## 1. Relationship Between Real GDP Growth and Inflation

## Source details

**Canonical URL:** [1. Relationship Between Real GDP Growth and Inflation](https://www.imf.org/-/media/files/publications/cr/2020/english/1zafea2020002.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2020/english/1zafea2020002.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2020/english/1zafea2020002.pdf.json)

---

### Introduction
- Real GDP growth (year on year) fell from 4.3 percent during 2000–07 to 2 percent during 2010–17 and was 0.8 percent in 2018.
- Structural rigidities depressed business confidence, private investment, and growth; monetary and fiscal policy provided support but could not fully offset these constraints.
- The SARB’s monthly CPI target is influenced by exchange rate performance; the rand is relatively volatile and SARB follows a policy of no foreign exchange intervention.
- Exchange rate pass-through to core inflation is estimated at around 10–15 percent.
- Monetary policy can influence inflation via the exchange rate and credibility channels; inflation expectations of businesses and trade unions closely follow inflation outturns with a lag.

### The Growth-Inflation Phillips Curve
- Historical pattern of the slope (coefficient of inflation on GDP growth):
  - 1970s: relationship around zero.
  - 1980s: slope turned negative.
  - 1990s: negative slope steepened markedly.
  - 2000s and 2010s: slope flattened; lower demand growth associated with more muted increases in inflation.
- Interpretations:
  - Growth from productivity improvements is disinflationary; structural reforms can boost productivity and growth and help anchor inflation.
  - Alternatively, inflation may dampen long-run growth, implying central bank priority on low and stable inflation.

### Econometric Analysis — Baseline Model and Results
- Model specification:
  - Four variables: (1) real domestic demand growth; (2) survey-based one-year-ahead inflation expectations (average for financial analysts, businesses, and trade unions); (3) real policy interest rate (deflated by inflation expectations); (4) nominal effective exchange rate (NEER).
  - Quarterly VAR estimated for 2000Q3–18Q2.
  - Variables detrended except inflation expectations; ordering: domestic demand growth, inflation expectations, real policy rate, NEER growth.
  - Model estimated separately for 2000Q3–2008Q2 and 2010Q3–2018Q2; three lags; generalized impulses reported.
- Key quantitative findings:
  - A 1 percentage point increase in the real policy rate is associated with a 2 percentage point decline in domestic demand growth during the 2000s (2000Q3–2008Q2).
  - The same 1 percentage point increase does not generate a systematic impact on domestic demand growth during the 2010s (2010Q3–2018Q2).
  - Monetary policy action reduces inflation expectations in both periods:
    - Impact during the 2000s: 1–1.5 percentage points reduction in inflation expectations.
    - Impact during the 2010s: 0.5 percentage point reduction in inflation expectations.
  - Comparison: Ramcharan (2008) finds a 1 percentage point increase in the nominal policy rate is associated with a 0.8 percentage point decline in GDP growth two quarters later.

### Robustness Checks
- Threshold effects:
  - Sample split into “high” and “low” real policy rate regimes using a threshold of 1.5 percent.
  - Findings: In the “high” real rate regime both demand growth and inflation expectations respond to monetary policy action; in the “low” real rate regime only inflation expectations respond.
- Role of core inflation:
  - Replacing inflation expectations with detrended realized core inflation (estimated for the 2010s only) shows core inflation declines with a lag after monetary tightening, but the relation is statistically insignificant.
- Overall interpretation: weakening of monetary policy transmission to demand growth and to core inflation, while inflation expectations remain responsive (albeit less than before).

### Conclusion and Policy Implications (Inflation–Growth section)
- Main conclusions:
  - During the 2010s, domestic demand growth responded little to monetary policy action; monetary policy transmission through demand has weakened.
  - Monetary policy continues to influence inflation expectations, though the effect diminished from the 2000s to the 2010s.
  - Exchange rate and credibility channels appear to remain operational.
  - The constrained impact of monetary policy on growth reflects underlying structural constraints that need to be addressed.
- Policy recommendation:
  - The SARB should continue efforts to anchor inflation and inflation expectations at a lower level while structural constraints to economic growth are removed.

*Prepared by Ken Miyajima; reviewed by Ana Lucía Coronel.*

### Main findings and policy implications (Rand volatility and investment)
- Policy tradeoffs related to a volatile rand are limited, even though SMEs suffer more.
- Rand volatility has some impact on investment for SMEs, and cross-border capital inflows, but these effects are generally relatively small, economically.
- To alleviate potential negative impact of rand volatility on SME investment, greater policy certainty and bank competition would reduce rand volatility and the cost of hedging for exchange rate risk.

