2008年-世界发展银行全球_The_Impact_of_Remittances_on_Poverty_and_Inequality_in_Ghana_41页_272kb

Summary of "The Impact of Remittances on Poverty and Inequality in Ghana"

Core Content

This paper investigates the impact of internal and international remittances on poverty and inequality in Ghana using data from the 2005/06 Ghana Living Standards Survey (GLSS 5). It employs a two-stage multinomial logit model with instrumental variables to address selection and endogeneity issues, focusing on variations in migration networks and remittances among ethno-religious groups.

Main Findings

• Poverty Reduction: Both internal and international remittances reduce poverty in Ghana, but the impact differs by type.
  • Households receiving international remittances experience a 88.1% reduction in poverty levels.
  • Households receiving internal remittances see a 69.4% reduction in poverty levels.
• Income Inequality: Both types of remittances increase income inequality.
  • International remittances raise the Gini coefficient by 17.4%.
  • Internal remittances raise the Gini coefficient by 4%.
• Migration and Remittances: Migration is not always associated with remittances. In Ghana, only about half of all migrants remit, and 50% of remittance-receiving households do not have a migrant.

Methodology

• The paper uses a two-stage multinomial logit model to account for selection bias and endogeneity.
• Instrumental variables are derived from variations in migration networks and remittances across ethno-religious groups.
• The first-stage equation models the probability of receiving remittances, incorporating:
  • Human capital variables (education levels of household members).
  • Household characteristics (age of head, household size, number of males over 15, number of children under 5).
  • Migration network variables (number of female-headed households receiving remittances, square of internal migrants).
  • Ethno-religious characteristics (square of group income).
  • Regional and ethnic variables.
• The second-stage equation estimates household expenditure, using similar variables to the first-stage, but focusing on expenditure rather than income.

Data Overview

• The study uses a sub-sample of 4,000 households from the GLSS 5 survey.
• 59 households were excluded due to missing data, resulting in a final sample of 3,941 households.
• The data includes:
  • Internal remittances (from within Ghana).
  • International remittances (from African or other countries).
  • Money, food, and non-food goods as forms of remittances.
• The sample is divided into three groups:
  • 64.7% of households receive no remittances.
  • 29.8% receive internal remittances.
  • 5.4% receive international remittances.

Key Variables and Instrumental Variables

• Human capital: Education levels of household members.
• Household characteristics: Age of head, household size, number of males over 15, number of children under 5.
• Migration networks: Number of female-headed households receiving remittances, square of internal migrants.
• Ethno-religious characteristics: Square of group income.
• Instrumental variables:
  • International remittances as a percentage of household income.
  • International migrants as a percentage of the population in the ethno-religious group.

Model Identification and Validation

• The model assumes that the instrumental variables are uncorrelated with the unobserved components of the expenditure equation.
• The Anderson test and Cragg-Donald test are used to validate the model.
• These tests show that the instrumental variables identify the second-stage equation and are reasonably strong.
• The model is based on a linear version for testing, but the non-linearity helps in identifying the selection term.

Conclusion

• The paper highlights the importance of using econometric methods to evaluate the impact of remittances on poverty and inequality.
• It emphasizes that remittances are not just income transfers but have complex effects on economic outcomes.
• The results suggest that international remittances have a more significant impact on poverty reduction compared to internal remittances, but both types increase income inequality.