Submission to the Financial System Inquiry Appendix C: Should Bank Supervision be Carried Out by the Central Bank or by a Separate Authority?

Introduction

87. This appendix reviews the literature on the institutional separation of monetary policy and bank supervision and discusses the background to some recent proposals for institutional change. In summary, it finds three main arguments for bank supervision to be combined with monetary policy in the central bank and three offsetting arguments which advocate bank supervision being placed in a separate institution.
88. The three main arguments for bank supervision to be combined with monetary policy in the central bank are:
    • a central bank has responsibility for financial system stability, and will have to play a central role in the resolution of a financial crisis due to its capacity to quickly provide system liquidity or lender-of-last-resort loans. It should, therefore, always be in a position to assess the health of the banking system, rather than having to rely on a report from another body;
    • a thorough knowledge of the health of the banking system contributes to better monetary policy; and
    • a combined authority is likely to give appropriate weight to economic efficiency in framing regulation.
89. The three main arguments for a separation of bank supervision and monetary policy are:
    • ultimately it is taxpayers' money that is put at risk when banks are supported, so it should be taxpayers (through their government) who make decisions about bank rescues;
    • there is a potential conflict between the goals of monetary policy and supervision; and
    • the reputation of the bank supervisor would be damaged by a bank failure. If the supervisor is in the central bank, the loss of reputation would also affect the central bank's credibility in its monetary policy responsibilities.
90. As in so many such discussions in economics, there is no unanimity about which set of arguments are more powerful. Similarly, there is no single institutional model; some central banks have direct responsibility for supervision of the banking system, while others do not. A recent survey by the International Monetary Fund found that central banks are the primary bank supervisor in over 60 per cent of member countries (Tuya and Zamalloa (1994)). In many of the countries in which the central bank is not the primary supervisor, it devotes considerable resources to bank supervision. Among G7 countries, the central banks are the principal or joint supervisor in six countries; Canada is the exception. A fuller account of arrangements in the main industrialised countries is given in Appendix B. The remainder of this Appendix provides a brief overview of the evolution of central banks' supervisory responsibilities and reviews the debate on whether supervision should be conducted by the central bank.

Some history

91. Central banks have long had responsibility for both monetary policy and the stability of the financial system, although the latter historically has played a more important role in their formation. The original European central banks had two initial roles: providing financial concessions to the government and facilitating the development of the payments system (by controlling note issue and the country's metallic reserves) (see Goodhart (1988)). Their ability to provide cash eventually saw them take on the role of provider of liquidity to the banking system. From this responsibility evolved a broader responsibility for the stability of the financial system.
92. For countries establishing central banks in the twentieth century, the maintenance of financial stability has often been the prime consideration from the outset. Indeed, in the US, concern about the stability of the banking system following the banking problems of 1907 was the principal reason for the founding of the Federal Reserve. The importance of supervision in maintaining system stability is reflected in the preamble to the Federal Reserve Act (1913) which calls on the Federal Reserve to establish more effective supervision of banking (see Volcker (1984)).
93. In Australia, the preservation of financial stability was also a factor in the origins of central banking, although it played a less explicit role than in the US. The proximate reason for providing the Commonwealth Bank with central-bank-type responsibilities (in 1924) was to gain greater control over note issue. This move took place in an environment in which the private banks were seen by many people as a major source of financial and economic instability, particularly after the banking problems in the 1890s. The establishment of central-bank functions in the Commonwealth Bank was seen as a tentative step towards creating a more stable financial system.
94. The Banking Commission of 1937 explicitly acknowledged that the central bank had a role in supervising, and maintaining the integrity of, the financial system. The Commission argued that the central bank should play an active role in the supervision of the system and should take control of any bank which was unable to meet its immediate obligations. In the following years, banks became subject to greater regulatory control, and by the time the Banking Act (1945) was drafted, depositor protection was seen as a key responsibility of the Commonwealth Bank. This responsibility was transferred to the Reserve Bank with the 1959 Banking Act (see Dwyer (1985) and Schedvin (1992)).
95. Prior to widespread deregulation of financial systems, prudential requirements were often a by-product of the regulations designed for macro-economic management. Controls over the structure of banks' balance sheets and their interest rates were designed to limit risk taking by banks and to serve a macro-economic role. As the financial system was deregulated, the mechanisms of monetary control became more clearly separated from those of prudential supervision. Deregulation also saw the focus of supervision move towards monitoring and understanding credit and market risk, rather than imposing restrictions on the activities that financial institutions can undertake.

Public inquiries and academic discussion

96. Recent years have seen increased attention given to the question of whether both monetary policy and bank supervision should be in the central bank, or whether bank supervision should be in a separate institution. Much of the debate has been stimulated by proposals for regulatory reform following problems in the banking industry in a number of countries. As a result, the main contributions have been from official sources – central banks, parliamentary committees and government departments and agencies. Academics have also contributed to the debate, but many have been more interested in the wider issue of whether regulation of the financial system is needed in the first place.

