HOW DO MIGRATION COSTS AFFECT PRODUCTIVITY?

Responsible researcher: Bruno Benevit

Original title: The Aggregate Productivity Effects of Internal Migration: Evidence from Indonesia

Authors: Gharad Bryan and Melanie Morten

Intervention Location: United States

Sample Size: 2,481,000 men

Sector: Work

Variable of Main Interest: Salary

Type of Intervention: Migration cost

Methodology: OLS, PPML

Summary

In developing countries, recent evidence indicates that greater ease of internal migration has the potential to increase productivity. The objective of this study was to verify how the reduction of internal migration costs in Indonesia impacts aggregate productivity in Indonesia, in addition to verifying how the presence or absence of amenities affects the average salary of the population in the destination location. Presenting a general equilibrium model and using various econometric methodologies, the authors identified that the elimination of barriers increases productivity and that the absence of amenities at the destination is compensated by higher wages.

1. Policy Problem

Several economists seek to identify how internal migration can affect decisions in the labor market. In developing countries, specifically, recent studies suggest that greater ease of internal migration has the potential to increase productivity. Although there is evidence of positive effects of seasonal migration on consumption (Bryan, Chowdhury and Mobarak, 2014), the aggregate impact of internal migration costs on wages involves a different decision-making process (Bryan and Morten, 2019).

According to the authors, when considering migration from one location to another, individuals compare not only the income perspective between the original location and the destination considered, but also the migration costs and the amenities that both locations offer. In this sense, individuals choose to migrate only if their earnings grow sufficiently to offset the costs of this migration. Reducing this type of barrier has the potential to improve the disposition of workers between different regions according to their skills (Bryan and Morten, 2019).

2. Policy Implementation Context

Migration in Indonesia is predominantly characterized by a single episode of permanent migration in adulthood. Among Indonesian male householders who move outside their province of birth, 69% make just one migration, 26% make two, and just 5% make three or more moves. Only 8% of migrations involve people under the age of 16, and half of second migrations are made by people returning home. These patterns are comparable to those observed in the United States, where the average number of location changes for male migrants is 1.98 and 50.2% of them return to their hometown.

The authors presented five stylized facts to understand how aggregate wages are associated with workers' location. The mechanisms elucidated by these motivational facts highlight the influence of distance on migration, its impact on productivity and the relationship between wages and local amenities. Taken together, the five facts suggest that increasing workforce mobility can result in productivity gains.

3. Assessment Details

This study used microdata from Indonesia and the United States. Data for Indonesia comes from the Intercensal Population Survey (SUPAS) and the National Socioeconomic Survey (SUSENAS) of 2011 and 2012. This dataset covers information on formal workers such as place of birth, current place of work and monthly income. To verify the effects considering the presence of self-employed workers, the study complemented the SUPAS/SUSENAS data with detailed information from the Indonesia Family Life Survey (IFLS), which covers a longer period and includes information on self-employed people. In the USA, data comes from the 1990 census and the 2010 American Community Survey. To obtain measures of amenity for locations in Indonesia, data from the Village Potential Statistics (PODES) were used. In order to observe how job-motivated migrations are affected, the authors restricted the final samples to male heads of households between 15 and 65 years old.

4. Method

To explain how migration and amenity costs affect productivity, the authors formalized five stylized facts associated with the relationship between the place of origin and destination of migration. The first fact establishes that the proportion of people who migrate to a given location decreases as the distance increases. The second fact stipulates that the average salary of people who migrated increases as the distance between the place of origin and destination increases. The third fact determines that the elasticity of average wages in relation to the share of the population of origin is negative. The fourth fact states that migration costs reduce productivity by reducing the selection of workers. Finally, the fifth fact establishes the existence of compensatory wage differentials, where places with greater amenities have lower aggregate wages.

Due to the census nature of the data used, the study's migration measure represents permanent migration based on a repeated cross-section. This assumption is corroborated by the analyzed IFLS data: migration in Indonesia can be characterized as a single episode of permanent migration. Mobility costs are defined as the straight-line distance between the place of origin (birth) and destination (for those who migrated to the current location). The variable that identifies amenities is defined as a single measure using six different amenity criteria (positive and negative).

To identify the relationship between migration costs and aggregate productivity, the ordinary least squares (OLS) method was used to estimate the five stylized facts for Indonesia. For the United States, only the first four facts were estimated.

Finally, the study presents a static general equilibrium model of migration, adapted from the work ordering model in Hsieh et al. (2019). The model formalizes that workers are born at a specific origin, acquire a skill for each destination, and select between destinations based on wages, amenities, and migration costs. Migration costs are relative to the place of birth, and wages and amenities are endogenous and adjust to ensure balance, so that locations have different sets of required skills. To estimate this model, a Poisson Pseudo Maximum Likelihood (PPML) model was used.

5. Main Results

The results of estimating the five stylized facts revealed the existence of a significant influence of movement costs and amenity differentials on migration and average wages across locations in Indonesia. Regarding movement costs, it was found that a 10% reduction in the distance between two locations resulted in a 7% increase in the proportion of migrants between these locations.

It has been observed that people who live further away from their birthplaces tend to have higher salaries, indicating that there is a need to financially compensate people to encourage them to move away from their hometowns. When the distance between origin and destination doubles, there is a 3% increase in average wages. These results suggest that moving costs play a key role in people's decision to move and in determining wages.

Furthermore, the analysis revealed the importance of selection effects, where the greater the proportion of people born in a given location who move to another, the lower the average salary of these migrants. The results for the United States showed similar behavior in terms of magnitude and significance. Regarding differences in amenities between locations, it was found that Indonesian workers in locations with low amenities receive higher wages, reflecting the need to compensate those who choose to live in areas with a lower quality of life.

The calibration of the structural model parameters indicates that the United States has lower migration costs compared to Indonesia. The model estimates showed moderate gains in aggregate productivity, showing great heterogeneity. Removing all barriers to migration is predicted to increase productivity by 22%, with greater gains for some origin locations, reaching 104% - these gains are most significant in locations where average wages have greater variation between destinations. Considering moving costs at US levels, counterfactual calculations indicate a 7.1% increase in average wages in Indonesia.

6. Public Policy Lessons

This article analyzed the impact of movement costs and convenience differentials on migration and labor productivity. Using econometric methods, the authors estimated five stylized facts associated with the topic. Additionally, the authors also presented a general equilibrium model of migration, considering migration costs and differentials in required skills and amenities between locations. Evidence from the study indicated that the costs of worker migration are offset by higher wages. In the same sense, the absence of amenities is also compensated. The results also indicate that migrant selection plays an important role, where a higher proportion of people moving to a specific destination is associated with lower average wages.

These results have relevant implications for the formulation of public policies. While migration that improves static labor allocation may not have as substantial impacts as some studies suggest, targeted policies can have significant effects on specific communities. Therefore, policies aimed at reducing movement costs can contribute to increasing productivity and improving living conditions in specific regions.

References

BRYAN, G.; CHOWDHURY, S.; MOBARAK, AM Underinvestment in a Profitable Technology: The Case of Seasonal Migration in Bangladesh. Econometrica , v. 82, no. 5, p. 1671–1748, 2014.

BRYAN, G.; MORTEN, M. The Aggregate Productivity Effects of Internal Migration: Evidence from Indonesia. Journal of Political Economy , vol. 127, no. 5, p. 2229–2268, Oct. 2019.

HSIEH, C.-T. et al. The Allocation of Talent and US Economic Growth. Econometrica , v. 87, no. 5, p. 1439–1474, 2019.