2014年-世界发展银行全球_Do_Poverty_Traps_Exist__27页_841kb

Do Poverty Traps Exist? Summary

Core Content

This paper by Aart Kraay and David McKenzie from the World Bank examines the existence and prevalence of poverty traps, which are self-reinforcing mechanisms that keep individuals or countries in poverty. The authors analyze both macro and micro-level theories and empirical evidence to determine whether poverty traps are a significant factor in persistent poverty.

Main Views on Poverty Persistence

There are three competing views on why poverty persists:

1. The American Dream Effect: Poverty can be overcome through hard work and thrift. This implies that with effort, individuals and countries can eventually escape poverty.
2. Poor Fundamentals: Poverty results from underlying issues such as weak institutions or low individual skills. Changing these fundamentals is necessary for long-term growth, but it may be slow and difficult.
3. Poverty Traps: Poverty begets poverty through self-reinforcing mechanisms. This view suggests that once in poverty, individuals or countries may be unable to escape without external intervention.

The third view is considered the most compelling for policy motivation, as it implies that poverty may not be inevitable and that targeted interventions could have lasting effects.

Empirical Evidence on Poverty Traps

Country-Level Data

• The paper highlights that per capita incomes in Burundi, Haiti, and Nicaragua have remained largely unchanged since 1960, suggesting persistent poverty.
• However, when analyzing 110 countries over the period 1960–2010, it is found that most countries experienced positive growth, and even the poorest countries showed average growth rates similar to the global average.
• The median country in the poorest quintile in 1960 would have reached the income level of the richest country in the second quintile in just 45 years at a 2.2% growth rate, indicating that poverty traps are not common in the data.

Individual and Household-Level Data

• Several studies have examined whether there is an S-shaped relationship in income or asset dynamics, which is a hallmark of poverty traps.
• While some studies (e.g., Barrett et al. 2006, Adato et al. 2006) found evidence of multiple equilibria in certain remote areas, others (e.g., Jalan and Ravallion 2004, Lokshin and Ravallion 2004, Nashold 2007) did not find such evidence.
• The lack of long-term panel data in most developing countries and measurement errors limit the persuasiveness of this evidence.

Key Mechanisms of Poverty Traps

1. Saving-Based Poverty Traps

• The idea is that saving rates are an S-shaped function of income, leading to a stable low-level equilibrium.
• However, empirical data shows that saving rates increase sharply at low levels of development, making a poverty trap unlikely.
• The Solow model with an exogenous saving function does not support the existence of a stable poverty trap in most cases.

2. Big-Push Models

• These models suggest that economies with increasing returns in the modern sector (e.g., manufacturing) can escape poverty through a large initial investment.
• While some studies (e.g., Graham and Temple 2006) found evidence of multiple equilibria, others (e.g., Caucutt and Kumar 2008) suggest that the threshold for escape is low, and subsidies are minimal.
• This raises the question of why large-scale aid has not been effective in breaking poverty traps.

3. Nutritional Poverty Traps

• These traps occur when the link between food intake and work capacity is non-linear, and individuals are too malnourished to work effectively.
• Empirical evidence suggests that calories are too cheap for most people to be in a nutritional trap, except in specific circumstances such as market breakdown or poor nutrient absorption.
• The non-linear relationship between nutrition and productivity is not strong enough to create a poverty trap in most cases.

4. Lumpy Investments and Borrowing Constraints

• These models suggest that individuals may be trapped due to the non-convex nature of production technologies and limited access to credit.
• In remote areas, such as rural Ethiopia and Kenya, pastoralists with small herds are found to be trapped in a low-level equilibrium, unable to invest in larger, more productive enterprises.
• These individuals are excluded from informal credit markets, reinforcing the poverty trap.

Conclusion

• The paper concludes that poverty traps are rare and largely limited to remote or disadvantaged areas.
• It suggests that policy interventions such as microfinance or massive aid increases may not be as effective as previously thought.
• Instead, promoting migration and targeting specific mechanisms such as non-convex production technologies and borrowing constraints may be more effective in addressing poverty.

Policy Implications

• Migration is highlighted as a promising area for policy action.
• Targeted interventions that address specific mechanisms (e.g., lumpy investments, nutritional deficiencies) may be more effective than broad-based poverty trap models.
• The concept of poverty traps is still useful for understanding certain forms of poverty, but the empirical evidence for widespread poverty traps is limited.

References

• Adato et al. (2006): Find evidence of multiple equilibria in asset dynamics in South Africa.
• Barrett et al. (2006): Find evidence of poverty traps in rural Kenya and Madagascar.
• Caucutt and Kumar (2008): Argue that the threshold for escape is low, and subsidies are minimal.
• Graham and Temple (2006): Find that many countries are in a low-level equilibrium.
• Jalan and Ravallion (2004): Find no evidence of poverty traps in China.
• Lokshin and Ravallion (2004): Find no evidence of poverty traps in Hungary and Russia.
• Nashold (2007): Finds no evidence of poverty traps in rural Pakistan and Ethiopia.
• Subramanian and Deaton (1996): Show that calories are too cheap for most poor individuals to be in a nutritional trap.
• Banerjee and Duflo (2011): Show that 2,400 calories can be obtained for 21 cents in the Philippines.
• Lybbert et al. (2004), Santos and Barrett (2011), Barrett et al. (2006): Provide evidence of lumpy investments and borrowing constraints in rural Ethiopia and Kenya.

Final Thoughts

• The empirical evidence for widespread poverty traps is limited.
• While theoretical models are useful, they may not always reflect real-world dynamics.
• The focus should shift from broad-based interventions to targeted policy actions that address specific constraints.