Official positions

97. The most extensive debate has occurred in the US. The issue received considerable attention in the early and mid 1980s as a result of the Task Group on Regulation of Financial Services Report (Bush (1984)). This group argued that removing the Federal Reserve from involvement in bank supervision might ‘undercut the long-term stability of the financial system’ (p. 48). Seven years later (1991), the US Treasury argued for a raft of reforms which included a proposal for two regulatory agencies, with the Federal Reserve being responsible only for the state-chartered banks. More recently, the Clinton Administration proposed to merge the four bank regulators into a single banking regulator outside the control of the Federal Reserve system. This latest proposal was largely motivated by the desire to reduce supervision and to reduce the possibility of ‘regulatory shopping’.
98. On each occasion the Federal Reserve argued strongly in favour of maintaining its supervisory role (see for example, Volcker (1984), Corrigan (199la), Federal Reserve Board (1994), Greenspan (1991, 1994) and Syron (1994)). The line of argument has been that the Federal Reserve's system-stability obligations require it to have practical knowledge of the banking system and the authority to influence banking organisations' actions. The Federal Reserve has also argued that the information gained as part of its supervisory responsibilities is important in the formulation of monetary policy, and that a stand-alone supervisor, without macroeconomic responsibilities, would have a long-term bias against risk taking, which would inhibit economic growth (see later).
99. The Bank of England's role in supervision has also been subject to considerable debate, again largely in response to the failure of financial institutions. In 1984, as a result of the collapse of Johnson Matthey Bankers, the government commissioned a report into supervision arrangements. The Committee and a subsequent government White Paper recommended that bank supervision remain with the Bank of England, and that this role be strengthened by the creation of the Board of Banking Supervision within the Bank (see UK Treasury (1986)). The Bingham (1992) Inquiry into the collapse of BCCI re-affirmed this arrangement.
100. In 1993 another inquiry was conducted, this time by the Treasury and Civil Service Committee of the House of Commons (this inquiry also considered the role of the Bank of England in monetary policy). It concluded that ‘there is no overwhelming case for separating out the responsibility for prudential supervision to a separate body’ (p. xxviii). The Committee recognised that the main argument against a separate regulator was that the Bank of England needed to undertake some oversight of the banking system in its capacity as the lender of last resort. It also concluded that conflicts between monetary policy and supervision are rare, and that when they do arise, they do so regardless of who is responsible for supervision.
101. In the wake of the collapse of Barings there were renewed calls for the Bank of England to lose its supervisory function. Again, the Bank of England argued against such a move with Governor George saying that he thought there were strong links between monetary policy and supervision, and that there is a synergy between ‘prophylactic supervision’ and the lender-of-last-resort function of central banks (Australian Financial Review (1996)).
102. In Norway, a parliamentary working party was established in 1992 to review regulatory arrangements (see Norges Bank (1992) and Tuya and Zamalloa (1994)). The working party recommended moving supervision into the central bank by incorporating the Banking, Insurance and Securities Commission into it. While the proposal was supported by the central bank, it was rejected by the parliament, on the grounds that there was a conflict of interest between supervision and monetary policy, and that it would centralise too much power within the central bank. The Banking, Insurance and Securities Commission therefore remains separate.
103. There have been two recent examples – Finland and Hong Kong – of responsibility for bank supervision being shifted into the central bank from an outside authority, both in 1993. The background to these cases is discussed in detail in Appendix B. In both cases, the decision reflected dissatisfaction with the performance of the previous supervisory arrangements and a view that there were important synergies between bank supervision and monetary policy responsibilities.
104. While most central banks have argued that they should be responsible for bank supervision, not all have done so. The Bundesbank is the most frequently quoted example. Tietmeyer (1991) has argued that ‘a successful monetary policy does not require that the central bank itself be given full control over the banking system and thus for banking supervision; indeed, this could even reduce the effectiveness of monetary policy. Experience also shows that central banks that are not charged with such additional responsibilities enjoy a higher degree of de facto independence’ (p. 185). This public statement is weakened by the fact that the Bundesbank has more staff involved in monitoring the condition of the banking industry (including bank inspections) than are employed in the main supervisory authority.
105. The Bank of Canada stands out because it has virtually no role in bank supervision. In 1986, the Inquiry into the Collapse of the CCB and Northland Bank (the Estey Inquiry) examined the structure of bank supervision in Canada. In submissions to the Inquiry, the Bank of Canada argued that there was an inherent conflict between the conduct of monetary policy and financial supervision. The Inquiry's Report also noted that there was no support in any of the submissions for the Bank of Canada to have control of supervision but did point out a number of problems arising with the Canadian arrangement (see later for more discussion).^[32]
106. The Swedish central bank also has little role in bank supervision. A 1989 committee of inquiry examined the question of whether the stand-alone bank supervisor (as it was at that time) should be brought under the central bank, and recommended against such a move. The main reason for the recommendation was a view that it could undermine the perceived independence of the central bank, because bank supervisory rulings would be subject to administrative review by other parts of the government.
107. The involvement of the central bank in supervision has also been extensively debated in South Africa. In the early 1990s, the South African Reserve Bank (which only began supervising banks in 1987) was criticised when a number of financial institutions failed. Subsequently, the Bank became a defendant in litigation when investors who had suffered losses claimed financial assistance from the Bank. While the Reserve Bank had no formal responsibility for deposit protection, it felt that there was a misguided perception that it had a ‘duty of care’ to protect depositors. It argued that this perception could undermine its credibility as a monetary authority.
108. In response, the government commissioned two inquiries. The Jacobs Commission on Equal Competition for Funds recommended that the government establish a Financial Regulation Policy Board and that the Registrar of Deposit-Taking Institutions (the Reserve Bank's supervision department) report to this Board, rather than to the Reserve Bank. The subsequent Melamet Inquiry into Financial Supervision argued that the government should establish a single regulatory body for the entire financial system (the Financial and Investment Services Commission). While both inquiries envisaged the Reserve Bank losing its supervisory responsibilities, this did not occur. The Registrar of Deposit-Taking Institutions remains within the Reserve Bank, although the government did establish a Policy Board to help formulate and co-ordinate regulatory policy (see South African Reserve Bank (1992 and 1993)).

