## Annex II: Breaks and Spells Stylized Facts with h = 8

## Source details

**Canonical URL:** [Annex II: Breaks and Spells Stylized Facts with h = 8](https://www.imf.org/-/media/files/publications/wp/2017/wp17202.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2017/wp17202.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2017/wp17202.pdf.json)

---

### Introduction and scope
- Focus: growth turning points (up-breaks and down-breaks) and growth spells in sub-Saharan Africa (SSA), compared with emerging and other developing countries.
- Sample and identification:
  - Period: 1950–2016.
  - Structural-break detection method: variant of Bai and Perron (1998) as in Antoshin, Berg, and Souto (2008).
  - Parameters used: minimum years between breaks h = 5 (primary), statistical significance p = 0.10; robustness noted for h = 8.
  - Identified breaks: 184 up-breaks and 185 down-breaks in GDP per capita growth across 146 countries over 1950–2016 (primary results); Annex II using h = 8 reports: Total: Up-Breaks 177 Down-Breaks 165.

### Stylized facts — Structural breaks
- Frequency and timing:
  - Growth turning points are common globally and in SSA, with substantial variation across time and income levels.
  - About 25 percent of total up-breaks and down-breaks occur in SSA (in line with SSA representing about 28 percent of countries in the sample).
  - SSA experienced many down-breaks during 1970–84 and many up-breaks during 1985–99; since 2000 SSA has seen relatively fewer down-breaks while maintaining about the same number of up-breaks.
- Magnitude and volatility:
  - Average break sizes (primary h = 5 results): per capita growth in SSA was 10 percentage points higher after an up-break versus before — the largest break size among groups examined (compare to 7 percent break size in the rest of emerging and developing economies).
  - Using h = 8 (Annex II, Table II.2): SSA: Up-Breaks 7.7- Down-Breaks -7.6 (percentage points).
  - SSA exhibits large swings between up-breaks and down-breaks, indicating high volatility.

### Stylized facts — Growth spells
- Definitions:
  - “Complete” growth spell: begins with an up-break followed by at least 2 percent average per capita income growth and ends with a down-break followed by less than 2 percent average per capita growth.
  - “Incomplete” spells: satisfy the first condition and are ongoing at sample end.
- Counts and concentration:
  - Total spells (primary sample): 80 complete and 67 incomplete spells identified.
  - Annex II (h = 8) reports: Complete Spells 366; Ongoing Spells 63 (full-sample h = 8 tabulation).
  - SSA recorded 40 growth spells between 1950 and 2016 — 27 percent of total spells; 25 of the 40 SSA spells are still ongoing (primary h = 5 results). Using h = 8, SSA: Complete Spells 62 (table entries fragmented).
  - Spells in SSA have been more concentrated after 1995.
- Duration and severity:
  - Median length of complete spells (primary h = 5):
    - SSA: 6 years (resource-intensive: 6 years; non-resource-intensive: 5 years).
    - Other developing economies: 10 years.
    - Emerging market economies: 8 years.
  - Annex II (h = 8) Complete Spell Duration — Median Duration of Spells (years) / Share of Spells longer than 10 years (percent):
    - Advanced: 15 — 100
    - Emerging: 10 — 80
    - Developing: 10 — 52
    - SSA: 9 — 33
    - LAC: 12 — 63
    - MENA: 9 — 43
  - Fewer SSA spells last long: only about one out of three SSA spells lasts at least 10 years (33 percent in h = 8 table) versus higher shares elsewhere.
- Growth swings before/during/after spells (complete spells, primary h = 5):
  - Median annual growth (SSA, complete spells): about −4 percent before a spell; 9 percent during spells; about −4 percent after the end of the spell.
  - Annex II (h = 8) Median Growth Before, During and After Complete Growth Spells:
    - SSA: Before −2.5 — During 5.8 — After −1.5 (percent).
  - Average annual growth after complete spells in SSA plummets from close to 10 percent during spells to −6 percent after (primary results), compared with −1 percent in other developing economies and −2 percent in emerging market economies. These patterns persist excluding conflict countries.

### Data and empirical methodology (overview)
- Determinant categories considered:
  - Exogenous: global liquidity (Fed Funds rate proxy), oil prices, terms of trade.
  - Macroeconomic: overall fiscal balance; public debt; government revenues; total and FDI; external debt; exchange rate regime/misalignment; trade openness; aid flows; inflation.
  - Institutions and social: political stability; democracy (Polity 2); internal conflicts; law and order; investment climate; schooling; human capital; infant mortality; inequality (Gini).
  - Structural/policy: export composition/concentration; financial liberalization; market distortions (relative price of investment goods); total factor productivity; technology adoption (imports of IT products; mobile use).
- Turning-point analysis:
  - One-sample t-tests on mean changes before vs after breaks; symmetry (two-sample t-) tests compare magnitudes between up-breaks and down-breaks.
  - No causal claims — descriptive stylized-fact focus.
- Duration analysis:
  - Survival (proportional hazard) models with Weibull baseline; hazard depends on lagged covariates.
  - Reported results use exponentiated coefficients (hazard ratios): >1 implies increases risk that spell ends; <1 implies protective effect.
  - Two-step approach: bivariate duration models controlling for initial income; then multivariate models with representative covariates.

### Turning points — Key empirical findings
- Exogenous factors:
  - Up-breaks in SSA and in emerging/developing economies coincide with looser global liquidity (cuts in U.S. federal funds rate proxy). Down-break evidence mixed for SSA.
  - SSA up-breaks associated with terms-of-trade improvements; down-breaks with terms-of-trade deteriorations.
  - Unlike other developing countries, SSA breaks are not significantly associated with oil price changes (reflecting mix of oil exporters/importers).
- Macroeconomic indicators:
  - Fiscal: up-breaks associated with higher government revenues; down-breaks in emerging/developing countries coincide with deteriorations.
  - Down-breaks are associated with increases in public debt-to-GDP ratios of about 11 to 16 percent of GDP (emerging/developing sample); SSA shows similar public-debt increases during down-breaks.
  - Investment: investment-to-GDP increases by 3.6 percentage points during up-breaks in emerging/developing economies and by 4.4 percentage points in SSA; investment does not significantly decline at down-breaks.
  - External debt: down-breaks associated with increases of about 15 percent of GDP in emerging/developing countries and 24 percent of GDP in SSA; up-breaks do not show significant external-debt reductions.
  - Exchange rates: SSA down-breaks associated with more flexible exchange rate regimes; undervalued real exchange rates are associated with up-breaks (significant for emerging/developing sample).
  - Trade openness: higher trade openness associates with up-breaks in both SSA and emerging/developing economies; aid flows to SSA increase during down-breaks.
  - Inflation: reductions in inflation coincide with up-breaks in SSA and emerging/developing economies; no symmetric inflation increases at down-breaks.
- Institutions and human development:
  - Political stability improves at up-breaks in SSA and elsewhere; no significant worsening at down-breaks.
  - Investment-profile (business climate) improves significantly at up-breaks, and modestly at down-breaks.
  - Internal-conflict risk declines at up-breaks; civil-war frequency does not show significant change at down-breaks.
  - Gini: inequality declines at up-breaks (more pronounced in SSA); down-breaks not associated with significant Gini increases.
  - Schooling, human capital, and infant mortality indicators improve during both up-breaks and down-breaks in SSA and emerging/developing economies.
- Structural factors:
  - Emerging/developing economies: exports of manufactures increase during both up-breaks and down-breaks; product concentration increases at up-breaks (partly explained by commodity booms).
  - SSA: evidence less conclusive due to data limits; deeper financial liberalization is associated with down-breaks only in SSA (possible policy response).
  - Total factor productivity: up-breaks coincide with significant TFP increases; down-breaks with TFP declines (robust across samples).
  - Technology proxies: mixed evidence; ICT imports accelerate in emerging/developing economies; mobile use declines at down-breaks and increases at up-breaks (not significant in SSA).

