CONTINUOUS RISK ASSESSMENT… UNVEILING ITS TRUE WORTH OR JUST HYPE?
Petro Appelo
Presentation
Relevant practice area: Life Insurance
Suggested audience knowledge level: Foundation
In this presentation, we explore the concept of continuous risk assessment (CRA), seeking to uncover its potential value amidst the ever-increasing flow of ongoing data in our lives. Join us as we examine the potential untapped opportunities that lie ahead for the life insurance industry. We believe that substantial changes can be made as to how insurance is sold, client interactions are conducted and portfolios are managed.
Continuous risk assessment goes beyond traditional, incidental or periodic assessments by integrating real-time or near-real-time data, including sensor and customer behaviour data. This data is often directly or indirectly related to the underlying health of individuals. Specifically, within the context of life insurance, CRA refers to the ongoing and dynamic evaluation of the health and claim risks associated with insured lives. Through the collection, analysis, and interpretation of data over time, we gain a comprehensive understanding of the ever-evolving risk landscape faced by insurers.
In its simplest form, CRA implies that we are able to understand risk – to some extent – on a more continuous and up-to-date basis. By understanding changes in the underlying risk profiles of customers, we are able to proactively identify emerging risks, detect anomalies or changes in risk profiles, and ideally, respond to mitigate potential losses. This dynamic approach to risk assessment enables insurers to make precise, timely decisions, enhance their risk management practices, and ultimately provide better protection and value to policyholders.
We stand at an important juncture, questioning how best to utilise this wealth of data. In theory, we can adapt underwriting, products, pricing, and risk management strategies accordingly. Through new and evolving technologies, consumers are obtaining increasingly credible health insights related to their own potential health futures and thus the asymmetry of knowledge between insurer and consumer will continue to diverge even further. However, it remains crucial to stay true to the core purpose of insurance – delivering comprehensive coverage when individuals need it most.
Join us in this presentation as we explore some of the opportunities that arise, prompting us to re-evaluate the traditional ways in which we have offered insurance to our customers illuminating its potential value and distinguishing it from mere hype.
ActuaryGPT: APPLICATIONS OF LARGE LANGUAGE MODELS TO ACTUARIAL WORK
Caesar Balona
Presentation with Paper
Relevant practice area: Data Science
Suggested audience knowledge level: Foundation
The recent advances in large language models, such as GPT-4, have generated much interest in their potential applications in various fields, including actuarial work. This paper presents a feasibility study of using large language models for actuarial work, both directly within actuarial modelling, and indirectly as an actuarial workflow assistant. We provide the actuarial audience an introduction to the concept of large language models and their potential applications in actuarial science. We will examine the specific areas of actuarial work where LLMs can be applied, including reserving, pricing, underwriting, capital management, product reviews, and more. We provide case studies with code that demonstrate the potential of large language models to enhance actuarial work. Overall, our results suggest that large language models have the potential to be a valuable tool for actuarial work, particularly in tasks that involve natural language processing or unstructured data analysis, as well as a coding assistant. However, there are also limitations and challenges to using these models in actuarial work, particularly on the side of professionalism and ethics, which we provide some high level guidance on.
The objectives of this paper are:
– To provide the actuarial audience an introduction to the concept of large language models and their potential applications in actuarial science.
– To review the current literature and case studies on the use of large language models in actuarial work, including their effectiveness in various areas such as reserving, pricing, underwriting, capital management, and product reviews.
-To provide several case studies } that demonstrate the potential of large language models for actuarial work.
-To assess the impact of large language models on the actuarial industry and explore potential future developments and applications.
-To provide a comprehensive overview of the current state of large language models in actuarial science and to offer insights into the potential benefits and limitations of their use in this field.
Overall, this research has the potential to contribute to the advancement of both actuarial work and the use of natural language processing in financial fields, with potential benefits for a wide range of industries and applications.
EVOLVING SOCIAL SECURITY IN SOUTH AFRICA: AN ACTUARIAL PERSPECTIVE
Megan Carswell, Andrea Bezuidenhout, Stephen Walker, Natalie van Zyl
Presentation with Paper
Relevant practice area: Retirement Matters / Professional Matters
Suggested audience knowledge level: Intermediate
This paper sets out South Africa’s current social security benefits using the Tiers of the International Labour Organization pensions model.
The analysis considers the benefits available under various adverse life circumstances. This highlights key reform considerations, some of which are under discussion while others are not yet fully explored. Two areas of policy reform considered in detail are the introduction of a Basic Income Grant (BIG) and improving retirement outcomes. Despite clear overlap, these issues have mostly been considered separately.
Universalisation of grants, including the Old Age Grant, could improve retirement outcomes by eliminating the poverty trap and encouraging preservation of occupational retirement savings, as well as appearing more fiscally fair. Despite stakeholder consensus, phasing out means-testing has not yet happened.
BIG became a serious policy issue when the COVID-19 pandemic struck in 2020. Most now agree on BIG being necessary, however the details are still debated. The benefit level is a major issue, particularly given that the existing “temporary” grant has become a de facto BIG over the last almost three years. Research shows that a BIG will likely have positive labour market effects but making it conditional on participation in labour activation programmes may be unnecessarily exclusionary. There remains debate about the economic impact of the redistributive effects of BIG. Funding the BIG represents a contentious issue. There is little appetite to increase government debt or to re-allocate the existing national budget, given existing constraints. Therefore, focus has been on how to provide funding by changing the tax system which poses political challenges. The difficulty in modelling BIG’s long-term costs hinders discussions.
Similarly, the idea of a National Social Security Fund to improve retirement outcomes has broad support, but there is no consensus on the details. Some in government prefer a partially-funded defined benefit approach whereas others argue that a funded Defined Contribution approach would provide more security, greater economic growth and more affordable contributions. Including informal sector workers in a suitable scheme structure with fair incentives is also keenly debated.
Actuaries want to be viewed by all as impartial trusted partners but are often seen as being purely business focused. We can play an important role by helping decision makers understand the implications of different courses of action and participating in inter-disciplinary research groups. We are striving to contribute meaningfully while being seen as professional, capable, respectful, curious and pragmatic.