### Stylized facts about rand volatility
- Rand volatility:
  - Relatively volatile in nominal effective terms and in absolute size of monthly movements; comparable to the Brazilian real and Turkish lira; Russian ruble and Argentine peso registered even larger movements.
  - Historically, rand weakened more than 70 percent year on year against the dollar in the mid-1980s during a debt standstill.
  - Currency pair weakened around 40–50 percent year on year in 2001, 2008 and 2015.
- Since 2003:
  - SARB’s forward position was short until 2003, with stock of short positions up to 20 percent of GDP through the late-1990s; after adoption of inflation targeting, short forward positions declined.
  - Gross official reserves increased gradually as the SARB bought foreign inflows opportunistically.
  - South Africa has used the rand, rather than official reserves, to absorb shocks—rand volatility is relatively high while volatility of official reserves relative to M1 is one of the lowest among EMs.
- Drivers of rand volatility:
  - Domestic: policy and political uncertainty increased rand volatility since late-2015; rand volatility rose and stayed above the VIX much of 2016 and 2017.
  - External: rand volatility generally co-moves with the VIX; the rand trades in large volumes globally and South Africa’s daily currency turnover in global markets scaled by official reserves is by far the highest among major EMs as the rand is traded as an EM proxy.
  - Nonresident holdings of local assets are large, increasing the impact of external shocks.
- Distributional link: periods of relatively high rand volatility (one standard deviation or more above mean) tend to be accompanied by relatively large rand depreciation.

### South Africa’s resilience to rand volatility and depreciation
- Balance sheets and hedges:
  - FX debt as a share of GDP in South Africa is intermediate compared to other EMs, limiting concerns about FX mismatches.
  - SOEs are mandated to hedge and large corporates that borrow in FX tend to have natural hedges or FX revenues; survey results indicate corporates, including some SMEs, tend to hedge exchange rate risk.
- Fiscal sector:
  - Historically a 10 percent rand movement in year-on-year terms would change the debt-to-GDP ratio by 0.4 percentage points, reflecting that South Africa's debt is largely denominated in rand and fiscal revenue/expenditure are mainly in local currency.
  - Long average maturities help reduce rollover pressures when rand depreciation raises the rand value of FX bonds.
- Banking system:
  - South African banks borrow and lend mainly in rand, making them little affected by rand movements.
  - Domestic banks are well capitalized, profitable, and have relatively low NPLs.
  - Some small banks have relatively high FX debt as a share of total assets, which could put those banks under pressure and transmit confidence effects to the broader system.
- Inflation pass-through:
  - Exchange rate pass-through to headline inflation declined from around 50 percent in the late-1990s to 20–25 percent in more recent years.
  - Forbes et al (2017) find pass-through fell from 15 percent in 2004–09 to 6 percent in 2010–15.
  - Staff analysis shows pass-through of rand volatility to inflation is relatively small.
- Domestic investor base:
  - South Africa’s local investor base as a share of GDP is one of the largest in EMs and has tended to reduce asset price volatility by buying local assets when valuations became more attractive.

### Impact of rand volatility on private investment
- Channels:
  - Higher volatility increases uncertainty and the risk premium, discouraging investment—especially for firms sensitive to uncertainty.
  - Higher currency volatility and depreciation negatively affect investors with currency mismatches; SMEs are more vulnerable due to financial constraints and less hedging access.
- Data and approach:
  - Investment measures: firm-level capital expenditure (capital stock accumulation) and non-resident portfolio investment (weekly net nonresident purchases, January 2000–April 2019, in billions of rand).
  - Firm-level data: 704 listed firms in South Africa for 2000–17 from Worldscope at 3-digit SIC level; median firm size separates SMEs from large firms.
  - Sector-level sensitivity to uncertainty constructed from US firms following Tong and Wei (2019) method.
- Firm investment results:
  - Higher rand volatility significantly reduces SME investment but has little impact on investment of large firms.
  - For an SME with sensitivity to uncertainty at the 90 percent threshold (0.046), an inter-quartile increase of rand volatility of 0.6 leads to a decline in the ratio of capital stock to assets by 1 percentage point.
  - Inter-quartile change of the capital/asset ratio is 3.4 percentage points.
  - Result may underestimate impact if smaller SMEs not in the dataset are more averse to uncertainty.
- Cross-border portfolio flows results:
  - Rand volatility negatively affects total portfolio net inflows, driven mainly by reduced equity inflows and to a lesser extent bond inflows.
  - Economic magnitude: an interquartile increase of rand volatility of 0.5 would reduce equity flows by R250 million.
  - Benchmarks:
    - One standard deviation of weekly net equity flows is R2.4 billion.
    - Over the past year, R4 billion worth of net portfolio flows left the country every week, on average.
    - SARB’s international reserve holdings are more than R700 billion.
  - Interpretation: impact on equity flows is statistically significant but not large economically.

### Conclusion (Rand volatility section)
- Analysis supports staff’s view that while rand volatility affects SMEs and cross-border portfolio flows, overall policy tradeoffs are limited.
- Policy measures to reduce negative effects on SMEs include enhancing policy certainty and promoting bank competition to lower rand volatility and hedging costs.

