## cr18247

## Source details

**Canonical URL:** [cr18247](https://www.imf.org/-/media/files/publications/cr/2018/cr18247.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2018/cr18247.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2018/cr18247.pdf.json)

---

### Debt dynamics, drivers, and projections
- National government debt: declined from 48 percent of GDP (FY94/95) to 26 percent of GDP (FY08/09); rose to 53 percent of GDP by FY17/18.
- Decomposition of change in the debt numerator (2007–17):
  - Fiscal deficit: 76 percent of the change.
  - Stock flow adjustment: 18 percent.
  - Accumulation of cash balances: 6 percent.
- Fiscal deficit increase drivers (2007–17):
  - Expenditure increase accounted for 90 percent of the deficit rise; revenue shortfalls contributed marginally.
  - Unreversed crisis-related expenditure increases (2008–09 vs 2007):
    - Wage bill: 1.5 percent of GDP.
    - Purchases of goods and services: 1.0 percent of GDP.
    - Social benefits: 0.5 percent of GDP.
- Revenue path:
  - Revenue declined by about 3 percentage points of GDP to 24.1 percent of GDP in 2010.
  - Revenue recovered to 26.4 percent of GDP in 2017.
- Staff projection (on current policies): growth outlook would remain lackluster and debt would continue to rise.
- Policy recommendation: stabilize debt dynamics at comfortable levels as soon as possible to strengthen economic resilience.

### General government and SOE debt, contingent liabilities
- General government debt: increased from 29.3 percent of GDP in 2007 to 55.7 percent of GDP in 2017.
- Implied rest-of-general-government debt: between 2 and 3 percent of GDP.
- SOE debt:
  - Total SOE loan and bond debt: 8 percent of GDP in 2007 to 15.6 percent of GDP in September 2017.
  - Non-financial SOEs: about 86 percent of SOE debt (13.5 percent of GDP) and 86 percent of the overall SOE debt increase.
  - Non-financial SOE assets: about 37 percent of GDP; financial SOE assets: about 5 percent of GDP.
  - Non-financial SOE cash flow deficits in 13 out of the 14 most recent years; average cash flow deficit about 1.6 percent of GDP in the last 10 years.
  - Average investment program about 3 percent of GDP per annum; operational revenue insufficient to sustain investment.
- SOE guarantees and contingent liabilities:
  - Guarantees: 2.9 percent of GDP in FY07/08 to about 9 percent in FY17/18.
  - Largest loan guarantees (Eskom + IPP): combined 7.3 percent of GDP in FY17/18.
  - Road Accident Fund claims: 1.4 percent of GDP in FY07/08 to 4 percent in FY17/18.
  - Medical assistance program and disputed bill claims: 2.2 percent of GDP in FY17/18 (declining over time).
- Measurement note: proxying general government debt by adding bond and loan data risks double counting due to lack of intra-government debt information.

### Composition of expenditure and efficiency drivers
- National government expenditure composition (contribution to increase in expenditure, 2007–17):
  - Wage bill: 64 percent of the increase.
  - Interest bill: 23 percent.
  - Social benefits: 13 percent.
- At general government level, contribution to total expenditure increase:
  - Wage bill: 55 percent.
  - Purchases of goods and services: 19 percent.
  - Interest bill: 17 percent.
  - Social benefits: 14 percent.
  - Net decline in remaining expenditure categories: about 5 percent of total expenditure.
- Capital spending and SOEs:
  - Budgetary capital expenditures share is lower than SADC and EM comparators partly because SOEs carried out about 3 percentage points of GDP of capital spending not captured in consolidated government figures.
- Wage bill and public employment:
  - Wage bill is on the high side of the EM distribution driven by high compensation.
  - Public employment as a share of the working-age population is lower than the EM average.
  - Both the wage bill as a share of GDP and of public expenditure are above the EM average.
- Sectoral efficiency signals:
  - Education: inefficiencies noted; several SADC countries spend less at primary level and achieve better outcomes; tertiary subsidies tend to be regressive in absolute terms.
  - Health: spending slightly above EM and SADC, but HALE and other outcome indicators are considerably below peers, indicating inefficiencies.
  - Social protection: spending considerably higher than OECD, EM, and SADC; coverage of poorest 20 percent is 83 percent in South Africa versus 59 percent in EMs and 79 percent in OECD.
  - Energy subsidies: IMF (2015) estimates energy subsidies at 13 percent of GDP (coal and petroleum 8 percent of GDP; electricity 1.2 percent of GDP); VAT exemption on fuel forgone revenue estimated at 1.5 percent of GDP.

### Was debt growth growth-enhancing?
- Crisis response:
  - Fiscal stimulus likely smoothed the global financial crisis impact: with multipliers estimated between 0.4 and 0.9, the stimulus likely prevented a sharper decline in growth of between 1.2 and 2.7 percentage points during the crisis.
  - Multiplier estimates tend to be below one in the short term and negligible in the longer term in most studies; alternative estimates (Makrelov et. al., 2018) suggest higher peak multipliers (2.5 and 3.5) when accounting for financial accelerators.
- Longer-term effects:
  - Expenditure increases over 2007–17 were mainly wage and interest payments and, to a lesser extent, social benefits.
  - Given spending composition and inefficiencies in health and education, the increase in spending did not have material permanent effects on growth.
- Implication: reallocating resources away from recurrent spending (wage increases and debt service) toward more productive outlays—such as highly productive capital expenditure and reforms in key network industries—would likely have yielded higher growth gains.

### SOE market access, borrowing composition, and policy implications
- SOE debt composition shift (non-financial SOEs):
  - 2007: about 72 percent bonds and 28 percent loans.
  - 2017: 51 percent bonds and 49 percent loans, with increasing loan placements with domestic financial institutions and multilateral financing.
- SOE financing dynamics:
  - Deteriorating finances and limited market appetite for non-financial SOE bonds led to increased government guarantees.
  - SOEs ran protracted deficits as a group during 13 of the last 14 years; most need government guarantees to borrow.
- Policy recommendations for SOEs:
  - Urgently implement turnaround plans for SOEs; condition any new government guarantees on meeting well-defined turnaround targets and efficient performance.
  - Make new guarantees contingent on efficient performance in line with well-defined turnaround targets.

