Development in Practice: Practical Applications of Economic Complexity in Ethiopia

By: Aziz Ben Baz (MPA/ID ’22)

Over the summer, I had the chance to work with the Growth Lab (GL) at Harvard’s Center for International Development (CID) on the Advancing Economic Diversification in Ethiopia project. One of the main reasons I chose to pursue the MPA/ID program at HKS is that I was a counterpart for the GL in my previous work. As such, it was a primary focus of mine to get immersed in the GL work during my time at HKS to get a higher exposure to the pioneering work done on International Development.

The Growth Lab started collaborating with multiple stakeholders in the Ethiopian Government in 2019. The culmination of GL’s initial findings on economic complexity in Ethiopia were put eloquently in the “Pathways for Productive Diversification in Ethiopia” paper which cites that one of the main binding constraints for economic development in Ethiopia is the access to foreign exchange. The two main frameworks used for such work were Growth Diagnostics and Economic Complexity. In this report, not only do policymakers get an informed analysis on which areas of economic policies to focus on, but they get a prioritized list of products that Ethiopia could target in order to unlock the next phase of economic growth. For Ethiopia, it is crucial to increase and diversify the export basket in order to afford imports required for further development, which would address the shortage of foreign exchange highlighted earlier.

My primary focus over the summer was to help the Ethiopian Investment Commission (EIC) take the Economic Complexity framework findings a step further and find practical applications. While the framework is useful as a first step to shortlist possible products a country could target, policymakers still need to do more due diligence on products viability and how to put such insights into action. For example, we conducted two deep dives for two different products, one related to Agriculture and the other to Machinery. The intention was to see how such products would require a different set of skills and possibly different policy levers to unlock and push forward. It is relatively easier (i.e., shorter distance in the Product Space) for Ethiopia to start targeting agro-processing due to its reliance on the Agriculture sector. Indeed, some of the industrial parks in Ethiopia are focused on fostering food and beverage processing by attracting more investors into the sector. But the trade-off is that such products are not particularly high in terms of economic complexity and would still face challenges to sustain economic growth. It is important to note that the intention is not to minimize the role of the Agriculture sector and the further development of the Value Chain. But rather how to “balance” the portfolio of targeted products by also selecting more complex ones (i.e., Machinery) and develop a strategy around that.

To that end, we tailored a specific approach to study the viability of the more complex products in the Economic Complexity framework, specifically by looking into Machinery. This was done by looking into the Economic Complexity metrics (see Growth Lab’s Atlas of Economic Complexity) and the dynamics of supply and demand (with emphasis on exports/imports) domestically, regionally, and globally. We also looked at the size and trends of investment flows to such sectors, major players, as well as the comparative advantage of Ethiopia and the tariffs dynamics versus its regional peers. We then identified specific tangible sub-products that are potentially viable and have a relatively reasonable existing footprint in terms of imports and exports, which could help in addressing the shortage of foreign exchange in Ethiopia.

We also leveraged a few more concepts into this work to compliment the Economic Complexity framework. For example, we’ve integrated concepts from the “Productive Ecosystems and the arrow of development” article by O’Clery, Yildrim, and Hausmann (2021). We also looked into the environmental aspects of the prioritized products by adding insights from the Green Transition Navigator. Lastly, we employed a Gravity Model that was developed by the GL to determine possible destinations for Ethiopia’s exports.

As for the next steps on this effort, there are still key areas that remain to be explored, especially ones related to understanding who are the current Ethiopian exporters and importers of the targeted sub-products and the challenges they face. Such essential missing information would inform the available policy options and test the viability of the selected sub-products in Ethiopia. It is important to acknowledge that applied economic research has to cater to realities on the ground. As such, the input from our counterparts in Ethiopia is very crucial. Many of the limitations and nuances can only be uncovered by those who live and breathe these problems every day.

It was refreshing to work with GL on such challenging development topics after an intense academic year at the MPA/ID program. The work done at the GL provides a sweet spot between academic rigor and applied policy research. It was a truly great learning experience. I’m very grateful for the mentorship of all the researchers, professor Ricardo Hausmann, my fellow classmates interning at the GL, and all of the GL staff who helped along the way.

