VIEWS: 103 PAGES: 32 POSTED ON: 2/6/2011
Studi e Note di Economia, Anno XIV, n. 2-2009, pagg. 155-186 GruppoMontepaschi Emergency Liquidity Assistance at work: both words and deeds matter* MICHELE MANNA** Emergency Liquidity Assistance (ELA) is a central banks’ function as old as these institutions can be. It is also a very modern instrument, as proved by some recent events. The ambition of this paper is first to outline some models of central banks’ communication on ELA; second to redress some common places about what Bagehot would have written on this topic in his Lombard Street; th.rd to review the decisions enacted by policy makers in the banking crises of Sweden (early 1990s), United Kingdom (1995), United States (2001) Austria (2006) and again United Kingdom and United States (2007-2008). Overall, the main message we can gather is that if crisis management is about (re)building trust in financial markets, neither in its words nor in its deeds the central bank should be perceived as hesitant. (J.E.L.: E42, E58, G21, N31, N33) “almost two hundred years after Thornton, the lender-of-last resort function remains an art not a science, to be applied on a case-by-case basis”. Giannini (1999) I. Introduction Emergency Liquidity Assistance (ELA) to financial firms in distress is a central banks’ function as old as these institutions can be, to judge from the fact that its classic and ever-green foundations were laid down by Thornton in 1802 and Bagehot in his 1873 Lombard Street book. Such an honourable history does not call the case closed on ELA and, in the midst of the market turbulence originated in 2007 in the US subprime credit segment, prominent central bank Governors felt appropriate to restate their views on emergency refinancing and on the lesson by Bagehot1. Arguably, issues such as the role of the instrument in crisis management or the balance between transparency in its communication and grip on moral hazard remain part of the debate on *Paper accepted in september 2008. ** Bank of Italy. E-mail: email@example.com My thanks are due to C.O. Gelsomino, who was the source for much food-for-thought on the topic of ELA, R. De Bonis, P. Maggio, C. Mastropasqua, S. Nobili, F. Panetta and the participants to an internal Bank of Italy seminar for helpful comments. I am also indebted to R. Potito for a number of editorial suggestions. The views expressed here are the sole responsibility of the author and do not necessarily reflect those of Bank of Italy. 1 Consider e.g. the speeches given by B. Bernanke, Chairman of the US Fed, on 13 May 2008 and G. Stevens, Governor of the Reserve Bank of Australia, on 15 April 2008. 156 Studi e Note di Economia, Anno XIV, n. 2-2009 central banking. The ambition of this paper is to journey through the topic of ELA by dealing with both theoretical and practical issues. To do so, the paper will revolve around four questions: 1. What is ELA? 2. How central banks communicate on ELA? 3. Which conditions may be applied on the emergency assistance? 4. How do central banks actually intervene in financial crises? The first question could seem a trivial one — basic intuition identifies ELA as central bank credit granted in emergency situations, as suggested by the acronym — if it were not the case that doubts linger on whether ELA is anything different from ordinary monetary policy operations. On a different ground, the emergency condition may be a necessary though not a sufficient one if, say, the Fed’s loosening of its liquidity supply on the aftermath of September 11th is usually reckoned more a form of “lending to the market” than, strictly speaking, ELA. If ELA is part of the armoury of a central bank, as most analysts would agree, this institution cannot escape the problem of how best to craft its com- munication on the instrument. A survey over the Statutes, Acts and other offi- cial documents of 18 central bank helps to identify some basic models of communication, ranging from near silence to formal open statements. The answer to the second question is complemented by an analysis of a number of recent memoranda of understanding on crisis management. Because these documents are less constrained by legalistic language, they offer good exam- ples of how central banks communicate their intended strategies on crisis management and thus on the extraordinary refinancing. Dealing with the third question almost unavoidably implies touching base on the triplet associated to the name of Bagehot, namely that at times of cri- sis the central bank ought to lend (i) at a high rate of interest, (ii) on good banking securities and (iii) freely. In fact, a reading of the original work by this author suggests more nuances than this would-be dictum suggests. Goodhart (1999) is a sure reference on how to dissect the accretion of myths from shallow readings of Lombard Street. So much about the words on ELA. The fourth question looks at the related deeds. Here, I follow a business case approach by reviewing the actions taken within the management of some recent financial crises. The focus will be on the crises of Sweden in the early 1990s, United Kingdom in 1995 (Barings), United States in 2001 (after September 11th), Austria in 2006 and again United Kingdom and United States in 2007-2008 (Northern Rock and Bear Stearns). By way of introduction, I should also define the boundaries of the present work and clarify a terminological element. With a view to striking a middle point between breadth and depth of the research, the list of questions I pro- posed above is not exhaustive and, say, other relevant topics include the inter- national domain of the lender of last resort (Giannini, 1999, and Jeanne and M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 157 Wyplosz, 2001) and the stress testing activity (Sorge, 2004). Furthermore, this is not one more survey on ELA. Not because writing a survey is less rel- evant, but because excellent ones are already available (Freixas et al., 2000, is among my favourites). I should also state my preference for using in this paper the term “emergency liquidity assistance” (ELA) vs. “lender of last resort” (LOLR). The two terms are often used interchangeably, although they may carry different nuances as the latter is perceived as broader in scope. Let us give here the floor to two central banks, Bank of Canada and Sveriges Riksbank, whose official publications featured an article on the “lender of last resort”, as made it clear in the title2. In both cases, the article covers a number of issues relating to the role of the central bank in financial crisis management; howev- er, within each of the two articles, the section referring to the provision of emer- gency credit by the central bank turns to the term “emergency liquidity assis- tance”. Accordingly, I will mainly speak of ELA since the paper chiefly deals with the act of providing liquidity under emergency. This notwithstanding, in citing the literature I will feel free to switch to LOLR if this is the term used by the author and by which her/his work is more often remembered, to avoid dis- tracting the reader with non-substantive elements of notational content. The rest of the paper is organised as follows. Second II to V try to answer to each of the questions outlined above, while in section VI some concluding remarks are put forward. II. What is ELA A basic issue a paper like this should try to address relates to what is ELA in the first place. In terms of the mechanics of the operation, ELA is about granting credit to a financial counterpart, typically a bank, under exceptional circumstances and usually against collateral. Most of these qualifications could equally apply to an ordinary central bank open market operation. For sure, the nature of the counterpart and the request for collateral are not pecu- liar. I will also argue below that the level of the applied rate of interest is not necessarily a discriminating element in recognising an ELA. Thus, one should not be surprised if some authors do not recognize the emergency lending as a separate central bank tool, since any liquidity provi- sion to the market falls, in their judgement, within the broad spectrum of open market operations. Arguably, a noticeable feature of ELA relates to the “extraordinary cir- cumstances” prevailing when the refinancing takes place, circumstances which apply to the financial firm receiving the assistance (and not necessar- ily to the market as a whole). 2Bank of Canada (2004): “Bank of Canada Lender-of-Last-Resort Policies”; Sveriges Riksbank (2003): “The Riksbank’s role as lender of last resort”. 158 Studi e Note di Economia, Anno XIV, n. 2-2009 This firm may be troubled by an IT or other technically-related failure which prevents its ordinary trading in the interbank market. The accident occurred at the Bank of New York in 1985 (a computer failure) proves that this is not academic theory; this being acknowledged, the instance of an ELA in response to a liquidity crisis triggered by an outage is a bit of a rarity. More common seems to be a scenario where some (non-technical) shock leads to heightened uncertainty and more risk averse attitudes. If information about the value of banks’ assets were freely available, market participants would be able to fully conduct peer monitoring and there should be nothing like an “illiquid but solvent bank”, because the latter condition (solvency) would ensure the access to the interbank credit and thus remove the former (illiquidity). In fact, information is often incomplete and costly and this is likely to be even more so during financial crises. To play safe, market partic- ipants curtail their credit lines, without much discriminating between coun- terparts actually affected by the shock and those that have no direct link with the firm setting off the crisis. Under these circumstances, even if liquidity supply is adequate at an aggregate level, some banks could no longer manage to borrow enough funds even though in principle they would be creditwor- thy. As a result, their reserve account with the central bank heads toward neg- ative territory, unless they receive an additional cash injection. Hence, in an ELA operation, we observe an ad hoc cash injection which is targeted by the central bank specifically towards the bank(s) in need, so that a more general increase in the aggregated liquidity supply would not nec- essarily be the right cure to subdue the crisis (although it may help)3. From this, one could gauge that the so-called lending to the market, i.e. when the central bank opts for keeping liquidity supply unusually ample in response to some major event, is not ELA. This is exactly because the central bank goal is to increase the overall supply, and not to modify its distribution at a micro level4. As in many areas of the debate on public intervention, views are split over whether the authority is better placed than market participants to selec- tively aid illiquid banks. In a well-known paper, Goodfriend and King (1988) conclude that there is little evidence that public lending to particular institu- tions is either necessary or appropriate, a view upheld also in Kaufman (1996). Their basic argument is that the central bank should be content with regulating the aggregate liquidity supply, leaving to banks the subsequent distribution of funds as a way to enhance peer monitoring. On a different line of thinking, supporters of free banking theories dispute the need for a central bank and thus, a fortiori, for discretionary forms of interventions. The expe- 3 A formal model that caters for illiquid but solvent banks is in Rochet and Vives (2004), a model of the lender of last resort in the context of a payment system crisis is in Flannery (1996). 