FAIRNESS OF ANNUITY PRICING FOR THE LOWER INCOME SECTOR
Joanna Combrink, Dale Taylor
Presentation with Paper
Relevant practice area: Retirement Matters / Professional Matters
Suggested audience knowledge level: Foundation
Background and purpose: National Treasury has stated publicly that the South African life annuity market does not present good value, particularly for low-income earners. The purpose of this paper is to look into the validity of this statement. This research is particularly relevant in South Africa today because of the ongoing retirement reform and its implications for the annuity market in future. With South Africa moving from a world of voluntary participation, voluntary preservation and largely voluntary annuitisation to a world where annuitisation (through legislative changes of 2021) and preservation (through proposed changes in 2024) are compulsory and there is intention of future compulsory participation, it is critical to ensure that the annuity products offered to retiring individuals are fairly priced.
Method: We first settled on a definition of fairness. We investigated research around the impact of income on mortality. We approached 4 large insurers for annuity quotations.
Results: Mortality investigations show that income has a direct impact on mortality. We estimate that the difference in mortality should account for a 24% difference in annuity pricing between low-income and high-income individuals.
There are two distinct annuity markets in South Africa: the “retail” market – through which individuals access annuity products, and the “institutional” market – through which entities outsource their pensioner liabilities to an insurer in bulk.
In the institutional market, income is taken into account in annuity pricing with the price per R1 of pension being 17%-33% higher for high-income individuals than for low-income individuals.
In the retail market, income is not generally taken into account in annuity pricing. The cost of an annuity for a low-income individual accessing a pension in the retail market is around 30% higher than the cost of the same annuity purchased by an entity in the institutional market. The difference is not fully explained by anti-selection (maximum of 15% based on our estimates) nor by explanations offered by insurers.
One insurer does take income into account in annuity pricing but the cost of this annuity for a low-income individual is not different to that offered by other insurers.
No similarities or helpful interventions were found in international markets.
Conclusion: Based on the findings, and relative to the definition of fairness used, we conclude that the there is certainly room for improvement to make the pricing of annuities fairer in the South African retail market.
ACTUARIAL REMUNERATION LANDSCAPE
Wilhelm de Wet, Henda Pretorius, Jurie Gouws
Presentation
SA3 has successfully conducted a salary survey amongst Actuarial professionals in South Africa for the past 7 years. In 2020 we shared our insights around the survey at the annual ASSA Actuarial Convention and our session was the second most popular session of the entire convention. This year we aim to dig into these results again, but with more data, from 2016 – 2022. We aim to cover the following in our presentation:
– Introduction to survey methodology and disclaimer
– Interesting data points
– Trends in Guaranteed and Variable remuneration over the last 7 years – Split between:
o Qualified Actuaries and
o Student.
– We will also give an overview of the demographics of actuaries that contributed to the survey. We look at the breakdown of:
o Practice area
o Location
o Gender
o BEE status
– Lasting impact of Covid 19
– Predictions for 2023 Salary Survey results
BENEFIT STATEMENTS FOR DEFINED CONTRIBUTION RETIREMENT FUNDS – INTERNATIONAL DEVELOPMENTS AND BEST PRACTICE
Tommie Doubell
Pre-recorded Oral Presentation
Relevant practice area: Retirement Matters
Suggested audience knowledge level: Intermediate
The presentation will aim to update delegates on the trends in communicating DC benefit statements and projections internationally.
– The presentation will recap on the FSCA draft conduct standard on benefit projections.
– It will look at international developments and best practice / guidelines.
– The UK’s move to automatic enrollment from 2018 resulted in a move to clarify communication of benefits to members, which could be relevant to us in SA.
Since the pensions industry in SA is still waiting on final guidance from the FSCA on benefit projections, it is a good time to provide an update on international trends and may influence actuaries to improve this aspect of communication to retirement fund members.
FROM BIAS TO BLACK BOXES: UNDERSTANDING AND MANAGING THE RISKS OF AI – AN ACTUARIAL PERSPECTIVE
Valerie du Preez, Ronald Richman, Shaun Bennet, Caesar Balona
Panel Discussion
Relevant practice area: Wider Fields
Suggested audience knowledge level: Foundational
We will explore key issues & challenges in AI risk management from an actuarial perspective based on discussions and outcomes from a transregional industry focus group. We will define the gaps and challenges faced when it comes to implementing and utilising modern modelling techniques from a risk-perspective, in the context of regulation. We will also discuss best practice guidelines and how to take a practical approach to AI risk management.
Practical outcomes include:
– An overview of key issues related to Data Science and AI regulation and risk management from an actuarial perspective
– Industry examples of the key modelling issues in practice in traditional and non-traditional actuarial work
– Best practice guidelines for Data Science and AI risk management
– Practical tips for Data Science and AI risk management
This discussion is aimed at participants who are looking to implement AI or Data Science within their functions or looking to improve their risk management in these domains. No prior technical knowledge on the topic is required.
Research findings from the working group is available on request at info@actuartech.com and we invite comments on our panel discussion either before or during the session.
INTERPRETING IFRS 17 KPIs: A PRACTICAL INTRODUCTION USING PYTHON
Valerie du Preez, Nico du Preez, Lara van Heerden, Patrick Moehrke
Workshop
Relevant practice area: Life/ General Insurance
Suggested audience knowledge level: Foundation
As the new IFRS 17 Standard becomes commonplace in industry, it is important for organisations to produce the BI necessary to compare their financial position to their peers. The differences in disclosures and reconciliations produced under IFRS 17 compared to IFRS 4 mean the traditional metrics may no longer apply. This presents a new challenge for actuaries and financial professionals looking to interpret results quickly. We will be hosting a workshop showcasing key reporting metrics of IFRS 17 including examples of visualisation and dashboarding.
Practical outcomes include:
– An overview of KPIs introduced by IFRS 17 and how they compare to IFRS 4
– Industry examples of IFRS 17 KPIs currently in use
– How to interpret and compare commonly used IFRS 17 KPIs
– An introduction to modern tools and techniques to visualise and dashboard KPIs, including examples of using Streamlit in Python
If delegates are interested in a summary of the Python libraries we’ll be showcasing, please contact us on info@actuartech.com.
MODELLING BANK RUNS: AN AGENT-BASED APPROACH
Johan Eybers
Presentation
Relevant practice area: Banking
Suggested audience knowledge level: Foundation
Description:
The presentation will start with a brief historical overview of bank runs. We trace their occurrence from the Great Depression to the Global Financial Crisis (GFC) and post-pandemic period, highlighting key characteristics and changes over time.
Building on the historical context we Introduce the influential Diamond-Dybvig model of bank runs. We explore its theoretical framework for explaining bank runs and the implications thereof.
We briefly mention other bank run modelling paradigms highlighting their advantages and disadvantages.