*Prepared by Ken Miyajima and Hui Tong; reviewed by Ana Lucía Coronel.*

### Growth performance: stylized facts and drivers
- Since 2007 South Africa’s growth performance has been persistently weaker than in other emerging markets due to a deteriorating business environment responding to weakening institutions, reduced competition, and low public spending efficiency, compounded by labor market rigidities.
- Real per capita GDP growth:
  - Averaged 1.2 percent since 1994 for South Africa.
  - Comparator EM median: 2.8 percent.
  - EM top performers: 5 percent.
- Unemployment:
  - Overall: about 29 percent.
  - Youth unemployment: 56 percent.
  - 40 percent of the unemployed (60 percent of the youth) have never had a job.
  - One third of the unemployed have not been employed for 5 years.
- Income from employment:
  - Accounts for 40 percent of total income for poor households.
  - Accounts for 80 percent for households that move out of poverty.

### Growth composition and trends
- Private investment and TFP:
  - Private investment’s contribution to growth was about half of the median EM and 40 percent of the top growth performers in 1994–2018.
  - Contribution of TFP was less than half of the median EM and one third of the top growth performers.
  - Change post-2007:
    - TFP growth: Above median EM in 1994–2007; negative and lower than the median in 2008–18.
    - Composition of investment: Private investment contribution fell by 30 percent while public investment doubled its contribution.
- Public investment and government role:
  - Public investment represented 31 percent of total investment in South Africa on average (1994–2018) versus 22 percent in the median EM and 24 percent in top growth performers.
  - Public investment peaked in 2009 at 36.7 percent (peak share noted).
  - Government expenditure increased by 6 percentage points of GDP in 2008–18.
  - In 1994–2007, government expenditure was on average about 2 percentage points of GDP above the median EM.
- Debt dynamics:
  - By end-2018, South Africa’s debt surpassed the median of EM comparators and top growth performers.

### Sectoral structure and external performance
- Sectoral shifts:
  - Share of private non-tradable sectors in total value added increased by about 14 percentage points since 1994, largely at the expense of tradable sectors.
  - Within tradable sectors, mining was the worst performing sub-sector; finance, business services, and real estate were top performers.
- Exports:
  - South Africa’s exports have remained broadly stagnant as a share of world goods exports, contrasting with increasing trends in the median EM and fastest growing EMs.
  - Export composition: importance of minerals and manufacturing sectors broadly unchanged.
  - Export complexity: South Africa has one of the least complex export structures among EMs per Growth Lab export complexity index.

### Demand-side contributions
- Consumption provides the greatest contribution to growth during 1994–2018.
- Fixed investment and net exports made weaker contributions relative to consumption.

### Growth diagnostics methodology and main constraints
- Methodology: growth diagnostics approach by Hausmann et al. (2008) complemented by a quantification exercise drawing from Prati et al. (2013) and Egert and Gal (2016).
- Hypotheses for weak private investment:
  1. Microeconomic features (market regulations, labor market legislation, governance weaknesses).
  2. Macroeconomic factors (unsustainable debt, high and volatile inflation, political risks, conflict).
  3. Inadequate provision/quality of complementary inputs (infrastructure, education, health, security).
  4. Market failure (coordination / information externalities).
  5. Financing constraints (national savings or international market access).
  6. Financial intermediation issues (high lending spreads and borrowing costs).
- Quantification exercise:
  - (1) Panel regressions for structural indicators on capital/labor/productivity across 184 countries (2000–18).
  - (2) Simulating improvement to 75th percentile.
  - (3) Aggregating component indicators for each reform.
- Main diagnostic result:
  - Access to finance or cost of financing were not binding constraints to private investment for most of 1994–2018.
  - Private savings exceeded private investment during most years excluding 2004–08 (except 2008–09).
  - Abundant access to international finance was available to cover gaps, evidenced by wide current account deficits that largely financed the general government and SOEs.
  - South Africa has a large pool of financial savings through pension funds, insurance companies, and investment banks (more than 200 percent of GDP).

### Financial intermediation, returns, and business environment
- Lending conditions:
  - Lending spread close to EM median before 2007 but below thereafter.
  - Real lending rate has continued to decline in line with the EM median for the most part.
- FDI and returns:
  - Five-year moving average of FDI consistently below the median level of other EMs.
  - Decline in FDI after 2009 was slightly above 50 percent compared to 20 percent in the median EM.
  - South African corporates increasingly sought investment opportunities abroad, suggesting low rate of return to private investment.
- Market failures and export structure:
  - Correlation coefficient between GDP per capita and the terms of trade is weak (0.2).
  - Export complexity is relatively low.
- Business environment deterioration:
  - Significant deterioration in Global Competitiveness indicators (GCI) and Doing Business Indicators (DBI) scores coincided with the decline in private investment after 2007.
  - Main areas of deterioration: macroeconomic environment, higher education and training, goods market efficiency, institutions, and infrastructure.
  - Long-standing relative weaknesses: health and primary education and labor market efficiency.
  - Security is a long-standing institutional weakness, adversely affecting cost of doing business, especially for SMEs and tourism.