4 This paper will cover however also the Fed response to September 11th, as an example of liquidity man- agement in a crisis situation, M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 159 riences of free banking in Scotland from 1792 to 1844 and Canada through- out the nineteenth century are generally rated as rather positive, although the extent to which these were fully free banking systems remains up to debate; a short summary of these two experiences is presented in annex. However, on the whole, the consensus is skeptical on the possibility that a banking systems may prove to be financially stable over a lengthy period without some authority ready to provide ad hoc external funding when cir- cumstances warrant. That is, “some public authority must provide the lender of last resort” (Bordo, 1989). The consensus view builds on two basic arguments. First, it is as a fact of life that embattled banks do receive some type of support more often than not. In a survey over 104 banking crises occurred mostly in the 1980s and 1990s over 24 developed countries, Goodhart and Shoenmaker (1995) count 81 instances were some form of external funding was made available to fail- ing banks and only 23 instances of no funding5. Second, on a more theoreti- cal level, the argument à la Goodfriend and King requires perfect repo and unsecured interbank markets (Freixas et al., 2004). In fact, in the aftermath of a shock the information advantage enjoyed by the monetary authority vis- à-vis ordinary participants is reckoned, if anything, to be widening. Furthermore, while the former weighs the cost of the selective lending (think to the moral hazard) against the benefit of reducing the likelihood of a sys- temic crisis, the latter is concerned with the financial risk-return profile of the deal at stake. As a result, even if the central bank and the single market par- ticipant availed of the same (limited) information set, the ensuing decision taken by these two pehes may turn out be different. III. When the central bank utters the E L A words Communication in this field could be a misnomer if “a lender of last resort should exist, but his presence should be doubted” (Kindleberger, 1978). As implicitly suggested by this quote, one option for central banks in designing their communication on ELA is not mentioning the issue altogether. By say- ing as little as possible on the conditions to receive the emergency lending, the illiquid bank will naturally tend to consider first better-known sources to tap its demand for funds and, hopefully, it will eventually turn to the central bank for ELA only when no alternative is in sight. Alternatively, the central bank can explicitly state its readiness to provide ELA when appropriate, although usually a number of caveats are embedded in the statement. 5 It is fair to acknowledge though that, from the viewpoint of an advocate of free banking, the banking system would find in itself the remedies to the crisis, with no need for external support, if there were no central bank in the first place. A survey of the so-called fiscal costs of a number of financial crises is in Dziobek and Pazarbasiogolu (1997). 160 Studi e Note di Economia, Anno XIV, n. 2-2009 Before proceeding further, it is honest to acknowledge that the gulf between the alternative policy options of near-silence and open disclosure could be narrower than it looks at first sight. As will be argued below, the preference for setting out public guidelines on emergency refinancing proba- bly owes more to earlier interventions on the occasion of a crisis, which make less meaningful the point of being (almost) silent, rather than to theoretical differences on the best communication policy (Sveriges Riksbank, 2003). References to ELA in the law, the Statute and other official documents of central banks A survey of publicly available documents issued by 18 central banks returns three benchmark models of communication on ELA: - in six central banks (Austria, France, Greece, Italy6, Luxembourg and Germany), few or no explicit references about ELA can be in found in their Statute or other legislative texts, nor many utterances have been made by the Governor and other members of the Board7; - in four central banks (Belgium8, Finland9, Ireland10 and Switzerland11), the Statute includes a reference to the duties of the central bank towards pre- serving financial stability, while a more explicit acknowledgement of its potential recourse to ELA can be found in other documents published by the central bank; in the case of the Netherlands the latter statement only applies12; 6 In its Financial Sector Assessment Program on Italy of 2006, the IMF staff states: “Although financial institutions may request emergency liquidity assistance (ELA) from the BI [Bank of Italy], they have not availed themselves of this option since the establishment of a new institutional framework in 1999, pur- suant to the establishment of the ESCB” (page 15; the document is available at http://www.imf.org/exter- nal/pubs/ft/scr/2006/cr0679.pdf). 7 The validity of this statement is confined to the speeches to which I had access browsing the central banks’ websites. 8 Statutes of the National Bank of Belgium, art. 23 (1): “The Bank shall contribute to the stability of the financial systems”. In its website, this central bank adds: “The National Bank holds a key position as the lender of last resort”. 9 Act on Bank of Finland, art. 3: “The Bank of Finland shall also .. participate on maintaining the relia- bility and efficiency of the payment system and overall financial system”. In its website, the Bank Finland states: “The Bank of Finland aims at maintaining confidence in the financial system without having to intervene. The Bank may, however, assume a more active role if the smooth functioning of the financial system is threatened. The last example of the Bank’s intervention was in the early 1900s, during Finland’s business crisis”. 10 Central Bank and Financial Services Authority of Ireland Act 2003, art. 5(b): “the objective of the Bank in contributing to the stability of the State’s financial system”. In its website this central bank clarifies: “The Bank’s responsibility to contribute to the overall stability of the financial system also involves: .. iv. Undertaking official financial operations in exceptional circumstances”. 11 Swiss National Bank Act, art. 5 (2)(e): “[the National Bank] shall contribute to the stability of the finan- cial system”. Swiss National Bank (2006) clarifies further that “the SNB also acts as lender of last resort. In this function, it can provide emergency liquidity assistance for one or more domestic banks”. 12 The Bank Act 1998 of De Nederlandsche Bank does not include explicit references to ELA / lending of last resort. In a speech given in 1992, Governor Wellink elaborating on “Central Banks as guardians of financial stability” emphasised that “Central banks have always been involved in trying to preserve finan- cial stability, for instance through their role as lender of last resort”. M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 161 - in further six central banks (Japan13, Norway14, Portugal15, United Kingdom16, Sweden17 and the United States18), explicit references to the ELA / LOLR can be found in the Statute and other official documents. In between the first and the second group of central banks, one could clas- sify the Banco de España whose Statute explicitly foresees duties in the field of financial stability, but it does not further detail whether this could actual- ly imply the offer of ELA, nor this can easily be inferred from other official documents relating to this central bank19. If an open stance on the subject is a viable option, as the survey suggests it is more common than often acknowledged, it should be of some interest to look more closely at the pre- cautions taken to control for the risk of moral hazard. Let us follow here two central banks that have been particularly transparent on the subject. Bank of Canada (2004) sets out a clear guideline: “the Bank provides ELA only to institutions judged to be solvent”. It also acknowledges howev- er that there may be contingencies which, though unlikely, cannot be dis- missed altogether and call for a different line of action: “In the extremely unlikely event of the failure of more than one LVTS [Large Value Transfer System] participant on the same day... the Bank could be obliged to lend to a failed institution, on a partially unsecured basis, to ensure settlement of the LVTS and so protect against systemic risk”. One could thus understand that, if the worse comes to the worst, the central bank would no longer necessari- 13 The Bank of Japan Law, art. 37.1: “The Bank of Japan .. may provide uncollateralized loans to finan- cial institutions .. and other financial business entities prescribed by a cabinet Order .. when they unex- pectedly experience a temporary shortage of funds for payment due to accidental causes”. The following art. 38.1: “The Prime Minister and the Minister of Finance may request that the Bank of Japan to conduct the business necessary to maintain an orderly financial system, including provision of loans, when it is believed to be especially necessary for the maintenance of an orderly financial system”. 14 Act of Bank of Norway, art. 19 (3): “when warranted by special circumstances, the Bank may grant credit on special terms”. 15 Banco de Portugal, Organic Law, art. 12: “it shall be particularly incumbent upon the Bank to: .. c) Provide for the stability of the national financial system, performing, for the purpose, in particular, the function of lender of last resort”. 16 Memorandum of Understanding between HM Treasury, the Bank of England and the Financial Services Authority, art. 14: “In exceptional circumstances, there may be a need for an operation which goes beyond the Bank’s published framework for operations in the money market. Such a support operation is expect- ed to happen very rarely and would normally only be undertaken in the case of a genuine threat to the sta- bility of the financial system to avoid a serious disturbance in the UK economy...”. 17 Sveriges Riksbank Act, Chapter 6, art. 8: “In exceptional circumstances, the Riksbank may, with the aim of supporting liquidity, grant credits or provide guarantees on special terms to banking institutions and Swedish companies subject to the supervision of the Financial Supervisory Authority”. 18 Federal Reserve Act, Section 10A – Emergency Advances to Groups of Member Banks, art. 1 (“Authority of Reserve Banks to Make Advances”): “any Federal reserve bank may make advances .. to groups of five or more member banks within its district .. provided the bank or banks which receive the proceeds of such advances as herein provided have no adequate amounts of eligible and acceptable assets available to enable such bank or banks to obtain sufficient credit accommodations from the Federal reserve through rediscounts or advances other than as provided in section 10b”. 19 Again, this statement is applicable only to the documents I browsed, i.e. in effect those available on the website of the Banco de España. 162 Studi e Note di Economia, Anno XIV, n. 2-2009 ly abide by the principle of “illiquid but insolvent” and its practical equiva- lent of full guarantee. As to the potential recipient of the credit, these are ordi- narily (if this is the right word in the context of a financial crisis) identified in the banks; however, “Under extreme conditions, the Bank can provide liq- uidity to any firm”. A somewhat similar path is followed in Sveriges Riksbank (2003): “To avoid the unnecessary retention of inefficient banks and distorting bank behaviour, the Riksbank should not support banks that lack long-term survival capacity”20 (emphasis in the original text). “At the same time, it is almost unavoidable that the tendency to refrain from granti- ng liquidity assistance declines where systemic risk is substantial, even if the assessment of the institution’s ability to pay is uncertain”. As to the recipi- ents, while “It is primarily if a bank suffers liquidity problems that the Riksbank has reason to intervene”, it is also the case that “other companies than banks can give rise to contagion risks in the financial system”, thus sug- gesting that the extraordinary support could be extended to other financial firms. The age of memoranda A remarkable feature of the nouvelle vague on the communication on cri- sis management is the signature of a Memorandum of Understanding (MoU) by the authorities entrusted with maintaining financial stability. All the con- sulted MoUs seem to lack the binding power of a law or even of a contract; they define however a commitment and as such they may be regarded as a piece of soft law (Goodhart and Shoenmaker, 2006). Details on thirteen MoUs are reported in Table 2, summary statistics in Table 1. Overall, they distribute evenly both along the domestic/cross border dimension and accord- ing to the nature of signatory parties: (i) central banks; (ii) central banks and banking supervisory authorities; (iii) central banks, banking supervisory authorities and ministries of finance. Four main elements recur in such documents. First, the MoU sets out the responsibilities of the involved parties in respect to financial stability and, notably, crisis management. In doing so, as a rule, most if not all the key ele- ments of the resulting outline are already defined, more or less explicitly, in the