### Poverty, inequality, labor, and fiscal incidence
- Poverty and inequality disparities:
  - Children and the elderly most affected by poverty; poverty levels highest for black Africans; black African women face the highest levels.
  - Poor households earn less than 40 percent of their income from employment; the remainder is government transfers.
- Employment and unemployment:
  - Unemployment at end-2017: 26.7 percent.
  - Youth unemployment (15–29): 52.2 percent.
  - Including discouraged workers: overall 36.4 percent; youth 67.4 percent.
  - Nearly 40 percent of unemployed have never had a job; for youth this rises to 60 percent.
- Spatial and provincial disparities (2015 figures shown in referenced figures):
  - Examples: Gauteng GDP = 97; Western Cape GDP = 85; Eastern Cape GDP = 49; Limpopo GDP = 50.
  - Poverty headcount and incidence in rural areas about twice urban levels.
- Financial inclusion:
  - Bottom quintile has half the number of bank accounts of the top quintile; access to loans and credit cards is one tenth that of the top quintile.
  - Bottom quintile accounts for 33 percent of loans from “mashonisas” compared to 8 percent for top quintile.
- Working-age population and employment projections (selected values from Table 2):
  - Working age population: 2016 = 36.7; 2030 = 38.9; 2050 = 43.0.
  - Total Employed: 2016 = 15.8; 2030 (Average2015 Elasticity) = 17.0; 2050 = 19.3; 2030 (2016 Elasticity) = 17.9; 2050 (2016 Elasticity) = 21.9.
  - Unemployed: 2016 = 5.8; 2030 = 6.5; 2050 = 8.0; alternative scenario values include 2030 = 6.8 and 2050 = 8.8.
  - Working age population expected to increase by over 2 million until 2030 and over 6 million by 2050.
  - Job creation at average 2010–16 levels would imply unemployed rising to 8 million by 2030; worst-case (2016 elasticity) would see unemployment exceed 10 million by 2030.
- Growth, poverty, inequality:
  - Growth is pro-poor in that extended growth periods reduce poverty, but has not reduced inequality because most jobs created are high-skilled.
  - On current trends, South Africa risks not achieving target of reducing Gini to 0.6 by 2030; comparable reductions in other EMs took 17–23 years.
- Fiscal incidence and progressivity:
  - Tax system is mildly progressive; average net impact of taxes on market Gini across four waves: Wave 1 = -0.03; Wave 2 = -0.02; Wave 3 = -0.02; Wave 4 = -0.02.
  - Social spending is progressive in absolute terms; adding cash benefits, health, education, free services, and housing subsidy reduces market Gini by an average of 0.18 across four waves.
  - Combined impact of taxes and social spending reduces Gini on average by 0.2 (about 30 percent of market income Gini) across four NIDS waves.
  - South Africa remains the country with the highest Gini for final income in the sample due to a highly unequal market distribution.
- Coverage and public services:
  - Social grants coverage expanded from 3 million in 2000 to 17 million in 2016 (over 30 percent of population).
  - Access to electricity and water has increased consistently between 1996 and 2016.

### Progressivity opportunities and tax policy considerations
- Personal income tax:
  - Highest marginal tax rate increased to 45 percent in the 2017 budget.
  - Literature suggests revenue from high incomes maximized at marginal rates of "50–60 percent"; 45 percent approaches that range.
  - Further increases may provoke tax avoidance and revenue loss in the context of capital mobility.
  - Limited PIT buoyancy observed in recent years.
- Policy options:
  - Seek additional revenue from property taxes which "have favorable distributional properties."
  - Address weaknesses in tax administration to raise revenue for progressive expenditure.
- Spatial policy:
  - Balance reducing spatial disparities with preserving agglomeration gains via de-concentration (human and physical infrastructure) and removing impediments to migration.
  - Network industries and easing urban land and housing constraints are critical enablers.
- Risks of intervention:
  - Well-intended policies (e.g., BEEP, land reform) can create regulatory uncertainty, reduce private investment, increase wage premia, and incentivize rent-seeking.

### Resilience, buffers, external and financial vulnerabilities
- Key vulnerabilities and trends:
  - Growth below "two percent" for four consecutive years as of 2017.
  - Sustained fiscal deficits of "around 4 percent of GDP or above" contributed to rising debt.
  - Public-sector debt doubled since 2007 to "about 53 percent of GDP in 2017."
  - SOE debt rose from "7 percent of GDP in FY2004 to around 14 percent of GDP in FY 2016."
  - Cumulative net non-resident portfolio investment flows (2012–17) estimated close to "$90 billion."
  - As of May 2018, 10-year yields remained below the January 2016 high of "9.6 percent."
  - As of March 2018, Eskom and South African Airlines combined outstanding government guarantees = "7.2 percent of GDP."
  - Scenario: if growth falls permanently "1 percentage point" below the baseline, public sector debt would likely surpass "70 percent of GDP by 2023."
- Sovereign ratings and market reactions:
  - April 2017: Fitch lowered local currency rating below investment grade.
  - November 2017: S&P reduced local currency rating to sub-investment grade; Moody’s remained as the sole investment-grade rater at that time.
  - Market impact of MTBPS (October 2017): sudden net portfolio outflows of "about $1 billion" in eight business days; "3½ percent" exchange rate depreciation; pickup in 10-year yields.
  - Ownership of local-currency government bonds (December 2017): Nonresidents "41%"; Banks "15%"; Insurers "7%"; Local pension funds "27%"; Other "10%."
- Financial sector and household vulnerabilities:
  - Household debt at 34 percent of GDP at end-2017; household debt amounted to 72 percent of disposable income in 2017.
  - Banking sector assets ~110 percent of GDP as of November 2017; five largest banks account for more than 90 percent of assets.
  - PIC assets under management: more than R2.1 trillion (about 46 percent of GDP) as of end-2017; PIC held about 20 percent of government domestic marketable bonds (R452 billion) and more than 50 percent of SOE debt securities.
  - Concentration risks: PIC exposures to SOEs could transmit a large SOE failure to the financial sector.
- External sector:
  - Current account deficit narrowed to 2.5 percent of GDP in 2017.
  - Total external debt nearly 50 percent of GDP as of end-2017.
  - Short-term debt at remaining maturity >14 percent of GDP at end-2017; >50 percent of that is foreign-currency external debt of monetary and non-financial private sectors.
  - In 2017 the current account deficit was financed entirely by net portfolio flows; FDI recorded a net outflow of -1.7 percent of GDP.
  - Net IIP positive at 12 percent of GDP at end-2017; excluding less liquid FDI, net IIP would be -20 percent of GDP.
- Reserves, ARA metric, and access to facilities:
  - End-2017 reserve adequacy (unadjusted ARA) = 64 percent of the ARA metric; with CFMs accounted for reserves = 70 percent of ARA.
  - South Africa was close to $30 billion short of reaching the lower bound of ARA adequacy at 100 percent.
  - Reserves at end-2017 above adequacy thresholds for short-term debt at remaining maturity (reserves at 102 percent) and close to 20 percent of broad money.
  - Access to global facilities: BRICS CRA total resources $100 billion; South Africa contributes $5 billion and has access to $10 billion (30 percent without IMF arrangement). Bilateral swap line with China: CNY30 billion (about $4.7 billion).
- Tail-risk and forced outflows:
  - A sub-investment grade local-currency rating by all three agencies could prompt capital outflows and higher funding costs.
  - Staff estimates: in 2016 about 20 percent of local-currency debt was held by IG-sensitive investors; by March 2018 this share fell to around 2 percent.
  - A Moody’s downgrade could prompt forced outflows of about $1½ billion (½ percent of GDP).
  - Estimated potential forced outflows for benchmark bonds (USD billion): 1.3, 1.7, 1.5; Percent of 2017 GDP: 0.4, 0.5, 0.4.
- Mitigating factors:
  - Flexible exchange rate, large domestic financial sector, low share of foreign currency government debt (~10 percent), long average debt maturity (~15 years), and reserves adequate relative to short-term debt and broad money metrics provide cushions.

