A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.



Balancing Panel Data and Removing Duplicate Unique IDs

11 minute read


Properly balancing panel data and removing duplicate unique identifiers is a recurring challenge for students in my Research Practicum course. Accordingly, I thought it would be helpful to provide an example for everyone.

Hello, World/Importing Data in R

less than 1 minute read


Hello, world! I have finally figured out GitHub and how to make a free website with Jekyll. It’s a great day!



Preventing the Negative Externalities of Development: Aid Compliance, State Capacity, and At-Risk Groups

This paper examines the potential negative externalities of foreign aid projects: that is, costs that accrue to people outside the aid transaction between the state (the implementer) and the aid financier (the supervisor). In contrast to most literature on foreign aid, I argue that compliance with policies to prevent negative externalities in aid projects mostly does not involve strategic calculations by donors or financiers. Instead, compliance with aid financiers’ social and environmental risk management policies mostly relates to two components of state capacity: aid recipients’ fiscal/taxation capacity, which proxies for states’ social contracts with their citizens and levels of economic informality; and the extent to which aid projects take place in urban areas. To test the two hypotheses, I compile a new dataset on states’ project-level compliance with World Bank safeguard policies on involuntary resettlement, indigenous peoples, and environmental protection. I find preliminary statistical support for the taxation hypothesis. Testing of the second hypothesis on urban projects is ongoing. The results will help scholars understand how development outcomes for at-risk groups are often dependent on aid agencies’ abilities to adopt supervisory roles for their principals: states.

Controlling Corruption: Foreign Aid in Low-Governance Environments

Can foreign aid succeed in corrupt environments? Conventional wisdom suggests that it cannot because institutional problems are intractable, aid is fungible, and most aid is political and thus inefficient. In this paper, I argue that multilateral aid can succeed in low-governance environments because corruption scandals create severe legitimacy costs that nowadays force aid agencies to take countermeasures. To curtail these legitimacy costs and thus ensure survival, multilateral aid agencies have invested in large anti-corruption infrastructures and work with recipients to include context-specific, remedial action plans at the project level when appropriate. Since these action plans address fungibility and elite capture through additional audits, procurement controls, staff oversight, and social accountability measures, aid can succeed in low governance environments. To test my argument, I individually coded all 3,437 World Bank investment projects approved from fiscal years 2001-2014 for the use of context-specific, project-level Governance and Anti-Corruption Action Plans (GAAPs). Using matching and Bayesian hierarchical models, I find that projects with GAAPs yields better project outcomes even in low-governance environments. Contrary to recent claims that the only way to control corruption is through collective action, monitoring based on the principal-agent model remains useful and relevant.

Institutional Autonomy and Donor Strategic Interest in Multilateral Foreign Aid: Rules vs. Informal Influence

Applications of principal-agent theory to the study of international organizations overwhelmingly suggest that agents only have as much autonomy as principals delegate to them. By contrast, this article argues that external shocks and agents’ contributions to underappreciated institutional design features have enabled agents to structure decision-making in line with their normative interests. In particular, principals have difficulty monitoring and controlling agents on tasks involving longer time horizons. This article analyzes the argument’s empirical relevance in Multilateral Development Bank (MDB) lending, a longer-term task/process that is of high strategic importance to powerful donor country principals. Consistent with the argument, the article shows that staff-led ratings of countries’ institutional environments at four MDBs are more important determinants of lending outcomes than measures of donor strategic interest. Moreover, the ratings are also consistently and significantly related to other non-lending outcomes in replicating many prior studies. Overall, agents’ formal rules, which are guided by their normative interests, enable multilateral aid to be less captured by powerful country principals’ informal influence than previous literature suggests. [Updated Paper]

Party Alignment and the Corruption-Reducing Effects of Lower Poverty and Violence: Evidence from Mexico

Clarity of responsibility theory suggests that once political parties align/match at higher and lower levels of government, voters can more easily discern who is responsible for corruption, and politicians react accordingly by reducing their corruption levels. By the same token, party alignment yields clientelistic resource advantages in newer democracies, and many countries have three-tiered political systems involving local, state, and national levels. In this paper, I examine the conditions under which three-tier party alignment and less clear alignment configurations condition politicians’ levels of corruption. To do so, I leverage new, objective corruption data from 11 years of Mexican municipal audits reports, and identify the causal effects of full-, partial-, non-alignment through a regression discontinuity design involving close elections. Results are currently preliminary, but I find some evidence that municipalities in which violence levels recently reduced exihibit lower levels of corruption as well. The paper will contribute to a better understanding of how poverty, violence, and political-institutional configurations interact to produce different levels of corruption in new democracies.

