## _sdn1108 - Executive Summary

## Source details

**Canonical URL:** [_sdn1108 - Executive Summary](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/sdn/2011/_sdn1108.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/sdn/2011/_sdn1108.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/sdn/2011/_sdn1108.pdf.json)

---

### Major themes and scope
- Focus: duration of growth spells (interval from growth upbreak to downbreak) and links between duration and policies, country characteristics, and income distribution.
- Structural emphasis: descriptive analysis of growth patterns (“hills, valleys, and plateaus”), statistical duration (hazard) modeling, empirical tables and figures, and policy implications.

### Central empirical findings on growth spells
- Definition and thresholds:
  - Growth spell minimum practical length: eight years.
  - Operational criteria: begins with a statistical upbreak followed by a period of at least two percent average real per capita growth; ends with a statistical downbreak followed by a period of less than 2 percent average growth or with the end of the sample.
  - Gini coefficient range: 0–100.
- Stylized observations:
  - Per capita incomes in developing countries show upbreaks and downbreaks rather than smooth multi-decade growth.
  - Initiating growth is easier than sustaining it.
- Tabulated summary (selected entries from Table 1 and Table 2):
  - Total Upbreaks: 78; Average break size: 6.9
  - Total Downbreaks: 96; Average break size: -6.3
  - Upbreaks by region (Total Upbreaks; Average break size):
    - Advanced countries: 11; 5.0
    - Emerging Asia: 19; 5.6
    - Latin America: 11; 4.0
    - Sub-Saharan Africa: 22; 10.3
    - Other developing: 15; 7.4
  - Downbreaks by region (Total Downbreaks; Average break size):
    - Advanced countries: 21; -5.3
    - Emerging Asia: 15; -6.0
    - Latin America: 13; -4.6
    - Sub-Saharan Africa: 26; -8.0
    - Other developing: 21; -6.9
  - Mean duration, complete spells (years) and percent lasting at least 10 years:
    - Advanced countries: Mean duration 13.0; % spells lasting at least 10 years = 100.0; Average growth before 3 years 3.3; during 6.0; after 1.2
    - Emerging Asia: Mean duration 18.0; % spells lasting at least 10 years = 33.3; Average growth before 3 years -0.7; during 9.1; after 1.4
    - Latin America: Mean duration 14.4; % spells lasting at least 10 years = 60.0; Average growth before 3 years 1.1; during 4.8; after 0.2
    - Sub-Saharan Africa: Mean duration 8.3; % spells lasting at least 10 years = 0.0; Average growth before 3 years -2.7; during 9.9; after -4.0
  - Mean duration, total (including incomplete spells), selected entries:
    - Advanced countries: Mean duration 24.4; % spells lasting at least 10 years = 100.0; Average growth during 5.7
    - Emerging Asia: Mean duration 24.2; % spells lasting at least 10 years = 87.5; Average growth during 5.8
    - Latin America: Mean duration 15.7; % spells lasting at least 10 years = 71.4; Average growth during 4.4
    - Sub-Saharan Africa: Mean duration 13.6; % spells lasting at least 10 years = 66.7; Average growth during 6.3

### Income distribution and growth duration — core quantitative relationships
- Central empirical association: longer growth spells are robustly associated with more equality in the income distribution.
- Selected quantified associations and thresholds:
  - Closing half the inequality gap between Latin America and emerging Asia: more than double the expected duration (central estimates).
  - A 10-percentile decrease in inequality: increases expected spell length by 50 percent.
  - Trade liberalization (closed -> open, Wacziarg and Welch dichotomous variable): associated with a 45 percent longer spell.
  - Autocracy reduction from 1 to 0 (Polity IV scale): associated with a 25 percent longer spell.
  - Decrease in real exchange rate overvaluation by 10 percentage points: associated with an 8 percent increase in expected spell length.
  - Increase in FDI liabilities from 8 to 12 percent of GDP: associated with a 15 percent longer spell.
  - Decrease in external debt ratio from 44 to 39 percent of GDP: associated with about a 2 percent increase in spell duration.
- Robustness: inequality remains significant when controlling for external shocks, initial income, institutional quality, openness to trade, and macroeconomic stability.

