Abstract
This Selected Issues paper discusses capital and liquidity regulations in Sweden. It recaps the recent debates on capital and liquidity buffers, and discusses a way to consider appropriate levels of capital and liquidity buffers in the case of Sweden. The paper estimates the government’s contingent liabilities from banks by different capital and liquidity levels. Also examined are options for Sweden in case the authorities face constraints to set buffers at their desirable levels.
II. Developing An Effective Macroprudential Policy Institution in Sweden1
1. Through the mid-2000s, Swedish banks actively expanded outward cross-border activity, including in the Baltics, with increasing reliance on short-term foreign currency funding. This raised vulnerabilities in Swedish banks just ahead of the 2008–09 global crisis period. Much of subsequent strains could have been attenuated if an appropriate institutional framework for macroprudential policy had been in place.2
A. Lessons Learnt from the 2008–09 Crisis
2. The Baltics was “subprime” for the Swedish banking system. In 2008–09, two of the largest banks, both increasingly funded on wholesale markets and exposed to the Baltics, faced sharp increases in loan losses and nonperforming loans, with their ratings marked down accordingly. Market concern toward these two banks spilled over to the Swedish banking system as a whole, as indicated by an increase in a join default probability indicator, thus the problem became a major systemic issue (for more details, see Ishi, 2010).
3. In response, the authorities implemented a range of bold crisis intervention measures. The Riksbank lowered the policy rate to a zero bound and introduced various new liquidity measures, including a US dollar lending facility. To boost international reserves, the National Debt Office (NDO) borrowed externally ($15 billion equivalent), and the Riksbank tapped US Fed’s and ECB’s currency swap arrangements. Meanwhile, the government doubled the deposit guarantee and introduced new bank recapitalization and debt guarantee schemes. These concerted efforts proved successful, and a full-brown crisis was prevented successfully. However, important questions remain.
Did the authorities appropriately anticipate a systemic risk and take preemptive measures?
4. The initial signs of the imbalances in Baltic economies, such as high credit growth and rapid housing price increases, were identified as early as 2005, but at that time, this was not recognized as a source of a systemic risk. The Riksbank noted that “borrowing in the Baltic states is still rising, as are defaults, but the ability to service debt is still judged to be good,” (Financial Stability Report 2005:1). Similarly, Finansinspecktionen (FI) noted, “the large banking groups were financially strong, and risks to systemic stability were low for the foreseeable future.” It also expressed caution about a potential risk but only in a longer perspective, “the expansion of their business abroad and in countries like the Baltic states and Russia may imply increasing risks,” (The Stability of the Swedish Financial Sector, 2005).
5. By contrast, Baltic authorities expressed their concern about growing imbalances at that time. The Estonian authorities attempted to persuade banks (mostly Finish, German, and Swedish bank affiliates) to slow lending growth and tighten lending standards, but without success. And they sent a letter to home supervisors, including the Riksbank and FI, to tighten regulations in December 2005. However, the home supervisor opposed to a change, noting that this would not be consistent with EU-wide harmonized rules (The Swedish National Audit Office, 2011).
6. In 2007, the authorities hinted with greater emphasis the potential risks arising from the Baltics, but decisive actions were not taken. The Riksbank assessed “risk lies in developments in the Baltic states, in particular Latvia but also Estonia and Lithuania. Signs of economic overheating there are becoming increasingly clear,” (Financial Stability Report 2007:1). FI reported, “the Baltic economies, albeit to varying degrees, are judged to be overheated… banks have enough capital to handle a situation that could entail extremely large credit losses in any of these countries,” (The Stability of the Swedish Financial Sector, 2007). Both the Riksbank and FI indicated at that time that Swedish banks had ample buffers to manage deteriorations in credit quality in the Baltics.
7. In the middle of 2008, the Riksbank expressed concern that “the risks in the Baltic countries have not diminished…. During the next result period, the banks may experience additional negative impacts.” And it added, “there is also a risk that this will affect capital strength In the event of continued unfavorable developments, liquidity risks will probably increase,” implying that banks needed more capital and liquidity buffers. But the Riksbank was apparently hesitant to make a formal recommendation to FI on this regard. In June 2008, FI’s Board discussed whether banks’ capital be raised. But it assessed the level of banks’ capital as adequate, and no action was taken (The Swedish National Audit Office, 2011).
8. Meanwhile, banks started raising capital in markets. However, as September 2008 approached, strains in financial markets increased sharply. At that point, it was too late to apply preemptive measures to reduce systemic risks.
Why did the authorities fail?