### Policy conclusions and prioritized recommendations
- Fiscal policy and public finances:
  - Pursue fiscal consolidation to rebuild policy buffers and stabilize debt dynamics.
  - Control personnel costs in education and health; prioritize efficiency improvements in service delivery.
  - Implement IMF Expenditure Assessment Tool (EAT) benchmarking to identify efficiency and reallocation opportunities.
  - Reorient spending toward productive capital expenditure and reforms in network industries.
- SOEs and contingent liabilities:
  - Urgently implement turnaround plans for SOEs; condition new guarantees on well-defined turnaround targets and efficient performance.
  - Limit unfunded contingent liability buildup; improve SOE governance and operational efficiency.
- Growth, jobs, and inclusion:
  - Implement structural reforms to lift growth, improve the business environment, and encourage private investment and job creation.
  - Use social assistance to protect the poor during fiscal adjustment; address distribution channel problems so benefits reach recipients.
  - Improve efficiency and quality of basic education and health to enhance redistributive impact.
  - Consider better-targeted tertiary financial support rather than broad cost recovery reductions.
- Resilience and buffers:
  - As opportunities arise and consistent with the inflation mandate, accumulate international reserves toward adequate levels.
  - Strengthen public-sector balance sheets, modernize infrastructure, and maintain an effective social safety net.

*Source: IMF staff extracts from cr18247 (chapter content).*

### 1. Local Currency Sovereign Bond Yield Spread to U.S. ___________________________________ 4

### 1. Local Currency Sovereign Bond Yield Spread to U.S.

### Introduction and key context
- From FY94/95 to FY08/09 national government debt declined from 48 percent of GDP to 26 percent of GDP; since the global financial crisis public debt roughly doubled, reaching 53 percent of GDP by FY17/18.
- In the last two fiscal years (prior to June 29, 2018), significant revenue shortfalls, slowing economic growth, governance weaknesses, increased spending pressures, and materializing contingent liabilities from SOEs raised financing needs.
- Favorable global financing conditions mitigated risks but became less benign as borrowing costs increased after sovereign downgrades; US monetary policy expectations and changes in investor appetite toward emerging markets further reduced benign conditions.
- Staff projection (on current policies) indicates growth outlook would remain lackluster and debt would continue to rise.
- Recommendation: stabilize debt dynamics at comfortable levels as soon as possible to strengthen economic resilience.

### What drove the accumulation of debt and contingent liabilities (2007–17)
- Decomposition of change in the debt numerator (2007–17):
  - Fiscal deficit accounted for 76 percent of the change in the numerator of the debt ratio.
  - Stock flow adjustment accounted for 18 percent.
  - Accumulation of cash balances accounted for 6 percent.
- The fiscal deficit increase was driven primarily by an increase in expenditure as a share of GDP, which accounted for 90 percent of the increase in the deficit; revenue shortfalls contributed only marginally.
- Specific expenditure changes around the 2008–09 crisis that were not reversed:
  - Wage bill increase: 1.5 percent of GDP (2008–09 vs 2007).
  - Purchases of goods and services increase: 1.0 percent of GDP.
  - Social benefits increase: 0.5 percent of GDP.
- Revenue evolution:
  - Revenue declined by about 3 percentage points of GDP to 24.1 percent of GDP in 2010.
  - Revenue recovered to 26.4 percent of GDP in 2017, buoyed by personal income, VAT, international trade tax collections, and tax policy measures (including bracket creep, changes in income tax rates, and adjustment of excise rates).

### National government expenditure composition and drivers of the deficit
- Economic classification decomposition (contribution to increase in expenditure, 2007–17):
  - Wage bill: 64 percent of the increase in expenditure.
  - Interest bill: 23 percent of the increase in expenditure.
  - Social benefits: 13 percent of the increase in expenditure.
  - Reduced capital expenditure and grants broadly offset increased expenditure on subsidies and, to a lesser extent, goods and services.
- Functional classification (largest areas of growth):
  - "General public services" is the largest area of growth (captures increasing wage and interest bills).
  - Education and health follow, likely reflecting large wage bill increases in those employment-heavy sectors.
  - Social protection spending increased reflecting rising social benefits.

### General government debt (2007–17)
- General government debt increased from 29.3 percent of GDP in 2007 to 55.7 percent of GDP in 2017.
- Implied that the debt of the rest of the general government (outside national government) is between 2 and 3 percent of GDP.
- Provincial governments and extra-budgetary funds are not allowed to borrow; social security funds maintained a surplus during the projection period.
- Contribution to total expenditure increase (general government level):
  - Wage bill: 55 percent.
  - Purchases of goods and services: 19 percent.
  - Interest bill: 17 percent.
  - Social benefits: 14 percent.
  - Net decline in remaining expenditure categories of about 5 percent of total expenditure.
- Note on measurement: general government debt was proxied by adding available bond and loan debt information from balance sheets of extra-budgetary funds, provincial governments, and local governments; this proxy risks double counting due to lack of intra-government debt information.

### State-Owned Enterprise (SOE) debt and contingent liabilities
- Total SOE loan and bond debt increased from 8 percent of GDP in 2007 to 15.6 percent of GDP in September 2017.
- Non-financial SOEs accounted for about 86 percent of SOE debt (13.5 percent of GDP) and 86 percent of the overall increase in SOE debt during the period.
- Asset size disparity: non-financial SOE sector assets about 37 percent of GDP; financial SOE sector assets about 5 percent of GDP.
- Non-financial SOE operational performance and investment:
  - Non-financial SOEs had cash flow deficits in 13 out of the 14 most recent years.
  - Average cash flow deficit about 1.6 percent of GDP in the last 10 years.
  - Average investment program about 3 percent of GDP per annum.
  - Operational revenue insufficient to sustain investment; expenditure grew by 0.6 percent of GDP driven by increased spending on goods and services and interest.
- Illustrative SOE issues (Eskom example): collection of arrears, procurement of key inputs at high costs, overstaffing, operational inefficiencies, delays and cost overruns in investment projects, and unfunded mandates.

### Policy considerations and tools highlighted
- Given the critical role of expenditure in debt accumulation, identify areas with room for efficiency gains or rationalization.
- Use of IMF’s Expenditure Assessment Tool (EAT) for benchmarking spending level and composition against comparators (SADC, Emerging and Middle-Income Countries (EM), and, in some cases, OECD).
- Implication: reallocating resources away from recurrent spending (wage increases and debt service) toward more productive outlays—such as highly productive capital expenditure and reforms in key network industries—would likely have yielded higher growth gains.

*Source: IMF staff study on South Africa (June 29, 2018).*

### 10.          SOE borrowing has increasingly

### 10. SOE borrowing has increasingly taken the form of loans

### SOE debt composition and market access
- In 2007, about 72 percent of non-financial SOEs obligations were bonds and 28 percent loans.
- In 2017 the mix was 51 percent for bonds and 49 percent for loans, with increasing loan placements with domestic financial institutions and reliance on multilateral financing.
- With deteriorating finances and limited market appetite for non-financial SOE bonds, a significant share of non-financial SOE debt is guaranteed by the government.

### Contingent liabilities from SOEs and other sources
- Guarantees on SOE loans are the most frequent and fastest-growing contingent liabilities.
- Guarantees amounted to 2.9 percent of GDP in FY07/08 and increased to about 9 percent in FY17/18.
- The largest loan guarantees were to Eskom and the independent power producers (IPP), which combined amounted to 7.3 percent of GDP in FY17/18.
- The IPP initiative was introduced in FY12/13 to alleviate energy shortages.
- Road Accident Fund claims rose from 1.4 percent of GDP in FY07/08 to 4 percent in FY17/18 and were the second most important source of contingent liability growth outside SOEs.
- Contingent liabilities related to the medical assistance program for civil servants and claims related to disputed bills together amounted to 2.2 percent of GDP in FY17/18 (and have been declining over time).

