## _cr16218

## Source details

**Canonical URL:** [_cr16218](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16218.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16218.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16218.pdf.json)

---

### A. Linkages between South Africa and China
- China now absorbs 10 percent of South African exports compared to around 2½ percent in the mid-2000s.
- China accounts for around 17 percent of global output in PPP terms compared to 16 percent for the US.
- In terms of imports, China accounted for approximately 16 percent of the world total compared to 14 percent for the U.S.
- More than 60 percent of the world’s traded iron ore—South Africa’s main mineral export—is absorbed by China.
- Commodity exports account for 34 percent of total goods exports (51 percent when manufactured commodities are included).
- Oil accounts for around 16 percent of total goods imports for South Africa.
- Capital flows and financial linkages with China:
  - Direct investment and portfolio flows from China are increasing but remain modest relative to capital flows from the UK and the US.
  - JSE equity index beta with China’s HSCEI index (after controlling for movements in the VIX and S&P500) has increased to 0.2 from less than 0.1 in 2005.
  - Share of mining companies in the JSE declined to 13 percent in April 2016 from 42 percent at end-2011.
  - FX beta with the Chinese renminbi (after controlling for a U.S. dollar index and the VIX) has increased to 1.3 since August 2015 from 0.9 in a longer sample starting in January 2013.
- Interpretation:
  - Increasing trade and financial linkages, and China’s role in global commodity markets, magnify China’s impact on South Africa through prices and financial market spillovers.

### B. Assessing the Impact of Lower Growth in China on South Africa
- VAR evidence:
  - A simple VAR on quarterly real GDP growth (U.S., EU, China, South Africa) from 2000 onwards suggests a 1 percentage point decline in China’s real GDP growth would lower South Africa’s growth by 0.3 percentage points after one quarter.
  - Restricting the sample to 2010-15, the impact of a 1 percentage point decline in China’s real GDP growth rises to 1 percentage point on South Africa.
  - Full-sample (2000-15) and latest-sample (2010-15) spillovers (impact of a 1 percentage point negative growth shock on South Africa's growth after 1 quarter):
    - Latest sample (2010-15): US -0.3; EU -0.2; China -1.0.
    - Full sample (2000-15): US -0.4; EU -0.7; China -0.3.
- Transmission channels (Bayoumi and Swiston 2008 approach; auxiliary VARs with exogenous channel variables):
  - The decline in South Africa’s export commodity prices resulting from a shock to China’s growth has a large and persistent impact on output growth.
  - Financial spillovers (proxied by U.S. financial conditions) are also important, indicating global confidence effects.
  - Spillovers to trade volumes are small and positive, suggesting exchange rate depreciation’s positive effect on exports can outweigh lower foreign demand.
  - Declining import commodity prices (mainly oil) provide a partial offset.

### C. Domestic Propagation of Commodity Price Shocks
- Input-output linkages:
  - The mining multiplier from the IO table: a R1 million increase in mining output raises output in the overall economy by R1.8 million (indirect effect R0.8 million).
  - Upstream linkage strongest: transportation.
  - Important downstream linkages: metal production; petroleum, chemical, and mineral production; and other manufacturing.
- Sectoral structural VAR (identified using IO multipliers):
  - Model components:
    - South Africa-specific commodity export price index (Ptex).
    - Growth contributions of real value-added in mining (VAtmining) and non-mining (VAtnon-mining).
    - Exogenous variables: commodity import price index, growth in main trading partners, and global financial conditions.
    - Identification uses IO-derived parameter restrictions (A matrix, coefficients aij, parameters bi).
  - Main result:
    - A 10 percent fall in the growth rate of export commodity prices would reduce real GDP growth by nearly 0.2 percentage points after two years.
    - Direct impact via the mining sector is relatively small and short-lived; indirect impacts on upstream and downstream industries are more prolonged and larger than implied by the IO table.
  - Caveat:
    - Structural bottlenecks (e.g., electricity shortages, protracted strikes) coincided with declining commodity prices and could affect estimates.
- Channels transmitting commodity price shocks to the broader economy:
  - Main channels: declines in employment and corporate profitability.
  - Decline in net wealth in household and corporate sectors also contributes.
  - REER does not play a substantial role (slow responsiveness of export volumes to REER depreciation).
- Quantitative example:
  - Quarterly growth impact of a 10 percent fall in commodity export prices (2000Q1-2015Q2 sample):
    - Aggregate impact nearly -0.2 percentage points on real GDP growth after two years.
  - Sectoral contributions:
    - Employment and corporate profits drive most of the overall effect on mining and non-mining growth; REER and net wealth play smaller roles in the model results.

### D. Conclusion (China shock and commodity-price transmission)
- China’s growth slowdown has large spillover effects on South Africa, transmitted mainly through export commodity prices and global financial conditions, and amplified by sectoral interlinkages.
- Rising trade and financial linkages with China, together with China’s large role in global commodity markets, have increased the impact of a Chinese growth slowdown so that it now exceeds that of a growth slowdown in the U.S. and the E.U. (in the 2010-15 sample).
- Domestically, falls in commodity export prices are amplified through linkages between mining and non-mining sectors and are transmitted to the economy via large falls in mining sector employment, corporate profitability, and wealth effects.

### Paper overview: MACRO-FINANCIAL LINKAGES: CAPITAL FLOWS, SOVEREIGN RATINGS, AND THE FINANCIAL SECTOR NEXUS
- Prepared by Hui Miao, Pablo Morra, and Yi Wu.
- Focus: key macro-financial linkages and downside scenarios for South Africa, emphasizing risks from high reliance on external finance, banks’ larger role in intermediating capital flows, and interactions among low growth, rising interest rates, fiscal risks, sovereign ratings, and the financial sector.

