I

The euro area is characterised by a peculiar governance structure: while Member States are in charge of national fiscal policies, the European Central Bank (ECB) conducts monetary policy for the euro area as a whole. As early as 1990 the Deutsche Bundesbank warned that this set-up was fragile and argued that only a European state would provide proper fiscal and economic foundations for the smooth functioning of a monetary union.Footnote ² However, instead of opting in favour of a more centralised policy approach, the designers of the Maastricht Treaty decided to mitigate the Bundesbank's concerns by introducing a strong separation between monetary and fiscal policy and by imposing a strict framework for the conduct of national fiscal policies. Key elements were the No-ECB credit to the public sectorFootnote ³ and the No-bail-out clauses in the Maastricht Treaty, the criteria on government deficits and debt for entering the euro area and the Stability and Growth Pact (SGP). The assumption was that this governance structure would make a European state redundant as there would be no need for either monetary financing of deficits or for fiscal transfers among euro area Member States. The crisis in euro area government bond markets has proved this assumption wrong. Most observers blame a failure of the fiscal framework, in particular the SGP, for this outcome (Issing Reference Issing2010a; Wyplosz Reference Wyplosz2010b).

This article provides a different interpretation of the crisis. In particular, it will be argued that the crisis reflects similar stability challenges to those that financial intermediaries are subject to when confronted with an adverse macroeconomic shock which makes their solvency questionable. Thus, when assessing the origins of and possible solutions to the crisis, it is useful to review the response of commercial banks to similar shocks. As euro area governments perform maturity transformation services without having access to a proper lender of last resort (LOLR) in times of crisis, the appropriate benchmark is a banking system operating under similar conditions. In the nineteenth century the United States had such a banking system which was regularly hit by financial crises. US banks responded by using clearing houses as co-insurance mechanisms for the ‘joint production of confidence’ (Gorton and Mullineaux Reference Gorton and Mullineaux1987; Kroszner Reference Kroszner2000; Tallmann and Moen 2010).

A comparison of nineteenth-century US commercial bank clearing houses (CBCHs) with twenty-first-century euro area governments reveals that most of the instruments CBCHs used are similar to the ones euro area governments have either recently decided upon, such as the establishment of the European Financial Stability Facility (EFSF), or are at present debating intensively, such as the issuance of joint euro bonds. This suggests that euro area countries have followed best private-sector practices of crisis management. However, inherent fragilities of co-insurance mechanisms, as well as the long maturity and elusive quality of government debt, present challenges in designing a sustainable solution to the crisis without compromising the original goal of euro area governance: ensuring sound fiscal policies as a prerequisite for maintaining price stability. Thus, a substantially more comprehensive economic and political union might be the prerequisite for the smooth functioning of the European Monetary Union.

The article is structured as follows. After this introduction, Section II compares commercial banks and governments as providers of maturity transformation services and their exposure to stability risks arising from macroeconomic shocks to solvency. Section III explains why the crisis in euro area government bond markets reflects the peculiar governance framework of the European Monetary Union. Section IV introduces and analyses clearing houses as an institutional mechanism used by US banks in the nineteenth century to contain financial crises under a similar macroeconomic policy set-up. Section V highlights the similarities and differences between nineteenth-century US banks and twenty-first-century euro area governments in containing financial crises. Section VI concludes by drawing lessons euro area governments can learn from US banks in fighting crises of confidence.

II

Financial markets and intermediaries provide qualitative asset transformation services, including maturity transformation (Bhattacharya and Thakor Reference Bhattacharya and Thakor1993).Footnote ⁴ The provision of these services fosters growth and productivity (Levine Reference Levine1997), but also entails substantial stability risks. This holds in particular for the provision of maturity transformation services. While Diamond and Dybvig (Reference Diamond and Dybvig1983) suggest that the fragility derives from the self-fulfilling prophecies of banks' depositors, the empirical evidence on bank runs (Calomiris and Gorton Reference Calomiris, Gorton and Hubbard1991) indicates that runs are caused by the asymmetry of information between depositors and bank managers about the quality of banks' assets (Diamond Reference Diamond1984). As a result, when confronted with noisy information about a decline in the quality of banks' assets, like recessions, substantial drops in house or stock prices or a failure of one bank, depositors prefer to play safe and decide not to extend their deposits. Thus, bank runs are triggered by rational investors observing a change in fundamentals that might have a negative impact on the solvency of the bank. As investors not only face credit, but also liquidity risk due to the ‘sequential service constraint’, they try to limit their potential losses by being ‘first’ in selling their claims, which causes the run.Footnote ⁵ However, due to the noisiness of the information about fundamentals, rational investors may turn out to be wrong ‘and force the bank to liquidate unnecessarily’ (Calomiris and Kahn Reference Calomiris and Kahn1991, p. 510).

Financial history provides many examples of this fragility. However, the frequency and magnitude of financial crises have varied considerably from country to country. This suggests that institutional features of the respective financial systems play a key role in explaining why the provision of maturity transformation services has been more crisis-prone in some countries and periods compared to others (Calomiris and Schweikart Reference Calomiris and Schweikart1991; Calomiris and Gorton Reference Calomiris, Gorton and Hubbard1991). The establishment of an LOLR has been the most widely used public-sector device to deal with this fragility. The central bank intervenes in the money market when it is convinced that the market is ‘wrong’ (Bagehot Reference Bagehot1873), even though it does not know for certain whether all the institutions benefiting from its interventions are indeed solvent (Goodhart Reference Goodhart1999). However, the private sector has also been active in designing stabilising instruments. US commercial bank clearing houses (CBCHs), operating in the nineteenth century, represent one of the most intensively researched instruments private banks have used to contain financial panics by granting credit to member banks unable to finance themselves on the open market.

