A. Willingness to Act

4. International experience has shown that certain institutional aspects can foster strong willingness to act (IMF-Bank for International Settlements (BIS)-Financial Stability Board (FSB), 2016). In particular, clear assignment of the macroprudential mandate, well-defined objectives for institutions involved in safeguarding financial stability, a strong role for the central bank, dedicated financial stability units, and adequate accountability and communication frameworks can all foster willingness to act and thereby overcome the inaction bias.

5. These institutional features are in place in South Africa. The Financial Sector Regulation (FSR) Act 9 of 2017 defined key features of the institutional framework, which are then further elaborated on in a policy framework agreed between the Minister of Finance and the Governor. The FSR Act established a strong mandate for the SARB to protect and enhance financial stability and restoring financial stability in case a systemic event has occurred. Hence, the SARB plays a pivotal role in the institutional framework for macroprudential policy, with the Governor serving as the key decision-maker, backed by external and internal advisory committees. Other financial regulators also have a financial stability mandate (or at least a role to assist in maintaining financial stability), resulting in South Africa having more than one macroprudential authority, albeit SARB plays a leading role.

6. The combination of a designated decision maker and advisory committee structure with clear financial stability mandates is likely to foster willingness to act. Given the leading position of the Governor in the institutional arrangements, a key concern—albeit not one that has materialized—is that at some point an individual with a strong inaction bias (or overly aggressive action preferences) would be appointed. Hence, it is important that the FSR Act set up two external advisory committees, which have clear mandates and advise the Governor on macroprudential policy, complementing the longer-standing internal Financial Stability Committee (Figure 1, top panel):

• The Financial Stability Oversight Committee (FSOC) is a Statutory Committee under the FSR Act and advises the Governor of the SARB on matters relating to financial stability. FSOC’s primary objectives are to: (a) support the SARB when it performs its functions in relation to financial stability; and (b) facilitate cooperation and collaboration between the SARB and other financial sector regulators. It includes representatives from the SARB, NT and all financial regulators. The committee consists of the Governor as Chairperson, the Deputy Governor responsible for financial stability matters, the Chief Executive Officer (CEO) of the Prudential Authority (PA), the Commissioner of the Financial Sector Conduct Authority (FSCA), the CEO of the National Credit Regulator (NCR), the Director-General of NT, the Director of the Financial Intelligence Centre (FIC), and up to three additional officials of the SARB appointed by the Governor. Normally, it meets biannually.

• The Financial Sector Contingency Forum (FSCF) is a statutory forum established by the Governor of the SARB to assist the FSOC with identifying risks that could lead to a systemic event; and coordinating plans, mechanisms and structures to mitigate those risks. The FSCF was created to ensure broad participation and engagement of stakeholder groups in defining and coordinating approaches to crisis management. The forum does not play an active role in managing systemic events. Rather, it supports the development and testing of contingency plans and works as an established network for coordinating interventions and communicating effectively during a systemic event. The forum was established by the Governor. It consists of a Deputy Governor as the Chairperson (designated by the Governor), representatives from each financial sector regulator, representatives from other organs of state and representatives from financial sector associations.

• The Financial Stability Committee (FSC) is an internal SARB committee which was established in 2000 and restructured and elevated in 2010 in terms of its membership and responsibilities. The purpose of the FSC is to formulate financial stability policy on behalf of the SARB in support of its mandate. The FSC comprises the Governor as Chairperson—who cannot be outvoted in the committee—the Deputy Governors, all members of the Monetary Policy Committee (MPC) and a maximum of seven other SARB officials. The SARB fulfils its responsibility to monitor and review the strengths and weaknesses of the financial system and any risks to financial stability, and take the necessary steps to mitigate these risks, through the FSC. The FSC meets four times a year, or as required, to monitor vulnerabilities in the global and domestic environments, assess their possible implications for domestic financial stability and decide whether any mitigating measures need to be taken. The content and discussions at FSC meetings are shared with the FSOC while the issues discussed by the FSC and any decisions taken are communicated in the biannual Financial Stability Review publication. Records of FSC meetings are not published.