Academic discussion

109. Principal contributions by academic economists have come from Goodhart and Schoenmaker (1993, 1995), Mishkin (1992 and 1994) and Goodfriend and King (1988). Goodhart and Schoenmaker conclude that there is no overwhelming case in either direction, with the strongest argument in favour of combined functions resting on the role that the central bank plays in preventing/resolving payment and financial system problems. On the other hand, they argue that the central issue in deciding who should be responsible for supervision should be who pays the bill in case of a bank failure. They conclude that the government's role as the ultimate source of funds swings the balance towards separate institutions, although they acknowledge that the institutions would need to work closely together. Against this, Schoenmaker (1995) has recently argued that since the European Central Bank will be the only body that could alleviate a general liquidity crisis in the European monetary system, it will need to be closely involved in bank supervision and need to develop a capacity to understand and monitor the potential for systemic risk.
110. Mishkin (1992) comes down strongly in favour of the central bank having supervisory responsibilities. There are three planks to his argument. First, he argues that financial crises are defined by periods in which asymmetric information becomes so intense that financial institutions are not able to perform their intermediation function adequately, with adverse implications for the real economy. Second, in such periods, the central bank has a legitimate and important responsibility to provide directed liquidity to the system through lender-of-last-resort loans. Third, if such loans are to be made, the central bank needs to have regulatory oversight of potential borrowers to reduce moral hazard problems and to ensure that borrowers are financially sound.
111. In general, economists who see an important role for the lender-of-last-resort facility also see the central bank as undertaking bank supervision (see, for example, Brimmer (1989) and Calomiris (1993)). In contrast, those who believe that there is no compelling rationale for the public provision of lender-of-last-resort loans, argue that a central bank should not have supervisory responsibilities. Perhaps the most frequently cited paper is Goodfriend and King (1988). They argue that the existence of lender-of-last-resort loans leads to an increase in risk taking, and that the system is better served if open market operations are used to inject liquidity into the system in a (potential) crisis.^[33] They conclude that because the central bank should not make lender-of-last-resort loans, it does not need to undertake supervision. By inference, a similar conclusion would probably be drawn by those who have recently argued for the abolition of the discount window in the US (for example, see Kaufman (1991) and Schwartz (1992)). These authors argue either that the lender-of-last-resort function should be performed by the private sector, or alternatively, replaced by central bank open market operations. The usual argument against the lender-of-last-resort facility is that when deciding to make loans the central bank will err by supporting insolvent institutions. This will blunt market discipline and produce riskier less-efficient banks (see Bordo (1990)).

The arguments in detail

112. The various issues and arguments can be classified under three broad, but related, headings:^[34]
     • the role that the central bank plays in preserving system stability;
     • the synergies and conflicts between supervision and monetary policy;
     • organisational and cultural issues.

The role of the central bank in preserving system stability

113. The strongest argument in favour of the central bank having supervisory responsibilities is that it is uniquely positioned to prevent and resolve financial system crises. By virtue of its macro-economic responsibilities, its daily involvement in financial markets and its central position in the payments system, the central bank has the ability to react quickly to financial disturbances. This makes it the best-placed institution to protect the real economy from financial system shocks.

A financial crisis?

