Abstract
Panama’s extensive trade and financial linkages make it vulnerable to adverse external shocks, and this would have a sizable impact on Panama’s real activity. In the absence of monetary policy, macroprudential policy tools could usefully complement microprudential tools. A macroprudential supervisory body must possess the ability or power to collect and analyze firm-, market-, and global-level data to detect risks before they develop into full-blown crises. This study analyzes Panama’s tax structure, performance, and administration in order to identify priority areas for further strengthening
Macroprudential Policy in Panama? Lessons from Cross-Country Experience1
A. Introduction
1. The Panamanian financial system is sound and has proven resilient to shocks. Domestic credit and output growth rebounded quickly following the global financial crisis, and financial sector indicators suggest that the banking sector—which dominates the financial system—remains well capitalized, liquid, and profitable. The 2011 FSAP found that banking sector supervision was broadly adequate, though it also noted that (i) the Superintendency of Banks of Panama (SBP) did not have a framework to monitor macroeconomic developments and their impact on the banking system; and (ii) the regulation of nonbanks suffered from important weaknesses as regards the legal framework and capacity of the supervisory agencies.
2. An important FSAP recommendation was to enhance offsite supervision to develop a view on macro-prudential and systemic risk trends.2 While the creation of a council of supervisors (Financial Coordination Committee, CCF) was welcome, the FSAP also indicated that the council needed an enabling legal framework to oversee the financial system effectively.3 Since then, the authorities have made significant progress in addressing the FSAP recommendations, including through initiating a quarterly financial stability report, and upgrading the legal framework and enhancing the technical and analytical capacity of nonbank supervisors, but much remains to be done to oversee the financial system as a whole.
3. This paper discusses how macroprudential policy could supplement the existing microprudential policy framework in Panama. This is done through (i) analyzing the strengths and weaknesses of existing institutional arrangements for macroprudential policy in a number of countries with a view to identifying best practices that could be applied in Panama; and (ii) discussing the advantages and disadvantages of different macroprudential policy instruments, based on the experience of other countries, and their potential usefulness in Panama.
B. Background
Financial Sector Structure, Performance, and Oversight
4. Panama is an important regional financial center. With total assets representing more than three times GDP, the financial system is the largest in Latin America, and is dominated by banks, which represent almost 80 percent of total system’s assets. Securities’ firms account for about 17 percent of total system’s assets (Table 1). The Panamanian banking sector includes a sizable offshore sector (some 20 percent of total bank assets), which is largely isolated from the rest of the financial system.4 While onshore banks pursue a traditional model of lending financed with deposit taking, foreign funding has been increasing: from 2005 to 2011, the share of foreign deposits in total retail deposits increased from about 20 percent to 30 percent. At more than 90 percent of GDP, bank credit to the private sector is high by Latin American standards (Figure 1).
Structure of the Financial System
(June 2012)
Structure of the Financial System
(June 2012)
| Entities | Number of entities | Assets in percent of total financial system assets | Assets in percent of GDP | Supervisory authority | |
|---|---|---|---|---|---|
| Cooperatives | 577 | 1.6 | 4.8 | Panamanian Autonomous Institute of Cooperatives (IPACOOP) | |
| Savings and credit associations | 4 | … | … | National Mortgage Bank (BHN) | |
| Insurance companies | 31 | 1.3 | 4.0 | Superintendency of Insurance | |
| Pension fund | 2 | 0.7 | 2.2 | Superintendency of Securities Markets | |
| Development banks | 2 | 0.5 | 1.5 | Agricultural Development Bank, and National Mortgage Bank (BHN) | |
| Financial companies | 161 | 0.8 | 2.3 | Ministry of Commerce and Industry | |
| Leasing companies | 118 | 0.3 | 0.9 | Ministry of Commerce and Industry | |
| Securities firms | 74 | 16.9 | 50.9 | Superintendency of Securities Markets | |
| Remittance companies | 15 | … | … | Ministry of Commerce and Industry | |
| Pawn shops | 280 | … | … | Ministry of Commerce and Industry | |
| Banks | 90 | 77.9 | 234.8 | Superintendency of Banks | |
| General license | 48 | 63.6 | 191.7 | Superintendency of Banks | |
| International license | 28 | 14.3 | 43.2 | Superintendency of Banks | |
| Representative license | 14 | … | … | Superintendency of Banks | |
| Total | 1,354 | 100.0 | 301.6 | ||
Structure of the Financial System
(June 2012)
| Entities | Number of entities | Assets in percent of total financial system assets | Assets in percent of GDP | Supervisory authority | |
|---|---|---|---|---|---|
| Cooperatives | 577 | 1.6 | 4.8 | Panamanian Autonomous Institute of Cooperatives (IPACOOP) | |
| Savings and credit associations | 4 | … | … | National Mortgage Bank (BHN) | |
| Insurance companies | 31 | 1.3 | 4.0 | Superintendency of Insurance | |
| Pension fund | 2 | 0.7 | 2.2 | Superintendency of Securities Markets | |
| Development banks | 2 | 0.5 | 1.5 | Agricultural Development Bank, and National Mortgage Bank (BHN) | |
| Financial companies | 161 | 0.8 | 2.3 | Ministry of Commerce and Industry | |
| Leasing companies | 118 | 0.3 | 0.9 | Ministry of Commerce and Industry | |
| Securities firms | 74 | 16.9 | 50.9 | Superintendency of Securities Markets | |
| Remittance companies | 15 | … | … | Ministry of Commerce and Industry | |
| Pawn shops | 280 | … | … | Ministry of Commerce and Industry | |
| Banks | 90 | 77.9 | 234.8 | Superintendency of Banks | |
| General license | 48 | 63.6 | 191.7 | Superintendency of Banks | |
| International license | 28 | 14.3 | 43.2 | Superintendency of Banks | |
| Representative license | 14 | … | … | Superintendency of Banks | |
| Total | 1,354 | 100.0 | 301.6 | ||
Latin America: Private Sector Credit
(percent of GDP, 2011)
Citation: IMF Staff Country Reports 2013, 089; 10.5089/9781484344477.002.A002
Sources: Haver, IFS, WEO, and IMF staff estimates.Latin America: Private Sector Credit
(percent of GDP, 2011)
Citation: IMF Staff Country Reports 2013, 089; 10.5089/9781484344477.002.A002
Sources: Haver, IFS, WEO, and IMF staff estimates.Latin America: Private Sector Credit
(percent of GDP, 2011)
Citation: IMF Staff Country Reports 2013, 089; 10.5089/9781484344477.002.A002
Sources: Haver, IFS, WEO, and IMF staff estimates.5. Banks are well capitalized, profitable, and liquid (see Table 5 of the accompanying staff report). The stress tests conducted during the 2011 FSAP suggest that the system could withstand a wide range of shocks, including a repeat of the Lehman episode. However, the analysis was constrained by data gaps: for example, the SBP does not collect regularly data on loan write-offs, construction sector, property prices, and loan-to-value ratios. Given that the interbank market is not operating smoothly and that important safety net elements such as a lender of last resort and deposit insurance are missing, banks hold significant liquidity buffers. On the other hand, some banks have a high degree of concentration in their interbank liquidity holdings, and some small banks appear vulnerable to liquidity shocks. Bank failures have been rare, of limited size and complexity, and have been effectively managed, generally with no losses to depositors or creditors.
