Central banks and financial oversight

Speech by Mr Fernando Restoy*, Chairman, Financial Stability Institute, Bank for International Settlements, at the Fundación Ramón Areces,  Madrid, Spain, 4 June 2018.

BIS, FSI speech  | 
18 June 2018


I would like to start by thanking the organisers for inviting me to participate in this event. It is certainly a pleasure to share this stage with Emilio Albi and Julio Segura on the occasion of the presentation of the study Reguladores y Supervisores económicos independientes: el caso español, coordinated by the latter.

I have been asked to focus my remarks on the general topic of the institutional organisation of financial supervision. I am happy to do so, although having discussed this issue so many times with the coordinator of this study, I have difficulties in adding new angles to the ones already covered in the volume.

I thought, however, that it could be of interest for this audience to focus on how supervisory models have changed since the financial crisis and, in particular, how the role of central banks has evolved in that regard. For that purpose I will draw on a recent publication by the Financial Stability Institute (Calvo et al (2018)) in which we analyse the changes observed in the organisation of financial sector supervision during the last decade across a broad sample of jurisdictions.

That piece of evidence will help me to offer some reflections on potential implications of the enhanced role of central banks in areas other than monetary policy with regard to their statute as independent authorities.

The choice of supervisory model

Financial supervisory agencies have traditionally performed mainly two types of function: the oversight of the safety and soundness of different types of financial institution, such as banks, insurance companies and securities firms (the microprudential function): and the protection of investors, which typically includes the supervision of conduct of business rules of intermediaries and the preservation of securities market integrity. Following the crisis, that set has been expanded by two new functions: the monitoring and mitigation of risks and vulnerabilities of the financial system as a whole (the macroprudential function); and the resolution of vulnerable banks able to generate systemic stress.

A supervisory model essentially consists of the allocation of these functions to different agencies, together with the definition of the institutional features of each agency (independence, governance, etc) and possible coordination mechanisms among them.

Existing supervisory models can be broadly classified into three different types: sectoral, integrated and partially integrated. In the sectoral model, agencies are responsible for the prudential and conduct of business oversight of all institutions of a specific sector: namely, banking, insurance and securities markets. Under the integrated model, at the other extreme, a single authority (located inside or outside the central bank) is responsible for both the prudential supervision and customer protection for all types of financial institution as well as for market integrity. The partially integrated models group responsibilities in different authorities according to different criteria. One example of the partially integrated model is the Twin Peaks model, now operating in countries like Australia, the United Kingdom or the Netherlands, where two authorities are responsible for the prudential and conduct of business oversight, respectively, of all institutions. Another example of a partially integrated model is the bi-sectoral model (or Two-Agency model as we call it in the FSI Insights paper) in which one authority is responsible for the prudential and conduct of business oversight of both banks and insurance companies and a second authority oversees securities business and markets. This model prevails, for example, in France and Italy.

The choice of supervisory model is typically the product of a range of considerations. In particular, the experience accumulated from the Great Financial Crisis has played a role in highlighting the political economy implications of having different allocations of responsibilities within the financial supervisory architecture. Yet, models score differently in two relevant dimensions: i) their ability to exploit synergies across different functions by grouping them within a specific authority: and ii) their power to minimise conflicts across objectives by assigning them to different institutions. The latter is, in principle, as important as the former, since conflicting objectives within the same authority may lead to the subordination of one to the other in a way that may not be socially desirable. Examples of possible conflicts of objectives are those arising from the simultaneous oversight of banks' solvency (as pursued by the microprudential function) and the protection of customers' interests (as part of the conduct of business oversight function).

The role of the central bank

An important feature of the supervisory models is the role assigned to central banks. There is plenty of literature analysing the pros and cons of giving central banks powers in the prudential domain (eg Goodhart (2003) or Wymeersch et al (2012)). On the positive side, there exist clear synergies between monetary policy and banking supervision as banks' balance sheets - which are directly affected by monetary policy decisions - are key components of the monetary transmission mechanism. Moreover, central banks' role as - mostly the sole - emergency liquidity provider makes them a key actor in banks' crisis management. That role can hardly be performed effectively if central banks do not have sufficiently comprehensive and timely information on the situation of individual banks, although, arguably, this may not require direct supervisory responsibilities if there is effective coordination between central banks and supervisors.

More generally, it appears to be an excessive analytical simplification to consider economic and financial stability as two different objectives that could be effectively pursued by different agencies, acting separately, using different instruments, such as interest rates for the former and prudential policies for the latter. Using mathematical terminology, the system of objectives-instruments is not composed by independent but by simultaneous equations and cannot normally be resolved recursively.

It is true, though, that the involvement of central banks in supervision may also generate conflicts of objectives. In particular, it has been argued that a specific mandate to preserve banks' solvency - or financial stability more generally - may crowd out the fundamental objective of the monetary authority, which is to pursue price (or economic) stability, and expose the central bank to legal or reputational risks that can jeopardise its ability to perform its core monetary policy function effectively.

