Published and Presented Works 

Mapping The Road Not Traveled: A Synthetic Control Evaluation of Health Care Reform in The Netherlands

Presented at the spring conference of The Association of Public Policy Analysis and Management · Apr 8, 2017. 

Motivation for Research:
Having been a staffer on Capitol Hill and watching passage of the Affordable Care Act were milestones in my career. However, I soon learned that passing policy was one thing, implementing it successfully was another. For almost a year, I worked on implementation of the Affordable Car Act for a healthcare Assocation. That experience, along with studying the cornucopia of systems in the rest of the world (it actually isn't a monolith of single-payer systems) made me curious about the comparative advantages, disadvantages and counterfactuals of adopting certain types of programs.  

Nations that guarantee universal access to healthcare for its citizens use one of three systems categories: single-payer, two-tiered (e.g., public options), or a private insurance mandate. While the advantages and disadvantages within each system have been debated in abstract and theoretical terms, empirical comparative research between these systems in lacking. In 2006, The Netherlands restructured its healthcare delivery by moving from a public option to a private insurance mandate. The goal of the reform was to improve efficiency in healthcare management. Numerous studies indicate the program achieved this goal as cost-controls and competition led to a decrease in spending rate increases. Currently, The Netherlands is ranked among the highest in the OECD bloc for both cost-effectiveness and patient satisfaction. However, the impact the reforms had on health outcomes has never been formally evaluated-- leaving open the question of whether health quality was sacrificed for cost savings.

This paper uses the synthetic control method to estimate the effects the law had on both short-term (potential years of life lost) and long-term (life expectancy at birth) health indicators. The results of the synthetic control were then validated using a fixed effects difference-in-differences estimator. Both models showed a statistically significant reduction in years of life lost. While life expectancy at birth improved, it was considered just above the threshold for statistical significance—likely the result of the small number of post-enactment observations. The results suggest that by switching to an insurance mandate, The Netherlands improved both patient outcomes and system efficiency. The implication is that regulated market forces can help lead to health outcome improvements in countries that provide universal healthcare. 

All That Glitters: A Meta-analysis of Microfinance in the Developing World
Published in The Public Purpose: Volume XIV · Apr 1, 2016. 

Motivation for Research:
When I lived in St Croix, I saw constant private and public investment flowing into the island yet poverty cycles still persisted. I was reminded of this when I took development economics in grad school and reading countless studies of repeated donor capital flows into the third world with no apparent long-term benefits for the indigenous people living there. I undertook this study to help shed light on why this was occurring.

Microfinance has become a popular tool over the last several decades to fight poverty in the developing world. However, broad systematic evaluation of the direct effects of these programs is lacking. Furthermore, individual studies show conflicting results both within countries and between regions. This paper helps fill this void by estimating the impacts of microfinance on poverty by using a meta-analysis combining data from randomized field experiments, quasi-experiments, and market reports from 38 Microfinance Institutions in Africa, Asia, North, and South America. This paper also tests the difference between two types of common financial products: micro-savings & micro-loans. The results indicate that while microfinance is potentially viable in the long term, it has not statistically significant effects of poverty rates in the short term (1-4 years). Entities that engage in development microfinance must be prepared to operate at a loss during this time. Furthermore, setting statistical significance aside, the results suggest that micro-savings programs are more effective than micro-loans at poverty reduction with the likely reason being that micro-loan participants do not usually possess complete information on both acceptable risk and effective financial management practices. The implications for stakeholders are that microfinance programs should be redesigned with the assumption that their effects will, on average, take 8-10 years to materialize and will need to be coupled with other interventions. This assumption is based on inferences from comparative extrapolation using similar data from the U.S. Federal Reserve. 


Working Papers 

Crossing the Potomac: A Regression-Discontinuity Analysis of Maryland's Medicaid Expansion

Author's Note:
A preliminary version of this paper was presented at the 2022 annual conference of the Society for Behavioral Medicine.