6. Data deficiencies make it difficult to assess the financial situation of some nonbank financial institutions, most of which (considered individually) are not systemically important. For example, cooperatives—some highly leveraged—do not produce financial statements according to IFRS. Savings and loans institutions produce financial statements according to IFRS, but are not subject prudential requirements to classify and provision assets.
7. Oversight responsibilities are fragmented, though the recently-established Financial Coordination Committee (CCF) has already contributed to enhance coordination among the supervisory agencies. Financial system supervision is split among eight different entities, half of which are autonomous and the other half are part of the government. The CCF was created in 2011 in order to improve coordination among all these entities and harmonize financial sector regulation, but the lack of a clear financial stability mandate and differences in capacity of member institutions may diminish its effectiveness (Table 1 and footnote 3).
Stability Challenges and the Role for Macroprudential Policy
8. The main financial stability challenges relate to the openness of the financial sector in a highly uncertain macroeconomic global environment, strong credit growth against the backdrop of very low interest rates, the highly dynamic real estate market, and increasing cross-sectional linkages. Chapter 1 of this Selected Issues Paper shows that the impact of external shocks on Panama’s output is amplified through their effect on domestic credit. Indeed, during the 2008-09 crises, foreign banks cut their credit lines to Panamanian banks, which significantly affected credit and real GDP growth in Panama. There are no signs of bubbles in the real estate market, though anecdotal evidence suggests that there may be oversupply in certain segments such as hotels and the high-end real estate market. In recent years, although overall credit growth has remained in line with economic activity, mortgage loans seem to have grown faster than wages, and credit to some other subsectors (e.g. tourism) seems to have grown more than these sectors’ revenue-generating capacity.
9. Financial stability challenges are mitigated by banks’ conservative lending practices. Having always operated in an open system without a lender of last resort, banks favor conservative business models, dominated by traditional deposit-taking (rather than relying on wholesale funding) and lending operations. Furthermore, banks maintain ample liquidity and capital buffers, apply prudent self-imposed loan-to-value ratios, and have so far largely avoided more sophisticated instruments such as loan securitization.
10. Nonetheless, a macro-prudential function could be useful in a fully dollarized economy like Panama, as a more efficient way to address systemic risks. Although self-discipline has served Panama well and the system has proved resilient to even large external shocks, individual institutions do not have an overview of the system as a whole. Self-imposed rules can easily be broken, particularly in times of increased competition, and individual decisions may not always be optimal for systemic stability. A stronger mandate for financial stability and the ability to implement macroprudential policies, be it as a start the formulation of guidelines for the financial system or specific subsectors, could be useful to prevent future problems, particularly in the absence of a monetary policy.
C. Existing Institutional Arrangements Around the World–Advantages and Disadvantages
11. The key institutional elements of a macroprudential policy framework include the mandate, powers, instruments, and coordination with microprudential and macroprudential policies. For example, a formal mandate can improve the clarity of decision making and avoid policy paralysis when views of stakeholders differ. A mandate normally comes with the power to collect information and adopt measures. Establishing accountability in conducting macroprudential policy is important given that there is no straightforward metric of success (Appendix I).5
12. Previous IMF studies identify three broad categories and seven stylized models of macroprudential policy. The three broad categories are differentiated mainly based on how objectives and functions of macroprudential, monetary, and microprudential policies are coordinated and how much information is available within the central bank: full integration means that all financial supervisory and regulatory functions are carried out by the central bank or by its subsidiaries; partial integration means that the securities supervisor or business conduct supervisor are separate entities, while prudential supervision of banks (and other institutions) is conducted by the central bank; and separation means that essentially all financial regulatory functions (other than payments oversight) are housed outside of the central bank (Nier and others, 2011 and Appendix II).
13. Institutional designs of macroprudential policy vary across countries and regions. The 2010 IMF survey suggests that, compared to other regions, the Western Hemisphere has the lowest share of financial stability and macroprudential policy mandates given to the central bank. This is because it is seen to conflict with the independence and mandate of the central bank as sanctioned in the constitution in some countries (Appendix III). In most of the countries without a formal macroprudential policy mandate, there is a formal financial stability mandate, usually shared among agencies (e.g. Canada, Chile, and Colombia).