### Duration analysis — Bivariate results (selected)
- Exogenous:
  - Tighter global financial conditions (U.S. interest rate proxy) significantly increase the probability growth spells end in SSA and in emerging/developing/world samples.
  - Oil-price fluctuations do not show a significant average impact on SSA spells (heterogeneous exporter/importer effects).
- Macroeconomic:
  - Better fiscal balances significantly increase the likelihood a spell continues in SSA; higher debt burdens accelerate spell ends.
  - Investment has a protective effect: higher investment-to-GDP ratios reduce the risk of spell termination.
  - Aid flows appear to significantly increase spell duration in SSA.
- Institutions and human development:
  - Political and economic institution proxies (political stability; investment profile; institutional constraints) strongly support spell duration in SSA, more than in other regions.
  - Armed conflicts shorten spell duration in SSA.
  - Human-capital indicators (education levels; reductions in infant mortality) are associated with longer spells in SSA.
- Structural/technology:
  - Export diversification (higher share of manufactures in exports) is positively associated with longer spells in SSA.
  - Deeper financial liberalization is associated with longer spells in SSA.
  - Technology-adoption proxies (imports of computers/IT) predict longer spells in emerging/developing/world samples but are not statistically significant in SSA.

### Duration analysis — Multivariate (baseline) results and quantitative magnitudes
- Baseline multivariate model covariates include: U.S. interest rate, terms-of-trade change, oil-price change, trade openness, capital-account openness, inflation, exchange-rate regime, debt-to-GDP ratio, change in investment rate, human capital, Polity 2, relative price of investment goods, average growth during spell, and indicator for 2000–08.
- Main baseline inferences for SSA:
  - Spell duration is prolonged by: higher investment-to-GDP ratio, more trade openness, lower inflation, lower debt-to-GDP ratio, improvements in political institutions (Polity 2), and reduced market distortions.
  - Higher previous-period average growth increases the probability a spell ends (overheating risk).
  - Results robust across specifications and distributional assumptions.
- Quantified effects (change in expected spell length for a 25-percentile-point improvement from median SSA values, holding others constant):
  - Terms-of-trade: +3.4 years.
  - Oil price: +5.7 years.
  - Change in investment-to-GDP: +1.3 years.
  - Reduction in debt-to-GDP ratio: +1 year.
  - Increase in trade openness: +6 years.
  - Reduction in inflation: +5.4 years.
  - Reduced market distortions (relative price of investment): +2.5 years.
  - Less overheating (reduction in GDP per capita growth): +2.8 years.
  - Improvement in Polity 2: +7.7 years.
- Key multivariate coefficient excerpts (Table 12, exponentiated coefficients / hazard ratios, SSA column):
  - U.S. interest rates 1.021 (standard error (0.110))
  - Oil price, change 0.935 *** (0.0187)
  - Terms of trade growth 0.959 ** (0.0161)
  - Investment to GDP, change 0.923 * (0.0403)
  - Total public debt to GDP 1.007 ** (0.00343)
  - Trade openness 0.98 * (0.0117)
  - Inflation 3.242 ** (1.770)
  - Polity 2, cumulative change 0.815 ** (0.0652)
  - Real GDP per capita growth 1.115 *** (0.0321)
  - Ancillary parameter α 3.956 ***; p-value (Ho: α = 1) 0.000.
- Interpretation:
  - Hazard ratio < 1 implies protective effect (prolongs spells); > 1 implies increases risk spell ends.
  - Estimated Weibull parameter α > 1 (statistically significant) → conditional baseline hazard rises with time (end of a spell becomes more likely as spell matures), controlling for covariates.
  - Model fit: likelihood-ratio tests and pseudo-R2 measures support baseline model fit over nested alternatives; results robust to alternative functional forms and inclusion of frailty/conflict/resource-exporter indicators.

### Key contrasts: triggers vs sustainers of growth
- Factors associated with triggering up-breaks are not always the same as those sustaining spells:
  - Associated with up-breaks but not necessarily sustaining spells: reductions in Fed Funds rate, FDI increases, fewer internal conflicts, improvements in investment profile and law/order, productivity growth, export-partner diversification.
  - Factors important for sustaining spells but not necessarily changing at up-breaks: contained public debt levels; lower market distortions (relative price of investment).
  - Some factors (terms-of-trade improvements, trade openness, lower inflation) help both to jump-start and to prolong growth episodes.

### Policy implications and conclusions
- Since 2000, growth up-breaks and spells have become more common in SSA, but spells remain shorter, more volatile, and more prone to hard landings than elsewhere.
- In the current less-supportive external environment (commodity-price slump, lower aid, changing global demand), impetus to revive and sustain growth in SSA must come primarily from domestic policy action.
- Policy priorities for countries in severe slowdowns:
  - Preserve macroeconomic stability: consistent monetary policy to contain inflation; sound fiscal policy to anchor public debt.
  - Foster investment, increase trade openness, and stabilize political environment to trigger turnarounds.
- Policy priorities for countries currently in growth spells:
  - Focus on prolonging spells and avoiding hard landings: address vulnerabilities, rebuild buffers (fiscal space), limit public-debt increases (mobilize domestic revenue), and avoid overheating.
- Cross-cutting reforms to unlock growth potential:
  - Economic diversification to increase resilience.
  - Deepening technology adoption.
  - Financial-sector reforms.
  - Strengthening macro-structural policies to reduce market distortions and improve the business climate.

