## _wp0925

## Source details

**Canonical URL:** [_wp0925](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0925.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0925.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0925.pdf.json)

---

### I. Introduction
- Between 1996 and 2006 annual average per capita GDP growth in South Africa was 1.7 percent versus 3.6 percent for low- and middle-income countries generally.
- Study objective: examine factors accounting for South Africa’s growth since the end of apartheid by comparing GDP components and saving and investment performance with a panel of faster-growing countries (the 10 fastest-growing economies meeting a population criterion and data availability).
- Core finding:
  - Sluggish investment over the last decade hindered faster growth in South Africa; underinvestment is partly explained by limited saving.
  - To achieve sustained acceleration in growth South Africa needs to:
    - raise its investment and savings rates,
    - raise labor productivity,
    - become more open to trade.

### II. Major constraints on growth (1996–2006): decompositions and stylized facts
- Overall comparison:
  - Real per capita GDP growth averaged 4.7 percent a year for the panel and 1.7 percent for South Africa (first decomposition panel definition).
  - Main consistent result across decompositions: the contribution of investment (capital) to growth is relatively low in South Africa.

- First Decomposition — Demand Components of GDP
  - South Africa is less export-driven than the panel:
    - South Africa: two-thirds of real growth of domestic production comes from domestic demand and one-third from foreign demand.
    - Panel: the opposite (foreign demand larger).
    - Trade flows (exports + imports)/GDP: panel average 63 percent; South Africa 53 percent.
  - Investment contributes less and consumption contributes more in South Africa:
    - Investment contributes 25 percent of GDP growth in South Africa versus 32 percent in the panel.
    - Gross domestic saving averaged 15 percent of GDP in South Africa versus 29 percent in the panel.
  - Table 1 (Normalized Contributions of Demand Components to GDP Growth (percent)):
    - South Africa: Absorption 100; Trade Balance –17; Private Consumption 73; Investment 25; Public Consumption 19.
    - Panel: Absorption 100; Trade Balance –4; Private Consumption 69; Investment 32; Public Consumption 9.

- Second Decomposition — Productivity and Labor Input Characteristics
  - GDP growth decomposed into: labor productivity, employment rate, participation rate, working age population, and population size.
  - Panel average real GDP growth per capita for this decomposition: 4.4 percent.
  - Key findings:
    - Labor productivity growth contributed much less in South Africa than in the panel. If South Africa’s labor productivity had grown at the panel rate, average annual GDP growth would have been 3 percentage points higher.
    - Employment vs. participation:
      - Employment made a negative contribution and participation a positive one in South Africa; the opposite occurred in the panel.
      - Employment grew 3.5 times more in South Africa than in the panel, but labor force grew five times the panel’s rate, leading to a decrease in the employment rate and increase in participation rate in South Africa.
  - Table 2 (Employment and Labor Force in South Africa and Comparators (average 1996–2006) (Percent)):
    - South Africa: Annual Employment Growth 2.8; Annual Labor Force Growth 3.2; Employment Rate 74.5; Participation Rate 52.6.
    - Panel: Annual Employment Growth 0.8; Annual Labor Force Growth 0.6; Employment Rate 91.0; Participation Rate 66.2.

- Third Decomposition — Capital, Labor, and Total Factor Productivity (TFP)
  - GDP growth broken into capital, labor, and TFP components.
  - Panel composition change shifts panel average real GDP growth per capita to 3.8 percent for this decomposition.
  - Main results:
    - Lesser contribution of capital in South Africa: capital contribution is one-third of that in the panel — the main explanation of the GDP growth gap.
    - TFP contribution in South Africa is about two-thirds of that in the panel; part of this gap may reflect lower labor quality and lower R&D investment.
      - R&D expenditures (1996–2006): South Africa 0.76 percentage points of GDP; panel 0.92 percent.
      - R&D researchers per million people (average): South Africa 307; panel 1,207.
    - When TFP is adjusted for labor quality, the difference in TFP contribution is reduced.
  - Table 3 (Results of the Third Decomposition (Percent)):
    - Without labor quality adjustment:
      - South Africa: GDP Growth 3.5; Capital Contribution 0.6; Labor Contribution 1.8; TFP Contribution 1.1.
      - Panel: GDP Growth 5.1; Capital Contribution 1.8; Labor Contribution 1.6; TFP Contribution 1.7.
    - With labor quality adjustment:
      - South Africa: GDP Growth 3.5; Capital Contribution 0.6; Labor Contribution 1.9; TFP Contribution 1.0.
      - Panel: GDP Growth 5.1; Capital Contribution 1.8; Labor Contribution 1.9; TFP Contribution 1.3.

### III. Investment determinants in South Africa compared with the panel
- Investment underperformance:
  - Between 1996 and 2001 investment grew sluggishly in South Africa; pickup since 2002 insufficient to close the gap with the panel.
  - On average the investment-to-GDP ratio in South Africa has been 10 percentage points less than in the panel for the period studied.
  - Annual real investment growth has been 1.5 percentage points less in South Africa than in the panel.
  - By 2007 the investment ratio in South Africa remained 9 percentage points below the panel ratio.

