top of page

Economic complexity

Economic complexity is a term we use to refer to the network structures that societies form in order to embed productive knowledge in different individuals and then bring it back together in productive organizations and in the networks these organizations form. Since some products (like computers or jet engines) can only be made in very complex societies and others (like T-shirts or cereals) can be made almost anywhere, the mix of things that a country can produce says something about the knowhow it possesses and can use effectively. Generally, the more complex the web of human interaction is in a country, the more prosperous the country. So how do we know how much knowhow is in a country? We examine two characteristics of a country’s productive output: ​​ Diversity of a country: Simply, how many products does a country make? Countries that make many distinct products are likely to be more complex Ubiquity of a product: Simply, how many countries are able to make a product? Products that are made by very few countries are likely to be more complex We can combine these two characteristics and iterate them. For example, we can calculate how ubiquitous are the products that a country makes or how diversified are the countries that make a product. We can keep at it, by calculating how diversified are the countries that make the products that a particular country makes or how ubiquitous are the products that are made by the countries that make a particular product. We can iterate this process an infinite number of times and calculate a measure of each country’s economic complexity: its Economic Complexity Index (ECI). We can do the same for products and calculate the Product Complexity Index (PCI). ​As one would expect, there is a connection between ECI and GDP per capita: complex countries are generally richer. But more interesting is the fact that the disparity between the actual GDP per capita of a country and the GDP it would be expected to have, given its ECI, is a strong predictor of economic growth. In fact, complexity-based economic variables contain much more predictive information than other commonly used variables such educational attainment, the quality of institutions, the depth of the financial system or the competitiveness of a country as measured by the World Economic Forum. How and why Economic Complexity impacts growth is the subject of ongoing research.


Dark matter


Dark matter is a term I coined with Federico Sturzenegger to refer to the ’invisible’ foreign assets built up from the export (mostly) of knowhow from one country to another. These types of assets do not usually show up in trade statistics, but they can provide the country exporting dark matter with a large and stable flow of income. At the same time, for the countries that import dark matter it works like an invisible foreign liability, and thus a form of expenditure. We estimate dark matter by looking at the disparity between a country’s net investment income and the net investment income that would have been caused by the accumulation of net foreign assets consistent with its observed current account, assuming a normal interest rate. Countries whose investment income is higher than would be presumed on the basis of their cumulative current account balance have dark matter assets, while countries for which the net investment income is too low, given that balance, will have dark matter liabilities. For example, between 1980 and 2012 the cumulative US current account deficit represented about 50 percent of 2012 GDP. But the net investment income of the US indicates that is not paying anything for it. By contrast, during the same period, Chile borrowed less than 15 percent of its GDP but pays about 7.5 percent of GDP for it, an implicit interest rate of more than 50 percent. The way the dark matter approach explains this puzzle is by noting that the US borrowed over 100 percent of GDP in gross debt, took about 50 percent of that to pay for its observed current account deficit and the other 50 percent to fund its Foreign Direct Investment (FDI) abroad. Because these investment mix money with knowhow they obtain a much higher return, say 7 percent, than the amount it pays on its liability, say 3 percent. The US can do this but not Chile, because the US has knowhow that it can productively mix with money to invest abroad. By contrast, Chile needs the knowhow of foreign investors to run its mines, power systems and banks and must pay for this knowhow in the form of excess returns on the FDI it receives. When Chile invests abroad, it does so preferentially in exchange-traded assets and these subject to a no-arbitrage condition in its pricing and hence deliver on average the same risk-adjusted return. One consequence of dark matter is that there is less reason to worry about global imbalances than is usually assumed: the world, as seen when dark matter is taken into account, looks less out of quilt than the picture that emerges using the standard approach. For emerging economies, their external liabilities look quite different than the standard measures in ways that are not equal across countries: Mexico looks much better than Chile when you include dark matter, reversing the picture that emerges using standard measures, that show Chile as almost a net creditor.


economic volatility

​I have dedicated many years to thinking about the causes and consequences of macroeconomic volatility and about the policies to tame it. This is an issue that is close to my heart because Venezuela has one of the most volatile macro-economies in the world in terms of the instability of output, inflation, the real exchange rate and so many other important economic variables that affect welfare, growth and inequality. ​ My first major report as Chief Economist of the Inter-American Development Bank (IADB) was entitled Overcoming Volatility. The idea is that volatility is the consequence of a country’s policy framework and of the impact of this framework on the way a country absorbs shocks, such as changes in the terms of trade, weather events or elections. This led me to study the pro-cyclicality of policies on all fronts. With Michael Gavin, Ernesto Talvi and Roberto Perotti I explored the pro-cyclicality of fiscal policy. With Michael Gavin I explored the issue of bank credit booms and crises. With Alberto Alesina, Rudi Hommes and Ernesto Talvi, I explored the role of budget institutions in determining fiscal outcomes. Finally, I worked on the implications of budget institutions for macroeconomic volatility in the 1997 flagship report of the IADB and published a book with the OECD on the same topic. The work on fiscal procyclicality led to an important finding: some countries have procyclical policies because their ability to borrow is also procyclical: they have market access in good times and not in bad ones. Why? The hypothesis I explored led me to the work on debt denomination and original sin. Countries that have a net foreign debt and cannot denominate it in local currency tend to adopt more rigid exchange rates, accumulate more foreign international reserves, have lower credit rating than their fiscal fundamentals would justify and face more volatile output and capital flows. In a recent paper with Ugo Panizza we showed that the recent improvement in emerging markets is due to the fact that they are borrowing less, not to the fact that they can now borrow in their own currency. ​ With Ugo Panizza and Roberto Rigobon I explored the causes of the excess volatility of developing countries. We found that it is not due mainly to the fact that they face larger external shock, but to the fact that they are less resilient to them. With Roberto Rigobon, I developed a theory of the resource curse based on the idea that the volatility of commodity prices leads to a volatile real exchange rate that makes investment in other tradables more risky and leads to inefficient specialization. In a paper with Francisco Rodriguez and Rodrigo Wagner we showed that Growth Collapses are associated with external shocks that are received by countries poorly positioned in the Product Space.



