How I Got My Research Ideas

Two years ago, Zach Ward wrote a Twitter thread on how he came up with his research ideas. He did this as part of a wrap up for submitting his tenure packet. I’m also submitting my tenure packet, and I wanted to help create a tradition/ritual of researchers passing on that tacit knowledge to the next generation. I started with a previous post on how I published four papers as an undergraduate.

Today, I’m detailing where I got the ideas for the seven publications that I had after getting my PhD and before getting tenure. At the end, I’ll review what lessons I think we can extract.

To be clear, the focus is not to review the results of the papers. If you’re interested in that, I have tl;dr summaries on my research page plus links to the papers.

Also, I had Dall-E 3 design some posters for some of them. Maybe I should have included key figures or tables, but this was not about the results. It was about the inspiration.

Smartphones and Child Injuries, Journal of Public Economics Vol. 156, pp. 200-213. 2017

This paper started before I even was actively doing research. I was working in a software development job, and my coworker was very excited that the 4G network was coming to our area. That was the moment I learned that mobile networks did not launch in every location on the same date, and I took note of it because it sounded a lot like the kind of things I was hearing in my economics classes and reading in my economics books. In the back of my head, I always wondered if there was something I could look at where the rollout of a cell phone network would be an identification strategy.

That was in 2010. In 2013, I was in my second year of graduate school, and I read an article in the Wall Street Journal documenting a rise in child injuries. Starting in 2008, the long-run trend of decreasing child injuries reversed, and children were more frequently ending up in the hospital. The primary suspect was smartphones. The timing worked out, but, more importantly, the mechanism seemed plausible too. The article detailed several anecdotes of children receiving serious injuries while their were parents scrolling on their phones.

I can remember exactly where I was sitting around my kitchen table when I read this. This was it! If the 3G network spread across the country in a staggered rollout like the 4G network did, then I would have an identification strategy to investigate this question. Shortly after that article, I had two accidents on the playground with my son thanks to phone-induced distractions (one my own, one someone else’s), and I knew there was something to the story.

This was also my first dive into the audacious nature of getting data. Finding the rollout data was hard, convincing the stewards of the injury data to give me their data took a whole year, but when I finally had it all together, the results immediately popped out. It was a fun experience.

Is Uber a substitute or complement to public transit? (with Jonathan Hall and Joseph Price). Journal of Urban Economics 108: 36-50. 2018

This paper is a direct descendant of the previous. Obviously, a headline like “economist shows smartphones send children to the hospital” is going to play big. Multiple outlets covered my paper, and it caught the attention of Jonathan Hall.

Jonathan and I both went to the same undergrad university, so I suspect something in that network drew his attention to my paper. But he was an urban economist thinking of a similar question, “How did Uber affect public transit?” When he saw my article, he had a similar moment where he realized this was the way to answer his question.

Fortunately, he happened to be coming to Yale for a conference around that time, so he sent me an email. We went out to pizza and he pitched me the idea. Since we were both young, we asked our friend Joe Price to join the paper as well to help us write the best paper possible.

Funny enough, we ended up not using the 3G network since it was mostly finished by the time Uber was expanding. But the work we did was impactful: it is my most-cited article and the Journal of Urban Economics’ most-cited article in the last five years.

Intermission - Becoming an Economic Historian

The rest of these papers are deeply tied with the reason I decided to get a PhD, so I think I should give you that background before talking about the papers.

From May 2007 to May 2009, I lived in Florida as a missionary for the Church of Jesus Christ of Latter-day Saints. The Church asked me to learn how to speak Haitian Creole and work with the congregations in Orlando and the surrounding areas. So for two years, I made lots of friends working with Haitians.

About six months after I finished, in January 2010, the earthquake hit Haiti. Since I spoke fluent Haitian Creole, I searched for ways to be involved in the recovery efforts. I did end up going to Haiti in the Summer, but one of the big revelations in these moments was the (naive) thought that if I understood more about economics and Haiti’s history, I could understand what caused its poverty and help build toward a solution.