- Four macroeconomic variables examined as investment determinants (excerpt focuses on saving and real interest rates):
  1. Saving:
     - Saving-to-GDP ratio averaged 14 percentage points lower in South Africa than in the panel.
     - The saving gap increased from an average of 12 percentage points for 1996–2001 to 16 percentage points for 2002–2006.
  2. Cost of capital (real interest rate):
     - Real interest rate was on average 0.2 percentage points a year higher in South Africa, with the gap widening over time:
       - Average gap –1.4 percentage points for 1996–2001;
       - +2.2 percentage points for 2002–2006.
     - Real interest rates declined significantly in South Africa (particularly 1996–2001) but declined more in the panel, negating South Africa’s initial advantage.
     - Econometric evidence: a simple equation with capacity utilization and real interest rate explains 60 percent of variability of real investment growth in South Africa since 1970.
       - Estimated effect: a 1 percentage point increase in the real interest rate decreases real investment growth after a year by 7 percentage points (other factors equal).
  3. Capacity utilization and micro/institutional factors are also analyzed in the full paper.

- Figures referenced (descriptive):
  - Figure 3 (Gross Capital Formation (Percent of GDP)) shows panel vs. South Africa series with South Africa persistently below the panel for 1995–2007.
  - Figure 4 (Real Interest Rate (Percent)) shows panel vs. South Africa series for 1996–2006 with divergent trends described above.

### IV. Evidence on financing constraints and investment returns
- Liquidity constraints on private financing do not seem to explain South Africa’s slow investment growth relative to the panel; financing appears more abundant in South Africa than in the panel.
- Financial depth evidence:
  - On average, over the decade the annual stock of domestic credit to the private sector amounted to 123 percent of GDP in South Africa and 51 percent in the panel.
- Indirect indicators of returns on investment:
  - Net operating surplus to capital stock at current cost averaged 17 percent between 1996 and 2006 with a steady increase over the period (South Africa).
  - FDI inflows (1996–2006): South Africa 1.6 percent of GDP; panel 2.3 percent of GDP (average).
  - Stock prices (1998–2006): South Africa grew by 18 percent; panel grew by 26 percent.
- Interpretation:
  - Despite deeper domestic credit, lower FDI inflows and lower stock market returns suggest returns on investment have been less attractive in South Africa than in the panel.

### V. Microeconomic business environment (Doing Business Survey, 2003–2006)
- South Africa outperformed the panel at each stage of procedures required to go into business; difficulties in doing business do not seem to account for lower investment in South Africa.
- Selected Doing Business indicators (2003–2006):
  - Starting a business:
    - Number of procedures: South Africa 9, Panel 10
    - Days required: South Africa 37, Panel 37
    - Cost (% of per capita income): South Africa 9, Panel 31
  - Registering property:
    - Number of procedures: South Africa 6, Panel 6
    - Days required: South Africa 23, Panel 114
  - Dealing with licenses:
    - Number of procedures to build a warehouse: South Africa 17, Panel 19
    - Days required to build a warehouse: South Africa 174, Panel 260
  - Employing workers:
    - Rigidity of employment index (0: less; 100: more): South Africa 44, Panel 43
  - Enforcing contracts:
    - Number of procedures: South Africa 26, Panel 39
    - Days required: South Africa 600, Panel 602
  - Protecting investors:
    - Disclosure index (0 less; 10 more): South Africa 8, Panel 6
  - Closing a business:
    - Years to resolve insolvency: South Africa 2, Panel 4
- Note: Data are only available since 2003.

### VI. National saving: stylized facts for South Africa
- Long-run trend:
  - Since 1980 national saving in South Africa has trended downward: the rate decreased from an average of 26 percentage points of gross national disposable income (GNDI) during 1960–1985 to 15 percent in 2007 (the large spike in saving around 1980 associated with a bubble in gold prices).
- Disaggregated changes (1996–2007):
  - National saving decreased by 1.8 percentage points of GNDI between 1996 and 2007.
  - This change reflected:
    - A fall in private saving by 8.1 percentage points of GNDI.
    - A rise in public saving by 6.2 percentage points of GNDI.
- Sectoral saving changes:
  - Within the private sector, the decline in saving between 1996 and 2007 was:
    - Corporations: 6 percentage points of GNDI.
    - Households: 2.1 percentage points of GNDI (starting from a lower level).
  - Corporate saving on average accounts for 84 percent of total private saving over the period (16 percent for household saving); in recent years the corporate share reached almost 90 percent.

### VII. South Africa versus the panel: levels and gaps in saving
- Overall comparisons (1996–2007):
  - The national saving rate was on average more than 10 percentage points lower in South Africa than in the panel.
  - Since 1996 the national saving rate decreased in South Africa but increased in the panel.
- Table 6 summary (Average 1996–2007; Evolution between 1996 and 2007, percentage points):
  - National saving:
    - Evolution: South Africa -1.8, Panel +2.9, Gap 4.7
    - Average (1996–2007): South Africa 15.7, Panel 27.7, Gap 12.0
  - Public saving:
    - Evolution: South Africa +6.2, Panel +6.1, Gap -0.1
    - Average: South Africa 0.2, Panel 2.0, Gap 1.8
  - Private saving:
    - Evolution: South Africa -8.1, Panel -3.2, Gap 4.9
    - Average: South Africa 15.6, Panel 25.6, Gap 10.0
- By international standards both corporate and household savings in South Africa seem to be low:
  - In 2006, household saving rate averaged 4.8 percentage points in OECD countries vs. –0.5 in South Africa.
  - Corporation saving gap: 11.5 percentage points in OECD countries vs. 10.1 in South Africa (comparison caveat due to data limitations).