### Annex I — Variable list and sources (selected definitions)
- Real GDP per capita growth: 2011 U.S. dollars at PPP, extended using WEO until 2016, percent change — PWT 9.0.
- Initial GDP per capita: U.S. dollars in PPP — PWT 9.0.
- Investment to GDP: In percent of GDP — PWT 9.0.
- Trade openness: Exports plus imports, in percent of GDP — PWT 9.0.
- Overall fiscal balance: In percent of GDP — WEO.
- Total public debt: In percent of GDP — FAD.
- External debt: In percent of GDP — WEO.
- Log (1 + inflation): Consumer price index — WEO.
- Crude oil price: Simple average of Brent, WTI, and Fateh spot prices — WEO.
- Terms of trade growth: Goods, U.S. dollars, percent change — WEO.
- Political stability (Polity 2): Modified polity score +10 to -10 — Polity IV dataset.
- Years of primary schooling: Average number of years — Barro and Lee database (2013).
- Infant mortality rate: Per 1,000 live births — WDI.
- Product concentration: Index — SPR.
- ICT goods imports: Percent of total goods imports — WDI.
- TFP (Total factor productivity): Constant national prices (2011=1) — PWT 9.0.
- Financial liberalization: Index — IMF (2008).
- Conflict level: Dummy 1 War: At least 1000 battle-related deaths in one calendar year. — Uppsala Conflict Data Program database.
- Acronym notes: FAD = IMF Fiscal Affairs Department; ICRG = International Country Risk Guide; ICT = Information and Communications Technology; PPP = Purchasing Power Parity; PWT = Penn World Tables; SPR = IMF, Strategy, Policy, and Review Department; WDI = World Bank, World Development Indicators; WEO = IMF, World Economic Outlook.

*Source: Annex II, “Breaks and Spells Stylized Facts with h = 8,” wp17202 (PDF chapter), 1950–2016.*

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

### References

### Figures
- 1. Figure 1: Sub-Saharan Africa Growth Spells ..................................................................... 26
- 2. Figure 2: Effects on Growth Spells ..................................................................................... 26

### Tables
- 1. Table 1: Selected Groups: Up-breaks and Down-breaks, 1950-2016 ................................ 27
- 2. Table 2: Change in Median Annual Per Capita Growth at Up-breaks and Down-breaks .. 27
- 3. Table 3: Growth Spells, 1950-2016 .................................................................................... 27
- 4. Table 4: Duration of Complete Spells, 1950-2016 ............................................................. 27
- 5. Table 5: Median Growth Before, During and After Completed Growth Spells ................. 27
- 6. Table 6: Changes in Macroeconomic Variables during Growth Breaks ............................ 28
- 7. Table 7: Effects of Exogenous Factors on Growth Spells .................................................. 29
- 8. Table 8: Effects of Macroeconomic Variables on Growth Spells ...................................... 29
- 9. Table 9: Effects of Institutional Variables on Growth Spells ............................................. 30
- 10. Table 10: Effects of Human Development Indicators on Growth Spells ......................... 30
- 11. Table 11: Effects of Structural Transformation and Technology on Growth Spells ........ 31
- 12. Table 12: Multivariate Model of Effects on Growth Spells ............................................. 32
- 13. Table 13: Nested Models and Fit ...................................................................................... 33
- 14. Table 14: Alternative Functional Forms ........................................................................... 33
- 15. Table 15: Alternative Model Specifications ..................................................................... 34

### Annex
- Annex I: Variable List and Sources ........................................................................................ 35

*Source: wp17202 - References (page listing of figures, tables, and annex).*

### Annex II: Breaks and Spells Stylized Facts with h = 8 .......................................................... 36

### Annex II: Breaks and Spells Stylized Facts with h = 8

### Introduction and scope
- Focus: growth turning points (up-breaks and down-breaks) and growth spells in sub-Saharan Africa (SSA), compared with emerging and other developing countries.
- Sample and identification:
  - Period: 1950–2016.
  - Structural-break detection method: variant of Bai and Perron (1998) as in Antoshin, Berg, and Souto (2008).
  - Parameters used: minimum years between breaks h = 5 (primary), statistical significance p = 0.10; robustness noted for h = 8.
  - Identified breaks: 184 up-breaks and 185 down-breaks in GDP per capita growth across 146 countries over 1950–2016.

### Stylized facts — Structural breaks
- Frequency and timing:
  - Growth turning points are common globally and in SSA, with substantial variation across time and income levels.
  - About 25 percent of total up-breaks and down-breaks occur in SSA (in line with SSA representing about 28 percent of countries in the sample).
  - SSA experienced many down-breaks during 1970–84 and many up-breaks during 1985–99; since 2000 SSA has seen relatively fewer down-breaks while maintaining about the same number of up-breaks.
- Magnitude and volatility:
  - Average break sizes: per capita growth in SSA was 10 percentage points higher after an up-break versus before — the largest break size among groups examined (compare to 7 percent break size in the rest of emerging and developing economies).
  - SSA exhibits large swings between up-breaks and down-breaks, indicating high volatility.

### Stylized facts — Growth spells
- Definitions:
  - “Complete” growth spell: begins with an up-break followed by at least 2 percent average per capita income growth and ends with a down-break followed by less than 2 percent average per capita growth.
  - “Incomplete” spells: satisfy the first condition and are ongoing at sample end.
- Counts and concentration:
  - Total spells (full sample): 80 complete and 67 incomplete spells identified.
  - SSA recorded 40 growth spells between 1950 and 2016 — 27 percent of total spells; 25 of the 40 SSA spells are still ongoing.
  - Spells in SSA have been more concentrated after 1995.
- Duration and severity:
  - Median length of complete spells:
    - SSA: 6 years (resource-intensive: 6 years; non-resource-intensive: 5 years).
    - Other developing economies: 10 years.
    - Emerging market economies: 8 years.
  - Fewer SSA spells last long: only about one out of three SSA spells lasts at least 10 years (vs more than half in developing economies and about 80 percent in advanced economies).
- Growth swings before/during/after spells (complete spells, SSA):
  - Median annual growth: about −4 percent before a spell; 9 percent during spells; about −4 percent after the end of the spell.
  - Average annual growth after complete spells in SSA plummets from close to 10 percent during spells to −6 percent after, compared with −1 percent in other developing economies and −2 percent in emerging market economies.
  - These patterns persist excluding conflict countries.

### Data and empirical methodology (overview)
- Determinant categories considered:
  - Exogenous: global liquidity (Fed Funds rate proxy), oil prices, terms of trade.
  - Macroeconomic: overall fiscal balance; public debt; government revenues; total and FDI; external debt; exchange rate regime/misalignment; trade openness; aid flows; inflation.
  - Institutions and social: political stability; democracy (Polity 2); internal conflicts; law and order; investment climate; schooling; human capital; infant mortality; inequality (Gini).
  - Structural/policy: export composition/concentration; financial liberalization; market distortions (relative price of investment goods); total factor productivity; technology adoption (imports of IT products; mobile use).
- Turning-point analysis:
  - One-sample t-tests on mean changes before vs after breaks; symmetry (two-sample t-) tests compare magnitudes between up-breaks and down-breaks.
  - No causal claims — descriptive stylized-fact focus.
- Duration analysis:
  - Survival (proportional hazard) models with Weibull baseline; hazard depends on lagged covariates.
  - Reported results use exponentiated coefficients (hazard ratios): >1 implies increases risk that spell ends; <1 implies protective effect.
  - Two-step approach: bivariate duration models controlling for initial income; then multivariate models with representative covariates (to preserve sample size).

