## 1bwaea2024003

## Source details

**Canonical URL:** [1bwaea2024003](https://www.imf.org/-/media/files/publications/cr/2024/english/1bwaea2024003.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2024/english/1bwaea2024003.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2024/english/1bwaea2024003.pdf.json)

---

### Executive Summary — Overview
- FSAP mission conducted risk analysis for large insurance companies and retirement funds as part of the 2023 Financial Sector Assessment Program (FSAP).
- Sample coverage:
  - Four life insurers, four short-term insurers and four retirement funds.
  - Market coverage between 80 and 95 percent in each sector (sample details provided in Stress Test Results section).
- Data limitations:
  - Incomplete reporting data complicated top-down modelling, specifically regarding the geographical breakdown of investments for insurers and the valuation of insurance liabilities.
- NBFIRA background and resources:
  - NBFIRA established by the Non-Bank Financial Institutions Regulatory Authority Act 2006.
  - As of March 2022, NBFIRA had a staff of 89 and an annual budget of around BWP 80m.
  - NBFIRA supervises the activities of around 800 entities.

### Insurance sector — structure, shocks, and solvency findings
- Focus of analysis:
  - Solvency under the adverse macrofinancial scenario used in the banking stress test; complemented by sensitivity analyses (interest rate and currency shocks, default of the largest banking counterparty).
- Asset and capital impacts in the adverse scenario:
  - Median life insurer: assets decrease by 7 percent.
  - Some life insurers: liabilities related to annuity business and unit-linked insurance decline at a similar amount as assets.
  - Median short-term insurer: asset values decrease by 1 percent.
  - Excess of assets over liabilities:
    - Median life insurer: shrinks by 12 percent.
    - Median short-term firm: shrinks by 2 percent.
  - Drivers: shocks to domestic bond and equity prices contribute most to the decline in capital.
  - BWP depreciation increases the value of foreign assets (a compensating factor for life insurers).
  - Large cash holdings result in overall low sensitivity to market risks, in particular for short-term insurers.
- Solvency levels:
  - Median life insurer solvency ratio: declines by 12 percentage points, from 325 percent to 313 percent.
  - 25th and 75th percentiles (life): decline each by around 40 percentage points.
  - Median short-term firm solvency ratio: decreases by 2 percentage points to 155 percent.
- Vulnerabilities and transmission channels:
  - Concentrated deposits with individual domestic banks present a significant transmission channel for systemic risk.
  - Largest deposits with a single banking counterparty amount to more than one third and up to half of shareholder capital for most insurers in the sample; for one firm the largest deposit exposure exceeds their capital.
- Modelling constraints:
  - Incomplete reporting on investments’ geography and insurance liability valuation limited the top-down modelling.
  - Reporting templates lack cumulative exposure toward a counterparty across different classes and omit detailed data on derivatives and granular credit ratings.
  - Potential reactive management actions (de-risking, reinsurance changes, dividend cuts) were not modeled.

### Retirement funds — design, asset allocation, and impact on members
- Scheme design and market size:
  - Retirement funds are almost entirely defined-contribution schemes, transferring investment risk to scheme members.
  - Retirement funds assets at end-2021: BWP 110bn.
  - Retirement funds contributions: BWP 5.6bn in 2021 (BWP 3.8bn in 2016).
- Asset allocation and domicile:
  - 68 percent invested in stocks (stocks / total investments: 67.9 percent, retirement funds FSI).
  - 66 percent invested abroad (foreign assets / total investments: 66.5 percent).
  - Offshore equity accounts for 52 percent of total assets.
- Representative member scenarios and impacts:
  - Member A: 10 years prior to retirement; BWP 1,000,000 accumulated pension savings; BWP 20,000 annual contributions.
    - Adverse scenario impact: future pension values decline by around 17 percent compared to the baseline.
  - Member B: 30 years prior to retirement; BWP 200,000 accumulated pension savings; BWP 12,000 annual contributions.
    - Adverse scenario impact: future pension values decline by 4 percent compared to the baseline.
  - Largest part of decline stems from lower stock prices in the scenario.
- Policy implication for members:
  - Members close to retirement should switch to a more conservative asset allocation to avoid inability to fully recover after shocks.
- Liquidity and market implications from legislative change:
  - The new Retirement Fund Act (2022) provides more (lump-sum) withdrawal options for pension scheme members, transferring liquidity from pension funds to households.
  - Payments upon retirement currently account for around 80 percent of all payouts; of these around 40 percent are paid as a lump-sum.
  - The Act reduces sequentially the cap on foreign assets, raising questions about the capacity of domestic capital markets to absorb repatriated investments.
  - Effects will be seen in other financial sectors, e.g., in life insurers’ annuity business.

### New Retirement Fund Act (2022) — liquidity and market implications
- Key features and implications:
  - More lump-sum withdrawal options upon retirement and prior to retirement (e.g., to repay mortgages, pay medical bills).
  - Transfers liquidity from pension funds to households, increasing the need for robust liquidity management by funds.
  - Considerable variation across funds depending on age and salary composition of members and employer industry cyclicality.
  - Smaller funds, and funds where benefits already exceed contributions, will need robust liquidity risk management.
  - Sequential reduction of the cap on foreign assets raises concerns about domestic market absorption of repatriated investments.
  - Potential spillovers to life insurers’ annuity business if pension funds transfer less accrued pension interest to annuity providers.