### VIII. Accounting decomposition of the corporate saving rate (1996–2007)
- Corporate saving decomposed into four components: gross operating profit rate; ratio of total corporate income to production income; retained earning ratio; ratio of profit after social contributions.
- Three main results:
  1. Increased corporate tax payments are the main reason for the decrease in corporate saving:
     - The ratio of profit after social contributions declined by 27 percentage points between 1996 and 2007.
     - The ratio of taxes to gross primary income increased by 25 percentage points between 1996 and 2007.
     - This occurred despite decreases in the statutory corporate income tax rate because better tax collection and broadening of the tax base (notably introduction of a capital gains tax in 2001) more than offset the rate decline.
  2. The increase in tax payments transferred saving from the corporate to the public sector, with negligible net effect on total saving (accounting identity).
     - Simulation: assuming the ratio of tax to corporate gross primary income held at its 1996 level, corporate saving in 2007 would have been 15 percent and public saving would have been –2 percent (other factors unchanged); these simulate close to the 1996 data (16 and –3).
  3. Decrease in retained earnings also reduced corporate saving:
     - The retained earnings ratio decreased by 21 percentage points between 1996 and 2007.
     - The drop was especially noticeable before 2001; since 2001, better investment opportunities seem to have increased incentives to retain dividends.

- Table 9 (Selected levels and variations):
  - 1996: S/VA = 29.0; (1) = 45.5; (2) = 103.7; (3) = 75.9; (4) = 80.9
  - 2001: S/VA = 21.1; (1) = 48.3; (2) = 107.2; (3) = 61.0; (4) = 66.7
  - 2007: S/VA = 17.4; (1) = 53.0; (2) = 111.5; (3) = 55.1; (4) = 53.5
  - Variation 1996–2007: S/VA = -11.5; (1) = +7.5; (2) = +7.9; (3) = -20.8; (4) = -27.4

### IX. Econometric analysis: behavioral determinants of private saving
- Method:
  - Apply parameter estimates from Loayza, Schmidt-Hebbel, and Serven (2000) private saving equation (GMM) to South Africa and the panel for 1996–2006.
  - Determinants include: domestic income (level and growth), real interest rates, financial depth, demographic variables (dependency ratios, urbanization), fiscal variables, measures of uncertainty (inflation, unemployment), and persistence (lagged saving rates).
- Two primary econometric results:
  1. Factors explaining why private saving is lower in South Africa than in the panel (average contributions, 1996–2006):
     - Actual private saving rate (av. 1996–2006): South Africa 19.7, Panel 30.1, Gap -10.4
     - Main explanatory differences (average long-term contributions):
       - Real growth rate income: South Africa 0.9, Panel 3.1, Gap -2.2
       - Real per capita income: South Africa 94.4, Panel 96.0, Gap -1.6
       - Real interest rate: South Africa -4.9, Panel -3.6, Gap -1.3
       - Terms of trade: South Africa 1.3, Panel 0.5, Gap 0.8
       - Urbanization ratio: South Africa -53.1, Panel -47.3, Gap -5.7
       - Old dependency ratio: South Africa -9.6, Panel -18.3, Gap 8.7
       - Young dependency ratio: South Africa -38.7, Panel -30.0, Gap -8.6
       - Gov saving: South Africa 0.0, Panel -1.7, Gap 1.7
       - Private credit: South Africa -8.2, Panel -6.4, Gap -1.8
       - Inflation: South Africa 2.4, Panel 5.0, Gap -2.6
       - Implied constant + residual: South Africa 35.0, Panel 31.2, Gap 3.8
     - Interpretation:
       - Four exogenous factors explain most of the gap: (1) a younger population, (2) a higher rate of urbanization, (3) lower growth in GDP per capita, (4) lower inflation.
       - The higher young dependency ratio in South Africa largely explains the gap, partially mitigated by a lower old dependency ratio.
       - Urbanization has a crucial but ambiguous role; lower inflation (less macro uncertainty) reduces precautionary saving.
  2. Factors explaining the decrease in South Africa’s private saving over the last decade (change between 1995–97 and 2004–06):
     - Actual decrease: -5.7
     - Estimated long-term decrease: -4.6
     - Estimated long-term contributions (selected):
       - Real growth rate income: 1.8
       - Real per capita income: 0.7
       - Real interest rate: 2.9
       - Terms of trade: 0.4
       - Urbanization ratio: -4.0
       - Old dependency ratio: -2.1
       - Young dependency ratio: 4.0
       - Government saving: -3.4
       - Private credit: -3.9
       - Inflation: -1.0
     - Interpretation:
       - The main explanatory variables for the decline are: relaxation of credit constraints (private credit term), increased urbanization, increased public saving, and aging (old-age dependency) effects.
       - The private credit term (measuring release of borrowing constraints after apartheid) is the dominant factor.
       - Increased urbanization also contributed to the decline.
       - Increased public saving is consistent with corporate-to-public saving transfers observed in accounting decomposition.
       - The increase in old-age dependency lowered private saving but was partly offset by a decrease in young-age dependency.

### X. Conclusions and policy implications
- Key constraints on growth since 1996:
  - Low investment rate, insufficient labor productivity gains, lower openness to trade, and slower technical progress relative to faster-growing countries.
  - Underinvestment is a common factor across growth decompositions and has significantly slowed growth since 1996.
- Saving and investment link:
  - Low saving has constrained investment expansion; private saving is limited by structural factors.
  - Structural determinants (demographics, urbanization, financial deepening) largely explain the low private saving rate and are not easily affected by policy.
  - The two most important factors explaining the private saving gap with the panel are demographic: the young dependency ratio and the urbanization rate.
- Policy implications:
  - Some macroeconomic variables can affect private saving but may have adverse effects on national saving or macro costs (e.g., decreasing public saving or increasing inflation).
  - Preferred policy to increase national saving is to continue raising public saving, preferably by containing the growth of public consumption.
    - An increase in public saving would raise national saving even if partially offset by the private sector.
    - Increasing the tax take (higher rates or broader base) is likely to be less effective in boosting national saving than slowing public consumption growth because private sector offset is likely to be higher and faster for tax increases (illustrated by corporate accounting analysis).
  - Ricardian considerations:
    - Decreasing public saving increases private saving by a smaller amount (Ricardian effect), so national saving falls.
    - Conversely, a reduction in public consumption has an indirect and potentially smaller offset on private saving, making it a more effective route to raise national saving than raising taxes.

