Date of Award
2024
Degree Type
Restricted to Claremont Colleges Dissertation
Degree Name
Economics, PhD
Program
School of Social Science, Politics, and Evaluation
Dissertation or Thesis Committee Member
Gregory DeAngelo
Dissertation or Thesis Committee Member
Scott Cunningham
Dissertation or Thesis Committee Member
Mark Hoekstra
Terms of Use & License Information
Rights Information
© 2024 Muhammad Salman Khalid
Keywords
Causal Inference, Marriage, Police, Police Training, Vaccination
Subject Categories
Criminology | Economics | Public Health
Abstract
Causal inference methods in economics serve as vital tools for disentangling cause-and-effect relationships. These methods enable economists to find the impact of policy interventions, and market changes on desired outcomes. Employing techniques such as randomized controlled trials, instrumental variables, difference-in-differences, and regression discontinuity designs, economists can isolate the causal effect of specific variables while controlling for confounding factors. This dissertation applies these methods in different settings, to draw the causal relationship answering the said question. The first chapter of the dissertation applies the event study design to address the effectiveness of police training in providing procedural justice. The second chapter explores the difference-in-difference technique to draw the causal relationship between the vaccination monitoring system in Pakistan on vaccine uptake and coverage. Finally, the third chapter uses callaway and sant’anna method and Goodman-bacon decomposition to show how difference-in-difference with staggered timing can yield biased results. The first chapter of the dissertation explores the effect of police training in providing procedural justice. The reasonable use of force by police officers has remained a hot topic among academics and policymakers after the George Floyd incident. This paper draws a causal relationship between police training and different measures of escalations like the use of force, escalated events, and arrests. These three types of police training are used for the analysis i.e. use of force training, de-escalation training, and crisis intervention training. Multiple data sets were obtained from the Dallas police department including 911 calls data, use of force data, arrest data, training data, and police personnel data. This paper uses the event study design to draw the causal estimates and finds a statistically insignificant effect of these trainings on different outcome measures. The paper found strong evidence of no effect of these trainings on providing procedural justice in the field. The second chapter focuses on the digital vaccination monitoring system in Pakistan and how it affects vaccine uptake and coverage. In late 2014, E-VACCS (an Electronic Vaccine Registration System) was implemented across the province of Punjab, one of the four provinces in Pakistan. We use yearly data (2010-2019) from the Pakistan Social and Living Standards Measurement Survey (PSLM) in a difference-in-differences framework to exploit the natural across-province over-time variation offered to us by Punjab’s adoption of E-VACCS. We analyze the impact of E-VACCS on complete vaccination coverage as well as antigen-specific vaccination rates. Treatment effects on vaccination coverage are largely driven by specific subpopulations. The intervention led to a 2.2% increase in the proportion of children having received at least 1 dose of vaccination. Across urban households, we find a 6.2% increase in vaccination coverage on the intensive margin. These treatment effects in urban areas are limited to mid-high income households. Our estimates for antigen-specific rates of vaccination coverage confirm similar trends. Among urban mid-high income households, improvements occurred in the antigen-specific coverage rates for vaccines administered towards the end of the childhood vaccination cycle (e.g. D-Tap-3 or Hep-B-3). However, for rural low-income households, we find negative treatment effects for antigen-specific vaccine coverage rates, particularly for the three hepatitis doses. The results suggest that the increase in vaccination coverage across urban mid-high-income households came at the expense of a significant decrease in coverage across low-income rural households, signaling a redirection of critical resources. Our estimates remain robust to several variations in the specifications. Further robustness is also confirmed through event study estimation methods. Finally, the third chapter uses callaway and sant’anna method and Goodman-bacon decomposition to show how difference-in-difference with staggered timing can yield biased results. Stevenson and Wolfers (2006) theorized that unilateral divorce laws shifted power from one party to a distribution of power across both parties in the marriage, effectively providing an application of Coasian Bargaining. Utilizing a two-way fixed effect (TWFE) difference-in-difference estimator, they found that unilateral divorce had significantly reduced suicide rates, domestic violence, and intimate homicide. Innovations in econometric theory have raised concerns regarding the use of TWFE with differential timing in the treatment variable, leading to biased estimation. We revisit Stevenson and Wolfers (2006) with more modern estimators for suicide and intimate homicide rates and utilize appropriate estimators to examine the effect of unilateral divorce laws on suicides and intimate partner homicide rates. In contrast to the original research, we do not find significant effects of unilateral divorce on suicide and intimate homicide rates. This indicates that on net these laws were not beneficial to women’s outcomes.
ISBN
9798382742007
Recommended Citation
Khalid, Muhammad Salman. (2024). Applications of Methods in Causal Inference. CGU Theses & Dissertations, 782. https://scholarship.claremont.edu/cgu_etd/782.