### Supervisory capacity, data, and analytical needs
- Identified needs for NBFIRA:
  - Review supervisory reporting templates and amend data items necessary for financial stability analysis (including alignment with IFRS 17).
  - Introduce a data management system allowing horizontal and time-series analysis and financial stability analysis.
  - Regular stress testing of large insurers and retirement funds and analysis of risks from concentrated bank exposures.
  - In-house actuarial expertise required, together with specific knowledge on reinsurance business and intensified cooperation with home supervisors of local reinsurance subsidiaries.
  - Analysis of the impact of IFRS 17 on life insurers’ business models and strategy.
- Capacity specifics:
  - As of March 2022, NBFIRA staff: 89; annual budget: around BWP 80m.
  - NBFIRA supervised a total of 812 entities as of end-2021 (including 87 retirement funds, 9 life insurers, 12 short-term insurers, 5 reinsurers).

### Stress test methodology, scenarios, and key parameters
- General approach:
  - Top-down stress test (TD) for large insurers and retirement funds using adverse macrofinancial scenario from the banking ST with added granularity for market shocks; complemented by sensitivity analyses.
- Scope and sample:
  - TD solvency stress test on solo-entity basis for:
    - Four life insurers (market coverage: 93 percent in life sector based on assets).
    - Four short-term insurers (market coverage: 77 percent in short-term sector based on gross written premiums).
    - Four retirement funds (cover 81 percent of retirement funds market based on assets); balance sheet assets of participating retirement funds amount to BWP 100bn.
  - Aggregated balance sheet assets of the four insurer participants: BWP 18.2bn (BWP 17bn from four life insurers).
- Reference date and timing:
  - Reference date: 30 June 2022.
  - For insurers, shocks applied instantaneously at start of first year (market shocks front-loaded).
  - For retirement funds, shocks spread over the first two years of projection horizon.
- Market risk shocks applied (values in percent):
  - Equity: Domestic -25.1
  - Equity: Offshore, advanced economies -28.7
  - Equity: Offshore, emerging markets -25.1
  - Bonds: Domestic Short-term (~1-3 years) -9.2
  - Bonds: Domestic Long-term (~5-10 years) -7.3
  - Bonds: Offshore -15.0
  - Property -15.0
  - Currency: External value of BWP against USD, EUR, GBP -20.6
    - Note: Negative value means a depreciation of the BWP.
- Additional sensitivity tests (single-factor shocks):
  - Currency: Increase of the BWP external value by 10 percent.
  - Equity: Decline of domestic and foreign equity prices by 40 percent.
  - Counterparty risk: Default of the largest banking counterparty.
  - Sensitivity results were not aggregated to the adverse scenario results.
- Valuation and PCT assumptions:
  - Current valuation regime and PCT introduced in 2012 form the basis for the insurance stress test.
  - Most assets valued mark-to-market; exceptions for group undertakings.
  - Liabilities: no detailed prescribed valuation method beyond “best-estimate valuation assumptions, adjusted by compulsory margins and possibly also by discretionary margins.”
  - IFRS 17 being implemented mainly by larger insurers; authorities have not adopted a strategy how to embed IFRS 17 in prudential regime—resulting in a parallel regime of prudential valuation and IFRS accounting.
  - PCT calculation differences between life and short-term insurers; capital requirement for short-term insurers excludes an operational risk module; explicit requirement for asset concentration risk (including counterparty concentration) is missing for both sectors.
  - Reporting limitation: mid-year PCT composition assumed same as end-2021 for stress test.
- Stress test mechanics:
  - Haircuts aligned with adverse scenario applied to asset classes; for life insurers interest rate shock applied to liabilities (separate for annuity and non-annuity business); short-term insurers do not discount future cash flows.
  - Unit-linked liabilities decline mirrored market value loss of underlying assets.
  - Available capital proxied by difference between post-stress assets and post-stress liabilities.
  - PCT partly recalculated after stress: market/investment risk component proportionally adjusted for reductions in exposures due to stress materialization.
- Data used:
  - Annual and quarterly prudential reporting templates for balance sheets, composition of PCT, and sensitivity analyses for bank counterparty exposures.

### Stress test results, sensitivities, and other risks
- Insurance stress test summary:
  - Adverse scenario, in isolation, does not cause any capital shortfalls.
  - Heterogeneity noted across life insurers due to business models, asset allocation, and asset and liability durations.
- Sensitivity analyses (insurance):
  - Equity shock (-40 percent domestic and foreign) would reduce excess of assets over liabilities for the median life insurer by 13 percent; for one firm reduction would be well more than 50 percent.
  - BWP appreciation by 10 percent: for the median life insurer excess of assets over liabilities would decline by 3 percent.
  - Concentrated bank exposures: largest deposit exposures can be a multiple of capital for some firms; further analysis should consider loss given default and second-round effects.
- Other risks:
  - Reinsurance: reinsurance business primarily channeled through domestic reinsurers that are subsidiaries of foreign groups; need for strong cooperation and data sharing with home authorities and development of specialized supervisory knowledge.
  - Climate risk: currently small in insured losses but Botswana faces higher temperatures and droughts; critical to evaluate insurability and protection gaps.
- Retirement funds stress test summary:
  - Representative member with 10 years to retirement: future pension values decline by around 17 percent compared to baseline.
  - Representative member with 30 years to retirement: future pension values decline by 4 percent.
  - Policy recommendation: timely switch of pre-retirement members to more conservative allocations; supervision intensity to increase for vulnerable funds and requirement of liquidity buffers where necessary.