### XI. Computation methods (summary)
- First decomposition:
  - YΔ = AΔ + TBΔ with A = C + G + I, TB = X − M.
  - Contributions normalized and averaged annually for 1996–2006.
- Second decomposition:
  - Y decomposed into labor productivity, employment rate, participation rate, working-age population share, and total population.
  - Contributions computed annually and averaged for 1996–2006; South Africa employment and labor force data corrected for 2000 survey break.
- Third decomposition:
  - Cobb-Douglas Y = A Kα L1−α with α = 1/3.
  - Capital stocks via perpetual inventory method with geometric depreciation at 5% and asset life 21 years.
  - Alternative specification transfers labor quality H from TFP to labor using education proxies (H = s^0.07).
- Corporate saving decomposition:
  - S/VA = (GOS/VA) * ((GPI + d)/GOS) * (GPI/(GPI + d)) * (S/GPI).
- Econometric application:
  - Loayza, Schmidt-Hebbel, and Serven (2000) GMM private saving equation applied out-of-sample to 1996–2006.
  - Long-term multipliers derived by dividing short-term coefficients by (1 − coefficient of the lagged dependent variable).

*Source: IMF staff calculations and referenced data as presented in _wp0925.*

### References..............................................................................................................

### _wp0925 - References..............................................................................................................

### I. INTRODUCTION
- Between 1996 and 2006 annual average per capita GDP growth in South Africa was 1.7 percent versus 3.6 percent for low- and middle-income countries generally.
- Study objective: examine factors accounting for South Africa’s growth since the end of apartheid by comparing GDP components and saving and investment performance with a panel of faster-growing countries (the 10 fastest-growing economies meeting a population criterion and data availability).
- Core finding (summary): sluggish investment over the last decade hindered faster growth in South Africa; underinvestment is partly explained by limited saving. To achieve sustained acceleration in growth South Africa needs to:
  - raise its investment and savings rates,
  - raise labor productivity,
  - become more open to trade.

*Source: IMF staff calculations and World Development Indicators as presented in the document.*

### II. MAJOR CONSTRAINTS ON GROWTH (1996–2006): DECOMPOSITIONS AND STYLIZED FACTS
- Overall comparison:
  - Real per capita GDP growth averaged 4.7 percent a year for the panel and 1.7 percent for South Africa (first decomposition panel definition).
  - The main consistent result across decompositions: the contribution of investment (capital) to growth is relatively low in South Africa.

A. First Decomposition — Demand Components of GDP
- Two key conclusions:
  - South Africa is less export-driven than the panel:
    - South Africa: two-thirds of real growth of domestic production comes from domestic demand and one-third from foreign demand.
    - Panel: the opposite (foreign demand larger).
    - Trade flows (exports + imports)/GDP: panel average 63 percent; South Africa 53 percent.
  - Investment contributes less and consumption contributes more in South Africa:
    - Investment contributes 25 percent of GDP growth in South Africa versus 32 percent in the panel (see Table 1).
    - Gross domestic saving averaged 15 percent of GDP in South Africa versus 29 percent in the panel.
- Table 1 (Normalized Contributions of Demand Components to GDP Growth (percent)):
  - South Africa: Absorption 100; Trade Balance –17; Private Consumption 73; Investment 25; Public Consumption 19.
  - Panel: Absorption 100; Trade Balance –4; Private Consumption 69; Investment 32; Public Consumption 9.

B. Second Decomposition — Productivity and Labor Input Characteristics
- GDP growth decomposed into: labor productivity, employment rate, participation rate, working age population, and population size.
- Panel average real GDP growth per capita for this decomposition: 4.4 percent.
- Key findings:
  - Labor productivity growth contributed much less in South Africa than in the panel. If South Africa’s labor productivity had grown at the panel rate, average annual GDP growth would have been 3 percentage points higher.
  - Employment vs. participation symmetry:
    - Employment made a negative contribution and participation a positive one in South Africa; the opposite occurred in the panel.
    - Employment grew 3.5 times more in South Africa than in the panel, but labor force grew five times the panel’s rate, leading to a decrease in the employment rate and increase in participation rate in South Africa.
- Table 2 (Employment and Labor Force in South Africa and Comparators (average 1996–2006) (Percent)):
  - South Africa: Annual Employment Growth 2.8; Annual Labor Force Growth 3.2; Employment Rate 74.5; Participation Rate 52.6.
  - Panel: Annual Employment Growth 0.8; Annual Labor Force Growth 0.6; Employment Rate 91.0; Participation Rate 66.2.