This leads to the proposal of expanding the Diamond-Dybvig model using agent-based modelling (ABM) to account for heterogeneity among market participants. This model is the subject of an ongoing PhD study utilizing ABMs to develop novel models for a deeper understanding of bank runs, emphasizing the ability to capture contagion effects, herd behaviour, and feedback loops at the systemic level.
The presentation ends with briefly commenting on the planned implementation of deposit insurance in South Africa and its likely implications for bank runs and depositor protection.
Outcomes that attendees can look forward to benefitting from include:
– A grounding in the history of bank runs and how it has evolved over time to the present day.
– An introduction to the influential Diamond-Dybvig model. This will lead to a greater appreciation of how maturity transformation creates the fundamental instability of the financial system.
– A grasp of different bank run modelling paradigms and their respective advantages and disadvantages.
– An introduction to new research on bank-run modelling. Specifically, how ABMs are used to capture complexity, heterogeneity and systemic effects
– An understanding of the likely effects of the new depositor insurance scheme in South Africa.
– An introduction to a powerful alternative modelling paradigm, namely ABMs, which can be applied to a range of other phenomena.
GETTING BACK TO BASICS – REDEFINING THE ROLE OF THE CRITICAL ILLNESS PRODUCT
Kadene Gibbs, Matthew Procter
Presentation
Relevant practice area: Life Insurance
Suggested audience knowledge level: Foundational
In 2019, there were more than 1,000 new projects established to focus on cancer research, and this is expected to increase over the coming years. This has resulted in significant advances in cancer screening, diagnoses and surgical procedures.
With the increased focus on cancer research, technologies such as liquid biopsy and molecular analysis – which allow for the practice of precision medicine – create effective treatment options for cancer patients and reduce the side effect of traditional treatments like chemotherapy.
Targeted treatments, while effective, are often times not covered by medical schemes and the public health service largely due to their cost.
With targeted treatment increasingly being used to treat a number of medical conditions, this presents an opportunity for the insurance industry to fill the gap and meet policyholders’ needs through its Critical Illness offering.
The role of the Critical Illness product has always been to meet the changing lifestyles and shortfall in medical expenses that one would expect as a result of a critical illness event. The SA market offers comprehensive Critical Illness products covering as many as 300+ conditions, with a number of these conditions acting as mere window dressing to the 4 major condition groups – cancer, stroke, heart attack and CABG. The Critical Illness product has, arguably, lost its way.
With cancer claims currently accounting for roughly 40% of the cost of the Critical Illness product, and with other conditions increasingly being treated with targeted treatment, simply increasing the current sums assureds to allow for targeted treatment would render the Critical Illness product unaffordable.
The session will cover our innovative design and pricing to a reimagined CI benefit that meets the needs of policyholders in the face of advancing technologies, all while maintaining the affordability of cover. The session will also provide our thoughts on how we can continue to innovate our CI offering so that the benefit remains true to its core purpose.
LOADSHEDDING WILL START IN 55 MINUTES – A SOUTH AFRICAN PROVERB
Zane Heyl
Presentation with Paper
Relevant practice area: Wider Fields
Suggested audience knowledge level: Foundational
South Africans faced 3,776 hours (4 straight months) of loadshedding in 2022 and we’re on track to beat that in 2023.
The design of the South African electricity grid is driven by a bimodal demand curve – most electricity demand happens in the mornings and in the evenings. Geysers drive a significant portion of the demand for electricity during those two peaks – the times of highest loadshedding stages. But, geysers do not have to be on continuously for people to have hot water when they want it.
Smart geyser devices allow customers to select when they want hot water and ensure that their geysers systematically turn on and off throughout the day. This saves money and ensures hot water when it’s wanted/needed. However, most geyser users would choose to have hot water at the same two times of the day: mornings and evenings.
A new startup company in South Africa has designed a new type of smart geyser device which considers each individual device as a participant in a broader network of same devices. This network benefit allows optimization of electricity demand for hot water by staggering that demand intelligently.
It therefore redistributes that demand to other times of the day without compromising on when people need hot water. At scale, this can alleviate 6,000 MWhs of demand from the grid, consistent with four stages of loadshedding.
Actuaries make a professional promise to consider and/or act in the public interest, and the energy crisis cannot be viewed is anything other than applicable.
REINSURANCE OPTIMISATION
Idelia Hoberg, Marguerite Fevrier, Celeste Rajool
Presentation
Relevant practice area: Life Insurance
Suggested audience knowledge level: Intermediate
In the South African non-life market, reinsurance is managed holistically and often using economic capital models to understand risk exposures. Reinsurance programs and structures are reviewed and challenged frequently. Reinsurance is a key tool in managing the business and capital.
In the life space, reinsurance is generally used in a less holistic way. It is mainly used to manage product risk exposures and as such the decision-making often sits within the pricing function. This does not enable the life insurers to implement optimal reinsurance structures, and requires insurers to revisit their approach to reinsurance as a whole.
In this presentation we will discuss the purposes of reinsurance, in particular the quantitative purposes, with examples. This will give the audience a sense of what the objectives of reinsurance may be.
We will then consider how life insurers can maximise the value they derive from reinsurance, by looking at different measures that can be implemented. These include establishing a best practice reinsurance strategy, holistically reviewing the existing reinsurance program, redefining the reinsurance operating model to be aligned to the reinsurance strategy, identifying and implementing optimal reinsurance structures and increasing the inwards reinsurance to be placed centrally.
We will go through all the steps to maximise the value life insurers can derive from reinsurance, making it practical for the audience to gain a clear understanding of how this may be achieved.
Practical outcomes:
1. Clarity on the different purposes reinsurance and reinsurance structures that can be used for those purposes.
2. An understanding of how life reinsurance may be used to manage risk and capital holistically.
3. Clarity on the steps life insurers can take to implement an optimal reinsurance strategy and target operating model.
4. Examples of how to identify optimal reinsurance structures and how these could work.
ACTUARIAL AND STATISTICAL PROBLEMS AROUND THE COVID PHENOMENON
Nick Hudson
Presentation
Relevant practice area: Healthcare
Suggested audience knowledge level: Intermediate
I’ll explore the many interesting challenges thrown up under the following headings, with a view to highlighting areas for future research and actuarial attention.
