## 11. Summary Regressions

## Source details

**Canonical URL:** [11. Summary Regressions](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0859.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0859.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0859.pdf.json)

---

### I. Introduction — motivation and approach
- Central question: how to generate sustained economic growth in low-income and emerging market countries, emphasizing the importance of growth duration rather than average growth.
- Core argument: developing-country output paths exhibit low persistence; studying turning points (accelerations and decelerations) and the duration of growth spells is more informative than cross-country averages.
- Methodological contribution: focus on duration of growth spells using duration analysis techniques (proportional hazard / survival models) and distinguishing initial conditions at the time of an acceleration from changes that occur during a spell.
- Related literature: builds on Pritchett (2000), Rodrik (1999), Hausmann et al. (2005, 2006), Jerzmanowski (2006), and others; differs by identifying two regimes via structural breaks and then applying duration analysis to many covariates simultaneously.

### II. Structural breaks and definition of growth spells
- Break detection method:
  - Variant of Bai-Perron (1998, 2003) with sample-specific critical values (accounting for heteroskedasticity and small sample size) and extended sequential testing.
  - Minimum interstitiary periods h set to either 5 or 8 years to balance power and reliability.
  - Break location chosen by minimizing sum of squared residuals; critical values from Monte Carlo with bootstrapped residuals.
- Data: income per capita growth series for 140 countries from Penn World Tables version 6.2 extended to 2006 with IMF WEO.
- Key counts and patterns:
  - At p = 0.10 and h = 5: total of 290 breaks — 139 upbreaks and 151 downbreaks.
  - Standard Bai-Perron algorithm (same data, p-value, h) identifies only 74 breaks.
  - At p = 0.10 and h = 8: 171 breaks — 76 upbreaks and 95 downbreaks.
  - Upbreaks concentrated in the 1950s–60s (Europe, Latin America) and the 1990s (Africa); downbreaks concentrated in the 1970s (high-income: early 1970s; Latin America: 1978–1983; Africa: 1970s–early 1980s).
- Growth spell definition (statistical + economic):
  - Begins with a statistical upbreak followed by at least g percent average per capita growth.
  - Ends with a statistical downbreak followed by less than g percent average growth (complete) or with the end of the sample (incomplete).
  - Growth cutoff g examined at g = 2, g = 2.5, g = 3; the paper focuses on g = 2.
- Frequency and duration of spells (Table 2 summary):
  - Using h = 5 and p = 0.10: total spells = 103 (complete + incomplete); using p = 0.25: total spells = 160.
  - Using h = 8 and p = 0.10: total spells = 62; p = 0.25: total spells = 91.
  - Regional patterns: mean spell length much shorter in Latin America and Africa versus industrial countries and emerging Asia; for h = 8, 80–100 percent of spells in high income countries and emerging Asia last 10 years or more, but only ~two thirds for Latin America and Africa.
  - Asymmetry: Latin America has many past (short) spells but few ongoing spells (3–5); Africa has many ongoing spells (15–23) initiated mid–late 1990s and thus short on average.
- Growth dynamics before/during/after spells (Table 3 summary):
  - Similar growth levels during spells across regions (except Latin America somewhat lower).
  - After spells: advanced countries and emerging Asia show "soft landings" (growth between -1 and 3 percent); African spells tend to end with deep collapses (average growth between -3 and -6 percent).
  - Before spells: high-income and Asia start from positive or slightly negative per capita growth; other developing regions often begin spells from crises (interstitiary rates prior to spells between -1.3 and -5 percent).

### III. Analyzing duration — empirical strategy and methodology
- Objective: relate expected duration (or hazard that a spell will end) to initial conditions and time-varying covariates; distinguish initial level vs. change within spell.
- Empirical challenges:
  - Model selection is atheoretical and sequential due to data limitations and small number of spells (sample of spells no more than about 75 under stringent significance).
  - Data availability constraints prevent large multivariate specifications; sequential testing with minimal controls preferred.
  - Endogeneity: time-varying covariates treated as predetermined by using lagged realizations and proportional hazard model with time-varying covariates to reduce contemporaneous feedback bias.
- Regression methodology:
  - Proportional hazard (survival) model estimated via maximum likelihood; baseline hazard parameterized (Weibull distribution assumed in practice).
  - Hazard ratios (exponentiated coefficients) reported: value > 1 increases risk; value < 1 is protective.
  - Main control variables included: terms of trade shocks, U.S. interest rate changes, initial income per capita; many robustness checks and alternative samples (h = 5 or 8; p = 0.10 or 0.25).