South Africa: Institutional Arrangements for Macroprudential Policy¹

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

¹ Models #1-#3 are labeled in accordance with the 2014 IMF SGN.Sources: IMF Macroprudential Policy Survey, IMF staff calculations and illustrations.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

South Africa: Institutional Arrangements for Macroprudential Policy¹

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

¹ Models #1-#3 are labeled in accordance with the 2014 IMF SGN.Sources: IMF Macroprudential Policy Survey, IMF staff calculations and illustrations.

• Download Figure
• Download figure as PowerPoint slide

South Africa: Institutional Arrangements for Macroprudential Policy¹

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

¹ Models #1-#3 are labeled in accordance with the 2014 IMF SGN.Sources: IMF Macroprudential Policy Survey, IMF staff calculations and illustrations.

• Download Figure
• Download figure as PowerPoint slide

7. This wide-ranging committee structure is what sets the South African institutional arrangements apart from many other emerging market economies (EMEs), but as noted above, when it is combined with a designated sole decision maker it should be a source of strength (Figure 1, lower panel). Furthermore, as noted below, it also has the benefit of supporting effective cooperation and coordination.

8. The Financial Stability Review serves as a key communication and accountability vehicle, with clear requirements set out in the FSR Act, which also should foster willingness to act. The Financial Stability Review must, at least biannually, set out: (a) SARB’s assessment of financial stability in the period under review; (b) its identification and assessment of the risks and vulnerabilities to financial stability in at least the next 12 months; (c) an overview of steps taken by it and the financial sector regulators to identify and manage risks, weaknesses or disruptions in the financial system during the period under review and that are envisaged to be taken during at least the next 12 months; and (d) an overview of recommendations made by it and the FSOC during the period under review and progress made in implementing those recommendations.² As such it serves as an effective assessment of risks and as a tool to communicate clearly on those and the policies taken to alleviate them.

9. The FSC does not provide a summary of its meetings, except what is included in the Financial Stability Review. As the FSC meets four times a year and the Review is only published biannually, it implies that no information about the FSC’s assessment of risks or need for policy action is published every other time it meets.

B. Ability to Act

10. Willingness to act is of limited use if not combined with powers to take action to mitigate systemic risk, as well as strong institutional capacity and influence. Those powers can be of a different type and also var y in terms of strength. In par ticular , they can refer to powers to collect information and close data gaps, determine the design and calibration of regulatory constraints, and powers to designate SIFIs and initiate changes in the regulatory perimeter to address leakages where financial intermediation migrates outside of the perimeter. In terms of strength, powers can be ‘hard’ (direct control), ‘semi-hard’ (recommendations with a ‘comply or explain’ mechanism), or ‘soft’ (express opinion without need for follow-up by other agencies). Institutional capacity and influence involve having the resources to assess when action is needed and the status, going beyond in legal terms, to overcome potential resistance against needed policy actions.

11. The SARB’s powers are somewhat constrained outside of declared systemic events, being mostly confined to designating SIFIs and setting macroprudential policy constraints on SIFIs, but SARB remains a strong institution which strengthens its overall ability to act. The Governor designates which financial institutions are SIFIs, after having consulted the FSOC. This decision must take into account, among other things, the size, complexity, interconnectedness, and substitutability of the institution in question. Based on this designation, the SARB can direct the PA, after having consulted with it, to impose various types of macroprudential regulations on SIFIs. These include capital and liquidity tools (including a countercyclical capital buffer, CCyB, which, different from other macroprudential tools, would apply to all banks, not just SIFIs), but also a broader set of constraints (Article 30 in the FSR Act). The SARB may also advice other financial regulators, and any organ of state, to take steps to mitigate risks, but without any ‘comply-or-explain’ mechanism attached to it. The SARB’s advice may carry extra weight due to its credibility in South Africa.

