Martin Hellwig

Financial Stability and Monetary Policy in a Time of Crisis

Columbia Law School, March 7, 2016

Abstract

The talk discusses the role of financial stability in monetary policy in the financial crisis of

2007-2009 and the European crisis since 2010. Major concerns are (i) the

microeconomic/financial underpinnings of the monetary system and the role of the central

bank as a lender of the last resort, (ii) the role of financial stability as an objective for

monetary policy, (iii) financial dominance and hidden fiscal dominance as threats to the

independence of monetary policy. The conceptual issues are illustrated by an account of the

European experiences and discussions, including the move to banking union and the legal

discussion about the European Central Bank’s purchases of government bonds.

IN THE GENERAL COURT OF THE EUROPEAN UNION

9 July 2015

APPLICATION FOR INITIATING PROCEEDINGS

_______________________

BETWEEN

Alcimos Consulting SMPC a company established under the laws of Greece, with its registered office at Vournazou 14, 11521 Athens, Greece (“Alcimos”)

AND

The European Central Bank having its address at Sonnemannstraße 20, 60314 Frankfurt am Main, Germany (“ECB”)

_______________________

Representative (name and status): Faidra Rodolaki, Attorney-at-Law Address for service: Leoforos Alexandras 207, GR-115 23 Athens, Greece Service to be effected to Ms. Faidra Rodolaki by telefax: +30 211 800 8380 or by email: frodolaki@alcimos.com

_______________________

1. The contested decisions

The Governing Council of the ECB decided on 28 June 2015 “to maintain the ceiling to the provision of emergency liquidity assistance (ELA) to Greek banks at the level decided on […] 26 June 2015” (https://www.ecb.europa.eu/press/pr/date/2015/html/pr150628.en.html). In a further decision on, 6 July 2015, the ECB governing council decided “to maintain the provision of emergency liquidity assistance (ELA) to Greek banks at the level decided on 26 June 2015 after discussing a proposal from the Bank of Greece” and “to adjust the haircuts on collateral accepted by the Bank of Greece for ELA”. (https://www.ecb.europa.eu/press/pr/date/2015/html/pr150706.en.html). The above decisions (the “Contested Decisions”), inescapably resulted in the imposition of a bank holiday and capital controls in Greece, by virtue of an “act of legislative content” published in the Government Gazette of the Hellenic Republic on 28 June 20151, as there was no other way to satisfy the demands for withdrawal of deposits from Greek banks.

1 Fascicule A, Issue No. 65; as amended by a further act dated 30 June 2015 (Issue No. 66) and extended and

amended by decisions of the Greek Minister of Finance dated 6 and 8 July 2015 (Govt Gaz. Fascicule B, Issues

Nos. 1391 and 1420).

2

2. The damage sustained

2.1 Alcimos holds a bank account with a Greek bank, Eurobank Ergasias S.A., with a balance of €10,582.84 on the day the capital controls were imposed. The capital controls and the bank holiday resulted in Alcimos not being able to freely use the funds available for transactions outside Greece. 2.2 In addition, Alcimos has been acting as a sponsor for a number of ventures which are inextricably linked to the Greek economy, the Greek financial services system and their prospects. In particular, Alcimos is a sponsor for an application submitted to the UK Listing Authority on 17 February 2015 for the approval of a prospectus pertaining to an initial public offering of up to €500m for an investment vehicle which would invest in Greek residential property. Furthermore, Alcimos is in the process of establishing a Greek lending platform for mortgages, auto finance and factoring. Lastly, Alcimos is in advanced discussions for the acquisition of a number of receivables portfolios.

2.3 All these transactions have not only been put on hold by the imposition of the capital controls and bank holiday, but also run the risk of being cancelled, as investors lose confidence in the Greek financial system and the Greek economy, which is also a result of the imposition of said measures. This would cause serious and irreparable damage to Alcimos. In fact, the viability of Alcimos is currently in jeopardy and it is doubtful that the company will remain a going concern, purely as a result of the imposition of the bank holiday and capital controls.

3. Legal Standing

3.1 Article 263 TFEU grants the Court of Justice of the European Union jurisdiction to review the legality of acts of the EU institutions, including the ECB, that are intended to produce legal effects vis-à-vis third parties. The ECB decisions of 28 June 2015 and 6 July 2015 rejecting the request of the Bank of Greece to increase ELA for Greek banks are binding acts and are therefore susceptible to review. 3.2 Further, according to Article 263 TFEU “any natural or legal person may […] institute proceedings against an act addressed to that person or which is of direct and individual concern to them, and against a regulatory act which is of direct concern to them and does not entail implementing measures”. 3.3 It should firstly be noted that the Contested Decisions are of direct concern to Alcimos, since an unbroken chain of causation exists between the act and the loss or damage suffered. Clearly, the Contested Decisions left the Hellenic Republic no discretion in implementation, as the imposition of a bank holiday and capital controls was a direct and inescapable consequence of the contested acts. Even if one argues that the Hellenic Republic could choose whether or not to impose capital controls on the Greek banks, the possibility that they would not do so is “purely theoretical”, as established in Piraiki-Patraiki v Commission (case 11/82). 3.4 Furthermore, it can be argued that Alcimos is also individually concerned. The Court of Justice set out its well-known test for individual concern in Plaumann & Co v Commission (Case 25/62). Persons other than those to whom a decision is addressed are individually concerned only if a decision affects them “by reason of certain attributes which are peculiar to them or by reason of circumstances in which they are differentiated from all other persons and by virtue of these factors distinguishes them individually just as in the case of the person addressed”. Applicants may also claim to be singled out if they can show that they are a member of a closed class of applicants that was fixed and ascertainable at the date the measure was adopted (International Fruit Co NV v Commission, Cases 41-44/70; Piraiki-Patraiki). Alcimos, by virtue of holding a deposit with a bank

3

operating in Greece is one out of the fixed and ascertainable closed class of depositors with banks operating in Greece which were affected by the imposition of the bank holiday and capital controls. 3.5 Furthermore, it can be argued that the Contested Decisions are regulatory acts; in this case, simply establishing direct concern suffices, assuming the act does not entail implementing measures. Indeed, the Contested Decisions did not entail any implementing measures and, even if they did, there was certainly no discretion on the part of the Hellenic Republic. 3.6 Lastly, it should be mentioned that in case the Court denied standing to Alcimos, this would result in Alcimos being denied the protections afforded by Article 47 of the Charter of Fundamental Rights of the European Union.

4. Legal basis for dispensing ELA

4.1 According to para. 14.4 of the Statute of the European System of Central Banks (ESCB) and of the European Central Bank (OJ [2010] C326/230) “National central banks may perform functions other than those specified in this Statute unless the Governing Council finds, by a majority of two thirds of the votes cast, that these interfere with the objectives and tasks of the ESCB. Such functions shall be performed on the responsibility and liability of national central banks and shall not be regarded as being part of the functions of the ESCB”. 4.2 Furthermore, on 17 October 2013 the ECB published on its website “the procedures underlying the Governing Council’s role pursuant to Article 14.4 of the Statute of the European System of Central Banks and of the European Central Bank with regard to the provision of ELA to individual credit institutions”. 4.3 According to said procedures “ELA means the provision by a Eurosystem national central bank (NCB) of (a) central bank money and/or (b) any other assistance that may lead to an increase in central bank money to a solvent financial institution, or group of solvent financial institutions, that is facing temporary liquidity problems, without such operation being part of the single monetary policy. Responsibility for the provision of ELA lies with the NCB(s) concerned. This means that any costs of, and the risks arising from, the provision of ELA are incurred by the relevant NCB”.

5. The ECB’s position regarding the provision of ELA

5.1 As mentioned in the Monthly Bulletin for the 10th anniversary of the ECB (http://www.ecb.europa.eu/pub/pdf/other/10thanniversaryoftheecbmb200806en.pdf) “the NCBs may provide – temporarily and against adequate collateral – emergency liquidity assistance (ELA) to illiquid but solvent credit institutions. The possible provision of ELA is undertaken at the discretion of the competent NCB, subject to the conditions set out in the Treaty relating to the prohibition of monetary financing, and only in exceptional circumstances”. 5.2 This is line with what the ECB had mentioned in its 1999 Annual Report (http://www.ecb.europa.eu/pub/pdf/annrep/ar1999en.pdf): “The main guiding principle is that the competent NCB takes the decision concerning the provision of ELA to an institution operating in its jurisdiction. This would take place under the responsibility and at the cost of the NCB in question. Mechanisms ensuring an adequate flow of information are in place in order that any potential liquidity impact can be managed in a manner consistent with the maintenance of the appropriate single monetary policy stance”.

4

5.3 In the Opinion of the ECB on 24 January 2012 on a guarantee scheme for the liabilities of Italian banks and on the exchange of lira banknotes (CON/2012/4 https://www.ecb.europa.eu/ecb/legal/pdf/en_con_2012_4_f.pdf) “[t]he ECB notes that emergency liquidity assistance, granted by the national central bank (NCB) independently and at its full discretion to a solvent credit institution on the basis of a collateral security […] is in principle possible, provided that a number of conditions are met in order to ensure the NCB’s compliance with the monetary financing prohibition under Article 123 of the Treaty”. 5.4 In a letter by ECB Chairman Mr. Mario Draghi addressed to Mr. Andreas Pitsillides, Member of the European Parliament, dated 28 January 2014 (https://www.ecb.europa.eu/pub/pdf/other/20140128_pitsillides.en.pdf) Mr. Draghi states: “Emergency liquidity assistance (ELA) operations are undertaken by national central banks under national responsibility. However, in order to prevent these operations from interfering with the tasks and objectives of the Eurosystem – notably, the implementation of the single monetary policy – the Governing Council of the ECB has established rules and procedures with regard to the provision of ELA to individual credit institutions. These rules and procedures are available on the ECB’s website and provide answers to some of the questions you raised. ELA is a specific tool available to central banks in crisis situations. Its aim is to provide liquidity support, in exceptional circumstances, to temporarily illiquid but solvent credit institutions which cannot obtain sufficient liquidity through the market and/or their participation in regular monetary policy operations.” 5.5 In a letter by ECB Chairman Mr. Mario Draghi addressed to Mr. Dimitrios Papadimoulis, Member of the European Parliament, dated 7 May 2015 (https://www.ecb.europa.eu/pub/pdf/other/150506letter_papadomoulis_2.en.pdf) Mr. Draghi states: “The rules of the Eurosystem governing the provision of liquidity to the banking system are intended to ensure the singleness of monetary policy in the euro area”. 5.6 In a letter by ECB Chairman Mr. Mario Draghi addressed to Mr. Sven Giegold, Member of the European Parliament, dated 17 June 2015 (https://www.ecb.europa.eu/pub/pdf/other/150618letter_giegold.en.pdf?5c28b50385f872645ddb5b18eb2cd14a), Mr. Draghi states:

“Responsibility for the provision of emergency liquidity assistance (ELA) to Cypriot banks lies with the Central Bank of Cyprus. The role of the Governing Council of the ECB as regards the provision of ELA is to ensure that such operations do not interfere with the integrity of monetary policy within the Eurosystem, including the primary objective of maintaining price stability. Furthermore, ELA should not interfere with the prohibition on monetary financing.

The ECB is a rule-based institution bound by the EU Treaties, which require the Eurosystem to lend only to solvent banks, against adequate collateral, and, as mentioned above, to refrain from financing governments. Any direct or indirect financing of a government is incompatible with the prohibition on monetary financing enshrined in Article 123 of the Treaty on the Functioning of the European Union. Let me clarify that provision of ELA by a national central bank is aimed at supporting solvent banks facing liquidity problems, rather than providing solvency support. Therefore, a key requirement in this context is that recipient credit institutions remain solvent. Any decision (non-objection) by the Governing Council related to the provision of ELA depends on the assessment of the conditions of the recipient credit institution”.

5.7 The ECB, therefore accepts that the only constraints to a NCB’s decision to dispense ELA are any interference with the implementation of the single monetary policy and, possibly, the prohibition on monetary financing enshrined in Article 123 TFEU. This is in line with the principal of conferral, as per Articles 4 and 5 TEU.

5

5.8 It is true that, recently, the ECB changed its stance on the subject: in an ECB publication dated July 2015 titled “The financial risk management of the Eurosystem’s monetary policy operations” (https://www.ecb.europa.eu/pub/pdf/other/financial_risk_management_of_eurosystem_monetary_policy_operations_201507.en.pdf) it is mentioned: “[t]he objective of ELA is to support solvent credit institutions that are facing temporary liquidity problems. ELA thus addresses short-term liquidity problems and does not aim to provide solvency support. ELA takes the form of central bank money and/or any other assistance that may lead to an increase in central bank money. ELA needs to be distinguished from the Eurosystem’s credit operations, which are designed to implement the monetary policy of the Eurosystem and with which ELA should not conflict. ELA should not conflict with the objectives and tasks of the ESCB. Interference with the objectives and tasks of the ESCB could, for instance, result from the following: (i) a threat to the singleness of monetary policy, (ii) a threat to the implementation of monetary policy, for example by making the steering of short-term rates more difficult, (iii) a threat to the financial independence of the NCB, for instance if ELA was not provided against sufficient collateral to safeguard such independence, (iv) an obvious concern about a possible breach of the monetary financing prohibition, or (v) provision of ELA at overly generous conditions, which, in turn, could increase the risk of moral hazard on the side of financial institutions or responsible authorities”. 5.9 The fact that the ECB decided to change (or, at least, publicize the change in) its position post factum, i.e., after the rejection of the request by the Greek NCB, is itself suspect. As we shall see, even assuming these new rules were intra vires, their application should not have prevented the ECB from agreeing to the request of the Greek NCB.

6. The ECB decisions of June 28 and July 2015

6.1 In the Contested Decisions, the ECB decided to leave the levels of ELA unchanged, despite requests by the Greek NCB. According to press reports, (http://www.reuters.com/article/2015/06/29/us-ecb-greece-idUSKCN0P91N420150629) the Greek NCB has submitted a request for an increase of ELA by €6bn, from a level previously of €89bn. 6.2 As discussed previously, there are only two possible grounds for rejecting the request: possible interference with the implementation of the single monetary policy or the monetary financing prohibition under Article 123 TFEU. 6.3 Let us start by examining the latter. As Mr. Draghi mentioned in a press conference in Nicosia on 5 March 2015 (https://www.ecb.europa.eu/press/pressconf/2015/html/is150305.en.html): “The ECB is a rule-based institution. It’s not a political institution. One of the rules that we comply with is contained in the Treaty, and it’s Article 123, and it’s the prohibition of monetary financing. Monetary financing is when the central bank of a country prints money to buy the government bonds in the primary market of that country, and it could be either direct or indirect, when banks bring collateral to the ECB in order to be financed in order to buy the sovereign debt of that country, and we are prohibited from doing that”. 6.4 However, according to press reports (http://www.bloomberg.com/news/articles/2015-03-24/ecb-said-to-limit-greek-lenders-treasury-bill-holdings) the ECB had already put a limit on purchases of T-bills by Greek banks as early as 24 March 2015. Since the Hellenic Republic is only issuing T-bills, there was no way for Greek banks to use liquidity provided through ELA to finance the Greek State, and thus no breach of the prohibition on monetary financing under Article 123 TFEU could have taken place.

6

6.5 Could, however, the request by the Greek NCB, if accepted by the ECB, have interfered with the implementation of the single monetary policy? First of all, it cannot be convincingly argued that an extension of an additional €6bn of ELA to Greek banks could have interfered with the implementation of the single monetary policy, as this amount is a tiny fraction of the Eurozone money supply, currently standing at approximately €10.5 trillion. 6.6 But let’s assume that even this tiny amount was high enough to influence the implementation of the single monetary policy. Would, in fact, the acceptance by the ECB of the request of the Greek NCB have interfered with it? 6.7 This would have been the case, if the NCB were extending credit against insufficient collateral: if the recipient of ELA was not able to repay the Greek NCB, the default would, indeed, result in the Greek NCB having created new money, thus influencing the implementation of the single monetary policy. 6.8 It is undisputable that the request to increase ELA was in response to a substantial increase in deposit outflows from Greece, as a result of the political instability. Any additional funding by the Greek NCB to Greek banks would be highly unlikely to be diverted to asset generation, as assets of Greek banks have been contracting steadily since February 2011 (according to data published by the Bank of Greece). 6.9 The mere fact, therefore, that one type of liability (ELA) was substituted for another type of liability (deposits) can have no effect on the solvency of Greek banks. The ultimate arbiter of the solvency of the four large major banks (which account for more than 90% of the total assets of the Greek banking system) is the Governing Council of the ECB, to which the Single Supervisory Mechanism reports. 6.10 Either Greek banks were, in the opinion of the Governing Council of the ECB, solvent at the time of the requests by the Greek NCB, or not. If they were indeed solvent, then the substitution of one type of liability for the other could not have affected their solvency. Greek banks would still be able to repay ELA; thus no new money would be created and the implementation of the single monetary policy would have not been affected. 6.11 If Greek banks were insolvent at the time, then it would be the duty of the ECB to exercise its supervisory powers, as conferred upon it by Article 16 of Council Regulation (EU) No 1024/2013. In this case, it is evident from statements made by senior ECB officers that Greek banks were solvent and, as such, the increase in ELA could not be lawfully refused on that ground. For instance Ms. Danièle Nouy, Chair of the Supervisory Board of the SSM, said as recently as 7 June 2015 that “[Greek] banks continue to be solvent and liquid. The Greek supervisors have done good work over the past years in order to recapitalise and restructure the financial sector. That was also visible in the outcome of our stress test. The Greek institutions have experienced difficult phases in the past. But they have never before been so well prepared for them” (https://www.bankingsupervision.europa.eu/press/interviews/date/2015/html/sn150607.en.html). It cannot, obviously, be the case that only twenty-one days later the same banks were in practice deemed to be insolvent: assuming that were the case, the ECB would have prescribed measures to rectify the situation. 6.12 It cannot, of course, be that the ECB, wearing its supervisory hat considers one institution solvent, while considering it insolvent when acting as guardian of the single monetary policy. Nevertheless this is, in fact, a situation which Ms. Nouy, considers perfectly possible, when saying (in the same interview): “Monetary policy and supervision work in strict separation. We have different

7

staff and are located in different buildings. We share access to data and work closely together in the field of financial stability. Otherwise, we only inform each other about facts of cases for which it is absolutely necessary. When it comes to monetary policy decisions such as emergency loans, it is therefore up to the ECB Governing Council to decide on which banks it classifies as solvent. We carry out our own examination independently.”

6.13 When asked further “what would [she] do if one ECB board still classified the Greek banks as solvent and the other one didn’t?” Ms. Nouy avoided the question by saying “That is a hypothetical question that I will not answer. I simply do my supervisory job and send the results to the ECB Governing Council”. 6.14 This hypothetical question is no longer hypothetical: this is exactly the situation the ECB has brought itself in this case. The Governing Council, wearing its supervisory hat, considers Greek banks solvent. Wearing a different hat, it considers them insolvent. This can only be the outcome of political influence on the ECB, which of course violates ECB’s independence, enshrined in Article 130 TFEU. Alternatively, therefore, as a result, the Contested Decisions were issued in violation of the law on that ground, too. 6.15 Let us now turn to the question of collateral. It needs to be said, that the application of collateral rules by the ECB has been inconsistent at best. Indeed, according to press reports “Belgium’s central bank accepted Fortis' branch network as collateral for an ELA advance back in 2008” (http://www.independent.ie/business/irish/state-of-emergency-some-truths-on-the-funds-keeping-our-banks-afloat-26679262.html). In an email dated 31 July 2012 Mr. Benoît Coeuré, member of the ECB’s executive board, suggests to the Governor of the Cypriot NCB as a potential measure “revising the ELA valuation and methodology, in particular for credit claims (non-tradable instruments)”. Mr. Coeuré further notes that “[t]he Central Bank of Cyprus has in principle the possibility to apply less stringent valuations and haircuts compared to the approach followed by the Eurosystem in credit operations. This should of course be done in a transparent and reasoned way, substantiating the claim that standard Eurosystem haircuts are not necessary in the case of certain types of collateral accepted under ELA”. 6.16 In a report requested by the European Parliament's Committee on Economic and Monetary Affairs titled “The ECB’s Collateral Policy and Its Future as Lender of Last Resort” and dated November 2014 (http://www.karlwhelan.com/EU-Dialogue/Whelan-November-2014.pdf) it is stated: “The rules for the provision of credit via ELA, and the conditions required for agreement from the ECB Governing Council are not at all clear. Indeed they appear to be completely ad hoc, with decisions or threats to end ELA programmes producing a number of controversies in recent years. In this important sense, the Eurosystem does not really have a comprehensive collateral policy because when the most difficult cases occur, its standard rule-book goes out the window”. The inconsistent application of collateral rules further substantiates the claim that the ECB’s independence has been violated, in breach of Article 130 TFEU. 6.17 However, it can be argued that collateral was needed for ELA operations when the ECB was not the supervisory authority for the ELA recipient institutions. Since the ECB at the time had no view on the solvency of these institutions, it needed to request adequate collateral to ensure that credit extended under ELA would indeed be repaid and thus no new money would be created. Now that the ECB (through the SSM) is the ultimate arbiter of the solvency of systemic European banks (including Eurobank Ergasias S.A. and the three other Greek banks which account for more than 90% of the assets of the Greek banking system), collateral could not be the decisive factor. Either Greek banks are solvent (in which case, by definition, even their unsecured deposits are safe) or they are

8

insolvent. If unsecured deposits are safe, then credit extended under ELA to replace said deposits should also be safe. 6.18 Lastly, let us deal with the five conditions, which according to the ECB could result in interference with the objectives and tasks of the ESCB:

(i) a threat to the singleness of monetary policy This has already been dealt with at length hereinabove.