114. Defining a financial crisis is difficult but a central feature must be that a crisis has the potential to cause a contraction in economic activity (see Corrigan (199lb) and Mishkin (1994)). Mishkin argues that the definition needs to incorporate some reference to the process of financial intermediation. He argues that: ‘a financial crisis is a disruption to financial markets in which adverse selection and moral hazard problems become much worse, so that financial markets are unable to efficiently channel funds to those who have the most productive investment opportunities’ (p. 9).
115. The financial system can be subject to disturbances from many sources: the collapse of a bank through poor commercial loans, a disturbance in securities markets, problems in the payments system and rumours concerning the solvency of a bank. Whether or not these disturbances constitute a financial crisis depends on their potential to undermine the process of financial intermediation. In some cases, a bank failure will have no implications for the rest of the financial system. In other cases, confidence effects and the direct links through the payments system can have important implications for other institutions. By increasing uncertainty as to the creditworthiness of counterparties, financial disturbances have the potential to undermine the confidence that institutions have in their credit assessment procedures and thereby damage the process of financial intermediation, perhaps with severe macro-economic consequences.
116. It is usual to think of a financial crisis as resulting from a bank failure (usually due to it making poor commercial loans). This is the sort of situation that is discussed in most of the literature dealing with crisis management (see later). However, it is not the only disturbance which could threaten to create a financial crisis. Increasingly, the source of the disturbance is something that occurs in a financial market, such as the ones listed below (see, for example, Brimmer (1989), Calomiris (1993), Corrigan (199lb), Mishkin (1994) and Federal Reserve Board (1994)):
     1. The failure of Penn Central Railroad (1970). The failure of Penn Central to meet its commercial paper obligations caused the commercial paper market to seize up. As a result, firms with maturing obligations could not issue new notes and therefore sought to borrow from banks. In these uncertain circumstances, there was some unwillingness on the part of banks to extend credit until the Federal Reserve made it clear to the banks that it would provide the necessary liquidity through the discount window, although the banks remained responsible for the credit risks involved in lending to their business borrowers. (A similar experience occurred in Sweden in 1990; when the commercial paper market crashed, even well-managed non-financial companies found it difficult to raise finance (Davis, 1992).)
     2. The Ohio and Maryland thrift problems (1985). ESM, a small securities firm in Florida, defaulted on its loans to an Ohio thrift when it incurred losses in its government securities trading business. This triggered runs on privately insured thrifts in Ohio and Maryland. The institutions were closed and their reopening was conditional on their being found eligible to access the Federal Reserve's discount window and obtaining federal insurance. Under the Federal Reserve's direction, examiners entered the institutions to evaluate assets that might serve as collateral and to monitor currency outflows. As a result, the Federal Reserve was able to expeditiously provide discount window loans and meet all demands for currency. Corrigan (199lb) argues that the problems in these thrifts came very close to producing full-scale grid-lock in the entire mortgage-backed securities market, and had the potential to produce systemic damage.
     3. The stock market crash (1987). The fall in share prices meant that large margin payments had to be paid by brokers. As a result, brokers sought extra loans from their bankers. However, as banks became more risk averse and more concerned about their capital ratios, they became less willing to extend this credit. This threatened a collapse of the clearing and settlements system and the possible failure of securities firms. To counteract this, the Federal Reserve announced that it was ready to serve as a source of liquidity to the system, making it clear that it would provide discount loans so that the banks could lend to their brokerage clients.
     4. The Drexel Burnham failure (1990). The problems at Drexel meant that market participants were extremely reluctant to deliver securities or to make payments to Drexel. This reluctance could have brought the liquidation of Drexel's mortgage-backed securities to a halt. Had this occurred, capital markets would have been interrupted and the financial system would have become more vulnerable, as all players became much more uncertain about the creditworthiness of their counterparties. This could have led to a major disruption to the process of financial intermediation. Given this risk, the Federal Reserve used its knowledge of the payments system, the institutions involved, and its close working relationship with key personnel to develop procedures with the banks and securities houses that allowed the orderly winding down of Drexel's securities.

117. In each case, developments outside the banking system threatened the stability of the financial system. Credit rationing intensified and liquidity threatened to dry up. In each case, the Federal Reserve's understanding of the linkages between securities markets, the payments system and commercial banking, and the linkages between the financial system and the macroeconomy, allowed it to design responses which removed these threats at no cost to the taxpayer.
118. Looking forward, the rapid pace of technological change and innovation in financial markets increases the chance that other shocks will originate outside the banking sector, but quickly involve the banks in their transmission to the wider economy. When new instruments are evolving rapidly, the potential for a re-evaluation of their pricing is high, particularly if the pricing is extremely complicated. Calomiris (1993) notes that one of the reasons that the failure of Penn Central had such a large effect on the commercial paper market was that the market had grown so quickly in the 1960s that few had stopped to evaluate the risks and appropriate pricing of commercial paper. When the problem arose, risk assessments had to be updated and few were sure how to do this. Information about credit-worthiness and the real value of commercial paper became scarce. Davis (1992) suggests a similar situation occurred in Sweden in 1990. The repricing of junk bonds in the US in the early l990s is another example.