### Key summary findings and quantitative facts (external position and capital flows)
- South Africa’s current account deficit averaged 4.1 percent of GDP in 2005-14.
- Current account deficit declined to 4.3 percent of GDP in 2015 from a peak of 5.7 percent of GDP in 2013.
- External debt stood at 39.4 percent of GDP as of December 2015, up from 25.8 percent of GDP in 2008.
- Annual gross external financing requirements stood at about 19 percent of GDP as of September 2015.
- About half of the external debt is denominated in local currency.
- Net international investment position shifted from -8 percent of GDP at end-2014 to 18 percent of GDP as of end-2015.
- As of end-2015:
  - foreign assets = 157 percent of GDP;
  - external liabilities = 139 percent of GDP.
- About 40 percent of total external assets are foreign direct investment; about one-half of liabilities are portfolio investments by nonresidents.
- Net foreign direct investment flows turned negative, averaging -0.8 percent of GDP in 2014–15.
- Other investment flows and unrecorded transactions rose significantly since 2013, financing the current account deficit.
  - Other investment flows increased from 1.5 percent of GDP in 2013 to 3 percent of GDP in 2015.
- Banks account for 23 percent of the total stock of external debt and about 50 percent of the external debt coming due within one year.
- Projection assumption: no unrecorded transactions (customary in IMF projections).
- South Africa’s external portfolio liabilities = 68 percent of GDP.
- Foreign investors’ holdings:
  - bonds = 20 percent of GDP;
  - equities = 48 percent of GDP.
- About 90 percent of loans are at floating rates, linked to the policy rate.
- Since December 2013, the policy repo rate has risen by 200 bps.
- One-year negotiable certificate of deposits (NCD) rate increased by 274bps since December 2013.
- The gap between the NCD rate and the policy rate stands at about 170 bps.
- Required Liquidity Coverage Ratio (LCR) will continue to rise by 10 percentage points each year to reach 100 percent in January 2019 (including up to 40 percent from the South Africa Reserve Bank’s (SARB) committed liquidity facility).
- The Net Stable Funding Ratio (NSFR) is to be fully implemented in 2018.
- Bank short-term wholesale funding currently accounts for about 60 percent of total bank liabilities.
- About 60 percent of NBFIs’ assets are invested in stocks.
- Many international firms listed on the Johannesburg Stock Exchange have foreign earnings accounting for about half of total earnings of all listed companies.
- About half of foreign earnings are from EMs.
- Major European banks account for 76 percent of total external bank lending to South Africa (BIS data).

### A. The Role of Capital Flows in the South African Economy
- South Africa is highly reliant on capital flows: large and persistent current account deficits and rising external debt.
- Exchange rate depreciation in 2015 facilitated a shift to a positive net international investment position by lowering the U.S. dollar value of rand-denominated liabilities.
- External financing mix shifted away from portfolio and direct investment toward other investment flows and unrecorded transactions since 2013.
- Other investment flows comprise primarily bank loans and deposits; remainder includes financial derivatives, trade credit, and other net receivables.
- A significant share of other investment flows are short-term tenor (less than one year) and denominated in FX, posing rollover risks and higher hedging costs in a shock.
- While banks’ FX position is small and within regulatory limits, maturity mismatches exist because some FX assets are corporate loans to fund acquisitions abroad.
- Sustainability of bank-intermediated flows is uncertain; the source and economic rationale for unrecorded transactions are unknown.

### B. Capital Flows and Macro-Financial Linkages
- Outlook:
  - Capital flows to emerging markets expected to be scarcer, more volatile, and costlier (per April 2016 Global Financial Stability Report).
  - For South Africa, expected continued negative net FDI flows and lower, more volatile portfolio flows as U.S. monetary policy normalizes and advanced economies’ policies diverge.
  - Fund staff project other investment flows to be the main source of external financing in coming years, but sustainability is uncertain as major European banks retrench.
- Potential foreign investor pullout risks and mitigating features:
  - Large foreign participation in domestic bond and equity markets could cause sizable outflows if nonresidents liquidate holdings.
  - Mitigants: floating exchange rate, deep financial markets, and about half of external debt in local currency.
  - Currency denomination varies by sector:
    - bulk of government external debt in local currency;
    - majority of SOEs’, banks’, and corporates’ external debt in FX.
  - SOEs required to hedge FX risk, but hedging may be rolling and imperfect; hedging costs and collateral requirements could rise with downgrades.
  - Nonresident liquidation would likely incur losses on asset prices, which could reduce outflow volumes.
- Transmission channels of capital outflows:
  - Depreciation of the rand, higher interest cost of debt, decline in local asset prices, negative wealth effects.
  - Depreciation may have limited export impact due to structural constraints and policy uncertainty.
  - Large outflow likely to force sharp import compression and sudden current account adjustment, reducing growth.
  - International firms listed on JSE and their foreign earnings (about half of total) provide some buffer against depreciation for stock prices.
  - Some importers (e.g., retailers) would suffer from rand weakness.
- Financial sector and fiscal feedback loops:
  - Rand depreciation would raise inflation, likely prompting SARB to hike rates; about 90 percent of loans are at floating rates, amplifying growth impact.
  - Initial boost to banks’ profitability from higher interest rates could reverse as funding and credit costs rise, leading to higher NPLs and lower profitability.
  - Decline in banks’ profitability would adversely affect fiscal revenue given heavy reliance on the financial sector for income tax.
  - Government likely able to secure domestic financing, but domestic portfolio rebalancing may depress equity prices (about 60 percent of NBFIs’ assets in stocks), increase interest rates, and reduce private lending (crowding-out), further depressing growth.
- Banking sector funding and regulatory dynamics:
  - Since December 2013, regulatory changes and higher policy rates have increased banks’ funding costs.
  - Rising LCR requirements and NSFR implementation will push banks toward more term funding and retail deposits; outstanding bank bond issuance projected to rise, increasing average funding costs.
  - SARB’s national discretion to apply a 35 percent available stable funding factor to short-term funding from financial corporate customers reduces NSFR challenge relative to prior assessments.
  - SARB has increased size of the money market liquidity shortage to enhance monetary policy transmission.
  - Capital outflows would exacerbate expected rises in banks’ funding costs.