### Key statistics and financial soundness indicators (selected figures)
- NBFI assets at end-2021: BWP 153bn.
- Retirement funds assets: BWP 110bn.
- Retirement funds contributions: BWP 5.6bn in 2021 (BWP 3.8bn in 2016).
- Retirement funds membership (Dec 2021): 293,015 active members; 51,806 deferred members; 13,187 pensioners.
- Gross written life premiums (2021): BWP 4.4bn; growth since 2016: 6.9 percent per year.
- Gross written short-term premiums (2021): BWP 1.6bn; growth since 2016: 4.6 percent per year.
- Reinsurance premiums (2021): BWP 466m (BWP 175m five years earlier).
- Investment income of retirement funds (2021): BWP 17bn; investment return 15.1 percent.
- Stress test sample coverage:
  - Life insurers: 93 percent of life sector assets.
  - Short-term insurers: 77 percent based on gross written premiums.
  - Retirement funds: four funds cover 81 percent of market by assets; aggregated participating funds’ balance sheet assets amount to BWP 100bn.
- Stress shock magnitudes (selected):
  - Domestic equity: -25.1 percent
  - Offshore equity (advanced economies): -28.7 percent
  - Offshore equity (emerging markets): -25.1 percent
  - Domestic bonds short-term (~1-3 years): -9.2 percent
  - Domestic bonds long-term (~5-10 years): -7.3 percent
  - Offshore bonds: -15.0 percent
  - Property: -15.0 percent
  - Pula depreciation vs USD/EUR/GBP: -20.6 percent
- Selected FSIs (2021 unless noted):
  - Assets / liabilities – life: 16.5
  - Assets / liabilities – short-term: 30.0
  - Solvency ratio – life: 669.4
  - Solvency ratio – short-term: 170.6
  - Investment income / total assets – life: 5.0
  - Cash and deposits / total assets – short-term: 49.8
  - Risk retention ratio – life (net premium / gross premium): 96.9
  - Risk retention ratio – short-term: 55.1
  - Retirement funds: Investment income / total investments (average): 15.1; Stocks / total investments: 67.9; Bonds / total investments: 13.7; Foreign assets / total investments: 66.5; Cash and deposits / total investments: 6.3; Contributions / benefits: 153.5

### Main recommendations (summary from Table 1)
- Data (Addressee; Timing*; Priority**)
  1. Ensure complete and consistent submission of reporting data (¶46) — NBFIRA; I-C; H
  2. Introduce a data management system which allows for horizontal and time-series analysis (¶46) — NBFIRA; I-C; H
  3. Review reporting templates and amend data items necessary for financial stability analysis (¶46) — NBFIRA; ST; H
  4. Review reporting templates to align with IFRS 17 (¶46) — NBFIRA; MT; M
- Risk Analysis
  5. Undertake a comprehensive impact assessment of the new Retirement Fund Act on the financial sector (¶45) — MoF, BoB, NBFIRA, FSC; I; H
  6. Analyze impact of concentrated bank exposures and relevant transmission channels (¶47) — NBFIRA, BoB, FSC; ST; H
  7. Conduct regular top-down stress tests for large insurers and retirement funds and feed results into the supervisory dialogue with firms (¶47) — NBFIRA; ST; M
- Other supervisory and capacity actions
  8. Intensify cooperation with home authorities on the supervision of local reinsurance subsidiaries, and build up specialized knowledge about reinsurance business (¶38) — NBFIRA; ST; M
  9. Establish a robust supervisory framework for liquidity risks of retirement funds and consider the use of liquidity buffers (¶44) — NBFIRA; I; H
  10. Expand number of (experienced) actuarial staff at NBFIRA (¶48) — MoF, NBFIRA; I; H
  11. Develop an understanding of the impact of IFRS 17 on life insurers business models and strategy (¶48) — NBFIRA; ST; M
- Timing codes:
  - C = Continuous; I = Immediate (within one year); ST = Short Term (within 1-3 years); MT = Medium Term (within 3-5 years).
- Priority codes:
  - H = High; M = Medium; L = Low.

*Source: IMF staff calculations and NBFIRA data as presented in the IMF FSAP technical note (executive summary and sections 4–5).*

### EXECUTIVE SUMMARY __________________________________________________________________________ 5

### EXECUTIVE SUMMARY

### Overview
- The FSAP mission conducted a risk analysis for large insurance companies and retirement funds as part of the 2023 Financial Sector Assessment Program (FSAP).
- Sample: four life insurers, four short-term insurers and four retirement funds, with a market coverage between 80 and 95 percent in each sector.
- Data limitations: Incomplete reporting data complicated top-down modelling, specifically regarding the geographical breakdown of investments for insurers and the valuation of insurance liabilities.
- NBFIRA background and resources:
  - NBFIRA was established by the Non-Bank Financial Institutions Regulatory Authority Act 2006.
  - As of March 2022, NBFIRA had a staff of 89 and an annual budget of around BWP 80m.
  - NBFIRA supervises the activities of around 800 entities.

### Insurance sector — structure, shocks, and solvency findings
- Focus of analysis: solvency under the adverse macrofinancial scenario also used in the banking stress test; complemented by sensitivity analyses (interest rate and currency shocks, default of the largest banking counterparty).
- Key asset and capital impacts in the adverse scenario:
  - For the median life insurer, assets decrease by 7 percent.
  - For some life insurers, liabilities—related to annuity business and unit-linked insurance—decline at a similar amount as assets.
  - Shocks to domestic bond and equity prices contribute most to the decline in capital.
  - Large cash holdings result in an overall low sensitivity to market risks, in particular for short-term insurers.
  - BWP depreciation increases the value of foreign assets.
- Overall solvency result:
  - The adverse scenario, in isolation, does not cause any capital shortfalls.
  - Vulnerability: concentrated deposits with individual domestic banks present a significant transmission channel for systemic risk.
- Modelling constraints: incomplete reporting on investments’ geography and insurance liability valuation limited the top-down modelling.