C. Third Decomposition — Capital, Labor, and Total Factor Productivity (TFP)
- GDP growth broken into capital, labor, and TFP components.
- Panel composition change shifts panel average real GDP growth per capita to 3.8 percent for this decomposition.
- Main results:
  - Lesser contribution of capital in South Africa: capital contribution is one-third of that in the panel — this is the main explanation of the GDP growth gap.
  - TFP contribution in South Africa is about two-thirds of that in the panel; part of this gap may reflect lower labor quality and lower R&D investment.
    - R&D expenditures (1996–2006): South Africa 0.76 percentage points of GDP; panel 0.92 percent.
    - R&D researchers per million people (average): South Africa 307; panel 1,207.
  - When TFP is adjusted for labor quality, the difference in TFP contribution is reduced.
- Table 3 (Results of the Third Decomposition (Percent)):
  - Without labor quality adjustment:
    - South Africa: GDP Growth 3.5; Capital Contribution 0.6; Labor Contribution 1.8; TFP Contribution 1.1.
    - Panel: GDP Growth 5.1; Capital Contribution 1.8; Labor Contribution 1.6; TFP Contribution 1.7.
  - With labor quality adjustment:
    - South Africa: GDP Growth 3.5; Capital Contribution 0.6; Labor Contribution 1.9; TFP Contribution 1.0.
    - Panel: GDP Growth 5.1; Capital Contribution 1.8; Labor Contribution 1.9; TFP Contribution 1.3.
- Comparative production-function results (Table 4) underline differences with earlier studies and sample periods; IMF (2008) results for 1996–2006 show GDP 3.5; Capital 0.6; Labor 1.8 (1.9 quality adjusted); TFP 1.1 (1.0 quality adjusted).

### III. INVESTMENT DETERMINANTS IN SOUTH AFRICA COMPARED WITH THE PANEL
- Investment underperformance:
  - Between 1996 and 2001 investment grew sluggishly in South Africa; pickup since 2002 insufficient to close the gap with the panel.
  - On average the investment-to-GDP ratio in South Africa has been 10 percentage points less than in the panel for the period studied.
  - Annual real investment growth has been 1.5 percentage points less in South Africa than in the panel.
  - By 2007 the investment ratio in South Africa remained 9 percentage points below the panel ratio.
- Four macroeconomic variables examined as investment determinants:
  1. Saving:
     - Saving-to-GDP ratio averaged 14 percentage points lower in South Africa than in the panel.
     - The saving gap increased from an average of 12 percentage points for 1996–2001 to 16 percentage points for 2002–2006.
  2. Cost of capital (real interest rate):
     - Real interest rate was on average 0.2 percentage points a year higher in South Africa, with the gap widening over time:
       - Average gap –1.4 percentage points for 1996–2001;
       - +2.2 percentage points for 2002–2006.
     - Real interest rates declined significantly in South Africa (particularly 1996–2001) but declined more in the panel, negating South Africa’s initial advantage.
     - Econometric evidence: a simple equation with capacity utilization and real interest rate explains 60 percent of variability of real investment growth in South Africa since 1970.
       - Estimated effect: a 1 percentage point increase in the real interest rate decreases real investment growth after a year by 7 percentage points (other factors equal).
  3. (Other determinants analyzed in the paper include capacity utilization and additional micro/institutional factors; the excerpt focuses on saving and real interest rates as prominent macro determinants.)
- Figures referenced:
  - Figure 3 (Gross Capital Formation (Percent of GDP)) shows panel vs. South Africa series with South Africa persistently below the panel for 1995–2007.
  - Figure 4 (Real Interest Rate (Percent)) shows panel vs. South Africa series for 1996–2006 with divergent trends described above.

*Source: World Development Indicators, World Economic Outlook, International Financial Statistics, and IMF staff calculations as cited in the document.*

### 3. In principle slow investment growth could also result from liquidity constraints on private

### _wp0925 - 3. In principle slow investment growth could also result from liquidity constraints on private

### Evidence on financing constraints and investment returns
- Liquidity constraints on private financing (banks reluctant to lend regardless of interest rates) do not seem to explain South Africa’s slow investment growth relative to the panel; financing appears more abundant in South Africa than in the panel.
- Financial depth evidence:
  - On average, over the decade the annual stock of domestic credit to the private sector amounted to 123 percent of GDP in South Africa and 51 percent in the panel.
- Indirect indicators of returns on investment:
  - A direct measure of returns can be computed for South Africa: the ratio of the net operating surplus of the domestic economy to capital stock at current cost averaged 17 percent between 1996 and 2006 with a steady increase over the period.
  - Foreign direct investment (FDI) inflows: between 1996 and 2006 FDI inflows on average amounted to 1.6 percent of GDP in South Africa and 2.3 percent in the panel.
  - Stock market price changes: between 1998 and 2006, stock prices grew by 18 percent in South Africa versus 26 percent in the panel.
- Interpretation:
  - Lower FDI inflows and lower stock market returns suggest returns on investment have been less attractive in South Africa than in the panel, despite deeper domestic credit.

### Microeconomic business environment (Doing Business Survey, 2003–2006)
- South Africa outperformed the panel at each stage of procedures required to go into business; difficulties in doing business do not seem to account for lower investment in South Africa.
- Selected Doing Business indicators (2003–2006):
  - Starting a business:
    - Number of procedures: South Africa 9, Panel 10
    - Days required: South Africa 37, Panel 37
    - Cost (% of per capita income): South Africa 9, Panel 31
  - Registering property:
    - Number of procedures: South Africa 6, Panel 6
    - Days required: South Africa 23, Panel 114
  - Dealing with licenses:
    - Number of procedures to build a warehouse: South Africa 17, Panel 19
    - Days required to build a warehouse: South Africa 174, Panel 260
  - Employing workers:
    - Rigidity of employment index (0: less; 100: more): South Africa 44, Panel 43
  - Enforcing contracts:
    - Number of procedures: South Africa 26, Panel 39
    - Days required: South Africa 600, Panel 602
  - Protecting investors:
    - Disclosure index (0 less; 10 more): South Africa 8, Panel 6
  - Closing a business:
    - Years to resolve insolvency: South Africa 2, Panel 4
- Note: Data are only available since 2003.