### Fiscal implications and SOE performance
- SOEs ran protracted deficits as a group during 13 of the last 14 years, and most SOEs need government guarantees to borrow, indicating difficulty sustaining investment plans.
- Turnaround plans for SOEs are urgently needed; new government guarantees should be made contingent on efficient performance in line with well-defined turnaround targets.

### Key statistics and charts referenced (textual)
- Figures referenced include: Non-Financial Public Enterprise Debt and Deficit, 2004–17; Composition of Non-Financial SOE Debt, 2007–17; Composition of Contingent Liabilities, 2007–17.

---

### C. Benchmarking of spending levels and composition

### Overall spending trends
- Total spending as a percent of GDP increased by about 5.5 percentage points of GDP in 2007–16, a rate considerably higher than observed in SADC, EM, and OECD countries.
- Total spending is now broadly at par with the level of spending in other EMs and significantly above the SADC average.
- National government spending increased by 4 percentage points of GDP in the last decade, of which 3 points occurred during the global financial crisis years.

### Compositional shifts and capital spending
- The share of budgetary capital expenditures is lower than in the SADC and EM countries, partly because SOEs carried out about 3 percentage points of GDP of capital spending on average that is not captured in consolidated government figures.
- Compared to SADC, South Africa spends more on interest and other current expenditure because of higher debt levels and more extensive social assistance expenditure.

### Wage bill and public employment
- The wage bill is on the high side of the EM distribution and is driven by high compensation.
- Public employment as a share of the working-age population is lower in South Africa than the EM average.
- Both the wage bill as a share of GDP and of public expenditure are above the EM average, indicating high compensation rather than employment levels drive the larger wage bill.
- An annex in the 2017 Medium Term Budget Policy Statement provides a related analysis on public employment and compensation.

### Infrastructure and productivity implications
- Infrastructure quality compares favorably with respect to other EMs and SADC countries but lags OECD countries, particularly on ports and railroads; air transport and roads rate relatively well.
- The expansion in the wage bill during the last decade exacerbated budget rigidities without raising productivity; interest bill increases associated with rising debt further compounded these rigidities.
- A meaningful reduction in the wage bill would improve budget flexibility and spread the burden of adjustment more fairly.

---

### Composition of expenditure and sectoral efficiency

### Education
- Signs of inefficiency in education spending: several SADC countries spend less at the primary education level and achieve better outcomes (e.g., some SADC countries attain a 95 percent enrollment rate with similar spending).
- Since the 2016 budget review, reallocations funded additional tertiary education subsidies.
- International evidence suggests tertiary education spending tends to benefit disproportionately the better off and is regressive in absolute terms; subsidizing students who would have paid crowds out resources for those in greater need.
- Shift toward free tertiary education spending is unlikely to benefit the most vulnerable.

### Health
- Health expenditure is slightly above EM and SADC countries, but healthy life expectancy indicators (HALE) are considerably below what EM and several SADC countries achieve.
- Other indicators show relatively high infant mortality, lower life expectancy at birth, and lower number of physicians per 1000 inhabitants compared to other EMs.
- These outcomes point to inefficiencies in health spending.

### Social protection
- Spending on social assistance is considerably higher than in the OECD, EM, and SADC countries.
- Coverage of the poorest 20 percent is 83 percent in South Africa, compared to 59 percent in EMs and 79 percent in OECD countries.
- Benefit incidence: the poorest 20 percent get 20 percent of benefits in South Africa compared to 28 percent in EMs and 34 percent in OECD countries.
- Distribution issues exist where intermediaries force beneficiaries to contract for other services, reducing effective grant amounts received.

### Energy subsidies
- IMF (2015) estimates energy subsidies at 13 percent of GDP.
- Subsidies derive primarily from coal and petroleum (8 percent of GDP) and electricity (1.2 percent of GDP).
- Subsidies stem primarily from the exemption of fuel from VAT, with forgone revenue estimated at 1.5 percent of GDP, and from the gap between excise collections on energy products and the valuation of the externalities they generate.

---

### D. Was debt good for growth?

### Crisis response and short-term impact
- Fiscal stimulus likely smoothed the impact of the global financial crisis:
  - With multipliers estimated between 0.4 and 0.9, the fiscal stimulus likely prevented a sharper decline in growth ranging between 1.2 and 2.7 percentage points during the crisis.
  - Most studies suggest an impact of the government spending expansion on economic activity during the global financial crisis in roughly the same order of magnitude estimated by IMF staff, but limited effects in the longer term.
  - Multiplier estimates tend to be below one in the short term and negligible in the longer term in most studies; Makrelov et. al. (2018) argue higher peak multipliers (2.5 and 3.5) are possible when accounting for financial accelerator channels.

### Longer-term effects and composition of spending
- The expenditure increase over 2007–17 was mainly explained by wage and interest payments outlays, and to a lesser extent social benefits.
- Given the composition and relative inefficiency of spending in health and education, the increase in spending that drove up debt did not have material permanent effects on growth.
- The significant resources spent on wages and associated borrowing costs could have been invested more productively.

---

### Policy conclusions and recommendations (summarized from findings)
- Urgently implement turnaround plans for SOEs to ensure their contribution to growth and mitigate fiscal risks; condition any new government guarantees on meeting well-defined turnaround targets and efficient performance.
- Control personnel costs in education and health, implement auditor general recommendations to reduce wasteful expenditure, and prioritize efficiency improvements in service delivery.
- Use social assistance benefits to protect the poor during fiscal adjustment — broad coverage and relatively good targeting make social assistance a valuable compensatory policy, contingent on addressing distribution channel problems.
- Reorient spending toward more productive uses to raise workforce productivity and reduce inequality: additional productive capital expenditure supporting private-sector activity at national and subnational levels; structural reforms to increase productivity, especially in network industries where large SOEs operate; and improvements in the efficiency of basic education and healthcare.

*International Monetary Fund — Selected extracts from cr18247 (chapter 10) provided in the source content.*

### 1. The aggregate data on poverty and inequality mask significant disparities across age,

### 1. The aggregate data on poverty and inequality mask significant disparities across age, race, and gender.

### Major disparities and household income composition
- Children and the elderly are the age groups most affected by poverty.
- Poverty levels and severity are highest for black Africans; black African women face the highest levels among all groups.
- Income inequality levels are highest among black Africans.
- Poor households earn less than 40 percent of their income from employment, with the rest being government transfers.
- Income from employment accounts for 80 percent of the total income for households that move in and out of poverty.
- Nearly 40 percent of unemployed have never had a job; this rises to 60 percent for the youth.
- One in three of the unemployed were last employed five years ago; the proportion rises to almost half for those aged between 50 and 65.

### Employment, youth unemployment, and education links
- Unemployment at end-2017 stood at 26.7 percent.
- Youth unemployment (15–29 age group) stood at 52.2 percent.
- Including discouraged workers, the rates increase to 36.4 percent (overall) and 67.4 percent (youth).
- Employment prospects are largely driven by education levels; unemployment is lowest for university graduates.
- The share of labor income in total income largely determines poverty and inequality levels (Leibbrandt et al, 2012).