In addition, the assignment of tasks relating to financial stability to the central bank may pose risks of a political economy nature. The pursuit of financial stability typically entails not only responsibilities in the fields of microprudential supervision, macroprudential policy and resolution but also regulatory actions that can be quite intrusive with respect to the structure and functioning of the banking industry. While there could be a strong case for assigning most, if not all, of these powers to one or more independent authorities, the argument has been made that the combination of these functions with monetary policy-making within the central bank may entail trade-offs between social objectives that are not supposed to be addressed without the involvement of elected authorities.

In sum, despite the analytical appeal of the integration of financial sector responsibilities within the central bank, the accumulation of power within a single agency may pose issues of political legitimacy (Tucker (2018)) which could end up being resolved by introducing excessive constraints on central banks' mandates and operational procedures. I will come back to this in a minute.

The facts: the evolution of institutional arrangements for financial supervision

How have the above considerations affected the organisation of financial oversight in the recent past?

As I have mentioned, we at the FSI have recently conducted a study on the evolution of institutional arrangements for financial supervision in a large sample of jurisdictions. The study shows that most jurisdictions have implemented incremental changes within existing supervisory models. These reforms include new macroprudential and resolution frameworks; stronger consumer and investor protection rules; and improved coordination for the monitoring of financial stability. However, the number of jurisdictions that have shifted their supervisory models, as I have defined them before, after the crisis is relatively limited. Indeed, out of the 79 jurisdictions examined, the institutional arrangements for the performance of the traditional financial oversight functions have changed since 2007 in 11 countries; 13 if we include South Africa and China, where new regimes are still in the implementation phase.

Data show that the sectoral model is still the prevailing approach around the globe, as half of the jurisdictions still follow this model of separation of supervisory responsibilities for banks, insurance companies and securities markets. The exception across geographical areas is Europe, where the model of an integrated supervisor - located either inside or outside the central bank - is the most common approach. Interestingly, partially integrated models (the Twin Peaks) are only adopted in one fifth of the jurisdictions.

Yet the study shows that most changes observed in the supervisory models - seven out of 11, or nine out of 13, if we include South Africa and China - have consisted in moving away from the sectoral model and implementing a more integrated approach by merging together different supervisory responsibilities within a smaller number of authorities. Indeed, both the single supervisor model, when this is located within the central bank, and the partially integrated model (both Twin Peaks and Bi-sectoral) have gained relevance.

Another salient finding of the analysis is that central banks have gained competences in financial oversight. Central banks are not only the prudential supervisor for banking in two thirds of the jurisdictions but are also performing that role for insurance companies in an increasing number of countries (from 14 before the crisis to 22 currently). In addition, central banks have become the primary macroprudential authority in close to 60% of the jurisdictions, although often other authorities are also involved by participating in inter-agency committees. Furthermore, central banks also host the resolution authority in close to 60% of the surveyed countries.

The study also shows that while most jurisdictions have strengthened their frameworks for the supervision of conduct of business by financial intermediaries, only a few of them have actually created agencies specialised in consumer or investor protection. Indeed, although it is employed in several relevant financial centres, the Twin Peaks model that explicitly envisages the separation of prudential and conduct of business responsibilities in two different agencies is followed in only eight jurisdictions (or nine if we include South Africa).

Therefore, the crisis has not led to a big overhaul of the institutional arrangements for financial supervision around the globe. Yet it seems to have strengthened somewhat the case for seeking more synergies across supervisory functions by pursuing a higher integration of responsibilities in the supervisory authorities. More importantly, central banks have tended to gain substantive competences in the microprudential, macroprudential and resolution domains, thereby pointing to the quest of ensuring an adequate coordination of the monetary policy, emergency lending, prudential oversight and resolution functions, particularly in crisis situations.

This phenomenon, together with the limited actions undertaken to separate conduct of business from prudential responsibilities, suggests that the objective of minimising conflicts of objectives within agencies has received a lower weight than the need to exploit complementarities, when deciding on the adjustments to be made in supervisory models.

The debate on central bank independence

The assignment of additional responsibilities to central banks has coincided with the re-emergence in the public arena of the debate on central bank independence, in which not only academics but also public officials have participated (Den Haan et al (2017), Mersch (2017), Issing (2018)).

Interestingly, that debate has not been triggered by a possible underperformance of central banks in helping stabilise the economies, in accordance with their mandates, during or after the Great Financial Crisis. Quite the contrary, it seems to be generally accepted that central banks have effectively contributed to reducing the economic stress in most industrialised economies and facilitated the recovery by pursuing ultra-accommodative monetary policies over a sustained period of time.