### National saving: stylized facts for South Africa
- Long-run trend:
  - Since 1980 national saving in South Africa has trended downward: the rate decreased from an average of 26 percentage points of gross national disposable income (GNDI) during 1960–1985 to 15 percent in 2007 (the large spike in saving around 1980 associated with a bubble in gold prices).
- Disaggregated changes (1996–2007):
  - National saving decreased by 1.8 percentage points of GNDI between 1996 and 2007.
  - This change reflected:
    - A fall in private saving by 8.1 percentage points of GNDI.
    - A rise in public saving by 6.2 percentage points of GNDI (tax revenue gains from robust economic growth and improved revenue administration more than offset public expenditure increases).
- Sectoral saving changes:
  - Within the private sector, the decline in saving between 1996 and 2007 was:
    - Corporations: 6 percentage points of GNDI.
    - Households: 2.1 percentage points of GNDI (starting from a lower level).
  - Corporate saving on average accounts for 84 percent of total private saving over the period (16 percent for household saving); in recent years the corporate share reached almost 90 percent.

### South Africa versus the panel: levels and gaps in saving
- Overall comparisons (1996–2007):
  - The national saving rate was on average more than 10 percentage points lower in South Africa than in the panel.
  - Since 1996 the national saving rate decreased in South Africa but increased in the panel.
- Table 6 summary (Average 1996–2007; Evolution between 1996 and 2007, percentage points):
  - National saving:
    - Evolution: South Africa -1.8, Panel +2.9, Gap 4.7
    - Average (1996–2007): South Africa 15.7, Panel 27.7, Gap 12.0
  - Public saving:
    - Evolution: South Africa +6.2, Panel +6.1, Gap -0.1
    - Average: South Africa 0.2, Panel 2.0, Gap 1.8
  - Private saving:
    - Evolution: South Africa -8.1, Panel -3.2, Gap 4.9
    - Average: South Africa 15.6, Panel 25.6, Gap 10.0
- By international standards both corporate and household savings in South Africa seem to be low:
  - In 2006, household saving rate averaged 4.8 percentage points in OECD countries vs. –0.5 in South Africa.
  - Corporation saving gap: 11.5 percentage points in OECD countries vs. 10.1 in South Africa (note: comparison caveat—household and corporate levels compared only for OECD countries due to data limitations).

### Accounting decomposition of the corporate saving rate (1996–2007)
- Corporate saving decomposition into four components: gross operating profit rate; ratio of total corporate income to production income; retained earning ratio; ratio of profit after social contributions.
- Three main results:
  1. Increased corporate tax payments are the main reason for the decrease in corporate saving:
     - The ratio of profit after social contributions declined by 27 percentage points between 1996 and 2007.
     - The ratio of taxes to gross primary income increased by 25 percentage points between 1996 and 2007.
     - This occurred despite decreases in the statutory corporate income tax rate because better tax collection and broadening of the tax base (notably introduction of a capital gains tax in 2001) more than offset the rate decline.
  2. The increase in tax payments transferred saving from the corporate to the public sector, with negligible net effect on total saving (accounting identity).
     - Simulation: assuming the ratio of tax to corporate gross primary income held at its 1996 level, corporate saving in 2007 would have been 15 percent and public saving would have been –2 percent (other factors unchanged); these simulate close to the 1996 data (16 and –3).
  3. Decrease in retained earnings also reduced corporate saving:
     - The retained earnings ratio decreased by 21 percentage points between 1996 and 2007.
     - The drop was especially noticeable before 2001; since 2001, better investment opportunities seem to have increased incentives to retain dividends.

### Econometric analysis: behavioral determinants of private saving
- Method:
  - Apply parameter estimates from Loayza, Schmidt-Hebbel, and Serven (2000) private saving equation (GMM) to South Africa and the panel for 1996–2006.
  - Standard determinants include: domestic income (level and growth), financial variables (interest rates, financial depth), demographic variables (dependency ratios, pension system), crowding-out terms (fiscal balance, external financing), measures of uncertainty (inflation, unemployment, urbanization), persistence (lagged saving rates).
- Two primary econometric results:
  1. Factors explaining why private saving is lower in South Africa than in the panel (average contributions, 1996–2006):
     - Actual private saving rate (av. 1996–2006): South Africa 19.7, Panel 30.1, Gap -10.4
     - Main explanatory differences (average long-term contributions):
       - Real growth rate income: South Africa 0.9, Panel 3.1, Gap -2.2
       - Real per capita income: South Africa 94.4, Panel 96.0, Gap -1.6
       - Real interest rate: South Africa -4.9, Panel -3.6, Gap -1.3
       - Terms of trade: South Africa 1.3, Panel 0.5, Gap 0.8
       - Urbanization ratio: South Africa -53.1, Panel -47.3, Gap -5.7
       - Old dependency ratio: South Africa -9.6, Panel -18.3, Gap 8.7
       - Young dependency ratio: South Africa -38.7, Panel -30.0, Gap -8.6
       - Gov saving: South Africa 0.0, Panel -1.7, Gap 1.7
       - Private credit: South Africa -8.2, Panel -6.4, Gap -1.8
       - Inflation: South Africa 2.4, Panel 5.0, Gap -2.6
       - Implied constant + residual: South Africa 35.0, Panel 31.2, Gap 3.8
     - Interpretation:
       - Four exogenous factors explain most of the gap: (1) a younger population, (2) a higher rate of urbanization, (3) lower growth in GDP per capita, (4) lower inflation.
       - The higher young dependency ratio in South Africa largely explains the gap, partially mitigated by a lower old dependency ratio.
       - Urbanization has a crucial but ambiguous role; lower inflation (less macro uncertainty) reduces precautionary saving.
  2. Factors explaining the decrease in South Africa’s private saving over the last decade (change between 1995–97 and 2004–06):
     - Actual decrease: -5.7
     - Estimated long-term decrease: -4.6
     - Estimated long-term contributions (selected):
       - Real growth rate income: 1.8
       - Real per capita income: 0.7
       - Real interest rate: 2.9
       - Terms of trade: 0.4
       - Urbanization ratio: -4.0
       - Old dependency ratio: -2.1
       - Young dependency ratio: 4.0
       - Government saving: -3.4
       - Private credit: -3.9
       - Inflation: -1.0
     - Interpretation:
       - The main explanatory variables for the decline are: relaxation of credit constraints (private credit term), increased urbanization, increased public saving, and aging (old-age dependency) effects.
       - The private credit term (measuring release of borrowing constraints after apartheid) is the dominant factor.
       - Increased urbanization also contributed to the decline.
       - Increased public saving is consistent with corporate-to-public saving transfers observed in accounting decomposition.
       - The increase in old-age dependency lowered private saving but was partly offset by a decrease in young-age dependency.

