Edmund Malesky, PhD is conducting research at the intersection of Comparative and International Political Economy. Professor Malesky is particularly interested in authoritarian political institutions and their consequences, the political influence of foreign direct investment and multinational corporations, as well as political institutions, private business development, and formalization. To increase the impact of his research around the world, Professor Malesky has been publicly sharing his research data and encouraging scholars in Southeast Asian countries and beyond to conduct replication studies and use the data in their work. Professor Malesky is a proponent of preregistration and registered reports, as solutions to address questionable research practices that threaten the integrity of research and public trust in science. His impactful research on measuring business environments, addressing corruption or improving business compliance with regulations has been widely used by the World Bank and USAID and has inspired governmental reforms that improved people’s lives in Southeast Asia. Under Malesky’s research and leadership, the Duke Center for International Development is amplifying its voice in supporting good governance and improve policy in developing countries through educating politicians and government officials, providing best policy practices and engagement. To date, DCID trained more than 6,000 individuals in over 130 countries.
Emilia Chiscop-Head: How have you become passionate about political science?
Edmund Malesky: I do what is called Political Economy, so I sit between political science and economics. As an undergraduate student at the Georgetown University’s Walsh School of Foreign Service, I majored in Economics. I realized how many of the economic outcomes that I was interested in - economic transition from a centrally planned systems to the private sector, redistribution, and public service delivery - required deep thinking about how the political process and institutions worked. Therefore, I became very interested in exactly those types of decisions: how political institutions mediate discussions of different actors, how the entrenched power of political actors matter for these types of decisions, how firms interact with the policy process through lobbying and through information, but also illegally through corruption; and how this interaction shapes the business environment in which they operate.
You conducted a lot of research on Vietnam’s political economy. Why Vietnam?
Edmund Malesky: After my undergraduate studies, when I was working for a year at the Department of Justice, my undergraduate Alma Mater, Georgetown University, nominated me for the Luce Fellowship, a fellowship with a great deal of prestige like the Rhodes or Marshall Scholars, but with a different set up: this scholarship was founded in 1974 by Henry Luce with the goal to immerse young leaders in Asian countries to help develop mutual understanding and strengthen their relationships with the US.
The work experience that I had during that time, working at the National Economics University in Hanoi for the first business school in Vietnam, shaped a lot of who I am as a scholar. It also helped me learn Vietnamese, and I have a lot of friends in Vietnam. I conducted significant research on how Vietnamese political system works, what a developing country and emerging market development look like in practice; I studied how authoritarian institutions and single-party regimes work. I am trying to do work that speaks beyond Vietnam. My language abilities and my connections allow me to do things in Vietnam that are more difficult to do in other places.
And after Vietnam you came to Duke for an MA and a PhD. What mentors shaped your career?
Edmund Malesky: During my undergraduate studies I was very interested in a scholar named Robert Keohane, whose wife, Nannerl Keohane, was the President of Duke University at that time. As Bob had just moved down from Harvard to Duke, and I could have not imagined to have another mentor, I came to Duke to work with him. I was very interested at that time in international political economy and international organizations. I would like to mention three other fantastic scholars that influenced me. Herbert Kitschelt was my dissertation advisor and did a lot of work on creating democratic institutions in the former Soviet Union and Eastern Europe; David Soskice, a very famous political economist who studied the differences between economic systems, and Karen Remmer, who worked at that time in Latin America, in local political institutions and decentralization, were also on my dissertation committee. I learned from them many theoretical and analytical approaches that I use today in my scholarly work. My time at University of San Diego, from 2005 to 2012, as an Assistant and Associate Professor, also shaped me as a scholar. I was exposed there to a large number of development economists, like Craig McIntosh and Gordon Hanson, who introduced me to causal inference tools, and randomized experiments.
“I try to help my graduate students maintain the passion that brought them to graduate school in the first place, while giving them the theoretical guidance and the skills necessary to do that.”
What is your mentoring philosophy?