### IV. Main empirical results (by category)
- External shocks (Table 4):
  - Terms of trade: hazard ratios ~0.97–0.98 — a one percent improvement in terms of trade reduces probability of a downbreak by 2–3 percent (not always precisely estimated).
  - U.S. interest rate changes: large significant effect; a one percentage point increase in U.S. rates estimated to increase probability that a growth spell will end next year by 25–50 percent (samples vary).
- Political and economic institutions (Table 5):
  - "Polity2" (scale -10 to +10): a one point improvement lowers hazard by about 3–9 percent (initial level and within-spell change have similar effects).
  - Collapsing initial and change terms to contemporaneous polity improves precision: reduction in hazard from one point increase about 4–8 percent.
  - Democracy vs. Autocracy components: autocracy score exhibits stronger effect — a one point increase in autocracy raises probability a spell ends by 10–20 percent.
  - Concept indices: executive recruitment, executive constraints, political competition generally protective (hazard ratios < 1 and significant in many subsamples).
  - ICRG "investment profile" (0–12 scale): suggests a protective effect (one point improvement associated with up to 35 percent reduction in hazard) but imprecisely estimated and significant in one subsample.
- Inequality and fractionalization (Table 6):
  - Inequality (Gini): strong association — a one percentage point higher Gini lowers hazard that growth will continue (i.e., increases risk) by between 4 and 14 percent (large effect given cross-sectional SD of Gini > 10 percentage points).
  - Effect primarily from cross-sectional initial Gini; within-spell Gini changes estimated imprecisely and not significant.
  - Ethnic fractionalization (Alesina et al., 2003): sign as expected but not robustly significant; estimates unstable across subsamples.
  - Direct violent conflict measures (UCDP/PRIO): no strong robust associations found; possible interpretation is conflict affects growth primarily after spells end.
- Social and physical indicators (Table 7):
  - Education: no robust significant associations; possible exception — within-spell improvements in primary education may be beneficial (imprecise).
  - Health: infant mortality increases associated with higher hazard (one death per 100 increase raises hazard by about 10 percent in larger samples when Gini excluded); estimates imprecise and sample-dependent.
  - Physical infrastructure: telephone mainlines per 100 people hazard ratios < 1 but significant in one subsample.
- Globalization (Table 8):
  - Trade liberalization (Wacziarg-Welch dummy): very large protective effect — liberalizers show ~70–80 percent reduction in hazard that a growth spell ends (robust across timing: initial level or change within spell).
  - Trade openness (adjusted): small protective effect — a one percent of GDP increase in openness reduces chance a spell ends by 1–2 percent; significant in only one sample.
  - Financial integration (sum of external assets and liabilities): weak effects; aggregate integration changes may marginally increase risk.
  - Disaggregated external finance: debt liabilities accumulation associated with slightly higher risk; FDI liabilities associated with protective effect — 1 percent of GDP increase in FDI liabilities reduces probability of a downbreak by 4–7 percent.
- Current account, competitiveness, and export structure (Table 9):
  - Current account surplus: one percentage point of GDP rise in surplus lowers probability a growth spell will end by 5–20 percent (imprecise).
  - Overvaluation: on h = 8 samples, each percentage point of overvaluation increases hazard by 1–2 percent.
  - Domestic savings: hazard ratios below 1 but not significant.
  - Manufacturing exports share:
    - Initial level less important than change: a one percentage point increase in manufacturing exports share during a spell associated with a 2–4 percent reduction in hazard.
  - Export sophistication measures:
    - Hausmann, Hwang and Rodrik sophistication: positive associations; change within spell significant in several samples.
    - "Open Forest" (Hausmann, Rodriguez, Wagner): strong protective effects; one standard deviation increase in log "Open Forest" lowers risk of spell ending by about 40–50 percent (large effect); change within spell highly significant.
- Macroeconomic stability (Table 10):
  - Inflation (100*ln(1+π)): a one point increase raises annual risk of a downbreak by 1–4 percent (sample dependent); moderate inflation is economically meaningful — e.g., inflation rising from 10 percent to 50 percent (≈ 30 point rise in 100*ln(1+π)) could raise annual risk by up to 120 percent relative to baseline.
  - Exchange rate depreciation (parallel rate): one point increase raises risk by 2–6 percent.
  - Results robust to dropping observations with inflation/depreciation > 50 percent (i.e., not driven only by hyperinflation outliers).
- Joint (summary) regressions (Table 11):
  - Joint inclusion weakens some individual effects (e.g., inflation may lose significance), but several variables remain robust predictors:
    - Income distribution (Gini) and autocracy score (or polity2) are among the most robust predictors of shorter duration.
    - "Open Forest" remains significant in a summary specification.
    - Terms of trade changes and U.S. interest rate changes retain importance.
  - Duration dependence (Weibull parameter p) typically positive but modest (p between ~0.95 and ~1.90 across models).
  - Results hold when Asian observations excluded, indicating findings reflect general features of growth booms, not driven solely by East Asia.

