De-internationalizing global banking?

McCauley, Robert

INTRODUCTION

The decentralized multinational model of banking proved more robust than the international banking model to the disruption to wholesale funding markets during the global financial crisis (Committee on the Global Financial System, 2010; McCauley et al, 2012). 'What we observed during this peak of the crisis was ... a run on cross-border banking operations' (Gudmundsson, 2011). The international model employs funds raised in one country to lend in another, often relying on wholesale funds or foreign exchange swap markets. In its pure form, the multinational model matches assets and liabilities in each country, with liabilities consisting largely of deposits.

Quite apart from the multinational model's surviving the global financial crisis better, regulatory and supervisory developments could give a further boost to this multinational model. This article reviews how the application in some jurisdictions of the newly agreed international bank liquidity standards and policies favoring subsidiaries over branches ('subsidiarization') could de-internationalize global banking in the sense of inducing banks to match assets and liabilities jurisdiction by jurisdiction.

Such a development could pose risks, both to macroeconomic growth and to financial stability. First, if banks cease to perform any role in cross-border capital flows, then all the potential growth benefits of gross and net capital flows must be achieved only through portfolio and direct investment flows, which have their own limitations. Second, authorities may ex ante overstate the efficacy of purely national regulations and, ex post, find themselves relaxing the rules to allow banks to provide support to affiliates in the event of crisis.

An alternative approach is international cooperation to ensure the strengthening of banks on a consolidated basis. If the stronger capital and liquidity standards as agreed in Basel III at the consolidated level were thoroughly implemented, they could temper the very understandable regulatory and supervisory impulse to require stand-alone subsidiaries. If feasible, this approach would be first-best.

This article is structured as follows. The next section draws on and updates McCauley et al. (2012) to sketch long-standing trends in global banking and the effect of the global financial crisis. The following section provides a selective analysis of how policy initiatives may tend to fragment (1) international banking. The final section concludes.

TRENDS IN GLOBAL BANKING AND THE GLOBAL FINANCIAL CRISIS

Long-standing trends: deepening and local currency banking

Probably for decades, the growth in banks' foreign claims outpaced that in economic activity (Figure 1, left-hand panel). Foreign claims are defined as the sum of cross-border claims (which would be picked up in balance of payments statistics) and local claims in all currencies booked at offices outside the home country (which would not be captured in balance of payments statistics). In the process, banking has become more global just like other industries.

[FIGURE 1 OMITTED]

Banking stands out, however, in its legal form and reliance on cross-border funding. In other industries, as a firm expands from its home market, it sets up subsidiaries abroad that borrow locally to finance assets: this is the multinational model (Aliber, 1993). Accordingly, the multinational bank operates sizeable foreign branches and subsidiaries in multiple jurisdictions (Jones, 1991) and, at the limit, funds those positions locally in the host countries. Banks that exemplify this model, like the major Spanish banks and two of the major British banks with colonial roots, operate independently capitalized subsidiaries that are fully subject to local regulation and benefit from host-country deposit insurance. In some cases, head office even hedges the currency exposure of the equity in the local subsidiary to avoid exchange-rate risk. In contrast, the international bank operates out of the home country or in a (major) financial centre and conducts mostly cross-border business. Banks that exemplify the international model, like Japanese and some German banks, mobilize their excess domestic deposits to fund their foreign currency lending but typically do so on a hedged basis, so that they do not take on first-order currency risk (McGuire and von Peter, 2012).

The special risks that banks run in lending abroad have influenced the choice between these two models. The term 'country risk' covers the potential legal, political or economic sources of loss that are common to a jurisdiction. In particular, 'transfer risk' arises when an otherwise sound borrower cannot buy the foreign currency needed for debt service. When big banks stepped up their lending to emerging market governments and firms in the 1970s, supervisors started to require systematic reporting of banks' country exposures. The 1980s Latin American debt crisis inflicted heavy losses on cross-border loans denominated in dollars. In response, to avoid transfer (if not country) risk, banks began to shift toward the multinational model by establishing or acquiring a local bank in order to borrow and lend locally.