(ii) a threat to the implementation of monetary policy, for example by making the steering of short-term rates more difficult It cannot be convincingly argued that the mere extension of ELA by €6bn would be “a threat to the implementation of monetary policy, for example by making the steering of short-term rates more difficult”, as the amounts involved are too small.

(iii) a threat to the financial independence of the NCB, for instance if ELA was not provided against sufficient collateral to safeguard such independence There is no obvious relationship between the financial independence of the NCB and the provision of ELA against inadequate collateral—unless there are suspicions that the granting of ELA was politically influenced. There are no facts substantiating such a claim, particularly in view of the fact that Greek banks remain, in the opinion of the ECB, solvent.

(iv) an obvious concern about a possible breach of the monetary financing prohibition This has been dealt with hereinabove in detail.

(v) provision of ELA at overly generous conditions, which, in turn, could increase the risk of moral hazard on the side of financial institutions or responsible authorities”

6.19 While the ECB makes no effort to define when such conditions would be “overly generous”, they can certainly not be overly generous in a situation in which not extending ELA to solvent institutions inescapably results in a bank holiday and the imposition of capital controls. 6.20 Furthermore, since the promotion of the smooth operation of payment systems per Article 127(2) TFEU is one of the four basic tasks to be carried out through the Eurosystem, it can be argued that the ECB’s decision which led to the disruption of payment systems in an entire Eurozone country fails the proportionality test. The extension of additional ELA to Greek banks with its potential minute effect on the implementation of the single monetary policy would have certainly interfered less with the tasks of the ECB. 6.21 In light of the arguments presented above, the applicant argues that ECB, in issuing the Contested Decisions, has acted in violation of para. 14.4 of the Statute of the European System of Central Banks, of Articles 4 and 5 TEU, and of Article 130 TFEU.

7. Pleas in law and main arguments in summary

7.1 Firstly, the applicant is alleging that the ECB acted in violation of para. 14.4 of the Statute of the European System of Central Banks, as the ECB’s nonconsent to the request by the Bank of Greece to increase ELA to Greek banks would not have interfered with the objectives and tasks of the ESCB. 7.2 Secondly, the applicant is alleging that the ECB acted in violation of Articles 4 and 5 TEU, as it was acting ultra vires when rejecting the request by the Bank of Greece. 7.3 Thirdly, the applicant is alleging that the ECB acted taking into account political considerations, and therefore violating Article 130 TFEU, which enshrines the independence of the ECB.

M A X P L A N C K S O C I E T Y

Preprints of theMax Planck Institute for

Research on Collective GoodsBonn 2014/12

Yes Virginia, There is a European Banking Union! But It May Not Make Your Wishes Come True

Martin Hellwig

Preprints of the Max Planck Institute for Research on Collective Goods Bonn 2014/12

Yes Virginia, There is a European Banking Union!But It May Not Make Your Wishes Come True

Martin Hellwig

May 2014, final August 2014

Max Planck Institute for Research on Collective Goods, Kurt-Schumacher-Str. 10, D-53113 Bonn http://www.coll.mpg.de

1

Yes Virginia, There is a European Banking Union! But It May Not Make Your Wishes Come True1

Martin F. Hellwig

Abstract

The paper discusses the prospects for European Banking Union as they appear in the summer

of 2014. The first part gives an overview over the problems that gave rise to the Banking Un-

ion initiative, the second part discusses the legislative measures that have been taken towards

this objective.

The euro area is currently suffering from low growth, high indebtedness of private households

and firms, banks, and governments, and the weakness of financial institutions. Weakness of

financial institutions affects the economy not only in countries with outright banking crises

and sovereign debt crises, but also in some of the core countries that so far have seemed more

stable. ECB policies have so far stabilized the system without solving the underlying prob-

lems. At the national level, political will to solve the underlying problems is missing; most

governments prefer procrastination over cleanups, some governments do not have the funds to

recapitalize the banks of their countries, and some governments like their banks to borrow

from the ECB and lend to them.

The European Banking Union comes with a promise of reducing cross-border externalities in

dealing with banks. However, the Single Supervisory Mechanism is hampered by the need to

apply national laws that implement European directives; this makes for fragmentation even if

the ECB is in charge. Moreover, procedures for the recovery and resolution of institutions in

difficulties are problematic: If banks with systemically important operations in several coun-

tries enter into resolution, there is no way to prevent the breakdown of these operations and to

limit the resulting systemic damage. Further, the legislation makes no provisions for the li-

quidity needed for maintaining systemically important operations at least temporarily. Finally,

there is no fiscal backstop. Because of the deficiencies, the “too-big-to-fail” syndrome is still

present. In view of the many legacy problems, this issue is critical. If the European Banking

Union is to work, further reforms will be needed shortly.

Key Words: European Banking Union, European Central Bank, banking supervision, bank

resolution, too-big-to-fail, sovereign debt

JEL Classifications: E58, F55, G21, G28, H63

1 Paper presented at the 42nd Economics Conference of the Austrian National Bank, Vienna May 12-13,

2014. I am grateful for very helpful comments from Christoph Engel, Paul Schempp, and Monika Ziol-kowska.

2

1. Introduction

The title of this paper alludes to an episode in 1897 when an eight-year-old girl had written a

letter to the New York Sun asking: “Is there a Santa Claus?” and that newspaper published a

full-page article under the headline “Yes, Virginia, there is a Santa Claus!” In listening to

speeches or reading documents about the European Banking Union, I sometimes get the feel-

ing that banking union is regarded as a kind of Santa Claus, which will make our wishes come

true and solve all the problems of the euro area financial system.2

As an academic, I am always impressed by the ability of people in office to make succinct

statements about problems and policies without explaining how the latter relate to the former.

The Euro Area Summit Statement of June 29, 2012 affirms “that it is imperative to break the

vicious circle between banks and sovereigns” and asks the Commission “to present proposals

… for a single supervisory mechanism” for banks, without explaining how the latter relates to

the former. Nor does it explain what precisely is meant by “the vicious circle between banks

and sovereigns”.

I am also impressed by the ability of people in office to congratulate themselves on having

come to an agreement or passed a law without worrying whether the agreement or the law

will actually work and whether the new arrangements will solve the problems they are sup-

posed to solve. The mere fact that a new arrangement has been put into place is treated as an

achievement. In terms of the political process, this assessment may be appropriate, but, if the

underlying problems are not addressed, the “achievement” may just be a way of wasting time

and exposing us to further risks. If the policy makers have got the analysis wrong, we may all

end up being the worse for it.

Right now, we are all congratulating ourselves on the steps that have been taken towards a

European banking union, the Single Supervisory Mechanism and the Single Resolution

Mechanism, together with the Banking Recovery and Resolution Directive, and previously the

Capital Requirements Directive IV and Capital Requirements Regulation, as well as the Regu-

lations establishing the European Supervisory Authorities. These are big steps forward. How-

ever, I have serious doubts whether they will substantially improve the future financial stabil-

ity in Europe. I even have doubts whether they will suffice to take us out of the straights we

are currently in. The reasons for these doubts will be laid out in this paper.

In Section 2 below, I begin with a brief overview over the problems of the euro area, the

weakness of financial institutions that underlies these problems, and the governance problems

in the euro area that have so far prevented a resolution of these problems. Subsequently, in

Section 3, I discuss the possible role of banking union in dealing with the governance of fi-

2 Full revelation: I was a co-author of ASC (2012), which can be read in this vein. However, ASC (2012)

and, subsequently, Sapir et al. (2012) are very clear about the need for a viable resolution regime; the dis-cussion of the shortcomings of banking union in this paper follows directly from the analysis in those re-ports.

3

nancial sector weakness. Major concerns are the remaining heterogeneities in supervision and

from the lack of credibility of resolution mechanisms.

2. What Are the Problems? Where Do They Come From? Why Are They Not Solved?

Fundamental Weaknesses

European economies today suffer from three interrelated weaknesses:

Economic growth is disappointingly low, not only in the periphery countries that pur-

sue austerity policies but also in the core countries of the euro area.

The levels of indebtedness of governments, nonfinancial companies and private

households are very high and in most areas still rising.

Financial institutions are weak, not only in the periphery countries but all over Europe.

These observations are reminiscent of the experience of Japan over the past two decades. In

Europe as in Japan, some of the weaknesses may be due to fundamentals such as population

aging and may therefore be unavoidable. However, some of them are also the result of flawed

policies and should be mitigated by political reform. In the euro area, the problems are exac-

erbated by the fact that the arm’s length relation between the central bank and the member

states puts limits on the authorities’ ability to deal with the banking problems effectively.

The poor growth performance is to a large extent due to the effects of overhanging debt and to

the weakness of financial institutions. Because of excessive indebtedness, governments that

were used to spending substantially more than they took in have been forced to retrench their

activities, to raise taxes, or to obtain new funding through financial repression. All this harms

economic growth. Weak financial institutions have reduced their lending, in particular to new

firms that might provide impulses for innovation and growth. Some of this retrenchment has

been a reaction to overexpansion before 2007, some of it has been imposed by financial re-

pression, and some of it reflects the banks’ own forbearance towards problem borrowers, mo-

tivated by a desire to avoid laying open the problems and taking the resulting losses on the

books.3

3 For extensive accounts of these issues, see ASC (2012), as well as Caprio and Klingebiel (1996, 1997)

and Hoshi and Kashyap (2004, 2010). On financial repression and biases in bank lending in Europe, see Acharya and Steffen (2013) as well as the chapters by Bruni, Caminal et al., and Borges in Dermine (1990).

4

Banks and Sovereigns: A Vicious Circle?

In this context, the formulation “vicious circle between banks and sovereigns” in the Euro

Area Summit Statement of June 29, 2012 is not helpful. We have seen – and continue to see –

contagion effects from sovereigns to banks in some countries and from banks to sovereigns in

others, but the picture of a doom loop between the two is more confusing than clarifying. The

so-called “euro crisis” is in fact composed of different kinds of crises reflecting different fail-

ures of governance in the relation between financial institutions and governments.4

Some countries had old-fashioned sovereign debt crises that were caused by the inability of

their politicians to set priorities and make hard choices so as to make ends meet. Examples are

given by Greece, Portugal and, to a lesser extent, Italy. As documented by Reinhart and

Rogoff (2009), this kind of crisis has a long tradition. Sovereign debt crises spill over into the

financial system if the sovereigns in question have used their power to induce “their” banks

into funding them and the sovereign’s default imposes large losses on these banks. An exam-

ple is given by Argentina in the 1990s and early 2000s. In the case of Greece, the 2012 haircut

on sovereign debt necessitated substantial ESM contributions to recapitalizing Greek banks in

order to save them from being insolvent.

Other countries had equally old-fashioned banking crises that were induced by boom-and-bust

developments in real-estate markets. Examples are given by Ireland and Spain. This kind of

crisis also has a long tradition. A little over twenty years ago, boom-and-bust developments in

real-estate markets (and in lending to nonfinancial companies) were major causes of the bank-

ing crises in Japan, the United States, the Scandinavian countries, and Switzerland.5 When

such developments occur, governments that find it necessary to support their financial institu-

tions may see their debt levels rise dramatically so that the financial crisis in turn may induce

a sovereign debt crisis. This was the experience of Ireland in 2010. Fear of such an experience

was the reason why in 2012, Spain asked for the ESM to recapitalize its banks.

Except for the case of Spain, where the impact of the financial crisis on government deficits

and debts in turn forced the government to increase its reliance on Spanish banks, there is lit-

tle that is “loopy” about these developments. The two kinds of crises that I have described

originate in quite different failures of governance. Conventional sovereign debt crises origi-

nate in failures of the political system; if these crises spill over into the financial sector, there

is not much of a spillover back to the sovereign, which probably is unable to provide a bailout

anyway. Conventional real-estate boom-and-bust and banking crises originate in failures of

risk control in banks and in failures of prudential supervision over banks; if such a financial

crisis spills over to the sovereign, a spillover back to the financial sector can occur if the ini-

tial financial crisis was localized, and the sovereign’s difficulties affect the rest of the finan-

4 For a more extensive discussion of the interplay between the different crises, see Hellwig (2011). 5 See Hellwig (1994, 2009). In the United Kingdom, at the time, the costs of the downturn in real-estate

markets and of the mortgage defaults were to some extent shifted to institutions in the insurance sector that had provided credit insurance to the building societies.

5

cial system, a constellation that seems to have been relevant for Spain, where the financial

crisis was concentrated in the cajas and their successor institutions, but not in Ireland, where

the entire banking system seems to have been affected from the beginning.

The Weakness of Financial Institutions is More Widespread

The notion of a “vicious circle between banks and sovereigns” diverts attention away from the

fact that the weakness of European financial institutions is not limited to countries where the

sovereign has problems. This weakness also plagues countries such as France and Germany,

where, so far at least, the sovereign has been able to bear the costs of the crisis. Quite general-

ly, banks suffer from the weakness of their equity positions, from excessive reliance on short-

term funding through wholesale markets, and from an inability to earn profits in an environ-

ment that is characterized by excess capacity and intense competition.6

The events of 2011 are paradigmatic. With the results of the stress test of July 2011, the Euro-

pean Banking Authority also divulged information about the different banks’ exposures to

sovereign risks. Investors realized that a haircut of 50 % or more on Greek sovereign debt,

which they considered likely,7 might push some major European banks into insolvency be-

cause the equity of these banks was too small to absorb the impending losses. Consequently,

investors withdrew their funding. When in September 2011 the need for a larger haircut was

officially acknowledged, the pressures intensified. They were reinforced by the banks’ own

defensive measures, such as asset sales, which contributed to the downturn in asset prices and

caused further losses in the banks’ trading books. The October Summit’s decision to raise

capital requirements accelerated the downturn because the requirement was initially formu-

lated in terms of ratios of equity to risk-weighted assets, and banks responded by further

deleveraging. The process was only stopped when the ECB’s Long-Term Refinancing Opera-

tion provided financial institutions – and markets – with an assurance that reliable funding

would be available in large amounts.

The impact of the Greek debt haircut on banks outside of Greece should be seen as evidence

of these banks’ weakness, rather than a doom loop between sovereigns and banks. As of late

2010, the Belgian-French bank Dexia had equity equal to less than 2 % of its assets. The bank

did not have much Greek debt in its portfolio, but with so little equity, the haircut on Greek

debt was enough to make the bank go under.

And fear of such an event will cause the wholesale short-term lenders to run. Dexia, which

did not have a strong deposit base, was particularly dependent on wholesale lenders. Intense

competition had forced this bank to engage in significant maturity transformation, using short-

6 For extensive discussions of these issues, see ASC (2012, 2014). 7 Investors greeted the announcement of the European Summit of July 21, 2011, which referred to volun-

tary private-sector involvement amounting to only €37 billion, with scorn. In a letter of August 3, 2011, written to the European Union’s heads of state and government, the President of the European Commis-sion indicated that he shared this skepticism. Publication of this letter accelerated the market implosion.

6

term funding of long-term investments (the excess coverage needed as collateral on covered

bonds) in order to improve its ability to compete on margins.

Dexia was perhaps an extreme case.8 Throughout these years, however, most large European

banks have exhibited very low equity ratios and most large European banks have significantly

relied on wholesale short-term funding.9 Many of them relied on funding from US money

market funds to expand their activities in the United States, or more generally, US Dollar

markets. This reliance – and the withdrawal of US money market funds – played a major role

in the events of 2011 as well as the post-Lehman turmoil in 2008.

Consciousness of the vulnerability of financial institutions has shaped political reactions

throughout. The sentence “This might be the next Lehman event” has been prominent in many

discussions. I suspect that the breach of the no-bailout clause of the Maastricht Treaty in 2010

was at least partly motivated by a fear that the exposures of weak banks in France and Ger-

many towards Greek debt might endanger these banks if there was a haircut.10 The ECB’s

Securities Markets Programs in 2010 and 2011, its Long-Term Refinancing Operation in

2011/2012, and last not least, its ECB’s announcement of Outright Monetary Transactions in

September 2012 all seem to have been motivated by a sense that financial institutions were

weak, financial markets were jittery, and financial instability was undermining the stability of

the financial system and the macroeconomy.

Current Stability Hides Underlying Problems

Since September 2012, the European financial system seems to have become somewhat more

stable. But this only means that we are no longer in an acute state of crisis. The underlying

problems have not been resolved. Indeed, there are substantial reasons to be concerned about

financial stability even now:

Overall debt levels of nonfinancial actors have continued to go up, in particular, public

debt levels. For debtors whose risks are considered to be small, the burden of this debt

8 The German bank Hypo Real Estate (HRE), which also did not have much of a deposit base, had pretty

much the same experience, except that, in 2010, HRE had put more than €170 billion of problem assets into FMS Wertmanagement, a “bad bank” owned by the German government, so that the costs of the Greek haircut did not affect HRE. Because of their reliance on wholesale short-term funding, both Dexia and HRE had previously been particularly hard hit by the breakdown of money markets in September 2008.

9 As discussed by Brealey et al. (2010) and by Demirgüc-Kunt and Detragiache (2010), unweighted equity ratios have been significantly better indicators of bank robustness than risk-weighted equity ratios. From the late 1990s until 2007, unweighted equity ratios of large European banks went down significantly while risk-weighted equity ratios remained roughly the same. Even after correcting for differences in ac-counting rules, unweighted equity ratios in Europe tend to be significantly lower than for commercial banks in the United States. For an account of European developments, see ASC (2014).

10 As indicated by the numbers published after the 2011 stress tests, these exposures of French and German banks were subsequently much reduced. One may conjecture that some of the sales of these positions contributed to the increases in Cypriot banks’ holdings of Greek debt, which occurred at about the same time and contributed greatly to the Cypriot banking crisis, which came in the aftermath of the Greek hair-cut of March 2012.

7

may be light because nominal interest rates are small. However, for debtors whose

risks are considered significant, private borrowers and sovereigns in the European pe-

riphery countries, the burden is significant. Moreover, there always is a risk that inves-

tors might become yet more pessimistic again and ask for even higher risk premia.

Such increases in risk premia would further increase the burden on borrowers, which

might end up confirming the pessimism of investors.

Endeavors to improve the competitiveness of periphery countries may further increase

the burden of their debt. Many commentators have suggested that periphery countries

would easily regain competitiveness if only they were allowed to devalue. Such com-

ments overlook the difficulty that devaluations raise the burden of debt denominated

in foreign currencies.11 The same difficulty arises if the real exchange rate is lowered

by domestic deflation, rather than a devaluation of the currency.

Many banks are still weak, in particular in the periphery countries. Specifically, many

banks still have little equity and rely on the ECB for substantial funding. Such banks

tend to concentrate their investments in their own governments’ debt and in tradable

securities. Lending, in particular, lending to new firms, tends to come from banks that

are better capitalized.12 As mentioned above, the diversion of funds away from lending

to nonfinancial companies is a drag on the macroeconomy, in particular on economic

growth.

In contrast to their counterparts in the US, European banks’ profits do not seem to

have recovered yet. This is problematic because retaining earnings is the easiest way

to rebuild equity. The ability to earn profits would also be the best means of restoring

market confidence, enabling banks to reduce their reliance on ECB funding. There

seem to be several reasons for this low profitability: First, banks may find it hard to

earn significant profits because, following the crisis, banking capacity has not been

much reduced and competition is still intense. The post-Lehman policy of bailing out

most banks has prevented the adjustment of market structures that would otherwise

have occurred. Second, the low-interest environment, while allowing for cheap fund-

ing, also reduces the rates banks can charge and may thus contribute to margins being

low. If so, we face the dilemma that higher interest rates might seem to provide for

better margins on new lending, but higher interest also raise the risks from high levels

of outstanding debt.

The low profitability of banks also raises questions about the skeletons that they may

still have in their closets. For a few years now, we have seen European banks earning

moderate profits in the first three quarters of the year and then showing sizeable losses

11 This problem is well known from the experience of Latin-American countries. Devaluation of the curren-

cy reduces the debt burden only if debt is denominated in the currency itself. On the inability to issue debt in the country’s own currency, see Eichengreen and Hausmann (1999).