### Banks’ asset quality, stress tests, and macrofeedbacks
- Household-sector asset quality and risks:
  - Households’ debt-to-disposable income ratio is 78 percent.
  - NPLs on the unsecured credit book have increased to 9.0 percent as of end-2015.
  - Staff analysis (IMF, 2014) suggests that a severely adverse scenario would:
    - lead to a significant increase in households’ probability of default, and
    - reduce the Common Equity Tier 1 (CET1) ratio by 2.5 percentage points.
- Corporate-sector credit stress:
  - Pockets of credit stress have emerged in construction, mining, steel, and agriculture sectors; credit to these sectors represents only 6 percent of bank loans.
  - Declining corporate sector profitability has weakened the interest coverage ratio (ICR) of South African corporates; the ICR stands at 3 in 2015Q1-Q3.
    - Mining and quarrying industry recorded an ICR below the benchmark of 2.
  - The expected default frequency (EDF) of South African firms has deteriorated.
  - IMF (Chow, 2015) stress analysis: a stress scenario similar to the global financial crisis increases NPLs only by 1-2 percentage points.
- Stress tests, capitalization, and resilience:
  - Recent SARB stress tests indicate banks are adequately capitalized to withstand significant credit losses under stress scenarios.
  - Even in the most adverse scenarios (protracted recession or excessive financial market volatility with capital flow reversal and significant depreciation), banks’ average CET1 capital adequacy ratio remains above the regulatory requirement.
  - Banks are expected to maintain access to international capital markets, though at higher costs, given their strong balance sheets.
- Financial-cycle, credit growth, and macrofeedbacks:
  - Bank credit growth to households moderated from a peak of 24.8 percent y/y in March 2008 to 2.3 percent in April 2016.
  - Credit growth to corporates was 12.1 percent y/y as of April 2016.
  - The financial cycle has rebounded from its 2011 trough but appears to have peaked at a low level.
  - The SARB determined no need to activate countercyclical capital buffers as the credit gap remains negative.
  - South Africa’s financial cycle has become increasingly correlated with the U.S. financial cycle (IMF, 2016).
- Potential macrofeedback from deleveraging:
  - In a severely adverse scenario, banks likely deleverage, reducing lending to the private sector (private sector accounts for 67 of their total assets).
  - Fund staff estimate that a 2½ percentage point decline in credit growth relative to the baseline (from 7 percent to 4½ percent) could reduce real GDP growth by 0.5–1 percentage points.
  - Cross-country distribution context (from table):
    - At credit growth 7%: 10th percentile GDP growth 0.5; 25th 2.6; 75th 5.9; 90th 7.5.
    - At credit growth 4.5%: 10th percentile GDP growth -0.5; 25th 1.6; 75th 5.4; 90th 6.8.
    - Growth difference by percentiles: 10th 1.0; 25th 1.0; 75th 0.5; 90th 0.7.
- Fiscal exposure and financial sector importance:
  - Financial sector assets amount to 300 percent of GDP.
  - Finance, real estate, and business services account for 20 percent of GDP and 22 percent of formal nonagricultural employment.
  - Financial sector accounts for about half of personal income tax and one third of corporate income tax.
  - A shock to the financial sector could have a large impact on fiscal revenue, forcing the government to restrain spending with adverse second-round effects on growth.

### Sovereign rating downgrade scenarios and likely market impacts
- Current sovereign ratings context:
  - South Africa’s sovereign credit rating is one to two notches above investment grade (IG) for FX debt, and two to three notches above IG for LC debt.
  - Two rating agencies (Moody’s and Standard & Poor’s) have a negative outlook.
- FX sovereign debt downgrade:
  - Market pricing and spreads:
    - South Africa’s 5-year CDS spreads trade in line with BB+ rated EMs (e.g., Turkey, Brazil, Russia).
    - 5-year CDS spread increased by about 80 bps since November 2015 to 320 bps at end-May 2016.
    - Past episodes: CDS spreads typically increase by about 100 bps before downgrade and an additional 10-20 bps after it materializes.
    - JP Morgan estimates forced sales could amount to about US$2 billion based on survey of institutional investors.
  - Impacts on SOEs and banks:
    - SOEs: about 40 percent of their total debt is denominated in FX; funding costs could rise significantly.
    - Banks: about 10 percent of liabilities are in foreign currency; US$13.5 billion (4 percent of GDP) of these foreign-currency liabilities will come due within one year.
    - Syndicated loan spreads over Libor have largely increased in line with sovereign CDS spreads; syndicated loans typically 3-year floating with fees and spreads tied to sovereign rating.
  - Regulatory capital effects for international banks:
    - For local banks, risk weight for holding local currency government bond of South Africa is zero regardless of sovereign ratings.
    - For South African bank branches of international banks, home regulators assign high risk weighting for bonds below investment grade.
      - Example: for BBB- rated government debt, foreign bank branches assign risk weight of 50 percent; if downgraded to BB+, risk weight doubles to 100 percent.
  - Investor base and capital-flow volatility:
    - Institutional investors tend to reduce their investment on average by 16 percent of assets under management and only return slowly following a downgrade to speculative grade.
    - Since 1980, 19 cases of sovereign FX downgrades from IG to speculative (by at least 2 rating agencies):
      - 6 sovereigns regained IG after an average of 8 years.
      - The remaining 13 stayed below IG, averaging 9 years since loss of IG status.
- LC sovereign debt downgrade:
  - LC ratings and co-movement:
    - Moody’s does not differentiate between LC and FX credit risk and rates both at the same level.
    - Standard & Poor’s rates South Africa’s LC sovereign debt two notches above the FX sovereign debt.
    - Fitch’s LC rating is one notch above the FX rating.
    - If FX rating is downgraded, LC ratings would likely be downgraded too, but may still remain IG.
  - Forced sales risk and index/mandate sensitivities (analysis of two benchmark LC sovereign bonds, 5- and 7-year):
    - Foreign holdings account for more than half of the outstanding amount; these two bonds account for 13 percent of total foreign holding of LC sovereign fixed-rate bonds.
    - More than half of nonresident investors of these bonds follow JP Morgan’s Emerging Market Bond Index (GBI-EM), which does not require IG credit rating.
    - 21 percent of nonresident investors are IG rating-sensitive due to index restrictions or mandates.
    - Staff estimate: nonresidents’ total forced sales could amount to about 2½ percent of GDP in the event of LC sovereign rating downgrade to speculative grade.
    - Other investors might sell preemptively; the above estimates may be a lower bound.
    - Historically, equity flows have displayed a high correlation with bond flows in times of stress.