Edmund Malesky: I aspire to be like my dissertation advisors. Advisors come in three shapes and sizes. One model is the advisor who has a research program to which trainees contribute, and their work is meant to push forward that research agenda. A second type of advisor doesn't try to fit the mentee into their research program, but does try to help shape their research question. This mentoring style is very common, and it is easy to see the imprint of that dissertation advisor in the final product. My advisors were very different, and I try to aspire to that model: they accepted my research question, and we both became interested in it. For example, Herbert Kitschelt would often say that a dissertation is good when a student falls in love with their research question, specifically their dependent variable, and are passionate to explain it.
My mentors helped me answer my question better, told me where I needed to look, and helped me with the methodological aspects of my dissertation. Therefore, as a mentor, I work inside out to make a question better rather than telling you that you should ask a different question. I think that it's a lot harder to advise that way. I try to help my graduate students maintain the passion that brought them to graduate school in the first place, while giving them the theoretical guidance and the skills necessary to do that.
How do you come up with the research idea and how do you establish your research priorities?
Edmund Malesky: First, I start with big questions that have a lot of theoretical support, but have not been tested well at the micro level. For example, there's a large body of theoretical work on how economic reforms take place. But economic reforms involve politicians, bureaucrats, businesses, and voters making micro level decisions. I like to see if those micro level decisions that are predicted by theory actually hold up in practices. I try to choose questions that are very precise, and that I can answer with an artillery of evidence; questions that are aiming to shed light on much larger theoretical projects.
Second, I survey my research agenda and try to move the research questions that I helped answer forward a little bit at a time. Many of my questions are inspired by my living and working experience in Vietnam and by the problems that people with whom I interacted are trying to solve. When I see a policy problem I start thinking backwards: what does the political economy literature have to say about this question? How can I bring the theoretical literature to bear on that policy question? Then, I ask myself if I can come up with a research design that helps answer that question. This is how I engage with all my research projects.
“By being transparent, people will take my non-resolved policy question further and they will not repeat the same failure.”
I found on the Harvard Dataverse Repository many of your papers accompanied by replication data files. What is your view on the importance of replication studies? What can institutions do to support data sharing and replication?
Edmund Malesky: I have very strong and opinionated views on this topic. I post all of my data and all of my replication files. For example, when I do a statistical project, I have usually written some computer code, sometimes STATA Do-Files or R scripts that take the data and then produces the results that you would see in a paper. I always want to make sure that anybody can look at my files and perfectly replicate everything I have done. I want people to trust the evidence that I've provided, because I’m trying to answer big, important policy questions. I also want people to build on it and to challenge me, by attempting to replicate my work. If you were to look at the downloads of my data sets you will see that a lot of my data is downloaded by scholars from Southeast Asia, including Vietnam, Cambodia, Indonesia, who use the data that I'm providing in their own scholarly work.
I also do something that is unusual for social scientists but it is becoming more common for physical scientists: I publish my hypotheses that I want to test on the OSF Registries before I collect and analyze the data. In my pre-analysis plans I include the hypothesis that I want to test, what I expect to find, what techniques I plan to use to measure it, what econometric specifications and robustness tests I intend to run ,and what subgroups I intend to examine. I am doing this for two reasons. First, because I want people to believe the findings that I'm producing. We live in a world with tremendous proliferation of data: if I want to, I can get the result I want. I can write, and “massage” the data in a way that justifies my expected results, I can always drop observations for reasons that I can justify in the paper. Second, because I have one foot in the policy space and I am trying to answer policy questions that will make people's lives better, I want people to be able to know what not only what interventions work, but what doesn’t work, so they don’t waste more time and money on dead-end ideas. I want them to say, “I thought this idea would be a good one but it didn’t work. What can we do better?” By being transparent, people will take my non-resolved policy question further, and they will not repeat the same failure.
It would be great if universities could make these data sharing processes a lot easier. It takes a lot of time do a pre-analysis plan or prepare post replication files. It is an extra layer of difficulty on top of complex research and it takes a lot of times. It’s especially hard in the social sciences because we do everything ourselves: I invent the theory, I collect the data, I run the data. I am a little one-man operation. Yes, sometimes I have co-authors and wonderful grad student collaborators, but I do not have a lab, I do not have biostatistics experts who are running my analyses for me. I'm doing it all! Anything that universities can do to help us with data sharing, including providing funding for it, would be very important.
What do you think are the most important ethical challenges in your work?