### Retirement funds — design, asset allocation, and impact on members
- Design: Retirement funds in Botswana are designed as defined contribution schemes, transferring investment risk to scheme members.
- Asset allocation and domicile:
  - 68 percent invested in stocks.
  - 66 percent invested abroad.
- Analysis focus: future pension values for representative members 10 and 30 years prior to retirement (rather than funding ratios).
- Scenario impacts:
  - The adverse scenario has a sizable impact on members with only a few years to retirement, reducing future pension values by more than 15 percent compared to the baseline.
  - For members with 30 years to retirement, the decline in pension values amounts to less than 5 percent.
  - The largest part of the decline stems from lower stock prices in the scenario.
- Policy implication for members: members close to retirement should switch to a more conservative asset allocation to avoid inability to fully recover after shocks.

### New Retirement Fund Act (2022) — liquidity and market implications
- The new Retirement Fund Act, adopted in 2022, is very likely to impact the liquidity of retirement funds, calling for robust liquidity management.
- Key features and implications:
  - The Act provides more (lump-sum) withdrawal options for pension scheme members, effectively transferring liquidity from pension funds to households.
  - Payments upon retirement currently account for only around 80 percent of all payouts; of these around 40 percent are paid as a lump-sum.
  - There is considerable variation between retirement funds depending on age and salary composition of members and the cyclicality of the employer’s industry.
  - Smaller funds, and those where benefits already exceed contributions, will need robust liquidity risk management.
  - The Act reduces sequentially the cap on foreign assets, raising questions about the capacity of domestic capital markets to absorb repatriated investments.
  - Effects will be seen in other financial sectors, e.g., in life insurers’ annuity business.

### Supervisory capacity, data, and analytical needs
- The risk analysis highlights the need for NBFIRA to substantially improve data analytics and resources.
- Specific needs:
  - A review of supervisory reporting templates and more robust processes to ensure complete and consistent reporting.
  - Introduction of a data management system allowing horizontal and financial stability analysis.
  - Regular stress testing of large insurers and retirement funds and analysis of risks from concentrated bank exposures.
  - In-house actuarial expertise is required at NBFIRA, together with specific knowledge on reinsurance business and intensified cooperation with home supervisors of local reinsurance subsidiaries.
  - Analysis of the impact of IFRS 17 on life insurers’ business models and strategy.

### Main recommendations (from Table 1)
- Data (Addressee; Timing*; Priority**)
  1. Ensure complete and consistent submission of reporting data (¶46) — NBFIRA; I-C; H
  2. Introduce a data management system which allows for horizontal and time-series analysis (¶46) — NBFIRA; I-C; H
  3. Review reporting templates and amend data items necessary for financial stability analysis (¶46) — NBFIRA; ST; H
  4. Review reporting templates to align with IFRS 17 (¶46) — NBFIRA; MT; M
- Risk Analysis
  5. Undertake a comprehensive impact assessment of the new Retirement Fund Act on the financial sector (¶45) — MoF, BoB, NBFIRA, FSC; I; H
  6. Analyze impact of concentrated bank exposures and relevant transmission channels (¶47) — NBFIRA, BoB, FSC; ST; H
  7. Conduct regular top-down stress tests for large insurers and retirement funds and feed results into the supervisory dialogue with firms (¶47) — NBFIRA; ST; M
- Other supervisory and capacity actions
  8. Intensify cooperation with home authorities on the supervision of local reinsurance subsidiaries, and build up specialized knowledge about reinsurance business (¶38) — NBFIRA; ST; M
  9. Establish a robust supervisory framework for liquidity risks of retirement funds and consider the use of liquidity buffers (¶44) — NBFIRA; I; H
  10. Expand number of (experienced) actuarial staff at NBFIRA (¶48) — MoF, NBFIRA; I; H
  11. Develop an understanding of the impact of IFRS 17 on life insurers business models and strategy (¶48) — NBFIRA; ST; M
- *Timing codes: C = Continuous; I = Immediate (within one year); ST = Short Term (within 1-3 years); MT = Medium Term (within 3-5 years).
- **Priority codes: H = High; M = Medium; L = Low.

*EXECUTIVE SUMMARY, IMF FSAP Technical Note*

### 4.      The non-bank financial sector in Botswana is sizable, accounting for around 57 percent

### 4.      The non-bank financial sector in Botswana is sizable, accounting for around 57 percent

### Overview
- At end-2021, assets of the non-bank financial sector (NBFI) amount to BWP 153bn (up from 115bn in 2016), or 78 percent of GDP.
- The NBFI sector corresponds to around 57 percent of the domestic financial system assets at end-2021.
- The largest NBFI sub-sector is retirement funds with accumulated balance sheet assets of BWP 110bn, exceeding the size of the banking sector by a small margin.
- NBFIRA supervised a total of 812 entities as of end-2021, including:
  - 87 retirement funds
  - 9 life insurers
  - 12 short-term (life) insurers
  - 5 reinsurers
- Given their size, retirement funds—and, to a lesser extent, life insurers—are important players and highly interconnected domestically and regionally (many large insurers are subsidiaries of South African insurance groups).