applicable legal framework. This is not meant to say that this section of the MoU is a trivial copy-and-paste: in fact, there is an effort in presenting the roles of the authorities in a plain, non-legalistic language. One could thus infer that, in effect, by signing a MoU the authorities mean to speak to an audience broader than the one versed in documents such as central banks’ statutes. Note that the text of the large majority (9 out of 13) of the consult- 20 Sveriges Riksbank (2003) clarifies that “it is not sufficient that the value of assets is greater than the lia- bilities (given a ‘fair’ evaluation) for an institution to be solvent. It is also necessary that the bank has the capacity to generate profits”. Tab. 1 - Consulted memoranda of understanding in the field of financial stability Place of Signatory authorities Public status Date of issuance Link application (1) Australia The Reserve Bank of Australia and the Australian Prudential full text available October 1998 http://www.rba.gov.au/MediaReleases/1998/Pdf/mr_98_joint_rba_ap Regulatory Authority ra_mou.pdf (2) Belgium and the National Bank of Belgium, Belgian Banking, Finance Press release June 2006 http://www.nbb.be/pub/01_00_00_00_00/01_06_00_00_00/01_06_0 Netherlands and Insurance Commission and De Nederlandsche Bank 1_00_00/20060616_memorandum_of_understanding.htm?t=ho&l=en (3) Denmark National Bank of Denmark and Finanstilynet (Danish Banking full text available December 2006 http://www.nationalbanken.dk/C1256BE900406EF3/sysOakFil/MoU Supervisory Authority) _11102005/$File/MoU_DN_FT_dec06.pdf (4) Denmark National Bank of Denmark, Danish Ministry of Finance full text available April 2005 http://www.nationalbanken.dk/C1256BE900406EF3/sysOakFil/MoU and Danish Ministry of Economic and Business Affairs _financial_supervision_april05/$File/Memorandum_uk.pdf (5) Estonia, Latvia, Central Banks of Estonia, Latvia, Lithuania and Sweden full text available December 2006 http://www.riksbank.com/upload/Dokument_riksbank/Kat_AFS/mou Lithuania _est_lat_lit_swe.pdf and Sweden (6) European nordic Central banks of Denmark, Finland, Iceland, full text available June 2003 http://www.nationalbanken.dk/C1256BE900406EF3/sysOakFil/MO Norway and Sweden countries U_eng0603/$File/MoU_eng.pdf (7) European Union Banking supervisors and central banks of the European Union Press release March 2003 http://www.ecb.int/press/pr/date/2003/html/pr030310_3.en.html (8) European Union (a) Banking supervisors, central banks and finance ministries Press release May 2005 (a) http://www.ecb.int/press/pr/date/2005/html/pr050518_1.en.html of the European Union (9) Finland Bank of Finland and Financial SupervisionAuthority reference March 2006 http://www.bof.fi/en/rahoitusmarkkinat/kv_yhteistyo/yhteistoimintap in the website oytakirjat.htm (10) Ireland Central Bank & Financial Services Authority of Ireland full text available October 2003 http://www.centralbank.ie/data/StaOtherFiles/moufs.pdf and Irish Financial Services Regulatory Authority (11) The Netherlands De Nederlandsche Bank and Minister of Finance full text available February 2007 http://www.dnb.nl/dnb/home/news_and_publications/news_and_arch ive/news_2007/en/47-151162-64.html (12) Sweden Ministry of Finance, Sveriges Riksbank and Finansinspektionen full text available June 2005 http://www.riksbank.com/upload/Dokument_riksbank/Kat_AFS/sam radsdok_kris_eng_0602.pdf (13) United Kingdom HM Treasury, the Bank of England and the full text available March 2006 (update) http://www.bankofengland.co.uk/financialstability/mou.pdf Financial Services Authority (a) On 4 April 2008, a new “Memorandum of Understanding on co-operation between the Financial Supervisory Authorities, Central Banks and Finance Ministries of the European Union on cross-border financial stability” was agreed. A detailed introduction is available at: http://www.eu2008.si/en/News_and_Documents/Press _Releases/April/0404ECOFIN_Memorandum.html. 164 Studi e Note di Economia, Anno XIV, n. 2-2009 Tab. 2 - Summary statistics on the consulted MoUs. Domestic Cross-border total Central banks only --- 2 2 Central banks and banking supervisory authorities 3 2 5 Central banks, banking supervisory authorities and 4 2 6 ministries of finance (treasuries) total 7 6 13 ed MoUs is publicly available; where this is not the case, a sufficiently detailed press release has been issued. Furthermore, given the subject of this paper, it is highly relevant that an explicit reference to ELA can be found in six of the nine MoUs whose text is public and no MoU rules out this form of support. Second, the bulk of the MoU deals with improving the cooperation and the exchange of information among the signing parties notably when a crisis breaks out. Because of this, substantial attention is devoted to aspects of coordination and some crisis management committee is established. To avoid that such committee becomes rusty in the interval of time between one crisis and the next, regular gatherings are convened: on a monthly basis in the UK, at least quarterly in Sweden, “periodically” (with no defined frequency at least as far the text of the MoU goes) in the Netherlands, at least once a year in Denmark, monthly “or more frequently” in Australia. By contrast and perhaps surprisingly, no such need for regular gatherings is mentioned in the international MoUs. Third, the MoU outlines the conditions under which the procedures laid down in the MoU itself should be activated. The crisis situa- tion is described in fairly general terms21. But one element that clearly emerges is that it holds a system’s dimension. 21 “if either the RBA [the Reserve Bank of Australia] or APRA [the Australian Prudential Regulation Authority] identifies a situation which it considers is likely to threaten the stability of the financial system, it will inform the other as a matter of urgency” (Australian MoU); “if the financial situation within a finan- cial enterprise or in a financial market is deemed to entail significant risks to financial stability, the issue shall be discussed by the Coordination Committee” (Danish MoU). 22 By way of example, consider the MoU signed by the central banks of the European Nordic countries: “The responsibility for managing a financial crisis in a bank rests primarily with its owners and manage- ment.. Thus, emergency liquidity assistance from the central banks will only be provided in exceptional circumstances”, while the Danish MoU states: “the four parties to the agreement shall seek to find solu- tions whereby owners of base capital and the board and management themselves bear the responsibility to the greatest possible extent”. Likewise, the press release announcing the 2005 EU MoU states that: “The memorandum of understanding should not be construed as representing an exception to (i) the principle of the firm’s owners’ / shareholders’ primary financial responsibility, (ii) the need for creditor vigilance, and (iii) the primacy of marketled solutions to solve a crisis situation in individual institutions”. M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 165 Fourth, besides the exchange of information, two main lines of action are set out: (i) the primacy of market-led solutions and the fact that losses ought to be borne by the owners of the troubled institution in the first place; (ii) when any such option has been verified, no alternative is in sight and excep- tional circumstances warrant, emergency lending assistance is called for22. In short, those that have led the bank to the disaster ought not to benefit from the central bank’s rescue through ELA (Goodhart, 1999). So much about the elements more frequently spelled out in the MoUs. What about those elements which one could expect to find there and, in fact, are less frequently mentioned? One such element is about the financial strings attached to the ELA (say, interest rate and maturity of the credit). As to the otherwise ubiquitous references on the illiquid-but-solvent issue, I was able to extract only one direct quote, from the Nordic MoU: “one of the con- ditions for a decision to provide emergency liquidity assistance is that the bank in question is not judged to be insolvent”. Another less-frequent ele- ment is the governance of the crisis management. With respect to the key question of which authority should have the final word on whether to grant the ELA or not, only a minority of MoUs offer clear guidance. One example is the UK MoU: “Ultimate responsibility for authorisation of support opera- tions in exceptional circumstances rests with the Chancellor”. So does, albeit with different result, the Danish MoU: “Where such public commitments are made, they shall be authorised by the competent authority. In this regard too, Danmarks Nationalbank shall independently make any decision concerning its own participation”. A somewhat intermediate solution seems to be offered by the Dutch MoU: first, it states that “DNB [the central bank] shall act as cri- sis manager. Within its powers, it shall take the measures it deems necessary if the urgency of the situation so requires, including the granting of special credit in exceptional cases”, and subsequently it clarifies that “Should devel- opments at a financial enterprise or otherwise threaten financial stability, pos- sibly requiring measures by the Minister, the Minister shall delegate a repre- sentative of the crisis management team formed by DNB”. More or less along the same lines, the Nordic MoU clarifies that “In cases where a bank is judged to be insolvent or where its solvency is deemed uncertain, the central banks are required to contact their respective Ministry of Finance immediately”. IV. Straight from Walter’s Mouth: Which conditions on ELA? It seems awkward to write a paper on ELA / LOLR without paying trib- ute to Walter Bagehot and his book “Lombard Street” (1873)23. Using a measure of web-popularity such as the number of entries returned by a major 23 I used two versions of the book: an hard-cover book published by The Economist in 1978 and an elec- tronic version at http://socserv.mcmaster.ca/econ/ugcm/3ll3/bagehot/lombard.html. 166 Studi e Note di Economia, Anno XIV, n. 2-2009 Tab. 3 - Web-popularity of Walter Bagehot (number of entries returned by Google on the string “first/family name”, in ‘000s). String Nobel Number of entries researched laureate in 25k 50k 75k 100k 125k 150k Walter Bagehot XXXXXXXXXXXXXXXXXXXXXX> 173k Edmund Phelps 2006 XXXXXXXXXXXXXXX> 112k Robert Aumann 2005 XXXXX> 50k Fiin Kydland 2004 XXX> 34k Robert Engle 2003 XXXX> 40k Daniel Kahneman 2002 XXXXXXXXXXXXXXXXXX // XXX> 259k George Akerlof 2001 XXXXXXXXX> 72k James Heckman 2000 XXXXXXXXXXXXXXXXXXXX> 143k Robert Mundell 1999 XXXXXXXXXXXXXXXXXX> 131k Amartya Sen 1998 XXXXXXXXXXXXXXXXXX // XXX> 1230k Milton Friedman 1976 XXXXXXXXXXXXXXXXXX // XXX> 1250k John M. Keynes (1) XXXXXXXXXXXXXXXXXX // XXX> 902k (1) Keynes died in 1946; the first Prize in Economic Science, more commonly referred to as the Nobel prize in economics, was awarded in 1969 to R. Frish and J. Tinbergen. search engine over his first and family names, Walter Bagehot fares very favourably against the economic Nobel laureates of the last ten years: only Amartya Sen and Daniel Kahneman (1998 and 2002 laureates) beat him in this ranking24. Not a bad performance for someone who became famous for his writings on economic and financial affairs – in 1860 he was became The Economist editor-in-chief – but was not an economist by profession. Alas, when going through the quotes from Bagehot, this author seems to be victim of his own success to the point that his sentences seem to be taken in an acritical way and occasionally one may wonder if second-hand quotes are not passed on. To start with, in Lombard Street Bagehot self-assigned the objective of writing about what he called concrete realities, because he explains that the money market can be described in plain words, and the read- er should not seek in his book for fixed rules, rather all nuances should be borne in mind to be appreciative of his lesson. Moreover, his analysis should be cast against the institutional settings prevailing at the time – notably, the gold standard and the then internal organization of the Bank of England – and some adaptations are in order to extract practical lessons for today’s world. Let’s revert to his original text on the scope of and conditions on emergency lending: “The end is to stay the panic; and the advances should, if possible, stay the 24 By the same yardstick, Bagehot is dwarfed by Milton Friedman and John Maynard Keynes, but so would do most other well-established