14. The models in the full and partial integration categories, where the central bank either alone or together with other agencies is in charge of macroprudential policy, are less relevant for Panama. The central bank becomes the owner of macroprudential policy when it is given the objective to safeguard financial stability (as in the Czech Republic and Singapore). Partial integration or twin peaks models involve close institutional integration between the functions of the central bank and the prudential supervisor, while the regulation of activities or “conduct” in retail and wholesale financial markets is conducted by another agency (e.g. the setups in Brazil, the Netherlands, the U.K., and the U.S.). The main advantages of the full or partial integration models relate to better flow of information and improved coordination across objectives and functions within one organization, which can increase effectiveness of decision-making. The main disadvantage relate to the lack of institutional mechanisms to challenge the “house views” formed within one institution.
15. The models falling under the separation category (models 5-7 in Appendix Table 1) are more relevant for Panama. The strengths of such a multiagency setup include (i) reduced risks for any one institution becoming unchallenged in its identification of risks or assessment of the appropriate policy response and (ii) keeping each agency focused on their main objective, which in itself may contribute to maintaining financial stability. Under this arrangement, policy making benefits from different perspectives on the sources of systemic risk, the potential for regulatory arbitrage, and appropriateness of measures (which may be housed in different agencies). The existing models in Canada, Chile, Mexico, Peru, as well as Australia, Hong Kong SAR, and Korea are examples of such stylized models.
16. This setup also faces a number of challenges in ensuring effectiveness of macroprudential policy. In particular, a collective responsibility for systemic risk mitigation can dilute accountability and incentives and may create a situation where no one institution has all the information needed to analyze all interlinked aspects of systemic risk (due to barriers to free flow of information, caused by rivalry or legal obstacles). This may increase the chances of risks remaining unaddressed and delays in taking remedial measures.
17. A key mechanism to address some of these weaknesses is the establishment of a coordinating committee. It can facilitate the exchange of information between agencies and foster the engagement on the part of each agency with the shared goal of financial stability. Formal arrangements, which are more visible to the public, can enhance these benefits. Specifically, more formal arrangements may allow the committee to issue public warnings and recommendations to constituent agencies (as in Mexico). This can foster effective use of macroprudential policy instruments even where such recommendations are not binding on the agency. However, a committee may not be able to fully address deep-rooted accountability and incentive problems that remain a concern for the effectiveness of this group of models.
18. An important risk is also that decisions may be subject to delay. This risk is greater where the committee’s membership is large or where the treasury occupies a strong role. Careful design of voting arrangements can reduce the risk that no action is taken as a result of persistent disagreement between constituent agencies or political economy pressures. Such voting should be subject to a simple majority or a qualified majority rule rather than requiring unanimity among all constituent agencies (Nier and others, 2011).
D. Advantages and Disadvantages of Individual Policy Instruments
19. This section describes briefly the most frequently used macroprudential measures and discusses their relevance for Panama. The measures considered here include loan-to-value ratios (LTV), debt (service)-to-income ratios (DTI), and dynamic provisioning (DP). Advantages and disadvantages of other measures are described in Appendix IV.6 These are for illustration purposes only, and actual assignment and specification of instruments has to take into account local considerations, such as legal and constitutional constraints, effectiveness of instruments to meet objectives, and the level of development, structure, and complexity of the financial system.7
LTV
20. LTV limits enhance banks’ resilience to credit risks by increasing the collateral backing of loans and thus restricting losses in the event of default. Generally, the ratio is set based on the historical volatility of the collateral value. It directly limits risky lending, slowing down the supply of credit to specific sectors (e.g. real estate, car lending, etc).8
21. Limits on LTV ratios have been increasingly applied to reduce systemic risk arising from boom-bust episodes, notably in real estate markets. By limiting the loan amount to well below the current value of the property, LTV limits can help rein in house price increases by putting the brakes on household leverage, reducing the financial accelerator effect. For example, Wong and others (2011) find that, for a given fall in prices, the incidence of mortgage default and bank losses bank losses are higher for countries without an LTV measure. Furthermore, this measure is less prone to international leakage as it is also applied to branches of foreign banks.
22. The ratio can be (and often is) applied countercyclically. Tightening the ratio during a boom restricts the accumulation of risks, thereby moderating the credit boom and house price increases. The caps would generally apply to new loans rather than the stock of existing loans to avoid the situation where borrowers would have to provide more collateral after a large fall in the price of collateral. Some countries have kept LTV rates constant to provide a minimum buffer against an unsustainable increase in house prices (Colombia, Lebanon, Malaysia, and Sweden). In other countries, LTV limits are adjusted in line with the cyclical position, with a tightening occurring during housing booms and a relaxation during downturns (China, Hong Kong SAR, and Korea). In some cases, the adjustments are made in a reactive, and not necessarily countercyclical, manner (Lim and others, 2011).
23. Like other measures, LTV limits have also a number of disadvantages. First, implementing this measure has costs associated with potential credit rationing. For example, new entrants to the housing and real estate market could be rationed out. In some countries (e.g. Hong Kong SAR), this problem is addressed with insurance programs for first-time home buyers. Accordingly, it is difficult to calibrate the trade-off between financial stability benefits, economic activity and societal preferences for home ownership.9 Second, the measure is susceptible to circumvention and could encourage obtaining second mortgages on the same property or unsecured loans such as credit card borrowing. Importantly, it has less impact on leverage of borrowers and banks.
24. Globally, this is the most frequently used tool. According to the 2010 IMF survey, 34 out of the 52 responding countries had this measure in place.10 LTV limits are particularly popular in Asian countries: 9 out of 12 surveyed countries had LTV limits. In the Western Hemisphere, LTV limits were applied in Brazil,11 Chile, Colombia, Mexico, and Paraguay as well Canada 12 and the U.S.
