Regmi, Ganesh (2015-05). Dynamic Relationships Between Immigrants and US Gross Domestic Product Using A Vector Error Correction Model (VECM). Master's Thesis. Thesis uri icon

abstract

  • Debate over immigration including the role visa policies and immigrants play in the US economy, especially effects on wages, gross domestic product (GDP), employment rate, and consumption remain unresolved. This study investigates the dynamic relationships among the selected economic variables and the number of immigrants to the United States. Variables included are annual total number of immigrants, US GDP, investment in education, national hourly wage rate, and energy consumption from 1964 to 2011. These variables are found to be non-stationary via augmented Dicky-Fuller tests and cointegrated with four cointegrating vectors. A vector error correction, therefore, is used in the analysis. Directed acyclic graphs are used to find contemporaneous causal relationships between the variables. DAGs showed, GDP and wage are source of information, energy both receives and provide information in the system, investment in education is only receiver of the information while immigrants are contemporaneously exogenous. Tests of exclusion find all the variables are in the cointegrating space suggesting all variables share long run relationships. Exogeneity test suggests that all variable responses to the perturbations in the long-run relationships. Result shows that in the short run, wage has a negative reaction to a shock in GDP. All variables except number of immigrants' response positively to one time innovations in investment in education. Increases in immigrants will has a negative effect on the other variables in short-run. The number of immigrants, in the short-run, do not respond in the innovations in the other variables. Similarly, any shock in energy consumption will not be responded by any the variables in short-run. Forecast error variance decompositions suggest in short-run a variable is mainly explained by itself; as one moves time ahead forecast the share of other variables becomes larger in explaining a variables forecast error. Wages explain a large amount of the variability in investment in education. All the variables are cointegrated and any policies implemented to increase or decrease a single variable has effect on other rest of the variables. So policy maker should consider the macroeconomic effect in the system.
  • Debate over immigration including the role visa policies and immigrants play in the US economy, especially effects on wages, gross domestic product (GDP), employment rate, and consumption remain unresolved. This study investigates the dynamic relationships among the selected economic variables and the number of immigrants to the United States. Variables included are annual total number of immigrants, US GDP, investment in education, national hourly wage rate, and energy consumption from 1964 to 2011. These variables are found to be non-stationary via augmented Dicky-Fuller tests and cointegrated with four cointegrating vectors. A vector error correction, therefore, is used in the analysis. Directed acyclic graphs are used to find contemporaneous causal relationships between the variables. DAGs showed, GDP and wage are source of information, energy both receives and provide information in the system, investment in education is only receiver of the information while immigrants are contemporaneously exogenous. Tests of exclusion find all the variables are in the cointegrating space suggesting all variables share long run relationships. Exogeneity test suggests that all variable responses to the perturbations in the long-run relationships.

    Result shows that in the short run, wage has a negative reaction to a shock in GDP. All variables except number of immigrants' response positively to one time innovations in investment in education. Increases in immigrants will has a negative effect on the other variables in short-run. The number of immigrants, in the short-run, do not respond in the innovations in the other variables. Similarly, any shock in energy consumption will not be responded by any the variables in short-run. Forecast error variance decompositions suggest in short-run a variable is mainly explained by itself; as one moves time ahead forecast the share of other variables becomes larger in explaining a variables forecast error. Wages explain a large amount of the variability in investment in education.

    All the variables are cointegrated and any policies implemented to increase or decrease a single variable has effect on other rest of the variables. So policy maker should consider the macroeconomic effect in the system.

publication date

  • May 2015