Governments also have to rely on bondholders' confidence in their ability to service the debt because they face similar roll-over risks to those of banks. Government assets are very long-term and illiquid, i.e. they consist of roads, buildings etc. Hence, it is mainly the ability and willingness of the respective taxpayers to service the debt which influences bondholders' confidence in the solvency of the respective government. Accordingly, roll-over risks increase (1) when government bonds are predominantly short-term, (2) when debt ratios cross benchmarks that make it practically impossible to repay all maturing debt, including long-term debt,Footnote ⁶ and (3) when growth prospects decline substantially (Domar Reference Domar1944).

As a result, government bonds face a similar liquidity risk to that to which bank deposits are exposed. If bondholders lose confidence in governments' solvency, for example as a result of macroeconomic or political shocks, there will be a run from government bonds as investors aim to avoid selling at distressed prices due to the drying up of liquidity. Thus, like bank runs, runs on government bonds are rational, even though bondholders might misjudge the ability and willingness of the taxpayer to service the bonds, in the same way as depositors might misjudge the quality of their banks' assets. As well as this, government bonds of a group of countries can become subject to contagion triggered by noisy signals about fundamentals that have the same impact on all the members of the respective group of countries. In addition, the default of a single government sharing broadly similar economic and fiscal characteristics with other countries can trigger a contagious run. Finally, given the importance of government bonds for the mature economies' financial systems as a whole, contagion can spread from bond markets to the banking systems of the countries involved (Arnold and Lemmen Reference Arnold and Lemmen2001; Bini Smaghi Reference Bini Smaghi2010b).

III

The risk of a run on government debt is most pronounced in settings where the government cannot rely on a central bank as a lender of last resort, either because the country is running a fixed exchange rate regime with open capital accounts (Giavazzi and Pagano Reference Giavazzi, Pagano, Dornbusch and Draghi1990; Obstfeld Reference Obstfeld1994) or when it has issued a substantial share of its debt in foreign currency (Eichengreen and Hausmann Reference Eichengreen and Hausmann1999; Rodrik and Velasco Reference Rodrik and Velasco1999). By contrast, with government debt issued in domestic currency and a central bank standing ready to act as an LOLR for government bonds, central bank and government are in a position to jointly guarantee their own stability and the stability of the financial system (Stella Reference Stella2010) – the government in terms of solvency, the central bank in terms of liquidity (Figure 1a). Thus, even in the case of strong concerns about government solvency, bondholders do not ‘run’ because the central bank reduces liquidity risks to zero. At the same time, default risk is transformed into inflation risk (Arnold and Lemmen Reference Arnold and Lemmen2001).Footnote ⁷

The euro area governance framework as laid down in the Maastricht Treaty implies that the ECB is unable to perform the role of LOLR for euro area government bond markets (Figure 1b). Thus, it dismantles the stability mechanism which characterises most mature economies because debt issued by euro area governments is put on the same level as foreign currency debt issued by emerging market governments and banks (Tabellini Reference Tabellini2010). This is because

• • the euro – while being the domestic currency of all Member States – is not the domestic currency of any individual Member State;

• • the No-ECB credit clause and the de facto No-ECB credit policy to the public sector provides a strong signal that the ECB will not operate as an LOLR for euro area governments.

Thus, euro area governance has been built on the assumptions that

• • a central bank does not need the government as the ultimate guarantor of its creditworthiness (Goodhart Reference Goodhart1988, Reference Goodhart1998) and

• • governments do not need and should not need the central bank to ensure the liquidity of their bonds, i.e. the smooth functioning of government bond markets, following the doctrine that in developed financial markets liquidity always follows solvency (Rochet and Vives Reference Rochet and Vives2004).

By contrast, there is no difference regarding the access of euro area commercial banks to LOLR facilities compared to their counterparts in other mature economies (Figures 1a and 1b). The ECB – like all central banks of mature economies – has performed the LOLR role in the global financial crisis, intervening in euro area money markets in a timely fashion and with unlimited amounts.Footnote ⁸ Moreover, in the post-Lehman period euro area governments, like all mature economy governments, have stabilised the banking system in the form of bank capitalisation. However, in strong contrast to other mature economies, euro area governments stabilised their banking systems and economies without backing by the central bank to contain potential negative liquidity effects. These effects might arise if the debt implications of government interventions triggered concerns about government solvency. Accordingly, euro area governance plays a key role in explaining why several euro area governments experienced a run from their debt, interpreting the spikes in credit default swaps and interest rates (Figure 2) as indicators of financial market distress (Tallman and Moen Reference Tallman and Moen2010).

Figure 1. Nation state and euro area governance – a financial stability perspective

Note: To keep the diagram simple it does not include banking regulation and supervision, the main anti-moral hazard instrument at the disposal of the LOLR. This is because there are different institutional settings in which banking regulation and supervision are conducted in nation states as well as in the euro area. Illustrating these differences in the figure would make it more complex without providing much additional insight.

Sources: Author's compilation, reflecting arguments contained in Goodhart (Reference Goodhart1988, Reference Goodhart1998) and Tabellini (Reference Tabellini2010).

Figure 2. Ten-year government bond yields of selected mature economies

Source: Reuters.

The alternative interpretation stressing the weaknesses in enforcing the euro area fiscal framework, i.e. the Stability and Growth Pact, and focusing on fiscal profligacy among the euro area crisis countries, i.e. Greece, Ireland, Portugal and Spain, has several shortcomings. Firstly, Ireland and Spain followed the SGP rules for most of the pre-crisis years. Moreover, several countries, i.e. France and Germany, violated the SGP rules but have not experienced turmoil in their respective government bond markets. Thus, the noise-to-signal ratio (Kaminsky et al. Reference Kaminsky, Lizondo and Reinhart1998) of the signal ‘violation of the SGP’ is quite large. Secondly, compared with other mature economies, the government debt ratios of the crisis countries are not expected to reach levels which, by their sheer magnitude, seem to be unsustainable. Government debt ratios of Japan and the US are higher than those of Spain and Portugal, while debt ratios of the UK and the US are expected to rise faster than those of Greece and Portugal (Table 1). However, government debt markets of the non-euro area mature economies have continued to function smoothly.

