Why is Switzerland so rich? Why is Portugal so poor? And what exactly is the recipe for going from poor to rich? These are the sorts of questions that development economists and policy folk from a thousand think-tanks spend their days and nights puzzling over. But according to a group of economists and data-crunchers at Harvard’s Center for International Development, the answers may not be as elusive as some fear.
The Harvard researchers have spent the last few years developing a predictive data set that suggests that present and future prosperity is not much to do with natural resources, or location, or infrastructure. It’s not even directly to do with skills – a relatively poor country like South Africa can have a great record of producing Nobel prize-winners, and may have some unique high-level skills that no one else can match, and still stay relatively poor. Rather it is to do with complexity and knowledge – how many different kinds of skills there are in the whole economy, and how well they connect up. Complex economies do better at making their citizens wealthy.
This is the argument of the project known as the Harvard Atlas of Economic Complexity. The Harvard economists claim that the Atlas can predict growth patterns better than any other rankings, including the World Bank indicators on governance, the World Economic Forum’s competitiveness indicators, or other rankings based on education or on the availability of finance.
The Harvard project is led by the Venezuelan development economist Professor Ricardo Hausmann. In a recent interview Hausmann summarised the conclusions: “We take two measures. How much a country knows, and how easy would it be for it to know more. Together those variables can predict how fast it will grow. Not just the country as a whole – we can predict how fast an industry can grow in a given location. We can predict the complete appearances and disappearances of industries, and the results are five to ten times more accurate than random guesses.”
The most important insight of the Atlas is that it is not just human or infrastructural resources that determine future growth, but the inter-relatedness of those resources. Complexity is a measure not only of what knowledge resides in an economy, but also how productive it is and how easy it is for an economy that manages to produce one kind of complex high value product to also produce a related product. Once there are enough ‘connected’ individuals to make something like advanced machinery, they will probably find it relatively easy also to use at least some of the same skills to make chemicals or electronics as well.
On the other hand, there are some industries that do not connect well with others. Natural resource industries are a case in point: a country may well have vast natural wealth but that does not mean that it will be able to put that wealth to work in developing more complex industries. Nigeria and Angola are among the least complex economies in the rankings, but are still wealthy – but only so long as the oil price holds up. Complex economies are not so vulnerable and tend to grow wealthier over time irrespective of cycles in resource prices.
This complexity analysis attempts to capture some very subtle qualities in the make-up of individual economies, by comparing countries on the basis of the sophistication and interconnectedness of knowledge. But it does this using very simple publically available trade figures which are turned into a dataset that can measure complexity and economic potential. It does this in three stages. First it calculates how many kinds of things a country exports – the diversity value. Then it adds a measure of how common these exports are among all exporters – the ubiquity measure. Finally it uses the first two values together with a standardized ranking of products by complexity, to calculate how readily a country can ascend the complexity scale: this is the measure of potential. The share of the economy taken by exports is not relevant, and nor is the openness of an economy to trade. Only the nature of exports is taken into account.
For example, Egypt and Switzerland have a roughly similar total GDP at purchasing power parity, but with its much smaller population Switzerland is about eight times richer than Egypt in per capita terms. How is this difference explained? As it happens the two countries export a similar number of categories of products, around 180. But the products that Egypt exports are much more ‘ubiquitous’ than the products exported by Switzerland (meaning that a large number of other countries are able to match the Egyptian products). In addition the products that Switzerland exports are only matched by other highly diversified economies, while Egypt’s exports tend to be matched by poorly diversified economies. Together these measures of diversity and ubiquity generate a ranking that measures the amount of productive knowledge in each of these economies. In the latest online release of the Atlas which is updated every two years, Switzerland is rated the second most complex economy in the world, and Egypt the 66th (the rankings based on data up to 2013 are available here).
But for policymakers the future is always more interesting than the past. They will look to the final part of the analysis – the Complexity Outlook Index – which contains at least some hints as to what a country’s options are when it comes to developing a more complex and valuable economy. Here the interesting conclusion is that it is not the most complex economies that have the highest growth potential, but rather those with complex economies that could easily become more complex (which helps explain why relatively sophisticated economies like those of Eastern Europe have experienced such rapid growth in recent years). Economies at the top of the complexity tree have no existing economic territory to move into: they face the challenge of creating entirely new industries if they are to grow fast.
Another interesting conclusion that emerges in the margins of the research is that it is not the complexity of products themselves that determines an economy’s potential, but the complexity of the skills required to produce them. Technologically dazzling products like micro-chips turn out not to require particularly complex skill sets (which is one reason why such production is so readily transferred to emerging economies). According to the Harvard Atlas the most diverse knowledge is called for in rather less glamorous products like industrial machinery and chemical analysis devices.
The complexity approach should explain exactly how much wealth a given collection of skill groups can produce. It almost achieves that: according to the Harvard Center the complexity analysis is 78% correct in accounting for prosperity. Setting aside the natural resource effect, which is stripped out of the data, the remaining discrepancy seems to down to non-economic factors, such as policies that constrict economic development, and political interventions that boost prosperity temporarily. As a result some countries are richer than the analysis says they should be, and some the reverse. Greece for example seems to be much richer than its knowledge base would justify, while India seems much poorer. Over time it seems likely that these anomalies will be corrected – something that is happening rather rapidly in the case of Greece.
It is at this point that the Harvard approach runs out of data and into the uncertainties of policy and implementation. How can an economy dependent on non-complex products develop higher-level knowledge industries? Some, like South Korea (currently ranked the fourth most complex economy in the world), have achieved it. Most have not, although that is not for want of trying. Complex knowledge tends to be awarded to those who already have it. Breaking that anti-development cycle remains the biggest challenge for the world’s economic policymakers.