### Mechanisms and other correlates identified
- Channels through which inequality may shorten growth spells:
  - Credit market imperfections limiting human-capital investment by the poor (secondary education achievement correlates negatively with inequality, controlling for per capita income).
  - Political economy effects: redistribution pressures, elite resistance, vote buying, corruption that deter investment.
  - Political instability reducing investment incentives and policy responsiveness.
- Additional bivariate and multivariate correlates of longer spells:
  - Better political institutions (constraints on the executive, political accountability).
  - Within-spell increases in education, health, and physical infrastructure (notably primary education improvements).
  - Financial development: increases in the ratio of bank deposits to GDP during the spell.
  - International integration: FDI inflows beneficial; growth of external debt harmful.
  - Trade liberalization and export structure: high manufacturing export shares and more “sophisticated” export structures associated with longer spells.
  - Macroeconomic stability: lower currency depreciation rates and inflation increase expected spell length.
  - External shocks: declines in terms of trade and increases in U.S. interest rates shorten spells.
  - Ethnic fractionalization: some association with shorter spells, but effects vary and are often not statistically significant.

### Empirical case evidence (Box 1) — selected country outcomes and hazard contributions
- Table of six spell-ends (growth in real per capita GDP during spells; next decade growth; hazard ratio; main contributing factors with shares):
  - Cameroon: Spell dates 1978–85; Growth during spell 6.6; Next decade -5.6; Hazard ratio 109. Main contributing factors: Inequality 0.49; Low FDI Inflow 0.33; Increased external debt -0.05; More autocracy 0.33; Over-valuation 0.09.
  - Colombia: Spell dates 1967–78; Growth during spell 3.4; Next decade 1.2; Hazard ratio 66. Main contributing factors: Inequality 0.73; Low FDI Inflow 0.46; Increased external debt -0.06; More autocracy -0.23; Over-valuation -0.03.
  - Guatemala: Spell dates 1958–79; Growth during spell 2.4; Next decade -1.3; Hazard ratio 56. Main contributing factors: Inequality 0.39; Low FDI Inflow 0.38; Increased external debt -0.07; More autocracy 0.13; Over-valuation -0.02.
  - Ecuador: Spell dates 1971–78; Growth during spell 7.2; Next decade -1.0; Hazard ratio 47. Main contributing factors: Inequality 1.05; Low FDI Inflow 0.34; Increased external debt -0.05; More autocracy 0.17; Over-valuation -0.13.
  - Panama: Spell dates 1959–80; Growth during spell 4.7; Next decade 0.0; Hazard ratio 42. Main contributing factors: Inequality 0.44; Low FDI Inflow -0.61; Increased external debt 0.62; More autocracy 0.28; Over-valuation 0.06.
  - Nigeria: Spell dates 1968–76; Growth during spell 5.9; Next decade -4.0; Hazard ratio 29. Main contributing factors: Inequality 0.27; Low FDI Inflow 0.41; Increased external debt -0.08; More autocracy 0.39; Over-valuation 0.47.
- Note: Hazard ratio definition—ratio of predicted probability that the spell would end during the five years prior to its actual end to the predicted probability of a spell ending for the average observation. Contributions shown are based on Model 1 of Table 12 of Berg, Ostry, and Zettelmeyer (2008), rescaled to sum to one.