12. To attain a wider range of powers, the Governor needs to declare an event as a ‘systemic event’. The FSR Act gives the Governor, after having consulted the Minister of Finance and potentially the FSOC, the power to declare that a specific event or circumstances is a ‘systemic event’, regardless of whether they have already taken place or not or are considered imminent. Following such a declaration, the SARB “must take all reasonable steps” to mitigate the risk and safeguard financial stability. If such a declaration is made, all financial sector regulators must provide the SARB with any relevant information and consult the SARB before exercising their own regulatory powers. In the case where a systemic event has already occurred or is imminent, the SARB in effect attains authority over regulators’ directives to address the event (explicitly in the case of the PA, FSCA, and the FIC).³

13. However, this is problematic in practice, may foster the inaction bias, and it is telling that the COVID-19 pandemic was not labeled as a systemic event. This reflects the stigma associated with such a declaration and the risk of further fueling financial dislocations by facilitating further concerns about the severity of the event. As a result, policy makers may be reluctant to make such a declaration, or at least from a macroprudential policy perspective, it risks action being taken too late. This institutional feature may also blur the distinction between macroprudential policy and crisis management policies, which may give rise to tensions between the two.

C. Coordination and Cooperation

14. Effective institutional arrangements to ensure close cooperation between regulators serve as important mitigants against the inaction bias and somewhat narrow powers of the SARB. This is reflected by in particular, the strong legal requirements for cooperation, the committee structures that entail overlap in membership (importantly also across to monetary policy), explicit financial stability mandates of institutions, and presence of information sharing mechanisms.

D. Recommendations

15. The SARBs power to put in place macroprudential measures should extend beyond SIFIs if needed, and the need to declare an event as systemic should be reconsidered. As noted above SARB’s authority is mostly confined to SIFIs, except when the Governor declared an event as systemic. While SIFIs can on their own pose systemic risks, such risks can also arise without any failur e of an individual financial institution where e.g., capital or liquidity stress arise and leads to the disruption of financial services (IMF-FSB-BIS, 2009). This arrangement also makes the institutional framework vulnerable to leakage, reflecting that financial intermediation may migrate outside of the SIFIs that are subject to macroprudential restrictions, in particular if regulatory constraints are tightened further. As discussed below, the macroprudential policy toolkit is narrow in South Africa compared to peers and bank centric. If the toolkit were to broaden, or become more tightly calibrated, the risk of domestic and cross-border leakage would likely rise, in particular if credit developments would start to improve considerably. Furthermore, the need to declare an event as systemic entails risk of unintended side-effects, including fostering too late policy actions, in particular in cases where policy actions need to go beyond the SIFIs to be effective. Hence, this could be reconsidered, in particular when other amendments to the FSR Act would be planned as well.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

• Download Figure
• Download figure as PowerPoint slide

• Download Figure
• Download figure as PowerPoint slide

16. The FSC could consider publishing a short record of its meeting, in particular when the Financial Stability Review is not being published, to further enhance transparency. This would provide market participants and the general public with more frequent assessment of systemic risk and how the FSC sees it evolving between meetings. Examples of such arrangements include Iceland and Romania. This could help overcoming the inaction bias as well as foster a broader understanding of macroprudential policy. Providing at least the FSOC with an opportunity to comment on the meeting records in a similar manner as they do with the Financial Stability Review could be continued.

A. Systemic Risk Monitoring Framework

18. SARB’s systemic risk monitoring framework is well advanced and has followed examples set by advanced and EME central banks alike, as well as Fund guidance. SARB utilizes a number of relevant macrofinancial indicators to build a heatmap of different types of risks. It also relies on financial conditions and financial cycle analysis, as well as the growth-at-risk framework. Stress test capacity has been built-up in line with the 2014 FSAP r eco mmenda tio ns and include b oth solvency and liquidity stress tests. Finally, the SARB summarizes its overall assessment in the risk assessment matrix (RAM) and how it has evolved over time in the Financial Stability Review (Figure 2, text figure).

South Africa: Systemic Risk Monitoring

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

Source: SARB Financial Stability Review 2021:1.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

South Africa: Systemic Risk Monitoring

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

Source: SARB Financial Stability Review 2021:1.

• Download Figure
• Download figure as PowerPoint slide

South Africa: Systemic Risk Monitoring

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

Source: SARB Financial Stability Review 2021:1.

