Kevin De Bruyne, a world-class midfielder, signed a new deal with Manchester City Football Club recently, extending his contract until 2025. Premier League players getting new contracts would hardly be worth a mention outside the sports pages but De Bruyne took a novel approach that would be quite familiar to the modern corporate risk manager: De Bruyne used analytical tools to establish his potential value to the club and used those numbers to negotiate with the club.
Instead of using an agent to negotiate, De Bruyne employed a data analytics company to analyze his past performance to project his likely future value. This strengthened his hand at negotiations and ultimately led to a 30% raise. Not a bad day’s work.
The use of analytics in sports is not exactly new. The movie Moneyball popularized the first known instance of analytics in player valuations in the 2002 Major League Baseball season when California’s Oakland A’s, operating on a smaller budget than their bigger rivals, used analytics to identify value-for-money players who came together to go on a record 20-game winning streak. Analytics are now used to some extent by all teams across the major leagues.
De Bruyne’s story came to mind recently after a large client turned to us for analytical support upon finding its total premium across all lines would go up by nearly 30%. Insurers also sought deductibles that were four times higher than the previous year on some of their biggest risks.
When viewed together with their total premium spending over the last decade and what they had recovered in claims over the period, the risk manager and the finance team intuitively felt that insurance was poor value. But how could they reach such a conclusion with a reasonable level of confidence? A decision like buying less insurance cannot be taken lightly and will come under heavy scrutiny if there is an adverse event that could have been insurable.
This is where our analytics came in. We looked at their historical claims experience for each line and the changes to their exposure metrics over the same period. We combined these finding with market data to build models that projected their likely future claim experience. Our model predicted not just their expected experience in an average year but also their likely losses at various probabilities – one in 10, one in 100, one in 200 return periods and so on – exactly the scenarios for which one buys insurance.
On top of the model results, we overlaid the insurance structure to forecast the losses that would be retained by the client within deductibles as well as in excess of the insurance limits and the losses that would be ceded to the insurers, both for individual lines as well as for all the lines together.
For the most part, the model confirmed what the risk manager and his team felt: The current insurance program was expensive. Even in an extreme one in 200 scenario, they would only recoup less than 10 years’ worth of premiums. There were certain lines where insurance was cost-effective, but we concluded there just wasn’t enough value at an aggregate level.
These insights allowed the client’s risk team and our brokers to do two things immediately:
By deploying analytics more effectively, they could identify pockets of value in the insurance market and ensure that their premium spend was directed to those lines. Not only that, much like De Bruyne was able to do with his team, the risk manager was able to demonstrate that the insurance they were going to buy in 2021 did in fact provide “value” – with the key benefit of an audit trail, thanks to the use of analytics.