12 This is shown in Acharya and Steffen (2013).

8

in the last quarter. These losses seem to be driven by write-offs that are calibrated so

that the overall result for the year is a black zero. While it is reassuring to see operat-

ing profits that enable them to pursue this strategy at all, one may wonder about the

writeoffs that have not yet been taken. This concern is particularly relevant for posi-

tions in the bank book. I am not convinced that the asset prices that underlie the collat-

eral valuations for German shipping loans or Irish or Spanish real-estate loans have

been properly adjusted to the realities of the asset markets in question, to the extent

that these markets are operating at all. I appreciate that asset markets are sometimes

excessively volatile but I also know that some of the asset price declines that we have

seen reflect a substantial asset overhang rather than any short-term market jitters. The

shipping crisis, for example, will not disappear before the excess of prevailing capaci-

ty over demand at marginal-cost prices has been removed and shippers are again able

to earn margins over average variable costs. This simple outcome of elementary mi-

croeconomic analysis has been neglected in all predictions from shippers and their

bankers that I have seen.13

Such concerns are also among the reasons why some banks still do not have much ac-

cess to market funding and why for others such funding may become jittery again. As

long as there are reasons to believe that a bank has not yet laid open all its losses, in-

vestors will also be concerned that the bank might be insolvent and will not be willing

to fund it unless they expect to be bailed out, by taxpayers or by the central bank.

Political Procrastination

Some of these problems lie beyond the purview of banking regulation and banking supervi-

sion. However, the persistent weakness of European financial institutions also reflects short-

comings in the policies that have been followed since the crisis. In particular, as mentioned,

the post-Lehman policy of bailing out most banks has prevented the adjustment of market

structure that is necessary if the intensity of competition is to be reduced to a level where

banks do not have to take unconscionable risks in order to survive because there is too much

capacity in the market.

An important role was also played by regulatory forbearance towards the problematic assets

that banks might have in their books. Closing one’s eyes to the fact that performance of loan

customers and collateral values may be questionable may seem a convenient way to avoid

disagreeable and potentially costly interventions. However, more often than not, the problems

13 I made this prediction in 2009 when, as chair of the Lenkungsrat Unternehmensfinanzierung, I was in-

volved with the applications of two major shipping companies for support from the German government’s Wirtschaftsfonds Deutschland. According to the documents we got at the time, the shipping crisis would run in parallel to the business cycle and was therefore predicted to be over by 2012. In 2013, when the governments of Hamburg and Schleswig-Holstein proposed to raise their second-loss guarantees for the asset portfolio of HSH Nordbank, the prediction was that the crisis would be over by the end of 2014, even as excess capacity in shipping was still building up; see Hellwig (2013).

9

do not disappear on their own and the delay is likely to make the intervention that much cost-

lier when it becomes unavoidable.14

There are several reasons for these shortcomings. First, intervention is always costly. If a bank

is in serious trouble, a recapitalization costs money, and resolution may bring turmoil to the

economy. Governments and supervisors must also fear public scandal as people ask why the

problems have been allowed to arise and why they have not been dealt with before. Kicking

the can down the road and hoping for the best may therefore seem more attractive. If the

banks in question are extremely large or if there are very many of them, the problem may also

be too big to handle because the public funds needed to avert the negative fallout from the

crisis may exceed the government’s fiscal capacity. Thus, when the Swedish government in-

tervened very promptly to clean up the banking system in 1992, it lacked the fiscal capacity to

also smooth the recession (which however was short, thanks to the cleanup of the banking

sector and to the trade effects of currency devaluation).

Second, banks are political. This is true in particular of public banks like the German Landes-

banken, whose lending policies are often tailored to the interests of the regional governments

that own them. More generally, political authorities tend to think of banks as institutions that

should serve to fund their policies, promoting the government’s industrial policies or simply

funding the government itself.15 In some cases, the government’s industrial policies have been

focused on the banks themselves, using financial institutions that attract funds from the rest of

the world and invest funds in the rest of the world as a tool for creating a fair number of high-

paying jobs very quickly.16 With such a policy stance, they are not likely to engage in active

interventions that would force the banks to lay open their losses and either recapitalize or re-

trench their activities.

Cross-Border Externalities in the European Union and the Euro Area

European integration also plays a role. In the European Union, and in particular in the euro

area, national policies towards banks are fraught with cross-border externalities. If a bank’s

activities in all countries of the European Union, indeed, in the European Economic Area, are

regulated and supervised under the home country principle, any bank’s customers and coun-

terparties depend on the home country’s authorities’ doing a good job to ensure the safety and

soundness of their banks. If the home country’s authorities are interested in using the banking

sector as a source of economic growth however, they may be willing to compromise on su-

pervisory standards.

14 On this point, see ASC (2012), as well as Caprio and Klingebiel (1996, 1997). 15 For a more detailed discussion, see Chapter 12 in Admati and Hellwig (2013). 16 This has been the experience of Iceland, Ireland, and Cyprus. More traditional financial centers, such as

the United Kingdom or Switzerland have also seen economic growth fuelled by promoting the financial sector as an export industry but their dependence on this sector has been somewhat less pronounced.

10

Such laxness played a role in Icelandic banks growing by acquiring deposits from customers

in the United Kingdom and the Netherlands. In the crisis, the costs of bailing out these deposi-

tors were borne by the United Kingdom and the Netherlands rather than the home country of

the failing banks.17 In Ireland, a promise of “light-touch” regulation and supervision was a

means of attracting financial business to Ireland, and funds from abroad fuelled the Irish real-

estate bubble. In the crisis, the Irish government ended up bailing out the senior unsecured

creditors, many of them banks from other European countries, but from what I have been told,

this decision was anything but a foregone conclusion and involved much pressure from Euro-

pean institutions and from other member states.

In the case of Ireland, the decision to bail out the senior unsecured creditors required the

country to seek help from the EFSF. The problem of cross-border externalities was thus shift-

ed from the level of cross-border externalities for investors to the level of cross-border exter-

nalities for other member states and European institutions. The Spanish request in 2012 for

ESM funding of bank recapitalization exhibits the same kind of externality. In the mid-2000s,

national authorities in Spain failed to interfere with banks fuelling a real-estate bubble. Ulti-

mately, this failure was at the origin of the need for ESM support in 2012.

In the summer of 2012, the other member states of the euro area had a substantial interest in

the matter. Markets were dominated by a sense of panic that threatened the funding of finan-

cial institutions all over Europe, as well as the funding of the Spanish sovereign. There were

substantial fears that the Spanish authorities had been less than incisive in dealing with the

problems of the cajas and their successors and that the hidden losses might exceed the sover-

eign’s capacity to bail out the banks’ creditors.

As in 2011, these developments put the ECB on the spot. Financial stability is not explicitly

mentioned in the Treaty as an objective of ECB policy, but banks are an important part of the

monetary system, and a banking crisis poses a serious threat to monetary stability. In Spain in

2012, markets were again jittery and the monetary system was under pressure. Even deposi-

tors, usually the most patient of investors, were moving their funds out of the country.18

17 Remarkably, the EFTA Court accepted the argument of the Icelandic government by which it was legiti-

mate to transfer Icelandic deposits but not foreign deposits from the failing banks to the successor institu-tions so that the government’s bailout measures benefited only domestic depositors.

18 In this context, it is helpful to go back to the simple quantity theory approach of Friedman and Schwartz (1963). For the years 1929 – 1933 in the United States, they observed that, while the monetary base grew by 15 %, the quantity of money (M1) contracted by 33 % because the banking crises induced behavioral changes among depositors as well as banks, raising both the currency-deposit ratio and the reserve-deposit ratio. In their assessment, monetary policy in these years was “contractionary” because the expan-sion of the monetary base failed to compensate for the implosion of deposits. Bernanke (1983) focuses on the credit channel for monetary transmission, arguing that bank closures caused the loss of information capital that had been accumulated in banks’ lending to firms and disappeared when the banks went under. Despite the differences in their accounts of the transmission mechanism, the different authors agree that commercial banks are an essential part of the monetary transmission mechanism and that a breakdown of banking calls for additional measures of the central bank.

11

Putting the ECB on the Spot

Throughout these years, with unorthodox measures in 2008, the Securities Markets Program

in 2010 and 2011, the Long-Term Refinancing Operation in 2011/2012, the announcement of

Outright Monetary Transactions in 2012, the ECB has repeatedly stepped in to preserve fi-

nancial and monetary stability by counteracting the effects of financial sector weaknesses. It

could do so because it was in a unique position to act without regard to funding constraints.

There are, however, substantial reasons to believe that the Long-Term Refinancing Operation

benefited not only healthy banks but also banks whose health was doubtful, perhaps even

banks that would have been insolvent if they had been forced to uncover their hidden losses.

In fact, reliance on ECB support was most important for those banks that had the weakest cap-

ital positions and the greatest difficulties in obtaining market funding.19

A decade ago, the various Memoranda of Understanding (MoU) for how to deal with banks in

difficulties provided for a clear division of tasks: Solvency problems were to be covered by

the national treasuries, liquidity problems of individual institutions by the national central

banks, and liquidity problems of the entire system by the ECB.20 If supervisory forbearance at

the national level enables de facto insolvent banks to benefit from ECB funding, these princi-

ples are violated, and there is little that the ECB can do about it.

The very strength of the ECB is a source of weakness. If the ECB is serious about monetary

stability, it is forced to follow a policy that effectively supports the financial system, including

those institutions that should be resolved but are not. Given the knowledge that the ECB will

support the system anyway, the pressure on national governments and national supervisors to

clean up their banking systems is that much weaker. Some politicians may in fact have come

to understand that the very weakness of their banks gives them an indirect access to the print-

ing press. After all, in the case of the Long-Term Refinancing Operation, a large part of the

money that banks got from the ECB was lent to the banks’ own governments.21

The division of tasks that was enshrined in those MoUs was naïve. Acting as a lender of the

last resort has always been an important role of central banks, and this role has always in-

volved the provision of implicit subsidies to the banks that received the support.22 One of the

more successful central banking operations of recent decades was the 1990 turnaround of US

monetary policy. When the large money center banks in the US were in a state of crisis, the

Federal Reserve lowered short-term interest rates quite drastically and allowed the troubled

banks to rebuild their equity by playing the yield curve for years. However, apart from the

implicit transfer of seigniorage from the central bank to the commercial banks, this policy had

the drawback that, as seen in 1994, commercial banks became very vulnerable to interest rate

19 Acharya and Steffen (2013). 20 For a critical discussion of this arrangement, see Hellwig (2007). 21 Acharya and Steffen (2013). 22 For a systematic discussion, see Hellwig (2014), with references to Goodhart (1988).

12

shocks and, more importantly, that they came to believe in the “Greenspan put” as protection

against any risks that they might run.

In the European context, the availability of ECB support has contributed to the maintenance

of market structures and the failure to close unprofitable banks. Here again, national govern-

ment policies involve significant cross-border externalities. Countries that expand their finan-

cial sectors as a means of industrial policy put pressure on bank margins Europe-wide. So do

countries that provide explicit or implicit guarantees to some or all of their banks. The fact

that banks like Dexia and HRE had to engage in wholesale short-term funding for the excess

of their portfolios over their covered-bond issues must in part be ascribed to the German legal

“reform” of 2005, which reduced barriers to entry into covered-bond markets, a measure that

allowed the Landesbanken to much expand their activities in this segment. While the crisis

has induced some retrenchment, many of the basic structures are still in place, ready to ex-

pand again when the occasion arises.

Maintenance of market structures with excess capacities through explicit or implicit guaran-

tees and other subsidies should in principle be prevented by the European Commission’s state

aid control. However, as shown by the decade-long fight over the public guarantees for the

Landesbanken, in the area of banking, where significant political stakes are involved, state aid

control is weak and slow.23 With the crisis, state aid control has become even weaker because

any government that wants to maintain a bank will simply claim that, if the bank is resolved,

financial stability will suffer.24 Such a claim may be dubious but the rules for state aid to fi-

nancial institutions that have been put in place since 2008 allow for financial stability consid-

erations, and it is not easy for the Commission to question whether the bank really poses a

threat to financial stability.

23 In the end, the 2001agreement between Commissioner Monti and the German government enabled the

European Commission to establish the prohibition of guarantees for public banks as a form of illicit state aid without having to go to court over the matter. For details, see the European Commission’s Press Re-lease IP/02/343 01/03/2002 of February 28, 2002, on the implementation of the understanding, available at http://europa.eu/rapid/press-release_IP-02-343_en.htm?locale=en. Under the agreement, the Commis-sion accepted a four-year transition period, which public banks used to raise significant additional fund-ing under government guarantees. Wasteful investment of the funds was a major reason for their difficul-ties in the crisis, from German banks Sachsen LB and West LB to Austrian Hypo Alpe Adria.

24 See for example the case made in 2013 by the German government to justify renewed support for HSH Nordbank. See

13

3. Will Banking Union Solve the Problems?

The Decision of the Euro Area Summit of June 2012

The many cross-border externalities in financial-sector regulation suggest that, in a monetary

union, a system with purely national control over financial institutions may not be viable.25

Given the importance of judgment in supervisory decisions, the mere harmonization of the

legal framework through regulations and directives may not be enough to eliminate moral

hazard and negative cross-border externalities. Recognition of these problems led many to

argue for the creation of a European banking union.26

However, the different participants in the June 2012 decision had different interests and were

pursuing different objectives. The European institutions saw banking union as a further deep-

ening of European integration and hoped that this would overcome the problems. In particu-

lar, the European Commission was pushing for a European deposit insurance system in order

to stop the outflow of deposits from countries that were perceived to be at risk. The European

Central Bank was pushing for a single supervisory mechanism in order to get out of the

straightjacket of having to tailor its monetary policy to the needs of financial stability that

were insufficiently taken into account by national authorities. Spain was pushing for ESM

support for recapitalizing its banks. Germany, it seems, was pushing for European control as a

prerequisite to making ESM funds available to Spanish banks, perhaps without appreciating

that this might also involve European control over German banks.

Developments since then have been much influenced by these differences in interests and ob-

jectives. They have also been influenced by differences in legislative procedures for the dif-

ferent components of the banking union. In the euro area, supervision will be handled by the

Single Supervisory Mechanism (SSM), which is created by a Regulation of the Council under

the auspices of Art. 127 (6) TFEU. Resolution in the euro area will be handled by the Single

Resolution Mechanism (SRM), which is created by an EU Regulation under the auspices of

Art. 114 TFEU, and will be funded by a Resolution Fund, which is created by an intergov-

ernmental agreement with the approval of the Commission and the Parliament. In the Europe-

an Union as a whole, procedures for dealing with banks in difficulties will be governed by the

Bank Recovery and Resolution Directive (BRRD), which still needs to be transposed into na-

tional laws. New rules for deposit insurance will also be governed by a directive. I am won-

dering to what extent the differences in legal foundations may end up affecting the viability of

the overall system.

25 Some of the problems with the previous arrangement were pointed out in Hellwig (2007). ASC (2012)

suggests that, even for the European Union as a whole, with the internal market in banking, a purely na-tional control over financial institutions, subject to European regulations and directives, is problematic.

26 Brussels-based Bruegel provided some of the key arguments and ideas. See, in particular, Pisani-Ferry et al. (2012).

14

Supervision: Heterogeneity of National Laws and Judicial Review

An important innovation of Council Regulation (EU) No 1024/2013, which creates the Single

Supervisory Mechanism, concerns the status of supervisory authorities. According to Art. 19

of the Regulation, the ECB and the competent national authorities shall be independent in

their supervisory activities. This is a welcome change from the status quo ante, which had at

least some supervisory authorities subordinated to their respective governments.27 This

change provides some hope that supervisory decisions will become less influenced by the na-

tional governments’ political interests.

As a practical matter, the shift to the SSM is unlikely to pose major problems. But even here,

there are pitfalls. One pitfall concerns the heterogeneity of laws and jurisdictions. One might

think that, in principle, there is just one set of rules for the entire European Union. However,

only regulations are directly applicable. Much of the relevant EU law takes the form of direc-

tives, which are not directly applicable but require transposition into national law. Art. 4 (3)

of the Regulation stipulates that “the ECB shall apply all relevant Union law, and where this

Union law is composed of Directives, the national legislation transposing those Directives.”28

This means that the ECB will have to apply 17 or more different laws. This heterogeneity

raises issues of consistency across member states. It also raises questions about the roles of

the Court of Justice of the European Union (CJEU) and the national courts in judicial review.

According to Recital 60 of the Regulation, judicial review is assigned to the CJEU, which will

assess “the legality of acts … of the ECB, other than recommendations and opinions, intended

to produce legal effects vis-à-vis third parties.” This formulation is taken almost verbatim

from Art. 263 TFEU, and the recital refers to this article. Even so, I am wondering how this

will work for decisions of the ECB that are taken under national laws even when these laws

implement European directives. Is the CJEU really qualified for this task? And does the refer-

ence to the Treaty provide a sufficient legal basis for this assignment? Presumably, the word

“legality” in the recital and in the Treaty refers to the legal norms of the European Union, not

to the legal norms of the member states. But the laws implementing the European directives

are legal norms of the member states.

In this context, I note that the different courts have different traditions of judicial review of

administrative decisions. For example, in the area of competition policy, the CJEU has tradi-

tionally given the European Commission wide scope for the exercise of judgment in the as-

sessment of facts. The court has typically refrained from making a judgment of its own about

the validity of the Commission’s conclusions and merely asked whether the evidence and the

27 A decade earlier, this had been a matter of dispute in the discussion about the European Constitution. The

ECB would have liked the Constitution to stipulate independence of central banks in all their activities, not only in matters of monetary policy. The Constitutional Convention did not accept the ECB’s proposal. See European Commission (2013).

28 The question of how to deal with legal norms that are codified at the European level in the form of Direc-tives was raised by Sapir et al. (2012) in a comment on the European Commission’s first draft of the Reg-ulation, which did not address the problem at all.

15

reasoning that the Commission had put forward were sufficient to support its decisions under

the relevant legal norms.29

By contrast, some of the national judiciaries are more restrictive. In Germany, for example,

administrative courts draw the lines for the exercise of judgment very narrowly and require

the administrative authority to provide material proof that its decision is derived from the le-

gal norms as they apply to the facts on hand. Underlying this practice is the consideration that,

under the democracy mandate of the Basic Law, i.e., the constitution, any decision by an ad-

ministrative authority should in principle be derived, at least indirectly, from choices taken by

the representatives of the voters.

In the context of the SSM, the difference in judicial attitudes to the exercise of judgment by

an administrative authority is important because much supervisory activity does involve such

an exercise of judgment, judgment about the quality of assets that a bank holds, about the

riskiness of a bank’s strategy and even the professional quality of its management. Indeed,

this exercise of judgment is where the governance of supervision matters most and where the

shift to a Single Supervisory Mechanism may be presumed to have the biggest impact.

One may hope that these issues will never arise because nobody goes to court. However, even

if nobody goes to court, the mere threat that affected parties might do so can have an effect.

Consider the public discussions that we have had after the crisis about supervisory laxness in

the preceding years, for example, the German supervisor’s acceptance of practices whereby

banks created special purpose vehicles to hold mortgage-backed (and other) securities without

backing them by equity, funding them through asset-backed commercial paper and providing

the creditors with liquidity guarantees for these vehicles. These vehicles and the commitments

that banks made to them played a major role in the buildup of risks before the crisis, and they

caused substantial losses. The German supervisor has maintained that they were aware of the

risks but, under the letter of the prevailing law and given the strictness of German administra-

tive courts, they did not see any room for prohibiting these practices. Other supervisors were

more restrictive and disallowed such practices.

Would this have played out any differently if the ECB had already been in charge? Or would

the ECB have issued an injunction against the use of special purpose vehicles for holding in-

vestments off-balance-sheet? If so, how would the decision have been treated by a German

administrative court? Or, if the appeal to a German administrative court was precluded by

Recital 60 of the SSM Regulation, how would the German Constitutional Court have looked

29 For example, in the famous Airtours-First Choice merger case of 2002, the Court overturned the Commis-

sion’s decision on the grounds that the Commission assessed the merger as being conducive to more col-lusion among the remaining suppliers in the market without giving a coherent argument as to why this would be so. The plaintiffs had argued that, according to standard economic theory, a variety of objective features of the markets and products involved would make collusion quite difficult. In the same year, the Commission’s prohibition of the Schneider-Legrand merger was turned down on the grounds that the Commission had failed to provide the empirical data needed to establish its case.

16

upon the denial of due process against an administrative measure taken under German law?30

For after all, the law on banking regulation that implements the European directives is a na-

tional law!

Such problems would of course be removed once and for all if all the relevant legal norms

were brought into the Regulation. I expect that, at some point, we will get there, and I hope

that, in the intervening time, the uncertainties and impracticalities associated with the hetero-

geneity of national laws will not be too costly.

Dealing with Banks in Difficulties: Recovery and Resolution Procedures

The most careful and most professional supervisor is helpless if there is no practical way of

dealing with problem banks.31 The Lehman experience has made us all very sensitive about

this issue. We learnt that “Too Big To Fail” is not a myth: Letting a bank fail can indeed have

catastrophic consequences, and can be much costlier than a bailout.

Since the Lehman crisis, authorities worldwide have been torn back and forth between two

concerns, on the one hand, the desire to avoid a repetition of the post-Lehman panic, on the

other hand, the desire to develop procedures for dealing with problem banks that would avoid

the kind of tsunami that we saw in September 2008. The BRRD and SRM are part of this pro-

gram.

However, I sometimes wonder whether improvements in resolution procedures are really

meant to make resolution viable, or whether they are meant as placebos to avert political pro-

test against a regime in which the financial industry has blackmailed taxpayers into providing

support, for fear that otherwise things might get much worse. Many of the reforms that have

been instituted are likely to prove impractical if we get into another crisis.