### Policy implications and recommended directions
- Overarching message:
  - Analyses of growth and income distribution should be integrated; distributional patterns materially affect growth sustainability.
  - Policies must balance short-run trade-offs: poorly designed redistribution can harm growth, but some policies can be “win-win.”
- Policy options and examples:
  - Better-targeted subsidies to goods consumed mainly by the poor to free resources for public infrastructure.
  - Improvements in economic opportunities for the poor and equality of opportunity through effective investments in health and education.
  - Active labor market policies that promote employment and foster job-richer recoveries.
  - Well-designed progressive taxation and measures that enhance bargaining power for labor, with attention to avoiding dual labor markets.
  - Caution: trade-offs vary; country-specific analysis is necessary.
- Implications for international institutions (e.g., the IMF):
  - Distributional analysis matters for assessing risks to growth; downbreaks can precede macro crises by several years.
  - Standard policy advice may need adjustment when rising inequality amplifies macro and financial risks.

_Italic: Source: _sdn1108 - Executive Summary._

### Executive Summary ......................................................................................................

### Executive Summary

### Major themes and sections
- I. Introduction (page 4)
- II. The Hills and Valleys of Growth (page 5)
- III. Income Distribution and Growth Sustainability (page 8)
- IV. Some Tentative Policy Implications (page 16)

### Empirical components (tables and figures referenced)
- Tables:
  - Table 1. Growth Breaks by Region and Decade (page 7)
  - Table 2. Characteristics of Growth Spells (page 8)
  - Table 3. The Ends of Six Spells (page 11)
- Figures:
  - Figure 1a. The Hills of Growth (page 5)
  - Figure 1b. The Hills, Valleys, and Plateaus of Growth (page 6)
  - Figure 2. Duration of Growth Spells and Inequality (page 9)
  - Figure 3. Effect of Increase of Different Factors on Growth Spell Duration (page 12)
- Box:
  - Box 1. Is It Really Income Distribution? A Closer Look at Country Cases (page 14)

### Structural focus for AI ingestion and search
- Emphasizes the structure of the chapter: descriptive analysis of growth patterns (hills, valleys, plateaus), the relationship between income distribution and growth sustainability, and policy implications.
- Contains empirical tables and figures linking growth spells, their characteristics, duration, and factors affecting duration, plus a case-focused box on income distribution.

*Source: _sdn1108 - Executive Summary (page references as listed in the source content).*

### EXECUTIVE SUMMARY

### _sdn1108 - EXECUTIVE SUMMARY

### Executive summary
- The relationship between income inequality and economic growth is complex: some inequality supports market incentives and investment, but inequality can also amplify crisis risk or impede the poor from investing in education.
- The empirical literature has mixed findings: some studies find higher average long-run growth with more initial equality; others find that increases in equality today tend to lower growth in the near term.
- A key missing feature in earlier work is the lack of persistence in per capita income growth in developing countries: periods of rapid growth are punctuated by collapses and stagnation (the “hills, valleys, and plateaus” of growth).
- This note focuses on the duration of growth spells—defined as the interval starting with a growth upbreak and ending with a downbreak—and the links between duration and policies and country characteristics, including income distribution.
- Many poor countries can initiate high-rate growth for a few years; what is rarer is the ability to sustain growth. The question becomes: what determines the length of growth spells, and what is the role of income inequality in duration?
- Central empirical finding: longer growth spells are robustly associated with more equality in the income distribution. Example: closing half the inequality gap between Latin America and emerging Asia would, according to the central estimates, more than double the expected duration of a growth spell.
- Inequality remains significant even when controlling for external shocks, initial income, institutional quality, openness to trade, and macroeconomic stability.
- Policy implication: analyses of growth and income distribution are difficult to separate. Poorly designed redistribution could harm growth; however, some “win-win” policies include better-targeted subsidies, improvements in economic opportunities for the poor, and active labor market policies that promote employment.
- While evidence is not decisive on short-run trade-offs between growth and distribution, attention to inequality may bring significant longer-run benefits for growth; over longer horizons, reduced inequality and sustained growth may be two sides of the same coin.