25. In Panama, there are no formal LTV requirements. Individual banks apply LTV ratios for their creditworthiness assessments based on the value of the underlying (financed) property. For example, for properties worth US$45,000, 95 percent is considered the maximum; for properties worth more than US$45,000 but less than US$120,000, 90 percent; and for more expensive property, 80 percent LTV ratio is considered a norm. Usually, these loans are provided against life insurance, which does not cover unemployment. However, increased competition may force banks to apply less stringent requirements. The authorities could usefully consider formal requirements, including lower ratios in zones where housing prices increase much faster than the national average. To be able to apply such a differentiated ratio, the authorities would need to collect and analyze information on housing prices as well as (actual) LTVs applied by banks.
DTI
26. When used alone, limits on DTI aim at safeguarding banks’ asset quality. They limit risky lending and reduce the probability of default. When used in conjunction with the LTV, the DTI can help further dampen the cyclicality of collateralized lending by adding another constraint on households’ capacity to borrow. As with LTV limits, adjustments in the DTI ceilings can be made in a counter-cyclical manner to address the time dimension of systemic risk (Lim and others, 2011).
27. Like in the case of the LTV, DTI may involve costs associated with potential credit rationing. Moreover, data requirements can be challenging and calibration is difficult. It is susceptible to circumvention.
28. In Panama, there are no formal DTI requirements, and banks apply fairly high DTI ratios, sometimes above fifty percent (when the debt burdens for mortgages and other consumer loans are combined), for their creditworthiness assessments. Although this is less of a problem in a booming economy, characterized by historically low interest rates and unemployment, the situation may change in downswings with higher unemployment rates, especially if it is accompanied with much higher interest rates.
DP
29. Dynamic provisioning (DP) is designed to distribute loan losses evenly over the credit cycle. It is based on the notion that provisions should account for expected loss over the long term (cycle) rather than incurred loss. Generally, the level of provisioning on this basis would be less subject to sharp swings stemming from the strength of economic activity because of the primacy of expected, rather than actual, losses in a dynamic provisioning approach. By requiring banks to build reserve buffers during an upswing, DP counterbalances the tendency of specific loan reserves to be low when credit quality is high. As a result, the marginal cost of loan-loss provisioning is smoothed significantly over the credit cycle. Overall, DP is more effective when applied to narrowly-defined categories at the beginning of the cycle. DP would be less effective if a bank incurs large losses in an upswing, reducing the available cushion in the form of accumulated reserves.
30. While DP has a number of beneficial properties, there are also limitations to what it can achieve. For example, it can help absorb reasonably large shocks to loan quality, reducing a bank’s probability of default, but it is not designed to cover large unexpected loan losses (for which, there is bank capital) or tail risks. While it contributes to smoothening credit cycle, it is not designed to rein in rapid credit growth. The overall impact on credit growth is muted as lending can be shifted to foreign (parent) banks and less-regulated intermediaries. For example, in Spain, the buffer of dynamic provisions was large enough to offset about half of the loan losses occurred during 2008-09 but not all delinquencies, since eventual loan losses exceeded expected losses. By contrast, the reserves coverage in Uruguay ballooned as the expected loan delinquencies on which the model was calibrated did not materialize (Lim and others, 2011).
31. Data requirements and calibration can become challenges. Some calibration does not take into account the credit risk profile of banks. Those involving probability of default estimations require granular data, which are missing in many countries. Moreover, data should cover a full credit cycle; data covering only the boom period would lead to underestimation of risks. There are strong overlaps with countercyclical capital buffers and variable risk weights tools.
32. DP is widely applied in Latin America. Following the introduction of the Spanish system in 2000, Uruguay (2001), Bolivia (2008), Peru (2008), Colombia (2009), Chile (2011), and Mexico (2011) implemented countercyclical provisioning tools.
33. The design of the DP systems varies significantly across countries.
In Spain and Uruguay, the system requires banks to (continuously) build up provisions against the average flow of provisions through the credit cycle.
In Colombia and Peru, the system does not require continuous provisioning, but rather includes an activation mechanism that triggers the accumulation of dynamic provisions during an economic upswing and the drawdown of these provisions during a downturn. Under the Peruvian system, which is based on GDP growth performance, the activation or deactivation of the mechanism is common to the whole system. In Colombia, given the system’s discretionary nature, there is no explicit variable used so far, although the authorities have announced that credit will be taken into account.
In Chile and Mexico, provisioning rates are set according to debtors’ classification or risk profile in terms of expected loss (Chilean banks are allowed to build additional countercyclical provisions to cover “unexpected losses;” Wezel and others, 2012).
34. In Panama, the authorities have initiated a project to adopt a DP system. The project is supported by IMF technical assistance.
E. Implications and Conclusions
35. In the absence of monetary policy, macroprudential policy tools could usefully complement microprudential and other tools in Panama. To potentially benefit from macroprudential policy tools, the authorities should define an agency responsible for the stability of the financial system as a whole, and start monitoring and analyzing systemic risks. This would involve providing adequate mandate and powers to the agency, collecting necessary data and building capacity to monitor systemic risk. On the former, the CCF could become such an agency.
The CCF should be given a clear a financial stability mandate. While institutional arrangements are largely shaped by country-specific circumstances and there is no “one size fits all,” fragmented institutional structures can create frictions in risk identification and mitigation. The CCF has already contributed to improving coordination among the supervisory agencies, but does not have a mandate for financial stability. A clear mandate would strengthen accountability and incentives to act, and reduce (potential) risks of delayed action due to political pressures or lobbying in the presence of multiple agencies.