• Download Figure
• Download figure as PowerPoint slide

19. However, SARB’s access to micro-level data is limited, providing limitations for assessment of tail risks, as well as to guide the calibration of macroprudential tools.⁴ Systemic vulnerabilities can arise against the backdrop of relatively slow-growing broad credit developments, indeed tail risk play an important role in systemic risk monitoring. Hence, there has been a growing awareness of complementing aggregate (or sectoral) data with micro-level data that allows for looking at the distribution across different types of borrowers. This allows for more timely identification of growing tail risks and to assess the need for policy action in a targeted manner. Access to micro-level data can faciliate the calibration of, in paricular, borrower-based, tools (Nier et al., 2019), assess the effects of a crisis on borrowers (Olafsson and Vignisdottir, 2012) or assist policy making more broadly (IFC Bulletin, 2021). Hence, despite the term, access to “micro-level” data is critical from a macroeconomic and financial stability perspective. The SARB is aware of the need to attain micro-level data and work is underway to close this data gap, although it remains uncertain how far these efforts will reach.

B. Recommendations

20. Continue the important work on developing a national central credit registry to faciliate risk analysis and calibration of macroprudential tools. This is a resource-demanding endeavor where different financial regulators need to collaborate and agree on the overall approach and set-up. What is important from the perspective of SARB is to attain timely and reliable micro-level data, preferably not only regardings loans, but also borrower characteristics, including importantly income. The coverage of the loans would need to include unsecured loans as well as mortgages, from both banks and other lenders, for the SARB to be able to effectively monitor risk and leakages. Whether to include corporate loans is a question of resources and the legal and practical challenges associated with that. Some countries have opted to only gather corporate loan data above a certain threshold.

A. Measures to Address the Sovereign-Financial Nexus

27. The sovereign-financial nexus has emerged as a challenge for financial stability in a number of countries, in particular as a result of financial distress and crises where financial institutions may need to step in to provide funding to the sovereign as other sources—domestic and foreign—dry up and/or as domestic private credit risk increases and attaining sovereign exposure becomes more attractive. As a result, distortions arise where private credit is crowded-out and systemic risk may materialize and work through various macrofinancial linkages and feedbacks that have been recognized in the literature (IMF, 2018).

28. Addressing the sovereign-financial nexus has proven difficult for a number of reasons, as also evidenced by the inability to reach international consensus (Basel Committee on Banking Supervision (BCBS), 2017). From a macroprudential perspective, this reflects trade-offs and thereby potential clashes with other policies, risks of unintended side-effects, and the lack of experience and agreement with utilizing macroprudential tools to address this risk. Hence, proceeding with a high degree of caution and aiming to achieve a broad domestic policy agreement is desirable to attain a desirable trade-off.

29. In South Africa, exposure of banks and the financial sector more broadly towards the sovereign is increasing, while still not being excessive in international comparison.⁸ In addition, there are important mitigants in place, such as a relatively benign currency and maturity profile of government debt. However, there are also potential amplifiers. In particular, the fiscal and growth outlook risk aggravating the nexus going forward, as well as if nonresidents’ holdings would continue to decline, which cannot be excluded given the credit rating developments and the potential of a tightening of global financial conditions. Indeed, the risk stemming from the nexus is reflected in the FSAP’s solvency stress test, which r eveals sizable losses resulting from sovereign exposures in the adverse scenario over the 3-year horizon. SARB has also flagged this as a key systemic risk (Figure 2 above, SARB Financial Stability Review 2021:1 andSmith et al., 2020).

South Africa: The Sovereign-Financial Nexus

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

South Africa: The Sovereign-Financial Nexus

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

• Download Figure
• Download figure as PowerPoint slide

South Africa: The Sovereign-Financial Nexus

Citation: IMF Staff Country Reports 2022, 186; 10.5089/9798400214653.002.A001

• Download Figure
• Download figure as PowerPoint slide

30. Fiscal consolidation is likely to have to be key in reducing the sovereign risk in the financial sector, but prudential policies can play an important complementary role. Ensuring debt sustainability and strengthening the growth outlook (which is discussed in the Article IV consultations) could reduce the risk associated with the sovereign-financial nexus and support a wider diversification toward the private sector on financial institutions’ balance sheet. An important question to ask is how prudential policies could complement that, being mindful that there is an underlying distortion arising from how sovereign exposures are generally treated under the Basel III framework.