### Introduction (role of sustained growth and recent context)
- Sustained growth is central to poverty reduction; rapid growth in countries such as China and India has led to unprecedented poverty reduction.
- The recent global crisis revived interest in links among income inequality, crises, and growth sustainability. Examples in the literature:
  - Piketty and Saez (2003): sharp rise in U.S. income inequality over recent decades.
  - Rajan (2010): political/economic pressures and credit dynamics linked to inequality and crisis.
  - Kumhof and Rancière (2010): mechanisms linking income distribution and financial excess across historical crises.
- Inequality can be productive for incentives (Chaudhuri and Ravallion, 2006) but excessive inequality may produce financial, political, and human-capital frictions that harm growth.
- Research question: Can growth be sustained with highly uneven income distribution? Does less inequality lengthen growth spells?

### The hills, valleys, and plateaus of growth
- Stylized observation: per capita incomes in developing countries rarely grow smoothly for decades; instead they show upbreaks and downbreaks.
- Growth spells are defined operationally:
  - Minimum length: eight years (practical operational choice).
  - A growth spell begins with a statistical upbreak followed by a period of at least two percent average real per capita growth and ends with a statistical downbreak followed by a period of less than 2 percent average growth or the end of the sample.
- Key empirical patterns:
  - Initiating growth is easier than sustaining it: almost all growth spells in advanced countries and emerging Asia last at least 10 years or more, but only about two-thirds of Latin American and African spells do.
  - Table 1 (growth breaks by region and decade) highlights:
    - Total Upbreaks: 78; Average break size: 6.9 (percentage point change)
    - Advanced countries: Total Upbreaks 11; Average break size 5.0
    - Emerging Asia: Total Upbreaks 19; Average break size 5.6
    - Latin America: Total Upbreaks 11; Average break size 4.0
    - Sub-Saharan Africa: Total Upbreaks 22; Average break size 10.3
    - Other developing: Total Upbreaks 15; Average break size 7.4
    - Total Downbreaks: 96; Average break size -6.3
    - Advanced countries (Downbreaks): 21; Average break size -5.3
    - Emerging Asia (Downbreaks): 15; Average break size -6.0
    - Latin America (Downbreaks): 13; Average break size -4.6
    - Sub-Saharan Africa (Downbreaks): 26; Average break size -8.0
    - Other developing (Downbreaks): 21; Average break size -6.9
  - Table 2 (characteristics of growth spells) highlights:
    - Complete spells, mean duration (years):
      - Advanced countries: Mean duration 13.0; % spells lasting at least 10 years = 100.0; Average growth before 3 years 3.3; during 6.0; after 1.2
      - Emerging Asia: Mean duration 18.0; % spells lasting at least 10 years = 33.3; Average growth before 3 years -0.7; during 9.1; after 1.4
      - Latin America: Mean duration 14.4; % spells lasting at least 10 years = 60.0; Average growth before 3 years 1.1; during 4.8; after 0.2
      - Sub-Saharan Africa: Mean duration 8.3; % spells lasting at least 10 years = 0.0; Average growth before 3 years -2.7; during 9.9; after -4.0
    - Total (including incomplete spells), mean duration (years):
      - Advanced countries: Mean duration 24.4; % spells lasting at least 10 years = 100.0; Average growth during 5.7
      - Emerging Asia: Mean duration 24.2; % spells lasting at least 10 years = 87.5; Average growth during 5.8
      - Latin America: Mean duration 15.7; % spells lasting at least 10 years = 71.4; Average growth during 4.4
      - Sub-Saharan Africa: Mean duration 13.6; % spells lasting at least 10 years = 66.7; Average growth during 6.3
- Regional contrast: African spells are rapid while ongoing, but they tend to end in deeper collapses (soft landings more common in advanced countries and Asia).