### Insurance Sector: Structure and Market Indicators
- Insurance density (2021):
  - USD 186 per person total
  - USD 136 life insurance premiums per person
  - USD 50 short-term (non-life) insurance premiums per person
- Insurance penetration (premiums as percent of GDP, 2021):
  - 2.3 percent in the life sector
  - 0.8 percent in the non-life sector
- Life insurance:
  - Business is relatively traditional (funeral insurance, credit life insurance; largest life insurer maintains a large annuity book).
  - Very concentrated: three largest life insurers hold a combined market share of more than 90 percent.
  - Gross written life premiums: BWP 4.4bn in 2021.
  - Average growth in gross written life premiums since 2016: 6.9 percent per year.
  - Reinsurance retention: around 97 percent retained (only around 3 percent ceded to reinsurers).
- Short-term insurance:
  - Gross written short-term premiums: BWP 1.6bn in 2021.
  - Average growth since 2016: 4.6 percent per year.
  - Largest lines of business: property (32 percent of gross written premiums) and motor (29 percent).
  - Concentration: three largest entities (out of twelve) hold a market share of 69 percent.
  - Retention rate in 2021: 55 percent (2016–20 average: 63 percent).
- Reinsurance:
  - Reinsurers wrote premiums of BWP 466m in 2021, up from BWP 175m five years earlier.
  - Number of reinsurers increased to five (two new reinsurers entered since 2020).
  - All reinsurers are subsidiaries of foreign groups (four headquartered in Zimbabwe, one in Nigeria).
  - Regulation: primary insurers ceding business must contract with domestic reinsurers unless domestic capacity is insufficient and foreign approval is granted by NBFIRA.
- Distribution and interlinkages:
  - Majority of insurers are subsidiaries of international groups and often interlinked with related entities within the group (cross-sectoral groups offering banking, insurance, asset management).
  - Insurance brokers and agents account for about 60 percent of total premiums.
  - As of March 2022, 144 insurance agents and 58 insurance brokers were active and licensed by NBFIRA.

### Insurance Sector: Asset Allocation, Profitability, and Solvency
- Asset allocation (stress test participants):
  - Life insurers: 27 percent in government bonds, 22 percent in corporate bonds, 13 percent in shares, 26 percent in unit-linked assets.
  - Short-term insurers: almost two thirds (65 percent) of assets in cash and deposits; equity and bonds each represent less than 5 percent.
- Profitability (2016–2021 and 2021 specifics):
  - Investment income for life insurers: 5 percent as share of total assets in both 2020 and 2021.
  - Increased mortality during COVID-19: net claims for life insurers increased in 2021 by 43 percent compared to 2020.
  - Short-term combined ratio: 96 percent in 2021 (below 100 percent for the last four years).
  - Reinsurers’ combined ratio: 82 percent in 2021; fluctuated between 74 percent and 115 percent in the five years prior.
  - Average return on equity (2016–2021): life insurers 16 percent; short-term insurers 10 percent; reinsurers 11 percent.
- Capital ratios:
  - Life insurers’ average capital ratio (shareholder equity to assets): around 16 percent (stable for last four years).
  - Short-term insurers’ capital ratios: declined since 2016 but at 30 percent in 2021 (significantly higher than life insurers).
  - Prescribed capital target (PCT) composition is not publicly published in full; more risk-based solvency ratios are not published.

### Retirement Funds: Structure, Membership, and Investments
- Scheme type and contributions:
  - Retirement funds are almost entirely defined-contribution schemes; investment risk borne by fund members.
  - Contributions increased from BWP 3.8bn in 2016 to BWP 5.6bn in 2021 (plus 50 percent).
  - Employer contributions constitute 69 percent of total contributions.
- Membership (as of December 2021):
  - 293,015 active members (59 percent of the working population)
  - 51,806 deferred members
  - 13,187 pensioners
  - Deferred members increased substantially in 2021 from around 18,000 at end-2020.
- Market concentration and fund size:
  - Four largest funds account for 81 percent of the market by balance sheet assets.
  - Among these four, three are single-employer funds (including the Botswana Public Officers Pension Fund) and one is a multi-employer fund.
  - Thirty-one funds recorded less than 200 members as of end-2021.
  - Thirty-seven funds managed less than BWP 100m.
  - Number of retirement funds increased from 89 in 2017 to 92 in 2021.
- Investment allocation and risks:
  - Investments are growth-oriented with large holdings in stocks; assets are mostly foreign.
  - Offshore equity accounts for 52 percent of total assets (offshore equity is the dominant asset class).
  - Retirement funds hold about 44 percent of outstanding domestic bonds and correspond to about 52 percent of the Batswana stock market capitalization (domestic firms only).
  - Currency risks from foreign-denominated assets are almost entirely unhedged.
  - Investment income is volatile due to market valuation and inclusion of unrealized gains and losses in investment income measurement.
  - Investment income: record BWP 17bn in 2021 (investment return of 15.1 percent), driven by market rebound after fiscal year-end 2020.