### Per-capita growth gains from reforms and policy priorities
- Estimated gains:
  - Product market reforms: slightly above 1 percentage point to per-capita growth once complete.
  - Improving the macroeconomic environment: about 0.7 percentage points to per-capita growth.
  - Complementary factors and labor market reforms: combined payoff over time could be potentially as large as macro and product market reforms but may take longer.
- Policy priorities (three-pronged approach):
  - Reduce micro distortions: Product Markets, Labor Markets.
  - Accelerate reforms to strengthen institutions.
  - Improve provision of complementary inputs and their quality (e.g., health, education).
  - Consolidate public finances.
  - Reduce policy and regulatory uncertainty.
- Rationale: product market and macroeconomic environment reforms are feasible near-term priorities with measurable per-capita growth gains and can help mobilize broader reform momentum.

### Box 1 — Main issues highlighted by GCI, 2008–2018 (key deviations)
- Institutional weakening: from 0.7 standard deviations above the median to 0.4 below the median on average.
- Property rights and investor protection: from 1.8 standard deviations above the median to 0.8 above the median on average.
- Transparency in policy making: from 1.2 standard deviations above the median to slightly below the median on average.
- Competition indicators: from 0.7 standard deviations above the median to 0.1 standard deviations below the median on average.
- Quality of infrastructure (SOE domination): from 1.1 standard deviations above the median to 0.1 above the median on average.
- Labor market indicators: despite a 0.5 standard deviation improvement, remain on average 1.3 standard deviations below the EM median in 2018.
- Higher education provision: from 1.4 standard deviations above the EM median to 0.3 above the EM median on average.
- Health provision: despite a 0.6 standard deviation improvement, indicators remained 2.1 standard deviations below the EM median in 2018 on average.
- Security provision: indicators remained on average 1.1 standard deviations below the EM median.

### Annex I — Crime and business impact (key points)
- Homicide rates compare unfavorably to EMs and satisfaction with law enforcement has declined.
- Crime is reported as a major obstacle in business surveys, particularly affecting SMEs and start-ups.
- Incidence of crime is higher than in other regions and is especially costly for smaller firms.
- Crime has damaging effects on tourism inflows.

*Source: IMF staff chapter content (South Africa).*

### 1. Relationship Between Real GDP Growth and Inflation __________________________________ 4

### 1. Relationship Between Real GDP Growth and Inflation

### Introduction
- South Africa experienced a marked slowdown in real GDP growth after the Global Financial Crisis:
  - Real GDP growth (year on year) fell from 4.3 percent during 2000–07 to 2 percent during 2010–17 and was 0.8 percent in 2018.
- Structural rigidities depressed business confidence, private investment, and growth; monetary and fiscal policy provided support but could not fully offset these constraints.
- The SARB’s monthly CPI target is influenced by exchange rate performance; the rand is relatively volatile and SARB follows a policy of no foreign exchange intervention.
- Exchange rate pass-through to core inflation is estimated at around 10–15 percent.
- Monetary policy can influence inflation via the exchange rate and credibility channels; inflation expectations of businesses and trade unions closely follow inflation outturns with a lag.

### The Growth-Inflation Phillips Curve
- Historical pattern of the slope (coefficient of inflation on GDP growth):
  - 1970s: relationship around zero.
  - 1980s: slope turned negative.
  - 1990s: negative slope steepened markedly.
  - 2000s and 2010s: slope flattened; lower demand growth associated with more muted increases in inflation.
- Interpretations of the observed negative relationship between higher GDP growth and lower inflation:
  - Growth stemming from productivity improvements is disinflationary; structural reforms can boost productivity and growth and help anchor inflation.
  - Alternatively, inflation may dampen long-run growth, suggesting the central bank should maintain low and stable inflation.

### Econometric Analysis — Baseline Model and Results
- Model specifics:
  - Four variables used: (1) real domestic demand growth; (2) survey-based one-year-ahead inflation expectations (average for financial analysts, businesses, and trade unions); (3) real policy interest rate (deflated by inflation expectations); (4) nominal effective exchange rate (NEER).
  - Quarterly VAR estimated for 2000Q3–18Q2.
  - Variables detrended except inflation expectations; ordering: domestic demand growth, inflation expectations, real policy rate, NEER growth.
  - Model estimated separately for 2000Q3–2008Q2 and 2010Q3–2018Q2 to assess pre- and post-2008 changes; three lags; generalized impulses reported.
- Key quantitative findings:
  - A 1 percentage point increase in the real policy rate is associated with a 2 percentage point decline in domestic demand growth during the 2000s (2000Q3–2008Q2).
  - The same 1 percentage point increase does not generate a systematic impact on domestic demand growth during the 2010s (2010Q3–2018Q2).
  - Monetary policy action reduces inflation expectations in both periods:
    - Impact during the 2000s: 1–1.5 percentage points reduction in inflation expectations.
    - Impact during the 2010s: 0.5 percentage point reduction in inflation expectations.
  - Comparison with literature: Ramcharan (2008) finds a 1 percentage point increase in the nominal policy rate is associated with a 0.8 percentage point decline in GDP growth two quarters later.