A macroprudential supervisory body must possess the ability or power to collect and analyze firm-, market-, and global-level data to detect risks before they develop into full-blown crises. For effective risk identification, it is important that all relevant data are available to all agencies, or at least to the agency that is in the lead in risk identification. In this context, the SBP’s chairing the CCF is welcome, but this function could be further expanded though building up the SBP’s capacity to monitor and analyze systemic risks, which may involve additional costs (human resources, software, etc.). Priority should be given to collecting and analyzing data on real estate prices, loan write-offs, LTVs, leverage indicators for households and corporates and building the capacity to analyze macro-financial linkages. The SBP’s recent initiative to produce financial stability reports could also be extended to cover nonbank sectors, with inputs from other supervisors.
The CCF should be given the power to adopt or recommend macroprudential measures as needed. Importantly, the CCF should be able to influence and be responsive to microprudential policies.
Any specific macroprudential measures that the authorities might adopt would depend on the types and expected impact of systemic risks they face. While banks apply (self imposed) LTV and DTI ratios, increased competition may force them to loosen lending standards. Thus, the authorities could consider adopting formal LTV and DTI requirements or at least recommend a range, taking into account leverage of the household sector. In adopting any measures, the authorities should weigh the benefits of the measures against their costs. The authorities should also continue their efforts to implement DP.
Appendix I. Some Relevant Elements of the Institutional Design of Macroprudential Policy1
Information and resources. To gauge accumulating systemic risks, it is essential that policy makers have access to information and data on the components of the financial system, including data on individual financial institutions, their exposures to other institutions, and developments in payments and settlements systems. When several bodies are involved, the arrangements for sharing information become complex, as some information are confidential and market sensitive. It is also important that adequate resources are available to process received information and develop measures or provide recommendations.
Mandate and powers. Advantages of developing a formal macroprudential mandate include establishing clear objectives, responsibilities, and powers for the agency (agencies) involved in macroprudential policy. The 2010 IMF macroprudential survey found that less than half of the respondents had a formal macroprudential mandate in place, beyond financial stability. A larger proportion of emerging markets economies (50 percent) than advanced economies (35 percent) has such a mandate, which may be related to the fact that emerging markets had more frequent financial crisis in the past compared to advanced economies. Of those without a formal mandate, about half have plans or are contemplating to adopt a mandate.
Powers to communicate risk warnings and to recommend regulatory instruments and actions are essential parts of policy making. Examples include the ability to issue non-binding recommendations to other authorities. The recommendations are often subject to a “comply or explain” mechanism (e.g., in EU, UK and US), sometimes strengthened by an ability to publish recommendations.
Accountability. An institutional design challenge is to establish accountability when the “costs” of macroprudential measures in the form restrictions on certain activities are felt immediately while “benefits” of lowering incidence of financial distress accrue over a long term and are hard to measure. This challenge is often compounded by the presence of multiple agencies in macroprudential policymaking that may differ in their primary objectives. This challenge highlights the importance of insulating the authorities in charge of macroprudential policy from pressures linked to the political cycle.
Transparency and clear communication of policy decisions to the public are central elements of accountability. This can include ex ante statements of strategy, publication of records of meetings, Financial Stability Reports and annual performance statements with an ex post assessment of policy effectiveness.
Appendix II. Some Key Distinguishing Dimensions of Real Life Macroprudential Policy Models (based on Nier and others, 2011)
Degree of institutional integration of central bank and financial regulatory functions. Institutional integration affects coordination across objectives and functions of macroprudential, monetary, and microprudential policies and how much information is available within the central bank. The degree of integration can be full, partial, or separation.
Ownership of macroprudential policy. Ownership of the macroprudential mandate can rest with the central bank or a committee related to the central bank or an independent committee or be shared by multiple agencies. If the mandate is given to multiple agencies, each agency is expected to take responsibility for mitigation of systemic risk arising in its domain.
Role of the treasury. The formal role of the treasury can be (i) active, if it plays a leading role in policymaking or coordinating committees; (ii) passive, if the treasury participates in such committees, but has no special role; or (iii) simply nonexistent.
Existence of a separate body coordinating across policies to address systemic risk. A separate coordinating committee is a common feature when the policy mandate is shared by multiple agencies.
Stylized Models for Macroprudential Policy 1
Stars are explained in the table.
Stylized Models for Macroprudential Policy 1
| Features of the model/Model | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 |
|---|---|---|---|---|---|---|---|
| 1. Degree of institutional integration of central bank and supervisory agencies | Full(at a central bank) | Partial | Partial | Partial | No | No (Partial*) | No |
| 2. Ownership of macroprudential policy mandate | Central bank | Committee “related” to central bank | Independent committee | Central bank | Multiple agencies | Multiple agencies | Multiple agencies |
| 3. Role of MOF/treasury/government | No (Active*) | Passive | Active | No | Passive | Active | No (Active*) |
| 4. Separation of policy decisions and control over instruments | No | In some areas | Yes | In some areas | No | No | No |
| 5. Existence of separate body coordinating across policies | No | No | No (Yes*) | No | Yes | Yes (de facto**) | No |
| Examples of specific model countries/regions | Czech Republic Ireland (new) Singapore* | Malaysia Romania Thailand UK (new) | Brazil*France (new) United States (new) | Belgium (new) Netherlands Serbia | Australia | Canada Chile HK SAR*Korea**LebanonMexico | Iceland Peru Switzerland |
Stars are explained in the table.
Stylized Models for Macroprudential Policy 1
| Features of the model/Model | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 |
|---|---|---|---|---|---|---|---|
| 1. Degree of institutional integration of central bank and supervisory agencies | Full(at a central bank) | Partial | Partial | Partial | No | No (Partial*) | No |
| 2. Ownership of macroprudential policy mandate | Central bank | Committee “related” to central bank | Independent committee | Central bank | Multiple agencies | Multiple agencies | Multiple agencies |
| 3. Role of MOF/treasury/government | No (Active*) | Passive | Active | No | Passive | Active | No (Active*) |
| 4. Separation of policy decisions and control over instruments | No | In some areas | Yes | In some areas | No | No | No |
| 5. Existence of separate body coordinating across policies | No | No | No (Yes*) | No | Yes | Yes (de facto**) | No |
| Examples of specific model countries/regions | Czech Republic Ireland (new) Singapore* | Malaysia Romania Thailand UK (new) | Brazil*France (new) United States (new) | Belgium (new) Netherlands Serbia | Australia | Canada Chile HK SAR*Korea**LebanonMexico | Iceland Peru Switzerland |
Stars are explained in the table.