Motivation for Research:
I will never forget watching from the House gallery the final passage of the Affordable Care Act as a Congressional Aide, being in the middle of history in the making. As a healthcare economics researcher, I am curious what the specific benefits of each component of the Affordable Care Act had at the community-level. Given how important the Medicaid expansion was in lowering the overall uninsured rates in the U.S., I am equally as curious as to what the counterfactual would be if this component wasn't included. 

Using a regression discontinuity research design, Crossing the Potomac analyzes county-level indicators (including economic, healthcare, and provider numbers) in comparable Maryland and Virginia counties before and after Maryland's Medicaid expansion (Virginia expanded Medicaid in 2019). As well as incorporating for endogeneity between dependent variables, it utilizes environmental, social and structural covariates to further isolate the effects the Medicaid expansion had on counties. 

Incubating Urbana: An Epidemiological Approach to Studying Gentrification

Authors Note:
A preliminary version of this was presented at the 2023 Annual Urban Affairs Association (UAA) conference. It is being developed as part of my acceptance into the inaugural cohort of up-and-coming authors within the UAA.

Motivation for Research:
Growing up in a 99 percent black working-class neighborhood, it wasn't easy for me to make friends. Infact, I could count them on one hand. Starting in mid-90s and continuing until the early 2010s, I slowly watched the neighborhood completely gentrify. This including housing, economic and cultural displacement. The first of my friends to leave was Tyson. He was the charismatic cool leader I secretly tried to emulate. His departure was sudden and I never knew where he ended up. Then brothers JR and Roni moved to Prince George's County, MD. My best friend Sean moved across the river to Anacostia. We would still visit each other; it was only a 10-minute drive but ride coordination wasn't the easiest thing and it would take an hour to metro to each other. The last to leave was Rafi circa 2006. He was able to "hang on" not because he was smarter or wanted it more than the others. He survived the first wave of gentrification because, well, he was a drug dealer and diversified his... clientele early on. But eventually economics (and the DEA) caught up to him. This dynamic has made me focus on cutting-edge workforce development (and income/housing supports while participants are upskilling) as the key to diversified community resilience.  

This paper uses 12 years of Census Bureau microdata to develop and new theoretical model of gentrification and its resulting effects on community characteristics. Furthermore, it uses epidemiological techniques to map and predict various "strains" of gentrification in American cities. 

None Need Apply: How Covid Altered the Balance Between Race, Gender and Class in Hiring

Motivation for Research:
It is no secret that there are biases in hiring decisions based on things like race, sex, age, incarceration status, education, residential status and other socio-economic factors. I didn't have to (but I did) read economic statistics and watch news reports to know that the Covid pandemic exacerbated and altered many of these fault lines-- During 2020 and 2021, I could see it out my bedroom window and walking down the streets of Richmond, Virginia. 

This paper uses data from the Virginia Employment Commission (VEC) and the State Council on Higher Education in Virginia (SCHEV) to study hiring patterns before and after the emergence of the Covid 19 pandemic.  It compares the dynamics both between and within groups based on demographics, income and geography. 

The Death and Life of Redemption: A Study of Recidivism

Motivation for Research:
Shawn was one of my best friends growing up in Potomac Avenue. Despite living in a notorious housing project, he was dreaming big of being a downtown DC lawyer. To make this happen, he worked a part time retail job near Capitol Hill while getting his pre-law bachelor's. One night after getting off a grueling shift, he had an encounter with an MPD officer with a reputation in the community for brutality and incitement. Shawn's hopes of achieving his dream ended that night with a questionable arrest and charges. After being released, he sank into depression and soon dropped off the grid--I never heard from him again. Along with reform of police culture, hopefully this study will provide insights into what we can do in the future to save people like Shawn. 

This paper uses longitudinal parole data supplemented with other government data to study how geography, regional economics, social networks, demographic characteristics, workforce development initiatives, mental health and health care access affect the likelihood of recidivism.   