### Robustness Checks
- Threshold effects:
  - Sample split into “high” and “low” real policy rate regimes using a threshold of 1.5 percent.
  - Findings: In the “high” real rate regime both demand growth and inflation expectations respond to monetary policy action; in the “low” real rate regime only inflation expectations respond.
- Role of core inflation:
  - Replacing inflation expectations with detrended realized core inflation (estimated for the 2010s only) shows core inflation declines with a lag after monetary tightening, but the relation is statistically insignificant.
- Overall interpretation of robustness checks: results confirm a weakening of monetary policy transmission to demand growth and to core inflation, while inflation expectations remain responsive (albeit less than before).

### Conclusion and Policy Implications
- Main conclusions:
  - During the 2010s, domestic demand growth responded little to monetary policy action; monetary policy transmission through demand has weakened.
  - Monetary policy continues to influence inflation expectations, though the effect diminished from the 2000s to the 2010s.
  - Exchange rate and credibility channels appear to remain operational.
  - The constrained impact of monetary policy on growth reflects underlying structural constraints that need to be addressed.
- Policy recommendation:
  - The SARB should continue efforts to anchor inflation and inflation expectations at a lower level while structural constraints to economic growth are removed.

*Prepared by Ken Miyajima; reviewed by Ana Lucía Coronel.*

### 4.      The main findings of this analysis continue to support staff’s view that the policy

### 4.      The main findings of this analysis continue to support staff’s view that the policy

### Main findings and policy implications
- Policy tradeoffs related to a volatile rand are limited, even though SMEs suffer more.
- Rand volatility has some impact on investment for SMEs, and cross-border capital inflows, but these effects are generally relatively small, economically.
- To alleviate the potential negative impact of rand volatility on investment by SMEs, greater policy certainty and bank competition would reduce rand volatility and the cost of hedging for exchange rate risk.

### Structure of the note
- Section B: Stylized facts about rand volatility.
- Section C: South Africa’s resilience to relatively large rand volatility.
- Section D: Econometric analysis of the potential impact of rand volatility on private investment.
- Section E: Conclusion.

### Stylized facts about rand volatility
- The rand has been relatively volatile in nominal effective terms and in absolute size of monthly movements, comparable to the Brazilian real and Turkish lira; the Russian ruble and Argentine peso have registered even larger movements.
- Historically, the rand weakened more than 70 percent year on year against the dollar in the mid-1980s during a debt standstill.
- The currency pair weakened around 40–50 percent year on year in 2001, 2008 and 2015.
- Since 2003 rand volatility has been determined by market forces:
  - The SARB’s forward position was short until 2003, with stock of short positions up to 20 percent of GDP through the late-1990s; after adoption of inflation targeting, short forward positions declined.
  - Gross official reserves increased gradually as the SARB bought foreign inflows opportunistically.
  - South Africa has used the rand, rather than official reserves, to absorb shocks—rand volatility is relatively high while volatility of official reserves relative to M1 is one of the lowest among EMs.
- Drivers of rand volatility:
  - Domestic: policy and political uncertainty increased rand volatility since late-2015; rand volatility rose and stayed above the VIX much of 2016 and 2017.
  - External: rand volatility generally co-moves with the VIX; the rand trades in large volumes globally and South Africa’s daily currency turnover in global markets scaled by official reserves is by far the highest among major EMs as the rand is traded as an EM proxy.
  - Nonresident holdings of local assets are large, increasing the impact of external shocks.
- Distributional link: periods of relatively high rand volatility (one standard deviation or more above mean) tend to be accompanied by relatively large rand depreciation.

### South Africa’s resilience to rand volatility and depreciation
- Balance sheets and hedges:
  - FX debt as a share of GDP in South Africa is intermediate compared to other EMs, limiting concerns about FX mismatches.
  - SOEs are mandated to hedge and large corporates that borrow in FX tend to have natural hedges or FX revenues; survey results indicate corporates, including some SMEs, tend to hedge exchange rate risk.
- Fiscal sector:
  - Historically a 10 percent rand movement in year-on-year terms would change the debt-to-GDP ratio by 0.4 percentage points, reflecting that South Africa's debt is largely denominated in rand and fiscal revenue/expenditure are mainly in local currency.
  - Long average maturities help reduce rollover pressures when rand depreciation raises the rand value of FX bonds.
- Banking system:
  - South African banks borrow and lend mainly in rand, making them little affected by rand movements.
  - Domestic banks are well capitalized, profitable, and have relatively low NPLs, providing buffers to absorb shocks.
  - System-wide resilience masks diversity: some small banks have relatively high FX debt as a share of total assets, which could put those banks under pressure and transmit confidence effects to the broader system.
- Inflation pass-through:
  - Exchange rate pass-through to headline inflation declined from around 50 percent in the late-1990s to 20–25 percent in more recent years (Kabundi and Mlachila, 2019).
  - Forbes et al (2017) find pass-through fell from 15 percent in 2004–09 to 6 percent in 2010–15.
  - Staff analysis shows pass-through of rand volatility to inflation is relatively small.
- Domestic investor base:
  - South Africa’s local investor base as a share of GDP is one of the largest in EMs and has tended to reduce asset price volatility by buying local assets when valuations became more attractive.