31. Any prudential measure that could be considered to address the nexus should be mindful of other prudential requirements and the broader need for financial institutions to hold government bonds. In particular, the need to hold high quality liquid assets (HQLAs) to fulfill the LCR requirements—and some voluntary buffer on top of that—and to fulfill banks role in the government bond market. Similarly, many NBFIs also need to hold government bonds, at least voluntarily, for similar liquidity purposes. This will need to be considered in the design of any prudential measure to address the nexus.

32. Similarly, any measure would need to be mindful of the importance of safeguarding the smooth function of the government bond market and banks’ liquidity and profitability. Potential tensions between prudential policies to address the nexus and fiscal policies’ objectives with regard to funding costs and smooth functioning of government bond markets are key to address in the design of any prudential measure aimed at the nexus. Similarly, the effects on financial institutions’ liquidity and profitability would need to be considered carefully, as these are the first line of buffers when shocks materialize—even those stemming from sovereign risk.

33. Hence, prudential measures should aim at (i) ensuring resilience through increased loss absorbing capacity and (ii) providing disincentives against excessive concentration of sovereign risk, while at the same time (iii) avoiding unintended side-effects. As discussed above, these unintended side-effects could take the form of an excessive reduction of liquidity, bond market pressures with resulting effects on government financing, or other unwarranted macro-financial dynamics (see paragraph 35 below). The discussion below will focus on the sovereign-bank nexus, as discussed in paragraph 24, measures on NBFIs are likely to reflect international agreements on strengthening their prudential regulations.

34. Various measures can be considered, but regardless, they would need to be phased-in, be carefully calibrated, and clearly communicated, for them to be likely to achieve their objectives:

• Higher risk weights: Given that a key source of distortions emanates from the preferential risk-weight treatment of sovereign exposures—in line with the national discretion embedded in the Basel framework—increasing RWA arises as a natural response. Importantly, IRB banks have already endogenously increased RWA on their sovereign holdings and thereby hold more capital against them. Similarly, the authorities could consider introducing positive RWA on banks that apply the standardized approach and thereby combine a two-fold solution. However, the potential for procyclical dynamics would need to be considered (Véron, 2017).

• Capital surcharges: Another option would be to apply Pillar 1 or 2 capital surcharges, which can be calibrated in a manner that is commensurate to the assessed risk. Importantly, they should allow for increased resilience and disincentives to building excessive concentration, while limiting risk of unintended side-effects. For instance, by applying positive surcharges only above a certain thresholds (to account for holdings of sovereign bonds to meet liquidity requirements, etc. as listed above), and then rise only gradually as financial institutions’ exposures increase as a share of assets (see an example of such a calibration in 2018 Romania FSSA).⁹ Potentially, capital surcharges—if they are being used as usable buffers—could be relaxed if sovereign risk would materialize, providing additional resilience. However, any such relaxation would need to be implemented carefully and judgement applied to the cost-benefit trade-offs (see IMF, 2014), including how useable such released buffers would be in practice.

• Concentration limits: A third option is to use a quantitative measure, instead of price-based ones, in the form of concentration limits. Such measures would reduce concentration, but unintended side-effects in the form of ‘cliff effects’ may materialize when banks would come close to the limits and might need to resort to fire sales of bonds with detrimental effects.

35. Quantitative assessments could guide the calibration and help with communication of prudential measures to address the sovereign-financial nexus. Stress tests can help with assessing the level of resilience which any measure should aim to provide, for instance, through higher RWs or capital surcharges. Not necessarily such that the measure should provide ‘insurance’ against the full set of losses in an adverse scenario, but to pass judgement on the desired level of resilience against these potential losses. Furthermore, quantitative impact assessments could look into the potential effects of any measure taken on at least four areas: (i) bank funding costs; (ii) bond yields; (iii) bank profitability; and (iv) bank liquidity position. For an example of such an analysis, see the2018 Romania FSSA.

36. Another challenge associated with introducing prudential measures to address the sovereign-financial nexus, is on timing. SARB already has a framework in place to guide the introduction of macroprudential tools, and how to weigh the cost-benefit trade-offs—including assessing the cost of inaction—which could be useful in this regard.