Appendix III. A Comparison of Regional Arrangements
National institutional designs of macroeconomic policy frameworks are largely influenced by legal traditions, pre-existing coordination arrangements, and objectives. In some countries in Latin America, for example, the central bank does not participate in the committee in charge of macroprudential policy because this is seen to conflict with the independence and mandate of the central bank as sanctioned in the constitution. The authorities in many countries try to capitalize on existing institutions and governance structures if they are working well.
In a number of advanced economies, in particular in Europe, macroprudential functions have been integrated into the central bank. Generally, these countries have adopted some form of “twin peaks” model, leaving conduct-of-business and securities market supervision as a responsibility of a separate agency (Belgium, France, the United Kingdom, and the United States).
A number of countries created dedicated policy-making committees. But the roles of the central bank and treasury in these committees differ across countries. For example, the Financial Policy Committee (FPC) in U.K. is chaired by the Governor of the Bank of England, while the Financial Stability Oversight Council (FSOC) in the U.S. is chaired by the U.S. Treasury.
In Latin America, the institutional arrangements for financial stability have broadly shaped the institutional arrangement for macroprudential policy framework. Here, countries can be classified in two distinct groups: the “Atlantic” model (includes Argentina, Brazil, and Uruguay), and the “Pacific” model (includes Chile, Colombia, and Peru, as well as Costa Rica and Mexico).1
In the Atlantic model, the central bank is implicitly in charge of macroprudential policies, although the precise institutional setup varies across countries. In Brazil, the National Monetary Council (CMN) is vested with broad powers, including potential decisions of macroprudential policy nature (based on recommendations from the central bank). In Uruguay, all financial regulation and supervision is fully integrated at the central bank. In Brazil, the government has the majority of members and chairs the CMN. In Argentina, the government has no representation on the central bank board, but, in practice, exercised influence over the central board, which led a high turnover of central bank governors. In all three countries, the government plays an active role in macroprudential policy.
In the Pacific model, both the central bank and the financial supervision agency take regulatory decisions that fall in the domain of macroprudential policy, creating challenges in ensuring appropriate accountability. In all the countries in this model, the governor of the central bank cannot be held accountable for financial stability because this responsibility is beyond the scope of their mandate. The role of the government varies across countries: in Chile and Mexico, the government plays a key role in macroprudential policy since the MoF chairs the committees. In Colombia, the government also plays an important role as the MoF is in charge of financial sector regulation, and the Financial Superintendence legally reports to the MoF. In Peru, the government plays no role on financial stability.
The new macroprudential committees in Latin America (Chile, 2011; Mexico, 2010; and Uruguay, 2011) are vested with powers to obtain information from all financial institutions and to play a coordinating role to secure the consistency of financial stability efforts. They have a mandate to prevent the buildup of systemic risks and, if necessary, recommend the implementation of macroprudential policies to the relevant agencies. In all three countries, the committee is presided by the MoF/Treasury,2 perhaps because crisis management is among its goals, and reflecting the fact these countries have had financial/banking crises that involved resolving insolvent institutions with public money. In particular, the financial stability committees in Mexico and Uruguay have explicit powers to manage financial crises. In Chile, the crisis management powers reside with the individual agencies and the Council operates as a coordinating device.
In European emerging market economies, the institutional setup varies across countries. In Hungary, the macroprudential policy committee comprises the central bank, treasury department, financial supervisory authority, and the chairmanship rotates. In the Czech Republic, on the contrary, the macroprudential policy framework is centered on the central bank (CNB), whose mandate includes both price and financial stability. There is no yet formal macroprudential policy mandate in Croatia and Poland, but the central bank in these countries has frequently used their financial stability mandate to take measures of macroprudential and capital flow management nature.
In Asia, institutional designs of macroprudential policy widely vary across countries. While Australia has a separation model (and does not have a formal macroprudential policy mandate),3 Singapore has a full integration model. Malaysia established (in 2009) a financial stability committee within the central bank structure, chaired by the central bank Governor; Thailand established a similar model in 2008. Hong Kong SAR and Korea have separation models, where policies are coordinated formally and informally.
Compared to other regions, Asia has the lowest share of macroprudential mandates being fixed in legislation and the highest ratio of the mandate being shared among multiple agencies. In all countries but Mongolia, the mandate, if there is one, is shared among several agencies. In all countries with a mandate, the central bank is given the mandate (means that the analytical capacity of the central bank is fully utilized.
