Essays in Labor Economics and Econometrics Público

Cahn, Yisroel (Spring 2022)

Permanent URL: https://etd.library.emory.edu/concern/etds/b8515p66n?locale=pt-BR
Published

Abstract

This dissertation explores methods of estimating and evaluating policy interventions that have heterogeneous effects on the outcome of interest.

Policy evaluation generally compares an observed outcome of interest with the estimated counterfactual outcome had the policy not been implemented. Chiefly , this is done by comparing means or means conditional on a subgroup of the population. However, if the policy has heterogeneous effects on the distribution of outcomes — for example, if a policy affects low-wage workers differently than high-wage workers, then simply comparing means masks the diversity of outcomes a policy maker might be interested in. This is particularly relevant if the policy maker is interested in inequality or poverty. The first chapter surveys the literature on interpreting heterogeneous outcomes of a policy intervention.

The second chapter looks at an example — minimum wage policy. Recent proposals to raise the U.S. Federal minimum wage to $15 an hour are designed to improve the welfare of low-wage workers, but may involve important economic trade-offs. Although the effects of minimum wage on employment and wages have been well studied, little is known about its effects on hours worked which may be responsive to minimum wage changes. One reason for the lack of research is that hours are only observed for those who are employed, but workers could be exiting or entering the market and biasing the results. I fill this gap in the literature by employing a Heckman-type selection model to estimate the effects of minimum wage on hours worked, accounting for possible employment effects. Using U.S. Current Population Survey data, I found that increases to the minimum wage increased the hours worked of low-wage workers. However, I also found that the effects varied by industry—fast food and accommodation service workers saw their hours decrease while most other industries’ minimum wage workers saw their hours increase, suggesting that market structure could be causing these findings. Additionally, I propose a new method to estimate jointly determined outcomes, showing that hours and wages both increased for low-wage workers.

The final chapter uses machine learning methods to predict intergenerational income mobility in the United States. The machine learning methods are (1) non-parametric and are not sensitive to functional form misspecification, (2) give an out-of-sample performance indicator, and (3) allow for predictors to be ranked by how importance they are to the overall prediction. I find that family wealth, and not parent income, is the most important predictor of child income for large increases in the income distribution. 

Table of Contents

Inequality and Policy Evaluation 1

Introduction 1

Facts about Inequality and Poverty in the US 3

ComparingDistributions 5

Measuring Inequality and the Social Welfare Function 5

Conclusion 12

Estimating Jointly Determined Outcomes: How Minimum Wage Affects Wages and Hours Worked 13

Introduction 13

The Minimum Wage Puzzle 13

Wages and Hours as a Joint Outcome 20

The Remainder of the Chapter 25

Literature Review 26

Predictions of Minimum Wage Effects on Employment 26

Empirical Results on Low-wage Workers 27

Distribution Effects of Minimum Wage 28

Employment Selection Model 30

Joint Distribution Model 31

Counterfactual Analysis Setting 31

Identification 33

Counterfactual Distribution 34

Estimation of the Counterfactual Distribution 35

Inference 37

Data 38

Data Cleaning 38

DataVisualization 39

Results 47

Effects on Minimum Wage Workers 47

Effects on Minimum Wage Workers by industry 50

Joint Distribution Results 50

Distribution Effects with Conditional Mean Methods 62

Conclusion 64

Who Needs Help: Off-the-shelf Machine Learning Methods to Predict Income Mobility 65

Introduction 65

Terminology 67

Data 67

Methods 71

Measuring Mobility 71

Random Forest 73

Gradient Boosting 74

Variable Importance 75

Results 76

Conclusion 83 

Appendix A Simulation 84

Appendix B Proofs 88

Appendix C Jointly Dependent Outcomes 90

Bibliography 93 

About this Dissertation

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
School
Department
Degree
Submission
Language
  • English
Research Field
Palavra-chave
Committee Chair / Thesis Advisor
Committee Members
Última modificação

Primary PDF

Supplemental Files