As a result, the share of local currency claims in foreign claims on emerging market economies rose from 7% in 1983 to 25%-30% in the 1990s (Figure 1, right-hand panel). (2) After the Asian financial crisis of 1997-1998, the Committee on the Global Financial System (which meets at the BIS) widened the group of reporting jurisdictions and began to collect data on worldwide exposures. The new data showed that the local currency share of claims on the advanced economies and hence globally was initially even higher than the share of such claims on emerging market economies. Local currency claims on both emerging markets and on all countries gained share into the 2000s.

The shift to local banking slowed in the mid-2000s. In emerging markets, the share of local currency claim in foreign currency claims levelled off at less than half, as cross-border bank flows resumed in response to higher yields and US dollar depreciation (Avdjiev et al., 2012). Elsewhere, the introduction of the euro, spurring an area-wide interbank market, and European banks' heavy investment in US asset-backed securities had a similar effect. European banks raised dollars inter alia from US non-banks like money market funds, only to purchase risky, high-leverage US mortgage-backed securities, in what Shin (2012) calls the 'banking glut'. For the first time in the history of the euro-dollar market, claims on US non-banks in the mid-2000s made up the bulk of all claims on non-banks (He and McCauley, 2012, Graph 2, right-hand panel).

Effect of the global financial crisis

The global financial crisis has had a mixed impact on these trends of deepening on the one hand and local currency lending on the other. In particular, the trend toward deepening was at least interrupted. Foreign claims have fallen back as a share of global GDP and there are few signs that the previous trend will return. This reversal of international financial deepening is not confined to global banking but also shows up in overall international asset and liability positions (Ma and McCauley, 2014, p. 13).

By contrast, the global trend to local currency claims has re-asserted itself. In particular, banks' local claims as a share of their foreign claims now exceed 40%. (3) European banks have retrenched both from their ill-advised extension of credit to US households and from cross-border claims on sovereigns, banks and firms in Europe. The local currency share of emerging market claims has not risen since the crisis as strong credit growth in the largest emerging markets has been partly financed with cross-border flows (He and McCauley, 2013). In aggregate, however, the multinational model, characterized by local currency loans funded locally with local currency liabilities, emerged stronger from the financial crisis and is again growing faster than cross-border claims.

The data reviewed mostly reflect the adverse impact of a global bank funding crisis on banks with heavy reliance on cross-border funding and the subsequent response of managements to this experience. They also reflect policy to the extent that state aid came with conditions to downsize or exit business lines.

Looking forward, however, there is the possibility that regulatory responses to the crisis further favor the multinational model. Before we turn to consider selected policy responses, let us consider whether the multinational model itself may have received a new interpretation as a result of the crisis.

A mutation in the multinational banking model?

While the multinational model weathered the storm that hit bank wholesale funding markets during the global financial crisis better than the international banking model, the multinational banking model itself has mutated under the force of crisis at home. Santander, the bank with the largest market capitalization in Europe, has turned its foreign subsidiaries into sources of strength by extracting equity capital from them. It has done so by 'carving-out' partial stakes in its subsidiaries, first in Brazil, then in Mexico, and offering them to global investors. It raised US$8 billion by selling a stake in its Brazilian subsidiary in 2009 and followed up with a $4 billion offering of 25% of Mexican subsidiary in 2012.

For investors, such offerings promise a combination of the business model and management control of the parent with the presumed faster growth of the emerging market 'pure play'. Santander, for its part, was able to anticipate and to respond to the European Banking Agency's 2011 call for filling its 15.3 billion [euro] capital shortfall without directly diluting its own shareholders with a secondary stock offering at a price below book value.

