How COVID-19 has reemphasized the need for actuarial transformation
by Deepti KalraMs. Kalra is transformational AI leader with EXL Service, and has over 10 years of extensive experience in leading AI-powered transformation initiatives across Insurance value chain such as underwriting, pricing, actuarial & claims functions. Visit exlservice.com
Nothing in the 21st Century has impacted the globe on the same scale as COVID-19. The state-imposed lockdowns and shelter-at-home restrictions slowed economic activity drastically. As a measure to mitigate the impact of disruption, the Federal Reserve cut interest rates to a range of 0-0.25%. Investors were unsure of what to expect next, as the market volatility indicator (VIX) reached an all-time high of 82.69 and capital outflows from emerging markets hit levels comparable to the 2008 financial crisis.
With uncertainty around the extent of damage that the pandemic could cause, the number of deaths, and the duration until restrictions may be lifted, insurance organizations rely on their actuarial function to effectively model exposure to different types of risks and perform analyses to ensure that the business is not financially stressed and that asset liability mismatch is efficiently managed.
To modernize, actuarial departments have begun upgrading systems and financial models to comply with Long Duration Targeted Improvements (LDTI) and Principle-Based Reserving (PBR) reporting changes. These regulatory requirements and transformation initiatives have also underlined the need to enhance existing capabilities by utilizing data and technology to their fullest. The rapidly evolving global economic and financial conditions due to COVID-19 have reaffirmed that automation and data analytics are the best ways forward to manage uncertain conditions ahead.
Change In Exposure To Risk Factors Due To COVID-19
Risks for a life insurance and annuities business can be largely classified as tactical or strategic. Tactical risks are the type that are actively managed by the organization through internal risk models. Strategic risks are driven by regulatory changes, technology disruptors, and legal or other external change factors.
As exposure to the risks listed above increases, the current, “business-as-usual” actuarial process must be upgraded to account for new scenarios, manual adjustments, additional sensitivity runs, deep dive analyses, controls and audits.
The “business as usual” actuarial process can be largely classified into five stages: input data consolidation, model preparation, model runs, post valuation processing and reporting.
- Input Data Consolidation
For each production cycle run, input data is received from different sources: policy administration systems, finance team, corporate assumptions team and ledger. COVID-19 impacts the timing of input arrival from different sources, either due to additional checks and controls or due to staffing constraints.
Proposed intervention: An automated data consolidation and reconciliation tool can enable reduction in turnaround time for data processing, and help eliminate the possibility of manual errors. It can act as a single source of truth that is available to all consumers.
- Model Data Preparation
Cash flow models require “in-force” population data, along with assumptions, scenarios and market data. Additional “worst case” scenarios must be generated to adjust for COVID-19 conditions and included in reserve calculations along with traditional deterministic or stochastic scenarios.
Proposed intervention: An integrated assumptions database can store all the assumptions, scenarios and market data at a cohort / seriatim level as prescribed or determined for different reporting standards. As a result, all the business units can leverage one single integrated database.
In general, reserve cash flow models are run 20 to 40 times on an average for different blocks of policies or scenarios, which may vary across organizations, business units etc. With PBR & LDTI regulatory requirements, and COVID-19 impact, this requires double the model runs within the same monthly cycle.
Proposed intervention: Robotics Automation (RPA) based run inventory can enable auto model runs for different scenarios & blocks of contracts. An integrated data mart can aggregate the outputs of these multiple reserve model runs at a seriatim level.
Post Valuation Processing: As COVID-19 scenario evolves at a faster pace, depending on mortality rates and market reactions to global trends, an active post valuation processing system would be required to translate those effects into reserve adjustments.
Proposed intervention: An efficient attribution model can help in attributing the impact due to COVID-19 conditions and regulatory changes on reserve changes. A post valuation tool can aggregate the contracts that fall out of automated model runs either due to deviation in product features or run-offs and separately calculate the reserves.