### Turning points — Key empirical findings
- Exogenous factors:
  - Up-breaks in SSA and in emerging/developing economies coincide with looser global liquidity (cuts in U.S. federal funds rate proxy). Down-break evidence mixed for SSA.
  - SSA up-breaks associated with terms-of-trade improvements; down-breaks with terms-of-trade deteriorations. Unlike other developing countries, SSA breaks are not significantly associated with oil price changes (reflecting mix of oil exporters/importers).
- Macroeconomic indicators:
  - Fiscal: up-breaks associated with higher government revenues; down-breaks in emerging/developing countries coincide with deteriorations. Down-breaks are associated with increases in public debt-to-GDP ratios of about 11 to 16 percent of GDP (emerging/developing sample); SSA shows similar public-debt increases during down-breaks.
  - Investment: investment-to-GDP increases by 3.6 percentage points during up-breaks in emerging/developing economies and by 4.4 percentage points in SSA; investment does not significantly decline at down-breaks.
  - External debt: down-breaks associated with increases of about 15 percent of GDP in emerging/developing countries and 24 percent of GDP in SSA; up-breaks do not show significant external-debt reductions.
  - Exchange rates: SSA down-breaks associated with more flexible exchange rate regimes; undervalued real exchange rates are associated with up-breaks (significant for emerging/developing sample).
  - Trade openness: higher trade openness associates with up-breaks in both SSA and emerging/developing economies; aid flows to SSA increase during down-breaks.
  - Inflation: reductions in inflation coincide with up-breaks in SSA and emerging/developing economies; no symmetric inflation increases at down-breaks.
- Institutions and human development:
  - Political stability improves at up-breaks in SSA and elsewhere; no significant worsening at down-breaks.
  - Investment-profile (business climate) improves significantly at up-breaks, and modestly at down-breaks.
  - Internal-conflict risk declines at up-breaks; civil-war frequency does not show significant change at down-breaks.
  - Gini: inequality declines at up-breaks (more pronounced in SSA); down-breaks not associated with significant Gini increases.
  - Schooling, human capital, and infant mortality indicators improve during both up-breaks and down-breaks in SSA and emerging/developing economies.
- Structural factors:
  - Emerging/developing economies: exports of manufactures increase during both up-breaks and down-breaks; product concentration increases at up-breaks (partly explained by commodity booms).
  - SSA: evidence less conclusive due to data limits; deeper financial liberalization is associated with down-breaks only in SSA (possible policy response).
  - Total factor productivity: up-breaks coincide with significant TFP increases; down-breaks with TFP declines (robust across samples).
  - Technology proxies: mixed evidence; ICT imports accelerate in emerging/developing economies; mobile use declines at down-breaks and increases at up-breaks (not significant in SSA).

### Duration analysis — Bivariate results (selected)
- Exogenous:
  - Tighter global financial conditions (U.S. interest rate proxy) significantly increase the probability growth spells end in SSA and in emerging/developing/world samples.
  - Oil-price fluctuations do not show a significant average impact on SSA spells (heterogeneous exporter/importer effects).
- Macroeconomic:
  - Better fiscal balances significantly increase the likelihood a spell continues in SSA; higher debt burdens accelerate spell ends.
  - Investment has a protective effect: higher investment-to-GDP ratios reduce the risk of spell termination (illustrative country examples: Chad, Lesotho).
  - Competitiveness, trade openness, inflation: significant in world/emerging/developing samples; less so in SSA bivariately.
  - Aid flows appear to significantly increase spell duration in SSA (new finding).
- Institutions and human development:
  - Political and economic institution proxies (political stability; investment profile; institutional constraints) strongly support spell duration in SSA, more than in other regions.
  - Armed conflicts shorten spell duration in SSA.
  - Human-capital indicators (education levels; reductions in infant mortality) are associated with longer spells in SSA.
- Structural/technology:
  - Export diversification (higher share of manufactures in exports) is positively associated with longer spells in SSA.
  - Deeper financial liberalization is associated with longer spells in SSA.
  - Technology-adoption proxies (imports of computers/IT) predict longer spells in emerging/developing/world samples but are not statistically significant in SSA.

### Duration analysis — Multivariate (baseline) results and quantitative magnitudes
- Baseline multivariate model covariates include: U.S. interest rate, terms-of-trade change, oil-price change, trade openness, capital-account openness, inflation, exchange-rate regime, debt-to-GDP ratio, change in investment rate, human capital, Polity 2, relative price of investment goods, average growth during spell, and indicator for 2000–08.
- Main baseline inferences for SSA:
  - Spell duration is prolonged by: higher investment-to-GDP ratio, more trade openness, lower inflation, lower debt-to-GDP ratio, improvements in political institutions (Polity 2), and reduced market distortions.
  - Higher previous-period average growth increases the probability a spell ends (overheating risk).
  - Results robust across specifications and distributional assumptions.
- Quantified effects (change in expected spell length for a 25-percentile-point improvement from median SSA values, holding others constant):
  - Terms-of-trade: +3.4 years.
  - Oil price: +5.7 years.
  - Change in investment-to-GDP: +1.3 years.
  - Reduction in debt-to-GDP ratio: +1 year.
  - Increase in trade openness: +6 years.
  - Reduction in inflation: +5.4 years.
  - Reduced market distortions (relative price of investment): +2.5 years.
  - Less overheating (reduction in GDP per capita growth): +2.8 years.
  - Improvement in Polity 2: +7.7 years.
- Duration dependence and model fit:
  - Estimated Weibull parameter α > 1 (statistically significant) → conditional baseline hazard rises with time (end of a spell becomes more likely as spell matures, controlling for covariates).
  - Likelihood-ratio tests and pseudo-R2 measures support baseline model fit over nested alternatives.
  - Coefficient estimates robust to alternative distributional assumptions (accelerated failure-time and other nonmonotonic baselines).
  - Including frailty terms (observation- or country-level) does not change coefficients; robustness checks adding conflict and resource-exporter indicators leave main results intact.