economists. M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 167 panic. And for this purpose there are two rules: First. That these loans should only be made at a very high rate of interest... Secondly. That at this rate these advances should be made on all good banking securities, and as largely as the public ask for them”25. Quotes like this are often summarized by stating that emergency lending should be granted (i) at a high rate of interest, (ii) on good banking securities and (iii) freely. I shall now look at each of these three elements in turn. The high rate of interest First of all, “high” should not be read as penalty. In no instance, the p- word is used in Lombard Street next to the concept of the applicable interest rate. Indeed, Bagehot does not provide many details on this specific issue and the only safe element we have is his insistence that the central bank (the Bank of England) ought to step in at a relatively early stage of the crisis, before ten- sions mount excessively26. Hence, the rate applied on ELA should be high compared to the pre-crisis levels but not necessarily higher than those pre- vailing at the peak of the crisis (if any market rate is then really available)27. To gather more insights, a good starting point is the emphasis placed by Bagehot on the demand for reserves of the Bank of England originated out- side and within the country, the foreign and domestic drain in his speak. With regard to the former, he argues in favour of a large reserve of legal tender – because of the gold standard, he was concerned with large withdrawals of bullions from the country – which needs to be protected by a rise in the rate of interest. In respect to domestic drain, Bagehot’s recipe is simple enough: lend freely to restore trust in the market. Thus, if one fancies taxonomies, then he would associate the “high interest rate” to foreign drains and the “free lending” to the domestic ones. Now, because as observed repeatedly by Bagehot, “the two maladies – an external drain and an internal – often attack the money market at once”, the two tools often end up being jointly applied in this author’s world. Perhaps for this reason, the tradition wants that, according to Bagehot, the interest rate should be raised in any crisis. In fact, disappeared the gold standard and if no external constraint such as some form of (soft or hard) currency pegging is binding, the foreign drain described by Bagehot becomes less likely. But then, a mainstay for charging high interest rates on the lending-of-last-resort fades. 25 Page 147 of the version of the book by The Economist (see fn. 23). 26 Bagehot clarifies that “The rate should be raised early in the panic, so that the fine may be paid early; that no one may borrow out of idle precaution without paying well for it; that the Banking reserve may be protected as far as possible”. 27 Goodhart (1999): “Certainly the rate should be above that in effect in the market prior to the panic, but not necessarily above the contemporaneous market rate”. 168 Studi e Note di Economia, Anno XIV, n. 2-2009 One possible additional element of confusion may be the view that the hike in the interest rate is a way to contain the moral hazard: banks would refrain from applying for lending of last resort because this credit is dear. This interpretation could, in turn, reflect the notion that the “high” interest means to punish the banker that dared to request ELA. If that were the case, it would probably represent a light punishment indeed. Let follow a simple numerical example based on 2006 balance-sheet data of one of Italy’s largest banks: firstly, take its gross outstanding exposure against all other banks; sec- ondly, assume that out of the blue all its counterparts in the interbank market withdraw their credit; thirdly, the bank applies for an equivalent extraordi- nary lending by the central bank; finally, the latter one charges on the loan the current overnight interest rate plus five percentage points. Well, putting all these elements together – a sequence which looks fairly extreme even against the vagaries of financial crises – the extraordinary credit needs a maturity of more than 40 days to chop off just one tenth of the net annual income of the troubled bank. Really a bargain for the management that led the bank to sail in such dire straits! More credible looks the argument that banks do not eas- ily apply for lending-of-last-resort because of a reputational risk. Indeed, it is widely accepted that even in a context of a widespread crisis, some elements of poor management should have been in place if the embattled bank (only one or a few out of many!) requested the extraordinary refinancing. Against this line of reasoning, the case for a high interest rate to any emergency lend- ing weakens considerably. As convincingly argued by Martin (2005), in the aftermath of September 11th the policy of the US Federal Reserve of keeping the federal funds rate very close to zero proved to be quite effective in man- aging a crisis potentially disruptive also from a financial standpoint. Weakening the case is not zeroing it. Bagehot argues in favour of a high rate of interest as a heavy fine on unreasonable timidity. His point is about denying the emergency credit to those firms that do not strictly require it. If more banks apply for the extraordinary refinancing and if, in the reckoning of the central bank, to meet only some of the bids suffices to tame the ten- sions, then a hike in the interest rate may well run the job of any price: the self-selection mechanism which eventually equals the demand for and the supply of a good – in this case, the ELA. Otherwise, that is when it is one bank applying for ELA or when the central bank’s action is geared towards a substantial loosening of the liquidity supply across the board (as the Fed after 11/9), it is not obvious why charging a high rate of interest should be the right policy in respect to ELA. On good banking securities As regards the guarantees, Bagehot’s prescription is that LOLR should be granted against whatever in normal times is understood as good banking securities. The author is well aware of the fact that the identification of such M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 169 a security becomes a complex task in crisis times and, it seems to me, he argues in favour of a pragmatic approach28. A vivid example of the difficulty of sorting out good assets from a firm which typically would have few of them at the time it bid for ELA is offered by the role played by Freddie Mac CMOs among the assets guaranteeing the loan extended by the Federal Reserve Bank of New York to Bear Stearns in 2008 (see next section, in par- ticular fn. 65). Asking questions on the quality of the guarantees backing an ELA is doomed to introducing a further reference to the issue of illiquid but solvent state of the bank, often associated with Bagehot’s dictum. To cut a long story short – can the central bank actually ascertain whether the bank applying for ELA is solvent or not within the tight time frame available for effective crisis management? – it may suffice to observe that Bagehot did not really directly touch on the issue of insolvency. In fact, the reader is brought to understand that in this author’s framework if a bank avails of good securi- ties, then it is also solvent: the first concept in effect absorbs the latter one. The request for good banking securities was aimed mainly at protecting the Bank of England from possible losses. Noticeably, when Bagehot refers to the Bank of England, he had in mind mainly its then Banking department which acted, more or less, as an ordinary bank bar for the higher prestige it had acquired over the centuries29. This may explain his concern about poten- tial losses borne by the Bank and thus the need for acquiring good banking securities before granting the lending of last resort. Only at the margins, he mentions the Issuing department which was authorized to issue banknotes against Government securities and was closer in concept to a public institu- tion. Shall this concern hold true today? Building on the results of an extensive survey of central banks’ financial conditions, Stella (2002) argues that if credibility is important for the success of its policies, the central bank should better be financially strong30. It is beyond the scope of this paper to enter the fray of which authority should ultimately bear the costs associated with the support of critically-ill banks. To offer a flavour of the debate, the matter may be about size, namely if the central bank can bear on its books the losses orig- 28 “if bankers for the most part advance on such security in common times, and if that security is indis- putably good, the ordinary practice of the Bank of England is immaterial. In ordinary times the Bank [of England] is only one of many lenders, whereas in a panic it is the sole lender, and we want, as far as we can, to bring back the unusual state of a time of panic to the common state of ordinary times”. 29 The Bank of England was a private institution and its capital strength was important as for any such institution. As to its Banking Department, this “can no more multiply or manufacture bank notes than any other bank can multiply them” even if “All London banks keep their principal reserve on deposit at the Banking Department of the Bank of England” (Bagehot, 1873). 30 The literature on central banks financial strength has gained new momentum in the light of the emer- gency lending carried out during the subprime crisis (e.g. Buiter, 2008). The Bank of Chile, which pro- vides a counter-example of a central bank with a positive record in monetary policy making while oper- ating with a negative capital, is widely regarded only as an exception to the rule. 170 Studi e Note di Economia, Anno XIV, n. 2-2009 inated by an ELA, then she will be in charge, otherwise it will up to govern- ment (albeit likely acting on the advise by the central bank and the supervi- sory authority) to decide on how to best handle the crisis (Goodhart, 1999). Alternatively and perhaps more conventionally, the financial position of the troubled bank is regarded as the discriminating element: if the bank is clear- ly insolvent, then the government is in charge (Freixas et al., 2000), other- wise there may be a presumption for a direct role of the central bank. Thirdly and unsurprisingly, the proponents of the free banking school argue against any public intervention at all (Dowd 1988). Lend freely The concept of lend freely contains a hard and a soft dimension. The for- mer is the one most typically remembered: the central bank should not fix quotas or maximum amounts of ELA but rather meet all the demands for the extraordinary refinancing, so long as the troubled bank is able to stack good banking securities and is willing to pay the required interest rate. It seems to me however that the key (but less often stressed) message rests with the lat- ter dimension, the soft one. This deals with communication. Let us turn once more to Bagehot: “What is wanted and what is necessary to stop a panic is to diffuse the impres- sion, that though money may be dear, still money is to be had... To lend a great deal, and yet not give the public confidence that you will lend sufficiently and effectually, is the worst of all policies; but it is the policy now pursued”31. The message cannot be misunderstood: what matters most is not the vol- ume of emergency lending actually granted but rather the attitude of the cen- tral bank towards the availability of meeting any request of such lending (obviously, at given financial and non-financial conditions) and the outside perception of this attitude. To sum it up: large volumes of lending which are not supported by ade- quate communication could bring about little effect; conversely, if the psy- chological element – the uncertainty which under a crisis leads market play- ers to stop dealing even with reasonably sound counterparties – is checked, then even small volumes of ELA may do wonders. The emphasis on com- munication raises a somewhat related policy issue. The set of actions to be pursued – the central bank steps in at a early stage of the crisis and does so by injecting confidence even before than funds – requires clear command lines so as to swiftly and effectively act under time pressure. In the end, the market expects to hear the voice of the authority at the helm loud and clear. 