Labor Market Effects of Refugees in the United States: Evidence from a Difference-in-Differences Analysis

Labor Market Effects of Refugees on Local Economies in the US

1. Introduction

The United States has been a big refuge to most refugees since the introduction of the Refugee Act of 1980 and has so far resettled over three million refugees. The role played by refugees in the U.S. labor market is not a recent topic, but the implications of such inflow remains a debatable topic in the literature of policies and economy. The years of 2015-2017 have seen a great amount of immigration into the U.S. and this is quite a significant year to conduct an empirical analysis. This federal policy that remained a natural experiment concerning policymaking and exploring the potential consequences of such a large influx of refugees on the labor market had become a natural experiment even before, as it lowered the influx of refugees into the country by radical means since 2017. Nevertheless, new economic studies have been trying to add a complementary advantage, where as refugees who lack labour markets can work in recalls, drive consumption demand and even entrepreneurship in the host economy. The actual impact of the labor market needs to be developed to formulate the evidence-based policy to balance the humanitarian policies and the economic factors.

The proposed study seeks to provide an estimate of the causal impact of inflows of refugees on the local labor markets in U.S. states and metropolitan areas in the period between 2013 and 2017. We employ a Difference-in-Differences (DiD) design to contrast the pre- and post-2015 results in the labor market in both high-inflow and low-inflow states against the results of the 2015 refugee wave. Interest outcomes involve the employment rates, the labor participation, and the real wages. The data will be obtained using various publicly available data sources, namely WRAPSNet of the refugee placement, American Community Survey (ACS) of microdata, and BLS Local Area Unemployment Statistics (LAUS) and Quarterly Census of Employment and Wages (QCEW) of area-based labor statistics.

This research combines microdata of individuals and aggregate data of refugee inflows to offer nationally representative evaluation of the faculties of local labor markets to assimilate refugee populations. Demographic controls like age, education, occupation are also considered in the empirical approach and the potential confounders are met by the fixed effects and robustness tests. The study is an addition to the literature since it provides causal estimates, tests policy-relevant hypothesis and provides advice on the economic impact of refugee resettlement.

2. Data

2.1 Data Sources

The study incorporates a number of publicly available data:

1.WRAPSNet / Refugee Processing Center (RPC): Data on 2013-2017 Annual refugee arrivals by state, with nationality, and place of resettlement. Based on these data, the treatment variable, which is the refugee inflows per 1,000 residents is constructed.

2.American Community Survey (ACS, IPUMS PUMS): Micro-data with individual factors of employment, income, education, occupation, and demographic data. The ACS permits individual-level estimation of the outcomes of the labor market and the necessary amount of controls can be included.

3.BLS Local Area Unemployment Statistics (LAUTS): Provides state and county level labor market data, such as the employment, payroll and wage data.

2.2 Variable Construction

Outcome Variables:

Employment rate: this is a measure of the employed.

Labor Force Participation: Empstat Indicator.

Wages (incwage): Univariate continuous variable that is the annual income, but is winsorized to 1 percent to minimize the effects of outliers and scaled to 2014 dollars.

Treatment Variable:

Refugee Inflows per 1,000 Residents: This has been calculated by summing up annual arrival of refugees by state and dividing it by the population.

Control Variables:

• Age, education (educd), occupation score (occscore), and gender, survey weights (perwt, hhwt).

Merging Process: Statefip and year were used to merge the refugee arrivals reorganized by state and year with the ACS microdata. State identifiers that were not available were dealt with by IPUMS state codes. The obtained dataset comprises more than 22 million observations at an individual level and this has provided adequate statistical power.

2.3 Data Limitations

There must be a number of limitations to consider:

·       Measurement error: The ACS microdata has the potential to give inaccurate results on employment status of recent arrivals, which can then bias the outcomes of incumbent workers.

·       Geographic aggregation: Refugee inflows are state based thus concealing heterogeneity within a state or metropolitan jurisdiction.