Edmund Malesky: Even if I do everything I can to make my process as transparent as possible, there are going to be times when I have to do additional analyses for a paper that is not pre-registered, and, in this world of data proliferation, I can find sixteen different ways to measure an outcome variable. So I need to stay disciplined and systematic about making choices when I know the answer that I would like to achieve. I've become very interested in machine learning techniques recently, because I've started to realize that they're very helpful in terms of disciplining me; tying my hands to the sets of outcome variables selected by algorithms.
When you work in countries like Vietnam, China, Cambodia which are non-democratic regimes – because you are interested in solving people's problems - the policy answers that you find also lead to greater legitimacy of the Vietnamese Government. Some people have confronted me directly on this topic and told me that I should not work in these countries because we want these regimes to fail. While we need to reflect deeply of these ethical questions, First of all, there are very big differences between the types of authoritarian regimes that we encounter: some regimes that we consider to be non-democratic, are actually quite responsive to their citizens, and have done a lot to improve citizens’ lives. Helping those types of regimes, I don't think is a bad thing, especially when they demonstrate an interest in better governance, transparency, responsiveness, and reducing corruption. Second, but even more importantly, our work makes peoples’ lives better, it improves their incomes, their access to social services. By helping their governments to be more transparent, we help people make better decisions for their families and for their businesses.
But there are countries where I absolutely won't work because I don't think that they're responsive governments. There are also places that we call democracies where I would feel uncomfortable working because I don't trust the actors that are in power. When I conduct research, I collaborate with the local partners, debrief my results with them, engage them at the policy level and not just at the academic level.
And this speaks to the social impact of your work…
Edmund Malesky: Some of the tools that I've demonstrated have been adopted by countries. Occasionally I see my work cited in World Bank’s or USAID’s calls for proposals regarding measuring business environment, corruption or improving business compliance with regulations. Seeing that my work contributes to policy reforms and international development assistance projects, that it is used in a very real way to improve peoples’ lives makes me happier than anything else. And much happier than the citations count.
What are some of the questionable research practices in your field and how can they be addressed?
Edmund Malesky: It comes to mind the reproducibility crisis in political science with famous pieces that could not be replicated, with core findings that didn't hold up to sensitivity tests. P-hacking and the abundance of positive findings are very dangerous for the credibility of research. I am a big proponent of pre-registration and registered reports. I actually co-authored a paper where we advocated for registered reports and did a pilot with registered reports in a really important journal in political science called Comparative Political Studies. With registered reports, the journal reviews your article without the results, determines whether the research design was sound, and accepts the research paper for publication before seeing the results. This relieves a lot of the anxiety that leads scholars to engage in p-hacking. In fact, I was tasked with writing the procedures for registered reports for a team proposal to edit our flagship journal, the American Political Science Review. Unfortunately, our proposal was not selected. However, recently another top journal, the Journal of Politics, decided to implement registered reports and many of our ideas for how to administer it in their own procedures. They will start doing results free review this year.
“People are more likely to comply with rules when they had some say in designing them.”
Could you tell me about your current research project on regulatory compliance in Thailand?
Edmund Malesky: I was inspired by the Rana Plaza disaster, which happened in Bangladesh in 2013, when a garment factory collapsed and 1,134 workers died. The company had taken shortcuts with safety violations as the owners were too powerful politically and paid bribes. The newspapers have plenty of similar incidents.
So, my co-author and I started asking: how can we get businesses to comply with regulations in settings where countries either do not have this strong enforcement capacity because their bureaucratic capacity is too weak, or the system is too corrupt? It is extremely difficult to enforce safety regulation in many states. For example, Vietnam has six hundred thousand registered formal private companies and over three million informal enterprises while the Labor Safety Office in the Ministry of Labor has less than 10 employees. The answer is self-compliance, but how can you get firms to self-comply in Vietnam, at a time when most of them believed that regulations were adopted only for bureaucrats to be able to extract bribes? In psychology, sociology and literature there is a theory called procedural justice, sometimes called the deliberative democracy theory in political science: it says that people are more likely to comply with rules when they had some say in designing them. We designed a randomized experiment with the Vietnam’s Ministry of Labor’s Safety Bureau and the Vietnamese Chamber of Commerce and Industry. We had one group of firms that just learned about a regulation on chemical labor safety, which was very important at the time, because there were a lot of firms that were storing hazardous chemicals improperly; a second group of firms, which learned about the regulation, but were also asked to provide comments on the draft law; and a placebo group that learned about the services offered by the Vietnamese Chamber of Commerce. One year later, when the draft moved closer to being a final regulation, we audited the factories and saw that factories, which provided comments, were 10% more likely to allow auditing of their safety procedures and 5% more likely to be in compliance with the final regulation. This was a big success. Then we asked - could this idea travel beyond Vietnam? What about externalized non-compliance, such as environmental violations where the company doesn't pay anything if it does not comply.