### Key contrasts: triggers vs sustainers of growth
- Factors associated with triggering up-breaks are not always the same as those sustaining spells:
  - Associated with up-breaks but not necessarily sustaining spells: reductions in Fed Funds rate, FDI increases, fewer internal conflicts, improvements in investment profile and law/order, productivity growth, export-partner diversification.
  - Factors important for sustaining spells but not necessarily changing at up-breaks: contained public debt levels; lower market distortions (relative price of investment).
- Some factors (terms-of-trade improvements, trade openness, lower inflation) help both to jump-start and to prolong growth episodes.

### Policy implications and conclusions
- Since 2000, growth up-breaks and spells have become more common in SSA, but spells remain shorter, more volatile, and more prone to hard landings than elsewhere.
- In the current less-supportive external environment (commodity-price slump, lower aid, changing global demand), impetus to revive and sustain growth in SSA must come primarily from domestic policy action.
- For countries in severe slowdowns:
  - Preserve macroeconomic stability (consistent monetary policy to contain inflation; sound fiscal policy to anchor public debt).
  - Foster investment, increase trade openness, and stabilize political environment to trigger turnarounds.
- For countries currently in growth spells:
  - Focus on prolonging spells and avoiding hard landings: address vulnerabilities, rebuild buffers (fiscal space), limit public-debt increases (mobilize domestic revenue), and avoid overheating.
- Across all countries, priority reforms to unlock growth potential:
  - Economic diversification to increase resilience.
  - Deepening technology adoption.
  - Financial-sector reforms.
  - Strengthening macro-structural policies to reduce market distortions and improve the business climate.

*Source: Annex II, “Breaks and Spells Stylized Facts with h = 8,” wp17202 (PDF chapter), 1950–2016.*

### References

### References

### Bibliographic entries
- Acemoglu, D., S. Johnson, J. Robinson, and Y. Thaicharoen. 2003. “Institutional Causes, Macroeconomic Symptoms: Volatility, Crises, and Growth.” Journal of Monetary Economics 50: 49–123.  
- Aizenman J., and M. Spiegel. 2010. “Takeoffs.” Review of Development Economics, 14(2): 177–196.  
- Antoshin, S., A. Berg, and M. Souto. 2008. “Testing for Structural Breaks in Small Samples.” Working Paper No. 75. International Monetary Fund, Washington, DC.  
- Bai, J., and P. Perron. 1998. Estimating and Testing Linear Models with Multiple Structural Change. Econometrica 66: 47–78.  
- Barro, R. J. 1991. “Economic Growth in a Cross Section of Countries.” The Quarterly Journal of Economics 106: 407–43.  
- Becker, T., and P. Mauro. 2006. “Output Drops and the Shocks that Matter.” IMF Working Paper No. 06/172. International Monetary Fund, Washington, DC.  
- Berg, A., J. Ostry, and J. Zettelmeyer. 2012. What Makes Growth Sustained? Journal of Development Economics 98: 149–66.  
- Bluhm, R., D. de Crombrugghe, and A. Szirmai. 2014. “Do Weak Institutions Prolong Crises? On the Identification, Characteristics, and Duration of Declines during Economic Slumps.” Center for Economic Studies and Information Working Paper No. 4594.  
- Hausmann, R., L. Pritchett, and D. Rodrik. 2005. “Growth Accelerations.” Journal of Economic Growth 10: 303–29.  
- Hausmann, R., F. Rodriguez, and R. Wagner. 2006. “Growth Collapses.” Working Paper No. 136. Center for International Development, Cambridge, MA.  
- Hsieh, C-T and P. Klenow. 2007. “Relative Prices and Relative Prosperity.” American Economic Review, 98(3): 562–85.  
- International Monetary Fund (IMF). 2008. “Structural Reforms and Economic Performance in Advanced and Developing Countries.” Research Department Board Paper. International Monetary Fund, Washington, DC.  
- International Monetary Fund (IMF). 2017. “Regional Economic Outlook for sub-Saharan Africa (April 2017).” African Department. International Monetary Fund, Washington, DC.  
- Jerzmanowski, M. 2006. “Empirics of Hills, Plateaus, Mountains and Plains: A Markov-Switching Approach to Growth.” Journal of Development Economics 81: 357–85.  
- Jones, B. F., and B. A. Olken. 2008. “The Anatomy of Start-Stop Growth.” Review of Economics and Statistics 90: 582–87.  
- Kerekes, M. 2012. “Growth Miracles and Failures in a Markov Switching Classification Model of Growth” Journal of Development Economics 98: 167–77.  
- Krippner, L. 2016. “Documentation for Measures of Monetary Policy.” Reserve Bank of New Zealand. Wellington, New Zealand.  
- Lederman, D., and W.F. Maloney. 2003. “Trade Structure and Growth.” Policy Research Working Paper No. 3025. World Bank, Washington, DC.  
- Mora, R., and G. Siotis. 2005. “External factors in emerging market recoveries: An empirical investigation.” European Economic Review, 49(3): 683-702.  
- Pritchett, L. 2000. “Understanding Patterns of Economic Growth: Searching for Hills Among Plateaus, Mountains, and Plains.” World Bank Economic Review 90: 221–50.  
- Rajan, R.G., and A. Subramanian. 2008. “Aid and Growth: What Does the Cross-Country Evidence Really Show?” Review of Economics and Statistics 90: 643–65.  
- Robinson, D., M. Gaertner, and C. Papageorgiou. 2011. “Tanzania: Growth Acceleration and Increased Public Spending with Macroeconomic Stability.” In Yes, Africa Can: Success Stories from a Dynamic Continent, edited by P. Chuhan-Pole and M. Angwafo. Washington, DC: World Bank.  
- Rodrik, D. 2008. “The Real Exchange Rate and Economic Growth.” Brookings Papers on Economic Activity: Fall.  
- Tsangarides, C. G. 2012. “Determinants of Growth Spells: Is Africa Different?” IMF Working Paper No. 12/227. International Monetary Fund, Washington, DC.  
- Wooldridge, J. M. 2002. Econometric Analysis of Cross-Section and Panel Data. Cambridge, MA: The MIT Press.

### Figures
- Figure 1. Sub-Saharan Africa Growth Spells  
- Figure 2. Effects on Growth Spells  
  - Axis / items shown (as presented): Trade openness; Inflation; Investment to GDP; Total public debt; Oil price, change; Terms of trade growth; Polity 2, cumulative change; Real GDP per capita growth; Relative price of investment to consumption; grouped under Macroeconomic Variables, External Shocks, Institutions, Other.  
  - Note (verbatim): "Each bar shows the change in the expected duration of a growth spell (in numbers of years) if a variable improves by 25 percentile points from the median value in the sample of sub-Saharan African countries, while holding other variables constant."  
  - Source line (verbatim): "Source: Authors’ calculations."  
  - Additional labels present in figure area: "No spells", "Oil exporters", "Other resource-intensive countries", "Nonresource-intensive countries", "1960–94", "1995–2016".  
  - Note (verbatim): "Note: Each bar shows the change in the expected duration of a growth spell (in numbers of years) if a variable improves by 25 percentile points starting from the median."