Reporting, Filings and Disclosures: On average, each business unit generates approximately 300-400 reports for internal analysis or disclosures, which may vary across organizations and functions. Additional requests from across the organization, such as finance, corporate controllers, etc., will increase to ensure that reserve estimates are adequately maintained.
Additional regulatory disclosures and reports would also be required to report on operational and financial preparedness.
Proposed Intervention: A full-blown analytics and reporting dashboard can aggregate the data at seriatim level for different products, business units and legal entities. This can serve as a dynamic plug and play dashboard that provides various requestors across the organization an ability to choose the metrics they need and a standardized output to generate survey reports.
New Developments Required To Support Actuarial Reporting
As the COVID-19 scenario unfolds, actuarial organizations will be required to make frequent updates to existing models and systems to account for emerging conditions. Systems should be enabled to react faster, remain flexible for a shorter “time-to-market” and ensure reliable testing.
- Input Databases Update
Longer database update cycles as each new product introduced into the system will have to be coded across different database systems: contracts, in-force, premiums, transactions and fund values to be able to reflect when the sales cycle and onboarding cycle becomes active.as new products have to be coded into the system.
- Model Data Updates
Additional data sets must be prepared for population samples for testing various scenarios, model upgrades and calculation changes. Additionally, frequent assumptions updates are required to account for experience studies which will be affected by the prolonged pandemic.
- Model Development and Updates
Model update cycles will become longer to account for model input data updates. New models development might be required for new products introduced. New testware development needed for granular model testing.
Proposed intervention: Agile operating model for new enhancements can help efficiently manage the developments in multiple sprints with frequent and timely feedback loops. Scalable & efficient testware/ independent validation tool embedded with Artificial Intelligence (AI) based sample generator can help test the model changes with varied sample sets representative of population or entire block of population, as required.
- Post Valuation Tool Development
Post valuation tools and proxy/challenger models will be needed to analyze the impact of policies dropped out of model runs due to new model updates.
Proposed intervention: Automated post valuation engine can help reduce the topside adjustments to be made to the reserve output. Product feature changes owing to regulatory requirements can be automatically ingested into the reserve calculation with this engine with minimal manual intervention.
- Reporting, Filings and Disclosures
The National Association of Insurance Commissioners (NAIC) has initiated a collaborative approach, conducting surveys across the industry by requesting updates on COVID-19-related claim requests and payments every two weeks. As a result, the necessary reports, files and documents must be prepared to comply.
Proposed intervention: Integrated multi-functional data mart and standardized reporting format specific to COVID-19 related premiums and claims can reduce the additional effort required to manage the requests from outside the organization.
Other Operational Impact
Finance: Credit spreads have widened making new purchases of corporate bonds possibly more attractive. Active ALM strategy to counter low interest rates and highly volatile markets requires frequent changes to the investment mix between corporate bonds and government bonds.
Proposed intervention: An automated transfer algorithm can optimize the investment mix. This algorithm-based tool can balance the trade-off between policyholder account value and business solvency by redistributing asset allocation in line with overall business
Employee Productivity: Due to remote work, employees will be operating at lower productivity, either due to technology-related challenges or reduced workforce owing to the pandemic. It has been observed that employee productivity is impacted by ~20% industry-wide at this time.
Proposed intervention: Integrated data factory, and RPA / automation based interventions like highlighted above reduces the manual processing of the tasks and help employees focus entirely on core actuarial activities, and enable them to manage the increased workload in lesser time.
The actuarial function will be impacted significantly due to input data timing issues, new model developments and frequent model runs, as well as topside adjustments, deep dive analyses and reporting requirements related to disclosures and filings.
By implementing the aforementioned analytical interventions i.e. process management, data, AI/ML and automation/RPA driven interventions, either for end to end transformation or customizations unique to products and reporting stages, insurance carriers can transform their actuarial function into an automation and analytics-driven business better prepared for challenges like COVID-19 in the future.