31 Page 79 of the version of the book by The Economist (see fn. 23). M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 171 V. Crisis management at work To draw lessons on the concrete actions pursued by central banks, I review a few recent banking crises: Sweden in the early 1990s, United Kingdom in 1995 (Barings), United States in 2001 (after September 11th), Austria in 2006 and again United Kingdom and United States in 2007-2008 (Northern Rock and Bear Stearns). In doing so, with the aim to focus on the actual crisis man- agement, I will try to be brief on the events which led to the crises. The Sveriges Riksbank and the crisis of 1992-1993 The Swedish crisis of 1992-1993 took place against the backdrop of a typ- ical boom-bust switch: after a prolonged phase of strong growth –featuring i.a. overinvestment, especially in the real estate sector, and lowered credit standards by banks chasing for higher market shares in a newly deregulated market – the economy entered a recession with diving property prices. Already by the end of 1990, reported credit losses were substantially higher than in earlier years (Englund 1999). In 1991, house prices started to dive and arrears from lending related to real estate to pile up. Finally, the fixed- exchange rate policy requested high interest rates to support the Swedish krona, with a hike on the occasion of the EMS crisis in 1992. The combina- tion of these events led the seven largest banks, accounting for nine tenth of the banking market, to accumulate losses for approximately 12 percent of Swedish. To manage the crisis, the Swedish Ministry of Finance, its central bank and the financial supervisory authority pursued four main lines of action. First, the stance of the public communication was kept open and transparent, geared towards the taxpayer who should feel reassured that her money was not going to be used to rescue the banks’ owners. Second, the aforementioned authorities set up a dedicated institution, the Bank Support Authority (BSA), to manage the support provided during the crisis, while facilitating the exchange of expertise between the three parties. Third, a broad political con- sensus was sought, to prevent the effectiveness of the adopted measures being jeopardised by political feuding32. Fourth, the parliament approved in December 1992 a “bank support guarantee”, which protected the banks’ cred- itors but not the banks’ shareholders; quite crucially, no upper limit was set for the State support (Ingves and Lind 1996). To customize the intervention, banks were ranked according to a three-tier scale (Andersson and Viotti 1999). A bank was classified as of type A if its capital ratio was forecast to decline toward but remaining above the 8 per cent limit, as of type B if the capital ratio could temporarily decline below the 8 percent limit (but not below nil) and in any case the bank was expected to 32 The political opposition was represented in the board of the BSA. 172 Studi e Note di Economia, Anno XIV, n. 2-2009 regain profitability in the medium term, and as of type C if no such expecta- tion was in place so that its capital ratio was eventually bound to become neg- ative. The source of the support changed accordingly: the problems of a A- Bank could be solved by its shareholders, possibly with some temporary exten- sion of the State guarantee; those of a B-Bank required, besides the capital injection by the shareholders, a more extensive State guarantee; finally, a C- bank was closed or merged with another bank in sounder financial conditions. Overall, the crisis management can be judged as a success story, if already in 1995 the Swedish banking system reported, on aggregate, strong profits. Over a shorter horizon, no major signs of a general lack of trust towards the domestic financial system mounted, although some occasional larger-than- usual cash withdrawals from banks were recorded. As to the wider effects on the economy, on the hindsight symptoms of credit crunch appear to have been fairly short lived. Because of the general bank guarantee provided by the State, “there was no need for more direct government intervention in indi- vidual banks. As it turned out, the banking system outside of Nordbanken and Gota recovered, partly with new equity from their owners” (Englund, 1999). The Bank of England and the collapse of Barings in 1995 Barings Bank was the oldest merchant banking company in London until its collapse in 1995 due to the financial positions taken by one of its employ- ees, Nick Leeson. The main facts behind the Barings case are well known33. In 1992 Leeson was appointed general manager of Baring Securities Limited, located in Singapore, where he was also head trader and de facto head of the back office. His trading in futures and options (run through the famous 88888 accounts) led to cumulated losses of some £ 208 million by December 31st, 1994. Such losses were hidden and he had arranged for reporting handsome profits. By February 27th, 1995, the accumulated losses spiralled to 830 mil- lion, leading to Barings’ demise. In a nutshell, the crisis ultimately owed to the “serious failure of controls and managerial confusion within Barings” (Bank of England, 1995a). From a crisis management point of view, the fact that the crisis escalated at speed light had the potential to foster widespread panic in the markets34. On February 24th, Friday Barings’ staff in Singapore phoned Barings’ headquarters in London with the news on the immense loss- es hidden by Leeson in his “five-eight” account. After a frantic search for more information and possible solutions, on late February 26th, Sunday the 33 Information on the Barings collapse is available from a number of sources, including The Bank of England Report into the Collapse of Barings Bank (1995) and Leeson own account of the crisis is his book “Rogue Trader” (1996, E. Whitley co-authored). 34 One-quarter of Barings’ gross losses were incurred on a single day (Reserve Bank of Australia, 1995) and it is reckoned that if only Leeson’s game had been unveiled by end 2004, Barings could have remained solvent. In the ensuing notes I follow a series of articles by E. Ipsen for the International Herald Tribune, appeared on 27 and 28 February and later on 3 March 1995. M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 173 Bank of England announced that there would have been no rescue for Barings, as a result of which the troubled bank was put into administration35. Two explanations were brought forward. First, the full scale of damage at Barings was unclear, an element which likely made other institutions poten- tially interested in the sale of Barings highly cautious. Second, Barings was judged not systemically relevant and not worth an injection of cash directly by the Bank itself; ELA was however one of the options on the table, even it is hard to assess from the outside how close was the extraordinary liquidity injection to become a reality36. However, panic reactions in financial markets could not be ruled out and on February 26th the Bank of England (1995b) announced its readiness “to provide liquidity to the banking system to ensure that it continues to function smoothly”. The aim was clearly one of building trust by making it clear that aggregate liquidity supply would have been kept adequate; incidentally, a message with little difference in the substance was echoed by the Fed in the aftermath of September 11th (see below). In fact, not much happened in the markets: on February 27th, Monday, spreads between Treasury and eligible bank bills widened only fractionally for a short time and swap spreads were unchanged, while in the equity market there was less selling than expected and the FT-SE 100 fell only by 0.4%. To recap, the Bank of England announced its readiness to provide ample liquidity, fulfilling the Bagehot’s tenet of “lend freely”; note that no mention was made of penalizing the demand for liquidity by hiking the interest rate (another element of similarity with the policy pursued by the Fed on 11/9). It also assessed the risk for financial stability at a system’s level: having judged Barings not systematically relevant, it rejected the option of a direct “cash injection” to the troubled bank. Once again, crafting an appropriate, non-hes- itant communication from the central bank towards the financial market likely played an important role in the overall crisis management. This should not be perceived as less pivotal only because Barings was not a major player, when measured by the scale of its operations in the city. Indeed, the news brought a shock if the Chancellor of the Exchequer, Mr. Kenneth Clarke, stated before the Parliament that “This failure is, of course, a blow, to the City of London”. The Federal Reserve in the aftermath of September 11th, 2001 An almost real-time effect of the attacks against the World Trade Center, obviously aside the huge loss of life and property, was the virtual halt of many New York-centred financial markets. From September 11th through 35 It was only a few days later that administrators for Barings PLC reported that they were in talks with the Dutch bank Internationale Nederlanden Groep (ING) NV. ING purchased Barings for the nominal sum of £1 along with assumption of all of Barings liabilities. 36 The reckoning of a low risk of contagion was restated by the then Governor of the Bank of England, Sir Eddie George, who was quoted by press sources that the collapse “does not indeed have serious wider implications for other institutions”. 174 Studi e Note di Economia, Anno XIV, n. 2-2009 17th, when eventually more ordinary conditions started to resume, the trans- fer of liquidity in the interbank market was at a standstill. Some intermedi- aries were materially not able to conduct business as a direct result of the dis- ruptions caused by the attacks – e.g. Bank of New York, one of the two main clearing banks, had its headquarters evacuated37. Other banks found it diffi- cult to locate borrowers, also because disruptions wracked telephone equip- ments (McAndrews and Potter, 2002). Finally, irrespective of the actual state of payment and communication systems, most banks and dealers took a pru- dent stance refraining from making payments until later than normal in the day before being reassured about their general cash position (Federal Reserve Bank of New York, 2002). Thus, market coordination failures added to objec- tive difficulties in trading. Some mitigating factors were at play, however. First, there was no run to banks in the United States. Currency in circulation increased only by 0.7 per cent in the first two days after the attack, of which two thirds of the increase reflected larger banks’ vault cash holdings38. Second, the incident that had brought the markets to a halt was clearly identifiable and it can be classified, in the dry taxonomy of financial crises, as an operational shock rather than as a credit shock; in a related way, solvency of relevant market players was not in doubt. The Federal Reserve reacted with firm communication and a substantial loosening of liquidity conditions. On September 11th, at 9:44 it broadcasted a first message over the Fedwire, declaring that the system was “fully opera- tional at this time and will remain open until an orderly closing can be achieved”. This was followed at 11:25 by a press release announcing: “The discount window is available to meet liquidity needs”. Because the ordinary policy is to limit the borrowing from this window, which is usually accepted only when an institution is deemed to have exhausted alternative sources of credit39, the announcement was (correctly) interpreted as a marked shift in stance by the Federal Reserve, towards a robust liquidity easing; this inter- pretation was supported by additional words from senior Fed officials40. 37 Also because of the communication and operations difficulties, this bank could not initiate payments on Fedwire, resulting in a large accumulation of funds (with a backlog of payments reckoned in the order of USD 100 billion, Lacker, 2004). 38 My calculations on data from Lacker (2004). 39 To avoid an easy arbitrage opportunity – at the time, the discount rate was normally set lower than the federal funds target rate and thus generally below also the money market rates – the funds obtained through this window were not meant to be re-lent in the interbank market and, more broadly, a strict moral suasion was enforced (in ordinary conditions). As a result, on aggregate the discount window usually con- tributed only marginally to the total Federal Reserve refinancing. 