My interest in doing research in Haiti waned a bit over the next two years. It was always on my mind, but all of the research opportunities I could find were totally disconnected from this direction.

Then two things happened. First, when I was visiting Yale, I heard Naomi Lamoreaux present on economic history at Yale. I approached her after her presentation and asked if she thought the economic history of Haiti would be interesting, and she was very encouraging. Second, in my first semester at Yale, I took European Economic History from Tim Guinnane. I also told him I was interested in Haiti’s history, and he not only encouraged me then, a few weeks later (after the class was over) he stopped me in the street and said, “You have to do Haiti.”

So with that, I decided to do work in the economic history of Haiti. A lot of people thought there wouldn’t be enough data to establish a serious career with that as my specialty. But look at me now!

Small Farms, Large Transaction Costs: Haiti's Missing Sugar Journal of Economic History, 81(2): 513--548. 2021

Between undergrad and grad school, I spent a summer as a research intern for a microfinance organization in Haiti. One thing that experience taught me was that I wasn’t very interested in doing modern development work, and I was already starting to think about history thanks to the presentation I heard at Yale.

But another thing I got from that experience was the seeds of this idea. Since I was spending a summer in Haiti, I decided to grab a book on Haiti’s history. While there, I read about Haiti’s land tenure system, where an individual would cultivate a piece of land, but his entire family had the right to decide whether he could sell it or not. This stood out. “That has to mess up the economy,” I said to myself.

I wanted to examine this question, but there was a big barrier. There was no data on farms in Haiti. The whole literature agreed, it was basically impossible to find anything useful. I decided to look in archives anyway, searching for anything useful, but always having this land question in the back of my mind.

I don’t know exactly how it happened, but at one point I was browsing historical copies of the Haitian government’s gazette. I noticed at the back of each issue, there was a list of farms and some details, including where they were located and how big they were. The lists said they were printed in accordance with a 1928 law, so I went and looked it up. I found that they were plots owned by the government and leased to private farmers. Since there was about 22 years of lists available, I thought this was my best shot at getting the data I needed.

One of the most satisfying parts of this project was being in Haiti and talking to a history graduate student doing similar research. He mentioned the lists and said, “I can’t figure out what these properties are.” I was excited to tell him exactly what they were, and I was more excited that I had found something that no one knew about. I was going to be doing something novel!

The paper ended up being a detailed look at land institutions in Haiti with a small sprinkle of using this data to test some hypotheses I had about how the institutions distort plot sizes.

The Medium-Run Effects of a Foreign Election Intervention: Haiti’s Presidential Elections, 2010-2015 Contemporary Economic Policy 40(2): 369-390. 2022

This one is actually a slight departure from the history papers, but it’s still a Haiti paper.

Around 2018, the major political drama was the theory that Trump beat Clinton because of Russian interference in the US election. I loved the irony, because in Haiti’s 2010 election, a crude celebrity with zero political experience beat a former First Lady because of interference from the international community. And the main face of that scandal was none other than Secretary of State Hillary Clinton.

Haiti’s 2016 election was also riddled with scandals. One of the biggest concerns was that turnout was allegedly much lower than even the 2010 election, where thousands couldn’t vote because of the difficulties after the earthquake.

So as I was laughing to myself about the irony of Clinton complaining about interference, I wondered if there was a connection between the interference and the low turnout. After all, voting in Haiti is costly and risky. If I thought that some external force was just going to throw away my vote, I might not be willing to vote at all.

Fortunately, I had already collected data on Haiti’s elections, so I was able to quickly test the hypothesis.

State Capacity, Property Rights, and External Revenues: Haiti, 1932–1949 Journal of Economic History, 83(3): 709-746, 2023.

