Actuaries & IT symbiosis for better data. Does it create free radicals or antioxidants?

Does the symbiosis of actuaries and IT working jointly lead to harmful/inert outcomes (free radicals) or to positive business outcomes and solutions (antioxidants)? Our heading implies this question. 

Free radicals are unstable atoms or molecules that have unpaired electrons, which makes them highly reactive with other substances in the body. Antioxidants are molecules that can donate an electron to a free radical without becoming unstable themselves, thereby neutralizing the free radical and preventing it from causing cell damage.

So, while oxidation can lead to the formation of free radicals, antioxidants work to counteract and neutralize their effects.

Actuaries are the cornerstone of the insurance and financial services industries, tasked with predicting and managing risk through meticulous analysis. However, in today's data-driven landscape, actuaries often find themselves entangled in the intricacies of data management—a role that can be both time-consuming and detracting from their core functions.

This predicament is a symptom of a larger issue: the critical yet often underdeveloped relationship between actuarial teams and IT departments.

Actuaries and IT must work together to establish clear, concise, and comprehensive data dictionaries. The accuracy of data mappings in the ETL process is also crucial for actuarial models. Comprehensive test plans are necessary to validate these mappings. The level of data granularity, seriatim versus cohort level—has significant implications for actuarial analysis. Seriatim data offers individual policy detail but comes with increased complexity and volume, while cohort data provides a more aggregated view, which can be less burdensome but may lack depth.

Actuaries rely on fund value roll-forwards to track and project policy values over time. Establishing these requires a robust control environment to ensure accuracy and reliability.

Automation can significantly reduce the manual workload involved in data management, freeing up actuaries to focus on analysis. Backcasting, the process of testing model predictions against actual historical outcomes, is an important validation technique for actuarial models.

Detailed data review techniques such as regression testing and inventory distribution reports are essential for verifying data accuracy. Regression testing ensures that new data or system changes do not adversely affect existing models, while inventory distribution reports provide a snapshot of data characteristics and anomalies. Data governance encompasses the policies, standards, and procedures that ensure data integrity and security. Actuaries and IT must work together to set up proper data governance frameworks that align with organizational objectives and regulatory requirements.

To gauge the effectiveness of the relationship between actuaries and IT, organizations can look at several indicators. These include the speed and accuracy of data management processes, the frequency of data-related issues in actuarial models, and the level of actuarial involvement in data system design and testing.

A strong partnership between actuaries and IT will manifest in streamlined operations, fewer data errors, and more time for actuaries to dedicate to their primary roles.

The symbiosis between actuaries and IT is rooted in the need for reliable data. Actuarial models are only as good as the data fed into them, which underscores the importance of effective data management. For actuaries, being bogged down by data management tasks can be a source of frustration, particularly when it takes them away from analysis and modeling—areas where their expertise is most valuable.

It's essential, therefore, that actuaries and IT collaborate to streamline the data management process.

Perhaps the best example of data systems and joint efforts between actuaries and IT is the almost worldwide implementation of IFRS17. Once insurers are done with the fundamentals for IFRS17 compliance in the right way, we can use IFRS17’s newly developed systems, tools and processes in the future, such as 2025 and onwards, to create a lot of innovative workstreams which we can't right now due to limited data pipelines currently under IFRS4 systems.

Capturing data at far more granularity and comprehensiveness can open doors to analytical permutations and computations in unintended ways. From 2012 to 2023, data science has predominantly focused on creating data pipelines, filtering data, and consolidating it in one place. However, the introduction of new datasets from IFRS17 enables regulators and insurers to shift their focus from data creation and ETL processes to modeling.

And as we all know, modeling with domain skills is the area in insurance where actuaries (or actuarial data scientists) shine the most. Innovation can be merged with regulatory compliance instead of being on opposite ends of the spectrum where one has to choose just one from these two as either or. This can transform actuarial consulting from being ‘regulatory burdens’ to ‘active partners’ of key business decision making.

The Data Science modeling applications are endless and require terraforming a modeling regime that is future-proof but some instances are:

• Principal Component Analysis -for correlation instead of correlation matrix;

• K-means clustering for level of aggregation and correlation and different cohorts making for pricing; it can also cluster for onerous contracts

• decision trees for arriving at optimum level of aggregation

• shift from triangulation chain ladder which is a huge simplification and has little practical capacity for prediction accuracy with individual claims approaches. Such as fitting distributions on frequency, severity and carrying out Monte Carlo simulations on them (especially for truncated small datasets of reinsurance) and applying risk metrics on them. This will optimize not just the liability side, but also the assets/investment and reinsurance side.

• Applying stochastic simulation not just on reserving but also on the pricing side which is currently less prevalent.

• IFRS17 has led to huge implementation costs from software with far more enhanced capabilities than under IFRS4. That has led to increased focus on automation. This automation wave will sweep to a lot of processes in actuarial as well which will mean that actuaries will be able to answer more questions than ever before. And limited scarce actuarial resources are more effectively utilized than spending 90% of their time on number crunching.

• Ethics is an afterthought practically because modeling is based on few factors and small datasets currently. But imagine if norms become like in UK that 400 factors being evaluated to deciding premium levels and ethics will become a huge issue on the forefront. Explainable analytics will also become more important as ‘black box’ are not acceptable regulation wise or even for insurers to not to be able to understand the reasons behind importance in modeling.

• Better data systems mean better capacity to execute product innovation. Products requiring better integrated data systems are poised to thrive. This includes online aggregation distribution channels with API integration, and wellness products that capture lifestyle behavior. The latter is key for life and health insurance, as it accounts for the majority of morbidity risk. Similarly, telematics for motor insurance, capturing driving behavior, play a crucial role in assessing accident risk. Product innovation will be able to get streamlined, generate big data and actuaries will lead the analytics of such big data numbers through modeling and domain skills in insurance.

Having said that, although IFRS17 will increase probability of better analytics, it is by no means a guarantee. Insurers can keep complying with the bare minimum and not have the ability or willingness to explore ways to enhance their analytics, keeping data unutilized in their systems waiting to be discovered. Some insurers will be ahead of the curve, while others will lag behind. Although IFRS17 offers improved analytics opportunities, leveraging these opportunities for competitive advantage will largely depend on each insurer's attitude.

In conclusion, we need to minimize damaging free radicals and maximize the beneficial antioxidants when actuaries and IT collaborate together in a symbiotic manner. The marriage of actuarial acumen with IT expertise facilitates a more streamlined and effective approach to data management, allowing actuaries to focus on what they do best; analyzing data to predict and manage risk.

As the insurance and financial industries continue to evolve in an increasingly digital world, the importance of a well-oiled partnership between actuaries and IT cannot be overstated. It is this partnership that will drive the development of more sophisticated, accurate, and reliable actuarial models, ultimately leading to better risk management and more informed decision-making. Through mutual understanding, clear communication, and a shared commitment to data excellence, actuaries and IT professionals can pave the way for a data-rich future that benefits not just their organizations but the broader landscape of financial services.