Table 1. Government debt in selected mature economies (as a percentage of GDP)

*Austria, Belgium, Finland, France, Germany, Italy, Luxembourg, Netherlands - unweighted average

**Australia, Canada, Denmark, Iceland, New Zealand, Norway, Sweden, Switzerland - unweighted average

Sources: IMF (WEO database, October 2010), author's calculations.

High debt ratios imply that governments engage heavily in maturity transformation services, i.e. have a strong need to roll over government debt. In 2010 and 2011 all euro area crisis countries, with the exception of Ireland, have had to roll over debt in substantial amounts (IMF 2010). However, the figures also suggest that the refinancing needs of other mature economy governments, within and outside the euro area, are of a similar size or even higher.

Euro area crisis countries differ substantially from their mature economy peers with regard to the size of the macroeconomic shock they have been facing. Growth prospects are low, compared to other mature economies, within and outside the euro area, and even more striking when compared to the respective countries' pre-crisis growth experience (Figure 3). Moreover, the pre-crisis growth model, based on capital inflows, credit growth and domestic demand, has collapsed. Switching to an export-led growth strategy is difficult, as the previous growth process was accompanied by rising wages and a loss of competitiveness compared to other euro area Member States, in particular Germany. Euro area membership implies that currency devaluation is not an instrument available to regain competitiveness and restore growth in a short and less painful process.Footnote ⁹ The negative growth outlook, in combination with high and rising levels of debt, rationalises the run on debt issued by the euro area crisis countries, as it puts long-run solvency into question. In the case of Greece, these doubts were compounded by recurrent revisions of faked statistics.

Figure 3. Pre- and post-crisis average GDP growth rates in selected mature economies

Sources: IMF (WEO database, October 2010), author's calculations.

Overall, euro area crisis countries are in a similar position to banks when depositors have doubts about banks' solvency. If depositors withdraw their funds, and this leads to a rise in deposit rates, then liquidity problems, reflecting doubts about solvency, lead to insolvency. The same applies to government bonds and government bondholders concerned about government solvency (Ostry et al. Reference Ostry, Gosh, Kim and Qureshi2010). This is most evident for countries with high levels of government debt. Interest rates reaching levels that are twice or even three times as high as before the crisis are inconsistent with economic recovery and, given the low growth prospects, also inconsistent with long-run government solvency (Bini Smaghi Reference Bini Smaghi2011). If these levels prevail, the insolvency of high-debt countries is almost a foregone conclusion.

In countries where fiscal and monetary authorities operate at the same level, for example in Japan, the US and the UK, governments do not face this vicious spiral of illiquidity and insolvency, as investors know that the respective central banks would intervene if interest rates became inconsistent with growth. As discussed above, euro area crisis countries were in a different situation, because the ECB did not provide signals that it would engage in LOLR interventions.Footnote ¹⁰ Accordingly, rational investors – facing the risk of insolvency and illiquidity – decided to run on Greek government debt which – due to a high degree of similarity in growth models pursued in the past and the growth outlook for the future – spilled over to Ireland, Portugal and Spain. The question which euro area governments had to decide was whether rational investors were right. In early May 2010 they decided that this was not the case and established the EFSF and – in December 2010 – the European Stability Mechanism (ESM). While this has been seen as a break with the No-bail-out clause of the Maastricht Treaty, the following sections will provide evidence that it is close to what US commercial banks did in periods of financial turmoil before the founding of the Federal Reserve.

IV

Nineteenth-century US financial history was markedly different from financial history in most mature economies in at least three aspects:

• • Firstly, the US was a latecomer in establishing a central bank (Goodhart Reference Goodhart1988). While in most European countries, institutions performing central bank functions, in particular the LOLR function, had emerged by the mid to late nineteenth century, the Federal Reserve was founded only in 1913.

• • Secondly, the US banking system was highly fragmented due to many limitations in bank consolidation and branching (Calomiris Reference Calomiris, Klausner and White1993).

• • Thirdly, financial crises were frequent. Calomiris and Gorton (Reference Calomiris, Gorton and Hubbard1991) list six banking panics between 1873 and 1907. This is substantially more than in any European country (Tilly Reference Tilly2008).

These aspects are important in order to understand the functioning of US commercial banks' clearing houses (CBCHs). In the mid nineteenth century cheques became the dominant means of payment when deposits replaced notes as the main refinancing instrument used by US banks. In the beginning, cheques were cleared bilaterally. However, banks quickly discovered that central multilateral clearing would be more efficient. Thus, they founded CBCHs, which operated as the institution through which all member banks cleared their cheques.Footnote ¹¹

The quality of each cheque was dependent on the solvency of the account holder and the soundness of the bank the cheque issuer had an account with. CBCHs did not have faith in the ability of the market, i.e. depositors, to control the quality of the institutions and prevent the overissuance of deposits. Thus, they began developing regulatory and supervisory instruments, including (Cannon Reference Cannon1910; Gorton and Mullineaux Reference Gorton and Mullineaux1987):

• • an admission test (based on certification of adequate capital);

• • periodic exams (audits) by the clearing house;

• • information requirements (publication of financial statements);

• • reserve requirements;

• • disciplinary actions (fines) in cases of misbehaviour and, for extreme violations of rules, exclusion from membership.

The main goal of these regulatory activities was to ensure the convertibility of member banks' deposits into specie at the exchange rate of 1:1, i.e. the liquidity of all deposits issued by member banks. Thus, clearing houses were seen ‘as an essential discipline on overissue’ (Goodhart Reference Goodhart1988, p. 31) and became a quality signal which member banks could send to their depositors. As a result, membership was most attractive for weaker banks or banks that might engage in moral hazard behaviour (Dowd Reference Dowd1994). By contrast, strong member banks had an incentive to enforce regulations strictly.

Figure 4a summarises the set-up from a financial stability perspective. With specie being the ultimate source of liquidity, commercial banks provide maturity transformation services without an LOLR. Commercial banks establish CBCHs to economise on transaction costs in the clearing and settlement of payments. The latter develop regulatory and supervisory instruments to contain moral hazard behaviour by member banks in the form of overissuance (Norman et al. Reference Norman, Shaw and Speight2011).