The role of the central bank

119. Many economists argue that a central bank with supervisory responsibilities is the right starting point for preserving a stable and efficient financial sector. In a crisis, the central bank can inject liquidity into the market through open market operations, or in more extreme circumstances, lender-of-last-resort loans. It can also have an important impact on developments through what it says. Statements along the lines that ‘the Bank will ensure that the system has sufficient liquidity’ or that ‘in the Bank's opinion, a particular institution is solvent’ can, in certain circumstances, supply the information that the market is seeking, and head off a potential crisis.^[35] The central bank can also play a leadership role in negotiating solutions to payments problems which threaten to bring down the system, and in negotiating the merger of a solvent institution with an institution that has closed.
120. Could another institution perform these roles of crisis containment and resolution? In principle it could, but in practice there would be difficulties. In many crisis situations prompt action is needed and this requires that the responsible institutions have an understanding of:
     • the potential macro-economic implications of the crisis;
     • the operation and dynamics of financial markets;
     • the potential problems in the payments system; and
     • the creditworthiness of the banks with which it is dealing.

121. For a central bank, the first three requirements are satisfied as a by-product of its other responsibilities, and the final requirement is met through it having supervisory responsibility for the banking system. An institution that was solely a supervisor would not meet the first three requirements, although it could try to keep itself well informed. It would also not have the daily contact with markets needed for prompt action, and so would have to rely on the central bank if this was required.
122. Knowledge of the quality of banks' balance sheets is essential, not only in responding to shocks originating in the banking system, but also to those in securities markets. If there is some shock that leads to liquidity drying up in a securities market, non-financial firms may turn to banks for liquidity support. Failure to provide that support might lead to major problems for non-financial firms. In such a case, the correct response may be for the central bank to lend to the private banks so that they have the funds for short-term liquidity support of solvent, but illiquid non-financial firms. In Mishkin's terminology, the central bank would be using the banks as its ‘delegated monitors’ – that is, the private banks would be monitoring the creditworthiness of their customers, while the central bank monitors the creditworthiness of the banks to which it is lending.
123. The central bank may also be helpful in leading a rescue package for troubled institutions, or in attenuating fears that a problem in one part of the system could spread to other parts. In the US examples discussed above, the financial markets and the banking sector looked to the Federal Reserve to ensure that the financial disturbances did not turn into financial crises. Where central banks do not have supervision responsibilities, a stand-alone supervisor could play that role, although perhaps less well (see next section).
124. There are numerous examples of central banks co-ordinating rescue packages for troubled banks. Perhaps the most well-known example is that of the Bank of England organising the ‘Lifeboat’ to support troubled ‘secondary banks’ during 1974–75, and more recently during 1991–93, when several small and medium-sized banks suffered a withdrawal of wholesale deposits. The Bank was also responsible for the rescue of Johnson Matthey Bankers in 1984; at the time, it feared that the failure of this institution could threaten London's position as the leading international gold bullion market. Other examples of central banks playing a leadership role in resolving bank failures include the Banca d'Italia (Banca Steinhauslin – 1981/82), De Nederlandsche Bank (Friesch-Groningse Hypotheekbank – 1982) and the Reserve Bank of Australia (Bank of Adelaide – 1979). Interestingly, in Germany in 1983, when Schröder Münchmeyer Hengst failed, it was the Bundesbank which initiated the rescue package, bringing in the Federal Banking Supervisory Office at a later stage (see Dale (1992)). Similarly, it was the Bundesbank which organised the winding up of Bank Herstatt in 1974. In contrast, in Belgium, Denmark and Switzerland, banking commissions have played the leadership role (for more details see Goodhart and Schoenmaker (1993)).
125. Goodhart and Schoenmaker (1993) examine 104 banking crises in a range of countries and conclude that where the central bank remains in charge of supervision, bank rescues are ‘somewhat more likely’ to be solved with financing from commercial banks, rather than the public purse. However, they also argue that as competition in banking increases, the central bank (or anyone else) may find it more difficult to ‘encourage’ private banks to help fund a rescue.
126. The international orientation of central banks is another argument in favour of assigning them supervisory responsibility (see Tuya and Zamalloa (1994)). This argument has two parts. First, financial shocks are increasingly likely to have an international element, so that it is helpful if the supervisory authority has an understanding of international financial markets and exchange rates. The central bank is likely to be better placed than other agencies, by virtue of its foreign exchange market operations and its responsibility for investing foreign reserves. Second, central banks are in frequent contact with one another and have well-established formal and informal lines of communication. This makes timely recognition of adverse developments in banks with international activities more likely. These relationships may also prove useful in formulating an international response to a world-wide financial disturbance.