### Income distribution and growth sustainability — mechanisms and hazards
- Figure 2 correlation: more inequality (measured by the Gini coefficient, range 0–100) is associated with shorter growth spells.
- Potential channels through which inequality affects growth sustainability:
  - Credit market imperfections: inequality limits poor households' ability to finance education; the data show a negative correlation between some human capital indicators (notably secondary education achievement) and income distribution, controlling for per capita income.
  - Political economy: redistribution pressures in unequal societies may create investment disincentives; elites’ resistance (vote buying, corruption) can be distortionary.
  - Political instability: inequality may increase risk of political instability, reducing investment incentives and hampering policy responses to shocks.
- Analytical approach: duration (hazard) modeling analogous to medical survival analysis; distinguishes conditions at onset of a spell and changes within spells (policy-relevant).
- Bivariate findings (each tested while controlling for initial income and spell length as a risk factor) indicate longer growth spells correlated with:
  - Better political institutions (constraints on the executive, political accountability).
  - Increases in education, health, and physical infrastructure—notably within-spell improvements in primary education; child mortality both initial level and increases reduce expected duration.
  - Financial development: increases in the ratio of bank deposits to GDP during the spell are protective.
  - Trade liberalization: significant large effect consistent with increased market size, competition, and know-how transmission.
  - International financial integration: FDI helps duration; growth of external debt hurts.
  - Competitiveness and export structure: avoidance of exchange rate overvaluation, high shares of manufacturing exports, and more “sophisticated” export structures correlate with longer spells.
  - Macroeconomic volatility: increasing currency depreciation rates and inflation reduce expected spell length.
  - External shocks: reductions in terms of trade and increases in U.S. interest rates are associated with shorter spells.
  - Inequality: low income inequality is a large and statistically significant correlate of longer growth spells.
- Multivariate (joint) analysis results (preferred specification):
  - Method: compute expected duration at sample median (50th percentile), then recalculate expected duration when each variable improves by 10 percentiles, holding others at median.
  - Main multivariate results:
    - Political institutions (autocracy measured by Polity IV): a reduction in autocracy from a rating of 1 (sample median) to 0 on the 10-point scale is associated with a 25 percent longer spell.
    - Trade liberalization (Wacziarg and Welch dichotomous variable): associated with a 45 percent longer spell when comparing open vs closed regimes.
    - Exchange rate overvaluation: a decrease in overvaluation by 10 percentage points of the real exchange rate is associated with an 8 percent increase in expected spell length.
    - Financial globalization (nature matters):
      - FDI: an increase from 8 to 12 percent of GDP in FDI liabilities is associated with an expected spell duration that is 15 percent longer.
      - External debt: a decrease from 44 to 39 percent in the ratio of external debt to GDP suggests an increase in spell duration of about 2 percent.
  - Despite inclusion of many variables, income distribution remains one of the most robust and important factors associated with growth duration:
    - A 10-percentile decrease in inequality (a level of improvement several countries have experienced during spells) increases the expected length of a growth spell by 50 percent.
    - Inequality retains similar statistical and economic significance across samples and specifications, suggesting it is not merely proxying for other factors.
  - Ethnic fractionalization shows some association with shorter growth spells, but effects vary across samples and are often not statistically significant; evidence is firmer for income inequality.

### Key quantitative thresholds and definitions (preserved)
- Growth spell minimum practical length: eight years.
- Growth spell operational criteria (footnote definition):
  - Begins with a statistical upbreak followed by a period of at least two percent average real per capita growth.
  - Ends with a statistical downbreak followed by a period of less than 2 percent average growth or with the end of the sample.
- Gini coefficient: varies from 0 (all households have the same income) to 100 (all income received by one household).
- Selected quantified associations:
  - Closing half the inequality gap between Latin America and emerging Asia: more than double expected duration (central estimates).
  - 10-percentile decrease in inequality: increases expected spell length by 50 percent.
  - Trade liberalization (closed -> open): associated with a 45 percent longer spell.
  - Autocracy reduction from 1 to 0 (Polity IV scale): associated with a 25 percent longer spell.
  - Decrease in real exchange rate overvaluation by 10 percentage points: associated with an 8 percent increase in expected spell length.
  - Increase in FDI liabilities from 8 to 12 percent of GDP: associated with a 15 percent longer spell.
  - Decrease in external debt ratio from 44 to 39 percent of GDP: associated with about a 2 percent increase in spell duration.