9. Sweden’s experience reveals challenges in effectively identifying and diagnosing systemic risk, and making intervention in a timely and effective manner.
When signs of building-up risks or growing imbalances are detected, there is substantial uncertainty as to how large they are and what problem they pose on a system. Thus, it is difficult to decide whether a countermeasure is warranted, let alone for policymakers to explain to stakeholders the need for such a measure (the situation in 2005–06).
When risks continue to grow, a consensus can be easily reached that some action is needed. However, it is still hard to reach a consensus on how serious the problem is and what measures should be applied. Moreover, given that imbalances typically grow in tandem with an economic boom, taking a counter measure is politically unpopular and could even face social backlash. Thus, policymakers tend to wait and see until clear risk emerges (the situation in 2007).
The risks could further grow to a point that most feel such risks. However, policymakers could face dilemma of weighing benefits and costs of bursting a bubble. And often, it is still hard to reach a consensus on which measures should be applied and by how much, especially if there still remains a view of an optimistic denial (the situation in the first half of 2008).
10. This said, in the case of Sweden, the delayed policy response partly reflected a drawback in its own institutional setting. Sweden’s financial stability framework is highly decentralized and composed of several independent agencies. FI has a statutory mandate for financial stability and to promote consumer protection. The Riksbank, on the other hand, does not have in its charter an explicit financial stability mandate but is responsible for promoting a safe and efficient payments system and actively provides analyses on financial stability by periodic publication of its Financial Stability Reports. The MOF bears the ultimate political responsibility as the fiscal authority but is also responsible for legislation in the financial sector and plays a leading role in crisis management (currently, the MOF chairs an inter-agency standing crisis management group). The NDO manages the stability fund, deposit insurance, and investor protection systems. It also serves as the support authority when public funds are conferred to credit institutions.
11. The main problem is that although each of these agencies is highly professional in addressing issues in their responsible areas, there is neither a formal coordination and accountability framework nor a incentive mechanism to coordinate.3
While the Riksbank has comparative advantage and resources in diagnosing and identifying risks from macro-financial perspectives, it does not have sufficient tools to address macro-financial risks;
Meanwhile, the FI focused on financial stability at an institutional level and consumer protection and did not devote large resources to overall macro-financial stability issues.
Thus, the Riksbank and the FI have partly overlapping tasks but have different ways and means of performing these tasks from different perspectives.
Moreover, pre-crisis, while regulatory reform initiatives in the EU moved towards more centralization in supervision frameworks, there was no EU-wide macroprudential policy framework. Now the European Systemic Risk Board (ESRB) has been established, an important issue would be how best to integrate Sweden’s macroprudential oversight into the European framework.
B. Alternative Institutional Models for Macroprudential Policy
12. Reflecting lessons from the crisis, a consensus is building that a new institutional framework that solely focuses on macro-financial stability and associated macroprudential policies will be needed in Sweden. As an interim first step, in January 2012, the Riksbank and FI signed a Memorandum of Understanding and newly established a council for cooperation on macroprudential policy. The council is a forum for information exchange and consultation, including for financial sector risk assessment and for macroprudential policy making.
13. Nonetheless, this initiative should go further.4 A formal macroprudential authority will need to be established by law. Such authority must be autonomous (particularly from the financial industry and political influence) and should have clear and strong mandate, responsibilities, and accountability. This is particularly important because macroprudential policy could suffer from an inaction bias (ESRB, 2012): the success of macroprudential policy measures is hard to observe; while macroprudential policymakers would have to take unpopular decisions by taking away the punch bowl just as the party gets going.
14. There is no one-size-fit-all model for all jurisdictions. While there is a general consensus that central banks should play a leading role given their expertise in macro- financial analysis and their role in preventing a crisis as emergency liquidity providers (IMF, 2011 and ESRB, 2012), the choice of a specific institutional set-up may depend on country’s context, reflecting differential financial structure, regulatory architecture, as well as historical and political factors. This said, in general, a strong macroprudential authority should be able to (i) effectively identify, analyze, and monitor systemic risks; (ii) timely and effectively use macroprudential tools; and (iii) (if macroprudential policy powers are shared) make coordination among responsible agencies effective in risk assessment, while preserving their individual autonomy (Nier et al, 2011).
15. Berntsson and Molin (2012) have proposed five alternative models that could be considered for the case of Sweden (Table 1). The first two models involve the establishment of a new macroprudential policy council, and the remaining three models assume the existing institutions to be unchanged.