Figure 4. Nineteenth-century US commercial banks – a financial stability perspective

Sources: author's compilation based on Gorton and Mullineaux (Reference Gorton and Mullineaux1987) and Goodhart (Reference Goodhart1988)

In times of crisis, clearing houses' member banks temporarily joined forces to fight off contagion effects that might arise if one member bank were to default. These co-insurance activities (Kroszner Reference Kroszner2000) had the following characteristics (Cannon Reference Cannon1910; Timberlake Reference Timberlake1984; Gorton and Mullineaux Reference Gorton and Mullineaux1987; Kroszner Reference Kroszner2000):

• • CBCHs issued clearing house loan certificates on the basis of eligible collateral and despite several legal obstacles to doing so. Clearing house loan certificates were liabilities issued by the CBCH and jointly guaranteed by all member banks. In general, they were used by member banks as a substitute for cash and specie to balance positions within the house. In the South and in the panic of 1907, loan certificates were also used as cash substitutes outside CBCHs, when member banks offered depositors the opportunity to exchange deposits into loan certificates instead of specie.

• • Clearing house loan certificates were issued when the loan committee deciding on the issuance was convinced that either (a) the market was wrong in its concerns about individual members' solvency or (b) the risk of contagion following an individual member bank's failure was high. In this case, member banks replaced the open capital market by an internal market.

• • Clearing house loan certificates were accepted by depositors because they transformed individual member banks' debt into debt where all the member banks pooled their risk. Thus, loan certificates were backed by a single institution with one large diversified portfolio (Gorton Reference Gorton2007), deliberately making credit risks of member banks homogeneous. In line with this, CBCHs started publishing a weekly statement for the house as a whole instead of publishing weekly financial statements for individual banks.

• • Due to the risk of contagion, no member bank was allowed to fail during a panic. While there was no official policy about avoiding the failure of participating members, de facto certificates were available to all members, including those ‘which the CBCH perhaps knew would certainly fail’ (Gorton and Mullineaux Reference Gorton and Mullineaux1987, p. 464). Moreover, CBCHs did not publish the identity of member banks who were borrowing through the loan certificate process.

• • CBCHs ordered special monitoring and regulations for those member banks which received loan certificates. For example, banks borrowing from the CBCH were required to cease new lending activities for as long as they had not repaid the loan certificates received. Moreover, CBCHs would demand additional security from particularly troubled banks.

• • Member banks were expelled from the CBCH if they failed to repay clearing house loans and loan certificates after the crisis was over and market conditions had returned to normal. Given the benefits related to CBCH membership referred to above, the threat of expulsion had a strong disciplinary effect on CBCH members (Goodfriend Reference Goodfriend1991). Indeed, the largest CBCH, that of New York City, did not record any losses in its issuing activity.

• • Any losses from CBCH interventions – exceeding the value of pledged collateral – were in the majority of cases shared by the CBCH members pro rata of capital and surplus.

• • The joint production of confidence was a temporary phenomenon enacted only in times of crisis. When the panic had subsided, member banks repaid the loans and loan certificates and again started running operations on an individual basis in the open capital market. Thus, CBCHs were not exposed to moral hazard issues arising from medium- to long-term financial assistance.

Figure 4b illustrates what happened when CBCHs switched into crisis mode. Member banks jointly guaranteed their debt and issued clearing house loans and loan certificates as a substitute for specie to contain the panic arising from doubts about the solvency of individual member banks and the associated risk of contagion.

There is a broad consensus that CBCHs were an effective institutional device set up by the private sector to contain financial panics (Timberlake Reference Timberlake1984; Greenspan Reference Greenspan1997; Kroszner Reference Kroszner2000). This consensus includes advocates of free banking conceding that information problems in financial markets might lead to the formation of banking clubs and the endogenous development of regulatory instruments (Dowd Reference Dowd1994, Reference Dowd1996). However, they argue that the emergence of CBCHs in the US was caused by prior government interventions in the form of severe restrictions on branching. If the US banking system had been allowed to operate without these constraints, the problems CBCHs – and at a later stage central banks – were founded to resolve would have been dealt with largely by mergers, i.e. the emergence of larger and more diversified banks with a unified ownership.Footnote ¹²

However, CBCHs did not operate perfectly. ‘Often the issue of loan certificates was either too late or of insufficient quantity to reassure bank depositors, …’ (Roberds Reference Roberds1995, p. 21). There were two reasons for the slow reaction of CBCHs (Goodhart Reference Goodhart1988). First, the decision to provide assistance was taken by a committee of major commercial banks with diverse interests, which by definition takes time.Footnote ¹³ Second, strong member banks faced a conflict of interest: on the one hand they wanted to contain contagion effects, on the other hand crises were also perceived as opportunities to reduce the number of competitors by refusing to issue certificates. Empirical evidence indicates that swift and forceful joint action was crucial in determining the degree of success CBCHs had in containing and mitigating panics, with a lack of coordination being the main reason for a crisis with a considerably larger number of bank failures (Calomiris and Schweikart Reference Calomiris and Schweikart1991; Moen and Tallman Reference Moen and Tallman2000). Thus, only an institution which does not act as a competitor to any commercial bank can perform – in exchange for wide ranging supervisory and regulatory powers – LOLR activities properly (Goodhart Reference Goodhart1988). According to this view, strongly rejected by free-banking advocates (Dowd Reference Dowd1994, Reference Dowd1996), CBCHs were imperfect predecessors of the modern public-sector central bank.

