Building useful analytical tools brings opportunity but it also takes time and money. If the financial opportunity can be expressed in terms of dollars and that measure can become a target, then it is much easier to justify the investment into analytics.

The term Data Quality Management (DQM) covers much of the improvement of data as described above. As the Casualty Actuarial Society White Paper ‘Data Quality Management in the P&C insurance sector’ emphasizes, senior support is necessary:

‘Sometimes the biggest obstacle to appropriately managing data quality, from an actuarial perspective, is a lack of agreement amongst key stakeholders within an organization as to the importance of DQM to the organization. Many P&C insurers find themselves simultaneously addressing several high-priority immediate concerns. If key decision makers do not see DQM as an issue needing immediate attention, then it is very unlikely that resources will be allocated to it.’ (2)

If you can show that improving a presently unmeasured metric will create a financial benefit of, say $200m over a four year period then it becomes much easier to get the senior support & funding for a new team of four analysts.

‘Annual $’s Optimized’

There is an opportunity to create targets using a measure of risk-differential improvement: ‘Annual $’s Optimized’

Alteryx encourages the use of their platform with a measure of ‘Hours Optimized’. This is largely a subjective measure which is encouraged and promoted by stakeholders. It is very successful in underpinning the perception of Alteryx’s value.

I would suggest the creation of an ‘Annual $’s Optimized’ measure. This needs to be both simple and meaningful. My suggestion is to base ‘Annual $’s Optimized’ on a simple calculated value:

Annual $’s Optimized = Absolute value( Loss Ratio - target Loss Ratio) x Premium

Loss Ratio obviously has many complexities and it varies by IBNR adjustment and year. Nonetheless reasonable choices can create meaningful Loss Ratio goals for a variety of groupings. Lines with high severity and low frequency may well be excluded, or only included in larger groupings where their noise is absorbed.

Analysis can be used to identify large differentials between the actual LossRatio and the target Loss Ratio using the ‘Annual $’s Optimized’ measure and these can become the prime focus of efforts. For instance – if a $300 million book of business is found to have a 40% differential between the Loss Ratios of the Urban and Rural groupings (and shown to be independent of other measures), and the knowledge spread among the pricing stakeholders then, it is reasonable that a significant reduction in that differential should be evident within 2 years. The opportunity would be seen as an ‘Annual $’s Optimized’ value of $120m/ year.

This measure is simply putting a $ number to the visible discrepancies between the (Perfect World) target Loss Ratios and the actual.

Senior level buy-in

The improvement in data as described above is a significant component of Data Quality Management (DQM). As the Casualty Actuarial Society White Paper ‘Data Quality Management in the P&C insurance sector’ emphasizes, senior support is necessary:

‘Sometimes the biggest obstacle to appropriately managing data quality, from an actuarial perspective, is a lack of agreement amongst key stakeholders within an organization as to the importance of DQM to the organization. Many P&C insurers find themselves simultaneously addressing several high-priority immediate concerns. If key decision makers do not see DQM as an issue needing immediate attention, then it is very unlikely that resources will be allocated to it.’ (2)

A framework for explicit targets that relate to dollars and profit will help to establish senior level buy-in.

References

(1) Casualty Actuarial Society Monographs: A MACHINE-LEARNING APPROACH TO PARAMETER ESTIMATION Jim Kunce and Som Chatterjee https://www.Casualty Actuarial Societyact.org/sites/default/files/2021-02/06-Kunce-Chatterjee.pdf

(2) Casualty Actuarial Society Monographs: DATA QUALITY MANAGEMENT IN THE P&C INSURANCE

SECTOR Graham Hall, BSc, FIA Mark Jones, BSc, MA, ACasualty Actuarial Society, MAAA Kevin Madigan, PhD, ACasualty Actuarial Society, CERA, MAAA Steve Zheng, ASA

https://www.Casualty Actuarial Societyact.org/sites/default/files/2021-02/09-hall-jones-madiganzheng_0.pdf