Analytics relating to pathogenic transmission
– Geospatial analysis and spread of infection
– Epidemiological modelling challenges
– Calibration for susceptibility and non-pathogenic factors
– Recognition of heterogeneity
– Accommodating policy responses
Mortality and morbidity attribution
– Bayesian analysis and syndemic characteristics
-Challenges in assessing fatality rates
-Interpreting test data
Evaluating policy responses
– Evaluation methods
– Assessment of modalities
– Controlling for confounding factors in observational studies
– Evaluating experimental design in randomized control trials
Pandemic preparedness
– Assessing likelihood of pandemics meriting response
– Evaluation of opportunity costs and adverse effects of policy
THE POWER OF PUBLIC PRIVATE PARTNERSHIPS
Johan Human, Jolly Mokorosi, Ndabe Mkhize, Theo Winter
Panel Discussion
Relevant practice area: Wider Fields
Suggested audience knowledge level: Intermediate
This topic looks at the importance of public private partnerships in solving our energy (power) crisis. Pension funds hold significant capital, and private equity funds can set up the necessary structures (e.g. private credit and private debt) to fund private sector investment into this space.
Practical outcomes:
– Demystifying the crisis in terms of the energy availability factor, etc.
– Outlining how deals can be structured from a funding perspective
– Establishing the size and needs of the market (which sectors, etc.)
– Understanding the concerns that prevent pension funds from entering into this space, and what their needs are (typically, pension funds want investments that generate cashflows for pensions in payment)
ADAPTING TO DISRUPT: EXPLORING THE DIGITAL READINESS OF INSURERS IN THE AGE OF NON-TRADITIONAL COMPETITION
Nida Ibrahim, Ashish Desai, Jodi Christensen
Presentation
Relevant practice area: Short-term Insurance
Suggested audience knowledge level: Foundational
The Insurance sector has historically been dominated by large, traditional insurers who have built up significant brand power in this space. More recently, we’ve seen a wave of innovation in this space, largely from less traditional players, such as bancassurers, retailers and telecommunication operators. There are a number of reasons as to why these players are managing to disrupt this space and acquire market share. A case study will be used to appropriately set this scene.
For these players, this ability to use technology as the enabler to innovate is facilitated by the rapid and continuous improvements we are seeing in technology. So how do players in the insurance market who are not set up to employ these sorts of strategies navigate this digital turbulence?
The following questions arise from looking at the current landscape:
- In what ways do insurers need to digitally transform to retain or grow their market share?
- How do they ensure digital readiness?
- What role do actuaries play?
Answering these questions forms the basis for this presentation. We will look at the common barriers to digital transformation, and their state of digital readiness, by performing research and providing our own insights into the digital maturity of insurers in the market. We propose the idea of disrupting on the edges, which affords insurance players the opportunity to test new markets, new channels, new products or new servicing mechanisms without major investment or changes to their original systems and processes.
The outcomes attendees can expect to benefit from:
– Gaining industry knowledge, a deeper understanding of how specific digitally-focused insurance players are managing to disrupt the insurance market
– A holistic understanding of what digital readiness for an organisation means
– An overview of the most common organisational barriers to digital readiness, looking most specifically at operational and culture-related factors
– Sight of a high-level assessment of the digital maturity of several organisations in South Africa, which provides context for where one’s own organisation is placing in the race to be digital
– Gain familiarity with the concept of disrupting from the edges, which enables organisations to test new markets or ideas
– Clarify the role actuaries can play in supporting an organisation’s digital strategies through the influence they possess
ARE WE SNIFFING GOM? SURELY WE DON’T NEED A SEPARATE LICENSING CATEGORY FOR MICROINSURERS!
Tashvir Kalawan, Tsz-Yin Chen, Bonolo Pelele, Keith Mberengwa
Facilitator: Melusi Baloyi
Debate
Relevant practice area: Microinsurance
Suggested audience knowledge level: Foundational
Proposal:
There will be two teams of two members each.
One will argue that microinsurance as a regulatory catagory (alongside traditional insurers, branches and groups) is not necessary and that microinsurers might as well be regulated as a traditional insurer.
Arguments will include:
– Patronising
– IFRS17 requires sophisticated actuarial techniques in any case
– Very little difference between GOM and GOI
– Very little difference between FSM and FSI
The other will argue that:
– End result of a long consultation process with industry
– Worldwide trend
– Future developments may provide additional need for separate regulation
– Differences between GOM / FSM and GOI/FSI material
– Restriction on new entrants (capital requirement)
Practical outcomes:
– Increasing awareness and understanding of difference in regulations for traditional and microinsurers.
– Better understanding of whether these difference are intended / necessary / empowering / limiting / desirable.
– Improved industry practices by identifying gaps and inefficiencies in the current regulatory environment for microinsurers.
– Potential eventual regulatory changes to address inconsistencies between the microinsurance and traditional regulatory practices and standards.
– Identification of gaps in knowledge and highlighting areas that require further research and education.
METAMORPHOSIS: UNLEASHING THE POWER OF ACTUARIAL TRANSFORMATION
Shaun Lawton
Presentation
Relevant practice area: Wider Fields
Suggested audience knowledge level: Foundational
In a world where transformation is the key to survival and success, “Metamorphosis: Unleashing the Power of Actuarial Transformation” offers a roadmap for actuaries to navigate the transformative journey, embrace innovation, and shape a prosperous future for themselves and their organizations as the next steps in the journey is to transform and modernise the actuarial performance and processes of your teams.
Embracing transformation is essential for growth and success. This presentation explores the concept of actuarial transformation (aka finance transformation) and its significance in driving organizational effectiveness and meeting evolving demands. Drawing from real-world experiences and practical insights, we uncover key strategies and lessons for unleashing the power of actuarial transformation. The need for actuarial transformation in a rapidly changing business environment is growing and emphasizes the importance of staying ahead of the curve and adapting to emerging trends and challenges. Actuarial professionals must recognize that transformation is not an option but a necessity to thrive in an ever-evolving industry to get away from playing catch up.
Many finance transformation project fail, or only deliver a part of what they originally set out to do. The focus area will be to delve into practical lessons learned from successful actuarial transformation projects – or in some cases, to look at what some projects did not succeed.
By the end of the presentation, participants will be equipped with actionable outcomes and practical strategies to initiate and drive actuarial transformation within their organizations. They will leave with a clear understanding of the challenges and opportunities associated with transformation and the confidence to embrace change, adapt with agility, and unlock the full potential of their roles in the actuarial profession.
This includes
– gaining a clear understanding of the challenges and opportunities in actuarial transformation.
– learning some practical strategies for gaining buy-in and aligning individual goals with transformation objectives.
– gaining insights into resource allocation, including leveraging contractors for BAU tasks and retaining expertise.
– exploring the potential of their existing actuarial tools through comprehensive training and optimization.