### V. Key substantive findings — concise bullets
- Structural breaks:
  - At p = 0.10, h = 5: 290 breaks identified (139 upbreaks, 151 downbreaks).
  - At p = 0.10, h = 8: 171 breaks identified (76 upbreaks, 95 downbreaks).
- Spell frequency and duration:
  - Total spells (g = 2): 103 (p = 0.10, h = 5); 160 (p = 0.25, h = 5); 62 (p = 0.10, h = 8); 91 (p = 0.25, h = 8).
  - High-income and emerging Asia spells more likely to last ≥ 10 years (80–100 percent for h = 8); Latin America and Africa much less so (about two thirds or lower).
- Predictors that prolong growth spells (robust across many specifications):
  - More equal income distribution (lower Gini).
  - Better political institutions / lower autocracy (higher polity2, executive constraints, political competition).
  - Trade liberalization (Wacziarg-Welch dummy) and trade openness (weaker).
  - FDI liabilities (composition of external finance matters: FDI protective; debt liabilities less so).
  - Export structure: rising manufacturing exports share during spells; higher export sophistication and higher "Open Forest".
  - Macroeconomic stability: lower inflation and fewer large exchange rate depreciations; current account surpluses.
- External shocks raise risk of spell termination:
  - Terms of trade declines increase hazard; U.S. interest rate increases have large estimated effects (one percentage point raise increases hazard by 25–50 percent in many samples).

### VI. Interpretations and implications
- Synthesis: duration of growth spells is associated with political economy and structural features — inequality, institutions, openness and export structure, macroeconomic stability, and external shocks all matter.
- Suggested channels:
  - Inequality and social fragmentation may shorten growth spells via social/political conflict and weaker institutions.
  - Export orientation toward manufacturing and product sophistication can build constituencies and generate positive spillovers that sustain growth.
  - Composition of external finance matters: FDI supports sustained growth more than debt accumulation.
  - Macroeconomic volatility (inflation, depreciation) and adverse external shocks increase risk of spell termination.
- Policy-relevant implications (inferred from empirical associations):
  - Policies that reduce income inequality and strengthen political accountability and constraints on executive power are associated with longer growth spells.
  - Trade liberalization combined with competitive exchange rates and policies that attract FDI (and limit destabilizing debt accumulation) are associated with sustained growth.
  - Promoting manufacturing and more sophisticated export structures is associated with longer-lasting growth episodes.
  - Maintaining macroeconomic stability (low inflation, stable exchange rate) can materially reduce the hazard of growth spell endings.
  - Managing exposure to external shocks (terms of trade volatility, external interest rate movements) is important for sustaining growth.

*IMF Working Paper _wp0859 — Summary Regressions (excerpt).*

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

### References

### Figure
- Frequency of Upbreaks and Downbreaks, by Region ...............................................................................................8

### Tables
- 1. Growth Breaks by Decade and Region ..........................................................................7
- 2. Frequency and Duration of Growth Spells ..................................................................10
- 3. Average Growth Before, During and After Growth Spells .........................................12
- 4. Duration Regressions: External Shocks.......................................................................17
- 5. Duration Regressions: Institutions ...............................................................................18
- 6. Duration Regressions: Inequality and Franctionalization............................................20
- 7. Duration Regressions: Social and Physical Indicators.................................................21
- 8. Duration Regressions: Globalization ...........................................................................23
- 9. Duration Regressions: Current Account, Competitiveness, and Export Structure ......25
- 10.       Duration       Regressions:       Macroeconomic Volatility.......................................................27

*Source: _wp0859 - References..............................................................................................................*

### 11.       Summary       Regressions.....................................................................................