Overall, the experience of CBCHs in the United States suggests that

• • private banks responded to the fragilities of maturity transformation by creating co-insurance mechanisms based on non-market instruments, including regulatory and supervisory rules as well as the temporary replacement of the open market by an internal capital market;

• • private co-insurance mechanisms are fragile if run by institutions which in non-crises times are competitors. Public-sector central banks eliminate this sort of fragility (Norman et al. Reference Norman, Shaw and Speight2011). However, operating under a flexible exchange rate regime, the establishment of these implies that the price and quantity of the money supply becomes a public policy decision. As a result, mature economy central banks have become institutions not only focusing on financial stability, as their private-sector predecessors, the CBCHs, did, but also on price stability (Goodhart Reference Goodhart1988; Friedman Reference Friedman1990; Folkerts-Landau and Garber Reference Folkerts-Landau and Garber1992).

V

To what extent is the experience of US CBCHs of relevance in understanding and assessing the crisis in euro area government bond markets and the response by euro area governments to the crisis? At first glance, the establishment and management of CBCHs seem to be a very different undertaking than the establishment and management of a single currency among independent nations. However, there are a number of characteristics with a high degree of similarity:

1. 1. The reduction of transaction costs was one of the economic motivations for the establishment of the euro area (Emerson et al. Reference Emerson, Gros and Italiener1992) as in a highly integrated economic area, like the European Union, it is inefficient to use different currencies for the exchange of goods and services. This is similar to the argument that led to the establishment of CBCHs in the US with regard to the clearing of cheques.

2. 2. Transaction cost savings depend on equal and constant quality of the assets in exchange. The efficiency of central clearing via a clearing house is based on a netting process. Thus, all debits and credits have to be of the same quality in order to allow the clearing house to arrive at the respective balances leading to a corresponding flow of specie at the exchange rate 1:1. The introduction of the euro implied that legacy currencies were regarded – expressed at the final conversion rates – as of equal quality. Moreover, Member States agreed that there would be no need in the future to adjust the (nominal) exchange rates among the participating currencies.

3. 3. In both cases participating members agreed on common regulations and supervision to ensure the stability of the centralised transaction arrangements. Like CBCHs, the euro area

   • • defines entry conditions, i.e. an admission test, in the form of the Maastricht criteria;

   • • specifies certain obligations for participating members, that they provide and publish data and relevant information, for example in the area of government finance statistics;

   • • requires members to comply with certain norms of behaviour, with the criteria laid down in the SGP constituting the main rules which euro area governance is built upon;Footnote ¹⁴

   • • examines stability programmes submitted by individual euro area Member States;

   • • is in a position to impose sanctions on members that do not comply with the rules, stopping short, however, of suspending membership.

4. 4. In both cases strong members have a keen interest in strict regulation and supervision, as membership is seen as a quality signal, making membership most attractive for comparatively weak members. In the case of the euro area, the SGP and the No-bail-out clause were proposed by the German government (Artis and Winkler Reference Artis and Winkler1997; Wyplosz Reference Wyplosz2009).

Given these similarities, it is striking how differently CBCHs reacted to periods of financial turbulence and how the euro area was designed to react in such times. The CBCHs perceived times of crisis as a market failure, with contagion being the most important form of failure requiring the need for joint action to replace market mechanisms. By contrast, the governance of the euro area explicitly wanted each Member State to deal with crisis situations on a stand-alone basis, i.e. without any assistance being provided by the ECB nor by (a group of) individual euro area governments or the European Commission. There are at least three reasons for this difference:

1. (a) The character of the institution. CBCHs, founded by nineteenth-century US commercial banks, and the ECB, founded by twentieth-century euro area governments, operate at different levels. While CBCHs issued (imperfect) substitutes for specie in extraordinary crisis times, the ECB was founded to issue high-powered money on a regular basis. Thus, the governance of the euro area had to deal explicitly with the problem of containing inflation, a matter completely absent in the CBCH case. This includes the Maastricht criteria, the SGP, the No-ECB credit to governments and the No-bail-out clauses (Artis and Winkler Reference Artis and Winkler1997).

2. (b) Absence of banking panics and other major financial turbulences in advanced economies. While inflation was a major issue in several EU Member States, banking or financial crises were extremely rare among EU Member States – just as they were in other mature economies – in the four decades preceding the signing of the Maastricht Treaty. Thus, achieving the goal ‘financial stability’, historically the main motivation for establishing a central bank, was not on the agenda when EU governments were designing the Treaty.

3. (c) Theoretical foundations. Euro area governance was largely designed on the basis of economic models which are void of market failures (Folkerts-Landau and Garber Reference Folkerts-Landau and Garber1992). The government was only seen as a major source of negative economic shocks, in particular for monetary policy, and had no role to play in stabilising the value of money (Goodhart Reference Goodhart1998). Accordingly, the Maastricht Treaty does not include instruments and regulations which prepare the euro area to deal with incidences of market failure, i.e. a loss of confidence leading to contagion effects (Pisani-Ferry Reference Pisani-Ferry2010). By contrast, the Treaty aimed at strengthening the disciplinary effects of capital markets on government finances. Moral hazard considerations suggest that these effects would only be forthcoming if the governments of euro area Member States could not count on any assistance from other governments, the Commission, or the ECB in times of crisis. Thus, the Maastricht Treaty introduced the two No-clauses referred to above (Wyplosz Reference Wyplosz2009).

Several observers continue to argue that the crisis reflects only government failure, explicitly ruling out contagion effects as a major component of the crisis (Issing Reference Issing2010c). Accordingly, the basic conclusions to be drawn from the crisis are the reinforcement of the No-clauses of the Treaty and strengthening the SGP. The only new element to be introduced would be a mechanism for the orderly restructuring of government debt (Gianviti et al. Reference Gianviti, Krueger, Pisani-Ferry, Sapir and von Hagen2010). This would create a rule for crisis management, at present lacking, without violating the principles of the Treaty.

However, the actual reaction of euro area governments to the crisis has become increasingly similar to the behaviour CBCHs displayed in the various US financial crises of the nineteenth century.

• • Contagion effects have become the major argument for joint action. This holds for the establishment of new institutions like the European Financial Stability Facility (Council of the European Union 2010) and the European Stability Mechanism (Eurogroup 2010), the latter explicitly introducing the conceptual distinction between illiquidity and insolvency of governments. It also holds for the interventions by the ECB on euro area government bond markets which were justified with reference to a market failure comparable to the one observed in money markets after the Lehman collapse (ECB 2010b; Trichet Reference Trichet2010; Bini Smaghi Reference Bini Smaghi2010a).