ASSA SHOULD ACTIVELY PARTICIPATE IN BROAD PUBLIC POLICY ANALYSIS AND DEBATE
Simon Louw, Lusani Mulaudzi
Moderator: Nalen Naidoo
Debate
Relevant practice area: Wider Fields
Suggested audience knowledge level: Foundational
Serving the public interest is seen as a cornerstone of professions. There are many ways to achieve this goal, one of which is to contribute to public policy. In 2019 ASSA appointed a Public Interest Actuary to drive ASSA’s involvement in public policy, with a mandate to explore this beyond the traditional financial services sector. This lead to the creation of ASSA’s Public Interest Strategy 2021 -2024 which takes a broad view of public policy involvement. Why is this so? Should ASSA be involved in this way? While there are opportunities, there are also significant challenges.
Join us as we debate this topic. The motion is: ASSA should actively participate in broad public policy analysis and debate. One team, lead by Lusani Mulaudzi, will argue for the motion. Another team, lead by Simon Louw, will argue against the motion. A chair will moderate the debate, including contributions from the audience, and will determine the winner of the debate.
What you can expect to gain from this debate:
– A deeper appreciation of the public interest work of ASSA and the Public Interest Actuary;
– An understanding of the opportunities and challenges this work brings;
– An opportunity to add to the debate;
– Some fireworks (not really, but it should be lively).
PROFESSIONAL JUSTICE
Mickey Lowther, Jonathan Mort
Presentation with Paper
Relevant practice area: Professional Matters
Suggested audience knowledge level: Foundational
This paper firstly looks at the principles of ‘professional justice’ as opposed to criminal or contractual justice, and then asks what the profession wants to achieve within this third strand (‘professional oversight’) of our professional promise. Lowther and McMillan (2006) explored the concept of professionalization of actuaries: how an actuarial association chooses to offer different types and amounts of strands to its public from time to time. Accordingly, ASSA, like many other National Actuarial Associations, (Louw and Mulaudzi, 2023) acknowledges a responsibility to the public interest – but takes on a grander public interest role for the Society itself with its own agency, as opposed to a more limited role for individual members. The Code of Professional Conduct, at #25, specifically confirms that a member meets public interest requirements by adhering to the Code and related standards. From this viewpoint then the aim of the Process is for ASSA to demonstrate to its public, and the members, that members will be held to the agreed-on standards. It is not a criminal process, nor to do with compensation.
This limited purpose is possible, inter alia, because ASSA is self-regulated. Sanders (20XX), a former financial services regulator and Australia’s ‘Chief Professionalist’, reflects on how professions are so much more effective when self-regulated, because of the complexities which will only increase in the future as fields of practice become wider.
The project will examine some comparative codes in other national actuarial associations and other professions. In particular the Institute and Faculty of Actuaries in the UK have substantially changed their disciplinary process, using ‘upholding the reputation of the profession’ as their yardstick. Their revised process focuses on material acts or omissions. It also follows improved regulatory practice, an area where professions have often been criticised.
Thirdly, and within the limits of confidentiality, the research will analyse recent statistics of complaints. A similar exercise examining the Institute of Chartered Accountants in Ireland (Canning and O’Dwyer, 2001), provides a useful framework as well as comparative data.
Practical outcomes that we would anticipate include:
– Contributing to a logical and fit for purpose disciplinary process
– Raise the awareness of how the AGB serves the members
– Member buy-in to the disciplinary process
– Clarifying issues such as ‘when is an actuary an actuary, and subject to discipline?’
– Highlighting related changes that may need to be made to the Code of Professional Conduct
ADVANCED APPROACHES TO ANALYSE DISABILITY INCOME EXPERIENCE USING EXPERIENCE FROM CANADA, THE UK, SOUTH AFRICA, AUSTRALIA AND NEW ZEALAND
Koop Lubbinge, Bruce Robertson, David Hatherell, Adele Groyer, Louis Rossouw
Presentation with Paper
Relevant practice area: Life Insurance
Suggested audience knowledge level: Intermediate
Present on paper:
This will be a presentation of the paper titled Advanced Approaches to Analyse Disability Income Experience Using Experience from Canada, the United Kingdom, South Africa, Australia and New Zealand
Prepared by David Hatherell, Adele Groyer and Louis Rossouw
First presented to the 2023 International Congress of Actuaries
Paper abstract:
Disability income insurance is sold across multiple markets with different product features.
Product design, as well as other environmental factors, can have a significant impact on claims costs.
This study was prompted by discussion of the challenges faced by the Australian market.
Gen Re has collected disability income experience data across various countries.
We compare the product features across Canada, the United Kingdom, South Africa, Australia and New Zealand.
We use data analytics in an attempt to understand which factors have the greatest impact on claims costs.
We show high level actual vs expected results for incidence and terminations across companies.
The main aim of this project is to enable different countries to learn from one another’s experience as most other studies focus on single markets.
We acknowledge and discuss the complexity of this given the variation in multiple factors across countries.
Audience Outcomes:
The audience will get insights on:
– Product designs across markets
– Experience across markets based on data from 20 insurers across 5 countries.
– Drivers of experience and analytics techniques used to analyse experience
THE ROLE OF ACTUARIES IN A PANDEMIC: OVERVIEW OF THE ASSA COVID-19 TASK TEAM WORK
Claiton Manikai, Adam Lowe, Karsten Roux, Siebert Benade, Roseanne Harris
Presentation
In this session we will review the work of the ASSA COVID-19 Task team which covered a review of the COVID-19 pandemic including the impact on mortality and life insurance, heath, and the market conduct issues arising.
The work made it clear that:
• COVID-19 was a deadly pandemic killing many South Africans.
• Some societal responses assisted in reducing the spread, and vaccines protected from severe outcomes of the disease.
• Life insurers also saw significant claims due to the pandemic. Medical schemes saw reduced impacts on claims during the pandemic.
• There are market conduct lessons to be learned from the pandemic.
The presentation will bring the international actuarial perspective, based on IAA papers on lessons learned from the pandemic.
There also remains lots of future work to be done with regard to the ongoing impacts on COVID-19, for example on mortality directly, long-COVID, an analysis of COVID-19 outcomes for HIV lives, as well as through delayed diagnoses and treatment of HIV/AIDS and cancer, for example. The health stream is also embarking on significant research using medical scheme data. The Society also needs to look to the future as we may need to revise our views and how we operate when we think of future pandemics.