### Policy implications and tentative conclusions
- Income distribution should be considered alongside institutions, trade openness, macro stability, human capital, and financial factors when analyzing growth sustainability.
- Policy design must weigh short-run trade-offs: policies that reduce inequality but grossly distort incentives could undermine growth and hurt the poor.
- Potential “win-win” policy options include:
  - Better-targeted subsidies.
  - Improvements in economic opportunities for the poor.
  - Active labor market policies that promote employment.
- Evidence here tilts toward the view that attention to inequality can yield significant longer-run benefits for sustained growth; over long horizons, reduced inequality and sustained growth may reinforce one another.

_ EXECUTIVE SUMMARY _

### Box 1. Is It Really Income Distribution? A Closer Look at Country Cases

### Box 1. Is It Really Income Distribution? A Closer Look at Country Cases

### Empirical case evidence from high-hazard spells
- For the 11 complete spells in the main sample, the predicted risk that the spell will end during the last five years of the spell is several multiples higher than for an average country in the sample.
- Table: The ends of six spells (periods refer to growth spells; growth figures are growth in real per capita GDP)
  - Cameroon: Spell dates 1978–85; Growth in real per capita GDP during spell 6.6; Next decade -5.6; Hazard ratio 109.
    - Main contributing factors (share of total hazard): Inequality 0.49; Low FDI Inflow 0.33; Increased external debt -0.05; More autocracy 0.33; Over-valuation 0.09.
  - Colombia: Spell dates 1967–78; Growth in real per capita GDP during spell 3.4; Next decade 1.2; Hazard ratio 66.
    - Main contributing factors: Inequality 0.73; Low FDI Inflow 0.46; Increased external debt -0.06; More autocracy -0.23; Over-valuation -0.03.
  - Guatemala: Spell dates 1958–79; Growth in real per capita GDP during spell 2.4; Next decade -1.3; Hazard ratio 56.
    - Main contributing factors: Inequality 0.39; Low FDI Inflow 0.38; Increased external debt -0.07; More autocracy 0.13; Over-valuation -0.02.
  - Ecuador: Spell dates 1971–78; Growth in real per capita GDP during spell 7.2; Next decade -1.0; Hazard ratio 47.
    - Main contributing factors: Inequality 1.05; Low FDI Inflow 0.34; Increased external debt -0.05; More autocracy 0.17; Over-valuation -0.13.
  - Panama: Spell dates 1959–80; Growth in real per capita GDP during spell 4.7; Next decade 0.0; Hazard ratio 42.
    - Main contributing factors: Inequality 0.44; Low FDI Inflow -0.61; Increased external debt 0.62; More autocracy 0.28; Over-valuation 0.06.
  - Nigeria: Spell dates 1968–76; Growth in real per capita GDP during spell 5.9; Next decade -4.0; Hazard ratio 29.
    - Main contributing factors: Inequality 0.27; Low FDI Inflow 0.41; Increased external debt -0.08; More autocracy 0.39; Over-valuation 0.47.
- Note on hazard ratio: The hazard ratio is the ratio of the predicted probability that the spell would end during the five years prior to its actual end to the predicted probability of a spell ending for the average observation in the entire sample. A hazard ratio of one implies no unusual risk. Contributions shown are based on Model 1 of Table 12 of Berg, Ostry, and Zettelmeyer (2008), rescaled to sum to one; only main factors for these observations are shown.