### NBFIs’ capacity to absorb government bond sales and associated risks
- NBFIs held total assets of 210 percent of GDP in 2015, of which less than 10 percent are local government bonds.
- NBFIs could provide a backstop to foreign bond sales and absorb new issuance, but portfolio rebalancing implications:
  - NBFIs’ assets comprise mainly domestic equities and banks’ NCDs, so shifting to government bonds could:
    - lead to asset price declines;
    - reduce funding for banks (via lower demand for banks’ NCDs);
    - cause liquidity tightening and a significant rise in funding costs for banks, as illustrated by jumps in short-term rates during market turmoil episodes in August and December of 2015.
- Mechanisms for a liquidity squeeze (including in a “closed rand system”):
  - Banks may need to provide funding for NBFIs switching from NCDs to government bonds by buying back NCDs from NBFIs, requiring liquidity obtained by selling other assets.
  - Market demand for short-term liquidity rises significantly in periods of high risk aversion; banks may need to raise term funding to meet LCR requirements under unfavorable market conditions.
  - SARB can intervene to provide liquidity, but banks may be limited by the availability of eligible collateral.
  - In a more severe capital outflow shock, banks could face medium-term liquidity challenges if NBFIs reduce bank exposure and corporates reduce bank deposits when profitability suffers.
  - Potential bank responses include retrenching credit to restore LCR ratios and passing on higher funding costs to customers, leading to tighter financial conditions and lower growth.
- Constraints on NBFIs’ stabilizing role:
  - NBFIs’ relatively low holdings of government bonds could reflect a preference for higher-yielding instruments; there is no minimum requirement for NBFIs to hold government bonds in South Africa.
  - NBFIs’ funding to banks can be relationship-driven because insurers and CIS tend to hold bank paper issued by their connected banks.
  - During market turmoil in August and December 2015, NBFIs did not substantially increase purchases of government bonds despite the large jump in bond yields.
  - Mandatory capital repatriation after breaching prudential limits on overseas asset holdings is subject to delay:
    - NBFIs’ overall foreign asset holdings averaged 15 percent, below the regulatory ceiling, though some large private nonfinancial institutions are reported to be at the limit.
    - When the limit is breached, NBFIs can no longer increase offshore allocation but have one year to repatriate funds, implying delayed stabilizing effects from repatriation.
  - Therefore, it may take significant bond price adjustments and rand depreciation for NBFIs to substantially increase allocations to government bonds via portfolio rebalancing.
- Asset allocation and offshore exposure (2015):
  - NBFIs’ allocation of total assets:
    - Equities: 57%
    - Fixed rate bonds: 28%
    - Cash and deposit: 7%
    - Loans: 5%
    - Other: 3%
  - Share of Foreign Assets in NBFIs’ Portfolios and Prudential Limits (Assets Under Management, Percent of offshore assets in total assets, Offshore limits, percent of total assets):
    - Collective Investment Schemes: 2 (ZAR trn); 20 (percent); 30 (percent)
    - Pension Funds (FSB-registered): 2.5 (ZAR trn); 19 (percent); 25 (percent)
    - Life Insurance Companies: 1.8 (ZAR trn); 14 (percent); 30 (percent)
  - Note: Assets under management cannot be added up due to overlapping ownership.
  - Footnote: An additional 5 percent of total assets is allowed for investment in the rest of Africa.

*Prepared by Manabu Nose, Magnus Saxegaard, and Jose Torres. South Africa, June 16, 2016.*

### References _______________________________________________________________________________ 10

### THE IMPACT OF CHINA’S GROWTH SLOWDOWN AND LOWER COMMODITY PRICES ON SOUTH AFRICA

### A. Linkages between South Africa and China
- China now absorbs 10 percent of South African exports compared to around 2½ percent in the mid-2000s.
- China accounts for around 17 percent of global output in PPP terms compared to 16 percent for the US.
- In terms of imports, China accounted for approximately 16 percent of the world total compared to 14 percent for the U.S.
- More than 60 percent of the world’s traded iron ore—South Africa’s main mineral export—is absorbed by China.
- Commodity exports account for 34 percent of total goods exports (51 percent when manufactured commodities are included).
- Oil accounts for around 16 percent of total goods imports for South Africa.
- Capital flows from China to South Africa:
  - Direct investment and portfolio flows from China are increasing but remain modest relative to capital flows from the UK and the US.
  - JSE equity index beta with China’s HSCEI index (after controlling for movements in the VIX and S&P500) has increased to 0.2 from less than 0.1 in 2005.
  - Share of mining companies in the JSE declined to 13 percent in April 2016 from 42 percent at end-2011.
  - FX beta with the Chinese renminbi (after controlling for a U.S. dollar index and the VIX) has increased to 1.3 since August 2015 from 0.9 in a longer sample starting in January 2013.
- Interpretation: Increasing trade and financial linkages, and China’s role in global commodity markets, magnify China’s impact on South Africa through prices and financial market spillovers.