Alternative Macroprudential Institutional Models (Berntsson and Molin, 2012)
Alternative Macroprudential Institutional Models (Berntsson and Molin, 2012)
| Model | Comments |
|---|---|
A. A new decision making macroprudential council under the Riksdag |
|
| A. A new decision making macroprudential council under the Riksdag |
|
| B. A new recommending macroprudential council under the Riksdag  |
|
C. Riksbank issues recommendations and FI makes a decision |
|
D. Riksbank is responsible for macroprudential policy |
|
| E. FI is responsible for macroprudential policy  |
|
Alternative Macroprudential Institutional Models (Berntsson and Molin, 2012)
| Model | Comments |
|---|---|
| A. A new decision making macroprudential council under the Riksdag |
|
| A. A new decision making macroprudential council under the Riksdag |
|
| B. A new recommending macroprudential council under the Riksdag |
|
| C. Riksbank issues recommendations and FI makes a decision |
|
| D. Riksbank is responsible for macroprudential policy |
|
| E. FI is responsible for macroprudential policy |
|
16. The main strengths and weaknesses of each model are evaluated in Table 2, based on the following six criteria: (i) whether a model ensures macroprudential policy autonomy; (ii) whether a model provides for efficient identification, diagnosis, and resource use; (iii) timely and effective use of tools; (iv) efficient use of a mix of tools; (v) accountability; (vi) democratic control; and (vii) whether a model facilitates international coordination.
Model A appears to be most effective and efficient if the council is chaired by Riksbank governor to ensure its autonomy. The council—consisting of representatives of the Riksbank, FI, and experts—has both a risk identification/diagnostic role and decision making powers on macroprudential tools. The council has a clear responsibility, and thus it is easier to establish its accountability framework. Given that the council includes representatives of both the Riksbank and FI, their resources can be efficiently used, while the inclusion of experts would help enhance democratic accountability. A key weakness is that the council does not have decision making powers on microprudential tools (as FI’s operational autonomy should be respected).
Model B is a weaker form of Model A, because a council does not have any decision making powers. Hence, the risk identification/diagnostic function is separate from the decision making function, weakening timely use of tools and accountability.
Models C-D appear to be further weaker than Models A-B. In each of these models, resources of the Riksbank and FI will not be used efficiently, and the risk identification/diagnosis function and the decision making function are separate.
In Model E, FI has full authorities on microprudential and macroprudential policies. While this model has advantages (e.g., FI is clearly responsible for macroprudential policy, should be accountable, and can decide on an efficient mix of macroprudential and microprudential tools), there are a number of weaknesses. The main drawbacks include: (i) this model will not ensure the autonomy of macroprudential policy; and (ii) the Riksbank’ s resources will not be used.
Strengths and Weaknesses of Different Models
Strong if Council (chaired by the Riksbank) or the Riksbank is a macroprudential policy decision making body.
Strong if both the Riksbank and FI are involved in macroprudential risk identification and diagnosis.
Strong if a risk identification agency and a decision making agency is the same on macroprudential policy.
Strong if a single agency has both macroprudential and microprudential tools.
Strong if a risk identification agency and a decision making agency is the same on macroprudential policy.
Strong if the Riksbank (representing at European Systemic Risk Board) leads macroprudential policy.
Strengths and Weaknesses of Different Models
| Model A | Model B | Model C | Model D | Model E | ||
|---|---|---|---|---|---|---|
| Primary agency responsible for macroprudential risk identification | Council | Council | Riksbank | Riksbank | FI | |
| Decision making agency | ||||||
| Macroprudential tools | Council | Riksbank | FI | Riksbank | FI | |
| Micro prudential tools | FI | FI | FI | FI | FI | |
| Macroprudential institutional autonomy 1/ | Strong | Strong | Weak | Strong | Weak | |
| Efficient identification, diagnosis, and resource use 2/ | Strong | Strong | Weak | Weak | Weak | |
| Timely and effective use of tools 3/ | Strong | Weak | Weak | Strong | Strong | |
| 28 Efficient use of a mix of tools 4/ | Weak | Weak | Strong | Weak | Strong | |
| Accountability 5/ | Strong | Weak | Weak | Strong | Strong | |
| Democratic control | Likely not an issue | Likely not an issue | Likely not an issue | Riksbank may be seen as too powerful | Likely not an issue | |
| International coordination 6/ | Strong | Strong | Strong | Strong | Weak | |
Strong if Council (chaired by the Riksbank) or the Riksbank is a macroprudential policy decision making body.
Strong if both the Riksbank and FI are involved in macroprudential risk identification and diagnosis.
Strong if a risk identification agency and a decision making agency is the same on macroprudential policy.
Strong if a single agency has both macroprudential and microprudential tools.