### Stress Test Methodology and Scope
- General approach:
  - FSAP conducted a top-down stress test (TD) for large insurance companies and retirement funds using the adverse macrofinancial scenario from the banking ST, with added granularity for market shocks; sensitivity analyses (interest rate, currency shocks, default of largest banking counterparty) complemented the analysis.
- Scope and sample:
  - TD solvency stress test on solo-entity basis for:
    - Four large life insurers (market coverage: 93 percent in life sector based on assets).
    - Four short-term insurers (market coverage: 77 percent in short-term sector based on gross written premiums).
    - Four retirement funds (defined-contribution; cover 81 percent of retirement funds market based on assets); balance sheet assets of participating retirement funds amount to BWP 100bn.
  - Aggregated balance sheet assets of the four insurer participants: BWP 18.2bn (BWP 17bn from four life insurers).
- Stress scenario narrative:
  - Adverse macrofinancial scenario: stagflationary environment with high inflation from volatile food and energy prices, commodity price increases due to spillovers from the war in Ukraine; tighter global monetary policy; employment declines; diamond price decreases; Pula depreciates due to higher domestic inflation relative to trading partners.
- Scenario adjustments for NBFI stress test:
  - For insurers, shocks are instantaneous at start of first year (30 June 2022 reference date); market shocks front-loaded to realize maximum drawdown immediately.
  - For retirement funds, shocks are spread over the first two years aligning with longer projection horizon for pension values.
- Market risk shocks applied (Table 4: Stress Test Specification — shocks in percent):
  - Equity: Domestic -25.1
  - Equity: Offshore, advanced economies -28.7
  - Equity: Offshore, emerging markets -25.1
  - Bonds: Domestic Short-term (~1-3 years) -9.2
  - Bonds: Domestic Long-term (~5-10 years) -7.3
  - Bonds: Offshore -15.0
  - Property -15.0
  - Currency: External value of BWP against USD, EUR, GBP -20.6
    - Note: Negative value means a depreciation of the BWP.
- Additional sensitivity tests (single-factor shocks):
  - Currencies: Increase of the BWP external value by 10 percent.
  - Equity: Decline of domestic and foreign equity prices by 40 percent.
  - Counterparty risk: Default of the largest banking counterparty.
  - Results of these sensitivities were not added to the adverse scenario results.
- Modeling and capital standard assumptions:
  - Current valuation regime and PCT introduced in 2012 form the basis for the insurance stress test.
  - Most assets valued mark-to-market; exceptions for valuation of group undertakings.
  - Liabilities: no detailed prescribed valuation method beyond “best-estimate valuation assumptions, adjusted by compulsory margins and possibly also by discretionary margins.”
  - IFRS 17 is being implemented mainly by larger insurers; authorities have not adopted a strategy how to embed IFRS 17 in prudential regime—resulting in a parallel regime of prudential valuation and IFRS accounting.
  - PCT calculation differences: methods differ between life and short-term insurers; capital requirement for short-term insurers excludes an operational risk module; explicit requirement for asset concentration risk (including counterparty concentration) is missing for both sectors.
  - Reporting limitation: mid-year PCT composition was assumed to be same as end-2021 for stress test (reference date 30 June 2022).
- Stress test mechanics:
  - Haircuts aligned with adverse scenario applied to asset classes; for life insurers interest rate shock applied to liabilities (separate for annuity and non-annuity business); short-term insurers do not discount future cash flows so liability valuation is impacted differently.
  - Unit-linked liabilities decline mirrored market value loss of underlying assets.
  - Available capital proxied by difference between post-stress assets and post-stress liabilities.
  - PCT partly recalculated after stress: market/investment risk component proportionally adjusted for reductions in exposures due to stress materialization.
- Data used:
  - Annual and quarterly prudential reporting templates for balance sheets, composition of PCT, and sensitivity analyses for bank counterparty exposures.

### Key Findings and Statistics (selected)
- NBFI assets at end-2021: BWP 153bn.
- Retirement funds assets: BWP 110bn.
- Retirement funds contributions: BWP 5.6bn in 2021 (BWP 3.8bn in 2016).
- Retirement funds membership (Dec 2021): 293,015 active members; 51,806 deferred members; 13,187 pensioners.
- Gross written life premiums (2021): BWP 4.4bn; growth since 2016: 6.9 percent per year.
- Gross written short-term premiums (2021): BWP 1.6bn; growth since 2016: 4.6 percent per year.
- Reinsurance premiums (2021): BWP 466m (BWP 175m five years earlier).
- Investment income of retirement funds (2021): BWP 17bn; investment return 15.1 percent.
- Stress test sample coverage:
  - Life insurers: 93 percent of life sector assets (by coverage in sample).
  - Short-term insurers: 77 percent based on gross written premiums.
  - Retirement funds: four funds cover 81 percent of market by assets; aggregated participating funds’ balance sheet assets amount to BWP 100bn.
- Stress shock magnitudes (selected):
  - Domestic equity: -25.1 percent
  - Offshore equity (advanced economies): -28.7 percent
  - Offshore equity (emerging markets): -25.1 percent
  - Domestic bonds short-term (~1-3 years): -9.2 percent
  - Domestic bonds long-term (~5-10 years): -7.3 percent
  - Offshore bonds: -15.0 percent
  - Property: -15.0 percent
  - Pula depreciation vs USD/EUR/GBP: -20.6 percent

### Policy-relevant Observations and Implications
- Concentration risks:
  - High concentration in both the life insurance sector (three largest life firms >90 percent market share) and the retirement funds sector (four largest funds account for 81 percent of assets) implies potential systemic implications from distress at large entities.
- Interconnectedness and cross-border exposures:
  - Many insurers are subsidiaries of foreign groups, with significant intragroup interlinkages; reinsurers are foreign-owned; large retirement funds hold substantial foreign assets—raising cross-border contagion and currency risk considerations.
- Market and currency risk exposures:
  - Retirement funds’ large foreign asset allocations and almost entirely unhedged currency exposures expose members to currency volatility.
  - Life insurers’ large holdings in bonds and unit-linked assets and retirement funds’ growth-oriented, equity-heavy portfolios create market sensitivity to equity and bond price shocks.
- Supervisory and reporting gaps:
  - Supervisory reporting limitations hinder granular analysis: lack of detailed domestic vs foreign breakdowns, limited credit risk detail (e.g., credit ratings), and limited asset-liability maturity/duration data.
  - PCT composition reporting only annual; limited transparency on more risk-based solvency measures.
- Solvency regime design considerations:
  - Differences in capital requirement methodologies between life and short-term insurers, absence of operational risk module for short-term insurers, and lack of explicit asset concentration/counterparty concentration requirement suggest areas for prudential review and enhancement.
- Accounting and prudential alignment:
  - Parallel implementation of IFRS 17 by some larger insurers and existing prudential valuation regime may require a defined strategy to align accounting and prudential frameworks and to consider implications for product design, pricing, asset-liability management, and solvency assessment.