### Impact of rand volatility on private investment
- Channels:
  - Higher volatility increases uncertainty and the risk premium, discouraging investment—especially for firms sensitive to uncertainty.
  - Higher currency volatility and depreciation negatively affect investors with currency mismatches, reducing resources available for investment; SMEs are more vulnerable due to financial constraints and less hedging access.
- Data and approach:
  - Two investment measures analyzed: domestic firm-level capital expenditure (firm-level capital stock accumulation) and non-resident portfolio investment (weekly net nonresident purchases, January 2000–April 2019, in billions of rand).
  - Firm-level data: 704 listed firms in South Africa for 2000–17 from Worldscope at 3-digit SIC level; median firm size separates SMEs from large firms.
  - Sector-level sensitivity to uncertainty constructed from US firms following Tong and Wei (2019) method.
- Firm investment results:
  - Higher rand volatility significantly reduces SME investment but has little impact on investment of large firms.
  - For an SME with sensitivity to uncertainty at the 90 percent threshold (0.046), an inter-quartile increase of rand volatility of 0.6 leads to a decline in the ratio of capital stock to assets by 1 percentage point.
  - Comparison: inter-quartile change of the capital/asset ratio is 3.4 percentage points.
  - Result may underestimate impact if smaller SMEs not in the dataset are more averse to uncertainty.
- Cross-border portfolio flows results:
  - Rand volatility negatively affects total portfolio net inflows, driven mainly by reduced equity inflows and to a lesser extent bond inflows.
  - Economic magnitude: an interquartile increase of rand volatility of 0.5 would reduce equity flows by R250 million.
  - Benchmarks for scale:
    - One standard deviation of weekly net equity flows is R2.4 billion (10 times larger than R250 million).
    - Over the past year, R4 billion worth of net portfolio flows left the country every week, on average.
    - SARB’s international reserve holdings are more than R700 billion.
  - Interpretation: impact on equity flows is statistically significant but not large economically, suggesting multiple factors affect equity flows.

### Conclusion (Section E)
- The analysis supports staff’s view that while rand volatility affects SMEs and cross-border portfolio flows, overall policy tradeoffs are limited.
- Policy measures to reduce the negative effects on SMEs include enhancing policy certainty and promoting bank competition to lower rand volatility and hedging costs.

*Prepared by Ken Miyajima and Hui Tong; reviewed by Ana Lucía Coronel.*

### 22.      Findings of this study suggest that rand volatility tends to have limited policy tradeoffs,

### 22.      Findings of this study suggest that rand volatility tends to have limited policy tradeoffs

### Rand volatility: main findings
- Rand volatility tends to have limited policy tradeoffs, even though SMEs tend to be adversely affected.
- The impact of rand volatility on economic indicators such as firm investment and portfolio flows is relatively small.
- Some subsectors are more vulnerable to rand volatility than others, notably SMEs.
- To reduce rand volatility affecting investment in vulnerable sectors, greater policy certainty is needed through:
  - Clearer communication of the policy agenda.
  - Decisive implementation of announced measures.
- It is critical to avoid measures that could negatively affect private investment and economic growth.
- Lower risk premium, in terms of inflation and sovereign, would reduce interest rate differentials with trading partners and the attendant premium for hedging against expected rand depreciation.
- Maintain the SARB’s policy of allowing the rand to freely absorb shocks while preserving its international reserve buffers.

### Empirical evidence on firm investment (Table 1)
- Estimation: Rand volatility * sensitivity to uncertainty
  - All firms: -0.31**  (standard error [0.15])
  - Small/Medium firm: -0.66***  (standard error [0.26])
  - Large firm: 0.054  (standard error [0.17])
- Fixed effects: Firm fixed effects Yes; Year fixed effects Yes
- Observations:
  - All firms: 5708
  - Small/Medium firm: 2,948
  - Large firm: 2760
- Number of firms:
  - All firms: 704
  - Small/Medium firm: 429
  - Large firm: 275
- R-squared: 0.02 for all columns

### Empirical evidence on portfolio inflows (Table 2)
- Dependent variables: total inflow; equity inflow; bond inflow
- % change of rand (lag):
  - total inflow: -19.3***  (standard error [5.34])
  - equity inflow: -2.82  (standard error [3.24])
  - bond inflow: -16.5***  (standard error [4.06])
- rand volatility:
  - total inflow: -0.51*  (standard error [0.28])
  - equity inflow: -0.51***  (standard error [0.17])
  - bond inflow: 0.0011  (standard error [0.21])
- Constant terms:
  - total inflow: 0.81**  (standard error [0.32])
  - equity inflow: 0.58***  (standard error [0.19])
  - bond inflow: 0.22  (standard error [0.24])
- Observations: 1,153 for all columns
- R-squared:
  - total inflow: 0.014
  - equity inflow: 0.009
  - bond inflow: 0.014
- Significance notation: *** p<0.01, ** p<0.05, * p<0.1
- Sources for Figure 13 and tables: BIS, Haver, and IMF staff calculations.