While collecting the land data from the government gazette, I noticed something interesting. The lists reported when the property was requested and when it was granted. I noticed that there was a flood of requests in 1938 (addressed in the next paper) and another flood of approvals in 1942. I was trying to figure out the flood in 1942, and it seemed awfully convenient that it was around when the US mobilized for World War II.

Why would WWII cause Haiti to change how it handles land rentals?

That led me to find a ton of interesting stuff. The US diverted trade, this caused a customs collapse in Haiti, and to shore up its revenue, the government of Haiti reformed its income tax.

This all was right in line with some popular models of state capacity, so I continued to explore the model’s predictions.

Two things to mention with this paper. First, it felt like my first truly unique contribution to the economic history of Haiti. There was so much I found that I hadn’t seen discussed anywhere else, so it was very satisfying. Second, when I started on this project, I presented some of the basic ideas to students and faculty at Yale. When I pointed out that I was really interested in the processing delays, one of the faculty yelled at me. “That is the least interesting thing in the data! That is unique to Haiti and has no external relevance.” Turns out it was interesting!

The Forces of Path Dependence: Haiti's Refugee Camps, 1937–2009. Explorations in Economic History, 89, 2023.

This paper started with a result that was present since I first started collecting the land rental data. I noticed that requests spiked in 1938, right after the Dominican Republic massacred Haitians. “Was there a refugee crisis after the massacre?” I didn’t know. So I started reading more about it and found that there was indeed a bunch of refugees who came across the border, and many even settled in refugee camps.

I tried to incorporate this into the previous papers. A demand shock that could help us look at how institutions affect farms! A demand shock that overwhelmed the state bureaucracy! While I mentioned it in both papers, it was never the main result.

Then I decided to ask questions about persistence. I knew that the refugee camps still existed today, which was strange because these were all ethnic Haitians who could have integrated into the broader economy. Maybe this land rental program could help explain why they stuck around.

"A Whirligig of Revolutionary Presidents": State Capacity, Political Stability, and Business in Haiti, 1905-1927 European Review of Economic History, accepted

While I was looking for the land data, I came across lists of business licenses held by foreigners. The lists reported the foreigner’s name, country of origin, and the industry the business was in. Anytime I saw anything that looked like interesting data, I saved it, and for years I held onto it.

My land rental data focused on the late 1920s through 1940s, but for some reason, years later, I was looking at earlier periods. I noticed that there were more lists of business licenses. I could see foreign business activity before the US occupied Haiti and see how it changed in 1915. That’s about it. This one wasn’t as much about inspiration as it was just the obvious question to explore with that data.

Lessons

What broader patterns stuck out to me as I wrote that?

1. Preparation meets opportunity. Many of these ideas were just opportunities that happened to be the perfect match for something I had thought of years before. The 3G expansion idea was something on my mind before I was even involved in research. The Uber idea came from someone else being prepared when they saw my paper. The Haiti stuff all came from reading books years earlier.

2. Making the most out of a data set. I collected 20 years worth of land rental data, and it provided the foundation for three of my publications (and there’s a fourth paper using the same data in the works). This is a common theme among researchers. When you get an interesting data set, try to extract as much insight from it as possible. You’ve already overcome the fixed costs.

3. Explore trends in the data. While collecting the data, I looked at aggregate trends just to get a sense of what was happening. When I saw the trends, it inspired research questions. And that’s natural because no one had ever looked at this data. Some people get a lot of mileage out of exploring unanswered questions, but you have to keep in mind that there are even more unasked questions. If you’re looking at stuff no one has ever been able to consider, just looking at patterns will inspire tons of questions.

4. Grad school ideas. When I was in my last year at Yale, a seminar speaker told me, “You probably already have everything you need to get tenure. If you just finish out the projects you started here, you’ll probably find you get enough publications.” She was totally right. Every paper can be traced back to a pivotal moment in graduate school. That ended up being super helpful, because Haiti descended into political chaos just as I was finishing grad school, and it would have been difficult, if not impossible, to go back for more data.