• • The risks of contagion were most acute when markets were hit by bad news about the severity of the shocks hitting the crisis countries and when markets perceived signals from euro area governments and the ECB that a joint initiative might not be forthcoming, implying that investors might face a haircut. The latter signals ran counter to the standard CBCH practice of intervening even on behalf of members who might indeed become insolvent. Consistently with this, euro area authorities quickly reversed their positions towards the CBCH practice when facing the corresponding market reaction.

• • The EFSF has become the twenty-first-century euro area equivalent to the nineteenth-century CBCH to fight the risk of contagion faced by member governments which provide maturity transformation services on a large scale. Like a CBCH, it replaces the open capital market by an internal market among Member States, i.e. allocating resources by nonmarket means (EFSF 2011).Footnote ¹⁵ The Facility will be replaced by the ESM in mid 2013.

• • EFSF and ESM would be copies of CBCHs if they were allowed to issue euro bonds in exchange for bonds issued by individual euro area Member States. Juncker and Tremonti (Reference Juncker and Tremonti2010) present a eurobond proposal that follows most closely the CBCH loan certificate model. They suggest that in times of crisis Member States would be allowed to finance up to 100 per cent of their debt via bonds jointly issued by a European Debt Agency (EDA), succeeding the EFSF.Footnote ¹⁶ Thus, bonds issued by this agency would represent the same extent of credit risk homogenisation among euro area countries as seen among clearing house members in the nineteenth-century US banking system. Concretely, weak Member States would be in a position to offer holders of their liabilities an exchange into euro bonds instead of cash. This would be equivalent to the CBCH case where weak member banks offered CBCH loan certificates instead of cash when confronted with a payment request by other CBCH members or depositors. A key difference would be the maturity of the joint liabilities issued: eurobonds would be long-term debt securities replacing long-term government bonds while CBCH loan certificates were short-term liabilities replacing short-term deposits. In general, however, euro area governance in times of crisis would become similar to that characterising the US banking system 150 years ago (Figure 4b) as the EDA would take a CBCH-like co-insurance role in times of crisis (Figures 4b and 5).

• • Euro area Member States receiving assistance from the EFSF and the ESM are subject to strong conditionality, just as CBCH members receiving assistance were put under a special monitoring and supervisory regime.

• • The legal foundations of euro area governments' response to the crisis are uncertain, comparable to the uncertainty surrounding CBCH actions in containing nineteenth-century financial crises.

• • The crisis response was slow, often following controversial discussions, and involved quantities that turned out to be insufficient. This is similar to the coordination problems the CBCH loan committees faced when deciding on the timing and size of financial assistance for member banks.Footnote ¹⁷

• • The crisis response is fragile. While euro area governments are not competitors in a strict sense, the crisis response has been incoherent reflecting the divergent interest of strong countries. Like strong member banks of CBCHs, strong euro area countries would like to limit contagion effects, but are hesitant to commit resources to bail out countries which might fail to honour their debt.Footnote ¹⁸ A concrete example was the decision by euro area governments to give the ESM preferred creditor status (Oakley Reference Oakley2011). While it protects the taxpayers of strong euro area countries in case of a weak euro area government's default, it also signals to private investors that default might occur. Thus, the decision reinforced the private investors' concerns about solvency that the ESM was designed to alleviate in the first place (Gros Reference Gros2010).

While the actual crisis responses of CBCHs and euro area governments have converged, there are still substantial differences. These include (1) the role of the ECB as an actual and potential LOLR, (2) the maturity of the liabilities being refinanced, and – closely related to this – (3) the moral hazard risks of the co-insurance mechanism.

Figure 5. Euro area governance with an ESM issuing euro bonds in times of crisis – a financial stability perspective

Source: author's compilation.

1. 1. In contrast to nineteenth-century US commercial banks, euro area governments have an institution that could serve as an LOLR to government bond markets, the ECB. Indeed, in early May 2010 the ECB started interventions via the Securities Market Programme (SMP), referring to signs of ‘severe tensions’ (ECB 2010b) in the markets, i.e. a liquidity crisis sparked by concerns about governments' solvency. However, interventions have been hesitant and limited, violating one of the key principles of an LOLR, namely to ‘lend freely’ to illiquid, but solvent institutions. Moreover, the ECB did not foster expectations that it would stand ready to intervene without limits in government bond markets. This is in stark contrast to the ECB's interventions in euro area money markets which started in August 2007. Thus, despite its interventions, the ECB has not properly performed an LOLR role for euro area government bond markets.

   There are three reasons for the different approach the ECB chose when addressing tensions in money markets compared to tensions in government bond markets.

   1. (a) The legal and self-imposed constraints on government bond purchases and sales as an instrument of monetary policy, i.e. the legacy of the Maastricht Treaty and the Bundesbank's choice of monetary policy instruments which did not include buying and selling of government bonds.

   2. (b) Dissenting opinions on the decision, most importantly by the President of the Deutsche Bundesbank, suggesting that interventions in government bond markets carry substantial risks to monetary stability (Weber Reference Weber2010).