### 11.       Summary       Regressions

### I. Introduction — motivation and approach
- Central question: how to generate sustained economic growth in low-income and emerging market countries, emphasizing the importance of growth duration rather than average growth.
- Core argument: developing-country output paths exhibit low persistence; studying turning points (accelerations and decelerations) and the duration of growth spells is more informative than cross-country averages.
- Methodological contribution: focus on duration of growth spells using duration analysis techniques (proportional hazard / survival models) and distinguishing initial conditions at the time of an acceleration from changes that occur during a spell.
- Related literature: builds on Pritchett (2000), Rodrik (1999), Hausmann et al. (2005, 2006), Jerzmanowski (2006), and others; differs by identifying two regimes via structural breaks and then applying duration analysis to many covariates simultaneously.

### II. Structural breaks and definition of growth spells
- Break detection method:
  - Variant of Bai-Perron (1998, 2003) with sample-specific critical values (accounting for heteroskedasticity and small sample size) and extended sequential testing.
  - Minimum interstitiary periods h set to either 5 or 8 years to balance power and reliability.
  - Break location chosen by minimizing sum of squared residuals; critical values from Monte Carlo with bootstrapped residuals.
- Data: income per capita growth series for 140 countries from Penn World Tables version 6.2 extended to 2006 with IMF WEO.
- Key counts and patterns:
  - At p = 0.10 and h = 5: total of 290 breaks — 139 upbreaks and 151 downbreaks.
  - Standard Bai-Perron algorithm (same data, p-value, h) identifies only 74 breaks.
  - At p = 0.10 and h = 8: 171 breaks — 76 upbreaks and 95 downbreaks.
  - Upbreaks concentrated in the 1950s–60s (Europe, Latin America) and the 1990s (Africa); downbreaks concentrated in the 1970s (high-income: early 1970s; Latin America: 1978–1983; Africa: 1970s–early 1980s).
- Growth spell definition (statistical + economic):
  - Begins with a statistical upbreak followed by at least g percent average per capita growth.
  - Ends with a statistical downbreak followed by less than g percent average growth (complete) or with the end of the sample (incomplete).
  - Growth cutoff g examined at g = 2, g = 2.5, g = 3; the paper focuses on g = 2.
- Frequency and duration of spells (Table 2 summary):
  - Using h = 5 and p = 0.10: total spells = 103 (complete + incomplete); using p = 0.25: total spells = 160.
  - Using h = 8 and p = 0.10: total spells = 62; p = 0.25: total spells = 91.
  - Regional patterns: mean spell length much shorter in Latin America and Africa versus industrial countries and emerging Asia; for h = 8, 80–100 percent of spells in high income countries and emerging Asia last 10 years or more, but only ~two thirds for Latin America and Africa.
  - Asymmetry: Latin America has many past (short) spells but few ongoing spells (3–5); Africa has many ongoing spells (15–23) initiated mid–late 1990s and thus short on average.
- Growth dynamics before/during/after spells (Table 3 summary):
  - Similar growth levels during spells across regions (except Latin America somewhat lower).
  - After spells: advanced countries and emerging Asia show "soft landings" (growth between -1 and 3 percent); African spells tend to end with deep collapses (average growth between -3 and -6 percent).
  - Before spells: high-income and Asia start from positive or slightly negative per capita growth; other developing regions often begin spells from crises (interstitiary rates prior to spells between -1.3 and -5 percent).

### III. Analyzing duration — empirical strategy and methodology
- Objective: relate expected duration (or hazard that a spell will end) to initial conditions and time-varying covariates; distinguish initial level vs. change within spell.
- Empirical challenges:
  - Model selection is atheoretical and sequential due to data limitations and small number of spells (sample of spells no more than about 75 under stringent significance).
  - Data availability constraints prevent large multivariate specifications; sequential testing with minimal controls preferred.
  - Endogeneity: time-varying covariates treated as predetermined by using lagged realizations and proportional hazard model with time-varying covariates to reduce contemporaneous feedback bias.
- Regression methodology:
  - Proportional hazard (survival) model estimated via maximum likelihood; baseline hazard parameterized (Weibull distribution assumed in practice).
  - Hazard ratios (exponentiated coefficients) reported: value > 1 increases risk; value < 1 is protective.
  - Main control variables included: terms of trade shocks, U.S. interest rate changes, initial income per capita; many robustness checks and alternative samples (h = 5 or 8; p = 0.10 or 0.25).