Views differ on how this development should be interpreted. Santander management argues that it looks to issue equity in every local market in which it operates, including not only Poland but also the United Kingdom. This extension of the multinational model from local debt funding to local equity funding can allow local management teams to be given equity stakes in local operations, possibly motivating them better than equity in the global enterprise. Others view this development as a convenient but ultimately temporary arrangement that affords the minority shareholders no effective voice in the enterprise. On this view, the global enterprise will have ample opportunity to redistribute profits away from the subsidiaries. Once the new shareholders realize their weak position, they will be willing to sell the equity back to the global enterprise when its overall finances permit it to do so. Time will tell which of these views is correct.

FRAGMENTATION OF INTERNATIONAL BANKING: POLICY INITIATIVES

National application of liquidity regulation

While uneven implementation is a risk for all international regulatory standards, the application of the Basel Committee's new liquidity standards carries more than the usual potential for fragmentation. The standards and monitoring are to apply to internationally active banks on a consolidated basis, but the scope of application allows for the local application on a legal entity basis (BCBS, 2009, p. 133; see Cecchetti et al, 2011). As reported in CGFS (2010), a number of regulators have announced their intention to apply liquidity requirements to the subsidiaries and branches of foreign banks in their jurisdictions.

For example, the UK Financial Services Authority (the FSA, now known as the Prudential Regulation Authority or PRA) takes such an approach in its new liquidity regime, which includes the principle of self-sufficiency and adequacy of liquid resources for UK entities. Subsidiaries and branches of foreign banks must monitor and manage their liquidity on a standalone basis and thereby become self-sufficient in liquidity (FSA, 2009).

This form of local liquidity regulation also involves restrictions on (cross-border) intragroup funding. To the FSA (2009), the term self-sufficiency means that subsidiaries and branches of foreign banks will not be permitted to rely on other parts of their group to satisfy the new liquidity requirement. However, the FSA anticipates that many firms will apply for, and receive, modifications of the self-sufficiency requirement, provided there are cross-border supervision arrangements in place that are non-discriminatory to the interest of UK creditors.

Such regulation might well accelerate the trend to the multinational model. While the structure of decentralized multinational banking groups appears well aligned with the emerging pattern of liquidity regulation, banks following more centralized or international models may have to adapt in the following dimensions (CGFS, 2010):

* Local regulation could make each legal entity hold liquid assets in the host country, partly in local currency, to meet local liquidity requirements. This could fragment liquidity holdings across jurisdictions. It may also require an overall increase in group-wide liquidity holdings, if the holdings of foreign subsidiaries do not count toward the fulfilment of the consolidated liquidity requirement in the home country.

* Local liquidity requirements may force banks to set up local treasuries to manage liquid assets. This change could induce greater decentralization of banking operations and increase operational costs.

* In sum, the regulatory segmentation of a banking group along national lines could put in question the sustainability of the intragroup activity characteristic of the centralized model. Proposed intragroup limits as well as currency-matched local liquidity rules could make unviable certain specialized operations, such as offices that collect funds for remittance to the central treasury, or offices in other jurisdictions that heavily rely on such funds.

The macroeconomic consequences of de-internationalizing global banking along these lines are not obvious. True, the propagation of problems in wholesale funding markets as seen in 2008 might be limited. But in other circumstances, flows of funds within global banks can help offset ('accommodate') other movements of funds like portfolio flows. How does international finance work without bank flows that are responsive to narrow return differentials?

Subsidiarization

A related question is whether all jurisdictions remain open to new branches of foreign banks (Cerutti et al, 2007). Experience during the crisis makes it easy to understand why authorities could have concerns about branches.

