Reputation risk contagion: How one company can drag a whole sector down

The March edition of Journal of Network Theory In Finance, reveals that it just takes negative coverage of one bank to affect perceptions of the industry as a whole.

A leading Risk professional, Peter Mitic, used alva’s daily company reputation ratings for 10 of the leading UK banks, comparing larger-than-average movements in the reputation of each bank. He examined how a significant shift in one bank’s rating was trailed by a rise or fall in another’s.

The effects of these reputation changes for one individual bank on all the others in the group were calculated by modelling the spread of reputation contagion in a DeGroot network. Mitic uses the term “network drag” to describe the impact reputational events at one brand can have on others in the sector.

While the effect was not universal across all companies analysed, Mitic found that some banks could experience declines of almost 24% in their own corporate reputation scores, due to events at a rival company. The positive effects were even more pronounced, with some banks receiving a reputational boost of over 40% related to reputational uplifts at their peers. Overall, Mitic found that between 10-15% of the company reputation of any one bank in the sector is attributable to network drift.

The importance of this becomes clear when Mitic explains that network drag can move a bank into the stressed condition that can result in sales increasing by up to 3.4% or falling by up to 7.9%.

This is just one practical example of how alva’s integrated metrics approach enables companies to combine the insights from media and social media analysis with their own internal data in order to understand the affect stakeholder activity has on the bottom line and business KPIs.

Read our white paper on integrated metrics to understand how you can get more from your data.

Full abstract of Peter Mitic’s paper:

The effects of the reputation of any single member of a group of agents on all the others in the group are calculated by modeling the spread of reputation contagion in a DeGroot network. The reputation of individual agents is measured by compiling a reputation index for each agent over an extended period. Transition probabilities within the network are assessed by considering extreme reputational events using a Bayesian approach.  

The results indicate that consensus is reached quickly, and influential agents can be easily identified. Agents in the network with a very positive reputation serve to mitigate the negative reputation of other agents in the network. Approximately 10–15% of the reputation of any agent in the network is attributable to network effects; positive reputations are deflated and negative reputations are inflated. The network effect on the sales of any single agent can be estimated once the reputation score has been translated to sales.

The full paper is available in the March 2017 edition of Journal of Network Theory in Finance