### IV. Main empirical results (by category)
- External shocks (Table 4):
  - Terms of trade: hazard ratios ~0.97–0.98 — a one percent improvement in terms of trade reduces probability of a downbreak by 2–3 percent (not always precisely estimated).
  - U.S. interest rate changes: large significant effect; a one percentage point increase in U.S. rates estimated to increase probability that a growth spell will end next year by 25–50 percent (samples vary).
- Political and economic institutions (Table 5):
  - "Polity2" (scale -10 to +10): a one point improvement lowers hazard by about 3–9 percent (initial level and within-spell change have similar effects).
  - Collapsing initial and change terms to contemporaneous polity improves precision: reduction in hazard from one point increase about 4–8 percent.
  - Democracy vs. Autocracy components: autocracy score exhibits stronger effect — a one point increase in autocracy raises probability a spell ends by 10–20 percent.
  - Concept indices: executive recruitment, executive constraints, political competition generally protective (hazard ratios < 1 and significant in many subsamples).
  - ICRG "investment profile" (0–12 scale): suggests a protective effect (one point improvement associated with up to 35 percent reduction in hazard) but imprecisely estimated and significant in one subsample.
- Inequality and fractionalization (Table 6):
  - Inequality (Gini): strong association — a one percentage point higher Gini lowers hazard that growth will continue (i.e., increases risk) by between 4 and 14 percent (large effect given cross-sectional SD of Gini > 10 percentage points).
  - Effect primarily from cross-sectional initial Gini; within-spell Gini changes estimated imprecisely and not significant.
  - Ethnic fractionalization (Alesina et al., 2003): sign as expected but not robustly significant; estimates unstable across subsamples.
  - Direct violent conflict measures (UCDP/PRIO): no strong robust associations found; possible interpretation is conflict affects growth primarily after spells end.
- Social and physical indicators (Table 7):
  - Education: no robust significant associations; possible exception — within-spell improvements in primary education may be beneficial (imprecise).
  - Health: infant mortality increases associated with higher hazard (one death per 100 increase raises hazard by about 10 percent in larger samples when Gini excluded); estimates imprecise and sample-dependent.
  - Physical infrastructure: telephone mainlines per 100 people hazard ratios < 1 but significant in one subsample.
- Globalization (Table 8):
  - Trade liberalization (Wacziarg-Welch dummy): very large protective effect — liberalizers show ~70–80 percent reduction in hazard that a growth spell ends (robust across timing: initial level or change within spell).
  - Trade openness (adjusted): small protective effect — a one percent of GDP increase in openness reduces chance a spell ends by 1–2 percent; significant in only one sample.
  - Financial integration (sum of external assets and liabilities): weak effects; aggregate integration changes may marginally increase risk.
  - Disaggregated external finance: debt liabilities accumulation associated with slightly higher risk; FDI liabilities associated with protective effect — 1 percent of GDP increase in FDI liabilities reduces probability of a downbreak by 4–7 percent.
- Current account, competitiveness, export structure (Table 9):
  - Current account surplus: one percentage point of GDP rise in surplus lowers probability a growth spell will end by 5–20 percent (imprecise).
  - Overvaluation: on h = 8 samples, each percentage point of overvaluation increases hazard by 1–2 percent.
  - Domestic savings: hazard ratios below 1 but not significant.
  - Manufacturing exports share:
    - Initial level less important than change: a one percentage point increase in manufacturing exports share during a spell associated with a 2–4 percent reduction in hazard.
  - Export sophistication measures:
    - Hausmann, Hwang and Rodrik sophistication: positive associations; change within spell significant in several samples.
    - "Open Forest" (Hausmann, Rodriguez, Wagner): strong protective effects; one standard deviation increase in log "Open Forest" lowers risk of spell ending by about 40–50 percent (large effect); change within spell highly significant.
- Macroeconomic stability (Table 10):
  - Inflation (100*ln(1+π)): a one point increase raises annual risk of a downbreak by 1–4 percent (sample dependent); moderate inflation is economically meaningful — e.g., inflation rising from 10 percent to 50 percent (≈ 30 point rise in 100*ln(1+π)) could raise annual risk by up to 120 percent relative to baseline.
  - Exchange rate depreciation (parallel rate): one point increase raises risk by 2–6 percent.
  - Results robust to dropping observations with inflation/depreciation > 50 percent (i.e., not driven only by hyperinflation outliers).
- Joint (summary) regressions (Table 11):
  - Joint inclusion weakens some individual effects (e.g., inflation may lose significance), but several variables remain robust predictors:
    - Income distribution (Gini) and autocracy score (or polity2) are among the most robust predictors of shorter duration.
    - "Open Forest" remains significant in a summary specification.
    - Terms of trade changes and U.S. interest rate changes retain importance.
  - Duration dependence (Weibull parameter p) typically positive but modest (p between ~0.95 and ~1.90 across models).
  - Results hold when Asian observations excluded, indicating findings reflect general features of growth booms, not driven solely by East Asia.