Further reading:
Market Conduct matters faced by the Actuarial Profession relating to COVID-19
Life Insurance WS Final report
ENHANCING AGGREGATE RESERVING TECHNIQUES THROUGH CLUSTERING-BASED SEGMENTATION
Stephan Marais
Presentation with Paper
Relevant practice area: Short-term Insurance
Suggested audience knowledge level: Intermediate / Advanced
Traditional aggregate reserving techniques assume that each data triangle used is composed of a relatively homogenous group of claims. However, the practical constraints of management and reporting requirements often hinder the adoption of a statistically optimal approach to segmentation. There is often a reliance on legacy class hierarchies, business processes or expert judgement due to the difficulty of determining an optimal segmentation. This can reduce homogeneity in triangles and affect the performance of aggregate reserving techniques.
To address these challenges, this paper proposes a new framework and methodology for automated clustering to allocate individual claims to homogenous sub-groupings suitable for projection with triangle reserving techniques. The proposed approach entails the following:
– Separation of open claims from closed claims to prevent their unknown development from being a source of bias in the clustering process.
– Definition of claim features that describe the development of individual closed claims to optimise the accuracy of reserve estimations produced by triangle reserving methodologies. Features are adjusted for trends, where necessary.
– Implementation of a clustering method on all closed claims for a given class.
– Allocation of open claims to the clusters by finding the most similar closed claims clusters, based on operational time bands which reflect claim characteristics as they vary over time.
– A testing methodology to determine if the clustering method improves the accuracy of reserve estimates.
– Practical considerations for constructing models in a way that allows for choice of segmentation method and for managing segmentation changes across reporting periods.
This paper includes an illustrative example of closed claims clustering employing the K-means algorithm. This includes the calculation of claim features, determining the optimal number of clusters, and the application of Principal Component Analysis to reduce the number of variables explaining the distance between data points.
Overall, by providing both a methodology and a practical framework for leveraging clustering techniques as part of a reserving process, this research aims to improve accuracy and reduce the need for intervention for actuaries managing insurers’ claims reserves.
EMBEDDING ORSA: EXPLORE IT AND DISCOVER PRACTICAL IMPROVEMENT IDEAS
Neels Maritz
Presentation with Paper
Relevant practice area: Enterprise and Financial Risk Management
Suggested audience knowledge level: Intermediate
– Insurance Act 18 of 2017 became effective on 1 July 2018. It became the new legal framework for the prudential regulation and supervision of insurance business in South Africa. Insurance business is life insurance, non-life insurance or reinsurance business.
– As per the Act, the Prudential Authority (PA) may prescribe governance principles and requirements relating to risk management, including in respect of Own Risk and Solvency Assessment (ORSA).
– 77% of insurers self-assessed their ORSAs as weak or needing improvement in 2014, just a few years preceding the commencement of the new Act.
– ORSA addresses foreseeable material risks and assesses the insurer’s solvency position projected for 3-5 years based on its business plan and specific risk profile, across a range of scenarios. The essence of ORSA is to protect policyholder interests.
– Prudential Standard GOI 3.1 requires insurers to conduct ORSA on at least an annual basis. It specifies:
o An insurer must be able to demonstrate that ORSA is widely used
o ORSA outputs should be embedded into strategic decision-making processes, especially with respect to capital planning and management. Management actions should be linked to the outputs of the ORSA.
– A holistic ORSA process consists of various integrated sub-processes across Actuarial, Risk, Finance and Operations. The paper explores differences between an insurer’s ORSA and Solvency Capital Requirement (SCR).
– Embedding ORSA is a necessary requirement, but there is no industry framework to inform how to better embed ORSA. Neels is experienced in the field of ORSAs, and explores the concept of embedding ORSA by researching South African and international publications.
– Insurers are affected by tumultuous times and risks evolve daily in globally connected markets. It is noted that ORSA is required in jurisdictions across the world including Africa, Europe, America and Asia.
The paper/presentation offers the following practical and interesting outcomes:
– An understanding of what embedding ORSA means, including ChatGPT’s understanding
– How an insurer can better embed ORSA
– New risks for insurers, as well as how ORSA can support with risk management
– An original ORSA process diagram, showing various integrated sub-processes
– Practical ideas to better embed each sub-process
– The governance role of actuaries in embedding ORSA
– Key lessons and themes to consider when embedding ORSA.
EPIGENETICS: THE HIDDEN KEY TO UNDERSTANDING HEALTH AND INSURANCE RISK?
Christelle Oosthuizen, Poonam Vala
Presentation
Relevant practice area: Life Insurance
Suggested audience knowledge level: Foundational
Epigenetics refers to changes in gene expression that occur without altering the primary DNA sequence.
Genetics is the study of heredity and collectively our genes are made up of DNA which is in turn made up of molecules A, G, C and T, and it is the specific arrangement of these molecules that are interpreted as genes.
The word epigenetics comes from the Greek word epi, meaning above and gene, to literally mean anything on top of or above the level of a gene.
The biological purpose of epigenetics is to control gene expression, in other words to turn genes on or off. Whereas our DNA or genetics are static, in other words, they do not change throughout the course of our life, our epigenetics will change.
By obtaining epigenetic information from a non-invasive saliva sample, an insurer can:
– attain information about an insured’s current health and wellness statuses
– resulting from their environment and behaviour (for example diet, exercise, alcohol, and drug use),
– while steering clear of the perceived unfairness of unchangeable genetic information.
It can also contribute to accurate risk classification, an alignment of interests between insureds and insurers while upholding the essential regulatory norms of information symmetry and transparency.
In this presentation spanning life and healthcare practice areas, we explore:
- the ethical, legal, and social implications for life insurers and healthcare funders when using epigenetics:
o to enhance the accuracy of underwriting, pricing, and managing insurance risk; or
o for targeted managed care interventions
o (to name a few uses).
Practical outcomes for the audience include an understanding of:
– What epigenetics is, how it differs from genetics, and why the distinction between genetics and epigenetics is important.
– What your epigenetic clock refers to, how it is determined and how it can be used by insurers.
– How epigenetics affects your health outcomes, i.e., to what extent your health can be affected by environmental and behavioural factors.
– How your epigenetics can be tested, the accuracy and costs of epigenetics tests, and how the possible benefits for insurers compare to the additional costs of using epigenetics.
– Possible applications for life insurers and healthcare funders, how the applications differ between them, and risks involved.
– Case studies of insurers that have tried to or are currently using epigenetics, which demonstrate the practical uses of epigenetics in the life and healthcare industry.