### Spatial disparities and provincial outcomes
- Regional gaps reflect apartheid-era restrictions on geographical settlement and associated government spending patterns.
- Poverty headcount and incidence in rural areas is about twice the level in urban areas.
- GDP per capita in the poorest provinces (Limpopo and Eastern Cape) is half the level of that in Gauteng.
- Provinces with the lowest participation in the labor force have the highest incidence of poverty.
- Provincial examples (2015 data shown in figures):
  - Northern Cape: Severity = 16.5; PG = 28.0; GDP = 71
  - Western Cape: Severity = 7.6; PG = 14.7; GDP = 85
  - North West: Severity = 19.8; PG = 32.2; GDP = 69
  - Free State: Severity = 14.2; PG = 25.1; GDP = 72
  - Kwazulu-Natal: Severity = 22.7; PG = 36.1; GDP = 59
  - Eastern Cape: Severity = 27.1; PG = 41.3; GDP = 49
  - Limpopo: Severity = 26.4; PG = 40.3; GDP = 50
  - Mpumalanga: Severity = 17.3; PG = 29.0; GDP = 69
  - Gauteng: Severity = 6.9; PG = 13.2; GDP = 97

### Financial inclusion and informal credit reliance
- The bottom quintile of households has half the number of bank accounts of the top quintile; many of these accounts are only used to receive social grants.
- The bottom quintile’s access to loans and credit cards is only one tenth of the access of the top quintile.
- Bottom quintile households account for 33 percent of loans from “mashonisas” (higher-cost informal lenders) compared to 8 percent for the top quintile.
- Table 1 shows detailed percent distributions of bank accounts, loans from banks, credit cards, loans from microlenders, and loans from mashonisas by income decile (data from NIDS and IMF staff calculations).

### Demographics, employment projections, and scenarios
- Working age population projections and employment (Table 2, STATS SA):
  - Working age population: 2016 = 36.7; 2030 = 38.9; 2050 = 43.0
  - Total Employed: 2016 = 15.8; Average2015 Elasticity 2030 = 17.0; 2050 = 19.3; 2030 (2016 Elasticity) = 17.9; 2050 (2016 Elasticity) = 21.9; 2030 (duplicate columns) = 15.9; 2050 = 16.2
  - Formal sector (non-agricultural): 2016 = 11.0; 2030 = 11.9; 2050 = 13.4; 2030 = 11.4; 2050 = 12.2; 2030 = 11.3; 2050 = 11.9
  - Informal sector (non-agricultural): 2016 = 2.6; 2030 = 2.8; 2050 = 3.2; 2030 = 3.5; 2050 = 5.2; 2030 = 2.5; 2050 = 2.3
  - Agriculture: 2016 = 0.9; 2030 = 1.0; 2050 = 1.3; 2030 = 1.6; 2050 = 3.1; 2030 = 0.9; 2050 = 0.8
  - Private households: 2016 = 1.3; 2030 = 1.3; 2050 = 1.4; 2030 = 1.5; 2050 = 1.9; 2030 = 1.3; 2050 = 1.2
  - Unemployed: 2016 = 5.8; 2030 = 6.5; 2050 = 8.0; 2030 = 6.8; 2050 = 8.8; 2030 = 7.3; 2050 = 10.1
- Working age population is expected to increase by over 2 million until 2030 and over 6 million by 2050.
- Job creation at the average levels observed between 2010–16 would imply a rise in the unemployed to 8 million by 2030.
- Worst-case scenario (job creation remains at a lower level, 2016 elasticity) would see unemployment exceed 10 million by 2030.

### Growth, poverty, and inequality dynamics
- Growth in South Africa is pro-poor (periods of extended growth contribute to declines in poverty), but growth has not contributed to reductions in inequality because most jobs created are high-skilled.
- Inequality increased even as income levels rose, underscoring fiscal policy’s role as a redistributive instrument.
- On current trends, South Africa runs the risk of not achieving its target of reducing income inequality (Gini) to 0.6 by 2030; meeting this target would require a radical pick-up in growth from current anemic levels.
- Comparable reductions in inequality in other EMs took between 17 and 23 years, except Argentina where a post-crisis recovery achieved rapid declines.

### Fiscal incidence, tax progressivity, and social spending
- Fiscal incidence analysis approach: starts from market income distribution per decile; recomputes Gini after taxes and benefits to estimate fiscal policy effect on inequality.
- Analysis covers: PIT, UIF, SDL (direct taxes); VAT, fuel excises, alcohol and tobacco excises (indirect taxes); health and education expenditure components; largest social grants (old age, disability, child); some indirect subsidies (housing subsidy, free water, electricity, sanitation for the poor).
- Tax system is mildly progressive:
  - Progressivity of direct taxes plus some progressivity of general fuel excise offsets regressive specific excises on alcohol and tobacco and the relatively less regressive VAT.
  - Average net impact of taxes on market Gini across four waves: Wave 1 = -0.03; Wave 2 = -0.02; Wave 3 = -0.02; Wave 4 = -0.02 (Table 3).
- Social spending is progressive in absolute terms:
  - Most progressive: direct cash transfers; followed by primary and secondary education and all health expenditure.
  - Tertiary education spending is progressive in relative terms due to a very skewed market distribution.
  - Adding cash benefits, health, education, free services, and housing subsidy to market income reduces the market Gini by an average of 0.18 across four waves.
  - Impact of social expenditure on market Gini by wave: Wave 1 net impact = -0.13; Wave 2 = -0.16; Wave 3 = -0.21; Wave 4 = -0.20 (Table 4).
- Combined impact of taxes and social spending reduces Gini substantially:
  - Market Income Gini and Gini after benefits (Table 5): Wave 1 Market Gini = 0.71; after benefits = 0.58; net impact = -0.13. Wave 2 = 0.76 to 0.59 (net -0.16). Wave 3 = 0.76 to 0.55 (net -0.21). Wave 4 = 0.69 to 0.50 (net -0.20).
- The Gini coefficient is reduced on average by 0.2 (about 30 percent of the market income Gini) across the four NIDS waves, compared with an average reduction of 0.06 in the Commitment to Equity Initiative sample.
- South Africa remains the country with the highest Gini for final income in the sample due to a highly unequal market distribution.
- Factors explaining fiscal policy’s significant impact:
  - Other EMs spend less on direct cash transfers.
  - Direct cash transfers in South Africa are relatively well targeted (categorical targeting works well).
  - Given unequal market income distribution, education and health spending tend to be better targeted to lower deciles.
  - South Africa collects more tax revenue through direct taxes (progressive PIT) versus other EMs that rely more on indirect taxes.

### Social program coverage and public services
- Coverage of social grants expanded from 3 million in 2000 to 17 million in 2016, reaching over 30 percent of the population.
- Share of households with access to electricity and water has increased consistently (figures show percent access in 1996, 2001, 2007, 2011, 2016).
- Black Economic Empowerment Program implemented to diversify ownership and increase employment opportunities.