### V. Key substantive findings — concise bullets
- Structural breaks:
  - At p = 0.10, h = 5: 290 breaks identified (139 upbreaks, 151 downbreaks).
  - At p = 0.10, h = 8: 171 breaks identified (76 upbreaks, 95 downbreaks).
- Spell frequency and duration:
  - Total spells (g = 2): 103 (p = 0.10, h = 5); 160 (p = 0.25, h = 5); 62 (p = 0.10, h = 8); 91 (p = 0.25, h = 8).
  - High-income and emerging Asia spells more likely to last ≥ 10 years (80–100 percent for h = 8); Latin America and Africa much less so (about two thirds or lower).
- Predictors that prolong growth spells (robust across many specifications):
  - More equal income distribution (lower Gini).
  - Better political institutions / lower autocracy (higher polity2, executive constraints, political competition).
  - Trade liberalization (Wacziarg-Welch dummy) and trade openness (weaker).
  - FDI liabilities (composition of external finance matters: FDI protective; debt liabilities less so).
  - Export structure: rising manufacturing exports share during spells; higher export sophistication and higher "Open Forest".
  - Macroeconomic stability: lower inflation and fewer large exchange rate depreciations; current account surpluses.
- External shocks raise risk of spell termination:
  - Terms of trade declines increase hazard; U.S. interest rate increases have large estimated effects (one percentage point raise increases hazard by 25–50 percent in many samples).

### VI. Interpretations and implications
- Synthesis: duration of growth spells is associated with political economy and structural features — inequality, institutions, openness and export structure, macroeconomic stability, and external shocks all matter.
- Suggested channels:
  - Inequality and social fragmentation may shorten growth spells via social/political conflict and weaker institutions.
  - Export orientation toward manufacturing and product sophistication can build constituencies and generate positive spillovers that sustain growth.
  - Composition of external finance matters: FDI supports sustained growth more than debt accumulation.
  - Macroeconomic volatility (inflation, depreciation) and adverse external shocks increase risk of spell termination.
- Policy-relevant implications (inferred from empirical associations):
  - Policies that reduce income inequality and strengthen political accountability and constraints on executive power are associated with longer growth spells.
  - Trade liberalization combined with competitive exchange rates and policies that attract FDI (and limit destabilizing debt accumulation) are associated with sustained growth.
  - Promoting manufacturing and more sophisticated export structures is associated with longer-lasting growth episodes.
  - Maintaining macroeconomic stability (low inflation, stable exchange rate) can materially reduce the hazard of growth spell endings.
  - Managing exposure to external shocks (terms of trade volatility, external interest rate movements) is important for sustaining growth.

*Italic source attribution: IMF Working Paper _wp0859 — Summary Regressions (excerpt).*

### References

### References

### Institutions and Development
- Acemoglu, Daron, Simon Johnson and James A. Robinson, 2001, “The Colonial Origins of Comparative Development: An Empirical Investigation,” American Economic Review, Vol. 91, No. 5 (December), pp. 1369–1401.
- ———, 2005, “Institutions as the Fundamental Cause of Long-Run Growth,” in Handbook of Economic Growth (P. Aghion and S. Durlauf, eds., North Holland).
- ———, and Yunyong Thaicharoen, 2003, “Institutional Causes, Macroeconomic Symptoms: Volatility, Crises, and Growth,” Journal of Monetary Economics, Vol. 50, pp. 49–123.
- Hausmann, Ricardo, and Bailey Klinger, 2007, “The Structure of the Product Space and the Evolution of Comparative Advantage,” CID Working Paper No. 146, (April) (Cambridge, Mass: Center for International Development, Harvard University).
- Rodrik, Dani, Arvind Subramanian and Francesco Trebbi, 2004, “Institutions Rule: The Primacy of Institutions over Geography and Integration in Economic Development.” Journal of Economic Growth 9.2 (June), pp. 131–65.
- Sokoloff, Kenneth L., and Stanley L. Engerman. 2000. “History Lessons: Institutions, Factor Endowments, and Paths of Development in the New World.” Journal of Economic Perspectives, Vol. 14, No. 3, pp. 217–32.
- Rajan, Raghuram, and Luigi Zingales, 2006, “The Persistence of Underdevelopment: Institutions, Human Capital, or Constituencies?” NBER Working Paper No. 12093 (Cambridge, Massachusetts: NBER).

