How to accelerate your analytics journey in the IFRS 17 world

As the 2021 deadline for IFRS 17 implementation rapidly approaches, insurers would be well-advised to consider a data hub strategy to accelerate their journey to analytics, a speaker stressed last week at the Canadian Insurance Accountant Association’s Fall Technology Seminar.

This was one of the “thoughts from the field” – or lessons learned – that can help insurers with their data infrastructure to set them along the path to achieving outcomes beyond standard compliance with the standard, said Daryl Senick, a partner and national industry lead for insurance with BDO IT Solutions. Creating this superior data architecture can open up a new world of advanced analytics and predictive modelling.

“When we engage with clients, many of them have data challenges and need to build compensating data infrastructure within their data pipeline to address that challenge,” Senick shared with attendees to the seminar, presented by BDO and Moody’s Analytics and held at Vantage Venues in Toronto. “Our recommendation is to take the time that you have available to you to fix [your data infrastructure] as close to the source as possible.

“That compensating logic in your data platform will just increase the complexity of the solution exponentially,” Senick added. “And if you can get ahead of that and fix it at the source, you reduce the amount of headaches you’re going to have.”

International Financial Reporting Standard (IFRS) 17, Insurance Contracts, requires insurers to open up their data infrastructure to connect various systems together. “The way you approach your data for IFRS 17 is really quite important because you could have compliance or you can have compliance and a ‘customer 360’ [experience] or you could have an AI-driven underwriting process,” Senick said. “The path you take around your data infrastructure is really important to ensure that you drive outcomes beyond just simple compliance.”

Senick describes a data hub strategy as “pulling all your data into a single source of truth. You are creating an infrastructure where there is a common data model and there is common interpretation around the data and you’re able to latch on to that capability, leveraging advanced analytics tools, business intelligence tools and AI-type tools.”

Organizations that have multiple systems of record are sometimes going against multiple general ledgers and different types of actuarial systems, Senick noted. “Take the time and unify the meaning of data through a common data model,” he said as an example of another lesson learned. “Our recommendation is to take the time and think about your data infrastructure from a client-centric [perspective]. If you do that, you’ll get your client 360s, you’ll get your AI automation, you’ll get those outcomes beyond just standard compliance.”

There are three principles to a great data architecture:

• Storage
• Analysis
• A prediction model

Using Erie Mutual as a case example, another speaker described how BDO built a customer-centric data model in the cloud for storage to seamlessly integrate into all the insurer’s systems, with an agile data pipeline to load the data into the cloud.

BDO also built a predictive model for all renewals the mutual insurer was processing, said Alex Ng, national practice lead, data & analytics, with BDO Canada. There were two predictions: the first was whether or not there would be a claim on a policy in the next year; the second was the impact of that claim (high, medium or low). BDO trained the model using significant weather events from 2016 and came up with some interesting statistics. For example, analytics showed that primary heating of home and distance from a fire hydrant “actually do have an impact on what claims is going to be made and the impact of that claim.” For auto insurance, the purchase cost of the car does affect whether it will be a high-impact claim, and there were more high-impact claims in the fall season.

“Using all of this data that we collected from their policy management system, we have created a customer perspective, a policy perspective and a reinsurance perspective of that data” to have a ‘claim 360’ perspective, Ng explained. “It allows your claims department to understand where all the different claims are coming from. You can easily toggle to a reinsurance perspective to see what that net amount is and where their not-recovered costs are.”