial centres has shifted away from banking toward non-bank financial intermediation. This was in keeping with the same shift in the global financial system. Today business with non-bank entities, such as fund managers and other NBFIs, accounts for the largest share of most centres' cross-border activity. Their interbank liabilities (including intragroup positions) fell from close to 40% of total liabilities in 1995 to about 5% in 2020 (Graph 7, left-hand panel, green bars). Over the same period, the liabilities of non-banks rose from 40% to over 80% (beige bars). FDI accounts for the largest share of these liabilities, mostly incurred by SPEs affiliated with multinational companies.

The different types of activity in which cross-border centres specialise is reflected in the instrument composition of their external assets and liabilities. The Cayman Islands (KY), Ireland (IE) and Luxembourg (LU) host a large number of investment funds, consistent with the high share of portfolio equity in their liabilities (Graph 7, right-hand panel, yellow bars). The British Virgin Islands (VG), the Cayman Islands (KY) and the Netherlands (NL) host financial subsidiaries of multinational firms, through which bonds are issued to international investors. Their portfolio debt liabilities are correspondingly high (beige bars). In the Bahamas (BS), Hong Kong SAR (HK) and Singapore (SG), banking is still relatively important ("other investment", in blue). In Panama (PA), much of the loan liabilities are owed by non-financial companies, particularly shipping firms. Panama is home to the largest ship registry in the world.

Since their emergence in the 1970s, cross-border financial centres have secured an outsize role in the global financial system. They have prospered from the shift in international intermediation away from banks towards NBFIs. Their persistence and adaptability point to the importance of factors, such as time zones, regulation and tax, that work against the natural tendency of financial activity to concentrate in a few global financial centres.

Cross-border centres offer benefits to the global financial system but also pose challenges to its smooth functioning. Their rise has put competitive pressure on global and national financial centres to reduce costs, innovate and improve the quality of their services (Rose and Spiegel (2007), Hines (2010)). At the same time, to the extent that cross-border financial centres facilitate regulatory arbitrage and obscure risks, they can undermine international efforts to strengthen the resilience of international financial intermediation.

While consolidated supervision acts as a safeguard against regulatory arbitrage in banking, the non-bank financial sector poses a greater challenge to regulatory consistency. An important step to strengthen the resilience of NBFIs is taking a less fragmented and more consolidated supervisory perspective on their activities (Carstens (2021)). The Financial Stability Board is leading international efforts to strengthen the resilience of non-bank financial intermediation (FSB (2021)).

Furthermore, cross-border centres complicate risk analysis by obscuring the ultimate source and destination of investments, and even the type of investment (Coppola et al (2021)). For example, estimates of US portfolio investment in emerging markets are about a third higher based on the nationality of borrowers as opposed to their residence, with investment via cross-border centres accounting for much of the difference (Bertaut et al (2021)). Similarly, in the case of countries where companies issue bonds via financial subsidiaries in cross-border centres, exposure to global financial market conditions may be underestimated because the bond proceeds show up as FDI (in the form of an intercompany loan) rather than portfolio liabilities. The opaqueness of activity in cross-border financial centres highlights the usefulness of complementing the residence perspective taken in the standard international statistical framework with a consolidated view based on the controlling parent (Avdjiev et al (2018)).

Avdjiev, S, M Everett, P Lane and H S Shin (2018): "Tracking the international footprints of global firms", BIS Quarterly Review, March, pp 47–66.

Bank for International Settlements (BIS) (1995): Guide to the BIS statistics on international banking, April.

Bertaut, C, B Bressler and S Curcuru (2021): "Globalisation and the reach of multinationals: implications for portfolio exposures, capital flows and home bias", Journal of Accounting and Finance, vol 21, no 5, pp 92–104.

Broner, F, T Didier, S Schmukler and G von Peter (2020): "Bilateral international investments: the Big Sur?", BIS Working Papers, no 911, December.

Borga, M and C Caliandro (2020): "Eliminating the pass-through: towards FDI statistics that better capture the financial and economic linkages between countries", in N Ahmad, B Moulton, J Richardson and P Van de Ven (eds), Challenges of Globalization in the Measurement of National Accounts, National Bureau of Economic Research.

Carstens, A (2021): "Non-bank financial sector: systemic regulation needed", BIS Quarterly Review, December, pp 1–6.

Cassis, Y (2006): Capitals of capital: a history of international financial centres, 1780–2005, Cambridge University Press.