A stand-alone supervisor and system stability

127. The alternative proposition to the one argued above is that the system stability objectives can be best satisfied with a stand-alone supervisor who is answerable either to parliament or to the Minister of Finance (see for example Hilton (1994) and Taylor (1995)). Proponents of this approach argue that the information collected by the supervisor could be passed onto the central bank where appropriate.
128. This argument is explicitly rejected by most central banks, and implicitly rejected by many others, who, even through they do not have formal responsibility for bank supervision, employ a large staff examining the health of the banking system. As the Federal Reserve (1994) notes ‘central banks in all but one G-7 country, in most cases de jure but always de facto, are closely involved with the supervision of banks in their countries and internationally’ (p. 22).
129. The one exception is Canada. Despite having lender-of-last-resort responsibilities, the Bank of Canada is reliant on the Office of the Supervisor of Financial Institutions (OSFI) for supervision of the banking system and assurances as to an individual bank's solvency. The Bank of Canada is comfortable with this arrangement and supports it publicly.
130. Critics of this type of arrangement see a number of problems. First, in a crisis, time is of the essence. Slight delays in the transmission of information or any ambiguity about its contents could have damaging effects. Second, without direct responsibility for supervision, the central bank's practical ‘hands-on’ feel for what is happening would diminish. It would not have the person-to-person contacts that would enable it to broker a deal in a crisis such as accompanied the 1987 share market crash in the US. The Federal Reserve has argued that if its staff were forced in to the position of simply reading a stand-alone supervisor's reports, or even being relegated to the junior member of a supervisory team, ‘the tendency would to be retreat into a kind of ivory tower, adversely affecting both monetary and supervisory policy’ (Volcker (1984) p. 549).
131. This point is illustrated by the German and Japanese arrangements. In Germany, while the Federal Banking Supervisory Office (FBSO) is the formal banking supervisor, the Bundesbank carries out on-site inspections and reviews and follows up on reports by external auditors before they are passed to the FBSO. Despite not being the formal supervisor, the Bundesbank employs more supervisors than the formal supervisory agency. A similar situation occurs in Japan, where the Ministry of Finance is the statutory supervisor, but the Bank of Japan conducts regular on-site inspections, and has a large supervision staff. This duplication of supervisory activities increases the total cost of bank supervision.
132. The Canadian experience in the mid 1980s provides an example of the practical problems which can arise when the central bank has no supervisory responsibilities. The Estey Inquiry noted that when CCB and Northland Bank were experiencing difficulties, ‘The Bank of Canada … found itself in an invidious position in the events surrounding the collapse of CCB. The Governor of the Bank of Canada was seen as the leader of the banking system. Naturally, therefore, he was looked to for leadership in times of crisis. Indeed, it was taken for granted by all participants that the Governor of the Bank of Canada was the appropriate person to preside over the 22 and 24 March meetings to determine the fate of CCB. Unfortunately, the Bank of Canada is not clothed with the necessary statutory powers or staff to select the appropriate program in such circumstances and to guide its performance.’ (Estey (1986), p. 165).
133. The Canadian example illustrates the point that if the central bank has no supervisory responsibility, it has little option but to rely on the advice of the supervisor. In this episode, the supervisor originally assessed the banks as solvent and the Bank of Canada arranged loans for the banks. The Bank of Canada was active in informing the public that the banks were solvent and in arranging liquidity support from other banks. This was despite its having no independent way of assessing the solvency of the institutions. Subsequently, the supervisor reversed its decision and the Bank of Canada stopped extending credit. The initial determination that the banks were solvent meant that depositor protection had to be extended to all depositors, not just insured deposits (those less than $60,000).
134. Some economists support the Canadian-style system on the argument that the central bank is already a powerful institution, by virtue of its monetary policy responsibilities and that granting it supervisory responsibilities gives it too much power and reduces the ‘checks and balances’ in the system. Another argument is that since taxpayers' money is often put at risk in bank rescue operations, it is the taxpayers, through their representative (the government-controlled supervisor), who should make any decision.
135. This latter line of argument can be criticised on a number of grounds. First, the taxpayer is going to be at risk whether the rescue is financed by a stand-alone supervisor or the central bank because the government owns both institutions. It is true that a stand-alone supervisor would probably be more directly answerable to the political process than a central bank, but that could cause a second problem in that it would increase the chance that political decisions would dominate economic ones. In Taylor's (1995) proposal for a stand-alone supervisor in the UK, he argues that the Treasury would need to indemnify the Bank of England for any support funds it provides at the request of the FSC (Financial Services Commission). This leads him to conclude that ‘The Treasury would therefore need to be closely involved in any decision to provide LLR support, although it would act on the advice of the FSC and the Bank’ (p. 14). This implies that a committee representing three organisations, with at least one member under possible political pressure, would have to make a decision about whether to lend to a bank under extreme time constraints. Finally, the perceived independence of the central bank may be damaged by a system whereby the government directs the central bank to make loans to banks, or alternatively, by the central bank being obliged to extend credit to the supervisor (so that it can make loans to banks). The notion that the central bank does not lend to the government (or its agencies) and that the government does not direct the central bank to make loans to private firms is a cornerstone of responsible monetary management, and should not be discarded lightly.