### B. Assessing the Impact of Lower Growth in China on South Africa
- VAR evidence:
  - A simple VAR on quarterly real GDP growth (U.S., EU, China, South Africa) from 2000 onwards suggests a 1 percentage point decline in China’s real GDP growth would lower South Africa’s growth by 0.3 percentage points after one quarter.
  - Restricting the sample to 2010-15, the impact of a 1 percentage point decline in China’s real GDP growth rises to 1 percentage point on South Africa.
  - Full-sample (2000-15) and latest-sample (2010-15) spillovers table (impact of a 1 percentage point negative growth shock on South Africa's growth after 1 quarter):
    - Latest sample (2010-15): US -0.3; EU -0.2; China -1.0.
    - Full sample (2000-15): US -0.4; EU -0.7; China -0.3.
- Transmission channels (Bayoumi and Swiston 2008 approach; auxiliary VARs with exogenous channel variables):
  - The decline in South Africa’s export commodity prices resulting from a shock to China’s growth has a large and persistent impact on output growth.
  - Financial spillovers (proxied by U.S. financial conditions) are also important, indicating global confidence effects.
  - Spillovers to trade volumes are small and positive, suggesting exchange rate depreciation’s positive effect on exports can outweigh lower foreign demand.
  - Declining import commodity prices (mainly oil) provide a partial offset.

### C. Domestic Propagation of Commodity Price Shocks
- Input-output linkages:
  - The mining multiplier from the IO table: a R1 million increase in mining output raises output in the overall economy by R1.8 million (indirect effect R0.8 million).
  - Upstream linkage strongest: transportation.
  - Important downstream linkages: metal production; petroleum, chemical, and mineral production; and other manufacturing.
- Sectoral structural VAR (identified using IO multipliers):
  - Model: structural VAR with South Africa-specific commodity export price index (Ptex), growth contributions of real value-added in mining (VAtmining) and non-mining (VAtnon-mining); exogenous variables include commodity import price index, growth in main trading partners, and global financial conditions.
  - Identification uses IO-derived parameter restrictions (A matrix, coefficients aij, parameters bi).
  - Result: a 10 percent fall in the growth rate of export commodity prices would reduce real GDP growth by nearly 0.2 percentage points after two years.
  - Direct impact via mining sector is relatively small and short-lived; indirect impacts on upstream and downstream industries are more prolonged and larger than implied by the IO table.
  - Caveat: structural bottlenecks (e.g., electricity shortages, protracted strikes) coincided with declining commodity prices and could affect estimates.
- Channels transmitting commodity price shocks to the broader economy:
  - Main channels: declines in employment and corporate profitability.
  - Decline in net wealth in household and corporate sectors also contributes.
  - REER does not play a substantial role (slow responsiveness of export volumes to REER depreciation).
- Quantitative example:
  - Quarterly growth impact of a 10 percent fall in commodity export prices (2000Q1-2015Q2 sample):
    - Direct impact, indirect impact, total growth impact shown in model results (aggregate impact nearly -0.2 percentage points on real GDP growth after two years).
  - Sectoral contributions to mining and non-mining growth from channels (REER, corporate profits, employment, net wealth) demonstrate employment and corporate profits drive most of the overall effect.

### D. Conclusion
- China’s growth slowdown has large spillover effects on South Africa, transmitted mainly through export commodity prices and global financial conditions, and amplified by sectoral interlinkages.
- Rising trade and financial linkages with China, together with China’s large role in global commodity markets, have increased the impact of a Chinese growth slowdown so that it now exceeds that of a growth slowdown in the U.S. and the E.U. (in the 2010-15 sample).
- Domestically, falls in commodity export prices are amplified through linkages between mining and non-mining sectors and are transmitted to the economy via large falls in mining sector employment, corporate profitability, and wealth effects.

*Prepared by Manabu Nose, Magnus Saxegaard, and Jose Torres. South Africa, June 16, 2016.*

### References

### _cr16218 - References

### References
- Bayoumi, Tamin, and Andrew Swiston (2008), “Spillovers Across NAFTA”, IMF Working Paper WP/08/3.
- Gruss, Bertrand (2014), “After the Boom: Commodity Prices and Economic Growth in Latin America and the Caribbean,” IMF Working Paper WP/14/154.
- IMF (2014), “The Unconventional Energy Boom in North America: Macroeconomic Implications and Challenges for Canada”, IMF Country Report No. 14/28.
- IMF (2014b), “South Africa’s Exports Performance: Any Role for Structural Factors?”, IMF Country Report No. 14/339.
- Pesaran, Hashem, and Yongcheol Shin (1998), “Generalized Impulse Response Analysis in Linear Multivariate Models”, Economic Letters Vol. 58, Issue 1.

### Paper overview: MACRO-FINANCIAL LINKAGES: CAPITAL FLOWS, SOVEREIGN RATINGS, AND THE FINANCIAL SECTOR NEXUS
- Prepared by Hui Miao, Pablo Morra, and Yi Wu.
- Focus: key macro-financial linkages and downside scenarios for South Africa, emphasizing risks from high reliance on external finance, banks’ larger role in intermediating capital flows, and interactions among low growth, rising interest rates, fiscal risks, sovereign ratings, and the financial sector.