An Icelandic bank used branches to enter the UK and the Netherlands under the European Union's 'single passport' policy. Since Iceland was a member of the European Economic Area, EU directives allowed Icelandic banks to operate throughout this area, including as branches. Armed with high ratings from the major credit rating agencies and able to access plentiful funding, the Icelandic banks expanded abroad (Guomundsson, 2011). When the online savings accounts (known as Icesave) of the Icelandic bank, Landsbanki, suffered a run, the UK authorities paid out 2.35 [pounds sterling] billion to depositors. (4)

Reportedly, UK policy has been discouraging entry through branches. It has been reported that, from 2007 through 2012, the FSA gave permission to just four banks to open branches, while permitting 14 new subsidiaries (Masters, 2012). Moreover, in September 2012, the FSA announced plans to force non-EU banks to use subsidiaries if they are from countries such as the US and Australia that prefer home depositors in the case of a bankruptcy. Indeed, FSA head Andrew Bailey told the Parliamentary Banking Standards Commission that he would seek an exception to EU rules that allow branching. 'I do not think that a bank should be able to branch in and the host authority have to accept it come what may' (Masters, 2012). (5)

The issue of the London entry of China's rapidly internationalizing state-owned banks surfaced in public. They wrote to the UK Treasury complaining about the FSA's refusal to let them open branches and said they were routing business to Luxembourg instead (Shafer, 2012). A city think tank published a review of Chinese banks in the UK (He, 2013, p. 9) that warned that current policy could 'limit the ability of foreign banks in the UK to finance large infrastructure projects'.

In the event, the UK authorities have opted not to enforce thorough subsidiarization. 'A general policy, not a China policy' was announced by Andrew Bailey, CEO of the Prudential Regulatory Authority and Deputy Governor of the Bank of England, in October 2013. Citing the economic benefits of free trade and capital flows, including the City of London's longstanding provision of trade finance, Bailey said:

   International wholesale banking is therefore an important part of
   maintaining and developing the world economy, just as it was in the
   nineteenth century ... we should not design the world as if
   fragmentation and balkanisation are inevitably always likely to be
   with us. That is a counsel of despair, and it contradicts ... the
   benefits of free trade ...

   ... we do not wish to see non-EEA [European Economic Area] branches
   undertaking critical retail banking functions (like taking
   deposits) ... we think that new non-EEA branches should stick to
   straightforward wholesale banking, of the type that supports world
   trade and capital flows. Resolution will be a key deciding factor
   in the PRA's judgments, and it is thus where we will place emphasis
   when forming a view on our risk appetite towards branches operating
   in the UK. We will therefore expect assurance from the home state
   regulator over the recovery and resolution plans and assess if the
   plans adequately cover the UK branch's activities.

Box 1: ING direct USA: A cautionary tale

The US subsidiary that caused ING such losses that the Dutch state
had to rescue the banking group offers a cautionary tale of the
hazards of banking fragmentation (Raise, 2009). ING's business
model featured trendy orange internet cafes that made depositing
'as easy as drinking a cup of coffee'. Within Europe insured
deposits could be raised in Germany and funnelled back to Amsterdam
for further use by the banking group. In the United States,
however, ING opted for a thrift charter and so submitted itself to
the requirement that its assets be housing-related to some extent.
At first, the subsidiary invested in mortgage-backed securities
issued by government sponsored agencies. At some point, perhaps to
diversify credit risk and certainly to raise yields, the subsidiary
began to invest in private-label mortgage-backed securities,
including those backed by 'low-documentation' Alt-A mortgages (also
known as 'liars' loans').

When the value of these securities declined, the Dutch state
injected 10 billion [euro]. The Dutch government paid 19.8 billion
[euro], well above the fair value of 13.5 billion [euro], for 80% of
26 billion [euro] in Alt-A securities; the difference was essentially
an additional state capital injection. The European Union competition
commissioner responded to the state aid by requiring ING to sell
its insurance and asset management businesses. Eventually, ING sold
its US subsidiary for $9 billion to Capital One, a bank
concentrated on extending loans to credit card customers, a
business model for which the $80 billion in deposits made a perfect
match.