Are there conflicts or synergies between supervision and monetary policy?

Conflicts

136. The argument is that if a central bank is the bank supervisor, then it will have divided loyalties, which will interfere with its monetary policy objective of maintaining low inflation. The usual situation envisaged is where the monetary authority wishes to raise interest rates for anti-inflationary purposes, but the regulatory authority opposes it because of the weakness of the banking system. This argument suggests that if the two functions are separate, the monetary authority would go ahead and raise interest rates, regardless of what happens to the banking system. On the other hand, if the two functions are in the same institution, an internal compromise would be reached whereby interest rates would not be raised as much (or at all), resulting in a less anti-inflationary monetary policy.
137. A problem with this argument is that it seems to imply that the monetary authority should not take into account the condition of the banking system in setting monetary policy. But if it failed to, it could inadvertently bring about a financial crisis which could have large effects on economic activity and inflation, and hence result in worse monetary policy than otherwise. One way or another, the condition of the banking sector is a relevant piece of information for monetary policy and should be factored into the decision making process (see later for more detail). The argument for separation would have to rest on a different and more restricted form of conflict, whereby the combined institution deliberately chooses a setting of the monetary instruments which will give a sub-optimal economic outcome (including higher inflation), but will ‘prop up’ some weak banks and hence preserve the combined institution's reputation as a supervisor. This more precise form of the argument recognises that it is correct to take into account the condition of the banking system when deciding on monetary policy settings, but believes that a combined institution may misuse the information to serve its own self-interest.
138. Is there any empirical evidence for the idea that central banks have acted in this way? To the best of our knowledge, neither the literature on the subject, nor our observation of recent history, provide any such evidence. Australia and New Zealand, for example, which both have monetary policy and bank supervision combined in the central bank, went through a period of anti-inflationary monetary policy at a time of substantial bank losses in the period between 1989 and 1992. Looking back over that period, no-one has suggested that either central bank ‘went easy’ on its monetary policy in order to ‘prop up’ its banks and so preserve its reputation as a supervisor. To the contrary, most current views of the monetary policy of that time (which have benefited from hindsight) accuse it of having erred on the tough side.
139. There are two episodes in the US that are sometimes quoted (see Goodhart and Schoenmaker). The US abandoned the non-borrowed reserve base scheme for setting monetary policy in 1982 and moved to a more pragmatic system, which quickly led to a fall in interest rates. Many people attribute this to worries about the solvency of US money-centre commercial banks as a result of their exposure to LDC debt, but virtually no-one now feels that the change in monetary policy procedure was a mistake, or that it was inflationary. In the early 1990s, the Federal Reserve was accused of giving a high weight to the weakness of the Savings and Loans Industry (S&Ls) in setting its monetary policy. Again, there are few who would now say that monetary policy was too easy at that time, and, besides, the Federal Reserve was not the supervisor of the S&Ls and could not be accused of trying to protect its own reputation.
140. An alternative empirical approach has been provided by Heller (1991) who found that the average inflation rate for countries with separation of supervision and monetary policy was lower than for countries with both combined in the central bank. This result was necessarily based on a small sample, and heavily influenced by some Latin American countries where the central bank was responsible for bank supervision. Goodhart and Schoenmaker (1993) did their own compilation excluding these countries and still found that the average inflation rate was lower for countries where the central bank was not the bank supervisor. This is not surprising since Germany, Switzerland, Japan and the US, which have relatively low inflation, were in the group of countries where the central bank is not the sole bank supervisor. But the central banks in each of these countries do have some responsibility for bank supervision, although it is shared with other institutions (see Attachment B). They, therefore, would have some reputation as a supervisor to protect, so the comparison between the two groups is an ambiguous one.
141. Neither author claims much rigour for this type of exercise and Goodhart and Schoenmaker sum up by saying, ‘Our final conclusion is therefore that a Central Bank's involvement in supervision does not necessarily weaken its stance on monetary policy; and consequently we consider a Central Bank's inflation performance and its role in supervision as two, more or less, separate issues’.