METASURANCE
Neil Parkin
Presentation
Relevant practice area: Life Insurance
Suggested audience knowledge level: Foundational
In the early 1990s, the internet seemed like an odd place. A place for young hackers or computer geeks, but with little significance to the mainstream world. After all, who would dare input credit card details or trust someone on the web? Most people would have scoffed at the idea that this new technology would be powerful enough to change the way future generations communicate, date, buy groceries and even watch TV. Let alone have the power to change governments, affect elections and even the narrative of wars.
We again find ourselves at an inflection point: the rise of a new internet. Welcome to the metaverse. Like the internet 30 years ago, it’s hard to understand this abstract notion and it’s easy to dismiss it as a gamer’s fantasy. However, the changes we will see are much deeper than a Virtual Reality headset, but extend into immersive environments, digital payments, virtual healthcare, decentralisation and a shift in the ownership of data to individuals. The very way we work and engage each other will fundamentally change.
Join me as I discuss the technologies being developed and the challenges and opportunities these present for life insurers across the insurance value chain, from marketing and distribution through to pricing, underwriting, claims and consumer engagement. I’ll share my thoughts on what Metasurance could look like, and how to prepare for the real digital revolution within our industry.
The audience can expect:
– An entertaining and thought provoking presentation
– Insight into technological shifts, including how the current generative AI explosion will feed into these broader changes
– What this all means for the insurance market and our customers
– Apart from the potential, also a review of the risks, including social equity
Finally, participants will walk away with ideas on how we should be thinking about the changing techno-socio environment from an insurance strategy point-of-view.
ACTUARIAL GENERAL INTELLIGENCE? EXPLORING THE POTENTIAL IMPLICATIONS OF LARGE LANGUAGE MODELS (LLMs) ON ACTUARIAL EDUCATION AND LEARNING
Maximilian Popescu, Francois Berry
Presentation
Relevant practice area: Education
Suggested audience knowledge level: Foundational / Intermediate
A very short introduction will be given to the latest developments in Large Language Models (LLMs) with a non-technical explanation of what they are (and what they are not)
A review of some of the tests that have been performed by LLMs on actuarial exams, contextualising results by comparison to other professions
One novel test on GI Fellowship, including the addition of prompt engineering and (possibly) additional relevant training data
Implications of the results on actuarial education and examinations. This will be contextualised against what is happening in other professional areas
Ideas on how LLMs (and applications built on top of these) can be used to augment actuarial learning outcomes and improve examination candidates’ results
Views of the audience of the usefulness of LLMs, as well as the potential impact of these tools on both actuarial education and learning
REFLECTING ON IFRS 17 IMPLEMENTATION: WHAT WENT WELL? WHAT DIDN’T GO AS EXPECTED? AND WHATS NEXT?
Thato Pule, Karen Muyengwa, Lufuno Netshitenzhe
Pre-recorded Oral Presentation
Relevant practice area: Life Insurance
Suggested audience knowledge level: Foundational
We all should have know that a standard that took decades to complete, would be no walk in the park. As some insurers were making progress, the IASB pushed back the effective date from 1 January 2021, to 1 January 2022 and eventually 1 January 2023. This was another sign that this standard was not a small change and that there would be a story to tell at the end of it all.
We are not quite at the end of it all but as of today, IFRS 17 is effective.
Over the last few years, we have been involved in multiple implementation projects from a consulting perspective, from an industry role perspective and wearing an audit hat, across local and global insurers. This presentation will go into what we have learnt from these experiences:
- What went well? We will outline the elements of the implementation projects that were a success, for example, leveraging off the work done implementing SAM / Solvency II (where possible), minimising the need to reinvent the wheel and therefore, effort required.
- What didn’t go well? We will outline the elements of the implementation projects that didn’t go as expected, for example, late auditor engagement which resulted in undue stress if auditors had opposing views and it may have been too late to start from scratch.
- What’s next? We will outline the opportunities that exist for insurers on the back of the work done to get IFRS 17 compliant, for example, the opportunities to optimise the actuarial and finance reporting processes.
We want the audience to take from the presentation:
- An opportunity to take a step back and reflect. A large part of the community involved in IFRS 17 implementation has been focused on getting over the line, we want the audience to take a few minutes and think about their experience on their implementation.
- To reflect with us on their own implementation experience, which “mistakes” did they also make? Which successful elements did they also implement?
- For this to be a learning opportunity for the next big change. Some of what didn’t go well with implementation, could have been avoided.
- To take off their implementation hat for a few minutes and give some thought to the “after IFRS 17 world” with respect to the opportunities to leverage off the work done to optimise their processes and functions.
DESIGNING A MARKET CAPITALISATION-WEIGHTED BENCHMARK FOR DOMESTIC EQUITY INVESTORS
Mark Randall, Davy Corubolo
Presentation with Paper
Relevant practice area: Investments
Suggested audience knowledge level: Foundational
An appropriate index is an essential component of an effective capital market, particularly for benchmarking investment performance for general equity portfolios. The paper will review the properties of an ideal benchmark. In the South African asset management industry, equity investors tend to use the similar indices as their performance benchmarks. These indices influence stock allocation, performance fees and a range of listed and unlisted instruments, not to mention the public discourse.
Traditionally, market capitalisation has been used as the appropriate weighting mechanism for broad market indices. Trends in index development have led to alternative weighting mechanisms and alternative index structures, which will be listed in this paper. The paper will consider the merits of market capitalisation weighted indices and will address challenges in constructing South African market indices, including debates over the definition of market capitalisation, questions around the inclusion of dual listed and foreign companies, and the impact of non-resident participants that can significantly alter the benchmark composition. Furthermore, capital markets in developing economies are prone to concentration risk, and this raises the question of applying weighting caps to dominant index constituents.
This paper considers the construction options for market capitalisation weighted benchmarks for domestic equity investors. The authors consider the existing FTSE/JSE All Share Index variants, and present three alternative constructs:
- Global weighting approach
- SA Register approach
- SA Resident approach
Analysis of these alternatives include various capping scenarios to deal with single name concentration.
Attendees can expect to gain foundational knowledge on the following:
– An understanding of various approaches in equity index design and issues in selecting a suitable benchmark
– An understanding of the local listed equity market landscape, including analysis by dual listed status, company domicile, investor domicile, size and liquidity distribution, supporting an analysis on what the investment opportunity set is for local equity investors.