Coppola, A, M Maggiori, B Neiman and J Schreger (2021): "Redrawing the map of global capital flows: the role of cross-border financing and tax havens", Quarterly Journal of Economics, vol 136, no 3, pp 1499–556.

Díaz-Struck, E, D Reuter, A Armendariz, J Cosic, K Kehoe, M Gutiérrez, M Williams and N Sadek (2022): "ICIJ publishes final batch of Pandora Papers data on more than 9,000 offshore companies, trusts and foundations", available online at

Dixon, L (2001): "Financial flows via offshore financial centres as part of the international financial system", Bank of England Financial Stability Review, June 2001.

Financial Stability Board (FSB) (2021): Enhancing the Resilience of Non-Bank Financial Intermediation: Progress report, November.

Garcia-Bernardo, J, J Fichtner, F Takes and E Heemskerk (2017): "Uncovering offshore financial centers: conduits and sinks in the global corporate ownership network", Scientific Reports, vol 7, no 6246.

Hines, J (2010): "Treasure islands", Journal of Economic Perspectives, vol 24, no 4, pp 103–26.

IMF (2000): "Offshore financial centers: IMF background paper", 23 June.

KPMG (2021): Captive insurance in the Cayman Islands, November.

Lane, P and G Milesi-Ferretti (2011): "Cross-border investment in small international financial centres", International Finance, vol 14, no 2, pp 301–30.

--- (2018): "The external wealth of nations revisited: International financial integration in the aftermath of the global financial crisis," IMF Economic Review, vol 66, pp 189–222. An updated version of the External Wealth of Nations database is available on the Brookings Institution's website.

McCauley, R, P McGuire and P Wooldridge (2021): "Seven decades of international banking", BIS Quarterly Review, September, pp 61–75.

McCauley, R and R Seth (1992): "Foreign bank credit to US corporations: the implication of offshore loans", FRBNY Quarterly Review, vol 17, no 1, pp 52–65.

Mian, A (2006): "Distance constraints: the limits of foreign lending in poor economies", Journal of Finance, vol 61, no 3, pp 1465–505.

Organisation for Economic Cooperation and Development (OECD) (2021): "Statement on a two-pillar solution to address the tax challenges arising from the digitalisation of the economy", 8 October.

Pogliani, P and P Wooldridge (2022): "Cross-border financial centres", BIS Working Papers, forthcoming.

Rose, A and M Spiegel (2007): "Offshore financial centres: parasites or symbionts?", The Economic Journal, vol 117, October, pp 1310–35.

Sinha, N and A Srivastava, A (2013): "Offshore financial centers and tax havens – an overview", in The Chamber of Tax Consultants, International Taxation: A Compendium, third edition, pp 988–1011.

Weyzig, F (2013): "Tax treaty shopping: structural determinants of foreign direct investment routed through the Netherlands", International Tax and Public Finance, vol 20, pp 910–37.

Zoromé, A (2007): "Concept of offshore financial centers: in search of an operational definition", IMF Working Papers, no WP/07/87, April.

Z/Yen Partners (2022): The Global Financial Centres Index, March.

1  The authors thank Stefan Avdjiev, Claudio Borio, Stijn Claessens, Bryan Hardy, Robert McCauley, Patrick McGuire, Nikola Tarashev and Christian Upper for helpful comments and discussion, and Swapan Kumar Pradhan for assistance with data. The views expressed in this article are those of the authors and do not necessarily reflect those of the Bank for International Settlements.

2  Zoromé (2007) and Lane and Milesi-Ferretti (2011, 2018) also follow this approach, although the specific measures they use differ (net exports of financial services to GDP and total external positions to GDP, respectively). Garcia-Bernardo et al (2017) use network analysis to identify financial centres. Z/Yen (2022) assess business characteristics to rank financial centres. For further discussion, see the box A and Pogliani and Wooldridge (2022).

3  For many countries, comprehensive data on external positions have a short history. Lane and Milesi-Ferretti (2018) constructed long series by combining national data on international investment positions with other sources, including cumulative gross capital flows from the balance of payments. Their coverage increases from 103 countries in 1970 to 211 in 2020, the final year in the December 2021 version of their External Wealth of Nations (EWN) database. Following Pogliani and Wooldridge (2022), we extend the EWN by incorporating data from the BIS locational banking statistics (LBS); where external assets and liabilities derived from the LBS exceed estimates in the EWN, we replace the estimates.

4  Outliers are detected using a boxplot adjusted for the skewness of the distribution. For further discussion of measures and methods for identifying cross-border financial centres, see Pogliani and Wooldridge (2022).