2005年-世界发展银行全球_Migration_Trade_and_Foreign_Direct_Investment_in_Mexico_24页_425kb

Summary of "Migration, Trade, and Foreign Direct Investment in Mexico"

Core Content

This article examines the relationship between trade, foreign direct investment (FDI), and migration within Mexico, using the 2000 census data to evaluate the mechanisms behind the expected impact of NAFTA on migration to the United States. The study addresses the challenge of measuring illegal migration directly by focusing on internal migration flows across Mexico’s 32 states.

Main Points

• Rationale for NAFTA: The North American Free Trade Agreement (NAFTA) was expected to increase trade and FDI, thereby creating jobs and reducing migration to the United States. However, due to the difficulty of measuring illegal migration, the study uses internal migration data as a proxy.

• Migration Determinants: The analysis incorporates variables such as cost of living, amenities, networks, and labor market conditions (wages, unemployment). The study finds that labor market variables significantly influence migration decisions, especially when credit constraints are controlled for.

• FDI and Trade Effects: Greater exposure to FDI and trade is associated with reduced outmigration, primarily through labor market channels. The study suggests that FDI and trade can act as substitutes for migration, deterring it by improving employment and wage opportunities.

• Liquidity Constraints: The study highlights the role of liquidity constraints in migration decisions. These constraints can limit the ability of individuals to move, but FDI and trade may alleviate them, potentially reducing migration.

• Empirical Findings: On average, a doubling of FDI inflows leads to a 1.5–2 percent drop in migration. This finding supports the idea that FDI and trade have a measurable impact on migration behavior.

Key Variables and Data

• Migration Flows: Based on the 2000 census, migration data reflect the state of residence five years earlier. This method captures internal migration but may miss those who left and returned.

• Moving Costs: Approximated by a quadratic function of distance, including opportunity costs and communication costs. These costs are found to have a negative and decreasing effect on migration.

• Network Effects: Measured by the proportion of the population that has migrated to a destination state from a source state more than five years earlier. Networks are found to be strongly significant in reducing migration.

• Cost of Living: Two indices are used, one based on a consumption basket and the other on housing prices. The cost of living variable is found to be statistically significant, though its interpretation remains ambiguous.

• Amenities: Captured through a new index that includes urbanization, health infrastructure, education, and other characteristics. Amenities are found to have a moderate effect on migration.

• FDI and Trade Variables:

  • FDI: Per capita FDI data are from the Central Bank of Mexico, with a dummy variable for the Federal District to account for measurement error.
  • Maquila Value Added: Serves as a proxy for both FDI and trade, with data from industrial surveys. It is moderately correlated with FDI.
  • Exports and Imports: Export data are from the Ministry of Finance, while import data are from Bancomex. Both variables are associated with migration, though their effects may be complex and multifaceted.

Methodology

• The study uses a multinomial probit model, which accounts for the independence of irrelevant alternatives (IIA) violation found in multinomial logit models.
• The index function $I^* = V_k - V_i - C$ is used to model migration decisions, where $V$ represents indirect utility, and $C$ represents costs.
• Relative and freestanding variables are used to differentiate between the effects of destination and origin characteristics.

Results

• Labor Market Variables: In the complete sample, most labor market variables (except freestanding unemployment) are significant. In the restricted sample, they show the expected signs and significance, indicating that previous studies may have failed to account for credit constraints.

• Integration Variables:

  • FDI and trade are found to reduce migration significantly.
  • Freestanding FDI and trade terms are significant in some specifications.
  • The results suggest that FDI and trade can deter migration by improving labor market conditions and reducing liquidity constraints.

• Amenities and Networks: Interaction between amenities and network variables is observed. Including network terms reduces the significance of relative amenities, suggesting potential endogeneity or correlation with omitted variables.

• Migration to the U.S.: The study offers tentative inferences that increased FDI in Mexico may reduce migration to the United States, though this effect is less direct and requires further investigation.

Conclusion

• The study provides robust evidence that FDI and trade have a deterrent effect on migration, primarily through labor market improvements.
• It challenges the conventional view that labor market variables have little impact on migration, showing their importance when credit constraints are considered.
• The findings support the claim that NAFTA may reduce incentives for migration, but the effects are not uniform and require careful interpretation due to the complexity of migration determinants.