### Policy considerations and recommendations
- Inclusive growth is critical to reduce inequality given limited fiscal space and the already substantial role of fiscal policy in redistribution.
- Achieving significant further inequality reduction requires increased growth and job creation.
- Options to improve the efficiency of fiscal policy in a constrained resource environment include:
  - Reconsider the reduction in cost recovery in tertiary education:
    - Tertiary education subsidies are relatively progressive but are an inefficient instrument to reduce inequality.
    - Focusing subsidies on those who are qualified but lack means would better target limited resources and likely achieve better results at lower cost.
  - Improve the efficiency of spending in education and health:
    - South Africa appears to spend more yet get weaker outcomes in education and health than countries that spend less.
    - Since basic education (primary and secondary) and health expenditures are progressive in absolute terms, improving quality and outcomes would increase their impact on reducing inequality.
  - Seek opportunities to leverage social grants:
    - Explore linking social grants to incentivize behaviors that improve education and health outcomes (examples from Brazil and Mexico include linking cash transfers to school enrollment and attendance at nutrition and health clinics).
    - Such links can reduce human capital inequalities (and future income inequalities) and current income inequalities.

*Prepared by Alejandro Simone and Vimal Thakoor; reviewed by Ana Lucía Coronel.*

### 21. Opportunities to achieve significant progressivity gains from the design of the

### 21. Opportunities to achieve significant progressivity gains from the design of the

### Personal income tax progressivity and constraints
- The highest marginal tax rate for higher incomes was increased to 45 percent in the 2017 budget.
- Literature suggests revenues from high income individuals are maximized at marginal rates of "50–60 percent" (see IMF (2014)), implying the current 45 percent rate is approaching that range.
- Further increases in top marginal rates may provoke increased tax avoidance and, in a context of significant capital mobility, could result in lower overall revenue.
- Limited buoyancy of the personal income tax has been observed in recent years, reducing the scope for progressivity gains via rate increases.

Policy implications / opportunities:
- Seek additional revenue from property taxes, which "have favorable distributional properties."
- Address weaknesses in tax administration to provide additional revenues that can finance progressive expenditure.

### Policy responses to spatial inequality
Findings:
- Policy needs to balance reducing spatial disparities with preserving agglomeration gains.
- Kanbur and Venables (2005) recommend a two-pronged strategy:
  - De-concentration of economic activity through the development of human and physical infrastructure.
  - Removal of impediments to migration of individuals and households to areas of high and rising well-being.

Implications for South Africa:
- Allow households with limited access to markets and public services to commute to centers of economic activity to share the benefits of growth.
- Network industries play a critical role in enabling access.
- Alleviating constraints to urban land and housing will facilitate migration of workers from rural areas to economic centers.

### Interventions addressing apartheid legacies — risks and trade-offs
- Well-intended policies (e.g., Black Economic Empowerment Program (BEEP) and land reform) can create regulatory and policy uncertainty that negatively impacts private investment and growth.
- Such policies can unintentionally worsen distributional indices by:
  - Increasing the wage premium.
  - Creating incentives for rent-seeking.

### Distributional evidence presented (figures summary)
- Figures 12–15 present distributional incidence for:
  - Indirect taxes: VAT, excise tax, fuel levy (cumulative share of benefits received by income decile shown against market income).
  - Social grants: Old Age Grant, Disability Grant, Child Support Grant (cumulative share of benefits received by income decile shown against market income).
  - Education: Primary, Secondary, Tertiary public spending incidence (cumulative share of benefits received by income decile shown against market income).
  - Health: Primary and Secondary, Tertiary public spending incidence (cumulative share of benefits received by income decile shown against market income).
- Sources for these figures: "Stats SA, NIDS, and IMF staff estimates."

### Definitions, data sources, and methodological caveats (Annex I)
Definitions of progressivity:
- Progressivity is measured by comparing the share of a specific tax/benefit collected/paid by each decile with the share of total income each decile receives and the 45-degree line.
- Relative progressivity/regressivity:
  - A tax is progressive/regressive in relative terms if the cumulative share of the tax paid by the households accounting for X percent of households is lower/higher than their share in income as shown by the market income curve.
  - A benefit is progressive/regressive in relative terms if the cumulative share of the benefit received by the households accounting for X percent of households is higher/lower than their share in income as shown by the market income curve.
- Absolute progressivity/regressivity:
  - A tax is progressive/regressive in absolute terms if the cumulative share of the tax paid by X percent of the households is less/more than X percent.
  - A benefit is progressive/regressive in absolute terms if the cumulative share of the benefit received by X percent of the households is more/less than X percent.

Data sources:
- National Income Dynamics Survey (NIDS) waves 1–4 covering 12-month periods in 2008, 2010, 2012, and 2014.
- Fiscal data from the National Treasury and the SARB.
- NIDS contains data on household income, expenditures, cash benefits, and utilization of education and health services.
- Budget spending data used to estimate values of certain benefits (e.g., average amount paid for grant types, average spending on levels of education and health).
- Personal income tax amounts in surveys are computed based on corresponding tax schedules obtained from national budgets due to limited tax amount reporting in surveys.

Methodological caveats:
- For health, education, and other free services, "the value of government services does not consider quality"; value to recipients may be less than government cost if quality is poor.
- Cash benefits valuations may overstate benefits if only a fraction reaches intended beneficiaries.
- The accounting approach does not consider behavioral responses to changes in taxes and spending.
- Analysis excludes important taxes (corporate income, international trade, and property taxes) and spending categories (infrastructure investments and certain grants with limited beneficiaries, e.g., care dependency grant) due to data and methodological constraints.

### Vulnerabilities and buffers: How resilient is the South African economy? — overview and public sector vulnerabilities
Context and main vulnerabilities:
- South Africa has weathered external and domestic shocks but faces weak growth, rising public sector debt, exposure to weak state-owned enterprises (SOEs), and a current account deficit financed by potentially volatile capital flows.
- A large share of external debt is denominated in rand, yet gross external financing needs are sizable.
- Downside risks (e.g., tighter global financial conditions, sovereign credit rating downgrade) could trigger abrupt capital outflows.
- The paper assesses domestic and external vulnerabilities and the size of buffers to absorb adverse shocks.

Key statistics and trends:
- Cumulative net non-resident portfolio investment flows (equity and debt) to South Africa during 2012–17 are estimated to have reached close to "$90 billion."
- As of May 2018, South African 10-year yields (monthly average) remained below the January 2016 high of "9.6 percent."
- 2017 marked the fourth consecutive year with growth below "two percent."
- Sustained fiscal deficits of "around 4 percent of GDP or above" during the past decade contributed to rising debt.
- Public-sector debt doubled since 2007, reaching "about 53 percent of GDP in 2017"—largely due to growing central government debt.
- Debt of SOEs rose from "7 percent of GDP in FY2004 to around 14 percent of GDP in FY 2016."
- As of March 2018, Eskom and South African Airlines had combined outstanding government guarantees corresponding to "7.2 percent of GDP."
- Scenario: If growth falls permanently "1 percentage point" below the baseline over the projection horizon, public sector debt would likely surpass "70 percent of GDP by 2023."