In the rhetoric accompanying such legislation, the proponents never show how the new legis-

lation would have worked in the Lehman crisis if it had been available then. If we want to

avoid a repetition, however, it is imperative that we recall precisely those problems and see

what can be done about them. Actually, the post-Lehman experience was different in different

countries:

30 Such constitutional concerns can also be raised about Art. 13 of the SSM Regulation, which relates to the

authorization by a judicial authority of an on-site inspection if such authorization is required under na-tional law and asserts that in such cases the national judiciary shall control that the measures taken in this context are not taken wilfully, but shall not decide on the lawfulness of the measures; lawfulness is to be assessed by the European Court of Justice. If the conditions for such inspections are specified in the na-tional law, this article transfers jurisdiction over the application of national law out of the domain of the national judiciary.

31 This point is very much emphasized in ASC (2012) and Sapir et al. (2012).

17

From the perspective of the United States, the post-Lehman experience is dominated

by the implosion of money market funds.32 The Lehman Brothers bankruptcy caused

Reserve Primary to break the buck. News of this event triggered a run on Reserve

Primary and on other money market funds. As a result, all money market funds with-

drew funding from banks, in particular US investment banks, which the Lehman

Brothers bankruptcy had made to appear more risky anyway. As banks all over the

world were scrambling for cash, they tried to sell assets, which sent asset prices into a

tailspin.

In the United Kingdom, the post-Lehman experience is dominated by the disappear-

ance of a key market maker in derivatives markets. Maintenance of systemic functions

was deemed to be impossible because there was no legal basis for doing so and be-

cause there was no funding. Lehman Brothers, London, was a legally independent

subsidiary, but the different subsidiaries in different countries had integrated cash

management. When authorities in the UK took over the bank, they found that there

was no cash because all cash had been sent to New York at the previous close of busi-

ness.

Three important difficulties emerge:

As different legal entities belonging to the same group go into different bankrupt-

cy/orderly liquidation/recovery and resolution procedures, each one in the country

where it is located, the integrity of corporate operations is destroyed, and this can de-

stroy the viability of systemically important functions. In the case of Lehman Broth-

ers, this was most noticeable for their integrated cash management. Potentially even

more important are integrated IT systems, where the entry of multiple resolution au-

thorities in multiple places raises the question of what is the legal or contractual basis,

and what are the rules and the pricing, for continued joint use of these systems which

is essential for the maintenance of systemically important operations.

Any maintenance of systemically important operations requires funding. Without

funding, such operations cannot be maintained. Market funding, however, is likely to

vanish unless creditors are given guarantees that they will not be harmed.

Systemic effects are not limited to domino effects from the breakdown of existing con-

tracts. The disappearance of contractual partners on whose availability one had count-

ed or the implosion of asset prices from fire sales may be much more important.

32 The AIG episode occurred at the same time but, as far as I can tell, this episode was not directly related to

the Lehman Brothers bankruptcy. For a more detailed account, see Admati and Hellwig (2013), chapter 5, and the references given there, in particular FCIC (2011).

18

In thinking about the maintenance of systemic functions, it is worth recalling that

Lehman Brothers had hundreds if not thousands of subsidiaries.33 If such subsidiaries

act in an integrated fashion, managing the system is a daunting task even for those

who know it. For the authorities replacing incumbent managing, the task is that much

more difficult.

Dealing with Banks in Difficulties: Banks with Systemically Important

Subsidiaries in Different Countries

On the key issue of how to maintain systemic functions of a bank with systemically important

operations in different countries, progress since 2008 has been miniscule. Multiple-point en-

try, i.e., the entry of different authorities of different countries into the legally independent

units located there, is still the prevailing legal rule. The United States and the United King-

dom have been negotiating about single-entry procedures, but they seem to be thinking more

of recovery than resolution, and the issue of loss sharing in resolution has not been settled.

The living will that Deutsche Bank has submitted to the Fed and FDIC proposes that US au-

thorities should let the German authorities deal with any crisis situation. However, the US

authorities do not seem to be convinced by this proposal. Their recent ruling that foreign

banks must organize their US subsidiaries so that US equity and liquidity requirements can be

imposed indicates that they are thinking of ring-fencing the US operations of foreign banks.

Given the experience of ring-fencing by European supervisors, e.g., the restrictions that Bafin

imposed on Unicredit Germany in 2012, one can hardly blame them.

Directive 2014/59/EU of the European Parliament and of the Council establishing a frame-

work for the recovery and resolution of credit institutions and investment firms (BRRD) pro-

vides for some coordination within the college of resolution authorities. However this coordi-

nation can hardly substitute for the organizational integration of operations in the bank as a

going concern. This basic problem remains unsolved. Therefore I predict that, if a bank like

Barclays, BNP Paribas or Deutsche Bank, with systemically important functions in different

countries, were to get into trouble, authorities would be unwilling to enter into a recovery and

resolution procedure, i.e., we would continue the post-Lehman practice of bailing banks out.

The SRM provides for a centralized procedure with single-entry resolution for large banks.

However the procedure is complex and provides much scope for participants to veto decisions

they do not like. For institutions of the importance and complexity of BNP Paribas or

Deutsche Bank, the mechanism will therefore be no more practical and trustworthy than the

provisions of the BRRD. Indeed, since the US and the UK do not participate in the SRM, a

major part of the multiple-entry problem is not even addressed.

33 Herring and Carasi (2010) mention 433 majority owned subsidiaries, Miller and Horowitz (2012) speak

about 8000 subsidiaries in over 40 countries.

19

Dealing with Banks in Difficulties: The Need for Interim Funding

Another shortcoming of the BRRD is its naiveté about the time needed for resolution and the

need for funding during this time. Recitals 103 - 105 note that such funding may be needed

and assert that it should be provided by resolution funds under the control of resolution au-

thorities. Given the numbers involved and given past experience, this is unrealistic.

The Single Resolution Fund for the SRM is targeted for a level of € 55 billion, the German

Bank Restructuring Fund for a level of € 70 billion, to be reached after many years. These

numbers are much too small to ensure interim funding of institutions like Deutsche Bank or

BNP Paribas, with liabilities on the order of € 2 trillion, a large part of which is wholesale and

short-term, i.e., easy to discontinue if counterparties get nervous. Promises of support from a

fund with € 55 or € 70 billion are not going to stop a run if creditors with claims amounting to

€ 1 trillion or more are worried about a bank. In fact, this is not just a problem for banks with

trillion-euro balance sheets. The problem also arises with banks like Commerzbank or the

Landesbanken, whose liabilities amount to several hundreds of billions of euros.

Discussions about the funding of recovery and resolution procedures usually pay too little

attention to the distinction between the need to fund operations as long as they are ongoing

and the need to allocate or to absorb ultimate losses. Resolution or restructuring fund target

levels in the double-digit-billion range may be sufficient to absorb ultimate losses, but they

stand in no realistic relation to the interim funding that is needed to keep systemically im-

portant operations going, at least for a while. The SRM will be able to borrow from the ESM

but the numbers that have been given there, like those for restructuring or resolution funds,

stand in no realistic relation to what is needed to maintain interim funding.

In ordinary insolvency law, the problem of interim funding for ongoing operations is usually

handled by giving new creditors, i.e., creditors who come in after the firm has entered into

insolvency proceedings, priority over previous creditors. For nonfinancial companies, this

arrangement is viable, at least for a while, because the funds needed to maintain ongoing op-

erations tend to be small relative to the firm’s assets.

For a bank, this arrangement is problematic, which is precisely why we need a procedure that

is different from ordinary insolvency procedures. Banks have a lot of short-term funding,

through wholesale loans as well as deposits. If these claims on the bank are frozen, there may

be substantial systemic damage. For example, a money market fund whose claims are frozen

may be run upon, as Reserve Primary was after the Lehman Brothers bankruptcy. As we saw

in September 2008, such runs on money market funds may endanger the entire system of

short-term wholesale bank funding. If the short-term claims on the bank are not frozen,

maintenance of bank funding requires that these claims be renewed or replaced. For a bank in

a resolution procedure, such renewal or replacement of funds will not be forthcoming unless

the lenders are given public guarantees. Priority over previously incurred liabilities of the

20

bank is not sufficient because the amount of such funding is large in relation to the bank’s

assets so that, without public guarantees, there is a risk for the lenders.

Nor is it sufficient to exempt secured claims and very short-term inter-institution claims from

bail-in, as the BRRD does. For lenders with secured claims, there is always a question wheth-

er the collateral is sufficient. With Bear Stearns and Lehman Brothers, doubt about the collat-

eral caused the “repo runs” on these institutions. Such doubts can be caused by concerns of

the collateral value itself. They can also be caused by concerns about re-hypothecation, i.e.,

the fact that the same securities are used as collateral for several loans. If such doubts cause

lenders to increase collateral haircuts, encumbrance of the bank’s assets by collateralization is

exacerbated – and the ability to maintain funding further endangered.

Exemptions of very short-term inter-institution claims are more clearcut but even so these

claims are vulnerable to the risk that the lenders themselves might be run upon, as happened

to US money market funds after the Lehman Brothers bankruptcy.

The problem of interim funding can be solved by providing resolution authorities with public

guarantees or by allowing these authorities to borrow from the public purse. Under the Dodd-

Frank Act in the United States, the FDIC can simply borrow from the Treasury.34 Under the

BRRD, however, the problem is not addressed. An important question will be whether nation-

al legislation will go beyond the BRRD and provide resolution authorities with sufficient ac-

cess to interim funding or with sufficient backstops so that they can give the guarantees that

are needed to maintain the systemically important functions of a bank at least for a while.

Dealing with Banks in Difficulties: Asset Valuation and Bail-Ins

To some extent, the neglect of interim funding problems seems to be due to the fact that the

BRRD has a very optimistic vision of how resolution is carried out: Some Friday, the supervi-

sory authority determines that a bank is likely to fail. It calls for the resolution authority to

take over. The resolution authority obtains an independent valuation of the bank’s assets and

liabilities. On the basis of that valuation, it writes down the bank’s equity, and it writes down

the bank’s liabilities or converts them into equity, following the hierarchy of claims under

insolvency law. If all this is done over the weekend, then by Monday the bank is again well

capitalized, and the resolution authority is in a good position to move forward. Perhaps it has

already used the weekend to sell the business or to set up a bridge bank.

This vision is too optimistic. First, asset valuation is problematic. At the time of entry of the

resolution authority into the bank, the bank’s prospects and the value of its assets are highly

uncertain. The uncertainty about the value of the assets may itself be a key factor in the diffi-

culties of the bank.

34 The German Bank Restructuring Act of 2010 also allows for borrowing from the public purse, but the

scale of the restructuring fund is by an order of magnitude smaller than interim needs for funding and/or guarantees.

21

What was the value of the United States S&L’s assets in 1990? What was the value of assets

and derivatives in the books of Long-Term Capital Management in September 1998? What

was the value of mortgages and mortgage-backed securities in the books of Lehman Brothers

or AIG in September 2008? What was the value of real-estate loans in the books of Spanish

cajas in 2012? The answers to these questions are highly sensitive to the chosen principles for

valuation. They are also a matter of judgment as to how long the current crisis is going to last.

Finally they depend on how quickly the assets in question have to be liquidated. In the case of

the US S&Ls, estimates of the costs to deposit insurance institutions were on the order of $

600 billion to $ 800 billion dollars around 1990; in the end, these costs came to $ 153 bil-

lion.35 In the case of LTCM, the Federal Reserve feared that a bankruptcy followed by a quick

liquidation of assets and derivatives might trigger an asset price implosion. Therefore, it put

pressure on other banks to provide interim funding for a slow liquidation. The strategy was

successful so that the banks involved did not actually lose money on the interim funding.

The directive does allow for a preliminary valuation as a basis for bail-ins, but it also asks for

an ex post valuation to be performed “as soon as possible”. In the case of the S&Ls, “as soon

as possible” would have been ten years later, which probably is not what is meant by the

BRRD. A reliable final loss allocation however does require a lot of time – unless the authori-

ties are willing to speed the procedure up, if necessary by selling assets prematurely.

Second, resolution involves more than a valuation of assets and a recapitalization on the basis

of writedowns and debt-to-equity conversions. Key questions concern the correction of past

management mistakes, the search for new owners, the decision as to which assets should be

part of the bank as a going concern and which ones should be separated and wound down.

Answering these questions takes time. During this time, uncertainty about the future of the

bank and about the value of its assets encumbers the bail-in mechanism and endangers fund-

ing – even from creditors whose claims are not subject to bail-in.

Dealing with Banks in Difficulties: Fiscal Backstops

Most legal reforms of recovery and resolution procedures that have been introduced since

2008 have come with a promise that never again will taxpayers have to foot the bill for bank

bailouts. The Dodd-Frank Act in the United States is one example, the German Bank Restruc-

turing Act of 2010 another. The BRRD follows the same principle, albeit somewhat less

stringently.

These promises are either naïve or cynical. If systemically important banks are in trouble and

the choice is whether to let them go under or to support them, the answer will be “We do not

want to have another Lehman experience!” This is the lesson learnt in 2008, and in many re-

spects it is the right lesson. The costs of the Lehman Brothers bankruptcy and the financial

turmoil it induced far outweighed whatever the fiscal costs of a bailout might have been.

35 See Curry and Shibut (2000).

22

In such a crisis situation, some public funds are needed, at least to maintain interim funding of

systemically important operations. Putting in public funds, even temporarily, puts taxpayers at

risk. If the bank is insolvent, somebody has to pay for the difference between liabilities and

assets.

Ostensibly, this is what restructuring or resolution funds and industry levies are there for.

However, in a crisis that affects the entire industry, these funds are likely to be too small to

cover the losses. Even if industry levies are increased ex post, there is no guarantee that it will

be sufficient to cover losses. In a crisis situation, the capacity of surviving industry members

to contribute to such a levy will be severely limited.36 Even if the charges are spread over

time, there is still a substantial burden, which affects the banks just as an excess of debt over-

hang would.

For example, in the S&L crisis of the 1980s in the United States, the industry was in such dif-

ficulties that it could not bear the costs of the crisis; the Federal Savings and Loans Insurance

Corporation (FSLIC) became insolvent and was merged with the FDIC. Out of $153 billion of

losses, in the end, the industry paid $29 billion and taxpayers $124 billion. If a comparable

systemic crisis was to happen today, in the US under the Dodd-Frank Act or in Germany un-

der Bank Restructuring Act, the experience would be repeated. This would be a breach of the

promises with which these laws were introduced but at least it would work.

The S&L example may be seen as atypical in that most S&L funding in the United States had

come in the form of deposits, which were federally insured. Thus there was little room for

clawbacks or bail-ins of creditors. One may therefore hope that ultimate losses in bank resolu-

tion will be smaller if more creditors are bailed in, i.e., if more creditors are forced to partici-

pate in losses as they would have to do if the bank entered a bankruptcy or insolvency proce-

dure.

On this point, the BRRD is not reassuring. The BRRD contains important statutory exceptions

from bail-ins: Covered deposits, secured liabilities and derivatives, and inter-institution lia-

bilities with maturities of less than seven days. The authorities can also grant additional ex-

ceptions on the spot if they deem such exceptions to be necessary to forestall contagion or

other forms of systemic risk. To ensure that, in spite of these exceptions, there is at least some

debt that can be bailed in, the directive requires that exempt liabilities amount to no more than

ninety-two percent of a bank’s funding. Loss absorption from equity and bail-in-able debt can

be as little as eight percent of total assets.

The Lehman crisis and the post-Lehman bailouts have created a strong lobby against any

creditor liability. Forcing creditors to bear losses, we are told, entails a danger of systemic

risks from domino effects, as those creditors themselves may be too weak to absorb those

losses, or as the realization that creditor liability must be taken seriously hurts funding condi-

36 In any event, it should be clear that the levy itself is a kind of tax, supporting institutions in difficulties at

the expense of institutions that have not seen risks materialize.

23

tions of other banks. This thinking has dominated public discussion and public policy for

quite a while, including initial discussions about the Cypriot crisis. The Cypriot crisis and the

treatment of SNS Reaal provided for some change, but as yet I am not convinced that these

events determine the new paradigm.

Indeed, given the uncertainties about how systemically important functions are to be main-

tained and funded, I expect that, in a clutch, most governments will decide that it is better to

avoid a resolution procedure altogether. Back to “Too Big To Fail”!

The BRRD leaves room for such avoidance by allowing recapitalizations of banks even be-

fore they enter into the recovery and resolution procedure. Such a recapitalization presumes

that the requisite funding is available, as is the case in countries with strong fiscal positions. If

the requisite funding is not available, the recovery and resolution procedure may still be

avoided if the authorities exert forbearance and procrastination as they have done in the past.

Without a fiscal backstop at the European level, I am not convinced that, on this account, the

SSM will change so much.

Dealing With Banks in Difficulties: Legacy Risks and Fiscal Responsibility

Ironically, the legislation for banking union took so long that the concrete problem that was of

concern to the June 2012 Summit, namely the recapitalization of Spanish banks, has been

dealt with even before the legislation had been passed, let alone entered into force. In

2012/2013 ESM provided some € 41 billion for the recapitalization of Spanish banks, with

conditionality for restructuring of the industry; by now, the Spanish government has declared

that no new assistance will be needed; the funds that were provided will be repaid over a peri-

od of more than a decade. However, in this process, the Spanish government remained (and

remains) liable as the ESM funds did not go directly to the banks but the Spanish govern-

ment’s recapitalization fund.

The question of national liability has been at the core of the political controversy. Whereas the

original Spanish proposal for direct recapitalization of Spanish banks through the ESM would

have provided for a Europeanization of legacy risks, channeling these funds through the Span-

ish government’s recapitalization fund implied that the Spanish government itself would be

liable for the debt service.

The BRRD and the SRM leave the principle of national fiscal responsibility for banks un-

touched. For the BRRD, which applies to the entire EU, this is a matter of course – as a di-

rective, the BRRD merely provides the legal background to the Internal Market in banking

and does not in itself promote the banking union. In the SRM, the issue is dealt with by deny-

ing that it is an issue at all. Claiming that recovery and resolution will be paid for by the in-

dustry without any imposition on taxpayers is a way to avoid taking a clear stand on fiscal

responsibility. In a crisis, if the institutions that are at risk are sufficiently important, if nation-

al governments are unable to provide the requisite backstops, and ESM loans are insufficient,

24

one may find out that the problem must be dealt with anyway. As in other contexts, the crisis

may be used as an occasion for further integration, albeit by hurried stopgap measures.

In the political debate about the issue, legacy assets and legacy risks have played an important

role. Even people who would in principle acknowledge that a mutualization of fiscal respon-

sibilities for banks might serve a useful insurance function have argued that you shouldn’t

provide insurance for a house that is already on fire, i.e., any mutualization of fiscal responsi-

bilities for risks in the financial system should not cover losses on existing assets. Given the

externalities from keeping those losses hidden and having the weakness of financial institu-

tions endanger financial stability and growth all over Europe, I do not find this argument alto-

gether convincing. However, it has played an important role in the debate.

One might also argue the issue with a view to moral hazard. National policies affect the safety

and soundness of banks in a given country, so fiscal responsibility for any bailouts would en-

sure that these risks are properly taken into account. But there is another side to the coin: Su-

pranational institutions for supervision and resolution take decisions that affect risks to tax-

payers. National fiscal responsibility may therefore generate moral hazard on the side of those

institutions. Indeed, until now, this argument has played a major role in justifying national

competence for supervision and even the subordination of supervision to the national finance

minister.37 As constituted at present, therefore, the new regime is bound to raise questions

about the legitimacy of decisions taken at the supranational level that impose fiscal burdens

on national treasuries.

In the short run, there is a danger that the maintenance of national fiscal responsibility will

deepen the split between “periphery” and “core” countries. There is also a danger that the

cleanup of the financial system will be further delayed. Countries with sufficient fiscal capaci-

ty will be able to use the recapitalization option under BRRD to preempt any recovery and

resolution procedure. At the level of the individual institution, this may be satisfactory, if

costly for national taxpayers, but the needed adjustment of market structure will not take

place. Countries that do not have the requisite fiscal capacity will try to continue sweeping

problems under the rug; if this is not possible, they may again be forced to have recourse to

ESM support. However, there will be enormous pressure on supervisors to exercise forbear-

ance and act as if the problems with some of the banks’ assets were merely temporary and

hopes for an eventual recovery would justify asset valuations at which the banks can be

deemed to be well capitalized.

From this perspective, it will be interesting to watch the Asset Quality Review that is to take

place later this year. On the one hand, the ECB has a strong interest in ensuring that the Asset

Quality Review is serious and that problems are laid open and remedied. Otherwise there is a

risk of problems emerging soon after the SSM begins to work, which would be disastrous for

the ECB’s credibility. On the other hand, national authorities, and to some extent the ECB

37 See Wissenschaftlicher Beirat (2008) and Hellwig (2011).

25

itself, have a strong interest in ensuring that not too many problems are laid open. Otherwise

national authorities will be blamed for past laxness; moreover, the needed remedies and ad-

justments may not be feasible for some of the participants. At this point, the outcome of this

conflict is up in the air.

