# What's the methodology behind "Liveability Ranking"?

## News

• Author: Carlos Alberto Gómez Grajales
• Date: 20 September 2016
• Copyright: Image appears courtesy of Getty Images

Last month, The Economist Intelligence Unit released its latest “Liveability Ranking” 2016 Report [1]. This time, the city of Melbourne is ranked at the top as the best place to live on earth. Vienna and Vancouver come next. In total, three Australian cities are in the top 10, along with 3 Canadian ones. As you can imagine, the media immediately started reporting how I should immediately move [2].

Since you may have read a lot by now about the actual results, I’d like to focus on something else, on the methodology behind the index. As a statistician, I’m a huge fan of every numerical exercise, but I have learnt through many years of work that it is always healthy to understand how an indicator is calculated before making decisions based on it. So let’s discuss what The Economist Intelligence Unit does before you start packing.

Sadly, the report distributed by The Economist Intelligence Unit lacks much depth in the methodology section [3]. As some of you may know, there are several statistical techniques that create mathematically sound indices, such as Principal Component Analysis or Factorial related models, though none of these are reported in the document. The report does not mention in detail the calculations, only the general background behind the index.

According to the document, the final score of a city is calculated as a weighted sum of the values of a number of variables. These variables are organized into 5 categories: Stability, Healthcare, Culture & environment, Education and Infrastructure [3]. The categories seem to have some nice and rounded weights, with education being the least important one for the index, accounting for only 10% of it and both Stability and Culture & Environment having the highest weights with 25%. The other two categories account for 20% each of the index. There is no indication anywhere on the report as to how these weights were obtained, rounded or adjusted.

Each of the variables that encompass a category is rated in one of five ordinal categories: acceptable, tolerable, uncomfortable, undesirable or intolerable. For the quantitative variables, a rating is calculated based on the relative performance a city has in that indicator, a performance which is assessed from an external data point, usually from a World Bank indicator or from another international agency, such as Transparency International. Out of the 30 factors that the “Liveability Ranking” considers, only 7 are calculated this way. The rest of the indicators are considered “qualitative indicators” and, as such, their results are more subjective. The report states that, for these 23 variables, a rating is awarded based on the judgment of in-house analysts and in-city contributors. It is not stated how many in-house analysts and in-city contributors are reached in order to attain the performance of individual cities, nor the selection process of these respondents. Still, about 80% of the index is based on these evaluations (these constitute the “Global liveability survey”, with its results sold separately). As such, the latest “Liveability Ranking” relies mostly on what a non-representative group of people around the world think of their cities, than on objective indicators of performance. Perhaps, only perhaps, in-house analysts in Australia tend to be less demanding than those living in the UK.

It is worth mentioning that this in no way diminishes the results of The Economist’s analysis, nor promotes the use of any of the other city indices available [4] [5]. By understanding how the index is created, its depth and weaknesses, anyone can understand its usefulness and importance in its due context. This is not only true for city indices. Any statistical prediction, model of indicator has to rely on solid grounds, not only mathematically speaking, but considering the relevant field expertise. As any statistician knows, it is important to document our work, so anyone can understand the basis of our work and its current limitations.

Full Disclosure: I have no idea where my own place of residence, Mexico City lies in the index, though I don’t imagine it is high in the list. I can totally agree with that.

References

[1] The world’s most liveable cities. The Economist Website (August, 2016)
http://www.economist.com/blogs/graphicdetail/2016/08/daily-chart-14

[2] Roy Choudhury, Saheli. These are the most livable cities in the world. CNBC Website (August, 2016)
http://www.cnbc.com/2016/08/17/melbourne-vienna-vancouver-ranked-top-three-most-livable-cities-by-the-economist-intelligence-unit.html

[3] Global Liveability Ranking 2016 Full Downloadable Report. The Economist Website (August, 2016)
http://www.eiu.com/public/topical_report.aspx?campaignid=liveability2016

[4] Mercer Quality of Living Rankings. Mercer Website (2016)
https://www.imercer.com/content/mobility/quality-of-living-city-rankings.html

[5] Monocle’s Most Livable Cities. Monocle Magazine Website (2016)
https://monocle.com/film/affairs/top-25-cities-2016/

View all

View all