Financial Liberalization and Debt Sustainability in Ethiopia

By: Keita Takemura (MPA/ID ’22)

As an intern with the Growth Lab, I was working with the National Bank of Ethiopia on macroeconomic issues. Ethiopia has achieved rapid economic growth from the early 2000s, alongside large capital account and fiscal deficits (when the deficit is broadly defined). In spite of this situation where depreciation would be necessary to clear excess demand for FX, the exchange rate has not been devalued in a manner consistent with the level of deficit monetization. The exchange rate control, together with restrictions on financial account transaction, have led to a persistent excess demand for FX in relation to supply.

This policy stance appears to have caused several serious problems. Since the controlled FX market cannot meet the demand for FX, the amount of FX has been scarce in Ethiopia, leading to a constraint on imports. Shortage of FX also has led to the emergence of black market for foreign exchange and a stubbornly high black market premium. The real exchange rate has been overvalued due to the insufficient depreciation, which has resulted in declining global competitiveness.

To attack these issues, Ethiopia may want to consider unification of its de facto dual FX regime by allowing larger depreciation, while loosening control on FX market. Opening the financial account, however, would raise several concerns. One of them is a possible increase in the debt service of the Ethiopian government. The interest rate on public debt in Ethiopia has been set low compared with the inflation rate, which has helped contain the increase in public debt service (a form of financial repression). After Ethiopia loosens FX control, however, it would be expected that the arbitrage between domestic and foreign markets should work, and the real interest rate on public debt would rise. This rise could be a serious issue in Ethiopia, where there is a concern toward debt sustainability.

Therefore, my internship theme was analyzing the effect of FX control loosening on debt service in Ethiopia. I conduct my research from two approaches; first, I explore the development of the real interest rate and the debt/GDP ratio after such liberalization in other countries. Second, I use Ethiopia’s data and develop a simulation of the debt/GDP ratio after FX liberalization in Ethiopia.

In cross-country analysis, I found that the real interest rate has increased in many developing countries, while the debt/GDP ratio has not necessarily increased. The key factors to contain the debt/GDP ratio are a positive primary balance, a real growth rate higher than the real interest rate, and a high inflation rate.

The simulation also confirms that the debt/GDP ratio would not increase in Ethiopia due to FX liberalization even with positive real interest rate, under the assumption of steady economic growth (5.0%) and inflation and depreciation rate in line with the level in the recent three years (15.3%). It is also true, however, that in the risk scenarios of (1) low growth rate and (2) high depreciation rate relative to the inflation rate, the ratio would destabilize. It should also be noted that high inflation can help contain the ratio under risk scenarios until the debt with fixed nominal interest rate is rolled over, though this containment would last only for five years.

Through the internship, I was able to brush up my analytical skill as an economist. Especially, my breakthrough was learning (1) how to conduct research on countries where there are little data available, (2) how to apply international macroeconomics to analysis in developing countries, and (3) how to communicate with policymakers.

Data
There are little data available in Ethiopia. For example, GDP is available only on an annual basis, and there are very limited macroeconomic data around the labor market. Therefore, we should depend on descriptive analysis or simple regressions rather than more advanced methods. It does not, of course, mean that our research was easy. We have to have clear logic behind our statement and develop our arguments with simple and clear-cut descriptive graphs. This process helped me refine my critical thinking skills to provide coherent results.

International Macroeconomics
Before HKS, my research focus was on Japan, where the capital account is always positive and the foreign reserve is abundant. Therefore, even though I studied international economics in the first year at MPA/ID program, I did not fully understand whether the international macroeconomic models were useful to the practical research. However, when I analyzed capital control and exchange rate issues in Ethiopia, macroeconomic models such as the Mundell Fleming model, PPP, and UIP simplified these complex issues intertwined with many other variables and helped organize my thoughts.

Communication
Since COVID-19 forced this internship to be conducted remotely and counterparts at the NBE and the government broadly has been involved in a major reform push while also preparing for national elections these past months, I did not have many chances to engage in iterative discussions on this research with government officials. However, when I was discussing with supervisors at the Growth Lab, I was always asked, “does your interest really matter to what the NBE is facing?” or “does your suggestion really help the policy management of the NBE?” These comments helped me to start thinking about the needs of the policymakers while I was conducting research. These experiences should help my future career as a researcher working in the public policy area.

