Undercover Economist

The tricky business of measuring growth

Two experts offer a new approach to weighing economic strength, posing many good questions about the practice.

The barrier to change is not too little caring; it is too much complexity,” Bill Gates once opined, and he was right: many problems in development cannot be solved simply by wanting solutions badly enough. And yet when it comes to one of the key development outcomes, economic growth, the problem is not too much complexity, but not enough.

Complexity plays no obvious role in mainstream economics. Under the surface of traditional accounts of economic growth there is a rather crude model: economies are a bit like loaves of bread. They are made of two or three key ingredients, and bigger loaves simply have a bit more of everything.

Compare the economy of the UK with the economy of the Democratic Republic of Congo, a country with a similar population, and the textbook will say that the UK simply has more physical capital (factories, buildings, roads) and more human capital (education, training) and perhaps even better “institutions”. Of course, everyone knows that you cannot simply turn the DRC economy into the British economy by doubling the quantities of all the ingredients. The British economy is a different and more complex kind of thing altogether.

The economist Ricardo Hausmann and the network physicist César Hidalgo have been trying to measure this complexity, and I’ve written before about their work. They argue that economies are collections of “capabilities”, building blocks that can be put together like Lego to produce different products. A trustworthy post office is a building block; so is high-speed internet; so are functional bankruptcy courts; so is a literate workforce; so is a fast lane at customs for processing perishable foodstuffs. It’s not clear how one would go about measuring all of these capabilities. Instead, Hausmann and Hidalgo measure them indirectly, tracking the shadows that they cast upon a country’s trade statistics.

Their latest work, “The Atlas of Economic Complexity”, includes analysis not just of the general method, but of the “complexity statistics” of 128 countries. Hidalgo and Hausmann show that their generic ranking of economic complexity is much better correlated with gross domestic product than traditional indicators, such as governance or educational standards. The authors seem pleased with this, but it is depressing that they are tempted to engage in such statistical arm-wrestling. Their research is far more interesting than that.

If we can measure economic complexity and find it is highly correlated with economic productivity, then the question is: how can economies become more complex, acquiring new capabilities? A couple of points suggest themselves. Modern economies require complex rules: the English version of EU law contains more than 55 million words, equivalent to about 100,000 pages. Some of this is no doubt useless, but I wonder how much. To shape such rules sensibly is no easy task.

Think of a business that wants to export cut flowers. That requires appropriate phytosanitary regulations, that fast lane at customs, quick transport links between farm and airport, laws governing irrigation and much else. Getting governments to think about all this is a tall order – especially for a business that simply will not exist until the building blocks themselves do.

The second point is linked to the first: Hidalgo and Hausmann find it is easier to develop new capabilities that have something in common with those you already have.

And what of those countries whose existing capabilities offer no obvious avenues for development? The complexity approach asks some good questions, but answers must wait.

Also published at ft.com.