Appendix IV. Selected Systemic Risk Manifestations and MaPP Tools in Other Countries
(Based on review of the literature, and Osiński and Jafarov, forthcoming)
| MaPP tools | Country of use | Pros | Cons |
|---|---|---|---|
| Leverage ratio | Canada, USA | Guards against underestimation of asset risk. Less susceptible to arbitrage and mis-measuring. | A blunt instrument that can constrain economic activity. No penalty for risk may create perverse incentives to “risk-up.” |
| Time varying countercyclical capital surcharges | Increases costs of borrowing while building loss-absorbing capacity to cope for the bust. May help moderate credit cycles. Under Basel III, the “leakage problem” is mitigated by introducing mutual recognition of national countercyclical buffers. In particular, it is envisaged that the national buffer measure will apply to the local exposure of foreign banks. This reciprocity is mandatory only for buffers of up to 2.5 percent. The leakage problem could be more significant in small countries with large and open financial systems. | Limited success in curtailing the incidence and duration of credit booms. It is a crude tool if exuberance is localized in particular sectors, and may even encourage “risking up.” May divert attention from the liabilities side of banks’ balance sheets and is subject to “international leakage.” In case of Panama, the relatively large share of foreign bank branches in the domestic financial system assets (20 percent in 2010) requires a close attention to the leakage problem. The effectiveness of this measure is very sensitive to appropriate risk weighting of assets. | |
| Ceiling on general credit or credit growth. | Bulgaria, China, Colombia, Croatia, Greece, Nigeria, Portugal, Serbia, Slovakia, Romania | Direct impact on credit. Limits rapid expansion and leverage. Have had some success in slowing down the pace of bank credit. | Susceptible to circumvention. Can be offset by increases in credit from nonbanks, leading to build up of systemic risk in often less-regulated intermediaries, and foreign borrowing by some borrowers. |
| Riskweights/sector dependent risk weights | Austria, Argentina, Brazil, Bulgaria, Croatia, Czech republic, Estonia, France, India, Lebanon, Malaysia, Mongolia, New Zealand, Norway, Poland, Spain, Turkey | Targeted approach. May provide sharper incentives than countercyclical capital buffer. Adjusting risk weights on flow of lending relative to its stock could restrain lending in booms or encourage lending in downturns | May displace risk to other parts of the system — a “water bed” effect. Implementation challenges to ensure consistent application across balance sheet. Data needs greater than with aggregate tools. |
| Sectoral credit growth or level limits | China, Colombia, Malaysia, Philippines, Portugal, Singapore | Targets specific sectors, with limited impact on other sectors. | Muted impact on overall lending growth. |
| Time-varying liquidity buffers | Argentina, China, Croatia, India, Indonesia, Lebanon, Nigeria, Norway, Portugal, Serbia, South Africa, Switzerland, | Direct effect on banks’ liquid asset holdings and maturity mismatch, increasing resilience. Harder to arbitrage than capital-based measures. May also help to moderate the credit cycle (Croatia) | Limited international experience with liquidity requirements Microprudential standards still under development. |
| Corefundingratios | Belgium, Greece, India, Indonesia, Switzerland, Italy, Lebanon, Mongolia, Netherlands, Poland, Portugal, Spain, Sweden, Uruguay | Affects the quality and amount of liabilities. Limits the ability of financial institutions to rely on risky sources to fund growth in upswings and thus minimizes the impact of liquidity crises in bust periods. | |
| Limits on interbank exposures | Belgium, Chile, Colombia, India, Mexico, Nigeria, Paraguay, Peru, Portugal, Sweden, Switzerland, Uruguay. |
| MaPP tools | Country of use | Pros | Cons |
|---|---|---|---|
| Leverage ratio | Canada, USA | Guards against underestimation of asset risk. Less susceptible to arbitrage and mis-measuring. | A blunt instrument that can constrain economic activity. No penalty for risk may create perverse incentives to “risk-up.” |
| Time varying countercyclical capital surcharges | Increases costs of borrowing while building loss-absorbing capacity to cope for the bust. May help moderate credit cycles. Under Basel III, the “leakage problem” is mitigated by introducing mutual recognition of national countercyclical buffers. In particular, it is envisaged that the national buffer measure will apply to the local exposure of foreign banks. This reciprocity is mandatory only for buffers of up to 2.5 percent. The leakage problem could be more significant in small countries with large and open financial systems. | Limited success in curtailing the incidence and duration of credit booms. It is a crude tool if exuberance is localized in particular sectors, and may even encourage “risking up.” May divert attention from the liabilities side of banks’ balance sheets and is subject to “international leakage.” In case of Panama, the relatively large share of foreign bank branches in the domestic financial system assets (20 percent in 2010) requires a close attention to the leakage problem. The effectiveness of this measure is very sensitive to appropriate risk weighting of assets. | |
| Ceiling on general credit or credit growth. | Bulgaria, China, Colombia, Croatia, Greece, Nigeria, Portugal, Serbia, Slovakia, Romania | Direct impact on credit. Limits rapid expansion and leverage. Have had some success in slowing down the pace of bank credit. | Susceptible to circumvention. Can be offset by increases in credit from nonbanks, leading to build up of systemic risk in often less-regulated intermediaries, and foreign borrowing by some borrowers. |
| Riskweights/sector dependent risk weights | Austria, Argentina, Brazil, Bulgaria, Croatia, Czech republic, Estonia, France, India, Lebanon, Malaysia, Mongolia, New Zealand, Norway, Poland, Spain, Turkey | Targeted approach. May provide sharper incentives than countercyclical capital buffer. Adjusting risk weights on flow of lending relative to its stock could restrain lending in booms or encourage lending in downturns | May displace risk to other parts of the system — a “water bed” effect. Implementation challenges to ensure consistent application across balance sheet. Data needs greater than with aggregate tools. |
| Sectoral credit growth or level limits | China, Colombia, Malaysia, Philippines, Portugal, Singapore | Targets specific sectors, with limited impact on other sectors. | Muted impact on overall lending growth. |
| Time-varying liquidity buffers | Argentina, China, Croatia, India, Indonesia, Lebanon, Nigeria, Norway, Portugal, Serbia, South Africa, Switzerland, | Direct effect on banks’ liquid asset holdings and maturity mismatch, increasing resilience. Harder to arbitrage than capital-based measures. May also help to moderate the credit cycle (Croatia) | Limited international experience with liquidity requirements Microprudential standards still under development. |
| Corefundingratios | Belgium, Greece, India, Indonesia, Switzerland, Italy, Lebanon, Mongolia, Netherlands, Poland, Portugal, Spain, Sweden, Uruguay | Affects the quality and amount of liabilities. Limits the ability of financial institutions to rely on risky sources to fund growth in upswings and thus minimizes the impact of liquidity crises in bust periods. | |
| Limits on interbank exposures | Belgium, Chile, Colombia, India, Mexico, Nigeria, Paraguay, Peru, Portugal, Sweden, Switzerland, Uruguay. |
References
Bank of England, 2011, “Instruments of Macroprudential Policy,” A Discussion Paper.