It must be stressed that ING chose a thrift charter for its US
subsidiary in full knowledge of the asset restrictions. And if the
heavy investment in private label mortgage securities was partly
motivated by a desire to diversify away from the credit risk of the
US agencies, there is some likeness to the UBS story. UBS risk
managers had flagged the credit concentration of the UBS liquidity
portfolio in Japanese government securities, and the funds were
subsequently reallocated to a portfolio of highly rated
private-label mortgage-backed securities. These proved not only
illiquid but also loss-making in 2007 (UBS, 2008). All that said,
it does seem likely that ING might not have needed a state rescue
if deposits raised in the United States could have been as readily
deployed by headquarters as deposits raised in Germany.

In sum, branches are welcome if they stick to wholesale funding and if they are covered in extremis.

It is hard to overstate the implications of this policy choice in the longstanding centre of international banking and the foreign exchange market. Its implications do not ultimately depend on the role of the renminbi's internationalization in the decision. Ernie Patrikis, former head counsel at the Federal Reserve Bank of New York and now with White & Case, puts it this way: 'Subsidiarisation would be the end of international banking' (Masters, 2012). See Box 1 for a cautionary tale on subsidiarization.

Dodd-Frank and subsidiaries of foreign banks in the United States

A final issue is Dodd-Frank's application of US capital rules to holding companies of foreign banks in the US on a stand-alone basis. This has often been misunderstood. What must be clear at the outset is that this issue concerns subsidiaries and not branches. (6) Since 1975, the Basel Concordat has always anticipated that the liquidity and solvency regulation of subsidiaries be the primary responsibility of the host regulators (Goodhart, 2011, Chapter 4, 'The Concordat'). There can be no doubt that host authorities have the right to regulate the capital of locally incorporated operations of foreign banks.

The background of the policy is twofold. For years foreign bank subsidiaries have been able to operate in the United States with small, in some cases even negative, capital cushions if the US authorities accepted their consolidated, global capitalization as adequate. That is, the US authorities have looked to the consolidated capital of the parent organization rather than the capital of foreign banks' subsidiaries in the United States. Thus, one could argue that they were not subject to national treatment, since US-based banks could not operate with such capital cushions. In addition, during the global financial crisis, reliance of foreign banks operating in the United States on funding from the Federal Reserve was extensive. Even though the Federal Reserve credit was collateralized, the US Congress understandably sought to limit the official exposure to foreign banks.

The Federal Reserve adopted new rules in February 2014 that require foreign banks' subsidiaries in the United States to be placed under a US holding company that would have to meet local capital and liquidity standards. The original proposal had led the EU commissioner for financial services to write a letter to the US authorities in objection. The final rule allowed foreign banks until July 2016 rather than July 2015 to comply and raised the threshold to $50 billion in assets from $10 billion in assets.

Equity analysts identified Deutsche Bank and Barclays as the foreign banks most affected, with 37% and 23%, respectively, of their balance sheets in US subsidiaries (Van Steenis et al, 2014). Both banks have built securities dealers in the United States on the basis of acquisitions, with Deutsche buying Bankers Trust in 1999 and Barclays the US piece of Lehman Brothers after its collapse in 2008. It is easy to overstate the prospective costs of the new rule to these banks, given their opportunities to move assets outside the United States or to their unaffected branches in the United States as well as to restructure claims by head office on subsidiaries.

What has been generally overlooked in much commentary on the Federal Reserve's final rule-making, however, is the statement by Federal Reserve Governor Tarullo explicitly foreswearing any requirement that foreign bank branches in the United States be made into subsidiaries. Many had read Tarullo (2012) as preparing the ground for forcing branches to become subsidiaries. (7) However, Tarullo (2014) said, 'the rule before us walks a middle road between the vulnerabilities of the status quo and a complete subsidiarization model by, for example, continuing to permit branching'.

IMPLICATIONS AND CONCLUSIONS

From the standpoint of international cooperation in the regulation and supervision of banking, the worst kind of fragmentation arises when some jurisdictions demand less than international standards. The potential for activity to shift to such less demanding jurisdictions can undermine the effectiveness of the international standard and undermine financial stability. The Basel standards have always been agreed as minima, and there are legitimate reasons for jurisdictions to set higher standards.