40 “The Fed is the lender of last resort. If credit is needed to make transactions go, the Fed will provide it”; “I am sure that central bankers everywhere will do everything possible to maintain calm and seek to ensure the world economy functions smoothly in the face of this horrendous deed” (statements from Fed Governor E. Gramlich and the New York Fed President W. McDonough, see Lacker, 2004). M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 175 Concerning the liquidity management, “From Wednesday [September 12th] through the following Monday [September 17th], the sizes of open market operations were aimed at satisfying all the financing that dealers wished to arrange with the Desk, in order to mitigate to the extent possible the disrup- tions to normal trading and settlement arrangements” (New York Federal Reserve, 2002). As a result, refinancing soared: on Friday 14th the outstand- ing stock of overnight repos peaked at USD 81 billion, when a stock of 40 billion is reckoned at the upper side of the range prevailing under ordinary conditions. Additionally, the New York Federal Reserve simplified in many ways its credit allotment procedures41. Where the Fed deviated from the would-be textbook recipe of crisis management and emergency lending was about the “high interest rate” rule. In fact, it lowered the interest rate at which liquidity was provided42. As a result, in the week until September 18th a sub- stantial amount of the trades in the (overnight) federal funds market took place in range 0 to 3 percent. With the hindsight, the Fed can be reputed to have done the right thing – in the light of the quick recovery in the activity of the inter- bank market and the circumstance that the shock did not lead to the default of market participants, nor to further disruptions in other financial markets. Moreover, the Fed’s policy of lowering interest rates can be reconciled with more classic approaches if one “recognize that Bagehot had in mind a commodity money world, while the Fed operates in a fiat money world” (Martin, 2005). The Österreichische Nationalbank and the rescue of Bawag in 2006 In 2006 BAWAG P.S.K.43, Austria’s fourth-largest bank at the time, faced lawsuits by creditors of the bankrupt U.S. broker Refco – of which the Bank had been a business partner – claiming up to USD 1.3 billion plus additional restrictive measures44. It quickly became clear that the BAWAG was no longer “in a position to ensure full compliance with capital adequacy provi- sions for the current year  and to close the accounts for 2005, as it would have to allocate substantial funds to provisions” (OeNB, 2007). 41 The Desk operated relatively late in the day to act more effectively as a lender of last resort; it arranged no outright operations and relied exclusively on temporary ones; repos were arranged as a single tranche without differentiating between collateral types. 42 “the Desk .. had to accept the vast majority of propositions – even those offered at rates well below the new 3 percent target level [in force as from September 17th] – in order to arrange RPs of sufficient size” (New York Federal Reserve, 2002). Moreover, the reliance on the discount window was, in itself, a way to reduce the average refinancing rate given the structure of interest rates in the Fed’s operational frame- work. Finally, the Fed waived daylight overdraft fees and overnight overnight penalties for the period September 11th through September 21st (Potter, 2002). 43 Short form for Bank für Arbeit und Wirtshaft und Österreichische Postsparkasse AG. 44 The claim by creditors included “a court-ordered temporary freeze of its [BAWAG] U.S. asset holdings, after having granted Refco a loan of hundreds of millions of U.S. dollars shortly before Refco filed for Chapter 11 bankruptcy protection in the fall of 2005” (OeNB, 2007). 176 Studi e Note di Economia, Anno XIV, n. 2-2009 On May 2nd, 2006, the OeNB released a press release on the “solution to Bawag P.S.K.” announcing an Austrian federal government guarantee of up to EUR 900 million, of which 450 million stood available to improve the cap- ital situation of the troubled bank. The bail-out was agreed shortly afterwards creditors of Refco settled the 1.3 billion lawsuit in New York. As part of the deal, Bawag and its owner, the Austrian trade union federation OeGB, would have sold their stake of a combined 20 percent in the Austria’s central bank to the Austrian government for a nominal amount of EUR 2.5 million (Reuters, 2006). The OeNB press release also stated that representatives of the leading Austria’s financial institutions had met under the guidance of the central bank itself to decide on further steps (OeNB, 2006a). On June 16th such steps found concrete applications in the injection by four credit institutions and four insurance companies of a total of EUR 450 million, a sum which allowed the Bawag to achieve again at a group level an adequate capital ratio45. This injection was however well known already as at May 2nd (Financial Times 2006). Three facts stand out in this crisis: (i) one bank was under fire; (ii) there was clearly an issue about the solvency of this bank46; (iii) the crisis was orig- inated by a credit-type shock due to choices by the bank’s management. Besides brokering the rescue47, the central bank was active on the external communication too. I mentioned above on the press release issued at the peak of the crisis. As to the earlier communication sent out in late April, following the surge in customer withdrawals, the OeNB had tried to assuage the con- cerns by announcing “that it would ensure that BAWAG would immediately have the necessary cash on hand in the event of a liquidity bottleneck” (OeNB, 2006a; I was not able to identify the exact date of this statement). Furthermore, on April 28th the central bank stated that the BAWAG had suf- ficient capital at its disposal and that deposits with the bank were not at risk. With the hindsight, this press release was not much effective in allaying fears: 45 The four credit institutions BA-CA, Erste Bank, ÖVAG and RZB and the four insurance corporations Allianz, Generali, Uniqa and Wiener Städtische founded two corporations on Wednesday to provide sup- port for BAWAG P.S.K. While BA-CA, Erste Bank and RZB will each contribute EUR 100 million and ÖVAG EUR 50 million in capital to one corporation, each of the four insurance companies will contribute EUR 25 million to the second corporation. BAWAG P.S.K. will hold a controlling stake of 20% in both corporations” (OeNB, 2006c). 46 “As the events unfolded, it became clear that BAWAG P.S.K. .. would not be in a position to ensure compliance with capital adequacy provisions for the current year and to close the balance sheet for 2005” (OeNB, 2006a). 47 In describing its role in the crisis, OeNB (2007) states: “in the case of BAWAG P.S.K. it also contributed substantially to finding a way out. In this process the OeNB benefited from the positive effects and syn- ergies arising from its ability to cross-check its macroprudential findings with hindsight and experience gained from microprudential supervision, and was able to strengthen the pivotal crisis management role that it has as a lender of last resort”. M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 177 cash withdrawals by banks’ customers continued, at a pace of as much as EUR 50 million daily, bringing some Bawag branches to run short of cash. The Bank of England and Northern Rock in 2007 The turbulence which in 2007 originated in the US subprime segment led to a sequence of events not uncommon in financial crises. After a prolonged phase of easy credit and asset inflation (with substantial hikes in house prices in this specific case), arrears in mortgage payments, forced house sales and lower prices started to compound each other. Feeding through the value of securitized assets, these patterns led to a dramatic change in market sentiment which brought about the illiquidity of many segments of financial markets48. As a result, institutions which relied more extensively on capital markets for their financing and invested on onceliquid assets felt the bruise more severe- ly. The English Northern Rock Plc, which until its demotion was a top-ten UK bank with a special prominence in the mortgage business, was the epito- me of such institutions49. On September 13th 2007, the BBC broadcasted the news that Northern Rock had asked for and received emergency financial support from the Bank of England. It did not take long for queues forming outside Northern Rock’s branches50. The following day the news was confirmed by a press release through which the Chancellor announced the authorization to the Bank of England to act as lender of last resort towards Northern Rock, against appro- priate collateral and at an interest rate premium51. On October 9th additional facilities were made available to Northern Rock through the Bank of England. In this case however, the ultimate risk of default was to be borne by the Treasury rather than the Bank. By late October, Northern Rock had drawn down £13-14 billion. In parallel, a number of measures were announced to guarantee the Northern Rock deposits, to cover eventually 100 per cent of the first £35000 of deposits52. Further extensions of the guarantee to wholesale obligations by the 48 Virtually any central bank has presented its analysis of the roots and the implications of the crisis. To mention just one reference, the reader could consult the IMF Global Financial Stability Report of April 2008. 49 At the end of 1997, Northern Rock had assets on a consolidated basis of £15.8 billion (euro 20 billion). It was one of two FTSE 100 headquartered in the North East England. 50 Report of the Treasury Committee of the House of Commons, “The Run on the Rock”, http://www.par- liament.the-stationery-office.co.uk/pa/cm200708/cmselect/cmtreasy/56/5602.htm, 24 January 2008. 51 Tripartite Statement by HM Treasury, Bank of England and Financial Services Authority, “Liquidity Support Facility for Northern Rock plc”, 14 September 2007, http://www.bankofengland.co.uk/publica- tions/news/2007/090.htm . 52 Tripartite statements by HM Treasury, Bank of England and Financial Services Authority, “Northern Rock plc Deposits” and “Extended guarantee and additional facility for Northern Rock plc”, respectively 9 and 11 October 2007, http://www.bankofengland.co.uk/publications/northernrock/index.htm. 178 Studi e Note di Economia, Anno XIV, n. 2-2009 English bank were announced on 18 December53. In the judgement of the Treasury Committee of the House of Common “the run on Northern Rock was largely triggered by the announcement of the Bank of England’s support operation”. Ultimately, the issue was one of “no sign of a communications strategy of the Tripartite authorities during the crisis of September 2007. We believe that this was a contributory cause of the run on the bank.” (Committee Report, see fn. 50, § 213 and 289). It is difficult to add much to the extensive analyses carried out by the British authorities on the Northern Rock crisis54. One element that stands out is the room taken in these analyses by elements relating to external commu- nication and crisis governance. In relative terms, elements such as the inter- est rate applied to the emergency credit and the specifics of its financial con- ditions drew much less attention, rightly so I believe. The Federal Reserve and Bear Stearns in 2008 A second well-known casualty of the 2007-2008 market turbulence has been the US firm Bear Stearns, one of the largest world investment banking and securities trading houses. Against the rapid deterioration of the liquidity position of the firm, on the evening of March 13th 2008 the SEC informed the Fed New York that it expected Bear Stearns to file for bankruptcy protection the next morning. The Fed responded with a three-pronged strategy55. As an immediate fix, on March 14th, Friday, it extended through its discount window a 13 billion dollar overnight loan to JPMorgan Chase, so that this bank could on-lend to Bears Stearns. The loan was repaid at maturity, that is on March 17th, Monday. Already on March 13th, the new Primary Dealer Credit Facility (PDCF) had been launched and, in effect, allowed primary dealers and thus also Bear Stearns to borrow from the Fed56,57. Beforehand, as being a non-depository 53 HM Treasury press release, “Northern Rock plc: Extension of wholesale guarantee arrangements”, 18 December 2007, http://www.hm-treasury.gov.uk/newsroom_and_speeches /press/2007/press_149_07.cfm. 54 Besides the Treasury Committee Report, as regards the views by the Bank of England on the matter the reader may consult e.g. the speech by the Bank’s Governor M. King on 9 October 2007, http://www.bankofengland.co.uk/publications/speeches/2007/speech324.pdf. As to the UK FSA, a handy text is the March 2008 summary of its internal review on the supervision of Northern Rock, http://www.fsa.gov.uk/pubs/other/exec_summary.pdf. Senior officials from both institutions also offered their evidence to the Treasury Committee, the link to the one by M. King is http://www.publications.par- liament.uk/pa/cm200708/cm select/cmtreasy/56/7092002.htm. 55 Testimony of T. F. Geithner, President and CEO of the Federal Reserve Bank of New York, before the US Senate Committee on Banking, Housing and Urban Affairs, 3 April 2008, “Actions by the New York Fed in response to liquidity pressures in financial markets”. 