In the medium run, I believe that banking union will require a Europeanization of fiscal re-

sponsibility. First, this would contribute to defusing the issue of loss sharing in dealing with

banks that have significant cross-border operations, making single-entry resolution more pal-

atable. Given that the US and the UK are not included, this would only be a small step, but

one that is nevertheless worthwhile. Second, a Europeanization of fiscal responsibility is nec-

essary for the protection of monetary policy. To the extent that national fiscal responsibility

prevents a cleanup of the financial system, the ECB remains hostage to the weakness of the

financial sector. In particular, there is little hope for overcoming the fragmentation of finan-

cial and monetary systems that we currently have. This fragmentation makes the ECB’s task

of ensuring monetary stability in the euro area all but impossible to fulfil.38

The question is whether a Europeanization of fiscal responsibility can be achieved without the

creation of a European fiscal sovereign. The fiscal backstop that is needed for the SRM to be

viable requires some tax base. So far, such a tax base does not exist. Will banking union be-

come a reason for moving forward in this direction?

4. Concluding Remarks

As indicated by the preceding discussion, I am skeptical whether banking union as it has been

designed so far will really allow us to deal with the problems that currently plague our finan-

cial sector. Whereas the Europeanization of supervision and the independence of supervision

from political authorities may eliminate some of the distortions in supervision that we have

seen in the past, the resolution regime remains nonviable in my view. “Too Big To Fail” is

still with us. Moreover, the maintenance of national fiscal responsibility for banks preserves

incentives to sweep problems under the rug, and preserves some of the factors that have been

responsible for the fragmentation of financial and monetary systems that is plaguing the mon-

etary union.

Politically, the development of banking union seems to involve a bet between the European

institutions, in particular, the ECB, and the member states. From the perspective of the ECB,

banking union holds the promise that, if it works, the ECB may get out of the straightjacket

where it has to provide funding to banks, even if they are suspected to be insolvent, which

then provide funding to their governments. From the perspective of those governments, bank-

38 In this context, it is worth noting that the German Constitutional Court’s indictment of Outright Monetary

Transactions placed particular weight on the selectiveness of the program, a selectiveness that seemed mandated by the fragmentation of the monetary systems but whose distributive implications the judges considered unpalatable.

26

ing union holds the promise that the ECB is drawn even more deeply into being responsible

for financial stability and therefore the indirect access to the printing press becomes even eas-

ier. Which side will win is unclear but it is by no means a foregone conclusion that it will be

the ECB.

To be sure, European arrangements have always evolved dynamically, dealing with problems

as they came along. One day’s problems have often become the next day’s reforms. In that

sense, my skeptical remarks can be read as an agenda for further reform. I hope that this re-

form will come before the problems become unmanageable.

At a deeper level though, I am wondering. Banks are political and have always been. The ex-

ample of Jakob Fugger financing Charles V’s election to be Holy Roman Emperor is para-

digmatic. So is the example of the Medici taking over the government of Florence in order to

protect their bank from bankruptcy. The symbiosis of banks and treasuries has for centuries

been a key element of sovereignty. Are member states really prepared to transfer this part of

their sovereignty to the European institutions? I consider this transfer to be necessary if mone-

tary union is to survive, but I wonder whether the political will is there.

However, if the European Monetary Union were to fall apart, the details could be ugly. In

those countries where people expect claims on euros to be devalued, we must expect to see

bank runs, breakdowns of banks and of payment systems, and severe economic and social

crises. Economic and social damage could be enormous. So could be the effects on people’s

feelings about European integration and all that it stands for.

27

5. References

Acharya, V.V., and Steffen, S. (2013), “The greatest carry trade ever? Understanding

European bank risk”, Working paper, New York University and European School

of Management and Technology.

Admati, A.R., and M.F. Hellwig (2013), The Bankers’ New Clothes, Princeton University

Press, Princeton, N.J.

ASC (2012), “Forbearance, resolution and deposit insurance”, Report 01 of the Advisory

Scientific Committee of the European Systemic Risk Board, Frankfurt, July 2012.

ASC (2014), “Is Europe Overbanked?”, Report 04 of the Advisory Scientific Committee of

the European Systemic Risk Board, Frankfurt, June 2014.

Bernanke, B.S. (1983), “Nonmonetary effects of the financial crisis in propagation of the

Great Depression”, American Economic Review 73 (3), 257 – 276.

Brealey, R.A., I. A. Cooper, and E. Kaplanis. (2011), “International Propagation of the Credit

Crisis.” Mimeo. London Business School, London.

Caprio, G., and D. Klingebiel. (1996), “Bank Insolvencies, Cross-Country Experiences.”

Policy Research Working Paper 1620. World Bank, Washington, DC.

Caprio, G., and D. Klingebiel (1997), “Bank Insolvency: Bad Luck, Bad Policy, or Bad

Banking?” Paper written for the Annual World Bank Conference on Development

Economics, April 25–26, 1996.

Curry, T., and L. Shibut. (2000), “The Costs of the Savings and Loan Crisis: Truth and

Consequences.” FDIC Banking Review 13, 26–35.

Demirgüc-Kunt, A., E.Detragiache, and O. Merrouche. (2010), “Bank Capital: Lessons from

the Financial Crisis.” Policy Research Working Paper 5473. World Bank, Washington,

DC.

Dermine, J. (1990), European Banking in the 1990s, Blackwell Publishers, Oxford.

Eichengreen, B., and R. Hausmann (1999), „Exchange Rates and Financial Fragility”, in:

New Challenges for Monetary Policy. Proceedings of a symposium sponsored by the

Federal Reserve Bank of Kansas City.

European Commission (2013), State Aid No SA.29338 (2013/C-30) (ex 2013N-504) —

Increase of the second-loss guarantee for HSH Nordbank AG , Invitation to submit

comments pursuant to Article 108(2) TFEU , Official Journal 2013/C315/81 –/91,

28

European Council (2012), Euro area summit statement, 29 June 2012, http://www.european-

council.europa.eu/home-page/highlights/euro-area-summit-statement.

FCIC (Financial Crisis Inquiry Commission) (2011), The Financial Crisis Inquiry Report.

Washington, DC: U.S. Government Printing Office.

Friedman, M., and Schwartz, A.J. (1963), A monetary history of the United States 1867 –

1960, Princeton University Press, Princeton, N.J.

Goodhart, C.A.G. (1988), The Evolution of Central Banking, MIT Press, Cambridge, MA.

Hellwig, M.F. (1994), “Liquidity Provision, Banking and the Allocation of Interest Rate

Risk”, European Economic Review 38: 1363–1389.

Hellwig, M.F. (2007), “Switzerland and Euroland: European Monetary Union, monetary

stability and financial stability”, in: The Swiss National Bank 1907 – 2007, Verlag Neue

Zürcher Zeitung, Zürich, 741 – 780.

Hellwig, M.F. (2009), “Systemic Risk in the Financial Sector: An Analysis of the Subprime-

Mortgage Financial Crisis.” The Economist 157, 129–207.

Hellwig, M.F. (2011) “Quo Vadis Euroland? European Monetary Union Between Crisis and

Reform”, in: F. Allen, E. Carletti, G. Corsetti (eds.), Life in the Eurozone: With or

Without Sovereign Default?, FIC Press, Wharton Financial Institutions Center, Phila-

delphia 2011, 59 – 76.

Hellwig, M.F. (2012), “The Problem of Bank Resolution Remains Unsolved: A Critique of

the German Bank Restructuring Law.” In Too big to fail—Brauchen wir ein Sonderin-

solvenzrecht für Banken?, ed. Patrick S. Kenadjian. Boston: De Gruyter. 35–62.

Hellwig, M.F. (2013), „Stellungnahme zur Anhörung des Haushaltsausschusses der Bürger-

schaft der Freien und Hansestadt Hamburg über die Wiedererhöhung der Ländergarantie

für HSH Nordbank am 30. April 2013“ (Statement before the Budget Committee of the

Parliament of the City of Hamburg in a Hearing Concerning the Increase of State Guar-

antees for HSH Nordbank), Max Planck Institute for Research on Collective Goods,

Bonn, http://www.coll.mpg.de/sites/www.coll.mpg.de/files/text/

Anhoerung_HSH_Nordbank.pdf

Hellwig, M.F. (2014), “Financial Stability, Monetary Policy, Banking Supervision and Cen-

tral Banking”, Paper presented at the First ECB Forum, Sintra, May 25-26, 2014,

mimeo, Max Planck Institute for Research on Collective Goods, Bonn, forthcoming in:

European Central Bank (ed.), Monetary Policy in a Changing Financial Landscape,

Proceedings of the First ECB Forum on Central Banking.

29

Herring, R. and J. Carmassi. 2010. “The Corporate Structure of International Financial

Conglomerates: Complexity and Its Implications for Safety and Soundness,” in: A.

Berger, P. Molyneux and J. Wilson (eds.) The Oxford Handbook of Banking, Oxford

University Press.

Hoshi, T., and Kashyap, A.K. (2004), “Japan’s financial crisis and economic stagnation”,

Journal of Economic Perspectives 18 (Winter), 3 -26.

Hoshi, T., and Kashyap, A.K. (2010), “Why did Japan stop growing?”, NBER working paper,

National Bureau of Economic Research, Cambridge, MA.

Miller, H.A., and M. Horowitz (2012), “A Better Solution is Needed for Failed Giants”,

New York Times, October 9, 2012.

Pisani-Ferry, J., A. Sapir, N. Véron, and G.B. Wolff (2012), What Kind of European Banking

Union?, Bruegel Policy Contribution 2012/12, Bruegel, Brussels.

Sapir, A., Hellwig, M.F., and Pagano, M. (2012), A contribution from the Chair and Vice-

Chairs of the Advisory Scientific Committee to the discussion on the European Com-

mission’s banking union proposals, Report No. 2 of the Advisory Scientific Committee

of the European Systemic Risk Board.

Wissenschaftlicher Beirat (2008), "Zur Finanzkrise", Brief des Wissenschaftlichen Beirats

beim Bundesministerium für Wirtschaft und Technologie an den Bundesminister für

Wirtschaft und Technologie Michael Glos vom 10. Oktober 2008 (“On the Financial

Crisis“, Letter of October 10, 2008 from the Academic Advisory Committee of the

Federal Ministry of Economic Affairs and Technology to the Federal Minister for

Economic Affairs and Technology Michael Glos),

http://www.bmwi.de/BMWi/Redaktion/PDF/Publikationen/brief-wissenschaftlicher-beirat-

finanzkrise,property=pdf,bereich=bmwi2012,sprache=de,rwb=true.pdf

Preprints of theMax Planck Institute for

Research on Collective GoodsBonn 2015/10

Financial Stability and Monetary Policy

Martin Hellwig

MAX PLANCK SOCIETY

Preprints of the Max Planck Institute for Research on Collective Goods Bonn 2015/10

Financial Stability and Monetary Policy

Martin Hellwig

April 1, 2015, revised August 1, 2015

Max Planck Institute for Research on Collective Goods, Kurt-Schumacher-Str. 10, D-53113 Bonn http://www.coll.mpg.de

1

Financial Stability and Monetary Policy1

April 1, 2015, revised August 1, 2015

Martin Hellwig

Abstract

The paper gives an overview over issues concerning the role of financial stability in monetary

policy. Historically, financial stability has figured highly among central banks’ objectives,

with policy measures ranging from interest rate stabilization to serving as a lender of the last

resort. With the ascent of macroeconomics, these traditional tasks of central banks have been

displaced by macroeconomic objectives, price stability, full employment, growth. The finan-

cial crisis has shifted the focus back to financial stability concerns. Along with these devel-

opments, the shift from a specie standard to a pure fiat money system has widened the scope

for central bank policies, which are no longer constrained by legal obligations attached to cen-

tral bank money.

The paper first surveys the evolution of financial-stability and macroeconomic-stability con-

cerns in central banking and monetary policy. Then it discusses two major challenges: (i)

What should be done to assess the relevance of financial stability concerns in any given situa-

tion? How should one deal with the fact that systemic interdependence takes multiple forms

and is changing all the time and that many contagion risks cannot be measured? (ii) What is

the relation between financial-stability and macroeconomic-stability objectives? To what ex-

tent do they coincide, to what extent are they in conflict? How should tradeoffs be handled

and what can be done to reduce the risk of the central bank’s succumbing to financial domi-

nance?

Key Words: Financial stability, monetary policy, systemic risk, central banking

JEL Classifications: E42, E44, E52, E58

1 Paper prepared for the Federal Reserve Bank of Atlanta Conference on Central Banking in the Shadows:

Monetary Policy and Financial Stability Postcrisis, March 30 – April 1, 2015. At the time when this paper was written, I was Vice-Chair of the Advisory Scientific Committee (ASC) of the European Systemic Risk Board (ESRB). The content of the paper does not in any way reflect the opinions of the ASC or the ESRB.

2

1. Introduction

In the fall of 2006, I participated in an evaluation of the work of the financial stability group

at the Swiss National Bank (SNB). A legislative reform of 2002 had given the SNB the man-

date “to contribute to financial stability” (in addition to ensuring price stability). Our assess-

ment was very positive, but we observed some few structural problems, namely:

– There was little interaction between the financial stability group and the macroeconom-ics/monetary policy group of the SNB.

– At the level of the Board, monetary policy and financial stability were discussed separate-ly, usually with just a short observation to the effect that there were no serious financial-stability concerns.

– Concerns of the financial stability group that the two Big Banks (UBS and Crédit Suisse) had extraordinarily high leverage did not translate into effective policy discussion.2

– There was little interaction between the financial stability group of the SNB and the Swiss Banking Commission, which was in charge of supervision.

At the time, we interpreted these observations as being indicative of difficulties involved in

getting people and institutions with different specializations to work with each other, especial-

ly when a new unit enters the game, and the financial-stability concerns of this unit do not

appear to be urgent. So we suggested that it would be (would have been) useful if the finan-

cial stability and the macroeconomics/monetary policy groups had a joint investigation of past

episodes where information about the financial system might have been relevant for monetary

policy. As examples, we gave developments of the 1980s and 1990s. For example, would it

have been possible to avoid the sharp inflation increase of 1989 if monetary policy in 1988

had taken account of the fact that changes in liquidity regulation and in interbank clearing

would raise the multipliers for money creation in the banking sector? Had there been early-

warning signals that were overlooked? Would it have been possible to avoid the low-growth

experience of the 1990s if monetary policy had taken account of the weakness of the banking

sector that was due to substantial losses in real-estate and business lending?3

What I have seen since then makes me believe that the problems we observed then appear

much more generally. By now, after the financial crisis of 2007-2009 and the “euro crisis”

that has been with us since 2010, the importance of financial stability concerns for central

banks has become obvious, but we still do not have a good idea of how to manage these con-

cerns along with traditional monetary policy.

We have a research agenda with questions, such as: What was the role of monetary policy in

the run-up to the financial crisis of 2008 and in the run-up to the ”euro crisis”? What becomes

2 Thus, Bichsel und Blum (2005) and Blum (2007) had warned that the Big Banks had used the model-

based approach to determining required equity in order to raise their leverage to the order of 50, and they had proposed the introduction of a leverage ratio regulation. In the crisis, their warnings proved to be jus-tified.

3 We also suggested a similar joint post-mortem exercise with the supervisor on the 2001 difficulties of Crédit Suisse.

3

of standard monetary policy when the central bank intervenes to prevent a liquidity crunch, or

even a complete meltdown of the financial system? To what extent and how should monetary

policy for the post-crisis low-growth economy take account of the weakness of the financial

sector? But we do not know enough to answer these questions with any degree of assurance.

The problems run more deeply than just bureaucratic or intellectual specialization and inertia.

They also involve issues of political economy, ideology, and power: What monetary or pru-

dential policy is feasible in good times, when nobody wants to hear about the risk of a crisis?

What becomes of the traditional monetarist hands-off policy stance when a financial crisis

looms and the central bank intervenes to deal it? And, most importantly, who is in charge?

Discussions about these questions sometimes involve differences in ideologies. But even

without differences in ideologies, different approaches often involve different modes of think-

ing: Traditional macroeconomic/monetary policy approaches do not have much room for fi-

nancial-stability analysis; traditional prudential analysis focuses on the individual institution,

rather than the system as a whole, let alone the macroeconomy. And at the level of objectives,

it is not clear whether we think of macroprudential policy as a policy designed to protect the

financial sector from the fallout of macroeconomic developments or whether we think of it as

a policy designed to protect the macroeconomy from the fallout of financial instability.

Two further examples are instructive: When reading the book Maestro, Bob Woodward’s ac-

count of Alan Greenspan’s 1980s and 1990s record as Chair of the Federal Reserve, I got the

impression that, over time, Greenspan was a master at intervening to counteract the damaging

effects of his previous interventions: After the stock market crash of 1987, he flooded the sys-

tem with liquidity in order to avert a financial crisis. When the expansion of the money supply

created strong inflationary pressures in 1988 and 1989, he sharply reversed the policy, raising

interest rates again to levels unseen since the early 1980s. When the fallout from this measure

threatened to bankrupt the big money center banks in 1990, he again reversed his policy,

sharply lowering short-term interest rates and allowing the banks to play the yield curve, us-

ing the very large difference between long-term and short-term interest rates to earn record

profits four years in a row and rebuild their equity. This stop-and-go routine seems to have

been driven by alternating concerns about financial stability and inflation, without much of an

attempt to take a comprehensive view.4

Currently, the European Central Bank (ECB) is engaging in a substantial program of mone-

tary expansion through open-market purchases. This program is intended to revive economic

growth from its current low, in some member states even negative, levels and to counteract

the threat of deflation and deflationary expectations. Little account seems to be taken of the

possibility that the observed weakness of the macroeconomy might be caused by weakness in

the financial sector and that open-market operations at the long end of the market might com-

4 Federal Reserve Policy in the years 2000 – 2008 exhibits a similar stop-and-go pattern, with the added

ingredient that, in the election year 2004, as in the election year 1972, interest rates were kept low for longer than can be explained by lags in problem recognition.

4

press the yield curve even more and thereby reduce the profitability of financial intermediar-

ies, making it difficult for them to rebuild their equity.

As yet, we do not have a conceptual framework for thinking about these issues, about the

proper specification of objectives, about the tradeoffs that are involved and about the risks

that need to be taken into account. In the following, I will try to lay out some of the issues that

need to be considered in developing such a framework. I cannot offer a blueprint, but at least I

will try to remove the impediments to clear thinking that are due to semantic ambiguities, lim-

itations imposed by modelling techniques, and, most importantly, an unwillingness to

acknowledge certain tradeoffs and certain facts. I begin with an overview over the respective

roles of financial-stability concerns and monetary policy in the past. Subsequently, in Section

3, I discuss some of the challenges that must be met.

2. Financial Stability and Monetary Policy: Taking Stock of the Past5

2.1 The Gold Standard Era

Originally, central banks were just banks, albeit banks with a strong connection to govern-

ments, serving as depositories for government funds and lending to their governments, work-

ing under special government charters and benefiting from government-granted privileges

such as the monopoly on issuing notes, sometimes with government mandates to stabilize in-

terest rates. 6 Under the gold standard, monetary policy involved the use of discount rates to

influence gold flows and the dynamics of the money supply.

As discussed by Goodhart (1988), central banks’ closeness to the government and their con-

trol over much of the specie in their countries contributed to their acquiring a central position

in the financial system. In particular, they came to act as banks for other banks, taking other

banks’ deposits and providing them with liquidity when needed.7

In the course of this development, central banks came to take on the role of a lender of the last

resort in times of crisis, when other banks could not obtain liquidity in the market. This role

was enshrined in Bagehot’s (1873) famous prescription that, in a crisis, the central bank

should be prepared to lend freely to solvent banks, against good collateral and at penalty rates.

As businesses, central banks could not just pursue whatever mandates they had. With funding

by notes that promised payment in specie, they needed to make sure to always maintain their

5 For an extensive account of the evolution of central banks, see Goodhart (1988). Hellwig (2015 a) draws

on Goodhart (1988) for background to a discussion of financial stability, monetary policy, and banking supervision in the European context.

6 Goodhart (1988, 114-122) mentions such a mandate for the Banque de France in particular. 7 Goodhart (1988) also points out that this development was helped by their not behaving like ordinary

profit-oriented banks and by their abstaining from competing in the other banks’ key markets, for lending to non-financial firms and for taking deposits from the public.

5

ability to fulfil their obligations. As for any other business, this requirement had a liquidity

dimension and a solvency dimension. The liquidity dimension required central banks to main-

tain gold reserves that would be sufficient to meet the demand for conversion of notes into

specie. The solvency dimension required them to avoid substantial losses that might endanger

their ability to pay, if not immediately, then over time.

However, acting as a lender of the last resort under Bagehot’s rule and lending to solvent

banks at high rates, could be very profitable. Profitability would also benefit if the support

given to the financial system actually managed to forestall a deeper crisis that might have

compromised the central bank’s other assets.