### Growth Dynamics, Accelerations, and Collapses
- Hausmann, Ricardo, Lant Prichett and Dani Rodrik, 2005 “Growth Accelerations,” Journal of Economic Growth, Vol. 10, pp. 303–29
- Hausmann, Ricardo, F. Rodriguez, and R. Wagner, 2006, “Growth Collapses,” Working Paper No. 136, (April) (Cambridge, Mass: Center for International Development).
- Jones, Benjamin F. and Benjamin A., Olken, 2005, “The Anatomy of Start-Stop Growth,” Working Paper 1152 (Cambridge, Massachusetts: NBER).
- Ben-David, Dan and David H. Papell, 1998, “Slowdown and Meltdowns: Postwar Growth Evidence from 74 Countries,” The Review of Economics and Statistics (MIT Press), Vol. 80, No. 4 (November), pp. 561–571.
- Rodrik, Dani, 1999, “Where Did All the Growth Go? Exernal Shocks, Social Conflict, and Growth Collapses,” Journal of Economic Growth, 4 (December), pp. 385–412.
- Jerzmanowski, Michal, 2005, “Empirics of Hills, Plateaus, Mountains and Plains: A Markov-Switching Approach to Growth,” Journal of Development Economics, (May).
- Pritchett, Lant, 2000, “Understanding Patterns of Economic Growth: Searching for Hills Among Plateaus, Mountains, and Plains,” World Bank Economic Review, pp. 221–50.
- Jones, Benjamin F. and Benjamin A., Olken, 2005, “The Anatomy of Start-Stop Growth,” Working Paper 1152 (Cambridge, Massachusetts: NBER).

### Trade, Openness, and Comparative Advantage
- Dollar, David, 1992, “Outward Oriented Economies Really Do Grow More Rapidly,” Economic Development and Cultural Change, Vol. 40, No. 3, pp. 523–44.
- Alesina, Alberto, Enrico Spolare, and Romain Wacziarg, 2005, “Trade, Growth, and the Size of Countries,” Handbook of Economic Growth (Philippe Aghion and Steven N. Durlauf, eds., North Holland).
- Wacziarg, Romain, and Karen H. Welch, “Trade Liberalization and Growth: New Evidence,” Working Paper No. 10152 (Cambridge, Massachusetts: NBER).
- Sala-i-Martin, Xavier, Gernot Doppelhofer, and Ronald I. Miller, 2004, “Determinants of Long-Term Growth: A Bayesian Averaging of Classical Estimates (BACE) Approach,” American Economic Review, Vol. 94, No. 4, pp. 813–35.
- Hausmann, Ricardo, Jason Hwang, and Dani Rodrik, 2006, “What You Export Matters” (unpublished, Kennedy School of Government).

### Financial Globalization, Capital Flows, and Macroeconomic Policy
- Kose, M. Ayhan, E. Prasad, S. Wei, and K. Rogoff, 2006, “Financial Globalization: A Reappraisal,” IMF Working Paper No. 189 (Washington: International Monetary Fund).
- Dell’Ariccia, Giovanni, J. di Giovanni, A. Faria, A. Kose, P. Mauro, J. D. Ostry, M. Schindler, and M. Terrones, 2008, “Reaping the Benefits of Financial Globalization,” Occasional Paper (Washington: International Monetary Fund).
- Prasad, Eswar, Raghuram Rajan, and Arvind Subramanian, 2007, “Foreign Capital and Economic Growth,” Brookings Papers on Economic Activity 1, pp. 153–209.
- Edison, Hali J., Michael W. Klein, Luca Antonio Ricci, And Torsten Sløk, 2004, “Capital Account Liberalization and Economic Performance: Survey and Synthesis,” IMF Staff Papers, Vol. 51, No. 2, pp. 220–56.
- Aizenman, Joshua, Brian Pinto, and Artur Radziwill, 2004, “Sources for Financing Domestic Capital—Is Foreign Saving a Viable Option for Developing Countries?,” Working Paper No. 10624 (Cambridge, Massachusetts: NBER).
- Henry, Peter, 2007, Capital Account Liberalization: “Theory, Evidency, and Speculation,” Journal of Economic Literature, Vol. XLV, pp. 887–935, December.
- Pattillo, Catherine, Sanjeev Gupta, and Kevin Carey, 2005, “Sustaining Growth Accelerations and Pro-Poor Growth in Africa,” IMF Working Paper No. 195 (Washington: International Monetary Fund).
- Rajan, Raghuram, and Arvind Subramanian, 2005, “What Undermines Aid's Impact on Growth? IMF Working Paper No. No. 126 (Washington: International Monetary Fund).