Measuring Corruption Using Governmental Audits: A New Framework and Dataset

For about 25 years, empirical scholarship on corruption has primarily relied on perceptions data, but the drawbacks of these measures are ample and well-known. More recently, analyses centered on Brazil have showcased the utility of randomly assigned audits as a more objective alternative to perception-based measures. However, Brazil is the only country with randomized audits and has many unique institutional features that limit the external validity of the numerous studies using the Brazil data. In this paper, I provide a new framework to assess the quality of audit data even when they are not randomly assigned. Specifically, I show that it is acceptable to use experimental or observational audit data to measure corruption when: 1) the auditing institution is legally independent from the executive branch; 2) the distribution of audits is not biased against opposition party politicians, especially following close elections; and 3) the intensity/dosage is consistent across similar types of audits. I demonstrate the utility of the framework by analyzing a massive new dataset of subnational audits from India, Mexico, Honduras, and Guatemala. The new data and framework proposed in this paper will help researchers undertake more objective analyses of governmental corruption around the world. [Draft Paper] [Presentation]

External Validity

Forthcoming at Annual Review of Political Science
(with Mike Findley and Kyosuke Kikuta)


ArcGIS - How to Make a Map


Have you ever wondered how to make a map in ArcGIS? In this tutorial, we cover all of the basics: coordinate systems, how to work with shapefiles, spatial joins, and much more. By the end of the workshop, you will be able to plot point data and color in your map based on spatial polygons values using ArcGIS’ symbology function. [Slides Here]


Applied Research: Political Science (Research Practicum, Semester 2)

Instructor, University of Texas at Austin, Department of Government (GOV 355D), 2019

This course is the second semester of a Research Practicum program that attempts to provide undergraduates with a fairly comprehensive introduction to the research process in the social sciences. Classroom instruction covers experiments, data structures, data cleaning, hypothesis testing, measurement challenges, linear regression, as well as the basics of panel data, regression discontinuity designs, difference-in-differences, synthetic controls, logistic regression, and network analysis. Training in Stata, R, LaTeX, Mendeley, and ArcGIS continues during the second semester of the course as well. At the end of the second semester, students complete their own research projects, write-up their results in a formal paper, and present their findings to the class. [Syllabus] [Spring 2020 Evaluation] [Spring 2019 Evaluation]

Applied Research Methods 1 (Research Practicum, Semester 1)

Instructor, University of Texas at Austin, Department of Government (GOV 355C), 2019

This course is the first semester of a Research Practicum program that attempts to provide undergraduates with a fairly comprehensive introduction to the research process in the social sciences. Classroom instruction covers arguments, concepts, measures, causality, and basic statistics. Given that knowledge of statistical software, text editors, reference management software, and mapping software is increasingly helpful for success in the social sciences, the course will also provide training in Stata, R, LaTeX, Mendeley, Excel, and ArcGIS. At the end of the first semester, students will hand-in their own well-developed Research Proposals in lieu of a final exam. [Syllabus] [Latest Evaluation]

Data Science for the Social World

Instructor (with Mike Findley), University of Texas at Austin, Department of Government (GOV 355M), 2020

This course provides students with a comprehensive introduction to data science for the political and economic world. By focusing on practical data skills coupled with strong social science reasoning, the course will enable students to acquire skills that will help them prepare for jobs in data science, industry, and academia. Organized around a set of substantive themes and practical tasks, each class topic is motivated by real-world problems and then backed with data science skills to solve those problems. Emphasis is placed on developing proficiency in cleaning, manipulating, wrangling, scraping, visualizing, and mapping data. Most work is conducted in the software programs R and Excel, and to a lesser extent through introductory exercises in other programs. In the process, students learn about good principles of working with data, including through version control with Github. The class takes place through asynchronous instruction, online coding practice problems, exams, and online instructor consultations. [Syllabus]