De-internationalizing global banking first puts at risk the role of the banking system as a conduit for gross and net international capital flows. In the limit, international risk sharing and the financing of international asset and liability positions would fall only on portfolio flows and direct investment. Even if it is recognized that bank flows can enable domestic credit booms to outrun domestic resources (Avdjiev et al, 2012 and Lane and McQuade, 2014), it is not clear that cutting banks out of international capital flows promotes efficiency or welfare.

Second, it is not clear that focusing liquidity and capital regulation on subsidiaries will serve in a crisis. It is true that, despite objections by bankers before the global financial crisis that regulation 'trapped' liquidity in subsidiaries, such liquidity often came in very handy. That is, in the absence of such regulation, liquid resources might well have been redistributed around the consolidated bank at an early stage of the crisis, perhaps in late 2007. In the event, the subsidiary from which the liquidity might have been transferred at an early stage often later itself experienced funding pressure in 2008. The seemingly 'trapped' liquidity often proved very useful in the event.

That said, it is worth considering the long-standing argument that, when push comes to shove, the consolidated banking enterprise is what matters. Former Citibank head Walter Wriston is rarely cited nowadays for his insight into international banking. Indeed, his pronouncement 'countries don't go bust' is widely recognized as dangerous complacency. Still, another of his views holds more water:

It is inconceivable that any major bank would walk away from any subsidiary of its holding company. If your name is on the door, all of your capital funds are going to be behind it in the real world. Lawyers can say you have separation, but the marketplace is persuasive, and it would not see it that way.

(US Senate, 1981)

Banks live as consolidated enterprises jealous of the reputation of each part. For example, when wholesale funding markets became strained in the summer of 2007, Bank of America sought and received permission for its bank to provide repo funding to its broker-dealer in amounts that exceeded intragroup limits (Board of Governors, 2007). These limits had been imposed precisely to prevent the bank's deposit base from serving as a source of strength to the securities operation. This example suggests that the consolidated entity can prove to be a source of strength to troubled subsidiaries. If, when push comes to shove, the consolidated enterprise is most relevant, then bank regulators and supervisors would do well not to lose sight of the consolidated enterprise in the pursuit of stronger subsidiaries.

Thorough implementation of the agreed international rules could allay the regulatory and supervisory impulse for national policies that can fragment banking. If bank regulators and supervisors have confidence in the efficacy of each other's efforts, they may be more willing to accept branches. The new peer reviews of implementation by the Basel Committee on Banking Supervision (2012) could contribute to greater confidence in home regulators' consistent application of what has been agreed in Basel.

Recent developments in major jurisdictions suggest more of a willingness to allow international banking than was evident as recently as mid-2013. Recalling the role of London in financing global trade, the UK authorities will allow branches engaged in wholesale banking with acceptable resolution plans. For their part, the US authorities explicitly accept the branches of foreign banks that represent a significant share of US commercial banking. Whether the US authorities would again allow these branches to provide large sums of dollar funding to their parents remains to be seen. For now, international banking will continue to compete with multinational banking.

Acknowledgements

The author thanks Bill Coen, Ben Cohen, Pat McGuire and Goetz von Peter for discussion. Views expressed are those of the author and not necessarily those of the Bank for International Settlements.

REFERENCES

Aliber, R. 1993: The multinational paradigm. MIT Press: Cambridge, MA.

Avdjiev, S, McCauley, R and McGuire, P. 2012: Rapid credit growth and international credit challenges to Asia. In: Pontines, V and Siregar, R (eds). Exchange Rate Appreciation, Capital Flows and Excess Liquidity: Adjustment and Effectiveness of Policy Responses. The SEACEN Center: Kuala Lumpur, August, pp. 215-244.

Bailey, A. 2013: Regulating international banks. Speech to the British Bankers Association International Banking Conference, London, 17 October.