Keeping watch in the gardens of the city: Evaluating police response times and distances to neighborhood greenspaces in Richmond, Virginia

with Amy Clifton-Mills


Incident response time is one of the most highly researched aspects of policing. Previous research focused on characteristics such as neighborhood disorganization and crime rates as major determinants of response times. Recently, the focus has shifted to the relationship between greenspace (i.e., parks) and crime rates. However, greenspace is also linked to other known factors influencing both crime rates and response times, including changing neighborhood demographics and economics (i.e., redevelopment, disinvestment, gentrification). Using data on calls for service, parks, neighborhood socioeconomics, and demographics, this study helps bridge the gap between these disparate foci.

In a first-stage analysis, we categorize data on incident characteristics (including type, severity, time of day and police response time), greenspace features (including distance to the incident, presence of fencing, playground equipment and open fields) and socioeconomics of the census tract the incident occurred in. Using GIS software, we test the hypotheses that the geo-spatial distribution of incidents will differ by the categories in each typology. In the second stage, using regression analysis at the census tract-level, we examine proposed associations among the number of calls for service per year, amount of greenspace in the tract, unemployment rates, resident demographics, median income, and average home value. We also analyze the relationship between police response time, distance between the incident and the nearest greenspace, type of incident, census tract socioeconomics, as well as temporal factors including time, day and month of the incident.

The findings highlight the nuanced intersectionality of crime, geography and socio-economics, especially gentrification as a mediator of the relationship between police response time and greenspace.

Where country roads take us: The COVID-19 crisis, rural Virgina, and health equity implications.
with Britanny "Brie" Haupt, Forthcoming in the International Journal of Disaster Risk Reduction.


The Coronavirus (COVID-19) was declared as a worldwide pandemic on March 11th, 2020. In the following weeks, many state governments in the United States would issue stay-at-home orders and other extraordinary measures to combat the spread of the COVID-19 pandemic. However, implementation of these measures throughout 2020 and 2021 was haphazard and uneven. There emerged many challenges and outcome disparities on top of traditional American fault lines of race, geography, culture and economics. The primary focus of this study is to understand why COVID-19 had disproportionate impacts, including increased infections, hospitalizations and death rates on marginalized populations in rural and minority communities.

In this context, the study examines how variations in structural factors such as wealth, workforce roles, health care infrastructure, risky personal behavior and underlying community morbidity interacted with race and rurality. After analyzing data collected from the United States Census Bureau, Virginia Department of Health, Robert Woods Johnson Foundation and Environmental Systems Research Institute, our results suggest that differences in workforce roles associated with rural and minority areas were the primary driver in increased COVID-19 infections while underlying health issues such as diabetes were the biggest factors in increased COVID-19 hospitalizations and deaths for those under 75 in these areas. Furthermore, these two dimensions are endogenous to each other, suggesting they need to be addressed in tandem. The implications of this study are that devoting additional resources to underlying community-level structural factors will lessen the severity of future pandemics.

Crisis Communication Strategies by Food Bank Network during a Global Pandemic.
with Britanny "Brie" Haupt and Lauren Azevedo, Forthcoming in the Journal of Homeland Security and Emergency Management 


During the novel coronavirus pandemic (COVID-19), nonprofits needed to quickly rethink their

strategies concerning the sustainability of their organizations along with communication

regarding the crisis. The pandemic presented unique opportunities for understanding how

service-oriented nonprofits, who must share timely and accurate information across stakeholders

and work closely with partner organizations, can deliver services during crisis periods. Utilizing

a case study approach, the focus of this study is the impact of COVID-19 on the Federation of

Virginia Food Banks (FVFB), which is a 501(c)(3) nonprofit state association of food banks

affiliated with Feeding America, by examining the use of crisis communication strategies

utilized or not-utilized during the crisis response and recovery activities. The findings present

practical implications for nonprofit organizations and their crisis response and recovery networks

and a better understanding of the needs for nonprofit organizations to engage in crisis

communication planning for diverse crises and planning resources.