### Growth performance: stylized facts and drivers
- Since 2007 South Africa’s growth performance has been persistently weaker than in other emerging markets due to a deteriorating business environment responding to weakening institutions, reduced competition, and low public spending efficiency, compounded by labor market rigidities.
- Real per capita GDP growth:
  - Averaged 1.2 percent since 1994 for South Africa.
  - Comparator EM median: 2.8 percent.
  - EM top performers: 5 percent.
- Unemployment:
  - Overall: about 29 percent.
  - Youth unemployment: 56 percent.
  - 40 percent of the unemployed (60 percent of the youth) have never had a job.
  - One third of the unemployed have not been employed for 5 years.
- Income from employment:
  - Accounts for 40 percent of total income for poor households.
  - Accounts for 80 percent for households that move out of poverty.

### Growth composition and trends
- Private investment and TFP:
  - Private investment’s contribution to growth was about half of the median EM and 40 percent of the top growth performers in 1994–2018.
  - Contribution of TFP was less than half of the median EM and one third of the top growth performers.
- Change post-2007:
  - TFP growth: Above median EM in 1994–2007; negative and lower than the median in 2008–18.
  - Composition of investment: Private investment contribution fell by 30 percent while public investment doubled its contribution.
- Public investment and government role:
  - Public investment represented 31 percent of total investment in South Africa on average (1994–2018) versus 22 percent in the median EM and 24 percent in top growth performers.
  - Public investment increased since 2007, peaked in 2009 at 36.7 percent (peak share noted), and declined in recent years.
  - Government expenditure increased by 6 percentage points of GDP in 2008–18.
  - In 1994–2007, government expenditure was on average about 2 percentage points of GDP above the median EM.
- Debt dynamics:
  - By end-2018, South Africa’s debt surpassed the median of EM comparators and top growth performers.

### Sectoral structure and external performance
- Sectoral shifts:
  - Share of private non-tradable sectors in total value added increased by about 14 percentage points since 1994, largely at the expense of tradable sectors.
  - Within tradable sectors, mining was the worst performing sub-sector; finance, business services, and real estate were top performers.
- Relative price of tradables exhibited a declining trend (index: 2010=100 reference noted).
- Exports:
  - South Africa’s exports have remained broadly stagnant as a share of world goods exports, contrasting with increasing trends in the median EM and fastest growing EMs.
  - Export composition: importance of minerals and manufacturing sectors broadly unchanged.
  - Export complexity: South Africa has one of the least complex export structures among EMs per Growth Lab export complexity index.

### Demand-side contributions
- Consumption provides the greatest contribution to growth during 1994–2018.
- Fixed investment and net exports made weaker contributions relative to consumption.

### Growth diagnostics methodology and main constraints
- Methodology based on growth diagnostics approach by Hausmann et al. (2008) and complemented by a quantification exercise drawing from Prati et al. (2013) and Egert and Gal (2016).
- Hypotheses considered for weak private investment:
  1. Microeconomic features (market regulations, labor market legislation, governance weaknesses).
  2. Macroeconomic factors (unsustainable debt, high and volatile inflation, political risks, conflict).
  3. Inadequate provision/quality of complementary inputs (infrastructure, education, health, security).
  4. Market failure (coordination / information externalities).
  5. Financing constraints (national savings or international market access).
  6. Financial intermediation issues (high lending spreads and borrowing costs).
- Quantification exercise: impact of reforms on per capita GDP growth estimated by (1) panel regressions for structural indicators on capital/labor/productivity across 184 countries (2000–18), (2) simulating improvement to 75th percentile, and (3) aggregating component indicators for each reform.
- Main result from diagnostics:
  - Access to finance or cost of financing were not binding constraints to private investment for most of 1994–2018.
  - Private savings exceeded private investment during most years excluding 2004–08 (except 2008–09).
  - Abundant access to international finance was available to cover gaps, evidenced by wide current account deficits that largely financed the general government and SOEs.
  - South Africa has a large pool of financial savings through pension funds, insurance companies, and investment banks (more than 200 percent of GDP).

*Source: IMF staff chapter content (South Africa).*

### 15.      Financial intermediation was not a major obstacle either.

### 15.      Financial intermediation was not a major obstacle either.

### Financial intermediation and lending conditions
- The lending spread had been close to the EM median before 2007 but below thereafter.
- After 2007, the real lending rate has continued to decline in line with the EM median for the most part.
- Even when the real lending rate was higher than the EM median, growth was considerably faster than in recent years.
- Note: access to international finance does not appear to have been a binding constraint (current account deficits persistent despite slowing growth).

### Foreign direct investment and returns to private investment
- The five-year moving average of FDI has been consistently below the median level of other EMs.
- The decline in FDI that followed 2009 was slightly above 50 percent compared to 20 percent in the median EM despite already low levels.
- South African corporates increasingly started to seek investment opportunities abroad.
- Interpretation: these developments suggest that a low rate of return to private investment has been a long-standing problem, worsening after 2009.

