Cross-border financial centres

BIS Working Papers  |  No 1035  | 
25 July 2022



When monitoring capital flows and analysing international interconnectedness, it is useful to distinguish financial centres that specialise in cross-border financial activity from other countries. Such centres are a channel for international investment, as opposed to an ultimate source or final destination for funds. Also, their external assets and liabilities are exceptionally large compared with their domestic economy. As a result, they account for a small share of the global economy but a large share of international financial activity. We propose a quantitative, transparent and replicable method for identifying cross-border financial centres.


Our method improves on previous classifications of financial centres, first, by focusing on the entrepôt nature of activity in cross-border financial centres. Second, we adjust for the high degree of dispersion in external positions when setting the threshold to distinguish between cross-border centres and other countries. Finally, we pool many years of data to capture longer-term trends in countries' financial activity.


We find that the group of countries identified as cross-border financial centres varies over time and with different measures of financial intermediation. In 2020, we identify a core set of seven countries as cross-border centres: the British Virgin Islands, Bermuda, the Cayman Islands, Guernsey, Jersey, Luxembourg and the Marshall Islands. Another five countries were included in this set at one point or another over the 1995–2020 period: The Bahamas, Curaçao, Gibraltar, Liberia and Mauritius. A somewhat different set of countries is identified when the focus is narrowed to intermediation through the interbank channel.


Financial centres that cater predominantly to non-residents – which we refer to as cross-border financial centres (XFCs) –are important intermediaries of cross-border financial flows. For analysing capital flows and international interconnectedness, it can be useful to distinguish countries that are home to XFCs from other countries. We improve on previous methodologies for identifying such centres by constructing a measure focussed on the intermediation activity inherent to XFCs and explicitly taking into account the non-normal distribution of this measure across countries when detecting outliers. We also minimise volatility in the set of countries identified as XFCs over time by de-trending the data and pooling years. Our methodology identifies a core set of 12 countries as XFCs over the 1995-2020 period, but the countries vary with time and different measures of activity.

JEL classification: F21, F36, G15.

Keywords: international financial centres, offshore banks, outlier detection.