Sovereign ratings and market reaction:
- April 2017: Fitch lowered South Africa's sovereign local currency rating below investment grade.
- November 2017: S&P reduced the local currency sovereign credit rating to sub-investment grade; as a result, only Moody’s rated South Africa as investment grade.
- Market impact of the October 2017 MTBPS release:
  - Sudden net portfolio outflows of "about $1 billion" during the eight business days following the release.
  - A "3½ percent" exchange rate depreciation during that time, increased exchange rate volatility, and a pickup in 10-year yields.
- With "more than 40 percent of local currency government bonds held by nonresidents as of end-2017," South Africa is exposed to sudden financial market reactions.
- Ownership of local-currency government bonds, December 2017 (percent of total): Nonresidents "41%", Banks "15%", Insurers "7%", Local pension funds "27%", Other "10%."

Financial flows and volatility:
- Events cited (cabinet reshuffle, MTBPS, S&P downgrade, ANC elective conference) correlate with spikes in rand volatility and net portfolio flows.
- Cumulative non-resident net portfolio inflows (billions of rand) show episodes of outflows and inflows tied to political and fiscal developments (e.g., "Cyril Ramaphosa wins ANC election" corresponded with significant net inflows).

*Source: cr18247 - 21. Opportunities to achieve significant progressivity gains from the design of the personal income tax going forward will likely be more difficult (PDF).*

### 7.      The resilience of the South African economy helped reduce the impact on economic

### 7.      The resilience of the South African economy helped reduce the impact on economic stability

### Market reaction and portfolio flows
- Initial market impact of domestic events was relatively short-lived; 10‑year yields remained below 10 percent.
- International companies listed on the Johannesburg Stock Exchange (JSE) with offshore earnings may have benefited from exchange rate depreciation, playing a stabilizing role for equities.
- Portfolio equity flows recorded a net inflow of $¼ billion during the three weeks following the S&P downgrade (after prior equity outflows during several months of 2017 until early October), coinciding with global net non-resident equity flows to other emerging markets.
- Net non-resident portfolio bond inflows of about $½ billion were recorded in the week following the positive market sentiment after the leadership change in the ruling party; bond inflows only made up for outflows that started in October 2017 after significant inflows in late February and early March.

### Non-financial private sector vulnerabilities
- Non-financial corporate debt has risen but compares favorably to other major emerging markets.
- More than half of corporate debt is denominated in local currency.
- About three quarters of corporate debt has maturity longer than one year.
- Profitability appears generally sound in a cross-country perspective, though profitability in some sectors (e.g. mining and manufacturing) has weakened during the past several years.
- Debt at risk among non-financial SOEs has worsened as measured by the interest coverage ratio.
- Eskom, as of May 21, 2018, had an outstanding stock of bonds and loans of about R400 billion (8½ percent of 2017 GDP), with about R190 billion issued in foreign currency.
- Given significant government guarantees among some SOEs, vulnerabilities in this sector are closely linked to those that can impact the public sector.

### Household vulnerabilities
- Aggregate household debt is moderately high in a cross-country perspective: household debt at 34 percent of GDP at end-2017.
- Household debt has steadily declined since the GFC and virtually all household debt is denominated in rand.
- More than half the population lives in poverty and high unemployment continues to constrain household financial positions.
- As of 2017, household debt amounted to 72 percent of disposable income.
- Households face an interest rate spread of around 250 basis points relative to the prime rate.
- Loans and advances to households remained subdued at around 4 percent year-on-year in April 2018.
- A further increase in interest rates could put pressure on household debt servicing and impact bank balance sheets.

### Financial sector structure and resilience
- The financial sector is large and highly integrated with other sectors; dominated by a highly concentrated banking sector.
- Assets of pension and provident funds (including both official and private self-administered funds) amounted to R 3.5 trillion (76 percent of GDP) as of Q3 2017, with more than half accounted for by official pension funds.
- Banking sector assets corresponded to around 110 percent of GDP as of November 2017, with more than 90 percent of assets accounted for by the five largest banks.
- Capital adequacy is above regulatory requirements and profitability remains strong in a cross-country perspective, though profitability has declined moderately amid weak economic growth.
- A further slowdown in growth could increase the share of non-performing loans (NPLs), feeding into bank balance sheets.
- The banking sector is exposed to sizable contingent liabilities, largely related to credit lines with international companies located in South Africa.

### Pockets of banking-sector vulnerabilities
- Banks outside the Top 5 hold equity capital of around 15 percent of unweighted assets—nearly twice the level observed among Top 5 banks.
- NPLs among smaller banks are significantly higher than among the Top 5, with some smaller banks reporting NPLs of above 20 percent of loans.
- The capital ratios and NPLs for a couple of medium-sized and small banks are near those observed for African Bank when it was moving toward curatorship in 2014.
- Non-resident foreign-currency and derivatives funding is relatively large outside the Top 5—at 13 percent of total assets on average, and up to 30–50 percent of assets in some cases.
- Medium-sized and small banks on average tend to rely more on retail deposits than the Top 5, but some individual banks rely heavily on wholesale deposits.
- While the banking system as a whole is well-capitalized and profitable, some smaller, individual banks create vulnerabilities that, in isolation, would likely not be of systemic importance.

### The Public Investment Corporation (PIC) and concentration risks
- The PIC is the largest asset manager on the African continent, with more than R2.1 trillion (about 46 percent of GDP) of assets under management as of end-2017; close to 90 percent of these are those of the Government Employees Pension Fund (GEPF).
- PIC assets under management are invested in a variety of asset classes, in particular fixed income and equity.
- PIC deposits with banks account for less than 2½ percent of total banking sector deposits but are concentrated in few banks.
- Fixed income: as of end-2017 PIC held about 20 percent of government domestic marketable bonds (R452 billion or close to 10 percent of GDP) and more than 50 percent of SOE debt securities.
- Equity: PIC is the largest single institutional investor on the JSE; for example, PIC owned about 12 percent of Standard Bank shares in 2017.
- Overall domestic financial sector vulnerabilities are largely related to exposure to SOEs: only about 2½ percent of banking sector bonds and loans are to SOEs, but PIC is heavily exposed to large SOEs including Eskom and Transnet.
- In a tail event, a large SOE failure could spill over to the financial sector through the PIC.
- A sudden withdrawal of PIC deposits (which amount to about 2 percent of GDP) could potentially put strains on funding in some banks with potentially systemic impact.

### External vulnerabilities
- Current account deficit narrowed to 2.5 percent of GDP in 2017 but remains high relative to peers and estimated norms.
- Total external debt reached nearly 50 percent of GDP as of end-2017—almost a doubling since end-2007.
- External debt service needs are significant.
- Short-term debt at remaining maturity amounted to more than 14 percent of GDP at end-2017, of which more than 50 percent is foreign-currency external debt of the monetary and non-financial private sectors.
- Public sector and SOE external debt is largely of maturity longer than one year.
- In 2017, the current account deficit was financed entirely by net portfolio flows (equity and debt); foreign direct investment (FDI) flows recorded a net outflow of -1.7 percent of GDP.
- The composition of current account financing—reliance on flows prone to reversals—is a source of vulnerability.
- Net international investment position (net IIP) was positive at 12 percent of GDP at end-2017 (valued in rand-terms), amid external assets of 149 percent of GDP and liabilities of 137 percent of GDP.
- Excluding the generally less liquid net FDI position, the net IIP would have been -20 percent of GDP.
- The 30 percentage point of GDP increase in the net IIP between 2010 and 2016 is highly related to a rise in the valuation of FDI assets in China, notably a significant increase in the valuation of an investment in Tencent.
- The end-2016 net IIP position of 7 percent of GDP would have been about 20 percent of GDP lower if the increase in FDI assets in China were excluded; hence the positive net IIP position should be interpreted with caution.