### Evidence on market failures and export structure
- Evidence that low returns to private investment are a result of market failures is limited.
- Countries where this problem is prevalent are expected to have significant correlations between GDP per capita and terms of trade; for South Africa the correlation coefficient between GDP per capita and the terms of trade is weak (0.2).
- South Africa’s export complexity is relatively low.

### Business environment and binding constraints
- A business environment not conducive to investment is the most likely factor underlying low and deteriorating rates of return to private investment.
- A significant deterioration in South Africa’s Global Competitiveness indicators (GCI) and Doing Business Indicators (DBI) scores coincided with the protracted decline in private investment after growth peaked in 2007.
- Main areas of deterioration in the GCI categories: the macroeconomic environment, higher education and training, goods market efficiency, and institutions.
- Deterioration in the infrastructure category is also observed.
- Most significant long-standing relative weaknesses compared with other EMs: health and primary education and labor market efficiency.
- Within institutions, a long-standing relative weakness in security is notable and is confirmed by other data pointing to adverse impacts on the cost of doing business, especially on SMEs and tourism.
- Complementary findings from literature and National Treasury (2019): political and policy uncertainty, institutional weaknesses, poorly managed state-owned companies, and barriers to entry that distort product markets are important factors contributing to the low growth environment.

### Per-capita growth gains from reform estimates
- Product market reforms are estimated to add slightly above 1 percentage point to per-capita growth once complete.
- Improving the macroeconomic environment would yield about 0.7 percentage points to per-capita growth.
- The distance to the frontier is lower for product market and macroeconomic environment reforms than for labor markets and complementary reforms.
- Complementary factors and labor market reforms: combined growth payoff over time could be potentially as large as those from macroeconomic environment and product market reforms, but may be more difficult to achieve and take longer.
- Successful implementation of product market and macroeconomic environment reforms can help create reform momentum to advance other reforms.

### Policy priorities and three-pronged approach
- Priority areas identified:
  - Reduce micro distortions: Product Markets, Labor Markets.
  - Accelerate reforms to strengthen institutions.
  - Improve provision of complementary inputs and their quality (e.g., health, education).
  - Consolidate public finances.
  - Reduce policy and regulatory uncertainty.
- Rationale: product market and macroeconomic environment reforms are feasible near-term priorities with measurable per-capita growth gains and can help mobilize broader reform momentum.

### Box 1 — Main issues highlighted by GCI, 2008–2018
- Institutional weakening
  - Issues: Diversion of public funds; Public trust in politicians; Favoritism in public decisions by public officials; Efficiency of government spending.
  - Deviation: Change from 0.7 standard deviations above the median to 0.4 below the median, on average.
  - Property rights; Strength of auditing and reporting standards; Strength of investor protection: Change from 1.8 standard deviations above the median to 0.8 above the median, on average.
- Increasing policy and regulatory uncertainty
  - Indicator: Transparency in policy making.
  - Deviation: Change from 1.2 standard deviations above the median to slightly below the median, on average.
- Lack of competition
  - Issues: Number of days to start a business; Prevalence of non-tariff barriers; Extent of market dominance; Business impact of rules of FDI; Effectiveness of anti-monopoly policies.
  - Deviation: Change from 0.7 standard deviations above the median to 0.1 standard deviations below the median, on average.
- SOE domination of network industries
  - Issue: Quality of infrastructure.
  - Deviation: Change from 1.1 standard deviations above the median to 0.1 above the median, on average.
- Labor market rigidities
  - Issues: Cooperation in labor employer relations; Flexibility in wage determination; Hiring and firing practices; Pay and productivity.
  - Deviation: Despite an improvement of 0.5 standard deviations, indicators remained, on average, below 1.3 standard deviations below the EM median in 2018.
- Deficiencies in higher education provision
  - Issues: Extent of staff training; Availability of research and training services.
  - Deviation: Change from 1.4 standard deviations above the EM median to 0.3 standard deviations above the EM median, on average.
- Deficiencies in health provision
  - Issues: Tuberculosis cases; Infant mortality; Life expectancy.
  - Deviation: Despite an improvement of 0.6 standard deviations, indicators remained 2.1 standard deviations below the EM median in 2018, on average.
- Deficiencies in security provision
  - Issues: Business cost of crime and violence; Organized crime; Reliability of police services.
  - Deviation: Despite some marginal improvement, indicators remained, on average, 1.1 standard deviations below the EM median.

### Annex I — Crime and business impact (key points)
- Homicide rates compare unfavorably to EMs and satisfaction with law enforcement has declined.
- Crime is reported as a major obstacle in business surveys, particularly affecting SMEs and start-ups.
- Incidence of crime is higher than in other regions and is especially costly for smaller firms.
- Crime has damaging effects on tourism inflows.

*Source: IMF staff report (chapter text provided).*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2020/english/1zafea2020002.pdf_