Strong if a risk identification agency and a decision making agency is the same on macroprudential policy.
Strong if the Riksbank (representing at European Systemic Risk Board) leads macroprudential policy.
Strengths and Weaknesses of Different Models
| Model A | Model B | Model C | Model D | Model E | ||
|---|---|---|---|---|---|---|
| Primary agency responsible for macroprudential risk identification | Council | Council | Riksbank | Riksbank | FI | |
| Decision making agency | ||||||
| Macroprudential tools | Council | Riksbank | FI | Riksbank | FI | |
| Micro prudential tools | FI | FI | FI | FI | FI | |
| Macroprudential institutional autonomy 1/ | Strong | Strong | Weak | Strong | Weak | |
| Efficient identification, diagnosis, and resource use 2/ | Strong | Strong | Weak | Weak | Weak | |
| Timely and effective use of tools 3/ | Strong | Weak | Weak | Strong | Strong | |
| 28 Efficient use of a mix of tools 4/ | Weak | Weak | Strong | Weak | Strong | |
| Accountability 5/ | Strong | Weak | Weak | Strong | Strong | |
| Democratic control | Likely not an issue | Likely not an issue | Likely not an issue | Riksbank may be seen as too powerful | Likely not an issue | |
| International coordination 6/ | Strong | Strong | Strong | Strong | Weak | |
Strong if Council (chaired by the Riksbank) or the Riksbank is a macroprudential policy decision making body.
Strong if both the Riksbank and FI are involved in macroprudential risk identification and diagnosis.
Strong if a risk identification agency and a decision making agency is the same on macroprudential policy.
Strong if a single agency has both macroprudential and microprudential tools.
Strong if a risk identification agency and a decision making agency is the same on macroprudential policy.
Strong if the Riksbank (representing at European Systemic Risk Board) leads macroprudential policy.
17. If Model A or B is chosen, a representative from the MOF can be included in the council (though with no right to vote), without undermining the autonomy of the Riksbank and FI. In general, advantages and costs of including the MOF can be summarized as following:
Potential advantages. The MOF can help creating new macroprudential tools through legislative changes or contribute to mitigation of macroprudential risks via tax changes. Furthermore, given that the MOF is ultimately responsible for a crisis management and resolution (which involves taxpayers’ money), its views should also be listened to by macroprudential policy makers. In addition, the inclusion of the MOF in the macroprudential policy making loop could facilitate a smooth transition to a crisis management operation, if a crisis happens.
Potential costs. The MOF is a government agency led by a politician. Thus, the participation of the MOF (even without right to vote in the committee) could (or be perceived to) undermine the autonomy of macroprudential policy making.
18. Berntsson and Molin (2012) argue for the merits of keeping the MOF at arm’s length to ensure the autonomy of macroprudential policy decision in Sweden. In other jurisdictions (e.g., UK, and see Table 3), the MOF has some role (though limited) in a macroprudential policy committee. Were the MOF is excluded from the membership of the council, accountability and transparency obligations of the council would need to be designed even more strongly to ensure the council’s democratic accountability.
Stylized Models for Macroprudential Policy
Stylized Models for Macroprudential Policy
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 | |
|---|---|---|---|---|---|---|---|
| Degree of institutional integration of central bank and supervisory agencies | Full (at a central bank) | Partial | Partial | Partial | No | No | No |
| Ownership of macroprudential policy mandate | Central bank | Committee related to central bank | Independent committee | Central bank | Multiple agencies | Multiple agencies | Multiple agencies |
| Role of fiscal authorities | No | Passive | Active | No | Passive | Active | No |
| Separation of policy decisions and control over instruments | No | In some areas | Yes | In some areas | No | No | No |
| Existence of separate body coordinating across policies | No | No | No | No | Yes | Yes | No |
| Country examples | Ireland (new) | UK (new) | France (new) US (new) | Belgium (new) | Australia | Canada | Iceland |
Stylized Models for Macroprudential Policy
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 | |
|---|---|---|---|---|---|---|---|
| Degree of institutional integration of central bank and supervisory agencies | Full (at a central bank) | Partial | Partial | Partial | No | No | No |
| Ownership of macroprudential policy mandate | Central bank | Committee related to central bank | Independent committee | Central bank | Multiple agencies | Multiple agencies | Multiple agencies |
| Role of fiscal authorities | No | Passive | Active | No | Passive | Active | No |
| Separation of policy decisions and control over instruments | No | In some areas | Yes | In some areas | No | No | No |
| Existence of separate body coordinating across policies | No | No | No | No | Yes | Yes | No |
| Country examples | Ireland (new) | UK (new) | France (new) US (new) | Belgium (new) | Australia | Canada | Iceland |
19. Furthermore, some incentive mechanism needs to be built in to reflect a voice of regional countries in macroprudential policy making in Sweden. Prior to the crisis, warning signals from the Baltics were not effectively heard of by Swedish authorities, in part reflecting “asymmetry”: while the size of Swedish banks’ operations ranged from 60 to over 100 percent of GDP in each Baltic country, it accounted for less than 6 percent of GDP in Sweden. Accordingly, subsidiaries may have received less attention from the parent banks and home supervisors due to their low impact on banks’ financial positions as groups, until unexpectedly large shocks in Baltics were recognized (for more discussions, see Ishi, 2010).