37. It is advisable to start the implementation before there is a significant further build-up of this risk in financial institutions’ balance sheets. Also, taking into account the normalization of pandemic-related relaxation of capital requirements and the broader economic context. A reasonable transition period will be needed to allow financial institutions time to adjust their balance sheets; but a near-term announcement of envisaged measures, with the applicable transition period, can help condition behavior and thus smooth the adjustment process.

38. Communication is going to be key to smooth implementation of any new measures. The SARB has already taken a crucial step in this regard by providing a thorough analysis of the risks associated with the nexus. Next step is to set out the objectives, weigh the different policy options against those objectives, and perform a quantitative analysis to guide the calibration. Once a suggested calibration is in place, the various platforms in the committee structures can be utilized to build consensus, followed by a more-wider communication and a phase-in period.

B. Borrower-Based Tools

39. Borrower-based tools are among the most widely used macroprudential policy tools around the world (Alam et al., 2019), but they have so far not been implemented in South Africa. This is somewhat understandable given the broader macro-financial context in the post-GFC period, as well as due to the time it often takes to be able to use such tools in an effective manner given that preparations have been underway by the SARB.

40. Going forward, such tools might be needed to complement the mostly capital tools that are in place. Experience shows that capital-based tools, while providing important resilience against shocks, are less effective than borrower-based tools in reducing credit growth and feedbacks between asset prices and credit. Borrower-based tools also have the benefit of being targeted and can thereby complement broader capital tools to achieve an improved cost-benefit trade-offs if broader credit developments are weak but the are pockets of vulnerabilities in certain sectors.

41. The SARB is well-aware of this, and as noted above, work has been underway to broaden the toolkit to include borrower-based tools, such as caps on loan-to-value (LTV), debt service to income (DSTI) or debt/loan to income (DTI/LTI). Also taking into account the resilience enhancement of borrowers embedded in the NCA legislation.

42. Closing data gaps is an important element to foster that broadening, including preferably through a credit registry, and making sure the legal framework allows for the use of these tools. The importance of closing data gaps was discussed above. In many instances the legal basis for allowing the use of these tools has to be strengthened. The FSR Act provides for the use of sectoral tools, but it is not clear whether that applies only to sectoral capital (and liquidity) tools or also borrower-based tools. Most likely, a 2^nd tier legislation to the FSR Act could establish a framework for the application of borrower based macroprudential tools.

43. As these tools can be quite effective, it is important to consider avoiding excessive costs, and for South Africa, not to undermine efforts to enhance financial inclusion. These cost include (healthy) borrowing being discouraged (potentially slowing down the post-COVID recovery), excessively undermining consumption growth and thereby economic growth by requiring too much saving, and undermining financial inclusion. Ultimately, if these costs become excessive, macroprudential policy could risk losing its autonomy. Various design features can be considered to achieve a more beneficial cost-benefit trade-off when using borrower-based tools. In particular,

• More generous LTV caps for first time buyers;

• Allow for “speed limits” where banks can deviate from the requirements for a certain share of new loans;

• Consider ways to allow banks to perform their own income assessments of borrowers to take all income into account which appears to be addressed in the NCA;

• Consider issuing guidelines on LTV and DSTI ahead of binding rules to be able to observe their impact and guide the calibration;

• Aim to start the caps at a relatively neutral rates and gradually tighten.

44. To be effective, it is generally advisable to rely on a portfolio of borrower-based tools rather than any one single tool. In particular, if sectoral credit growth is strong an income-related measure (such as a DSTI or LTI) is likely to be needed as those are less affected by interactions between asset prices and credit as the LTV.

C. Recommendations

45. Consider carefully calibrated prudential measures to address the sovereign-financial nexus. The measure(s) should have clear objectives (as suggested above), be calibrated and implemented manner to achieve those objectives, and be clearly communicated to foster understanding of the need for policy action and diminish concerns of unintended side-effects.

46. Continue preparations to use borrower-based tools and consider design features that could allow for a more beneficial cost-benefit tradeoffs, taking South African country-specific circumstances into account.