Bank of England. 2013: News release - Closure of Cyprus Popular Bank Public Co Ltd (Laiki Bank UK) and transfer of all deposits to Bank of Cyprus UK, 2 April.

Basel Committee on Banking Supervision. 2009: Consultative proposals to strengthen the resilience of the banking sector: International framework for liquidity risk measurement, standards and monitoring.

Basel Committee on Banking Supervision. 2012: Basel III regulatory consistency assessment programme. April.

Board of Governors of the Federal Reserve System. 2007: Letter from Robert DeV Frierson to Patrick S Antrim, Bank of America, 20 August, www.federalreserve.gov/boarddocs/legalint/ FederalReserveAct/2007/20070820a/20070820a.pdf.

Cecchetti, S, Domanski, D and von Peter, G. 2011: New regulation and the new world of global banking. National Institute Economic Review 216(1), April: R29-R40, http://ner.sagepub.com/ content/216/1 /R29. abstract.

Cerutti, E, Dell'Ariccia, G and Martinez Pena, MS. 2007: How banks go abroad: Branches or subsidiaries? Journal of Banking & Finance 31(6): 1669-1692.

Committee on the Global Financial System. 2010: Funding patterns and liquidity management of internationally active banks. CGFS Publications: no 39, May.

Duxbury, C and Forelle, C. 2013: Iceland wins case on deposit guarantees. Wall Street Journal, 28 January, http://online.wsj.eom/news/articles/SB10001424127887323375204578269550368102278# printMode.

Financial Services Authority. 2009: Strengthening liquidity standards. Policy Statement, 09/16.

Goodhart, C. 2011: The Basel Committee on banking supervision: A history of the early years, 1974-1997. Cambridge University Press: Cambridge, UK.

Gudmundsson, M. 2011: The fault lines in cross-border banking: Lessons from the Icelandic case. OECD Journal: Financial Market Trends 2011(2): 1-10.

He, D and McCauley, R. 2012: Eurodollar banking and currency internationalisation. BIS Quarterly Review, June, pp. 33-46, http://www.bis.org/publ/qtrpdf/r_qtl206f.pdf.

He, D and McCauley, R. 2013: Transmitting global liquidity to East Asia: Policy rates, bond yields, currencies and dollar credit. BIS Working paper no. 431, October, Basel.

He, Y. 2013: China's Banks in London. Centre for the Study of Financial Innovation: no 111: London, July.

Jones, G (ed) 1991: Multinational and international banking. Edward Elgar Publishing: Aldershot.

Kalse, E. 2009: 1NG Direct was source of success and problems for Dutch bank. NRC Handelsblad, 30 October, http://vorige.nrc.nl/international/article2401602.ece/ING_Direct_was_source_of_success_and_problems_for_Dutch_bank.

Kreicher, L, McCauley, R and McGuire, P. 2013: The 2011 FDIC assessment on banks managed liabilities: interest rate and balance-sheet responses. BIS Working Papers no 413, May, Basel.

Lane, PR and McQuade, P. 2014: Domestic credit growth and international capital flows. The Scandinavian Journal of Economics 116(1): 218-252.

Ma, G and McCauley, R. 2014: Global and euro imbalances: China and Germany. China and World Economy 22(1): 1-29.

Masters, B. 2012: Britain tightens grip on foreign banks, Financial Times 9 December, http://www .ft.com/cms/s/0/3edf0b3a-41ef-lle2-979e-00144feabdc0.html#axzz2WUlq0kGy.

McCauley, R and McGuire, P. 2014: Non-US banks' claims on the Federal Reserve. BIS Quarterly Review, March, pp. 89-97, http://www.bis.org/publ/qtrpdf/r_qtl403i.pdf.

McCauley, R, McGuire, P and von Peter, G. 2012: After the global financial crisis: From international to multinational banking? Journal of Economics and Business 64(7): 7-23.

McGuire, P and von Peter, G. 2012: The dollar shortage in global banking and the international policy response. International Finance 15(2): 155-178.