### Institutions, Ethnicity, Conflict, and Social Cohesion
- Alesina, Alberto, A. Devleeschauwer, W. Easterly, S. Kurlat and R. Wacziarg, 2003, “Fractionalization,” Journal of Economic Growth, Vol. 8, pp. 155–94.
- Fearon, James D, 2003, “Ethnic and Cultural Diversity by Country,” Journal of Economic Growth, Vol. 8, pp. 195–222.
- Gleditsch, Nils Petter; Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg & Håvard Strand, 2002, “Armed Conflict 1946-2001: A New Dataset.” Journal of Peace Research, Vol. 39, No. 5, pp. 615–37.
- Easterly, William, Jozef Ritzen, and Michael Woolcock, 2006, “Social Cohesion, Institutions, and Growth,” Economics and Politics, Vol. 18, No. 2 (July) pp. 103–20.
- Sachs, Jeffrey, 1989, “Social Conflict and Populist Policies in Latin America,” Working Paper No. 2897 (Cambridge, Massachusetts: NBER).
- Rodrik, Dani, 2007, “The Real Exchange Rate and Economic Growth: Theory and Evidence” (unpublished, Kennedy School of Government).

### Empirical Methods, Structural Breaks, and Measurement
- Bai, Jushan, and Pierre Perron, 1998, “Estimating and Testing Linear Models with Multiple Structural Changes,” Econometrica, Vol. 66, No 1 (January), pp. 47–78.
- ———, 2003, “Computation and Analysis of Multiple Structural Change Models,” Journal of Applied Econometrics, Vol. 18, pp. 1–22.
- Antoshin, Sergei, Andrew Berg, and Marcos Souto, 2008, “Testing for Structural Breaks in Small Samples,” unpublished manuscript.
- Fernández, Carmen, E. Ley, and M. Steel, 2001, “Model Uncertainty in Cross-Country Growth Regressions,” Journal of Applied Econometrics, Vol. 16, No. 5, pp. 563–76.
- Bai, Jushan, and Pierre Perron, 1998, “Estimating and Testing Linear Models with Multiple Structural Changes,” Econometrica, Vol. 66, No 1 (January), pp. 47–78.
- Ghosh, Atish and Steven Phillips, 1998, “Warning: Inflation May Be Harmful to Your Growth,” IMF Staff Papers (Washington: International Monetary Fund).
- Sarel, Michael, 1996, “Nonlinear Effects of Inflation on Economic Growth,” Staff Papers, International Monetary Fund, Vol. 43 (March), pp. 199–215.

### Policy, Historical Perspectives, and Development Strategies
- Easterly, William, 2005, “National Policies and Economic Growth: A Reappraisal,” in Handbook of Economic Growth, Vol. 1A, ed. by Philippe Aghion and Steven N. Durlauf (Amsterdam and London: Elsevier).
- Easterly, William, Michael Kremer, Lant Pritchett and Lawerence H. Summers, 1993, “Good Policy or Good Luck? Country Growth Performance and Temporary Shocks,” Journal of Monetary Economics 32, pp. 459–83.
- Sachs, Jeffrey, and Andrew Warner, 1995, “Economic Reform and the Process of Global Integration,” Brookings Papers on Economic Activity 1, pp. 1–118.
- Becker, Torbjörn, and Paolo Mauro, 2006, “Output Drops and the Shocks That Matter,” IMF Working Paper No. 172 (Washington: International Monetary Fund).
- Berg, Andrew and Jeffrey D. Sachs, 1988, “The Debt Crisis: Structural Explanations of Country Performance”, Journal of Development Economics, No. 29(3), pp. 271–306.
- Johnson, Simon, Jonathan D. Ostry, and Arvind Subramanian, 2006, “Levers for Growth: Policy Lessons from Earlier Bouts of Growth in Developing Countries,” Finance and Development (March), pp. 28–32.
- Johnson, Simon, Jonathan D. Ostry, and Arvind Subramanian, 2007, “The Prospects for Sustained Growth in Africa: Benchmarking the Constraints,” Working Paper 13120 (Cambridge, Massachusetts: NBER).
- Pattillo, Catherine, Sanjeev Gupta, and Kevin Carey, 2005, “Sustaining Growth Accelerations and Pro-Poor Growth in Africa,” IMF Working Paper No. 195 (Washington: International Monetary Fund).
- Reddy, Sanjay, and Camelia Minoui, 2007, “Real Income Stagnation of Countries, 1960–2001,” forthcoming, Journal of Development Studies.

*Source: _wp0859 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0859.pdf_