How to use the Economic Complexity Index to guide innovation plans
On October 13th, 2022 Data Science Milan held a Meetup at Vedrai headquarters with Mauro Pelucchi as a speaker.
“How to use the Economic Complexity Index to guide innovation plans”, by Mauro Pelucchi, Head of Global Data Science at Lightcast
Intro
Mauro kicked off the meeting introducing the main topic of his speech: the Economic Complexity Theory. From the economic field, he stated, we can learn a lot of interesting metrics that we can apply to several contexts, for instance to the Labor Market one. During the meet-up, Mauro talked about the Economic Complexity Index, a network measures that provide insights into economic development patterns.
Economic Complexity Theory
Let’s start with a bit of background about the Economic Complexity Theory: this concept was firstly developed by Hidalgo and Hausmann in 2009 to identify countries comparative advantage.
Today this theory is used to capture the process of knowledge creation in an economy, by looking at the combination of 2 metrics:
· Diversity: how many different sectors does an area specializes in?
· Ubiquity: how many areas are specialized in a given sector?
On average, it has been observed that economies with a higher Economic Complexity Index tend to have a competitive advantage in several highly complex sectors. In fact, area that can sustain a diverse range of activities know-how, including sophisticated, unique know-how, are more likely to be able to do a wider diversity of activities, including complex ones that few other areas can make.
Mauro suggested to use the Economic Complexity Theory to measure the activity space of a country. For instance, you can observe the activity space for Italy in Fig.1: the colors depend on the sector activity whereas the size is not determined by the frequency of the activity, but rather by a measure of importance that we are going to see later.
We can apply the Economy complexity theory to the labor market and compute the complexity of an occupation based on the diversity of skills it requires, together with their ubiquity. Occupations that request to know a diverse range of know-hows, including sophisticated, unique know-how, are found to be able to do a wide diversity of skills, including complex one that few other occupations can make.
The Economy Complexity Theory is based on 4 indexes:
· Specialization: significant presence or absence of the skills for the occupation
· Diversity: how many different types of skills an occupation requires
· Ubiquity: the number of occupations that require a skill
· Proximity: the probability that an occupation requires skill A given that it requires skill B, or vice versa
If we mix these concepts, we arrive to the Economic Complexity Index. Let’s see an example of the application of the Economic Complexity Index to UK: we see in Fig.2 the evolution of the economy complexity in UK by using labor market data, like job posting data scraped from the web. Mauro said that the grey area is Aberdeen because there are lots of Oil&Gas companies, so the complexity of Aberdeen is constant. London on the other hand, seems to have increased its high degree of complexity.
How is Complexity mathematically defined?
From a mathematical point of view we can define Complexity as:
where Kc, Kp, Mcp, are respectively the complexity of an area, the complexity of a product, the specialization matrix.
Mauro provided an intuition that computing the Economic Complexity Index is in a way similar to a graph application because, if we compute iteratively the formula, we will end up composing a kind of graph. If for instance we apply this theory to the labor market data, we will obtain the results in Fig. 3 for UK: you can see that if we compute ubiquity of different skills, we have some skills going up and others down. Let’s take Agile Methodology for which the ubiquity goes up: it means that this skill is becoming more widespread and, therefore, it’s less complex.
Mauro said that under the hood of the ubiquity computation we have the RCA (Revealed Comparative Advantage), an index used in international economics for calculating the relative advantage or disadvantage of a certain country in a certain class of goods or services.
We can apply this index also to specific skills, for instance Excel is very spread in every occupation, but it does not mean that is very important in every occupation; on the other hand, aerodynamics is very important for specific occupations, but it is not present in every occupation. RCA measures therefore the specialization of a skill: for our example, Excel will have a very low RCA because it’s present in every occupation!
What problems are we trying to solve with the Economic Complexity Theory?
Let’s look at some examples:
· Microsoft Excel is often found among the top ten required skills for every occupation. But what level of knowledge is required for Microsoft Excel? Should a Data Scientist know Excel functions at the same level as a sales representative of accountant?
· How much of robotics does a software developer need to know? What about an automation engineer? What are the most intensive skills an engineer needs to know?
· Is an occupation equally complex between different communities? Can we identify proximities between communities that have common sectors and occupations?
Let’s see in Fig.4 an example of UK Economy Complexity, comparing the probability to increase the employability of an occupation between 2017 and 2021. We see for instance that in Aberdeen mechanical engineers are much more specialized in 2021 compared to 2017. It’s a kind of projection of the labor market that we can have in the future.
We can also measure the diversity of a specialization (see Fig.5), in a way that it helps us measure the evolution of role complexity over time.
Finally, we can also understand the intensity of the relation between 2 skills, like excel and another skill (for instance SQL) — see Fig.6. For instance, we can see how many times Excel and SQL are present together. Mauro says that when Excel and SQL are found together the salary goes up, so it’s a proficient combination of skills.
Recording & Slides
· Video
· Slides