Shafer, D. 2012: Chinese banks flee London's tough rules, Financial Times 28 October, http://www .ft.com/cms/s/0/3cabad56-2105-lle2-9720-00144feabdc0.html#axzz2WUlq0kGy.

Shin, HS. 2012: Global banking glut and loan risk premium. IMF Economic Review 60(2): 155-192.

Van Steenis, H, Lam, H, Manners, C, Graseck, B and Cyprys, M. 2014: Balkanisation weighs on returns: US foreign bank rules likely finalised 18 Feb. Morgan Stanley Research Europe Wholesale Banks, 13 February.

Tarullo, DK. 2012: Regulation of foreign banking organizations. Remarks to the Yale School of Management Leaders Forum, New Haven Connecticut, 28 November.

Tarullo, DK. 2014: Opening statement by Governor Daniel K. Tarullo, 18 February, http://www .federalreserve.gov/newsevents/press/bcreg/bcreg20140218a-tamllo-statement.htm.

UBS. 2008: Shareholder report on UBS's write-downs, UBS: Zurich, 18 April.

United States, Senate Committee on Banking, Housing, and Urban Affairs. 1981: Financial Institutions Restructuring and Services Act of 1981, Hearings on S. 1686, S. 1703, S. 1720 and S. 1721, 97th Congress, 1st Session, Part 11, pp. 589-590.

ROBERT McCAULEY

Bank for International Settlements, 2 Centralbahnplatz, Basel, 4002, Switzerland.

E-mail: robert.mccauley@bis.org

(1) The subject of this article is often referred to as fragmentation of international banking. This term is often used in a very specific way with regard to euro area financial markets to mean, for instance, the sharp decline in cross-border transactions and positions in the interbank market. This article discusses more general aspects of financial fragmentation.

(2) There was also a secular rise in the share of local claims in all currencies. Local claims refer to claims booked by foreign offices vis-a-vis residents of the host country. Foreign claims sum all cross-border claims and local claims booked in offices outside the bank's home country.

(3) The trend is not evident in claims on emerging markets, where low interest rates in the major currencies have spurred foreign currency borrowing into 2013. See He and McCauley (2013).

(4) According to Duxbury and Forelle (2013), 'As it became clearer that Landesbanki's assets would cover most, if not all, of the balance, the impact of the dispute [between the UK and Netherlands, on one side, and Iceland, on the other] faded'. They report that the court of the European Free Trade Association ruled that the relevant EU directive 'does not lay down an obligation on the State and its authorities to ensure compensation if a deposit-guarantee scheme is unable to cope with its obligation in the event of a systemic crisis'.

(5) Defenders of the UK policy would point to the role of a Cyprus bank's UK subsidiary in the so-called 'British solution' of a Cyprus bank's UK branch deposits (Bank of England, 2013).

(6) But Tarullo (2012) also raised questions about the treatment of branches and agencies of foreign banks in the United States, which in the past have raised funds from US money market funds and onlent the funds to affiliates abroad. They will be subject to 'certain additional measures', especially regarding liquidity. See below.

(7) Tarullo (2012) included: 'Foreign banks as a group moved from a position of receiving funding from their parents on a net basis in 1999 to providing significant funding to non-US affiliates by the mid-2000s--more than $700 billion on a net basis by 2008 .... There should also be liquidity standards for foreign bank branch and agency networks in the United States, although they may be less stringent, in recognition of the integration of branches and agencies into the global bank as a whole'. In fact, by late 2012, the position of non-US banks' branches in the United States had swung around from a net due from affiliates outside the United States to a net due to such affiliates. This resulted from an FDIC levy on short-term wholesale liabilities enacted in Dodd-Frank that applied to US-chartered banks, but not to non-US banks branches in the United States. The competitive dynamics of this could shift banking assets away from US-chartered banks, including subsidiaries of foreign banks, to non-US banks' branches. See Kreicher et al. (2013) and McCauley and McGuire (2014).