The debate seems to have been related rather to precisely the political and social implications of the instruments used by central banks in order to provide the economy with the desired degree of monetary accommodation. In particular, the argument has been made that by purchasing large amounts of public debt in secondary markets as part of their quantitative easing programmes, central banks were de facto performing a quasi-fiscal function. Moreover, it has been stressed that those policies have had significant distributional effects across different segments of the population (eg between investors and debtors and between investors in different types of assets). The above arguments do not seem to point to new undesirable effects of unconventional monetary policies but simply stress that the effects of accommodative monetary policies - both intended and unintended - have become especially pronounced in a context in which the degree of accommodation required to stabilise the economies was particularly elevated. Yet there appears to be currently a higher social sensitivity to possible unintended distributional effects of public policies which may have some influence on the public debate on central bank independence.

The debate is also taking place in a context of protracted moderation of inflationary pressures despite the generalised economic recovery and the situation of almost full employment in some large jurisdictions. In that conjuncture, the absence of an immediate inflation threat has made the traditional arguments that supported the case for central bank independence in the academic literature somewhat less influential in the current policy debate.

In any event, the ongoing addition of responsibilities to central banks in the current context of growing scepticism about central banking independence may pose risks for the preservation of the degree of autonomy that central banks around the globe had gained prior to the Great Financial Crisis.

In my view, the arguments against government interference in tasks like the conduct of monetary policy or prudential supervision remain fully valid. Moreover, as we have seen, the experience with the Great Financial Crisis indicates that there could be strong arguments for avoiding an artificial separation of central banks from tasks related to the preservation of financial stability. At the same time, it may make sense to seek ways of adjusting the policy framework somewhat in order to mitigate the political economy concerns related to the accumulation of power by central banks and to weaken the arguments in favour of increasing the control over their decisions and actions by governments or parliaments.

In that regard, a number of proposals have been recently made to enhance transparency and accountability (Eichengreen (2015), Best (2016), Tucker (2018), Powell (2018)), to establish specialised committees within central banks for decisions in different policy domains (Tucker (2018)), to put in place coordination mechanisms with fiscal and other government authorities in specific circumstances (Balls et al (2016)), among other contributions. All these ideas are worth exploring in depth.

Yet, while there are good arguments for assigning central banks responsibilities in the financial stability domain, there could also be strong reasons to withdraw from their mandates certain functions for which no obvious synergies exist with monetary policy. This may certainly be the case, for example, of the responsibilities that many (actually most) central banks hold for the protection of consumers of banking (and in some jurisdictions other financial) services. Those responsibilities also entail, as I have argued before, some conflicts with the obligation to ensure that the banking system remains safe and sound. That function could therefore well be transferred to a specialised agency outside the central bank.

Finally, the current debate on the possibility that the central bank could issue digital currencies or, more generally, accept deposits from the public must also be analysed in the context of its implications for the central bank's role in the economy, as it could affect the case for maintaining its functional independence. It is clear that a massive transfer of customers' deposits from commercial banks to the central bank and the corresponding enlargement of its balance sheet would make it unavoidable for the central bank to make sensitive decisions on the allocation of all those additional resources across sectors and agents. The significant extra power that central banks would thereby acquire may further affect the willingness of politicians and possibly also the population at large to leave the functional autonomy with which central banks currently operate unchanged.

To conclude, some adjustments to the current operational and governance procedures as well as some restraint in the specific functions assumed by central banks may help in striking the right balance between the benefits of the expansion of their mandates beyond price stability and the need to maintain their legitimacy as independent authorities.


Balls, E, J Howat and A Stansbury (2016): "Central bank independence revisited: after the financial crisis, what should a model central bank look like?", M-RCBG Associate Working Paper Series, No 67, November.

Best, J (2016): "Why we need better central bank accountability", speri.comment: the political economy blog, 7 July.

Calvo, D, J C Crisanto, S Hohl and O Pascual Gutiérrez (2018): "Financial supervisory architecture: what has changed after the crisis?", FSI Insights on policy implementation, no 8, April.

Den Haan, W, M Ellison, E Ilzetzki, M McMahon and R Reis (2017): "The future of central bank independence: results of the CFM-CEPR survey", VoxEU, 10 January.

Eichengreen, B (2015): "How can we improve central bank transparency?", Weforum, 6 January.

Goodhart, C (2003): "The organizational structure of banking supervision", Economic notes, vol 31, no 1, pp 1-32, December.

Issing, O (2018): "The uncertain future of central bank independence", VoxEU, 2 April.

Mersch, Y (2017): "Central bank independence revisited", keynote address at the "Symposium on building the financial system of the 21st century: an agenda for Europe and the United States", Frankfurt, 30 March.

Powell, J (2018): "Financial stability and central bank transparency", speech at the Sveriges Riksbank anniversary conference on "350 years of central banking: the past, the present and the future", Stockholm, 25 May.

Tucker, P (2018): Unelected power, Princeton University Press.

Wymeersch E, K Hopt and G Ferrarini (2012): Financial regulation and supervision: a post-crisis analysis, Oxford University Press.

*       I am grateful to David Archer and colleagues at the FSI for helpful discussions and comments on an earlier draft. The views expressed are, in any case, exclusively my own and do not represent those of the BIS or the Basel-based standard-setting bodies.

Related information