More generally though, the need to maintain the ability to fulfil their obligations could inter-

fere with the central banks’ stabilization functions. The problem is most obvious in the use of

high interest rates to protect or expand the central bank’s gold reserve. High interest rates

served to maintain the convertibility of the central bank’s notes, but they also did harm to fi-

nancial institutions, the overall economy, and the government budget. For example, the very

high discount rates used by the Reichsbank in the crisis of 1931 contributed materially to the

difficulties of the banking sector and to the subsequent large downturn in bank lending and

economic activity.8

Eichengreen’s (1992) account of the gold standard in the interwar years suggests that the

depth of the Great Depression is largely explained by a lack of international cooperation and

by the contractive monetary policies that different central banks used to maintain or even in-

crease their gold reserves. The desire to maintain convertibility took precedence over the

needs of the macroeconomy; moreover, those central banks that did try to ease their monetary

policies, experienced substantial outflows of gold that quickly led them to reconsider their

stance.

Similarly, the constraints imposed by the gold standard (and associated legal rules) prevented

the Reichsbank and the Federal Reserve from forestalling the banking crises of 1931 in Ger-

many and of 1931 and 1933 in the United States. Whatever financial stability concerns they

might have had were preempted by the rules of the gold standard. In the case of the Reichs-

bank, these rules continued to shape thinking and policy even after exchange controls had

been imposed, and the risk of a run on the Reichsbank had been eliminated.9

8 These high interest rates also enabled the Reichsbank to earn very high profits. These profits subsequently

enabled the Reichsbank to contribute substantially towards recapitalizing the private banks. See Born (1967, p. 170).

9 According to Eichengreen (1992), the leadership of the Reichsbank worried that any departure from the rules of gold standard might induce a loss of confidence and lead to unconscionable inflation. He shows that countries that had experienced serious inflations earlier on in the 1920s were late in abandoning the gold standard and even after they had done so continued to think about monetary policy in terms of the gold standard, with the result that they were the last to recover from the Depression. In this context, Ger-many is the exception that proves the rule: The change of leadership of the Reichsbank that was imposed by the Nazis in 1933 also led to a change of thinking and indeed prepared the road for the second hyperin-flation, this one from Nazi finance of rearmament and the war through the printing press.

6

2.2. Fiat Money and the Macroeconomic Orientation of Monetary Policy

For central banks, the Great Depression had two major effects: The gold standard disap-

peared, and thinking about monetary policy became an integral part of the newly developing

field of macroeconomics. The two changes were complementary; the macroeconomic orienta-

tion of monetary policy was made possible by the fact that central banks no longer had to

worry about the convertibility of their currencies into gold.

For the United States, the gold standard was effectively replaced by a pure fiat money system,

a paper currency that was not backed by specie or anything else.10 The Bretton-Woods system

of fixed exchange rates did not oblige the Federal Reserve to maintain the exchange rate.

Maintenance of exchange rates was left to the central banks of other countries, which were

thus operating under a dollar standard. For macroeconomists, the constraints that this obliga-

tion implied seemed inappropriate because they harmed or eliminated these central banks’

ability to pursue their own macroeconomic objectives.11 This assessment contributed to the

movement towards flexible exchange rates that culminated in the 1973 abolition of the Bret-

ton-Woods system. After that, most countries had pure fiat currencies, without any microeco-

nomic restrictions on the central banks.

The major debates about monetary policy in the second half of the twentieth century have all

focused on macroeconomics. What are the respective roles of monetary and fiscal policy in

stabilizing aggregate demand? What is the role of monetary policy in reducing unemploy-

ment? To what extent is monetary policy driven by the need to monetize government debt?

How can we avoid inflationary tendencies from fiscal dominance and establish monetary

dominance instead? All these questions involve macroeconomic indicators and macroeconom-

ic aggregates, rather than the details of what central banks do and how the details of what they

do feed into the macro-economy.

Without questioning the high quality of the debates, we should also see that they involved a

strong ideological element. The Great Depression had raised serious doubts about the viability

of a laissez-faire market economy, and the Keynesianism that triggered the development of

macroeconomics called for active government policies to promote macroeconomic activity

and growth. The monetarist counter-movement was strongly motivated by an aversion to gov-

ernment interventionism and a belief in the strength of a laissez-faire market economy. From

this perspective, monetary policy was seen as less harmful, because less discretionary than

fiscal policy, but monetary policy itself should also not be too activist. The seeming paradox

that, in the debate about the use of inflation to fight unemployment, the opposition to such use

of monetary policy came from “monetarists” such as Milton Friedman is easily resolved by

10 Often-heard references to the dollar’s being backed by the “full faith and credit of the American govern-

ment” should be understood as poetry rather than substantive statements. Holders of dollar bills issued by the Federal Reserve have no legal claims against the American government, except possibly to have old dollar bills replaced by new ones.

11 See, in particular, Fleming (1962), Mundell (1962).

7

observing that the term “monetarist” came from the earlier debates about fiscal versus mone-

tary policy and that the real dissent was about government activism.12

For someone opposed to government activism, the very existence of a central bank that issues

fiat money is a scandal. Money that can be produced at zero cost and yet has purchasing pow-

er provides the central bank – and the government – with the means to intervene at will. It

also provides a source of funding that is not subject to parliamentary control or market disci-

pline and that undermines the disciplinary role of the government budget constraint. This rea-

soning provides the rationale behind Hayek’s Denationalisation of Money (1976), a proposal

to introduce a system without a government-owned or government-licensed central bank, with

competing private banks issuing notes that are convertible into real resources as specified in

the underlying legal contract. Politically, Hayek’s proposal has never had a chance but some

of the underlying thinking is still relevant today.13

Political abuse of the power to print fiat money has of course been frequent. Such abuses

caused not only the many hyperinflations of the 1920s, but also the suppressed hyperinflation

in Nazi Germany, where the erosion of the purchasing power of money was hidden by the

imposition of price controls and rationing.

In the United States, in the 1940s, the Federal Reserve was mandated by the Treasury to mon-

etize government debt. The Treasury-Federal Reserve Accord of 1951 eliminated this man-

date and made the Federal Reserve independent of the Treasury, but then interest rate target-

ing provided a basis for continued monetization of government debt. Whenever interest rates

went up, the Federal Reserve was under strong political pressure to lower them. In the 1970s,

at last, this policy induced a Wicksellian spiral of inflation and artificially low real interest

rates that was only reined in when, under Chairman Volcker in 1979, the Federal Reserve

shifted to targeting the money supply.

In other countries, the freedom provided by the end of the Bretton-Woods regime was used to

get central banks to provide direct government funding, For example, in Italy, the Banca

d’Italia was legally bound to buy up all government debt issues that the private sector would

not buy. This use of the printing press for government funding was a major cause of the high

inflation in Italy in the 1970s.

The overall experience of the 1970s was sufficiently negative to provide political support for

a return to a more stability-oriented monetary policy. In Europe, this move took the form of

exchange rate arrangements in which central banks were effectively called upon to follow

12 On the role of monetary policy compare Friedman (1968) and Modigliani (1977). In this context, it is of

interest to observe that the asymmetric information “explanation” of the Phillips Curve in Lucas (1972) rests on the assumption that noise comes from the central bank. If the additional noise came from the real economy, the slope of the “Phillips Curve” in the theoretical analysis would be reversed. When I once asked Robert Lucas about this, I got the impression that having the noise come from the government ra-ther than the real economy reflected his views of how the world works.

13 See for example Vaubel (1985). Winkler (2015) gives an interpretation of Bundesbank President Weidmann’s opposition to Quantitative Easing by the ECB as based on Hayekian concerns.

8

German monetary policy, first the “snake” and then the Exchange Rate Mechanism. German

monetary policy was shaped by the Bundesbank, an independent institution with a mandate

solely to preserve price stability.14 With a strong commitment to this mandate, the Bundes-

bank repeatedly put the brakes on monetary policy even when a recession was already in

sight, as it believed that a tough monetary policy was needed to fight inflation and to bring the

other participants of the macroeconomic policy game, trade unions and governments, to their

senses.15

At the request of Germany, independence of the central bank and an exclusive focus on price

stability were also written into the Treaty for European Monetary Union.16 The exclusive fo-

cus on price stability stands in contrast to the mandate for maximum employment, stable pric-

es and moderate interest rate that the Humphrey-Hawkins Bill has given to the Federal Re-

serve in the United States.

2.3 Macroeconomic Monetary Policy and Financial Stability

In the macroeconomic debates about monetary policy, neither the problem of financial stabil-

ity nor the place of the central bank in the financial system played much of a role. These were

debates about monetary aggregates and interest rates, unemployment and inflation, and the

proper conditions for sustained economic growth.

Yet under the surface, the financial system did play a role after all. For example, how do we

assess whether monetary policy is expansionary? What is the role of monetary aggregates? If

we look at monetary aggregates, should we look at the monetary base, i.e. the money that has

been issued by the central bank? Or should we take account of the fact that, from the perspec-

tive of investors, some of the debt instruments that are issued by financial institutions are

close substitutes to central-bank money? If we do, which instruments should we consider as

money-like and which ones should we exclude?

In their monumental Monetary History of the United States, Friedman and Schwartz (1963)

focus on the monetary aggregate M1, the sum of central bank money and demand deposits

held by non-banks in the economy. They do so on the grounds that the relation between M1

14 Independence of the German central bank was originally imposed by the Allies in 1922 and became part

of the rules under the Dawes and Young plans. Independence was re-imposed by the Allies in 1948 and subsequently codified in the Bundesbank Law of 1957. Political support for this arrangement was based on the experience of the two hyperinflations, one in the early 1920s and one under the Nazis. However, without the European Monetary Union, it is not clear how much longer this support would have lasted. The independence of the Bundesbank was questioned by members of the outgoing Schmidt government in 1982 and again by members of the incoming Schröder government in 1998. By that time, however, the independence of the central bank had become part of the Maastricht Treaty and its status had been changed from a simple law to a provision of the constitution.

15 Restrictive monetary policy explains why the recessions of 1974, 1982, and 1992/3 were particularly pronounced in Germany. In each instance, the Bundesbank was fighting the high inflation of the day without taking account that the cycle was already turning.

16 At the time, academic research also suggested that independence of the central bank was conducive to price stability, see Alesina and Summers (1992), Grilli et al. (1991).

9

and nominal income (or aggregate demand) was very stable over the period they considered.

By choosing M1, rather than the monetary base, as their key monetary aggregate, they implic-

itly bring in the banking system, for M1 depends on the extent to which banks and their cus-

tomers create deposits.

In their analysis of the Great Depression, this dependence is crucial as Friedman and Schwartz

argue that monetary policy was contractionary and much exacerbated the economic decline.

Over the period 1929, M1 declined by 33 percent whereas the monetary base increased by 15

percent. The difference was due to financial instability, in particular, the various banking cri-

ses. These crises induced non-banks to move out of deposits and into cash. They also induced

banks to hold higher reserves for their deposits in order to have better protection against cus-

tomer withdrawals. The increase in the monetary base, i.e. in the money that was provided by

the Federal Reserve was insufficient to neutralize these contractionary effects.

As an analytical device, the use of M1 as a measure of monetary policy is problematic because

this measure is a compound of different components that depend on the behavior of commer-

cial banks and of non-banks as well as the central bank. As a rhetorical device, however, it

enabled Friedman and Schwartz to refute (Keynesian) claims about the ineffectiveness of

monetary policy in the Great Depression and to blame the Federal Reserve for having failed to

stabilize the “money stock”. In the grander scheme of their book, the focus on this monetary

aggregate provided them with a basis on which to criticize the practice of targeting interest

rates, which the Federal Reserve had used since the Treasury-Fed Accord of 1951.

But there is a paradox involved. Targeting M1 rather than interest rates is consistent with a

laissez-faire approach that wants to avoid the distortions of relative intertemporal prices that

might be implied by an interest rate policy of the central bank. At the same time, targeting M1

requires an interventionist monetary policy whenever the relation between M1 and the mone-

tary base is unstable, for example in a financial crisis. The criticism that, in the Great Depres-

sion, the Federal Reserve did not do enough to counteract the contraction of M1 translates into

a call for activist monetary policy at a time of financial crisis. It certainly was understood as

such by Chairman Bernanke in 2008.

From different sides of the political spectrum and the academic spectrum, the activism of cen-

tral bankers since 2007 has been much applauded and much criticized. Applause came from

those who would favor monetary activism anyway, as well as those who believed that this

was a special situation and that the central banks’ interventions prevented a disaster whose

damage would have rivalled that of the Great Depression.17

Criticism came from Hayekian fundamentalists appalled at the sheer power of an institution

that could create trillions of dollars of additional money at the stroke of a pen or, in this elec-

tronic age, a push of a few buttons. Criticism also came from monetarists who had forgotten

the account of the Great Depression in Friedman and Schwartz (1963), perhaps because such

17 See, for example, Eichengreen (2015)

10

a crisis had not been seen for decades, perhaps also because such forgetfulness was conven-

ient as a way of avoiding the tension between the call for an interventionist monetary policy

in a crisis and an overall laissez-faire approach to macroeconomic policy.

Ordinarily of course, this tension is irrelevant because changes in the monetary base go along

with changes in other monetary aggregates so that a policy recommendation of steady growth

in aggregates also calls for steady growth in the monetary base. Large increases in central

bank money may then be assumed to induce large increases in aggregate demand and large

increases in consumer prices. Many of the warnings that were put forward in recent years

about the inflationary impact of central bank policies expanding the monetary base have been

based on this narrative.18

In fact, the large increases in central bank money that we have seen since 2007 have not in-

duced similar increases in wider monetary aggregates or in consumer prices. For example,

from 2008 to 2013, the monetary base in the euro area doubled, but cumulative growth of M3,

the monetary aggregate that nowadays is considered most important, amounted to only 10

percent. Cumulative inflation also amounted only to 10 percent. The discrepancy between the

growth rates of the monetary base and of the wider monetary aggregates reflects a change in

the behavior of commercial banks. They increased their holdings of central bank money by a

lot. The breakdown of money markets in 2008 had taught them that central bank money was a

more reliable source of liquidity than interbank borrowing. In the US, this behavior change

was reinforced by the Federal Reserve’s paying interest on the commercial banks’ deposits.

A striking episode is provided by the ECB’s Long Term Refinancing Operation (LTRO) of

2011/2012. Under this operation banks received three-year loans at cheap rates from the ECB.

The monetary base increased by about a trillion euros, from two to three trillion. Yet, there

was no comparable increase in M3 or in consumer prices. The Bundesbank, forever concerned

about the inflationary consequences of monetary expansion, suggested that, if the risk of infla-

tion had not yet materialized, the reason must be that, as we know from Milton Friedman, lags

in monetary policy transmission are long and variable.19 By mid-2014, however, the expan-

sion of the monetary base had been completely reversed, again without any recognizable ef-

fects on wider aggregates and prices.20

To understand this episode, it is useful to recall that, in the second half of 2011, financial

markets and financial institutions in Europe were in turmoil. As the extent and the impact of

the upcoming haircut on Greek debt became known, concerns about the solvency of European

banks had caused market funding to erode. As banks sold assets to get cash, asset prices de-

clined, which exacerbated the solvency concerns. The downward spiral accelerated when, in

November 2011, the authorities imposed an increase in equity requirements, to take effect by

18 See, for example Bundesbank (2012), Sinn (2012, 193-196, 2013). 19 Bundesbank (2012), Weidmann (2014). 20 I have yet to see an acknowledgement of this reversal on the side of those who warned against inflation

from the expansion.

11

June 30, 2012, and they did so by fixing a ratio, rather than the absolute amount implied by

current holdings, thus contributing to the banks’ deleveraging – and to further declines in as-

set prices. The LTRO stopped this downward spiral as it gave banks and markets an assurance

of reliable and cheap funding over a substantial period of time. Subsequently, after market

confidence had returned – and market rates of interest had further declined – those banks that

had access to market funding were actually eager to repay their LTRO loans as market fund-

ing had become even cheaper.

The years since 2008 have made us appreciate again that the financial sector and, in particu-

lar, the banking system, plays an important role in the monetary transmission mechanism.

Whatever macroeconomic objectives a central bank may be called on to pursue, full employ-

ment, stable prices or stable growth, the impact of a central bank’s action on its ultimate ob-

jectives depends on how those actions translate into behavior in the financial sector and in the

real economy. In situations where the financial sector is impaired, the very pursuit of macroe-

conomic objectives may require the central bank to support financial stability, in particular of

course by preventing an implosion of the monetary system that might be associated with a

financial crisis.

One may doubt whether the monetary aggregates suggested by Friedman and Schwartz (1963)

and their followers are the right indicators for monetary policy and its potential macroeco-

nomic impact. For the Great Depression, the Friedman-Schwartz account has been comple-

mented, perhaps even superseded by Bernanke’s (1983, 1995) suggestion that bank defaults

and bank closures had such disastrous effects in the Great Depression because they destroyed

the information capital that banks had accumulated about their borrowers and thereby dam-

aged the provision of credit to nonfinancial companies. However, if we see the credit channel,

rather than the “money stock”, as the key to the transmission of monetary policy to the nonfi-

nancial sector of the economy, the conclusion is the same, namely that banks are an important

part of the system and that banking crises can have a strong negative impact on monetary

transmission and the overall macroeconomy.21

2.4 The Central Bank as a Lender of the Last Resort in a Fiat Money System

By now, we have come full circle from Bagehot’s discussion of the central bank as a lender of

the last resort and the breakdown of this role in the Great Depression to a macroeconomic

version of monetary policy with a focus on price stability, inflation, and perhaps full employ-

ment and back to a concern for financial stability, now on the presumption that, financial sta-

bility is a prerequisite for macroeconomic stability or at least the reliability of the monetary

transmission mechanism. In the years since 2007, central banks have repeatedly taken on the

role of a lender of the last resort, for banks and, at least indirectly, also for governments.

21 On the credit channel, see Bernanke (1995) and Bernanke, Gertler, Gilchrist (1999).

12

However, central bank support for the financial sector in the crisis departed from Bagehot’s

rule in many respects. Whereas central banks lent freely, they did so to banks of dubious sol-

vency as well as banks that were clearly solvent; moreover, they lowered quality standards for

collateral, and they did not charge penalty rates. For example, the ECB’s LTRO was most

beneficial for banks with weak equity positions in periphery countries and allowed these

banks to borrow at rates far below those that they would have had to pay in the market (if they

were able to get market funding at all).22 What are we to make of this disregard of Bagehot’s

rule?

Bagehot’s rule serves three purposes: First, it protects the central bank from losses. Second, it

minimizes moral hazard on the side of commercial banks. Third, it serves to avoid bailouts of

insolvent banks. All these purposes are important but the question is whether we shouldn’t

also consider the benefits of central-bank intervention and the tradeoffs that may result.

Bagehot was writing about a central bank that was operating under the gold standard and

needed to maintain the convertibility of its notes. A central bank that issues fiat money need

not worry about the possibility of default. Could it be that, in a paper money economy, some

departure from Bagehot’s rule is actually desirable?

From a Hayekian perspective, merely to ask the question is anathema, promoting an abuse

that is made possible by the government’s usurping a role in money creation and using its

power to impose a kind of money that is not even backed by anything substantial. In this

view, the use of fiat money to support an impaired banking system distorts market outcomes

and keeps banks alive that really are insolvent and should be eliminated from the system. Im-

plicitly, seigniorage from the creation of fiat money is used to provide illicit subsidies to fi-

nancial institutions.

Hayekian fundamentalism rests on the assumption that central banking as such is only based

on government fiat and serves no economic purpose that is independent of the government’s

interests. As shown by Goodhart (1988), this view does not match the historical evolution of

central banks. The role of a lender of the last resort did not initially come about by govern-

ment fiat but resulted from the inability of the private sector to deal with collective-action

problems in crises, in particular liquidity crises. In crisis situations, competitive interests of

individual institutions usually stood in the way of collective arrangements that would have

allowed the industry to smooth over the crisis. The central bank could step in and help be-

cause it was not driven by such competitive concerns, e.g. the aim to take over a failing

bank’s business.23

22 On this point, see Acharya and Steffen (2015). 23 In a comment on Vaubel (1985), Hellwig (1985) points out that fiat money also fulfils an allocative pur-

pose, namely, it serves as a store of value that can be traded back and forth between participants without anybody’s having to hold real assets. The point is most obvious in the economizing on specie that be-comes feasible as the economy moves e.g. from a gold standard to a paper currency. Because of Pareto-relevant pecuniary externalities, such an outcome may not be obtained by laissez-faire competition. In-

13

Given the benefits from such interventions, one must address the tradeoff between these bene-

fits and the concerns behind Bagehot’s rule. An example may illustrate the issue: In 1931, the

German banking crisis began with large withdrawals from Danat Bank and Dresdner Bank,

two banks that were known to be subject to substantial risk from bad loans. After the bank-

ruptcy of Nordwolle, a textile company to which Danat Bank was greatly exposed, the with-

drawals turned into a run. The Reichsbank provided the banks with liquidity through the dis-

count mechanism even after Danat bank had exhausted its discountable material and even af-

ter it had become clear that Danat Bank might be insolvent. On July 9, 1931, this support had

to be stopped when the Reichsbank itself ran afoul of rules for the backing of its currency by

specie reserves. On July 13, Danat Bank closed its doors, and there was a general run on all

German banks. To cope with the situation, the government imposed a bank holiday, which

was followed by three weeks in which the payment system was impaired. In the six months

that followed, bank lending imploded, GDP declined from ca. 80 percent to ca. 60 percent of

its pre-crisis level, and employment declined by another 2 million people.