   3. (c) A different exposure to credit risk. When intervening in the money market, ECB credit risk is limited by the implicit – and in the current crisis: explicit – promise of governments to bail out systemically important financial institutions. Such backing is not available for euro area government bonds as the No-bail-out clause precluded euro area governments from organising a co-insurance mechanism for government bonds.Footnote ¹⁹ Thus, when conducting interventions in euro area government bond markets, the ECB operates in a different credit risk setting compared to money market interventions.Footnote ²⁰

   From a financial stability perspective, the third reason is the most important one. Accordingly, the ECB has called for the rules and instruments of the European Stability Mechanism to be designed as flexibly as possible, in terms of quantities as well as qualities (Trichet Reference Trichet2011). This is because a strong and flexible ESM would enhance the co-insurance mechanism among euro area governments, i.e. it would maximise the probability that – although available in principle – ECB interventions will not be needed to address contagion effects. Thus, a strong ESM would allow a return to the separation principle between the monetary and fiscal policy of the original Maastricht Treaty for all but the most extreme scenarios of financial instability (Figure 5 – separation). Even in those extreme cases, a strong ESM would facilitate ECB interventions by exposing the central bank to a diversified credit risk encompassing all the euro area Member States (Figure 5 – reciprocal stabilisation) instead of the idiosyncratic credit risk represented by government bonds issued by weak euro area Member States only, as is currently the case.Footnote ²¹ Thus, a strong and flexible ESM would transform euro area governance either to one resembling the crisis mode of US CBCHs operating without an LOLR (Figure 4b), if the separation principle was reinstalled (Stella Reference Stella2010), or to one resembling a mature economy nation state with a central bank and government, the latter in the form of the ESM, providing each other with reciprocal stability services in terms of liquidity and solvency (Figure 1a).

2. 2. CBCHs co-insured short-term deposits backed by longer-term financial assets, while the EFSF and ESM will provide a co-insurance mechanism for long-term government bonds only, backed by the ability and the willingness of the respective countries' taxpayers to honour the commitments made by their governments. These differences suggest that the joint production of confidence by euro area governments will become a long-term project. Unlike CBCHs who were able to switch quickly from the ‘state-contingent integration’ (Kroszner Reference Kroszner2000) mode to pre-crisis mode of behaviour when the panic mood had vanished and member banks had repaid short-term clearing house loans and loan certificates, euro area governments have to be prepared to replace the external capital market by an internal one for a considerably longer period. The newly established ESM reflects this insight as it has been created as a permanent crisis mechanism replacing the EFSF, which was founded as a short-term crisis mechanism extending loans only until 2013.

   Moreover, the joint production of confidence is likely to be exposed to a higher degree of fragility. Unlike the case of commercial bank assets where market data provides evidence of a rise in value, the value of the underlying ‘asset’ government bonds are ‘financing’ is more elusive. Strong economic growth in the crisis countries may be the only convincing indicator to suggest that the value of these ‘assets’ has recovered to an extent that concerns about solvency vanish (Bini Smaghi Reference Bini Smaghi2011). Without growth, the required long-term assurance to markets that their doubts about weak countries' solvency are unfounded is likely to be repeatedly questioned. This concern reflects the fact that many market participants use the debt-to-GDP ratio as a key indicator to assess the solvency of governments. However, like the non-performing loan ratio in banking, this indicator is procyclical, sketching a scenario too rosy in the boom and too negative in the bust. As the policies of consolidation and adjustment pursued in weak euro area Member States are likely to imply several years of low growth, the debt-to-GDP ratio might continue to rise in the years to come despite significant progress in consolidation, reinforcing doubts about the solvency of weak euro area Member States. As these doubts might be shared by strong euro area governments providing emergency financing, markets will also question the willingness of strong governments to stand by their original commitments to support weak Member States. Thus, economic and political developments in Member States, in strong as well as in weak countries, could trigger new rounds of doubts about solvency, thereby perpetuating the crisis mode in which euro area governance will be operating. As a result, the fragilities the short-term CBCH crisis mode has already been subject to, i.e. the timing and the size of commitments, are more serious in the case of the euro area. This holds even more as euro area governments have not followed the CBCH example in issuing jointly guaranteed (euro)bonds which weak Member States could easily exchange for their previously issued liabilities. Thus, markets lack a strong and credible signal from euro area governments that could dispel doubts about the solvency of individual countries.

3. 3. The risk of moral hazard behaviour by weak members is substantially larger than in the CBCH case. There are two reasons for this. First, as just mentioned, financial assistance will have to be of a long-term nature in order to be credible in the first place.Footnote ²² Thus, the strong conditionality put on weak Member States must not only be exercised in difficult but also in good times. The SGP experience with sanctions suggests that this will be difficult to implement in the given political non-competitive environment. It is against this background that a strong and flexible ESM and the issuance of eurobonds face strong resistance from academics and policy makers alike (Franz et al. Reference Franz, Fuest, Hellwig and Sinn2010; Deutsche Bundesbank 2011 a, b). They fear that the main advantage of the ESM and eurobonds in times of crisis, namely the replacement of the open capital market by an internal one, will become the key disadvantage in normal times: fiscal policy in Member States would no longer be subject to the disciplinary forces of the market.Footnote ²³ As a result, many eurobond proposals (see Eijffinger Reference Eijffinger2011 for an overview) focus on limiting the moral hazard risks that could arise from jointly guaranteed bonds rather than on the benefits of the homogenisation of credit risk among euro area countries for crisis resolution. Indeed, some of the proposals (e.g. Delpla and von Weizsäcker Reference Delpla and von Weizsäcker2011) have little in common with the approach CBCHs took when responding to financial crisis 150 years ago. The second reason refers to the fact that unlike CBCHs, euro area governance does not operate the ultimate sanction, expulsion from membership, should a Member State default on its payment obligations to the ESM. Thus, a key instrument ensuring long-term incentive compatible behaviour of CBCH members is absent in the case of the euro area.Footnote ²⁴

VI

This article compares the experience of twenty-first-century euro area governments and nineteenth-century US commercial banks in times of financial crisis. The comparison is motivated by three observations:

1. 1. commercial banks and governments provide maturity transformation services;

2. 2. nineteenth-century US commercial banks and twenty-first-century euro area governments operate without a proper LOLR;

3. 3. commercial banks' liabilities and government debt may become subject to runs when macroeconomic shocks trigger doubts about the solvency of the respective debtors.