Growth of Caquetá, Guaviare, and Putumayo – Growth Lab Summer Internship

By: Alejandro Rueda-Sanz (MPA/ID ’22)

As the only road-accessible entry points to the Colombian Amazon, the departments of Caquetá, Guaviare, and Putumayo are facing stark challenges for their future growth. These departments have been at the center stage of the armed conflict and concentrate today most of the deforestation in the Colombian Amazon. Growth, however, has lagged in all three departments relative to its Colombian peers. This puzzle set the stage for my MPA/ID summer internship  at the Growth Lab at Harvard’s Center for International Development.

I began working on this engagement in the earliest stages of this project as a sequel of the Growth Lab’s Promoting Sustainable Economic Growth and Structural Transformation in the Amazon Region of Loreto, Peru project  – both funded by the Gordon and Betty Moore Foundation. I was tasked to use my background knowledge as a Colombian who had worked in policy, paired with the skills and frameworks from my first year of the MPA/ID to structure the prior stylized facts and assumptions for a Growth Diagnostic, an Economic Complexity Analysis, and to glean the causes of deforestation.

The Growth Lab provided the opportunity to achieve one of my objectives when joining the Harvard Kennedy School: shifting my career to strategic development policy. To work with the Growth Diagnostic and Economic Complexity frameworks, I received mentorship from my manager Tim Cheston (an MPA/ID alumn), on the practicalities of Growth Diagnostics, Economic Complexity, and the soft communication skills to present them.

Below, I discuss some initial findings and highlights from the internship.

Setting the priors – linking my experience to new frameworks and ideas
In the first weeks, I used the Colombian Atlas of Economic Complexity, which I used before joining HKS, and administrative data to draw the growth trajectory of the three departments. We learned these departments had different trajectories:

According to our estimates, it would take between 20 and 60 years for these departments to catch up with Colombia’s 2020 income per capita.

Satellite data showed us the extent of deforestation in the three departments. Between 2002 – 2018, Caquetá, Guaviare, and Putumayo had depleted 7.8%, 6.2%, and 11.9% of their forest cover, respectively. Moreover, through recent analyses and speaking with local experts, we learned that this trend had accelerated since the signature of the 2016 Peace Agreement with the FARC through cattle ranching and land price speculation. This trend, however, seemed puzzling given the slow growth in the region.

We presented the findings on the growth trajectory to the Growth Lab’s leadership, the project’s new team and discussed some stylized facts I had developed with my manager. Then, leveraging the Growth Lab’s expertise from the Loreto Project in Peru, we kickstarted discussions to build a growth question for the project and the diagnostic tree. At these discussions, I learned how researchers at the Growth Lab approach diagnostic frameworks, interpret information, and how they thought through adapting the framework to the circumstances of ecologically sensitive and remote geographies.

Being a part of the Lab: practice and learning about new trends and ideas
Setting the ground for the project was also a unique opportunity to learn from how the Lab approaches different countries. I learned what different teams did in other national and subnational contexts at discussions on ground-breaking development issues. For example, at the discussion on remoteness, led by Professor Hausmann, I learned how the Lab defines and approached this condition in diverse contexts, including Albania, Namibia, Loreto (Peru), and Ethiopia. I also reflected on how these analyses could enrich our discussion on the Colombian Amazon, given its unique circumstances where conflict and property rights play a critical role.

Working at the Growth Lab during the COVID-19 pandemic was further challenging and rewarding. Traveling to these sites in Colombia was difficult given global and local circumstances. Remote work reinforced the critical importance of communication and structured information management. This hurdle invited collaboration, creativity, and thought as the team navigated through data, qualitative information, and conversations with stakeholders. Furthermore, this circumstance provided a unique opportunity to contribute to the project with my experience in the field studying agricultural initiatives, especially in Putumayo.

Moving forward: linking growth diagnostics with deforestation
As the Colombia Amazon project moves forward, I am eager to learn about how the project will link the development of Amazonian departments with the protection of the rainforest as complements.  Given the urgency of ecosystem and biodiversity loss, its links to climate change and its effects globally, the project comes at a unique moment to halt and reverse damage by providing some analysis on the economic inclusion opportunities for these regions. Furthermore, the insights developed will be critical for the opportunities of rural populations and ethnic minorities in Colombia that have faced the armed conflict, waves of natural resource depletion, and the effects of the country’s rural-urban gap.