### Selected table captions and key numeric excerpts (verbatim)
- Table 1. Selected Groups: Up-breaks and Down-breaks, 1950-2016  
- Table 2. Change in Median Annual Per Capita Growth at Up-breaks and Down-breaks, percentage points  
- Table 3. Growth Spells, 1950-2016  
- Table 4. Duration of Complete Spells, 1950-2016  
- Table 5. Median Growth Before, During and After Completed Growth Spells, 1950-2016  
  - Excerpt (verbatim rows/columns as presented):  
    - Up-Breaks  Down-Breaks  Up-Breaks  Down-Breaks  Up-Breaks  Down-Breaks  Up-Breaks  Down-Breaks  
      Total3018478067404047  
      Advanced836168629  
      Emerging4331197105  
      Developing1812385350272833  
      SSA5  5 1220191411 8  
      LAC54111513759  
      Asia20975176  
      MENA6 3 611135 510  
      1950-19691970-841985-992000-16  
    - RegionsUp-Breaks  Down-Breaks  
      Total 7.3-7.4  
      Advanced3.8-4.6  
      Emerging8.0-7.2  
      Developing8.0-8.6  
      SSA9.9-9.3  
      LAC5.7-6.8  
      Asia6.4-8.7  
      MENA9.7-10.4  
    - Regions Complete Spells Ongoing Spells  
      Total8067  
      Advanced138  
      Emerging913  
      Developing5846  
      SSA1525  
      LAC189  
      Asia78  
      MENA184  
    - Regions Median Duration of Spells (years) Share of Spells longer than 10 years (percent)  
      Advanced1577  
      Emerging833  
      Developing845  
      SSA633  
      LAC844  
      Asia1357  
      MENA1050  
    - Regions Before During After  
      Advanced1.55.10.4  
      Emerging-1.68.0-1.8  
      Developing-1.37.4-0.4  
      SSA-3.89.2-3.9  
      LAC-0.75.90.3  
      Asia-0.15.9-0.4  
      MENA-0.48.40.1

- Table 6. Changes in Macroeconomic Variables during Growth Breaks  
  - Excerpts (verbatim numeric entries and notes as presented):  
    - U.S. federal funds rate -2.4 *** -0.8 *** +++ -2.9 *** -0.8 +++  
    - Crude oil price, change 2.7 * -8.7 *** +++ -1.5 -3.9  
    - Terms of trade, change 0.5 *** -0.4 *** 0.5 ** -0.6 *  
    - Overall fiscal balance to GDP 5.3 * -1.8 ** 8.4 -0.7  
    - Total public debt to GDP 0.01 11.5 *** +++ -9.0 15.7 **  
    - Investment to GDP, ratio 3.6 *** 0.1 +++ 4.4 *** -0.2 +  
    - External debt to GDP -0.3 15.5 *** +++ -9.4 23.7 ***  
    - Trade to GDP 9.3 *** 3.2 ** +++ 9.0 ** 3.9  
    - Aid to GDP -0.6 0.9 -0.1 5.0 *** ++  
    - Political stability 6.5 *** 1.1 +++ 4.1 ** 2.1  
    - Years of Schooling 1.3 *** 1.2 *** 1.3 *** 1.3 ***  
    - Infant mortality rate -24.8 *** -20.4 *** ++ -27.3 *** -23.6 ***  
  - Footnotes (verbatim): "1 changes multiplied by 100. 2 changes multiplied by 1000."  
  - Source/Note (verbatim): "Source: Author's calculations. Note: *, **, and *** indicate that the changes in means are significant at the 10, 5 and 1 percent levels. +, ++, and +++ indicate that the absolute magnitudes of changes in means between up-breaks and down-breaks are significantly different at the 10, 5, and 1 percent levels."

- Table 7. Effect of Exogenous Factors on Growth Spells  
  - Excerpt (verbatim):  
    - Variables US interest rates 1.11 * 1.082 ** 1.078 ** 1.072  
    - Oil price, percent change 0.999 1.006 *** 1.006 *** 1.007 ***  
    - Terms of trade, percent change 0.997 0.991 0.992 0.986  
    - Spells 371 141 347 7  
    - Failures 125 66 844  
    - Observations 6,201 1,904 2,334 1,284  
  - Note (verbatim): "The table reports exponentiated regression coefficients (so-called “hazard ratios”), where a hazard ratio larger than 1 implies that increases in the associated variable shortens spells, while a ratio smaller than 1 implies that the covariate has a “protective” effect, that is, it helps sustain the spell. EMEDEV = all emerging and developing economies; EMEDEV x SSA = all emerging and developing economies excluding sub-Saharan Africa; SSA = sub-Saharan Africa. *, **, and *** indicate statistical significance at the 10, 5, and 1 percent levels."

- Table 8. Effect of Macroeconomic Variables on Growth Spells  
  - Excerpts (verbatim rows):  
    - Overall fiscal balance to GDP, initial 0.995 *** 0.998 0.998 1.030 (Spells 215 258 31 ; Failures 418 211 14 ; Observations 311 668 803 357)  
    - Total public debt to GDP 1.004 * 1.001 1.001 0.993 (Spells 381 221 4384 ; Failures 136 275 49 ; Observations 634 1,972 2,474 1,338)  
    - Government revenue to GDP, change 0.982 0.983 * 0.983 * 0.923 * (Spells 329 911 467 ; Failures 740 473 3 ; Observations 458 1,371 1,666 913)  
    - Investment to GDP, change 0.941 *** 0.948 *** 0.955 ** 0.956 (Spells 401 261 4786 ; Failures 156 780 52 ; Observations 648 2,085 2,589 1,437)  
    - Exchange rate regime (de facto) 0.893 0.851 ** 0.868 * 0.832 ** (Spells 351 031 1968 ; Failures 942 4933 ; Observations 529 1,602 1,994 1,073)  
    - Aid percent of GDP 0.905 *** 0.977 0.983 1.044 (Spells 371 201 2683 ; Failures 115 961 48 ; Observations 610 1,906 2,070 1,296)  
    - Log (1+ inflation) 0.933 1.084 1.110 1.129 (Spells 371 191 4082 ; Failures 125 972 47 ; Observations 617 1,980 2,477 1,363)

- Table 9. Effect of Institutional Variables on Growth Spells  
  - Excerpts (verbatim):  
    - Political stability (Polity 2) 0.815 *** 0.929 *** 0.93 *** 0.948 ** (Spells 381 181 3880 ; Failures 146 073 46 ; Observations 636 1,991 2,446 1,355)  
    - Political stability (Polity 2, cumulative change) 0.88 *** 0.946 ** 0.954 ** 0.969 (Spells 351 111 2976 ; Failures 135 870 45 ; Observations 534 1,776 2,163 1,242)  
    - More constraints on the Executive (initial) 0.693 * 0.835 ** 0.876 ** 0.85 ** (Spells 311 061 2575 ; Failures 105 467 44 ; Observations 505 1,730 2,159 1,225)  
    - Fewer civil war 0.305 *** 0.897 0.937 1.465 (Spells 216 779 46 ; Failures 318 2215 ; Observations 244 792 937 548)