### Key summary findings and quantitative facts
- South Africa’s current account deficit averaged 4.1 percent of GDP in 2005-14.
- Current account deficit declined to 4.3 percent of GDP in 2015 from a peak of 5.7 percent of GDP in 2013.
- External debt stood at 39.4 percent of GDP as of December 2015, up from 25.8 percent of GDP in 2008.
- Annual gross external financing requirements stood at about 19 percent of GDP as of September 2015.
- About half of the external debt is denominated in local currency.
- Net international investment position shifted from -8 percent of GDP at end-2014 to 18 percent of GDP as of end-2015.
- As of end-2015: foreign assets = 157 percent of GDP; external liabilities = 139 percent of GDP.
- About 40 percent of total external assets are foreign direct investment; about one-half of liabilities are portfolio investments by nonresidents.
- Net foreign direct investment flows turned negative, averaging -0.8 percent of GDP in 2014–15.
- Other investment flows and unrecorded transactions rose significantly since 2013, financing the current account deficit.
- Other investment flows increased from 1.5 percent of GDP in 2013 to 3 percent of GDP in 2015.
- Banks account for 23 percent of the total stock of external debt and about 50 percent of the external debt coming due within one year.
- Projection assumption: no unrecorded transactions (customary in IMF projections).
- South Africa’s external portfolio liabilities = 68 percent of GDP.
- Foreign investors’ holdings: bonds = 20 percent of GDP; equities = 48 percent of GDP.
- About half of the total external debt is denominated in local currency.
- About 90 percent of loans are at floating rates, linked to the policy rate.
- Since December 2013, the policy repo rate has risen by 200 bps.
- One-year negotiable certificate of deposits (NCD) rate increased by 274bps since December 2013.
- The gap between the NCD rate and the policy rate stands at about 170 bps.
- Required Liquidity Coverage Ratio (LCR) will continue to rise by 10 percentage points each year to reach 100 percent in January 2019 (including up to 40 percent from the South Africa Reserve Bank’s (SARB) committed liquidity facility).
- The Net Stable Funding Ratio (NSFR) is to be fully implemented in 2018.
- Bank short-term wholesale funding currently accounts for about 60 percent of total bank liabilities.
- About 60 percent of NBFIs’ assets are invested in stocks.
- Many international firms listed on the Johannesburg Stock Exchange have foreign earnings accounting for about half of total earnings of all listed companies.
- About half of foreign earnings are from EMs.
- Major European banks account for 76 percent of total external bank lending to South Africa (BIS data).

### A. The Role of Capital Flows in the South African Economy
- South Africa is highly reliant on capital flows: large and persistent current account deficits and rising external debt.
- Exchange rate depreciation in 2015 facilitated a shift to a positive net international investment position by lowering the U.S. dollar value of rand-denominated liabilities.
- External financing mix shifted away from portfolio and direct investment toward other investment flows and unrecorded transactions since 2013.
- Other investment flows comprise primarily bank loans and deposits; remainder includes financial derivatives, trade credit, and other net receivables.
- A significant share of other investment flows are short-term tenor (less than one year) and denominated in FX, posing rollover risks and higher hedging costs in a shock.
- While banks’ FX position is small and within regulatory limits, maturity mismatches exist because some FX assets are corporate loans to fund acquisitions abroad.
- Sustainability of bank-intermediated flows is uncertain; the source and economic rationale for unrecorded transactions are unknown.

### B. Capital Flows and Macro-Financial Linkages
- Outlook: capital flows to emerging markets expected to be scarcer, more volatile, and costlier (per April 2016 Global Financial Stability Report).
- For South Africa, expected continued negative net FDI flows and lower, more volatile portfolio flows as U.S. monetary policy normalizes and advanced economies’ policies diverge.
- Fund staff project other investment flows to be the main source of external financing in coming years, but sustainability is uncertain as major European banks retrench.
- Potential foreign investor pullout risks:
  - Large foreign participation in domestic bond and equity markets could cause sizable outflows if nonresidents liquidate holdings.
  - Favorable features mitigating shocks: floating exchange rate, deep financial markets, and about half of external debt in local currency.
  - Currency denomination varies by sector: bulk of government external debt in local currency; majority of SOEs’, banks’, and corporates’ external debt in FX.
  - SOEs required to hedge FX risk, but hedging may be rolling and imperfect; hedging costs and collateral requirements could rise with downgrades.
  - Nonresident liquidation would likely incur losses on asset prices, which could reduce outflow volumes.
- Transmission channels of capital outflows:
  - Depreciation of the rand, higher interest cost of debt, decline in local asset prices, negative wealth effects.
  - Depreciation may have limited export impact due to structural constraints and policy uncertainty.
  - Large outflow likely to force sharp import compression and sudden current account adjustment, reducing growth.
  - International firms listed on JSE and their foreign earnings (about half of total) provide some buffer against depreciation for stock prices.
  - Some importers (e.g., retailers) would suffer from rand weakness.
- Financial sector and fiscal feedback loops:
  - Rand depreciation would raise inflation, likely prompting SARB to hike rates; about 90 percent of loans are at floating rates, amplifying growth impact.
  - Initial boost to banks’ profitability from higher interest rates could reverse as funding and credit costs rise, leading to higher NPLs and lower profitability.
  - Decline in banks’ profitability would adversely affect fiscal revenue given heavy reliance on the financial sector for income tax.
  - Government likely able to secure domestic financing, but domestic portfolio rebalancing may depress equity prices (about 60 percent of NBFIs’ assets in stocks), increase interest rates, and reduce private lending (crowding-out), further depressing growth.
- Banking sector funding and regulatory dynamics:
  - Since December 2013, regulatory changes and higher policy rates have increased banks’ funding costs.
  - Rising LCR requirements and NSFR implementation will push banks toward more term funding and retail deposits; outstanding bank bond issuance projected to rise, increasing average funding costs.
  - SARB’s national discretion to apply a 35 percent available stable funding factor to short-term funding from financial corporate customers reduces NSFR challenge relative to prior assessments.
  - SARB has increased size of the money market liquidity shortage to enhance monetary policy transmission.
  - Capital outflows would exacerbate expected rises in banks’ funding costs.

*Source: _cr16218 - References (IMF staff prepared material).*

### 11.      Banks’ asset quality is likely to deteriorate even in the baseline scenario, though banks

### 11.      Banks’ asset quality is likely to deteriorate even in the baseline scenario, though banks should be able to withstand sizable credit shocks.

### Household-sector asset quality and risks
- Households’ debt-to-disposable income ratio is 78 percent.
- NPLs on the unsecured credit book have increased to 9.0 percent as of end-2015.
- After a few years of a falling NPL ratio, some bank analysts and credit rating agencies expect it to rise modestly over the next few years.
- Staff analysis (IMF, 2014) suggests that a severely adverse scenario would:
  - lead to a significant increase in households’ probability of default, and
  - reduce the Common Equity Tier 1 (CET1) ratio by 2.5 percentage points.