In this example, developments after the banking crisis were so bad that it would probably

have been better if the Reichsbank had been able to continue its support for the banks, includ-

ing Danat Bank – in violation of Bagehot’s rule. At the time, this was not possible because the

Reichsbank was constrained by legal rules for the backing of its currency. Central banks today

do not have to satisfy such rules.

In terms of the ultimate macroeconomic objectives of central bank policies, it seems clear

that, if a catastrophic crisis is looming, then the ability of the central bank to create fiat money

at will should be used even if the support benefits banks of dubious solvency as well as banks

that are clearly solvent but are caught up in the general crisis. Bagehot’s criteria should not be

treated as objectives in and of themselves but as means to improving on the central bank’s

ultimate objectives, which means that they should be abandoned in situations where their ap-

plication would be harmful for those ultimate objectives.

Some criticisms of central bank interventions in the financial crisis of 2007 – 2009 and again

in the euro crisis have focused on the possibility that these interventions might cause losses

for the central banks. Warnings of such losses have also figured prominently in the legal and

political disputes about the ECB’s open-market purchases of government debt. If the securi-

ties purchased lose in value, the argument goes, say because a government defaults, the loss

will ultimately be borne by taxpayers. Even if there is no need to recapitalize the central bank,

any losses on assets acquired by the central bank will detract from the profit distributions that

can be made to the central banks’ owners, i.e., ultimately the taxpayers.24

Such warnings reflect a profound misunderstanding of the wealth effects of money creation in

a fiat money system. Using newly produced fiat money to acquire risky assets provides a

deed, as suggested by Friedman (1969), a tax-financed subsidy to money holding may be needed to im-plement an efficient allocation.

24 Bundesbank 2012, Sinn 2013, Weidmann 2014.

14

windfall to the central bank and, ultimately, its owners. The windfall does not appear on the

central bank’s balance sheet because the new money issue is listed as a liability. However, if

one looks at the expected discounted present value of cash flows associated with the different

positions, it is clear that the expected discounted present value of cash flows from the ac-

quired assets is positive and the expected discounted present value of cash flows (“debt” ser-

vice) associated with the newly issued money is zero. Any losses that are subsequently made

on the acquired assets detract from the windfall but cannot make the net return from the asset

acquisition negative.25

The real victims of such interventions would not be taxpayers but the holders of money and of

assets denominated in units of money because the additional money may have an inflationary

impact. Any intervention to support the financial system must be balanced against the other

objectives of the central bank, in particular the objective of price stability. However, in a true

crisis situation, price stability is usually not much of an issue. If financial institutions are weak

and endangered, lending and investment are likely to be constrained, aggregate demand is

likely to be weak, and consumer prices are likely to be subject to deflationary rather than in-

flationary pressures.

2.5 Interim Summary

The results of the discussion so far can briefly be summarized as follows.

– Since the nineteenth century, concerns for financial stability have been on the agenda of central banks, partly because private arrangements did not work, partly because acting as a lender of the last resort could be profitable, and partly because central banks had a sta-bilization mandate.

– However, under the gold standard, and later the Bretton-Woods system of fixed exchange rates, such stability concerns had to be subordinated to the need to maintain convertibility or the need to support the exchange rate.

– Since the Great Depression, monetary policy has been governed by macroeconomic man-dates, for price stability, full employment, stable growth, or stable interest rates, which on the surface have nothing to do with financial stability.

– Some financial stability concerns are, however, implicit in macroeconomic mandates be-cause the impact of monetary policy on the macroeconomy depends on the functioning of the financial system, in particular private money creation and lending by the banking sec-tor.

– The scope that central banks have to deal with stability issues, macroeconomic or finan-cial, has been greatly increased by the shift to a pure fiat money, first in abandoning the gold standard and, later in abandoning the Bretton-Woods system of fixed exchange rates.

– Shifting to a pure fiat money has also raised concerns about the unrestricted power of the central bank (and possibly the government that stands behind it) to appropriate resources

25 Hellwig (2015 b). The locus classicus for this discussion is of course Patinkin (1965).

15

merely by printing money. Such concerns have motivated the search for rules that would constrain the use and prevent the abuse of this power.

– However, any mandates for macroeconomic or financial stability may require discretion-ary interventions, in particular discretionary interventions in a crisis. From a welfare per-spective, such interventions are desirable if the prospective damage from leaving the sys-tem alone is very large.

3. Challenges

The preceding discussion does not say anything about how monetary policy should take ac-

count of financial stability. The examples given in the introduction show some of the difficul-

ties involved. As yet, we do not have a clear understanding of how these difficulties should be

dealt with. In this second half of the paper, I will discuss two major challenges:

– Understanding what is going on. – Clarifying objectives and strategies and avoiding financial dominance.

3.1 Understanding What is Going On

Monetary and supervisory authorities must develop an understanding of what is going on in

the financial system and what are the implications of ongoing developments for monetary pol-

icy and financial stability. This is not a once-for-all exercise but a challenge that must be met

over and over again as the financial sector is forever changing.

In 2008 and since, the question has often been raised why economists did not see the crisis

coming. Some, like Rajan and Shiller, have become heroes because in 2005 and 2006 already,

they warned of risk developing in US real-estate markets. However, even these heroes did not

warn of the follow-on effects that the real-estate crisis would have.

Indeed, these follow-on effects were not even recognized when the real-estate crisis was well

underway and understood. The IMF’s Global Financial Stability Report of April 2007 con-

tains a very good analysis of the real-estate and subprime-mortgage crisis but concludes with

the assessment that these developments in this relatively small segment of the global financial

system were unlikely to cause major harm for the worldwide economy. In June 2007, the An-

nual Report of the Bank for International Settlements (BIS) echoed this analysis and sent the

same message.

Monetary policy reflected this (lack of) understanding. Short-term interest rates in the United

States were kept high until the systemic nature of the crisis had become clear, in August 2007,

a year after real-estate markets had begun to turn down and the system of mortgage securitiza-

tion had begun to sputter. The ECB actually raised its interest rates in the summer of 2008

when the systemic crisis was well under way, and it did so again in the summer of 2011 even

though at that time the euro area was again hit by a systemic crisis.

16

I am not mentioning these past failures because I want to put blame on the institutions in-

volved but because it is important to understand the reasons for these failures. Regardless of

what the strategy for monetary policy may be, it is important that the central bank should un-

derstand what is going on. To get to that point, we must reflect on what went wrong in the

past.

I see three reasons for the lack of understanding in the run-up to the financial crisis:26

– Observers did not think in systemic terms. The assessment that subprime lending was only a small segment of the global financial system was quite correct. Indeed, the as-sessment of losses that was given at the peak of the crisis in the fall of 2008 was not larg-er than the losses that had been recorded in the Japanese crisis of the 1990s or the losses that had been feared in the S&L crisis in the United States in the early 1990s.27 The sub-prime crisis differed from these earlier crises in that the mortgages had been securitized and the mortgage-backed securities (MBS) and collateralized debt obligations (CDOs) were held by financial institutions worldwide, usually in the trading book, which required immediate writedowns as market prices declined, and usually with very little equity.

– The scope of contagion risk was unknown. About $ 1000 billion of MBS and CDOs were held in special purpose vehicles (SPVs) of regulated financial institutions, conduits and structured-investment vehicles (SIVs). When MBS and CDOs were downgraded in Au-gust 2007, funding for these vehicles evaporated, the sponsoring banks had to take these securities into their own books, and found that they had too little equity. The resulting deleveraging process contributed to the downward dynamics of asset prices in the year that followed. In August 2007, the overall size of the positions held in SPVs came as a surprise. The practice had been known but nobody appreciated its scope, let alone the systemic impact that a breakdown of funding for these SPVs would have. From what I have come to understand, the numbers would actually have been available but they were not collected and analyzed.

– The relation between monetary policy and the financial system had not been thought through. This criticism applies in the run-up to the crisis, when the looseness of monetary policy greatly encouraged the subprime-lending boom.28 It also applies to the central banks’ initial reactions in 2007 and 2008. In August 2007 and later, they understood the immediate liquidity concerns but they failed to recognize the extent of the crisis until well into 2008. In the case of the ECB, the delay in understanding was enhanced by the notion

26 For an extensive account of systemic risk in the crisis, see Hellwig (2009). 27 The IMF’s Global Financial Stability Report of October 2008 gives a number of $ 500 billion for losses

from subprime-related securities (on a total volume of $ 1200 billion of subprime mortgages that had been securitized). The IMF’s total estimate of losses in that crisis is much higher, but the difference is due to systemic follow-on effects that had not been considered initially.

28 In 2003/2004, when the Federal Funds rate stood at roughly 1.5 percent, the margin between (fixed-rate) subprime mortgages and money market borrowing amounted to 500 to 600 basis points. In Hellwig (2009), I had suggested that investors were eager to go into mortgage-related securities, MBS and CDOs, because they wanted to benefit from this huge spread. As shown by Acharya, Schnabl and Suarez (2013), however, the SPVs that invested in such securities with funding from the money market only earned about 10 to 30 basis points. The remainder seems to have gone to the mortgage banks, investment banks, rating agencies, and law firms that engineered the securitization. This observation may explain why risky lending to homeowners developed so differently from risky lending to corporations; see Demyanyk and Van Hemert (2009).

17

that you can use interest rate policy to pursue price stability and unorthodox measures, such as changing the rules for lending to banks, to support the financial system.29

Since the crisis, we have all learnt that we must think in systemic terms. In the European Un-

ion, the European Systemic Risk Board (ESRB) has the task to do precisely that and to pro-

vide the relevant authorities with appropriate warnings and recommendations when systemic

risks loom.30 However, understanding systemic risk is more easily said than done. If the task

is approached with the traditional tools of either central bankers or microprudential supervi-

sors, there is a danger of falling into a routine of ticking off items on a dash board without

seeing what is actually going on. We must recognize that systemic risk transcends the scope

of macroeconomic modelling as well as the supervisors’ assessments of individual institu-

tions.

There are many mechanisms of contagion, and their number is probably growing. At this

point, I can think of the following mechanisms:31

– Contractual dominos: The Lehman Brother bankruptcy imposed a loss on the Reserve Primary money market mutual fund and caused it to “break the buck”.

– Disappearance of contracting opportunities: When Lehman Brothers went bankrupt, it ceased to act as a market maker for many derivatives that other participants had been counting on for their risk management. When Reserve Primary broke the buck, investors whom it had counted on to fund its operations withdrew their funding. When investor withdrawals forced money market funds to reduce their activities, banks such as Dexia or Hypo Real Estate that had relied on the money market to fund the assets needed for the excess coverage for their covered-bond issues, had severe liquidity problems.

– Information contagion: When Reserve Primary broke the buck, investors inferred that other money market funds might also have problems so those funds also suffered runs. When Lehman Brothers was not bailed out, investors inferred that government support of banks could not be taken for granted and withdrew wholesale money market funding.

– Fire sale externalities: If financial institutions react to problems by selling assets, they exert downward pressure on the prices of those assets. Price declines force those institu-tions that also hold these assets to take writedowns, at least if they hold the assets in their trading books, and to register losses, which may force them to take defensive actions as well. Between August 2007 and September 2008, we saw such effects driven by delever-aging in response to a lack of equity; in September and October 2008, the process was driven by a scramble for cash and became torrential.

– Credit Crunch Contagion: If banks in difficulties reduce their lending to the real econo-my, investment and aggregate demand are negatively affected. This development in turn can feed back into nonfinancial firms’ debt service to banks and into bank profits.

29 See Hellwig (2015a). 30 From May 1, 2011 to April 30, 2015, the author has been first Chair and then Vice Chair of the Advisory

Scientific Committee and in this function was also a member of the General Board of the ESRB. 31 For a systematic discussion, see Hellwig (2014 a).

18

In practices, these different mechanisms are likely to appear at the same time, with interac-

tions that very much magnify the systemic effects. For example, the Lehman bankruptcy had

a direct domino effect on the Reserve Primary money market fund. Reserve Primary’s losses

triggered two interdependent runs, the run of money market fund shareholders on the money

market funds and the run of the money market funds on the banks, in particular, US invest-

ment banks. The breakdown of money market funding caused a scramble for cash by banks,

with fire sales of assets that much depressed asset prices, inducing stock market losses on the

order of twenty trillion dollars in the course of a few weeks. As a metaphor for this process,

the term “system meltdown” is very apt.

The overall complexity of the potential interactions between the different contagion effects is

awesome. Any notion that we can predict these interactions in advance is illusory. The fol-

lowing observations elucidate some of the difficulties involved in analyzing systemic-risk:

– Different processes occur on different time scales. Standard macroeconomic analysis works with a notion of a “period” as referring to a quarter or even a year. Asset market adjustments, however, occur in real time, so fast that there is no scope for relying on flows of new profits or new savings to balance disequilibria at the level of stock varia-bles. So far, the problems associated with differences in time scales have prevented mac-roeconomists from properly integrating financial developments into their quantitative models.

– Some of the contagion effects are highly contingent. For example, a bank’s fire sales will have different effects depending on whether the other market participants are jittery or exuberant. In the LTCM episode in 1998, fire sale externalities were feared to be extraor-dinary because markets were so jittery. This contingency, i.e., markets being jittery and reacting dramatically to asset liquidations, can hardly be assessed with any degree of reli-ability ex ante.

– Patterns of risk allocation in the financial sector change over time and so do the mecha-nisms by which systemic risks may realize. Many of the banking crises of the early 1980s and the early 1990s were due to different institutions’ being similarly exposed to certain macro shocks, in particular the fallout from interest rate increases. By 2007, direct expo-sures to such macro shocks were limited because, after the crises of the 1990s and with the advent of Basel II, banks had begun to hedge these risks to get them off their books. System risk then was hidden in correlations, correlations between underlying risks and counterparty credit risks in hedge contracts, or correlations between the jitteriness of markets and the institutions’ risk exposures.

– If systemic effects are hidden in correlations, the individual institutions’ balance sheets and risk models do not reveal the systemic risks. There may in fact be no data for such an assessment at all.

– If data are available at all, the data series are short and nonstationary. Moreover, many of the risks change in endogenous ways. For example, the counterparty risk of a credit de-fault swap with AIG depends on how many other such contracts AIG has written and how the risks for which these contracts provide insurance are correlated.

19

These difficulties should not, however, be taken as an indication that there is nothing to be

done. They should, rather, be taken as an indication that we need to think again about the

questions we ask in order to understand what is going on. Going back to the years 2004-2007,

here are a few questions that might have been asked but were not:

– Theory predicts that institutions with long-term liabilities, such as insurance companies and pension funds, should have a comparative advantage in holding long-term securi-ties.32 Why then did such a large part of the output of the securitization industry find its way into the portfolios of banks and their off-balance-sheet vehicles, rather than insur-ance companies and pension funds?

– Theory predicts that there may be negative incentive effects from securitization unless the originating or the securitizing banks hold on to the equity tranches.33 Why then were so many equity tranches traded? Or: Was there any substitute for the disciplining effects as-sociated with holding on to the equity tranches?34

– What were the causes and what were the potential risks of the growth of short-term wholesale funding of banks? What were the causes and what were the potential risks in-volved in the growing dependence of banking institutions on money market funds?

– What were the causes and what were the potential risks of the decline in equity funding of banks?35

– With an intermediation chain leading from final investors to money market funds, from money market funds to structured-investment vehicles holding CDO’s, from structured-investment vehicles holding CDOs to SPVs holding MBS and creating CDOs, from SPVs holding MBS to SPVs holding mortgages and creating MBS. From SPVs holding mort-gages to mortgage banks and from mortgage banks to homeowners, who was bearing the risks associated with maturity transformation, liquidity transformation and lending? To what extent might participants be fooling themselves as the complexity of the chain served to hide those risks?36

If these questions had been posed and answered in 2005, we might have obtained an idea

about the actual system risk exposure much sooner than we actually did.

32 For an extensive discussion, see Hellwig (2009). 33 See Hellwig (1994 a, 1998). 34 Before, the early 2000s, this seems to have been the role of the guarantees that Fannie Mae and Freddie

Mac, the government-sponsored enterprises, provided, along with the quality standards that they imposed for the securitization of prime mortgages. When the investment banks entered the business, focusing on subprime and not providing any guarantees, these safeguards disappeared. For an extensive discussion, see Hellwig (2009).

35 For Europe, ASC (2014) shows that, just in the years 1998 – 2007, there was a substantial further decline in equity funding of the largest banks. Whereas in 1998, equity below 4 percent of total assets was the ex-ception (and the two banks involved needed government support in the crisis), by 2007, equity above 4 percent of total assets was the exception, and some of the major banks had equity below two percent of total assets.

36 For an example of misperceptions about risk, see Gorton (2010). Gorton claims that subprime mortgages did not involve significant maturity transformation because the contract was bound to be renegotiated af-ter a short period of time. This assessment neglects the correlation between interest rate movements and defaults resulting from the fact that the borrower might be unable to pay higher rates of interest. For a discussion of the problem see Hellwig (1994 a,b, 1995).

20

Methodologically, the approach taken in these questions is not to start from a fixed model, but

to look at new developments, in particular developments that run counter to theoretical intui-

tion and to try and find out what is the underlying story. In other areas of applied economics,

e.g. in competition policy, we take it for granted that there is no one-size-fits-all model and

that we first have to find out, from the little material we have, which model or set of models

might be relevant or, more crudely, what is the story behind what we see.

In the practice of the different authorities, some of the studies of the IMF and the BIS come

close to what I am suggesting. The newly created macroprudential authorities should engage

in this kind of work. However, so far, much of what they do is too mechanical, focusing on

various indices, e.g. for credit growth or for developments in real-estate markets, and on cer-

tain risks to individual institutions, e.g. risks from foreign currency lending to non-financial

borrowers or from large-scale short-term funding. As yet, I have not seen much on potential

second-round systemic effects. Nor have I seen anything on the question how different devel-

opments fit together and where the unseen risks might be hidden.

Development of a capacity and a routine for such work is important. It is also important that

such analyses should be done without dependence on the powers that be in central banks or

supervisory authorities, powers that may have too much of an interest in making sure that the

results of the analysis conform to their policy choices or to the regulatory instruments over

which they have control. Before one can sensibly talk about the appropriate policy or the ap-

propriate mix of instruments, one first needs to understand what is going on. In situations

where policy makers and supervisors are tempted to be complacent, the unit or institution that

is charged with the analysis of systemic risk and more generally risk to financial stability

ought to be given a role of devil’s advocate, questioning the assumptions behind the compla-

cency in order to enable an early recognition of problems.

Much of the preceding discussion has focused on systemic risk in a crisis. The basic idea,

however, applies in normal times as well, in particular in the context of what is the role of the

financial system in the transmission mechanism for monetary policy. Three examples may

illustrate the point:

In the 1970s, institutional and technological innovations reduced the demand for money in the

sense of M1, so that the stable relation between M1 and aggregate demand that Friedman and

Schwartz (1963) had postulated was no longer observed. Among the reasons were sophisticat-

ed cash management systems that firms used to economize on cash, as well as investors shift-

ing out of demand deposits with banks and into money market funds, which offered close

substitutes for demand deposits but were counted in M1. The Federal Reserve’s slowness in

appreciating this development contributed to the inflation experience of those years.

The Swiss examples mentioned in the introduction are also pertinent. Surprisingly large infla-

tionary pressures in 1988/89 were due to the fact that the Swiss National Bank had underesti-

mated the implications of changes in the banking system for the impact of changes in the

21

monetary base on money creation by commercial banks. In the 1990s, the Swiss National

Bank’s underestimating the effects of its restrictive monetary policy on economic growth

seems to have been related to the failure to recognize the macroeconomic impact of the weak-

ness of commercial banks after the interest hike of 1990 and the problems in real-estate and

SME loan performance.

Turning to the current situation, I wonder to what extent the weakness of lending and growth

that we observe in many countries might be a reflection of debt overhang and financial-sector

weakness. Since the crisis, private households, nonfinancial companies and governments have

had substantial debt overhang, debt that was incurred with expectations of growth that failed

to materialize and that was not written down through bankruptcy, resolution, or restructuring

procedures. Meanwhile financial institutions have somewhat reduced their leverage but re-

main weak because of remaining risks from the crisis and of low profitability of ongoing ac-

tivities.37 If this assessment were correct, the current low interest rates, as well as devices that

flatten the yield curve, such as forward guidance and open-market purchases of long-term

securities, might delay the recovery.

3.2 Clarifying Objectives and Strategies, Avoiding Financial Dominance

Central banks need to be clear about their objectives and about their strategies. With multiple

objectives, there is a question about (possibly contingent) priorities and tradeoffs and about

accountability.

The side-by-side of financial-stability and macroeconomic concerns leaves open the question

of which of them should have priority. In some situations, the question is moot because there

is no conflict. For example, a central-bank intervention that forestalls an acute financial crisis

will also serve macroeconomic objectives such as protection of the monetary transmission

mechanism, full employment, or price stability (in this case, avoidance of deflation). The

ECB, which does not have an explicit financial-stability mandate, justified its interventions in

2008 and again in 2011 and 2012 by referring to the need to avoid deflation and the need to

maintain the viability of the monetary system. These arguments make sense but, if they are

taken literally, they can become dangerous.