56 Primary dealers are banks and securities brokers-dealers that trade in U.S. government securities with the Federal Reserve Bank of New York. 57 In many ways, the PDCF is close in spirit to the discount window. Compared to the open market oper- ations, the recourse to the PDCF is allowed on condition of posting a much broader range of collateral; moreover the rate applied is fixed and is not the outcome of an auction (“Understanding the Recent M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 179 institution, Bear Stearns had no such direct access. The establishment of the facility took place under Section 13(3) of the Federal Reserve Act.58 Moreover, the New York Fed eased the acquisition of Bear Stearns by JPMorgan Chase (the only institution which would have expressed interest in the ailing firm), by both brokering and financing the deal. As to the financ- ing, the Fed extended a loan of 29 billion dollar to JPMorgan, secured with a pledge of Bear Stearns valued at approximately 30 billion (as of March 14)59. This financing too took place under Section 13(3) of the Federal Reserve Act. The deal between the two firms was finally struck on 24 March, in the early morning hours. For the purposes of this paper, four elements of this crisis and of its man- agement are noteworthy. First, it is a good reminder that central banks can play multiple roles in a liquidity crisis, besides providing the funds. Second, the decision by the Fed to lend to a non-bank financial firm such as Bear Stearns was reckoned as extreme compared to measures enacted in past financial crises besides the Great depression60. The degree of development of the US financial system might suggest that such an upshot of the crisis is unique. In fact, other central banks are legally empowered to take a similar decision: statements in this sense can be found in or inferred from the Statutes / Acts of the central banks of Canada61, Norway62, Sweden63 and Changes to Federal Reserve Liquidity Provision”, Federal Reserve Bank of New York, http://www.newyorkfed.org/markets/Understanding_Fed_Lending.html). 58 “In unusual and exigent circumstances, the Board of Governors of the Federal Reserve System, by the affirmative vote of not less than five members, may authorize any Federal reserve bank .. to discount for any individual, partnership, or corporation, notes, drafts, and bills of exchange when such notes, drafts, and bills of exchange are indorsed or otherwise secured to the satisfaction of the Federal Reserve bank..”. 59 The loan took the name of “Tranche A Loan facility” to differentiate from a further financing to the pur- chase, the “Tranche B Loan Facility”, provided directly by JPMorgan in an amount equal to 1 billion, http://www.newyorkfed.org/newsevents/speeches/2008/Contract.pdf. The commitment to the loan by the Fed was “in principle”, since it was clarified that the credit would have been actually extended only at the time of the merger. 60 In an interview to Reuters, the former Federal Reserve Chairman Volcker stated: “We’ve seen the Federal Reserve take more extreme measures in some respects that any that have been taken in the past to deal with a financial crisis, which raises real questions not only for the Federal Reserve and its authori- ties, but for the structure of the financial system..”. The Federal Reserve acted under the authority grant- ed by Section 13(3) of its Act, a clause added in 1932 in the midst of depression era. After WWII, the clause was invoked during the 1960s although, based on available information, it cannot be stated that credit was actually granted. On the occasion of the crisis of Long Term Capital Management (LTCM) in 1998, the agreement which prevented this fund from going bankrupt by 14 banks was largely a result of an initiative of the Federal Reserve, which however it did not contribute any funds on the event. 61 Bank of Canada Act, Art. 18 (g): “if the Governor is of the opinion that there is a severe or unusual stress on a financial market or financial system, buy and sell any other securities, treasury bills, obliga- tions, bills of exchange or promissory notes, to the extent determined necessary by the Governor for the purpose of promoting the stability of the Canadian financial system” and (h) “[The Bank may] make loans or advances for periods not exceeding six months to members of the Canadian Payments Association on taking security in any property that the institution to which the loan or advance is made is authorized to hold”. A broad range of institutions is eligible to become member of the Association, including life insur- ance companies and securities dealers, http://www.cdnpay.ca/membership/eligibility.asp. 180 Studi e Note di Economia, Anno XIV, n. 2-2009 Switzerland64. By the same token, one cannot rule out altogether the option of emergency refinancing to a nonbank also where (nearly) no official state- ment on ELA is available on ELA (see section IV). Third, granting ELA implies for the central banking taking on a financial risk and the Fed took much care in explaining publicly all the measures taken to mitigate, to the extent possible, this risk65. Fourth, most decisions will be taken under extreme time pressure and almost unavoidably be clouded by uncertainty66. A sound governance of the decision procedure may at least avoid wasting time in procedural elements. VI. Concluding remarks This paper has examined a number of issues relating to Emergency Liquidity Assistance (ELA). ELA is about liquidity management by the central bank at a micro level, where the cash injection is directed to a firm (typically but not necessarily a bank) which is illiquid and which otherwise could default. Thus, by means of an ELA the central bank “corrects” the working of mar- ket forces which, at times, do not channel an adequate aggregate liquidity supply to needy (and otherwise generally worthy) market participants. When in March 2008 the Fed enters into a loan with Bear Stearns, it is exactly that firm that should receive the funds. 62 Norges Bank Act, Section 22: “In special circumstances the Bank may make loans and extend other forms of credit to entities other than banks in the financial sector ...” 63 Sveriges Riksbank Act, Chapter 6, Article 8: “In exceptional circumstances, the Riksbank may, with the aim of supporting liquidity, grant credits or provide guarantees on special terms to banking institutions and Swedish companies subject to the supervision of the Financial Supervisory Authority.” 64 Federal Act on the Swiss National Bank (National Bank Act), Art. 9 (1): “In performing its monetary tasks pursuant to article 5, paragraphs 1 and 2 [which entrusts the Bank i.a. with a duty to contribute to the stability of the financial system], the National Bank may .. e. enter into credit transactions with banks and other financial market participants on condition that sufficient collateral is provided for the loans”. Note however that Swiss National Bank (2006) refers to banks only in this chapter on ELA, perhaps sug- gesting that the provision of ELA to an “other financial market participant” is regarded only a remote pos- sibility. 65 “The portfolio supporting the credit extensions consists largely of mortgage related assets. In particu- lar, it includes cash assets as well as related hedges. The cash assets consist of investment grade securities (i.e. securities rated BBB- or higher [..]) and residential or commercial mortgage loans classified as “per- forming”. [..] All securities are domiciled and issued in the U.S. and denominated in U.S. dollars. The port- folio consists of collateralized mortgage obligations (CMOs), the majority of which are obligations of gov- ernment-sponsored entities (GSEs), such as the Federal Home Loan Mortgage Corporation (“Freddie Mac”), as well as asset-backed securities, adjustable-rate mortgages, commercial mortgage-backed secu- rities, non-GSE CMOs, collateralized bond obligations, and various other loan obligations. The assets were reviewed by the Federal Reserve and its advisor, BlackRock Financial Management. The assets were not individually selected by JPMorgan Chase or Bear Stearns.” New York Fed, http://www.newyork- fed.org/newsevents/speeches/2008/AnnexII.html. 66 In his testimony before the US Senate (see fn. 55) T. F. Geithner states that on the evening of March 13th he was informed by the SEC that Bear Stearns was about to file for bankruptcy, while it was at a sub- sequent teleconference held at 5 am of March 14th that the first corrective measures were agreed. M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 181 The paper has devoted a substantial attention to communication aspects. Three items of this communication have been considered: the ex ante com- munication channelled by central banks through their Statutes / Acts as well as Governors’ speeches; the less legalistic (though still ex ante and fairly for- mal) communication provided jointly by central banks and other authorities through a number of memoranda of understanding on crisis management; the communication issued on the event of actual crises. The central bank may prefer to be (almost) silent on the instrument of ELA, notably ex ante. By doing so, it would try to enhance uncertainty on the actual use of the instrument at next crisis and on the attached strings. The down side of this option is that by not speaking up before, and thus also not pre-announcing any limit in the scope for ELA, on the event of a crisis the central bank could eventually be pressured to be more lenient on the use of the instrument that it had actually envisaged. Moreover, if there is a recent record of banking crises, as e.g. in Sweden in the early 1990s, there is no clear advantage in denying what is absolutely evident, i.e. that the central bank would consider supporting the embattled bank(s). The alternative option of a more transparent communication is not plain sailing too, because of what could be defined as a double line of defence. To strike a balance between openness and control for the moral hazard, the cen- tral bank tends initially to be fairly strict in its statements, by announcing e.g. that it will support only banks which are illiquid but solvent and able to post adequate collateral. At the same time, reckoning that effective crisis manage- ment unavoidably calls for flexibility and pragmatism, this central bank would also add that, as an exception to the exception, it could go beyond this set of conditions (say, by not being too pushy on the collateral or granting the credit also to non-bank financial intermediaries). Obviously, the market could test the credibility of the central bank’s announcement to stop at the first limit. The paper has then touched on the conditions to be applied on ELA. Here, the work by Bagehot – who in turn followed on that of Thornton – is usual- ly summarised in the sentence that emergency credit by the central bank should be granted (i) at a high rate of interest, (ii) on good banking securities and (iii) freely. Alas, this author seems to be victim of his own success to the point that his sentences seems to be taken without retaining all nuances of his text or are read out of context. In any case, Bagehot was not describing a generic central bank but was referring more specifically to the banking department of the Bank of England as operating in the second half of the XIX century. As to (i), high is not penalty. The interest rate should be high enough so as to offer banks an incentive to seek for more ordinary ways to tap their demand of liquidity (“no one may borrow out of idle precaution without pay- ing well for it”, in Bagehot words). However, if the central bank does not 182 Studi e Note di Economia, Anno XIV, n. 2-2009 reckon such incentive to be important in the specific conditions of a crisis – it may even conclude that it would send a signal of stinginess in extending the support –, setting the interest rate on ELA above market levels may not be