### Corporate-sector credit stress
- Pockets of credit stress have emerged in construction, mining, steel, and agriculture sectors; credit to these sectors represents only 6 percent of bank loans.
- Declining corporate sector profitability has weakened the interest coverage ratio (ICR) of South African corporates; the ICR stands at 3 in 2015Q1-Q3.
  - Mining and quarrying industry recorded an ICR below the benchmark of 2.
- The expected default frequency (EDF) of South African firms has deteriorated.
- IMF (Chow, 2015) stress analysis: a stress scenario similar to the global financial crisis increases NPLs only by 1-2 percentage points.
- Banks’ buffer and profitability context:
  - CET1 to risk-weighted assets at 13.6 percent.

### Stress tests, capitalization, and resilience
- Recent SARB stress tests indicate banks are adequately capitalized to withstand significant credit losses under stress scenarios.
- Even in the most adverse scenarios (protracted recession or excessive financial market volatility with capital flow reversal and significant depreciation), banks’ average CET1 capital adequacy ratio remains above the regulatory requirement.
- Banks are expected to maintain access to international capital markets, though at higher costs, given their strong balance sheets.

### Financial-cycle, credit growth, and macrofeedbacks
- Financial cycle observations:
  - Bank credit growth to households moderated from a peak of 24.8 percent y/y in March 2008 to 2.3 percent in April 2016.
  - Credit growth to corporates was 12.1 percent y/y as of April 2016.
  - The financial cycle has rebounded from its 2011 trough but appears to have peaked at a low level.
  - The SARB determined no need to activate countercyclical capital buffers as the credit gap remains negative.
  - South Africa’s financial cycle has become increasingly correlated with the U.S. financial cycle (IMF, 2016).
- Role of capital flows:
  - Buoyant post-global financial crisis capital flows helped finance rising government debt and part of banks’ operation, allowing higher domestic funding than otherwise.
- Potential macrofeedback from deleveraging:
  - In a severely adverse scenario, banks likely deleverage, reducing lending to the private sector (private sector accounts for 67 of their total assets).
  - Fund staff estimate that a 2½ percentage point decline in credit growth relative to the baseline (from 7 percent to 4½ percent) could reduce real GDP growth by 0.5–1 percentage points.
  - Cross-country distribution context (from table):
    - At credit growth 7%: 10th percentile GDP growth 0.5; 25th 2.6; 75th 5.9; 90th 7.5.
    - At credit growth 4.5%: 10th percentile GDP growth -0.5; 25th 1.6; 75th 5.4; 90th 6.8.
    - Growth difference by percentiles: 10th 1.0; 25th 1.0; 75th 0.5; 90th 0.7.

### Fiscal exposure and financial sector importance
- Financial sector size and fiscal relevance:
  - Financial sector assets amount to 300 percent of GDP.
  - Finance, real estate, and business services account for 20 percent of GDP and 22 percent of formal nonagricultural employment.
  - Financial sector accounts for about half of personal income tax and one third of corporate income tax.
- A shock to the financial sector could have a large impact on fiscal revenue, forcing the government to restrain spending with adverse second-round effects on growth.

### Impact of potential sovereign debt rating downgrades — overview
- Current sovereign ratings context:
  - South Africa’s sovereign credit rating is one to two notches above investment grade (IG) for FX debt, and two to three notches above IG for LC debt.
  - Two rating agencies (Moody’s and Standard & Poor’s) have a negative outlook.
- Markets assign a high probability that South Africa’s sovereign FX debt credit rating will be downgraded to speculative grade; LC downgrade to speculative grade is not in staff’s baseline.

### Impact of a foreign currency (FX) sovereign debt rating downgrade
- Market pricing and spreads:
  - South Africa’s 5-year CDS spreads trade in line with BB+ rated EMs (e.g., Turkey, Brazil, Russia).
  - 5-year CDS spread increased by about 80 bps since November 2015 to 320 bps at end-May 2016.
  - Past episodes: CDS spreads typically increase by about 100 bps before downgrade and an additional 10-20 bps after it materializes.
  - Most of South Africa’s FX rating downgrade risk is probably priced in, but additional widening is possible when the downgrade materializes due to forced sales by investors with IG-only mandates.
  - JP Morgan estimates forced sales could amount to about US$2 billion based on survey of institutional investors.
- Impacts on SOEs and banks:
  - A sovereign FX downgrade would likely be followed by downgrades of most other South African FX debt issuers, including SOEs and banks.
  - SOEs: about 40 percent of their total debt is denominated in FX; funding costs could rise significantly.
  - Eskom’s funding costs have risen significantly since last November (relative timing in source).
  - Some SOE bonds/loans may have covenants contingent on maintaining an IG credit rating; covenants may trigger cost escalation or loan callability.
  - Banks: about 10 percent of liabilities are in foreign currency; US$13.5 billion (4 percent of GDP) of these foreign-currency liabilities will come due within one year.
  - Syndicated loan spreads over Libor have largely increased in line with sovereign CDS spreads; syndicated loans typically 3-year floating with fees and spreads tied to sovereign rating.
- Regulatory capital effects for international banks:
  - For local banks, risk weight for holding local currency government bond of South Africa is zero regardless of sovereign ratings.
  - For South African bank branches of international banks, home regulators assign high risk weighting for bonds below investment grade.
    - Example: for BBB- rated government debt, foreign bank branches assign risk weight of 50 percent; if downgraded to BB+, risk weight doubles to 100 percent.
  - International banks could see higher capital charges, potentially leading to higher government bond yields.
- Investor base and capital-flow volatility:
  - An FX debt downgrade would likely worsen the profile of the nonresident investor base toward more short-term, high-yield-seeking investors.
  - IMF study: institutional investors tend to reduce their investment on average by 16 percent of assets under management and only return slowly following a downgrade to speculative grade.
- Historical duration of sub-IG spells:
  - Since 1980, 19 cases of sovereign FX downgrades from IG to speculative (by at least 2 rating agencies).
    - 6 sovereigns regained IG after an average of 8 years.
    - The remaining 13 stayed below IG, averaging 9 years since loss of IG status.