### Conclusions and policy implications
- Key constraints on growth since 1996:
  - Low investment rate, insufficient labor productivity gains, lower openness to trade, and slower technical progress relative to faster-growing countries.
  - Underinvestment is a common factor across growth decompositions and has significantly slowed growth since 1996.
- Saving and investment link:
  - Low saving has constrained investment expansion; private saving is limited by structural factors.
  - Structural determinants (demographics, urbanization, financial deepening) largely explain the low private saving rate and are not easily affected by policy.
  - The two most important factors explaining the private saving gap with the panel are demographic: the young dependency ratio and the urbanization rate.
- Policy implications:
  - Some macroeconomic variables can affect private saving but may have adverse effects on national saving or macro costs (e.g., decreasing public saving or increasing inflation).
  - The preferred policy to increase national saving is to continue raising public saving, preferably by containing the growth of public consumption.
    - An increase in public saving would raise national saving even if partially offset by the private sector.
    - Increasing the tax take (higher rates or broader base) is likely to be less effective in boosting national saving than slowing public consumption growth because private sector offset is likely to be higher and faster for tax increases (illustrated by corporate accounting analysis).
  - Ricardian considerations:
    - Decreasing public saving increases private saving by a smaller amount (Ricardian effect), so national saving falls.
    - Conversely, a reduction in public consumption has an indirect and potentially smaller offset on private saving, making it a more effective route to raise national saving than raising taxes.

*Source: IMF staff summary of chapter content in _wp0925*

### 1.      Computation of the First Growth Decomposition

### 1.      Computation of the First Growth Decomposition

### Method
- GDP growth is decomposed into two components: the contribution of absorption (A) and the contribution of the trade balance (TB).
- Absorption A = household consumption (C) + government consumption (G) + investment (I).
- Trade balance TB = exports (X) − imports (M).
- Algebraic expression:
  - YΔ = AΔ + TBΔ
  - AΔ = (C/Y)Δ + (G/Y)Δ + (I/Y)Δ
  - TBΔ = (X/Y)Δ − (M/Y)Δ
- Two components identified: (1) contribution of domestic demand for domestic products, and (2) contribution of foreign demand for domestic products.

### Presentation and aggregation
- Results are presented in normalized form by dividing each side of the equation by GDP growth:
  - (X/Y)Δ/(YΔ) + (−M/Y)Δ/(YΔ) + (A/Y)Δ/(YΔ) = %100
- Contributions are computed annually and then averaged for the period 1996–2006.

---

### 2.      Computation of the Second Growth Decomposition

### Method
- Real GDP is written as the product of ratios: labor productivity, employment rate, participation rate, working age population share, and total population.
- Identity (levels):
  - Y/Ltot * Lempl/Lforce * Lforce/Lworking-age * Lworking-age/Ltotal * Ltotal = Y
  - Notation: L = employment; Lempl = labor force (emplL); Lforce = working age population (forceL); Ltot = total population (6415−L in source notation).
- Growth approximation (contributions to growth):
  - Ŷ ≈ (1) productivity + (2) employment rate + (3) participation rate + (4) “demographic dividend” term (working age population) + (5) population size
  - Expressed in source notation as:
    - Ŷ ≈ ∧(1) + ∧(2) + ∧(3) + ∧(4) + ∧(5) with the specific mapping described in the source.

### Presentation and application
- Unlike the first decomposition, contributions are not normalized.
- Contributions are computed annually and then averaged for 1996–2006.
- Decomposition is applied to South Africa after correcting the official employment and labor force data (note on 2000 survey break and correction referenced).

---

### 3.      Computation of the Third Growth Decomposition

### First decomposition (K, L, A)
- Real GDP growth is decomposed into capital (K), labor (L), and TFP (A) contributions using a Cobb-Douglas function: Y = A Kα L1−α.
- Growth decomposition expression:
  - YΔ = AΔ + (1−α) LΔ + α KΔ      (equation (1) in source)
- Assumptions for application:
  - α = 1/3 in all countries.
  - L = Employment or Labor force (depending on availability). The 2000 employment growth rate is adjusted for South Africa (see Box 3 in source).
  - Capital stocks computed via perpetual inventory method with geometric depreciation at 5% and asset life 21 years, chosen to be consistent with a degree of declining balance to depreciation equal to 1.05. The SARB straight-line capital stock is not used; the described method is applied to South Africa.