the right policy. As to (ii), the then Bank of England banking department was in many respects a private bank. Hence, the need to protect its share- holders through adequate backing to the credit. In today’s world, the taxpay- er, who represents the ultimate shareholder of a central bank abstracting from the details of the legal status of this institution in each country, wants to be reassured that all efforts have been made to reduce the risks associated with using public money. In practice, a broader balance of all pros and cons of the deal should be considered, where the availability of collateral is one of the many elements under consideration. Finally, as regards (iii), there is an important psychological element insofar the central bank should not be per- ceived as hesitant or faltering in providing the emergency credit; the free lend point is not bound to necessarily imply substantial increases in the liquidity supply. The paper ends reviewing some recent financial crises which differ in their underlying causes, triggering events as well as their management, dura- tion and upshot. If anything, this is a healthy reminder that hardly any crisis fully matches a previous one. Based on this review, the central bank may play multiple roles in the crisis management, from being a centre for macroeco- nomic and financial analyses, to lender of last resort, broker of private efforts, coordinating actor to other public institutions. Moreover, while some finan- cial support by the public authorities may generally be expected when a financial crisis breaks out, one ought not to take for granted that such readi- ness is tantamount to insulating the shareholders and the management of the troubled bank(s) (and possibly also other stakeholders too) from bearing sub- stantial losses. The fate of Barings and the C-tier banks in Sweden make this threat credible. Finally, once more, communication is key. If “liquidity exists when investors are confident in their ability to transact and where risks are quan- tifiable” (Warsch, 2007), then in order to restore market liquidity the central bank ought to restore investor’s confidence first. The example set by the Fed on 11/9 stands out. This is because, in Bagehot’s words, “To lend a great deal, and yet not give the public confidence that you will lend sufficiently and effectually, is the worst of all policies”. It is noticeable that both sentences I am quoting here use the term confiden(t/ce) as a pivot towards the resump- tion of ordinary conditions. M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 183 ANNEX Some notes on the Scottish and Canadian experiences of free banking Enquiries on the Scottish experience often end up with positive notes, to the point that this is regarded as a yardstick of how an unregulated banking system could work. However, a number of deviations from laissez-faire call into question the appropri- ateness of defining this experience as a fully-fledged free banking system. Following Cowen and Kroszner (1989), three deviations may be considered as especially rele- vant. First, the Act of 1765 outlawed the issuance of notes under £1, a measure which reduced competition, notably putting at a disadvantage smaller non-bank providers of financial services. Second, only three banking firms were chartered corporations (Bank of Scotland, Royal Bank of Scotland and British Linen Company) and thus enjoyed limitation of liability. Again, this dented the level-playing-field in the mar- ket since all other banks found it harder to raise capital because of their limited lia- bility status. Third, the Bank of England effectively acted as a shadow central bank towards the Scottish banking system in difficult times, such as in 1792-1793 and again in 1830 (in this respect too there is an element of inequality since only the three chartered banks depended directly upon the Bank as a lender of last resort). The Canadian experience in the nineteenth century and first three decades of the twentieth century without a central bank turned out to be much less prone to finan- cial crises than its big southern neighbour, notably it avoided the problem of season- al liquidity crises. This partly owed to the stance taken by the Canadian Bankers Association which acted as an effective self-policing agency. The Canadian nation- wide branch system likely helped too, since it diluted regional problems. In any case, two lines of defence were available. Firstly, Canadian banks kept reserves on call in the New York money market; secondly, when in 1907 and 1914 such reserves did not suffice, the Government of Canada stepped in to provide additional funds (Bordo, 1989). Hence, one could conclude that the system worked with external sources of emergency funding. At the same time, a measure of the positive record of this expe- rience is the reading of the emergence of the Bank of Canada in 1935 as owing more to political factors than to dissatisfaction with the performance of the system itself (Bordo and Redish, 1987). 184 Studi e Note di Economia, Anno XIV, n. 2-2009 REFERENCES M. ANDERSSON - S. VIOTTI (1999), Managing and Preventing Financial Crises – Lessons from the Swedish Experience, Sveriges Riksbank Quarterly Review, 1, 71-89. W. BAGEHOT (1873), Lombard Street. In N. St John-Stevas (ed.), The collected works of Walter Bagehot, volume nine, The Economist, London, 1978. An elec- tronic version of the book is available at http://socserv.mcmaster.ca/econ/ ugcm/3ll3/bagehot/lombard.html. BANK OF CANADA (2004), Bank of Canada Lender-of-Last-Resort Policies, Bank of Canada Financial System Review, December, pp. 49-55. BANK OF ENGLAND (1995a), Report of the Board of banking Supervision Inquiry into the circumstances of the collapse of Barings, Bank of England, London. BANK OF ENGLAND (1995b), Quarterly Bulletin, May. B. B. BERNANKE (2008), Liquidity provision by the Federal Reserve, speech at the Federal Reserve Bank of Atlanta Financial Markets Conference, 13 May. M. D. BORDO (1989), The lender of last resort: some historical insights, NBER Working Paper, 3011. M. D. BORDO - A. REDISH (1987), Why Did the Bank of Canada Emerge in 1935?, The Journal of Economic History, 47:2, pp. 405-417. W. H. BUITER (2008), Can central banks go broke? CEPR discussion paper series, no. 6827. T. COWEN - R. KROSZNER (1989), Scottish Banking Before 1845: A Model for Laissez-Faire, Journal of Money, Credit and Banking, 21:2, pp. 221-231. K. DOWD (1988), Private Money: The Path to Monetary Stability, Institute of Economic Affairs Hobart Paper, 112. C. DZIOBEK and C. PAZARBASIOGOLU (1997), Lessons from Systemic Bank Restructuring: A Survey of 24 Countries. IMF Working Paper, 1997/161. P. ENGLUND (1999), The Swedish Banking Crisis: Roots and Consequences, Oxford Review of Economic Policy, 15:3, pp. 80-97. FEDERAL RESERVE BANK OF NEW YORK (2001), press release issued on September 11. FEDERAL RESERVE BANK OF NEW YORK (2002), Domestic open market oper- ations during 2001. FEDERAL RESERVE BOARD (2001), press releases issued on September 11, 13, 14 and 17. FINANCIAL TIMES (2006), Bawag is rescued by bail-out and Refco deal, May 3. X. FREIXAS, C. GIANNINI and G. HOGGARTH - F. SOUSSA (2000), Lender of Last Resort: What Have We Learned Since Bagehot?, Journal of Financial M. Manna - Emergency Liquidity Assistance at work: both words and deeds matter 185 Services Research, 18:1, pp. 63-84. X. FREIXAS, J.C. ROCHET and B. M. PARIGI (2004), The Lender of Last Resort: A Twenty-First Century Approach, Journal of the European Economic Association, 2:6, pp. 1085-1115. C. GIANNINI (1999), Enemy of None but a Common Friend of All? An international Perspective on the Lender-of-Last-Resort Function, Princeton University, Essays in International Finance, 214. M. GOODFRIEND, R.G. KING (1988), Financial Deregulation, Monetary Policy and Central Banking, Federal Reserve Bank of Richmond Economic Review, 74:3. C. A. E. GOODHART (1999), Myths about the Lender of Last Resort, International Finance, 2:3, pp. 339-360. C. A. E. GOODHART, D. SCHOENMAKER (1995), Should the Functions of Monetary Policy and Bank Supervision Be Separated?, Oxford Economic Papers, 47:4, pp. 539-560. C. A. E. GOODHART, D. SCHOENMAKER (2006), Burden sharing in a banking crisis in Europe, LSE Financial Markets Group Special Paper, 164. S. INGVES, G. LIND (1996), The management of the bank crisis - in retrospect, Sveriges Riksbank Quarterly Review, 1. INTERNATIONAL HERALD TRIBUNE (1995), a series of articles by E. Ipsen appeared on the issues of 27 February, 28 February and 3 March. O. JEANNE, C. WYPLOSZ (2001), The international lender of last resort: how large is large enough?, NBER Working Paper, 8381. G.C. KAUFMAN (1996), Comment on Financial Crises, Payment System Problems, and Discount Window Lending, Journal of Money, Credit and Banking, 28:4, pp. 825-831. C. KINDLEBERGER (1978), Manias, Panics and Crashes: A History of Financial Crises, Basic Books, New York. J. LACKER (2004), Payment System Disruptions and the Federal Reserve Following September 11, 2001, Journal of Monetary Economics, 51:5, pp. 935-965. A. MARTIN (2005), Reconciling Bagehot with the Fed’s Response to September 11, Federal Reserve Bank of New York Staff Report, 217. J. MCANDREWS, S. POTTER (2002), Liquidity effects of the events of September 11, 2001, Federal Reserve Bank of New York Economic Policy Review, 8:2, pp. 59-79. N.W. LEESON and E. WHITLEY (1996), Rogue trader: how I brought down Barings Bank and shook the financial world, Little, Brown, Boston. NORGES BANK (2005), Norges Bank’s role in the event of liquidity crises in the financial sector, Norges Bank Economic Bulletin, Q2, pp. 80-89. ÖSTERREICHISCHE NATIONALBANK (2006a), Financial Stability Report, 11, June. ÖSTERREICHISCHE NATIONALBANK (2006b), OENB Liebscher: BAWAG sol- ventes Unternehmen, 28 April. ÖSTERREICHISCHE NATIONALBANK (2006c), Österreichische Finanzwirtshaft berät in OeNB BAWAG-P.S.K.-Lösung, 2 May. ÖSTERREICHISCHE NATIONALBANK (2007), Annual Report 2006. RESERVE BANK OF AUSTRIALIA (1995), Implications of the Barings Collapse 186 Studi e Note di Economia, Anno XIV, n. 2-2009 for Bank Supervisors. Bulletin, November. REUTERS (2006), article by H. Prammer and B. Groendahl, released at 12:22, 2 May. J.-C. ROCHET, X. VIVES (2004), Coordination Failures and the Lender of Last Resort: Was Bagehot Right After All?, Journal of the European Economic Association, 2:6, pp. 1116-1147. M. SORGE (2004), Stress-testing financial systems: an overview of current method- ologies, BIS Working Paper, 165. P. STELLA (2002), Central Bank Financial Strength, Transparency, and Policy Credibility, IMF Working Paper, 02/137. G. STEVENS (2008), Liquidity and the lender of last resort, speech at the Seventh Annual Sir Leslie Melville Lecture, 15 April. SVERIGES RIKSBANK (2003). The Riksbank’s role as lender of last resort, Financial Stability Report 2/2003, pp. 57-73. SWISS NATIONAL BANK (2006), Guidelines of the Swiss National Bank (SNB) on Monetary Policy Instruments (as at 29 June 2006), Zurich. H. THORNTON (1802), An Enquiry into the Nature and Effects of the Paper Credit of Great Britain, Hatchard, London. K. M. WARSCH (2007), Market liquidity - definitions and implications, Speech offered at the Institute of International Bankers Annual Washington Conference, Washington DC, 5 March. ADDITIONAL RELEVANT OFFICIAL DOCUMENTS AUSTRIA - The Federal Act on the Oesterreichische Nationalbank (1984). BELGIUM - Organic Act DD 22 February 1998 of the National Bank of Belgium; Statutes of the National Bank of Belgium. FINLAND - Act on the Bank of Finland (214/1998). IRELAND - Central Bank and Financial Services Authority of Ireland Act 2003. ITALY – Statute (2006); IMF Financial Sector Assessment Program (2006). JAPAN - The Bank of Japan Law (87/2005). NORWAY - The Norges Bank Act (28/1995). PORTUGAL - Banco de Portugal, Organic Law (2007). SPAIN - Law of Autonomy of the Banco de España (13/1994). SWEDEN - The Sveriges Riksbank Act. SWITZERLAND – National Bank Act 3/10/2003); Guidelines of the Swiss National Bank (SNB) on Monetary Policy Instruments. THE UNITED KINGDOM - Memorandum of Understanding between HM Treasury, the Bank of England and the Financial Services Authority. UNITED STATES OF AMERICA - Federal Reserve Act. These documents are available at the website of the corresponding central bank (among those not mentioned in the list, also the websites of the central banks of France, Germany, Greece, Luxembourg and the Netherlands have been consulted). The Central Bank and Financial Services Authority of Ireland Act is available at http://www.irishstatutebook.ie/2003/en/act/pub/0012/index.html.