### Country narratives linking inequality, shocks, and institutional factors
- Colombia (spell end in 1978)
  - Spark: crackdown on drug cartels leading to a long civil conflict.
  - High Gini (53 vs. sample average of 38) accounts for most of the higher predicted risk; baseline model predicts risk 66 times higher than average.
  - Cárdenas (2007) interpretation: a fortuitous event interacted with high levels of inequality and poverty and weak state presence to change growth trajectory.
- Guatemala
  - Civil war from 1960 until 1996; war peaked in 1979, coincident with end of growth spell.
  - Multivariate model predicts risk about 55 times higher than average during 1974–79; higher-than-average income inequality and low FDI were main drivers.
  - Political repression and reactions to economic conditionality (subsidy cuts, transport price rises) exacerbated unrest and policy reversals.
- Cameroon, Nigeria, Ecuador
  - External shocks (notably oil booms and busts) interacted with high inequality and weak political coalitions.
  - Oil wealth financed large public-sector and wage increases that were hard to reverse when oil prices fell, producing unpopular spending cuts that preserved elite privileges.
  - Nigeria: volatile politics, social incohesion, and external shocks; populist rent distribution collapsed when oil revenues dwindled.
  - Model hazard ratios very high: Cameroon >100, Nigeria 29; regression attributes important roles to high inequality, autocracy, and low FDI.
- Panama
  - Model attributes high risk mainly to rising external debt and inequality.
  - Military dictatorship sustained transfers to government workers via external borrowing; global crisis of early 1980s hit Panama hard, consistent with borrowing to bridge societal conflict.

### Interpretation of channels and timing
- Variety and complexity: different channels matter in different cases (e.g., crime in Colombia, oil shocks in Cameroon/Ecuador/Nigeria, debt in Panama).
- Timing nuances:
  - In Panama, debt grew prior to crisis and shows up as a predictor in hazard regressions.
  - In some oil exporters, debt grew mainly after the end of the spell as an initial unsustainable response to negative commodity shocks, helping convert shocks into sustained downturns.
- Ethnic fractionalization plays a role in some cases.
- Regressions suggest inequality is an underlying feature that increases the likelihood that multiple adverse factors come together to end a growth spell.

### Key policy implications at the country level
- Main results summarized:
  - (i) Increasing the length of growth spells, rather than just initiating growth, is critical to achieving long-term income gains.
  - (ii) Countries with more equal income distributions tend to have significantly longer growth spells.
- Cautions:
  - Income distribution is only part of the story; many other variables are involved and causality is difficult to disentangle.
  - Some policies may increase growth while affecting distribution; trade-offs require careful evaluation.
- Examples of policy approaches that can improve distribution without undermining incentives:
  - Better targeting of subsidies to goods consumed mainly by the poor can free capacity for public infrastructure while protecting the poor (Coady et al., 2010).
  - Active labor market policies to foster job-richer recoveries, making recoveries more sustainable (ILO, 2011); rising unemployment associates with deteriorations in income distribution (Heathcote, Perri, and Violante, 2010).
  - Equality of opportunity via effective investments in health and education (human capital) can promote durable growth and equity while avoiding short-run disincentive effects (Gupta et al., 1999; World Bank, 2005).
  - Propoor policies can markedly reduce income inequality while supporting growth (example cited: Brazil after 1994 reforms and social assistance, Ravallion, 2009).
  - Well-designed progressive taxation and adequate bargaining power for labor can promote equity, with attention to avoiding dual labor markets.

### Implications for international institutions (e.g., the IMF)
- Distributional analysis matters for understanding risks to growth at the country level; crises and growth downbreaks are not tightly correlated and downbreaks can precede macro crises by several years.
- Some developing-country debt crises of the 1980s and 1990s may have stemmed from borrowing premised on continued high growth.
- Based on current IMF World Economic Outlook forecasts at the time of writing, there were no downbreaks in 2008-2009, suggesting the global financial crisis will not end ongoing growth spells—however, identification of a spell end is only confident several years after it occurs.
- The efficacy of standard policy recommendations may depend on distributional forces; analysis of macroeconomic and financial sector linkages should account for rising inequality to yield useful policy advice.

_Italic: Source: Box 1. Is It Really Income Distribution? A Closer Look at Country Cases (excerpt)._

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/sdn/2011/_sdn1108.pdf_