The comparison reveals a flaw in the original euro area governance framework as it ignores these similarities of banks and governments. This can partly be explained by the different forms of institutions US commercial banks and euro area governments established to economise on transaction costs: clearing houses on the one hand and a supranational central bank on the other. Reflecting this difference, the Maastricht Treaty design pursued good intentions, namely to protect the ECB and its conduct of monetary policy from the unsound fiscal policies of euro area Member States. However, it has led to a governance framework exposing euro area governments to a higher risk of becoming illiquid than commercial banks. This risk materialised in the aftermath of the financial crisis, when deficits and debt ratios of euro area governments rose substantially, reflecting measures intended to stabilise national banking systems and domestic economies. Given substantial differences in the magnitude of banking sector problems – relative to the countries' GDP –, as well as the size of the macroeconomic shock and government debt levels inherited from pre-crisis times, the deterioration of government finances raised concerns about the solvency of individual Member States.

As a result, the euro area has faced turmoil in government bond markets not seen in other mature economies with similar fiscal and economic challenges. The comparison with nineteenth-century US commercial banking suggests that the No-bail-out clause has proved to be the weakest link in the chain designed by the Maastricht Treaty to separate monetary from fiscal policy. Indeed, the first lesson euro area governments can learn from US commercial banks is that a strong and flexible co-insurance mechanism is a prerequisite for meeting the original goal of the Maastricht Treaty, namely protecting the ECB and its conduct of monetary policy in times of crisis from unsound fiscal policies pursued by individual Member States. This lesson applies regardless of whether the setting-up of the ESM will be accompanied by a return to the principle of strict separation between monetary and fiscal policies, i.e. imitating the no LOLR situation, or will serve as a step towards the model prevailing in most mature economies where government and central bank jointly guarantee their mutual liquidity and solvency. In the first case, timely and forceful ESM action, imitating that taken by CBCHs about 150 years ago, would provide the opportunity to calm markets and keep the interest rates of weak euro area Member States on levels consistent with long-term solvency without the help of ECB interventions. In the second case, the ESM would limit the ECB exposure to credit risk if interventions were deemed to be appropriate.

The CBCH experience offers a checklist for the design of a mechanism to allow for the joint production of confidence. This includes:

1. (a) the replacement of the open capital market by an internal market;

2. (b) the homogenisation of the credit risk of government debt issued by euro area Member States;

3. (c) the guarantee to bail out all Member States in times of crisis;

4. (d) the adjustment of rules and regulations governing the euro area in normal times in the light of the factors identified as crisis causes, i.e. going beyond a simple strengthening of the fiscal rules of the SGP (Pisani-Ferry Reference Pisani-Ferry2010).

The CBCH experience also provides the lesson that – at least in principle – there is no contradiction between the repeal of the No-bail-out clause and the establishment, as well as the enforcement, of strict rules and supervision of euro area Member States, including of fiscal and other economic policies. This holds in particular for Member States with government debt in distress. Adjustment and bail-out are two sides of the same coin. Finally, the CBCH example shows that a long and controversial debate about concrete steps to mitigate the crisis is an inherent characteristic of self-insurance mechanisms set up by peer institutions.

At the same time, the slow and hesitant response by euro area governments has failed to calm markets and to reduce interest rates on government bonds issued by weak euro area Member States. However, a high level of interest rates carries the risk that concerns about governments' solvency and the associated liquidity problems diagnosed in May 2010 will be transformed into a problem of insolvency, including the failure of a Member State to honour its commitments to the EFSF or the ESM. CBCHs experiencing such an outcome excluded the failing institution from membership after the crisis mood had vanished, while the remaining members shared the losses. For the euro area the first option is not available, largely for two reasons. Firstly, the euro is a not only an economic but also a political project. Secondly, given the economic and political implications of one country leaving the euro area it seems almost impossible to design an orderly exit of a Member State in normal times without triggering a new round of contagious effects (Eichengreen Reference Eichengreen2007). Against this background, it is striking that euro area governments have discussed intensively how to involve the private sector in burden sharing in any form, a step CBCHs explicitly rejected, while barely explaining to their electorate the need for loss-sharing arrangements among solvent euro area governments, should a Member State default on its obligations to the EFSF or the ESM. Finally, (1) the long-term nature of the liabilities to be co-insured, (2) the elusive character of the ‘asset’ backing these liabilities and (3) the moral hazard risks arising from both characteristics, as well the political environment the co-insurance mechanism will be placed in, present additional challenges for euro area governance reform which CBCHs did not have to deal with.

Overall, the comparison of the crisis response by nineteenth-century US commercial banks and twenty-first-century euro area governments suggests that in many respects the latter are following the former in containing financial panic. This is reassuring, as it indicates that the measures taken by euro area governments do not run counter to fundamental market principles, a claim often heard not only, but particularly, in Germany (Proissl Reference Proissl2010), but follow to a large extent best private-sector practices of crisis management. However, the comparison also suggests that the balancing act between the ‘hierarchy and maintenance of market incentives’ (Gorton and Mullineaux Reference Gorton and Mullineaux1987, p. 466) is much more difficult to perform for euro area governments today than it was for CBCHs 150 years ago.

Against this background, the argument of the Deutsche Bundesbank (1990) that a smooth functioning of the European Monetary Union is likely to depend on establishing a ‘more far-reaching association, in the form of a comprehensive political union’ receives new backing, albeit for a different reason than originally intended (Goodhart Reference Goodhart1998).Footnote ²⁵ The reason is that the nation state presents a clear solution to the challenge that euro area governments are currently facing by establishing hierarchy as a control mechanism in crisis and in normal times. Taking the comparison between governments and commercial banks to the extreme, the creation of a European nation state with a common electorate would be the equivalent of a merger between banks with a unified ownership to internalise the negative external effects of their behaviour – here: contagion – but avoiding the moral hazard risks of co-insurance (Dowd Reference Dowd1994). Moreover, establishing a European nation state could restore the financial stability governance structure of modern mature economies, with the central bank and the government jointly guaranteeing financial and monetary stability, even in the most extreme states of the world. Historical evidence suggests that the nation state does not provide a guarantee of success, as there have been many examples of failures to achieve this objective. However, so far it is the only solution to provide evidence that it can be done.