·       Confounding effects: Local shocks (e.g. economic booms, natural disasters) could also be a factor and may potentially lead to bias in estimates when linked with the resettlement of refugees.

2.4 Summary Statistics

This table shows that there is variation among states and years. The patterns of inflow allow observing the significance of using heterogeneity in analyzing data.

3. Empirical Strategy: Difference-in-Differences (≈450 words)

3.1 Why DiD?

The DiD framework allows causal inference by comparing changes in outcomes over time between treated (high-inflow) and control (low-inflow) units. By including fixed effects for states (γ_i) and years (δ_t), this approach controls for unobserved time-invariant state characteristics and common time shocks.

3.2 Model Specification

The primary empirical model is:

   =  (* ) +  +  +   +

Where:

• = labor market outcome (employment, participation, wage) for state iii in year ttt
•  = refugees per 1,000 residents
• = indicator for post-2015 period
•  ​ = vector of individual-level controls (age, education, occupation)
• , ​ = state and year fixed effects
• ​ = clustered error term at the state level

Coefficient Interpretation:  represents the causal effect of refugee inflows on labor market outcomes.

3.3 Key Assumptions

1. Parallel Pre-Trends: High- and low-inflow states would have followed similar trends absent refugee inflows.
2. No Spill overs: Refugee inflows in one state do not affect outcomes in other states.

3.      Exogeneity of Refugee Placement: The assignment of refugees is not related systematically with unobservable shocks of the labor markets.

4. Results and Robustness

4.1 Summary Statistics

• Average employment rate: 0.62
• Labor force participation: 0.68
• Real hourly wage: $23.45
• Refugee arrivals per 1,000 residents: 4.5

Interpretation: The overall equilibrium in labor markets is divided between the high-and low-inflow states. The inflows of refugees are heavily varied, and it is possible to identify them.

4.2 DiD Estimates

Example Regression Output (Incwage):

Interpretation:

• Employment and participation remain stable, suggesting no crowding out.
• Wages increase modestly, consistent with complementarity between refugees and local labor.

4.3 Event-Study Analysis

• Pre-trends are parallel, validating DiD.
• Effects concentrated in high-inflow states but small in magnitude.

4.4 Robustness Checks

1.County-Level Analysis: Replacing the measures of the inflows at the state level with the measures at the county level gives the same results.

2.Instrumental Variable: The results are not materially different when lagged refugee allocations are used.

3.Placebo-Time-Test: Pre-2015 Years are null, proving DiD assumptions.

5. Discussion and Policy Implications

According to it, the inflows of refugees between 20152017 had no or slightly positive impacts on the labor markets. Participation and employment levels were not volatile, with the wages growing at a tame rate. The high-inflow states did not take away jobs afforded to incumbent workers, which reflects labor market flexibility and economic complementarities.

Policy Implications:

Policy Implications:

• Resettlement Programs: Refugees can be integrated without labor market disruption.
• Targeted Training Programs: Ensuring refugees complement local skills may amplify positive outcomes.

·       Legal: Higher inflows allow the state to have policy which promotes entrepreneurship and consumption.

Limitations:

• State-level aggregation may mask local heterogeneity.
• Employment misclassification in ACS may introduce measurement error.

·       Various shocks may be unobserved.

6. Conclusion

This paper presents a strict examination of inflows of refugees concerning the U.S. local labor markets with the framework of DiD. It uses the WRAPSNet data on refugees combined with ACS microdata on refugees and BLS data to analyze the impacts on employment, participation, and wage. According to the findings, the inflow of the refugees, which occurred in the country in 201517, did not affect the labour market performance negatively but, in certain instances, raised wages marginally. Strongness tests are tests that are done on the basis of county level data, IV and placebo test is used to help in the interpretation of causation.

 Appendix

Panel Setup & Lagged Variables:

DiD Regression:

Time Placebo Check