The opportunity came up in Thailand, a country which is moving all its ‘call and comment notice’ (when draft laws are posted for public comment) online, under the leadership of the Digital Governance Agency. Thailand is very interested in helping us test this theory and two other theories: one regarding how to improve the quality of public participation as well as the responsiveness theory, both inspired by our previous research in Vietnam.
In Vietnam, firms would provide comments, but the comments weren't useful to the drafting committee, because they were off the mark or they were not provided in a legal way. As part of the experiment in Thailand, a subset of firms are working with an expert to redraft their comments in a way that makes them more useful.
With the responsiveness theory we try to find out how receiving a response from the drafting committee as opposed to receiving no draft at all influence compliance rates. As I said earlier, my research is composed of small steps: I answer a policy question and then I move forward with additional incremental advances.
Is improved management a key strategy that should be used more widely to tackle corruption?
Edmund Malesky: This is a great question that I am also interested in. In Vietnam, I’ve been working since 2005 on one of the longest continuously funded project in the history of USAID, to produce an index of the business environment. It's called the Provincial Competitiveness Index, which I’ve been able to replicate in other places as diverse as Malaysia, Myanmar, and El Salvador. There is one pattern that shows up always in the data: that most productive firms pay the least amount of money in bribes.
Some firms use bribes to make up for competitive disadvantage: for example, someone is bidding for a procurement project, they don't have the lowest cost or best proposal, but if they pay 10% percent of the contract they might be more likely to get it. The pattern is that unproductive firms take shortcuts on regulation, don't follow environmental or safety regulations, because they know they can always pay a bribe to get out of it.
The question is how to increase productivity in order to lower corruption? I hope that business schools are right that better management leads to higher productivity. In Vietnam, we are working with 1,200 Vietnamese restaurants, which were drastically hit by the COVID 19 pandemic, in a randomized experiment. In one group, firms take a management class on how to improve their business, and we will test whether this will lead to greater productivity and reduce bribery in complying with regulations. In a second group, firms are exposed to a set of courses that don't have a direct impact on productivity. We do not have the answer yet to this policy research question.
Saving lives as a result of educating future leaders around the world
What is your long term vision for the Duke Center for International Development?
Edmund Malesky: We educate future leaders, mid-career professionals from around the world that come here to upgrade their skills, so they can go back and be successful technocrats or successful politicians. We are also doing rigorous academic research that's moving forward the policy envelope and helping inform policymakers. And we have our engagement portfolio through providing executive training to help inform development policy and events that communicate to various stakeholders around the world about our research. My vision for DCID is to construct a robust research portfolio which is producing cutting edge academic findings, and then delivers those findings directly into the classroom to future governmental officials and bureaucrats, but also to the private sector, to foster new engagement, so that our research findings are directly informing policymakers.
What is one concrete impactful success story that shows the DCID impact around the world that you are most proud of?
Edmund Malesky: My proudest moment as the director of DCID came when I learned how one of our MIDP alumni, Anil Bhardwaj helped his district of Faridabad addressed a shortage of oxygen tanks at the height of the first major wave of COVID infections in India. Anil developed an easy-to-use online system that allowed residents to request oxygen tank refills, receiving digital tokens and instructions on how to use their tanks. In 9 days, they were able to quickly service over 5,500 requests, saving hundreds, if not thousands of lives. His ideas were immediately replicated throughout India, compounding his influence. Anil’s initiative combined the creativity, analysis, precision, and public spiritedness that we pride ourselves on here at DCID. He is a perfect example of how we hope to make people’s lives better one good idea at a time.