If you are anything like me, the word ‘Economics’ alone elicits feelings of fear. Immediately memories of studying abstract graphs and concepts come tumbling into my mind.
And that was just basic economics. Not even close to the advanced stuff.
There is a reason why economists earn big-money – it is a very difficult, very competitive field…
A Quick Reminder – What are the Ig-Nobel Prizes?
If you have read any of our previous articles about the Ig-Nobel Prizes – you can skip this section! If, however, you have never heard of this award ceremony then keep reading for a quick introduction.
The Ig Nobel Prizes were set up to honour those people and projects that make people laugh, then think. The prizes are handed out for the weird, the imaginative, the unusual – anything that can spur people’s interest in science, medicine or technology. Usually, the ceremony is held every September as a gala-type event in Harvard University’s Sanders Theatre. Unfortunately, in a Covid blighted world this was not possible so the entire event was streamed online.
It is possible to watch the full event for free by clicking here. Previous ceremonies have had the prizes presented to recipients by Nobel Laureates in front of the massed 1100 spectators. Thousands more watch along online.
The Economics Prize – Background
The Economics Prize was awarded to researchers from France, Switzerland, Australia, Austria, The Czech Republic, and the UK – with the award being presented to Pavlo Blavatskyy for discovering that the obesity of a country’s politicians may be a good indicator of that country’s corruption. The prize was awarded by Nobel laureate Rich Roberts (Physiology/Medicine, 1993) at the timestamp of 35:20. Pavlo offers a brilliant and concise explanation of the work that won the award during the acceptance speech, which we recommend watching.
Political corruption is incredibly difficult and dangerous to measure directly, for obvious reasons. Conventional indirect measures usually rely on aggregated opinion surveys among foreign experts such as the Transparency International Corruption Perceptions Index – but this is subject to perception bias.
Post-Soviet states are used as corruption in the region is perceived to be a serious problem. According to the index aforementioned – Uzbekistan, Tajikistan, and Turkmenistan are ranked 157th, 161st, and 167th in the world for corruption out of 180 countries, where a lower place on the scale indicates a greater level of corruption.
A Brief Explanation of the Methodology
The dataset for this research comprised of 299 frontal face images of cabinet ministers in office in 2017 from 15 post-Soviet states. Cabinet reshuffles presented a problem for this work, but the issue of having several individuals occupying the same position was solved by simply selecting the individual who held the position for the longest period. To find the images of ministers Google Image search engine was used by entering the ‘Name Surname + 2017’ in the language of their country. Whenever possible, photographs that resembled passport photographs were selected that offered an unobscured frontal view, all taken in 2017.
These images were then used to estimate individual minister’s body-mass index using a computer vision algorithm developed by an external source, with robustness checked by another algorithm also developed externally.
Essentially, the algorithm builds an artificial neural network that is first trained to recognise human faces in a comparable fashion to human visual recognition. Following on from this the artificial neural network is then trained to associate the recognised faces with body-mass indexes – a generalisation of the process by which humans learn to classify others into discrete BMI categories; underweight, normal, overweight, and obese.
The Results – Corruption Visible on the Faces of Politicians?
The high cost for society extorted by political corruption means it must be measured as accurately as possible, with current methodologies for doing so being severely limited. Novel objective measures of indicators have been used over the years, like the import value of luxury Swiss watches although this has fallen out of use in recent times.
The results of this paper propose another solid indicator of political corruption – BMI. The estimates of BMI were found to be strongly correlated with conventional measures of corruption. This creates an interesting situation in which the level of corruption is proposed to quite literally be evident on the faces of politicians (however, this is not to suggest that individual obese politicians are more corrupt than others).
As photographs of top members of government are widely available across the globe the findings of this research are of significant importance. In countries where conducting micro-level surveys is not a possibility and foreign-experts do not have access, this method can be applied. Further to this, it is now a distinct possibility to measure historical corruption from archived photographs – something not possible before.
Figure 1. Scatterplot of median estimated ministers’ BMI against Transparency International Corruption Perceptions Index (2017), lower values indicate higher levels of corruption
Figure 2. Scatterplot of median estimated ministers’ BMI against World Bank worldwide governance indicator Control of Corruption (2017), lower values indicate higher levels of corruption