- Table 10. Effect of Human Development Indicators on Growth Spells  
  - Excerpts (verbatim):  
    - Years of primary schooling 0.933 0.964 *** 0.973 ** 0.964 ** (Spells 261 031 2477 ; Failures 953 6644 ; Observations 458 1,776 2,288 1,318)  
    - Human capital 0.154 * 0.286 *** 0.294 *** 0.297 *** (Spells 341 111 3277 ; Failures 135 770 44 ; Observations 523 1,808 2,312 1,285)  
    - Human capital (Initial) 0.208 0.359 ** 0.357 *** 0.345 ** (Spells 341 111 3277 ; Failures 135 770 44 ; Observations 544 1,862 2,374 1,318)  
    - Infant mortality rate 1.024 *** 1.013 *** 1.015 *** 1.011 *** (Spells 401 251 4585 ; Failures 156 678 51 ; Observations 669 2,118 2,617 1,449)

- Table 11. Effect of Structural Transformation and Technology Adoption on Growth Spells  
  - Excerpts (verbatim):  
    - Exports of manufactures 0.953 ** 0.989 * 0.988 ** 0.988 * (Spells 311 141 3583 ; Failures 854 674 46 ; Observations 383 1,566 2,050 1,183)  
    - Product concentration 1.242 1.25 * 1.331 *** 1.336 ** (Spells 361 211 4285 ; Failures 146 577 51 ; Observations 458 1,706 2,169 1,248)  
    - Partner concentration 2.344 *** 1.548 ** 1.653 *** 1.374 (Spells 361 211 4285 ; Failures 146 578 51 ; Observations 458 1,706 2,156 1,248)  
    - Financial liberalization 0.000128 ** 0.122 ** 0.119 *** 0.212 (Spells 146 781 53 ; Failures 625 2919 ; Observations 118 660 944 542)  
    - Imports of computers 0.322 0.492 ** 0.538 *** 0.493 ** (Spells 278 910 562 ; Failures 432 4028 ; Observations 340 1,100 1,389 760)

- Table 12. Multivariate Model of Effects on Growth Spells  
  - Excerpts (verbatim):  
    - SSA EMEDEV EMEDEV without SSA (column headers)  
    - External Factors: U.S. interest rates 1.021 1.046 1.037 (standard errors (0.110)(0.0437)(0.0529))  
    - Oil price, change 0.935 *** 1.005 1.032 (0.0187)(0.0209)(0.0293)  
    - Terms of trade growth 0.959 ** 0.996 0.996 (0.0161)(0.0135)(0.0165)  
    - Macroeconomic Variables: Investment to GDP, change 0.923 * 0.962 0.953 (0.0403)(0.0429)(0.0332)  
    - Total public debt to GDP 1.007 ** 1.006 ** 1.007 (0.00343)(0.00258)(0.00812)  
    - Trade openness 0.98 * 0.992 0.994 (0.0117)(0.00686)(0.0109)  
    - Inflation 3.242 ** 1.472 * 1.527 * (1.770)(0.302)(0.343)  
    - Institutions: Polity 2, cumulative change 0.815 ** 0.963 1.003 (0.0652)(0.0280)(0.0311)  
    - Social and Human Development: Human capital 0.159 0.464 0.355 (0.355)(0.333)(0.322)  
    - Other: Real GDP per capita growth 1.115 *** 1.049 *** 1.064 ** (0.0321)(0.0138)(0.0286)  
    - Initial GDP per capita 4.691 1.531 1.052 (6.359)(0.454)(0.422)  
    - Observations 326 1,076 750; Spells 288 153; Failures 829 21  
    - Ancillary parameter α 3.956 *** 1.557 *** 1.342 ***  
    - Ancillary parameter p-value (Ho: α = 1) 0.000 .000 .07  
  - Note (verbatim): "The table reports exponentiated regression coefficients (so-called “hazard ratios”), where a hazard ratio larger than 1 implies that increases in the associated variable shortens spells, while a ratio smaller than 1 implies that the covariate has a “protective” effect, that is, it helps sustain the spell. EMEDEV = all emerging and developing countries; SSA = sub–Saharan Africa. *, **, and *** indicate statistical significance at the 10, 5, and 1 percent levels."

- Table 13. Nested Models and Fit  
  - Excerpt (verbatim):  
    - Models Baseline Shocks Macro & Shocks Macro & Shocks & Social & Institutions No Institutions No Growth  
    - Log Likelihood -6.327 -17.990 -12.940 -10.240 -8.799 -9.558  
    - LR test statistic 23.340 13.220 7.828 4.944 6.464  
    - LR p-value 0.0160 0.0210 0.0200 0.0260 0.011  
    - Pseudo R-squared 0.693 0.126 0.372 0.503 0.573 0.536

- Table 14. Alternative Functional Forms  
  - Excerpt (verbatim): baseline and alternative functional-form coefficient excerpts (see table for detailed entries such as Oil price, change 0.935 *** 1.017 ** 1.03 *** 1.025 *** with associated standard errors, and Ancillary parameter¹ and p-values².)

- Table 15. Alternative Model Specifications  
  - Excerpt (verbatim): columns Baseline Conflict Resource Exporters Conflict and Resource Exporters No decade dummy with selected coefficient excerpts such as:  
    - U.S. interest rates 1.021 1.022 1.017 1.031 1.019 (standard errors shown)  
    - Oil price, change 0.935 *** 0.944 *** 0.942 *** 0.948 *** 0.939 **  
    - Terms of trade growth 0.959 ** 0.951 ** 0.96 * 0.953 * 0.959 **  
    - Observations 326 across columns; Spells 282; Failures 88  
    - Ancillary parameter α 3.956 4.502 4.095 4.729 3.845  
    - Ancillary parameter p-value (Ho: α = 1) 4.89e-07 5.90e-05 9.66e-08 9.84e-05 3.57e-06

*Source: wp17202 - References (authors’ calculations and table/figure captions as presented in the source PDF).*