C. Tentative conclusions
20. The authorities’ intention to formally establish a macroprudential policy authority is welcome. For a country with diversified supervision architecture like Sweden, one of the important aims in developing a macroprudential policy institution is to foster cooperation among various agencies in risk assessment and diagnosis and to make timely action possible. In this regard, the establishment of a formal arrangement by law will enhance cooperation, by addressing overlaps and gaps in risk identification and analysis and create a mutual understanding on which agency should tackle a problem that might otherwise fall between the cracks.
21. Various types of models can be considered, and each of which has potential advantages and disadvantages. Among the five models presented by Berntsson and Molin (2012), Model A, with the establishment of a macroprudential policy council, appears to have most advantages. However, for this model to work effectively, the following factors should be taken into account.
A macroprudential policy council should be autonomous with a clear financial stability mandate. To assure autonomy, the council should be chaired by Riksbank governor. At the same time, to ensure democratic control, the membership of the council can be evenly split among representatives from the Riksbank, FI, and external experts.
The governance structure of the council, including meeting and voting rules, should be clearly stipulated. For example, voting rules can be carefully designed to reduce the risks of no action, as a result of persistent disagreement among constituent agencies.
The council should be given a clear set of macroprudential policy tools, separate from micro prudential policy tools, avoiding overlapping powers to the extent possible.
A strong accountability and transparency framework should be in place. This should be even stronger if the MOF is excluded from the council.
Two complementary measures are important. First, the existing crisis management group (chaired by the MOF) should be formally institutionalized, with meeting schedule regularized. In addition, its mandate should be reconstituted to focus on contingency planning and crisis management. Second, establishing a regional macroprudential policy framework (on which the authorities have just initiated) will be equally important, given the lessons learnt from the Baltic crisis.
References
Berntsson, C. N., and J. Molin, 2012, “A Swedish framework for Macroprudential Policy,” Sveriges Riksbank Economic Review 2012:1, February.
European Systemic Risk Board, 2012, “The Macro-prudential Mandate of National Authorities,” March.
Finansinspektionen, 2005, Stability of the Swedish Financial Sector Report 2005:10, 2005:10–18.
Finansinspektionen, 2007, Stability of the Swedish Financial Sector Report 2007:16, 2007- 10–15.
FSB, IMF, and BIS, 2011, “Macroprudential Tools and Frameworks,” Update to G-20 Financial Ministers and Central Bank Governors, February.
International Monetary Fund, 2011, Macroprudential Policy: An Organizing Framework.
Ishi, K., 2010, “Regulating Financial Sector Outward and Inward Spillovers,” Attachment II, Sweden: Staff Report for the 2010 Article IV Consultation, IMF Country Report, No. 10/220.
Nier, E. W., J. Osiński, L. Jácome, and P. Madrid, 2011, “Institutional Models for Macroprudential Policy,” IMF Staff Discussion Note, SDN/11/18, November.
Sveriges Riksbank, 2005, Financial Stability Report 2005:1, June.
Sveriges Riksbank, 2007, Financial Stability Report 2007:1, June.
Sveriges Riksbank, 2008, Financial Stability Report 2008:1, June.
Swedish National Audit Office, 2011, Managing Financial Stability in Sweden: Experiences from the Swedish Banks’ Expansion in the Baltics, RiR 2011:9.
Prepared by Kotaro Ishi (kishi@imf.org).
In this note, macroprudential policy is defined as the policy that uses primarily prudential tools to limit systemic or system-wide financial risks, in line with IMF (2011) and FSB, IMF, and BIS (2011).
Informal cooperation arrangements existed prior to the crisis. In June 2005, the Memorandum of Understanding was signed among the MOF, the Riksbank, and FI. Later, in May 2009, this Memorandum of Understanding was expanded to include the NDO.
2011 FSAP Update recommended that the authorities establish a high level systemic financial stability council to focus solely on financial stability and related macroprudential policies.