Crowe, Christopher W., Deniz Igan, Giovanni Dell’Ariccia, and Pau Rabanal, 2011, “How to Deal With Real Estate Booms,” IMF SDN/11/02.
Dell’Ariccia, Giovanni, Deniz Igan, Luc Laeven, and Hui Tong, with Bas Bakker and Jerome Vandenbussche, 2012, “Policies for Macrofinancial Stability: How to Deal with Credit Booms,” IMF SDN/12/06.
FSB, IMF, and BIS, 2011, “Macroprudential Tools and Frameworks,” Update to G-20 Finance Ministers and Central Bank Governors (February).
Igan, Deniz, and Heedon Kang, 2011, “Do Loan-to-Value and Debt-to-Income Limits Work? Evidence from Korea,” IMF Working Paper No. 11/297 (Washington: International Monetary Fund).
Jácome, Luis I., Erlend W. Nier, and Patrick Imam, 2012, “Building Blocks for Effective Macroprudential Policies in Latin America: Institutional Considerations,” IMF Working Paper No. 12/183 (Washington: International Monetary Fund).
Lim, Cheng Hoon, Francesco Columba, Alejo Costa, P. Kongsamut, A. Otani, M. Saiyid, T. Wezel, and X. Wu, 2011, “Macroprudential Policy: What Instruments and How to Use Them? Lessons from Country Experiences,” IMF Working Paper No. 11/238 (Washington: International Monetary Fund).
Nier, Erlend W., Jacek Osiński, Luis I. Jácome, and Pamela Madrid, 2011, “Institutional Models for Macroprudential Policy,” IMF SDN/11/18.
Osiński, Jacek, Katharine Seal, and Lex Hoogduin, 2012, “Macroprudential and Microprudential Policies: Towards Cohabitation,” IMF SDN, forthcoming.
Wezel, Torsten, Jorge A. Chan-Lau, and Francesco Columba, 2012, “Dynamic Loan Loss Provisioning: Simulations on Effectiveness and Guide to Implementation,” IMF Working Paper No. 12/110 (Washington, DC: International Monetary Fund).
Wong, Eric, Tom Fong, Ka-fai Li, and Henry Choi, 2011, Loan-to-Value Ratio as a Macroprudential Tools—Hong Kong’s Experience and Cross-Country Evidence,” Hong Kong Monetary Authority Working Paper, 01/2011.
Prepared by Etibar Jafarov.
Systemic risk is defined here as the risk of disruptions in the provision of key financial services that can have serious consequences for the real economy. It is related to the interconnectedness of financial institutions and markets, common exposures to economic variables, and procyclical behaviors (IMF, FSB, BIS, 2011).
The CCF was created in 2011 to improve coordination among financial sector supervisors and harmonize regulation. Its members include the Superintendents of Banks (SBP), Insurance and Reinsurance (SSRP), the Capital Markets (SMV), as well as of head of the Panamanian Autonomous Institute of Cooperatives (IPACOOP), the Director of Financial Companies of the Ministry of Industry and Trade (MICI), and the Director of the Public Sector Workers’ Pension Funds (CICAP).
Although offshore banks can conduct interbank transactions with onshore banks, volumes of these transactions are reportedly small.
The “costs” of macroprudential measures in the of form restrictions on certain activities are felt immediately while “benefits” of lowering incidence of financial distress accrue over a long term and are difficult to measure.
Most macroprudential measures can be (and are) applied also for microprudential purposes. Both policies exist to correct market failures and externalities related to them. Generally, microprudential policy looks at individual institutions, while macroprudential at a financial system as a whole. In practice, overlaps are possible in the areas of perimeter, toolkit, and its transmission mechanism. Osiński and others, 2012 offer several approaches to deal with the problem of borderlines and potential tensions and conflicts.
These instruments are used to address the time dimension of systemic risks. See Appendix IV for instruments that are used for addressing the cross-sectional dimension of systemic risks.
Crowe and others (2011) find that tighter LTVs lead to lower house price increases, at least in the short run. Igan and Kang (2011) find similar results.
Many countries differentiate LTV limits on mortgage loans based on the purpose or value of the property (e.g. for commercial investors in Canada, Turkey, and Singapore or luxury or speculative investments in Hong Kong, Malaysia, and Singapore). Some Asian countries have adopted more granular features: Hong Kong relates the maximum LTV to the value of residential properties, while rates in Korea are based on whether or not a property is located in a speculative zone.
In addition, several countries such as Australia, Canada, Korea, Latvia, Thailand, and United Kingdom had granular capital requirements based on LTVs.
In Brazil, caps on LTV were abolished in December 2011.
In Chile, the maximum LTV ratio for covered bond-type mortgages raised from 75% to 100% for debtors with higher credit ratings in 2009. In Colombia, an LTV cap at 70 percent was introduced in 1999. In Canada, the authorities selectively tightened the LTV ceilings on cash-out refinancing transactions and investment property loans (in February 2010) and reduced the maximum amortization period for new government-backed insured mortgages with LTV ratios of more than 80% to 30 years from 35 years (in April 2011).
Based on FSB, IMF, BIS (2011).
Jácome and others, 2012.
The other members are the heads of the financial supervisory agencies and the central bank except in Chile, where the governor is invited to participate but is not formally a member of the Council.
In Australia, microprudential regulators take macroprudential considerations into account. The Council of Financial Regulators, chaired by the central bank, co-ordinates work of the country’s main financial regulatory agencies. The Council is non-statutory and has no regulatory functions separate from those of its members.