### Annex I. Variable list and Sources

### Annex I. Variable list and Sources

### Variable definitions and sources
- Real GDP per capita growth: 2011 U.S. dollars at PPP, extended using WEO until 2016, percent change — PWT 9.0
- Initial GDP per capita: U.S. dollars in PPP — PWT 9.0
- Investment to GDP: In percent of GDP — PWT 9.0
- Trade openness: Exports plus imports, in percent of GDP — PWT 9.0
- Population growth: Percent change — PWT 9.0
- Overall fiscal balance: In percent of GDP — WEO
- Total government revenue: In percent of GDP — WEO
- Foreign direct investment, net: In percent of GDP — WEO
- Total public debt: In percent of GDP — FAD
- External debt: In percent of GDP — WEO
- Log (1 + inflation): Consumer price index — WEO
- Crude oil price: Simple average of Brent, WTI, and Fateh spot prices — WEO
- Terms of trade growth: Goods, U.S. dollars, percent change — WEO
- Capital account openness: Normalized from 0 to 1 — Chinn-Ito database (2014)
- Shadow federal funds rate: Interest rate — Krippner (2016)
- Political stability (Polity 2): Modified polity score +10 (strongly democratic) to -10 (strongly autocratic) — Polity IV dataset
- More constraints on the Executive (Executive constraints): 0–7 scale, 1 is unlimited authority and 7 is executive parity — Polity IV dataset
- Political stability (Political risk): Composite index, 0–100 scale, higher numbers suggest improvement — ICRG dataset
- Investment profile: 0–12 scale, higher numbers suggest improvement — ICRG dataset
- Fewer internal conflict (Internal conflict): 0–12 scale, higher numbers suggest improvement — ICRG dataset
- Law & order: 0–6 scale, higher numbers suggest improvement — ICRG dataset
- Fewer civil war (Civil war): 0–4 scale, higher numbers suggest improvement — ICRG dataset
- Years of primary schooling: Average number of years — Barro and Lee database (2013)
- Human capital: Index — PWT 9.0
- Aid: Net official development assistance and aid received in percent of GDP — WDI
- Gini coefficient: Gini index (World Bank estimate) — WDI
- Infant mortality rate: Per 1,000 live births — WDI
- De facto exchange rate regime: DF: hard=1 conventional=2 basket=3 band=4 crawl=5 managed=6 independent=7 — October 2016 REO

### Exchange rate misalignment calculations (Chapter 2)
- Exchange rate misalignment: In percent of equilibrium level / Dummy variable — IMF staff calculation/PWT
- Product concentration: Index — SPR
- Partner concentration: Index — SPR
- Manufactured exports: Percent of merchandise exports — WDI
- ICT goods imports: Percent of total goods imports — WDI
- Mobile phone lines: Per 100 people — WDI
- Imports of computers: Percent of total imports — Comtrade
- TFP (Total factor productivity): Constant national prices (2011=1) — PWT 9.0
- Relative price of investment to consumption: Ratio — PWT 9.0
- Financial liberalization: Index — IMF (2008)
- Fewer financial restrictions on current account: Index — IMF (2008)
- Conflict level: Dummy 1 War: At least 1000 battle-related deaths in one calendar year. — Uppsala Conflict Data Program database
- Energy exporters: Dummy 0-1 — April 2016 REO Ch2

### Notes on acronyms
- FAD = IMF Fiscal Affairs Department; ICRG = International Country Risk Guide; ICT = Information and Communications Technology; PPP = Purchasing Power Parity; PWT = Penn World Tables; SPR = IMF, Strategy, Policy, and Review Department; WDI = World Bank, World Development Indicators; WEO = IMF, World Economic Outlook.

### Annex II. Breaks and Spells Stylized Facts with h = 8

### Table II.1. Selected Groups: Up-breaks and Down-breaks, 1950-2016
- Total: Up-Breaks 177 Down-Breaks 165
- Advanced: Up-Breaks 4 Down-Breaks 22
- Emerging: Up-Breaks 5 Down-Breaks 20
- Developing: Up-Breaks 83 Down-Breaks 143
- SSA: Up-Breaks 21 Down-Breaks 51
- LAC: Up-Breaks 51 Down-Breaks 11
- Asia: Up-Breaks 0 Down-Breaks 0 6 2 5 2 4 3 (table displays fragmented entries)
- MENA: Up-Breaks 1 1 2 6 1 4 6 2 5 (table displays fragmented entries)
- Periods shown: 1950-1969, 1970-84, 1985-99, 2000-16
- Source: Authors' calculations.
- Note: LA C = Latin America and the Caribbean; M EN A = Middle East and North Africa; SS A = Sub–Saharan Africa. Calculations using h= 8 and p=0.10.

### Table II.2. Change in Median Annual Per Capita Growth at Up-breaks and Down-breaks with h = 8, percentage points
- Regions: Up-Breaks 5.9- Down-Breaks -5.4
- Advanced: Up-Breaks 4.0- Down-Breaks -3.3
- Emerging: Up-Breaks 4.5- Down-Breaks -5.3
- Developing: Up-Breaks 6.6- Down-Breaks -6.9
- SSA: Up-Breaks 7.7- Down-Breaks -7.6
- LAC: Up-Breaks 4.9- Down-Breaks -5.2
- Asia: Up-Breaks 5.7- Down-Breaks -6.6
- MENA: Up-Breaks 10.2- Down-Breaks -10.4
- Source: Authors' calculations. Results using h=8 and p=0.10.

### Table II.3. Growth Spells with h = 8, 1950-2016
- Complete Spells: 366
- Ongoing Spells: 63
- Advanced: Complete Spells 6 2 (table fragmented)
- Emerging: Complete Spells 5 11 (table fragmented)
- Developing: Complete Spells 255 Ongoing Spells 0
- SSA: Complete Spells 62 5 (table fragmented)
- LAC: Complete Spells 8 10 (table fragmented)
- Asia: Complete Spells 48
- MENA: Complete Spells 7 7 (table fragmented)
- Source: Authors' calculations. Results using h=8 and p=0.10.

### Table II.4. Complete Spell Duration with h = 8, 1950-2016
- Regions — Median Duration of Spells — Share of Spells longer than 10 years
- Advanced: 15 — 100
- Emerging: 10 — 80
- Developing: 10 — 52
- SSA: 9 — 33
- LAC: 12 — 63
- Asia: 2 27 5 (table shows fragmented entry)
- MENA: 9 — 43
- Source: Authors' calculations. Results using h=8 and p=0.10.

### Table II.5. Median Growth Before, During and After Complete Growth Spells with h = 8, 1950-2016
- Regions — Before — During — After
- Advanced: 1.2 — 5.2 — 1.1
- Emerging: -0.4 — 5.5 — -1.6
- Developing: 0.2 — 6.4 — -0.2
- SSA: -2.5 — 5.8 — -1.5
- LAC: 0.5 — 5.6 — -0.1
- Asia: 0.3 — 6.2 — 0.2
- MENA: -0.1 — 11.0 — 0.6
- Source: Authors' calculations. Results using h=8 and p=0.10.

*Source: wp17202 - Annex I. Variable list and Sources*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2017/wp17202.pdf_