*Source: IMF staff report (chapter content).*

### 18.      While vulnerabilities expose the South African economy to risks, mitigating factors

### 18.      While vulnerabilities expose the South African economy to risks, mitigating factors and existing buffers can help cushion the impact of adverse shocks, should they materialize

### Mitigating Factors
- Recent favorable global market conditions and capital flows that are highly correlated with global flows have benefited South Africa.
- Financial market reactions during periods of stress have been largely temporary, reflecting the economy’s resilience.
- Flexible exchange rate provides an automatic cushion in response to shocks.
- Large domestic financial sector provides a domestic investor base for government bonds; assets corresponding to more than 40 percent of GDP are managed by the PIC, which can provide a buffer in the event of sudden nonresident capital outflows.
- Low share of foreign currency-denominated government debt (about 10 percent) and long average term to maturity of debt (about 15 years) limit public sector exposure to exchange rate risk.
- The share of overall external debt denominated in foreign currency has moderated since the GFC and is now below 50 percent of total external debt.
- While short-term external debt at remaining maturity is non-negligible, more than 40 percent is denominated in rand.
- About 20 percent of short-term debt (original maturity) is related to FDI debt.

### Existing Buffers
- Reserve adequacy metrics and thresholds:
  - Assessing Reserve Adequacy (ARA) metric: reserves considered adequate when between 100 and 150 percent of the ARA metric.
  - Reserves relative to short-term debt at remaining maturity: adequacy threshold of 100 percent.
  - Reserves relative to broad money: adequacy range of 5–20 percent.
- ARA metric weights for countries with floating exchange rates (IMF, 2016a):
  - Short-term external debt at remaining maturity: weight 0.3.
  - Other external liabilities (sum of portfolio and other investment liabilities, less short-term external debt at remaining maturity): weight 0.15.
  - Broad money: weight 0.05 for countries without capital flows measures and 0.025 for countries with capital flow measures.
  - Exports: weight 0.05.
- Recent reserve adequacy developments:
  - End-2017 reserve adequacy (as measured by the ARA) declined to 64 percent of the unadjusted ARA metric.
  - After accounting for existing capital-flow measures (CFMs), reserves were 70 percent of the ARA metric.
  - South Africa was close to $30 billion short of reaching the lower bound of the ARA adequacy range at 100 percent of the ARA metric.
  - Reserves at end-2017 were above adequacy thresholds for:
    - Short-term debt at remaining maturity: reserves were at 102 percent of short-term debt at remaining maturity.
    - Broad money: reserves were at close to 20 percent of broad money.
- Access to other parts of the global financial safety net:
  - BRICS Contingent Reserve Arrangement (CRA): resources amount to $100 billion; South Africa contributes $5 billion. Total access for South Africa amounts to $10 billion, of which 30 percent can be accessed without an IMF arrangement.
  - Bilateral swap line with China: SARB agreement amounts to CNY30 billion (about $4.7 billion). Purpose: support trade and investment and act as a mitigating resource for short term balance of payment pressures. Initial agreement for three years; extended in 2018 for a further three-year period.

### Downside Scenarios
- Tail-risk scenario overview:
  - Tighter global financial conditions, potentially combined with a sovereign credit rating downgrade, could prompt sudden and significant non-resident capital outflows with material impact on the economy.
  - Materialization of a large contingent liability could prompt an increase in borrowing costs and a higher deficit and debt ratio, weighing on growth.
  - In a severe downside scenario with significant capital outflows and exchange rate depreciation, borrowing costs would rise for SOEs and banks.
  - If nonresident demand for government bonds dries up, the PIC could potentially step in to support demand, possibly at the expense of funding to banks, curtailing banking sector credit extension and negatively affecting investment and growth.
  - GEPF members’ benefits are guaranteed by the state; if pensioners are not paid benefits payable from PIC’s investments, the government would have an obligation to fill any remaining gap.
- Sovereign downgrade and forced outflows:
  - A sub-investment grade rating of South Africa’s sovereign local-currency credit rating by all three credit rating agencies could prompt capital outflows and increased funding costs.
  - In March 2018 Moody’s maintained South Africa’s sovereign rating at Baa3 and changed the outlook to stable.
  - Exclusion from Citigroup’s World Government Bond Index (WGBI) would result in forced sales of domestic government bonds if both S&P and Moody’s assign sub-investment grade ratings.
  - Staff estimates:
    - In 2016 about 20 percent of local currency government debt was held by IG-sensitive investors (IMF, 2016b); by March 2018 this share had fallen to around 2 percent.
    - A sovereign downgrade by Moody’s could prompt forced outflows of about $1½ billion or ½ percent of GDP.
    - Table estimates for potential forced outflows (for two benchmark bonds, R186 and 2023): Estimated potential forced outflows (USD billion): 1.3, 1.7, 1.5; Percent of 2017 GDP: 0.4, 0.5, 0.4.
  - Actual outflows could exceed forced sales due to negative market sentiment and hedge fund participation in the South African bond market.
- Cross-country experiences:
  - Brazil (Q4 2008): net capital outflows of $15½ billion (about 0.9 percent of 2008 GDP).
  - South Africa (Q4 2008): portfolio outflows contributed $7½ billion in nonresident capital outflows, but resident inflows and errors and omissions more than offset them; South Africa did not experience net outflows in any quarter during 2008–09.
- Characteristics of a South African adverse scenario:
  - Amid weak growth and compressed imports, the current account deficit has narrowed; in 2017 the trade and services balance was positive at 1.4 percent of GDP.
  - The current account deficit in 2017 resulted from deficits on the income and transfer accounts amounting to 3 and 0.8 percent of GDP, respectively.
  - Significant capital outflows could prompt a decline in imports and a short-term improvement in the income account due to reduced nonresident holdings of assets, but external debt servicing costs could rise markedly given significant gross external financing needs.

### Concluding Remarks and Policy Recommendations
- Assessment of resilience and vulnerabilities:
  - The South African economy is resilient and has weathered bouts of volatility in 2017, aided by the flexible exchange rate and a large financial sector providing a domestic investor base for government bonds.
  - Vulnerabilities exist in both domestic and external sectors: deteriorated domestic public-sector finances over the past decade, rising fiscal risks from SOEs, significant sovereign-financial linkages, and large external financing needs with current account deficit financed by flows subject to sudden reversals.
- Policy priorities to strengthen resilience:
  - Implement structural reforms to lift growth and reduce vulnerabilities.
  - Pursue fiscal consolidation to rebuild policy buffers.
  - Improve the business environment to encourage private investment and support job creation.
  - Strengthen public-sector balance sheets and maintain an effective social safety net.
  - Modernize infrastructure to support investment.
  - As opportunities arise and subject to meeting the inflation mandate, accumulate international reserves toward adequate levels to strengthen buffers.

*Source: IMF staff summary of chapter titled "While vulnerabilities expose the South African economy to risks, mitigating factors and existing buffers can help cushion the impact of adverse shocks, should they materialize."*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2018/cr18247.pdf_