*Source: IMF staff calculations and NBFIRA data as presented in the IMF technical note (section 4).*

### 5.5 for life insurers, and 6.1–6.5 for short-term insurers). These reporting templates   lack a

### Stress Test Results – Insurance and Retirement Funds

### A. Results of the Solvency Stress Test (Insurance)
- Context: heightened market volatility during February/March 2022 following the war in Ukraine; asset values and available capital of some insurers as of end-2022 are expected to be lower than in the prior year.
- Valuation impacts (adverse scenario):
  - Median life insurer: assets decrease by 7 percent.
  - Some life insurers: liabilities—related to annuity business and unit-linked insurance—decline at a similar amount as assets.
  - Median short-term insurer: asset values decrease by 1 percent.
  - Excess of assets over liabilities:
    - Median life insurer: shrinks by 12 percent.
    - Median short-term firm: shrinks by 2 percent.
  - Heterogeneity: life sector shows large variations across firms due to differences in business models, asset allocation, and asset and liability durations.
- Solvency levels:
  - Median life insurer solvency ratio: declines by 12 percentage points, from 325 percent to 313 percent.
  - 25th and 75th percentiles (life): decline each by around 40 percentage points.
  - Median short-term firm solvency ratio: decreases by 2 percentage points to 155 percent.
- Drivers of capital decline:
  - For life insurers: shocks to domestic bond and equity prices account for 89 percent of the net overall change in capital.
  - BWP depreciation is a major compensating factor in the life sector (increases value of foreign assets).
  - For short-term insurers: 78 percent of the capital reduction stems from the shock on short-term domestic bonds; remainder from property exposures.
- Reporting and data limitations affecting the exercise:
  - Reporting templates lack cumulative exposure toward a counterparty across different classes; counterparty exposures only included if listed among largest individual exposures.
  - Supervisory reporting data granularity does not allow very granular modeling of adverse scenario impacts on assets and liabilities of an insurer or retirement fund.
  - Derivative hedging cannot be modeled due to missing reporting data on such instruments.
  - Potential reactive management actions (de-risking investment allocation, changes to reinsurance program, dividend cuts) were not modeled.

### B. Sensitivity Analyses (Insurance)
- Equity shock:
  - General equity price shock of -40 percent (domestic and foreign) would reduce the excess of assets over liabilities for the median life insurer by 13 percent; for one firm the reduction would be well more than 50 percent.
  - Among short-term insurers, impact of this equity shock limited to a small number of firms.
- FX shock:
  - An appreciation of the Pula (BWP) by 10 percent would have a limited impact: for the median life insurer, excess of assets over liabilities would decline by 3 percent.
- Concentrated bank exposures:
  - Largest deposits with a single banking counterparty amount to more than one third and up to half of shareholder capital for most insurers in the sample.
  - For one firm, the largest (deposit) exposure even exceeds their capital.
  - Reporting templates report largest exposures in each of five asset classes (deposits, bonds, property, shares, other loans and securities); smaller exposures across classes may be sizable but remain undetected.
  - Further analysis should consider potential loss given default and second-round effects (interbank market, sovereign bond spreads) if a large domestic bank defaults.

### C. Other Risks (Insurance)
- Reinsurance and supervision:
  - Reinsurance business is channeled primarily through domestic reinsurers that are subsidiaries of foreign groups.
  - Need for a robust supervisory approach with strong cooperation and data sharing with home authorities, joint on-site inspections, and regular exchange on supervisory findings.
  - NBFIRA should ensure staff acquires and maintains specialized knowledge about reinsurance business.
- Climate risk:
  - Climate risks borne by the Batswana insurance sector are currently small because many perils are not insured.
  - Botswana faces higher temperatures and droughts; critical to evaluate insurability of such risks and how protection gaps could be reduced.