One of the elements that accompany the development process is the reduction in the difference between the societal roles played by men and women. As work becomes more mechanized and more mental and as fertility rates decline, the old reasons that caused major differences in the work lives of women and men tend to decline. As a consequence, gender gaps in academic achievement, career choices and labor force participation have tended to decline. But progress is quite heterogeneous across countries, levels of development and aspects of gender disparities. I have worked with Laura Tyson,  Saadia Zahidi and the World Economic Forum to create the Gender Gap Report, that measures every year the gaps in education, health, employment and political representation for over 120 countries. In most Latin American countries, for the cohorts who are now under 50 years old, women have more education than men. With Ina Ganguli and Martina Viarengo, I looked at the impact of this change on who marries whom and who stays single. We find that educated women tend to marry less than educated men and uneducated women and when they do mary they do so unusually frequently with less educated men. We also looked across the world at how women’s labor force participation changes with education, marriage and motherhood.  We are currently working on why gender gaps emerge over time in the hierarchical positions within firms. I hope to share the results soon.



Growth Diagnostics is a methodology I developed with Dani Rodrik and Andrés Velasco to determine the obstacles to a country’s capacity to grow. It is a unified framework for identifying the binding constraints to growth, which is key to formulating growth strategies. The main idea is that each country may be bumping against different potential constraints but each constellation of constraints must be giving off a different collection of symptoms or signals. These symptoms can then be used to perform a differential diagnosis. Potential constraints are scrutinized systematically using a decision tree. The, researchers need then to come up with a syndrome that can account for the identified constellation of symptoms and can account for their persistence. By using Growth Diagnostics, policymakers can develop a clearer theory of change by designing policies that can take the country out of (or work around) its current syndrome and relax its most binding constraints. I subsequently wrote a how-to handbook with Bailey Klinger and Rodrigo Wagner. I have participated in growth diagnostic exercises in a bunch of countries including Algeria, Armenia, Belize, Botswana, Brazil, Dominican Republic, Ecuador, El Salvador, Mexico, Morocco, Pakistan, Peru and South Africa, among others. I have also collaborated with the World Bank (setting up the PREM project) and the Inter-American Development Bank (that eventually became a book on the application of Growth Diagnostics to Latin America). Dani Rodrik keeps a website with papers related to Growth Diagnostics. Also, the Millennium Challenge Corporation has adopted an approach based on growth diagnostics that they call constraints analysis, requiring its application to all partner countries.


Original Sin

Original sin is a term I coined in 1998 and analyzed its consequences first in a paper with Barry Eichengreen and then together in a book with Ugo Panizza to describe a situation in which the residents (or government) of a country are unable to acquire foreign debt that is denominated in their own currency. If a country that suffers from original sin accumulates a net foreign debt, as developing countries are expected to do, it will have an aggregate currency mismatch on its balance sheet. When economic conditions worsen the country’s real exchange rate typically depreciates, making its foreign debt more expensive precisely when it is harder to pay. Original sin is a key determinant of the stability of output, the volatility of capital flows, the management of exchange rates, and of a country’s credit ratings. After controlling for the level of development, of tax revenues and of public debt, original sin is associated with a credit rating that is significantly lower than in the absence of this problem. We entered into a debate with Ken Rogoff, Carmen Reinhart and Miguel Savastano, who argued in favor of a hypothesis of debt intolerance; and with Morris Goldstein and Philippe Turner, who argued for currency mismatches. With Ugo Panizza I also updated the evidence on the evolution of original sin and its consequences in a recent paper in the Journal of Globalization and Development. My current research on this issue involves the implications of the Euro for original sin: does a euro member suffer from original sin? Is this part of the cause of the current euro crisis? More on this soon.



Self-discovery (a.k.a. cost discovery) is a term I coined with Dani Rodrik and refers the process whereby entrepreneurs invest in activities that expand the set of products a country makes by exploring the feasibility of new products. It is not about inventing products that are new to the world but about finding whether they can be competitively made in a particular place. The discovery is about the place’s capabilities, more than about the product: hence the term. With this concept, Dani Rodrik and I contested the idea that growth will automatically result from the import of foreign technologies and the existence of stable economic institutions. Instead, self-discovery faces market failures that need to be addressed actively through some kind of activist industrial policy. Pioneer entrepreneurs, i.e. those who make investments in new industries, provide benefits to other entrepreneurs, whether they fail or succeed. In particular, they provide information regarding what works or does not and may solve problems that future imitators need not address. Because of this externality, a market economy may no provide adequate incentives to engage in self-discovery and countries may get stuck in a narrow set of low productivity activities. My paper with Dani Rodrik and later with our student Jason Hwang, allowed us to show a connection between a country’s export mix and its level of development and growth. At the core of the debate was whether the problem of self-discovery is associated to coordination failures or to the information spillovers we emphasized in our 2003 paper. Subsequent empirical research in which I participated with a team put together by the Inter-American Development Bank showed that coordination failures were more important. Also, I started to think that the order in which countries diversified could not be random, as we had assumed in the 2003 paper. If the problem is coordination failures due to missing markets, then there has to be a certain logic why some products become feasible before others, given a country’s initial conditions. This insight led to my work on economic complexity and the product space.

bottom of page