Synergies

142. For monetary policy to achieve its objectives, it is important that the central bank has a good understanding of the transmission mechanism – the chain of events that lead from a change in the instrument of monetary policy to the final outcomes in terms of inflation and output. This is not an easy task as there is no mechanical link that applies over time; each cycle is different in intensity and the length of lag.
143. One important element in the transmission process is how the banking sector will react in its vital role of providing credit. The provision of credit always involves taking risks, and banks' attitude to risk will depend importantly on the health of their balance sheets. Will they pass on to their customers the full increase in official interest rates? Will they still compete aggressively for new business? Will they stay with their present credit standards, will they lower them to get new business, or will they raise them to protect themselves from future losses? In different circumstances, the results will be different. For example, in 1988 and 1989, the banks seemed oblivious to high interest rates and still competed vigorously for business, lowering credit standards by more than they had in earlier cycles. On the other hand, they became extremely risk averse in the 1990-to-1993 period, despite the lowest interest rates for 20 years.
144. Thus, a given tightening of the monetary instrument will have different effects depending on the health of the banking system. If it is in good shape, a tightening may produce a modest slowing, but if it is in bad shape, it could result in a severe contraction. In order to do the job of monetary policy properly, it is important to have as thorough a knowledge as possible of the health of the banking system, and this would be most efficiently achieved by being the bank supervisor.
145. The argument that monetary policy will be better conducted if the central bank does not have responsibility for supervision amounts to saying that it will do a better job if it does not have access to full information. The alternative view – and the one the RBA would support – stresses the synergies between the two responsibilities. It believes that monetary policy would do a better job if it had all the available information at its disposal.
146. A counter to the above argument is to agree that the central bank should take into account the health of the banking system, but to point out that it could receive its advice on this matter from a separate supervisor rather than doing the job itself. Such advice would obviously be a help, but it is hardly the same thing as having direct knowledge of the situation. A stand-alone supervisor with no economic responsibilities or market involvement is bound to have a different perspective from an economic agency such as a central bank. The stand-alone supervisor would probably concentrate on judging the health as it was at a point of time, rather than assessing its susceptibility to possible changes in monetary policy.
147. The RBA attitude to this question is perhaps jaundiced by its recollection of the Pyramid Building Society episode. The Reserve Bank had no supervisory responsibilities in that case and had to rely on the assessments of the Victorian Registrar of Building Societies.

Organisational and cultural issues

The reputational argument for separation

148. A good supervisor does not attempt to prevent individual banks from failing if they are badly managed and if the failure has no significant systemic implications. However, this is not widely perceived by the public, who tend to regard any failure as a sign of incompetence on the part of the supervisor. Supervision is a thankless task – there being no rewards during the long period of stability but often severe recriminations from the public, the press and politicians if a failure occurs.
149. For this reason, some commentators feel that it would be better if the reputation of only the supervisor was sullied when an institution failed, rather than having it rub off on to the monetary policy arm of a central bank as well. This is an argument which has been used often in the case of the UK, where the Bank of England has been severely criticised for the failure of Johnson Matthey Bankers, BCCI and Barings. There is some weight in this argument, although it has never been strong enough for a central bank to propose divesting itself of its supervisory functions.

Cultural and incentive issues as arguments for a combined entity

150. There are two arguments here of relevance, both of which favour a combined organisation rather than separation. The first argument is that a single regulator with a focus on safety, and subject to the same sort of pressures described above, will have a long-term bias towards excessively tight supervision and will give less weight to the importance of flexibility and capacity to adjust to market forces. It would receive little, if any, credit for the ability of the banks to adjust to market forces and contribute to growth, but would be subject to severe criticism if an institution failed. These incentives could lead to supervisory policies which, at the margin, inhibit economic performance and possibly encourage the growth of financing outside the supervised sector. This is an argument that has been put regularly by the Federal Reserve in the US. It sees itself as having a better capacity to trade off safety versus adaptability than a more-narrowly focused regulator.
151. A related argument is that a stand-alone supervisor would develop a culture based on rules and their interpretation, and hence tend to be dominated by lawyers, and possibly accountants. In a central bank, however, while the supervisory wing may have some of these tendencies, it will have to battle constantly with the economists and financial markets professionals in the monetary-policy wing of the central bank. This creative tension would tend to reduce the extent to which an overly rules-based and legalistic approach was followed.
152. On balance, it is hard to weigh up these arguments, and much will depend on the point of view of the person doing the weighing up. If the aim is light regulation, and emphasis is on the need for market forces to play a large role, a combined supervisor would have the advantage. On the other hand, if the emphasis is on minimising the risk of failure, a separate supervisor would probably be superior.

Footnotes

Crow (1993) provides a brief summary of the Bank of Canada's approach to financial stability. [32]

See Corrigan (1991b) and Summers (1991) for why open market operations are not sufficient to deal with a crisis. [33]

Goodhart and Schoenmaker (1995), Tuya and Zamalloa (1994) and Federal Reserve Bank (1994) provide the most detailed discussion of the arguments. See also Shull (1993), Swinburne and Castello-Branco (1991) and de Swaan (1994). [34]

As was done in the cases of the Bank of Melbourne and Metway Bank in 1990, and the Rural and Industry Bank of Western Australia in 1992. [35]