### Alternative decomposition (labor quality)
- TFP as residual includes changes in labor quality; labor quality can be transferred from TFP to L.
- Method follows Barro (1999) and Bosworth-Collins (2003), with labor quality H (human capital) often proxied by average years of schooling s and a functional form where H = s^0.07 (as represented in source).
- Production function including labor quality (source notation):
  - A Kα (L H)1−α = Y  (leading to equation (2) in source)
  - Transferred decomposition: ∧Y = ∧K + ∧L + ∧H + ∧A* with algebraic adjustments as in source.
- Consequences of transferring labor quality from TFP to L:
  - (1) Increases the labor contribution and decreases the TFP contribution for both South Africa and the panel.
  - (2) Because labor quality (education) improvement appears better in the panel than in South Africa over the period, transferring labor quality increases the L contribution relatively more and decreases TFP relatively more in the panel, narrowing the gap between South Africa and the panel.

---

### 4.      Accounting Decomposition of the Corporate Saving Rate

### Definition and decomposition
- Corporate saving rate S/VA (ratio of corporate saving S to corporate value added VA) is decomposed into four multiplicative components reflecting successive deductions from corporate value added:
  - S/VA = (GOS/VA) * ((GPI + d)/GOS) * (GPI/(GPI + d)) * (S/GPI)
  - Where:
    - (1) GOS/VA = gross operating profit rate (GOS to VA) — deduction by workers.
    - (2) (GPI + d)/GOS = ratio of gross primary income before dividends to GOS — reflects lenders’ deduction (net interest paid).
    - (3) GPI/(GPI + d) = retained earnings ratio — reflects shareholders’ deduction (dividends paid).
    - (4) S/GPI = ratio of profit after social contributions — reflects government’s net deduction for taxes, other levies, and transfers.

### Empirical levels and variation (Table 9 in source)
- Table 9: Results — Level and Variations of the Accounting Components of the Corporate Saving Rate
  - Columns correspond to: S/VA; (1); (2); (3); (4)
  - 1996: S/VA = 29.0; (1) = 45.5; (2) = 103.7; (3) = 75.9; (4) = 80.9
  - 2001: S/VA = 21.1; (1) = 48.3; (2) = 107.2; (3) = 61.0; (4) = 66.7
  - 2007: S/VA = 17.4; (1) = 53.0; (2) = 111.5; (3) = 55.1; (4) = 53.5
  - Variation 1996–2007: S/VA = -11.5; (1) = +7.5; (2) = +7.9; (3) = -20.8; (4) = -27.4
- Data sources: Quarterly Bulletin of the South African Reserve Bank (S-125 and S-126) and IMF staff calculations.

---

### 5.      Econometric Analysis of the Private Saving Rate

### Estimation background and usage
- Uses Loayza, Schmidt-Hebbel, and Serven (2000) GMM system estimator specification estimated for a panel of 150 countries (1965–1994). Applying their estimates to 1996–2006 is an out-of-sample forecast per source.
- The selected private saving rate equation includes the following variables:
  - Lagged private saving rate: PS/Y
  - Real per capita GPDI: YR/L
  - Real growth rate of per capita GPDI: dlnYR/L
  - Real lending interest rate: R
  - Terms of trade: TOT
  - Urbanization ratio: URB
  - Old dependency ratio: OLD (population above 64 to total population)
  - Young dependency ratio: YOUNG (population under 15 to total population)
  - Ratio of government saving to GPDI: GS/Y
  - Ratio of private credit flow to GPDI: C/Y
  - Inflation rate of the GDP deflator: Π
- The ratio M2 to national income is noted as not statistically significant in the adopted specification.

### Estimated equation (short-term coefficients as in source)
- The estimated private saving equation (in source notation) is:
  - PS/Y = 0.143 * ln(PS/Y) + ...  [full equation in source]
  - Explicit functional form from source (preserving source formatting and coefficients):
    - )1ln(*143.0/*318.0/*285.0*299.0*655.0*382.0
      )ln(*078.0)1ln(*253.0)/ln(*45.0)/ln(*049.0/*587.0/
      1
      Π++−−−−−
      ++−++=
      −
      YCYGSYOUNGOLDUR
      TOTRLYRdLYRYPSYPS
  - (The source presents the full set of short-term coefficients and variable transformations in this expression.)

### Applications and computation of contributions
- The estimated World Bank equation is applied separately to South Africa and to the panel to:
  - Compute the contribution of explanatory variables to the average level of private saving in South Africa and in the panel and to the gap between them.
  - Compute the contribution of explanatory variables to the decline in South African private saving over the last decade, with contributions computed between intervals 1995–1997 and 2004–2006 to reduce sensitivity to start/end dates.
- Long-term contributions (because of lagged dependent variable) computed in two stages:
  - (1) Long-term multipliers derived by dividing each short-term coefficient by (1 − coefficient of the lagged dependent variable).
  - (2) Long-term contributions = multiplier × average value of explanatory variable over the period.
- By definition, the gap between the actual saving rate and the sum of explanatory variable contributions equals the sum of the estimated residual and the estimated constant. World Bank estimates are not available at country level, so the constant peculiar to South Africa (and to the panel) is unknown; therefore it is not possible to separate the estimated residual from the estimated constant in the source application.

*Source: _wp0925 - 1.      Computation of the First Growth Decomposition (IMF PDF chapter).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0925.pdf_