### Stress Test Results – Retirement Funds (Future Pension Values)
- System design:
  - Retirement funds in Botswana are defined contribution schemes; investment risks borne by scheme members.
- Representative member profiles used for simulation:
  - Member A: 10 years prior to retirement; BWP 1,000,000 accumulated pension savings; BWP 20,000 annual contributions.
  - Member B: 30 years prior to retirement; BWP 200,000 accumulated pension savings; BWP 12,000 annual contributions.
- Methodology highlights:
  - Analysis focuses on difference between baseline and adverse scenario, not exact future pension value.
  - Adverse scenario modified: market risks unfold over the first two years of projection horizon; afterwards investment performance same as baseline.
  - Baseline after t+2y: future investment returns projected via bootstrapping historic returns by asset class based on each fund’s asset allocation as of the reference date.
  - Assumptions: funds maintain asset allocation over full horizon and re-balance annually; contributions increase at 4.5 percent per year.
- Impacts:
  - Representative member with 10 years to retirement: future pension values decline by around 17 percent compared to baseline.
  - Representative member with 30 years to retirement: future pension values decline by 4 percent.
  - Policy implication: need for timely switch of pre-retirement members to more conservative asset allocation to lock-in accrued pension values.
- Liquidity implications of the new Retirement Fund Act (2022):
  - The Act provides more lump-sum withdrawal options upon retirement and prior to that (e.g., to repay mortgages, pay medical bills), transferring liquidity from pension funds to households.
  - Payments upon retirement currently account for around 80 percent of all payouts; of these around 40 percent are paid as a lump-sum.
  - The other 20 percent of payouts are for other reasons (payments upon dismissals or resignations most important in this category).
- Vulnerable funds under the new Act (likely to be particularly affected):
  - Smaller funds;
  - Funds serving only one employer or active in only one economic sector (especially highly cyclical sectors);
  - Funds serving the lower-income sector;
  - Funds with a high ratio of deferred to active members;
  - Funds where benefits are of a similar size as contributions or already exceeding contributions.
- Additional Act implications:
  - The Act reduces sequentially the cap on foreign assets, raising questions about domestic capital markets’ capacity to absorb repatriated investments.
  - Effects will be seen in other financial sectors (e.g., life insurers’ annuity business if pension funds transfer less accrued pension interest to annuity providers upon retirement).
- Supervision recommendation: NBFIRA should increase intensity of supervision for vulnerable funds and require liquidity buffers where deemed necessary.

### Stress Test Scenarios and Key Parameters (from STeM)
- Reference date: 30 June 2022.
- Sample size and market shares:
  - 4 life insurers: ABSA Life, Botswana Life, Hollard Life, Metropolitan Life.
  - 4 short-term insurers: Botswana Insurance, Hollard, Old Mutual, Sesiro.
  - 4 retirement funds: Alexander Forbes Retirement Fund, Bank of Botswana Pension Fund, Botswana Public Officers Pension Fund, Debswana Pension Fund.
  - Market share (by balance sheet assets or gross written premiums): Life insurance: 93 percent; Short-term insurance: 79 percent; Retirement funds: 81 percent.
- Scenario (Global market and emerging markets stress):
  - Share prices: -25.1 percent (domestic and offshore emerging markets), -28.7 percent (offshore advanced economies).
  - Bond prices: -9.2 percent (domestic short-term), -7.3 percent (domestic long-term), -15.0 percent (offshore).
  - Property prices: -15 percent.
  - Currency: +20.6 percent (BWP appreciation).
- Sensitivity analyses included:
  - Market shocks: increase in domestic interest rates, and BWP depreciation.
  - Default of largest bank counterparty.
- Risk factors assessed: market risks (bond, share, property prices, FX rates); credit risks (default of largest bank counterparty); summation of risks.
- Reporting and standards: statutory returns; National GAAP.
- Reporting format for results: impact on valuation of assets; impact on available capital and solvency position (insurance); impact on future pension values (retirement funds); dispersion across companies; contribution of individual shocks.

### Key Financial Soundness Indicators (selected figures)
- Solvency and liquidity (stress test participants only):
  - Median life insurer solvency ratio (pre-stress to post-stress): 325 percent to 313 percent.
  - Median short-term insurer solvency ratio (post-stress): 155 percent.
- Insurance sector FSIs (selected series, values as reported for 2021 unless otherwise noted):
  - Assets / liabilities – life: 16.5 (2021).
  - Assets / liabilities – short-term: 30.0 (2021).
  - Solvency ratio – life: 669.4 (2021).
  - Solvency ratio – short-term: 170.6 (2021).
  - Investment income / total assets – life: 5.0 (2021).
  - Cash and deposits / total assets – short-term: 49.8 (2021).
  - Risk retention ratio – life (net premium / gross premium): 96.9 (2021).
  - Risk retention ratio – short-term: 55.1 (2021).
- Retirement funds FSIs (selected series, values as reported for 2021 unless otherwise noted):
  - Investment income / total investments (average): 15.1 (2021).
  - Stocks / total investments: 67.9 (2021).
  - Bonds / total investments: 13.7 (2021).
  - Foreign assets / total investments: 66.5 (2021).
  - Cash and deposits / total investments: 6.3 (2021).
  - Contributions / benefits: 153.5 (2021).
- Notes: All FSI data reported as in the source tables; stress test participants only indicated where specified.

### Summary and Policy Recommendations
- Data and analytics:
  - Review supervisory reporting templates (also to accommodate for IFRS 17).
  - Implement more robust processes to ensure complete and consistent reporting.
  - Implement data systems allowing for horizontal and financial stability analysis.
- Stress testing program:
  - Conduct regular stress tests of large insurers and retirement funds.
  - Analyze risks from concentrated bank exposures; coordinate such analysis with the Bank of Botswana and the Financial Stability Council.
  - Approach: initially top-down stress tests by NBFIRA based on enhanced supervisory reporting data; periodically use bottom-up stress tests to calibrate top-down methodology and proliferate advanced stress testing methods among insurers and retirement funds.
- Capacity building:
  - NBFIRA should build in-house actuarial expertise for regular and deep-dive analyses of underwriting and asset-liability risks (microprudential and macroprudential), policy advice on IFRS 17 implementation, and valuation and solvency framework reviews.
  - While external experts may be used, senior in-house staff capable of overseeing outsourced work is important.

*Source: IMF staff calculations based on NBFIRA data (content as presented in the provided chapter).*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2024/english/1bwaea2024003.pdf_