Macroeconomic and financial-stability objectives are sometimes aligned and sometimes in

conflict. The two sets of objectives are aligned if the financial system is on the verge of an

acute crisis that can cause a lot of damage and pull down the entire economy. They are also

aligned if aggregate demand is rapidly expanding and the financial system is increasing its

risk exposures; in this case, monetary-policy measures and macroprudential measures that

slow down the expansion can serve both sets of objectives, except that a slowdown may be

37 In their study of European banks, Acharya and Steffen (2015) observe that the weaker a bank’s equity

position is, the more likely it is to rely on ECB funding and the more likely it is to use this funding to lend to its sovereign or to invest in securities markets, rather than lend to nonfinancial companies.

22

unpopular, so central banks and macroprudential authorities may not want to take these

measures.

The two sets of objectives can be in conflict, however, if the economy and the financial sys-

tem are both weak, but no acute crisis is looming. In this case, the pursuit of macroeconomic

objectives might call for an expansion of bank lending so as to stimulate investment and ag-

gregate demand. However, avoiding banks’ getting more deeply into trouble might require the

very opposite, a defensive strategy with reduced lending, possibly also with high margins that

might enable the banks to earn profits enabling them to repair their balance sheets. In such

situations, there is a need for setting priorities; this is particularly important in continental Eu-

rope where banks are central to the financial system and monetary policy works primarily

through its effects on bank lending.

In raising this issue in discussions with policy makers in positions of responsibility, I have

received four answers:

– We must respect the financial sector’s own needs and must not use macroprudential or other instruments to turn the financial sector into a tool of macroeconomic stabilization.

– We must do precisely that, i.e. turn the financial sector into a tool for macroeconomic stabilization, especially in the euro area, where monetary policy applies to all countries alike, and macroprudential regulation is the only means for dealing with country-specific problems.

– There is no conflict because, if the economy is doing badly, the financial sector will be doing badly as well.

– There is no conflict because, if the problems in the financial sector are not cleaned up, the economy will never recover.

These statements are mutually incompatible. Having them side by side makes clear that the

problem needs more thorough thinking. There are in fact two issues: First, should the finan-

cial sector be treated as a tool for macroeconomic stabilization? Second, how should we deal

with the fact that economic activity and financial-sector health depend on each other?

On the first issue, we should recognize that, if we induce financial institutions to take risky

actions because we hope that such actions may contribute to stimulating the macroeconomy,

then we must also be prepared to support those institutions if the risks turn out badly. The

ECB’s Targeted Long-Term Financing Operation (TLTRO), announced in the fall of 2014,

offers cheap long-term loans to commercial banks on condition that the funds from these

loans are used for lending to the real economy (rather than lending to sovereigns or investing

in securities markets, as happened with the funds from the 2011/12 LTRO).38 In fact, com-

mercial banks have not been eager to make use of this offer, perhaps because they do not need

the funding, perhaps because they see little scope for earning suitable returns, including risk

premia, from expanding their lending to the real economy. Suppose however, that they did

38 See Acharya and Steffen (2015).

23

take up the ECB’s offer, and that, despite the increase in lending, the real economy continued

to go down, and those loans went sour. Wouldn’t the ECB have some sort of co-responsibility

for the damage? And wouldn’t the ECB feel morally compelled to help the commercial banks

in overcoming the difficulties from the adverse developments?

Such an arrangement can be dangerous. Not so much because, in the end, the need to support

banks from the fallout of risks taken under the influence of the central bank is in conflict with

the central bank’s macroeconomic objectives; quite likely, the fallout from the risks has a

macroeconomic dimension because the macroeconomy itself is doing badly. The real danger

comes from the moral hazard that arises if commercial banks get a sense that the central bank

will always support them. The proverbial “Greenspan put” is deemed to have contributed to

the financial crisis by fostering such moral hazard. The emergence of the term “Draghi put” in

connection with the ECB’s interventions in 2012 raises the question whether the bailout expe-

riences of the crisis haven’t reinforced the moral hazard effect.

The underlying issue is one of governance. In a market economy, participants are free to do

what they want, but they must bear the consequences. Governments and central banks can set

the rules, in particular, the rules of prudential regulation, but they should not impose any par-

ticular business strategies. If central banks interfere with the commercial banks’ choices of

business strategies, providing the banks with an explicit or implicit promise of support if

things go wrong, the two sets of institutions may develop a give-and-take routine that defies

any notion of accountability and responsibility. In this symbiosis, the commercial banks may

be asked to help the macroeconomy and in turn be subsidized by implicit guarantees. Moral

hazard effects of such an arrangement may well become uncontrollable.

Turning to the interdependence of economic activity and financial sector health, the question

of how to deal with a situation when both are weak should really be treated as a question

about sequencing. Should we first clean up the financial sector, hoping that once this is done,

prospects for economic recovery will be greatly improved? Or should we delay a financial-

sector cleanup, hoping that, as banks are kept going, their continued lending will soften the

downturn in the real economy? An immediate cleanup of the financial system may sharpen

the current recession but bears the risk that a full recovery may be much delayed.

When faced with this dilemma, most authorities tend to go for delaying the cleanup. Forbear-

ance seems to offer some hope that financial-sector problems might disappear on their own

and that the current credit crunch and recession will be softened. However, the empirical liter-

ature on the subject suggests that such softening of the current credit crunch and recession

through forbearance towards banks may have a large cost in the form of an even bigger credit

crunch and recession in the future.39

39 For a systematic discussion, see ASC (2012) and the literature cited there. In discussions about the sub-

ject, the Latin American Debt Crisis of the 1980s is often cited as a counterexample, the argument being that in 1982 many of the major banks in the US and Europe may well have been insolvent and the system was better served by delaying the necessary writedowns until around 1990, by which time the banks were

24

The two alternatives may loosely be described as the Swedish and the Japanese strategies for

dealing with the crises of 1992. The Swedish authorities intervened promptly to clean up their

banking system; the cost was a very sharp recession, the benefit a very quick recovery. In

contrast, the Japanese authorities’ forbearance towards their banks has contributed greatly to

the low growth of the Japanese economy over the past two decades. For example, forbearance

of the authorities towards the banks involved allowing the banks themselves to exercise for-

bearance towards their loan customers so as to avoid writedowns on those loans that might

show the banks to be insolvent. Such forbearance towards loan customers however, possibly

with a provision of further funds that these customers needed for payments, contributed to low

growth because it made entry of new, more innovative firms more difficult as the incumbents

with whom these firms would be competing were supported by the banks.40

Thinking about the problem is complicated by the fact that financial-sector weaknesses typi-

cally involve stock variables and that, if one relies on natural flows of new profits, it takes

time to repair weaknesses in stock variables. For example, in the early 1990s, when US mone-

tary policy allowed banks to earn large margins from maturity transformation, it took about

two years of high profits for banks to get clearly out of the range where their solvency could

be said to be in doubt; during these two years, the recovery of the real economy was stalled.

For faster repairs of bank balance sheets, the problems have to be addressed at the level of

stock variables, through immediate recapitalization or through a resolution procedure impos-

ing writedowns on legacy assets as well as the banks’ shareholders and creditors. Recapitali-

zation tends to be strongly resisted by incumbent shareholders and management;41 entry into a

resolution procedure may be resisted even by the authorities if they fear that such a measure

might have damaging systemic repercussions.42 However, if neither approach to dealing with

the problem at the level of stock variables is used, the weakness of the financial sector can

persist for quite a long time.

The notion that financial-sector weakness may disappear on its own, without a cleanup, can

actually be self-defeating. If problems are due to excess capacity in the industry preventing

banks from earning appropriate margins, then a strategy of closing one’s eyes in order to

avoid bank resolution and bank closures may cause the excess capacity to persist so that the

much better capitalized. As stated, the argument overlooks the very large flows of money that the major Latin American debtors made to their creditors in the second half of the 1980s, much of it funded be offi-cial lending to these debtors.

40 See Hoshi and Kashyap (2004, 2010) 41 As explained by Admati et al. (2013), this resistance is largely due to a debt overhang effect implying that

the new funds from the recapitalization benefit incumbent debt holders as well as shareholders, with the consequence that the market value of total equity rises by less than the amount obtained by the recapitali-zation. Admati et al. (2012) shows that recapitalization can actually be used to address balance sheet problems at the level of stock variables. If banks raise equity through rights offerings and if they use the proceeds to buy other marketed assets, the net cash flows into the banking sector that are involved are small; all that happens is a reshuffling of the set of financial instruments that allocate rights to the returns of real assets in the economy.

42 For a systematic overview, see ASC (2012), for an assessment of current resolution regimes, see Hellwig (2014 b).

25

root cause of financial-sector weakness is not addressed. In such a situation also the Green-

span strategy of allowing banks to earn high profits in order to rebuild equity will not work.

In this context, it is pertinent to observe that banking in Europe is still characterized by low

profitability. This low profitability reflects the large capacity growth that European banking

has seen and the increased intensity of competition since the 1990s. It may also reflect the

flatness of the yield curve, at which point we must think about the effects of monetary policy

measures such as forward guidance and quantitative easing on the profitability of banks. We

thus come back to the conflict between the macroeconomic stability objective of making cred-

it to the real economy plentiful and cheap and the financial stability objective of allowing

banks to repair their balance sheets so as to better serve their macroeconomic functions in the

future.

At this point, we do not have a good conceptual framework for dealing with this conflict and

handling the implied tradeoffs. Thinking about the conflict in terms of tradeoffs may actually

be problematic because, for reasons given in the preceding subsection of this paper, the au-

thorities are unlikely to have very good information about the actual state of the financial sys-

tem. With insufficient information, one should not go for fine-tuning of tradeoffs.

With due caution, the following may seem like a reasonable agenda for proceeding:

– In normal times, let monetary policy serve its macroeconomic objectives without paying much attention to financial stability.

– At the same time, make sure that microprudential, macroprudential and monetary authori-ties understand what is going on in the financial system as well as the monetary system.

– If risks in the financial sector are building up, consider the use of macroprudential regula-tion to restrain the buildup. Targeted tools such as bounds on loan-to-value ratios and in-terest-to-income ratios in real-estate finance may ease the task of monetary policy in such situations.

– In an acute crisis, allow for financial stability concerns to take precedence and support the financial system.

– At the same time, make sure that solvency problems in the financial sector are not al-lowed to linger but are addressed right away even if there is a fear that the cleanup of the financial sector may exacerbate the weakness of the real economy.

– Try to address the financial sector weaknesses at the level of stock variables, in particu-lar, bank equity so as to avoid the slowness of a recovery that relies on profit flows. Im-posing recapitalization, inside resolution or outside, will be fought by the industry but is very likely needed to get out of the current mess.

– Think about monetary-policy measures and about macroprudential measures in a com-prehensive way, considering their effects on both the macroeconomy and the health of fi-nancial institutions.

This agenda actually minimizes the explicit support for financial stability that monetary poli-

cy should give, assigning much of the task instead to macroprudential regulation and to recap-

26

italization and resolution measures when the financial system is weak. In a serious systemic

crisis, too-big-to-fail policies intended to keep the system going are unavoidable and are in

fact desirable. However, recognizing the need for such support in a crisis should not become a

pretext for having central banks to always stand ready to support the industry. Moreover, cen-

tral bank support should not be abused to avoid necessary cleanups.

Without such limitations on the financial-stability mandate of the central bank, there is serious

risk of monetary policy succumbing to financial dominance, where central banks take finan-

cial-sector doings as fixed and adapt their monetary strategies to minimize systemic damage

from financial-sector risks. And there is a serious risk of hidden fiscal dominance if financial-

stability concerns induce the central bank to provide plentiful and cheap funding to weak

banks, and these banks use the central-banking funding to lend to their governments, a con-

stellation that has characterized the euro area, at least in 2012.43

For central banks not to be exposed to a risk of financial dominance, we need strong arrange-

ments for prevention, high equity requirement and macroprudential arrangements that make a

crisis unlikely, and we need strict rules for imposing cleanups. We also need to avoid a re-

gime in which the central bank makes private institutions subservient to its macroeconomic

needs while assuring them of support if risks turn out badly. Such a regime would exacerbate

moral hazard. Financial dominance then might become unavoidable and might persist for a

long time.

43 See Acharya and Steffen (2015). In 2012, the carry trade involved in borrowing from the central bank and

lending to one’s own government could be very profitable, with margins of 400 or so basis points. Since then, the success of the ECB’s OMT program, though never used, has eroded the greater part of these margins.

27

References

Acharya, V.V., Schnabl P. and Suarez, G. (2013), Securitization without Risk Transfer, Jour-

nal of Financial Economics, Volume 107, Issue 3, March 2013, pages 515-536.

Acharya, V.V., and Steffen, S. (2015), The greatest carry trade ever? Understanding European

bank risk, Journal of Financial Economics 115, 215-236

Admati, A.R., P. Conti-Brown, and P. Pfleiderer (2012), Liability Holding Companies, UCLA

Law Review 59, 852-913

Admati, A.R., P.M. DeMarzo, M.F. Hellwig, and P. Pfleiderer (2013), The Leverage Ratchet

Effect, Rock Center for Corporate Governance at Stanford University Working Paper

No. 146.

Alesina, A., and Summers, L. J. (1992), Central bank independence and macroeconomic per-

formance: Some comparative evidence, Journal of Money, Credit and Banking 25 (2),

151 – 162.

ASC (2012), Forbearance, Resolution, and Deposit Insurance, Reports of the Advisory Scien-

tific Committee of the European Systemic Risk Board Nr. 01, Frankfurt.

ASC (2014), Is Europe Overbanked?, Reports of the Advisory Scientific Committee of the

European Systemic Risk Board No. 04, Frankfurt.

Bagehot, W. [1873] 1906. Lombard Street: A Description of the Money Market. New York:

Charles Scribner's Sons.

Bernanke, B.S. (1983), Nonmonetary effects of the financial crisis in propagation of the Great

Depression, American Economic Review 73 (3), 257 – 276.

Bernanke, B.S. (1995), The Macroeconomics of the Great Depression: A Comparative Ap-

proach. Journal of Money, Credit and Banking 27(1): 1-28.

Bernanke, B.S., Gertler, M., and S. Gilchrist (1999), The financial accelerator in a quantita-

tive business cycle framework, in: J. B. Taylor and M. Woodford (eds.), Handbook of

Macroeconomics Vol. 1, Ch. 21, 1341-1393, Elsevier. .

Bichsel, R., and J. Blum (2005), Eigenkapitalregulierung für. Banken, Bestandesaufnahme

und Ausblick, Schweizerische Nationalbank, Quartalsheft 04/2005, 42 – 51.

Bank for International Settlements (2007), 77th Annual Report, June 2007, Basel, Switzer-

land.

Blum, J. (2007), Why 'Basel II' May Need a Leverage Ratio Restriction, Swiss National Bank

Working Paper 04/2007.

28

Born, K.E. (1967), Die deutsche Bankenkrise 1931, München, Piper.

Demyanyk, Y., and Van Hemert, O. (2009), Understanding the Subprime Mortgage Crisis,

Review of Financial Studies 24 (6): 1848-1880.

Deutsche Bundesbank (2012), Stellungnahme gegenüber dem Bundesverfassungsgericht zu

den Verfahren mit den Az. 2 BvR 2728/13, (Statement to the German Constitutional

Court in the Proceedings concerning the ECB’s OMT program).

Eichengreen, B.J. (1992), Golden Fetters: The Gold Standard and the Great Depression,

1919-1939. Oxford University Press, Oxford, U.K.

Eichengreen, B.J. (2015), Hall of Mirrors, Oxford University Press, Oxford, UK

Fleming, J. M. (1962). Domestic financial policies under fixed and floating exchange rates,

IMF Staff Papers 9: 369–379.

Friedman, M., and Schwartz, A.J. (1963), A monetary history of the United States 1867 –

1960, Princeton University Press, Princeton, N.J.

Friedman, M. (1968), The Role of Monetary Policy, American Economic Review 58, 1-17.

Friedman, M. (1969), The Optimum Quantity of Money, in: The Optimum Quantity of Money

and Other Essays, Chicago University Press, Chicago.

Goodhart, C.A.G.(1988), The evolution of central banks, MIT Press, Cambridge, MA.

Gorton, G. (2010), Slapped by the Invisible Hand: The Panic of 2007, Oxford University

Press, Oxford, UK

Grilli, V., Masciandaro, D., and Tabellini, G. (1991), Institutions and policies, Economic Pol-

icy 6 (13), 341 – 392.

Hayek, F.A. (1976), Denationalization of Money - The Argument Refined, The Institute of

Economic Affairs, London, U.K.

Hellwig, M.F. (1985), What do we know about Currency Competition?, Zeitschrift für

Wirtschafts- und Sozialwissenschaften 105, 565-589.

Hellwig, M.F. (1994 a), Liquidity Provision, Banking, and the Allocation of Interest Rate

Risk, European Economic Review 38 (1994), 1363 – 1389.

Hellwig, M.F. (1994 b), Banking and Finance at the End of the Twentieth Century, WWZ

Discussion Paper 9426, University of Basel, Switzerland.

29

Hellwig, M.F. (1995), Systemic Aspects of Risk Management in Banking and Finance”,

Schweizerische Zeitschrift für Volkswirtschaft und Statistik/Swiss Journal of Economics

and Statistics 131, 723 - 737.

Hellwig, M.F. (1998), Banks, Markets, and the Allocation of Risks, Journal of Institutional

and Theoretical Economics 154: 328-351.

Hellwig, M.F. (2009), Systemic Risk in the Financial Sector: An Analysis of the Subprime-

Mortgage Financial Crisis, De Economist 157: 129-207.

Hellwig, M.F. (2014 a), Systemic Risk and Macroprudential Policy, in: A. Houben, R.

Kijskens, M. Teunissen (eds.), Putting Macroprudential Policy to Work, Occasional

Studies 12-7, De Nederlandse Bank, Amsterdam 2014, 42 – 77.

Hellwig, M.F. (2014 b), Yes Virginia, There is a European Banking Union! But It May Not

Make Your Wishes Come True, in: ÖNB: Österreichische Nationalbank, Towards a Eu-

ropean Banking Union: Taking Stock, 42nd Economics Conference 2014, 156 – 181.

Hellwig, M.F. (2015 a), Financial stability, monetary policy, banking supervision, and central

banking, in: Monetary Policy in a Changing Landscape: Proceedings of the First ECB

Forum on Central Banking, 21 – 54.

Hellwig, M.F. (2015 b), Richtige und falsche Ängste vor einer expansiven Geldpolitik, Frank-

furter Allgemeine Sonntagszeitung, Nr. 10, March 8, 2015

Hoshi, T., and Kashyap, K. (2004), Japan’s financial crisis and economic stagnation, Journal

of Economic Perspectives 18 (Winter), 3 -26.

Hoshi, T., and Kashyap, K. (2010), Why did Japan stop growing?, NBER working paper, Na-

tional Bureau of Economic Research, Cambridge, MA.

International Monetary Fund (2007), Market Developments and Issues, Global Financial Sta-

bility Report, Washington D.C., April.

International Monetary Fund (2008), Financial Stress and Deleveraging: Macro-Financial

Implications and Policy, Global Financial Stability Report, Washington D.C., October.

Lucas, R.E. (1972), Expectations and the Neutrality of Money, Journal of Economic Theory

4, 103–124.

Modigliani, F. (1977), The Monetarist Controversy or, Should We Forsake Stabilization Poli-

cies?, American Economic Review 67, 1-19.

Mundell, R.A. (1962), The Appropiate Use of Monetary and Fiscal Policy for Internal and

External Stability, IMF Staff Papers 9, 70-79.

30

Patinkin, D. (1965), Money, Interest and Prices, 2nd Edition, Harper&Row, New York.

Sinn, H.-W. (2012), Die Target-Falle, Hanser-Verlag, München.

Sinn, H.-W. (2013), Verantwortung der Staaten und Notenbanken in der Eurokrise:

Gutachten im Auftrag des Bundesverfassungsgerichts, Zweiter Senat

Verfassungsbeschwerden 2 BvR 1390/12, 2 BvR 1439/12 und 2 BvR 1824/12

Organstreitverfahren 2 BvE 6/12 (Responsibilities of States and Central Banks in the Euro

Crisis: Statement before the German Constitutional Court in the Proceedings concerning

the ECB’s OMT program), mimeo, CES-ifo Group, Munich, June 2013.

Vaubel, R. (1985), Competing Currencies: The Case for Free Entry, Zeitschrift für

Wirtschafts- und Sozialwissenschaften, 105, 547-565.

Weidmann, J. (2014), Für die Verluste haften die Steuerzahler (Taxpayers are liable for los-ses), Frankfurter Allgemeine Sonntagszeitung, Dec. 28, 2014.

Winkler, A. (2015), Argumentation gegen EZB-Staatsanleihenkäufe, Wirtschaftsdienst 95,

178 – 185. Woodward, B. (2000), Maestro: Greenspan’s Fed and the American Boom, Simon&Schuster,

New York.