As part of our interview series, we ask renowned experts in the field about the future of research in development economics, and for their advice to young researchers. For this interview, we got the opportunity to talk to Prof. Douglas Gollin.
Douglas Gollin is the Jason P. and Chloe Epstein Professor of Economics at Tufts University. His research focuses on economic development and growth, with an emphasis on the structural transformations that occur during the growth process.
ETRM: We’ll start with a general question we ask all our guests: What first inspired your interest in development economics?
Doug: I got into economics through an unconventional path. I started with a degree in history and literature of the Renaissance and Reformation, then worked as a newspaper reporter in Omaha covering the farm crisis of the mid-1980s. Writing about agriculture led me to interview agricultural economists, which sparked my interest in the field. I went to Yale for an interdisciplinary master’s program where I took my first economics courses, focusing on agricultural development. After working with an NGO in agriculture and development, I realized I needed a PhD to work effectively in policy.
ETRM: What you see as the most exciting frontiers or emerging challenges in your field of research?
Doug: I feel that for a long time, both academic and policy discourse around growth have been constrained by our models. The one-sector growth model has been a powerful workhorse for how we think about growth, but within that framework, there’s limited scope—countries grow by increasing the Solow residual, which is both unhelpful and offers few policy levers.
Over the last 15-20 years, I think we’ve started unpacking structural transformation more broadly, looking at how countries change across multiple dimensions as they grow. This points us to more questions and potential policy levers. For instance, we’re now talking about gender norms and barriers to women’s work—20 years ago, you wouldn’t have seen that as macro-relevant. We increasingly understand that growth embeds different transformations in what activity takes place in the home versus market or formal versus informal sectors. We see more margins where policy can have impact.
At the same time, there’s been enormous micro development work giving us high-resolution clarity on causal impacts of interventions. I think there’s a great agenda in integrating what we learn from those tools with our macro models. So, the exciting area of work I see is integrating micro and macro methods for studying development and growth.
There’s a role for descriptive work, and I believe there’s also a role for qualitative research too. That’s a harder sell within the profession, but I think if we really want to understand what’s happening in the world, we need to be open to different methodological approaches and what they can tell us.
ETRM: How do you think a new perspective on the role of agriculture can inform other fields in development economics? For example, in your talks, you emphasize the difference between food and agriculture, a difference that can inform research on nutrition.
Doug: I think we need to do a better job in our models and our thinking in distinguishing between food and agriculture. We have a lot of models in which the agricultural sector produces a good called “agriculture” and that identical good enters consumer preferences. The reality is that a lot of what’s produced on farms isn’t consumed directly by households, even in low-income countries. And a lot of what’s consumed by households is only pretty tenuously connected to farms.
For instance, farms produce a lot of goods that are not really consumed directly as food for people – such as cotton or tobacco. Farms also produce a lot of goods that might in principle be food but in practice are fed to animals. This would include a lot of maize (corn) and soy.
At the other extreme, lots of foods – think of diet soft drinks, in the extreme case, or some of my favorite snack foods (e.g., cheese puffs) that are only very remotely linked to farm products.
We don’t yet have many good models of the whole agri-food value chain, that include processing, packaging, distribution, and so forth. A few economists – people like Tom Reardon – have been writing about this, but it hasn’t really made it fully into our models.
For development economics, I think that a good understanding of the industry structure of agriculture might help us to think more realistically about what we can achieve from, say, an increase in staple food productivity at the farm level. The structural change mechanisms will be much weaker in a world where the farm share of food consumption is only 15 percent, as opposed to our current models where the farm share is typically assumed to be 100 percent.
ETRM: Given your extensive experience working across various African countries, how have you seen the continent’s economic landscape evolve- particularly in light of its historical reliance on agriculture? In your view, what are the key steps needed to ensure a successful structural transformation going forward?
Doug: I’ve been thinking a lot recently about the disconnect between urban growth and agricultural development. A lot of African countries have seen substantial development – in the broadest sense of that word – in urban areas. When I first went to Accra, in Ghana, in 1987 or 1988, it was a fairly sleepy place. In the following 35 years, there has been extraordinary growth, not only in the size of the city, but also in the income and wealth of urban households. What is striking is that this growth has not fully fed back into rural living standards. Many rural areas look very similar to how they did 35 years ago, even while the cities are unrecognizable.
I suspect that part of the reason for the weak backward linkages from urban growth to rural development is that Accra is on the coast, and so urban consumers have access to imported food and various types of food-like items. This limits their connection to interior locations.
In my view, a successful structural transformation will need some significant investments in infrastructure – and especially in building the connectivity from urban areas to rural locations.
ETRM: Last question – In your recent work you used detailed information on location or irrigation, household, firm and villages. Could you tell us about how you accessed this information and what to consider when working with such detailed data?
First, I always ask how datasets are generated. Are they as granular as they appear? Weather data, for example, often involves heavy modeling—there aren’t many actual rain gauges, especially in places like the African Sahel, so the data includes significant interpolation and extrapolation based on satellite images and climate models. This introduces patterns of spatial correlation you need to be aware of.
Similarly, FAO GAEZ agricultural data on potential and actual yields has limited ground truthing. That doesn’t make the data bad, but I think you need to consider carefully whether you’re using it in a way that’s sensitive to how it was generated.
Second, consider what’s missing from your data and who might be missing. Spatial data can have systematic gaps or biases in coverage.
Third, what I’ve found is that agriculture is inherently spatial—it uses land in a non-trivial way and produces bulky, physical outputs. This means mobility frictions matter enormously. You need to account for the costs of moving inputs and outputs across space, and how farmers’ terms of trade worsen with distance from markets.
My key message is: be a critical, thoughtful user of data. Learn how your data is produced and ask how the production process might bias your analysis. All data is imperfect, but we need to acknowledge and work with those limitations.
ETRM: Thank you so much for your time and insights. We really appreciate you joining us today!
Doug: My pleasure – Thank you!