### Impact of a local currency (LC) sovereign debt rating downgrade
- LC ratings and likely co-movement:
  - Moody’s does not differentiate between LC and FX credit risk and rates both at the same level.
  - Standard & Poor’s rates South Africa’s LC sovereign debt two notches above the FX sovereign debt (maximum spread S&P allows).
  - Fitch’s LC rating is one notch above the FX rating.
  - If FX rating is downgraded, LC ratings would likely be downgraded too, but may still remain IG.
  - Financial markets do not appear to be pricing in a downgrade of South Africa’s LC sovereign rating below IG level at the time of the source.
- Forced sales risk and index/mandate sensitivities:
  - Bloomberg bond ownership database analysis for two benchmark LC sovereign bonds (5- and 7-year):
    - Foreign holdings account for more than half of the outstanding amount; these two bonds account for 13 percent of total foreign holding of LC sovereign fixed-rate bonds.
    - More than half of nonresident investors of these bonds follow JP Morgan’s Emerging Market Bond Index (GBI-EM), which does not require IG credit rating.
    - 21 percent of nonresident investors are IG rating-sensitive due to index restrictions or mandates.
    - Some global bond indices require LC IG ratings (Citigroup WGBI, Vanguard International Bond Index, Barclays Global Aggregate).
    - Some bond funds require 80 percent of investments in IG-rated bonds.
  - Staff estimate: nonresidents’ total forced sales could amount to about 2½ percent of GDP in the event of LC sovereign rating downgrade to speculative grade.
  - Other investors might sell preemptively; the above estimates may be a lower bound.
  - Historically, equity flows have displayed a high correlation with bond flows in times of stress.

*Source: IMF staff report excerpt (chapter on banks’ asset quality, stress tests, financial cycle, and sovereign rating downgrade impacts).*

### 24.      NBFIs have the balance sheet to absorb the government bonds sold by the

### _cr16218 - 24.      NBFIs have the balance sheet to absorb the government bonds sold by the

### NBFIs' capacity to absorb foreign sales and associated risks
- NBFIs held total assets of 210 percent of GDP in 2015, of which less than 10 percent are local government bonds.
- NBFIs could provide a backstop to foreign bond sales and absorb new issuance.
- NBFIs’ assets comprise mainly domestic equities and banks’ NCDs, so portfolio rebalancing to increase government bond allocations could:
  - lead to asset price declines;
  - reduce funding for banks (via lower demand for banks’ NCDs);
  - cause liquidity tightening and a significant rise in funding costs for banks, as illustrated by the jump in short-term rates observed during market turmoil episodes in August and December of 2015.

### Mechanisms for a liquidity squeeze (including in a “closed rand system”)
- Even with capital controls on residents and a floating exchange rate (the “closed rand system”), capital outflows (e.g., foreigners selling LC bonds) can trigger bank funding needs:
  - Banks may need to provide funding for NBFIs switching from NCDs to government bonds, for example by buying back NCDs from NBFIs.
  - To buy back NCDs, banks would have to find liquidity by selling other assets.
  - Market demand for short-term liquidity rises significantly in periods of high risk aversion, and banks may need to raise term funding to meet LCR requirements under unfavorable market conditions.
  - SARB can intervene to provide liquidity, but banks may be limited by the availability of eligible collateral.
  - In a more severe capital outflow shock, banks could face medium-term liquidity challenges if:
    - NBFIs reduce their bank exposure as part of portfolio rebalancing; and
    - corporates reduce bank deposits when profitability suffers.
- Potential bank responses include retrenching credit to restore LCR ratios and passing on higher funding costs to customers, leading to significantly tighter financial conditions and lower growth, even if all agents ultimately get funding.

### Constraints on NBFIs’ stabilizing role and likely market adjustments
- NBFIs’ relatively low holdings of government bonds could reflect a preference for higher-yielding instruments; there is no minimum requirement for NBFIs to hold government bonds in South Africa.
- NBFIs’ funding to banks can be relationship-driven because insurers and CIS tend to hold bank paper issued by their connected banks.
- During market turmoil in August and December 2015, NBFIs did not substantially increase purchases of government bonds despite the large jump in bond yields.
- Mandatory capital repatriation after breaching prudential limits on overseas asset holdings is subject to delay:
  - NBFIs’ overall foreign asset holdings averaged 15 percent, below the regulatory ceiling, though some large private nonfinancial institutions are reported to be at the limit.
  - When the limit is breached, NBFIs can no longer increase offshore allocation but have one year to repatriate funds, implying delayed stabilizing effects from repatriation.
- Therefore, it may take significant bond price adjustments and rand depreciation for NBFIs to substantially increase allocations to government bonds via portfolio rebalancing.

### Asset allocation and offshore exposure (2015)
- NBFIs allocate the majority of assets to equities (percent of total assets, 2015):
  - Equities: 57%
  - Fixed rate bonds: 28%
  - Cash and deposit: 7%
  - Loans: 5%
  - Other: 3%
- Share of Foreign Assets in NBFIs’ Portfolios and Prudential Limits (Assets Under Management, Percent of offshore assets in total assets, Offshore limits, percent of total assets):
  - Collective Investment Schemes: 2 (ZAR trn); 20 (percent); 30 (percent)
  - Pension Funds (FSB-registered): 2.5 (ZAR trn); 19 (percent); 25 (percent)
  - Life Insurance Companies: 1.8 (ZAR trn); 14 (percent); 30 (percent)
- Note: Assets under management cannot be added up due to overlapping ownership.
- Footnote: An additional 5 percent of total assets is allowed for investment in the rest of Africa.

*Source: SARB and FSB.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16218.pdf_