– An investigation into dual listed instruments, including topics of fungibility and register transfers
– An investigation into demand drivers for locally listed equity, particularly from non-resident investors, and the extent that this should impact on the local benchmark
– An explanation of the idiosyncrasies inherent in the existing FTSE/JSE All Share market benchmarks
– An appreciation of how various approaches to defining market capitalization can influence the structural composition of a benchmark, and a comparison of the resulting index metrics
– An analysis on benchmark concentration, and the impact of applying caps on constituent weightings
CLAIMS MODELLING WITH CLIMATE DATA
Ronald Richman, Rendani Mbuvha, Kovlin Perumal, Adam Balusik, Caesar Balona
Presentation
Relevant practice area: Short-term Insurance
Suggested audience knowledge level: Intermediate
Traditional short-term (or property and casualty) insurance pricing involves the use of statistical techniques such as Generalised Linear Models and policyholder specific rating factors to estimate the expected claims cost associated with a relevant risk profile. In recent years, the short-term insurance industry has been increasingly impacted by weather related events, which are expected to be exacerbated by both climate change and meteorological phenomena such as the El Niño–Southern Oscillation (ENSO) La Nina. Consideration of this risk has traditionally been captured in the actuarial pricing process indirectly through the use of geographical features such as area codes. In this working paper, we directly link an exposure dataset of buildings risk and associated claims experience to a high-resolution gridded precipitation dataset to investigate the predictive power of this feature on both an actual and forecasted basis. We establish a modelling framework that allows for the estimation of both the frequency and severity of buildings claims given the combined dataset. We consider the relative importance of the added precipitation feature against traditionally used rating factors and the sensitivity of the underlying claims frequency and severity to changes in precipitation. Finally, by considering different precipitation scenarios, we show how the risks of excessive precipitation can be quantified, allowing for more accurate forecasts of financial performance to be made, and risk mitigation strategies to be investigated.
Outcomes
The audience will gain a good understanding of weather factors that can influence short term insurance claims and how these can be analysed in an actuarial framework.
SMOOTHNESS AND MONOTONICITY CONSTRAINTS FOR NEURAL NETWORKS USING ICEnet
Ronald Richman, Mario Wuthrich
Presentation with Paper
Relevant practice area: Shirt-term Insurance
Suggested audience knowledge level: Intermediate
Abstract: Deep neural networks have become an important tool for use in actuarial tasks, due to the significant gains in accuracy provided by these techniques compared to traditional methods, but also due to the close connection of these models to the Generalized Linear Models (GLMs) currently used in industry. Whereas constraining GLM parameters relating to insurance risk factors to be smooth or exhibit monotonicity is trivial, methods to incorporate such constraints into deep neural networks have not yet been developed. This is a barrier for the adoption of neural networks in insurance practice since actuaries often impose these constraints for commercial or statistical reasons. In this work, we present a novel method for enforcing constraints within deep neural network models, and we show how these models can be trained. Moreover, we provide example applications using real-world datasets. We call our proposed method ICEnet to emphasize the close link of our proposal to the individual conditional expectation (ICE) model interpretability technique.
Outcomes
1) The audience will understand different methods for applying machine learning interpretability techniques
2) We will illustrate that neural networks may produce predictions that do not follow intuition
3) The audience will understand the new proposed method for solving this problem, and how predictions can be made to follow commercially sensible rules
4) The audience will gain knowledge on how machine learning models can be modified for practical use
PROTECTION AND VALUE OFFERED BY LIFE INSURANCE PRODUCT REWARDS SCHEMES
Lafras Eksteen, Riaan van Reenen, Nathea Nicolay, Clyde Parsons, Andrew Warren
Debate
Relevant practice area: Life Insurance
Suggested audience knowledge level: Intermediate
This topic is being submitted by the LAC Market Conduct Subcommittee, with the objective of stimulating discussion around market conduct topics within the profession. The debate focuses on the intersection between life insurance products and rewards programmes.
Loyalty, wellness and cross-product schemes (rewards schemes) are becoming a feature of the pure risk life insurance landscape, both locally and globally. These schemes are innovative and have been effective at driving positive behaviour amongst its participants. They generally operate outside the insurance contract, but observed behaviour (gathered through these schemes) is used to provide discounts on life insurance products. These discounts provide more affordable insurance for those whose scheme scores indicate desirable outcomes, e.g. better health metrics, but the application of these schemes to life insurance products is not constrained by the same guidance, regulation and governance requirements as life insurance premium reviews. This session will see two teams of experienced actuaries debate for and against the following proposition:
Reward schemes, despite not being subject to the same governance requirements as life insurance premium reviews, provide more value to policyholders than to shareholders.
Practical outcomes:
– Gain insight into the impact of rewards programmes on life insurance products
– Understand the interaction of rewards programmes with the governance structures of life insurance products, in particular those relating to premium reviews
– Understand the benefits and risks of rewards programmes to shareholders
– Understand the benefits and risks of rewards programmes to policyholders
CLIMATE RISK STRESS TESTING IN BANKING AND GENERAL INSURANCE
Christopher Sterley, Jerome Mahadeo
Presentation
Relevant practice area: Wider Fields
Suggested audience knowledge level: Foundational / Intermediate / Advanced
The aim of this presentation is to give insight into:
– How to conduct a Climate Risk stress test.
– Climate Risk measurement in banking and general insurance.
– The banking-insurance interplay.
– The need for a macroeconomic impact.
– Considerations in Physical-Transition Risk portfolio management.
The outcomes of this session include knowledge around how to conduct a Climate Risk stress test, as well as special considerations in applying a stress test in practice.
A GENERATIONAL LENS INTO TRANSFORMATION WITHIN ASSA: THE FUTURE LOOKS BRIGHT
Katlego Thaba, Jeanine Wilson
Panel Discussion
Relevant practice area: Transformation
Suggested audience knowledge level: Foundational
This panel discussion is a continuation of the Transformation, Diversity and Inclusion series that we have hosted at the last four editions of the ASSA Convention.
The most recent session, in which we hosted a discussion on the topic “Gender and Sexual Diversity within the Actuarial profession, was very well attended in Cape Town and proved to be quite impactful both at the Convention and with subsequent social media conversations/feedback around the talk.
For the 5th edition of the Transformation series, this year’s topic will include a panel of individuals from various generations within the actuarial profession, who will share their perspectives in relation to transformation both in ASSA and within the South African society.
The panel will be constituted as follows:
– Young member (aged 20-25)
– Experienced member (aged 30-40)
– Older member (aged 50-60)
– Retired member (aged 70+).
Practical outcomes:
– We will explore how each of the panellists have been impacted positively yet differently by the rapid transformation journey in our profession, our community and our country
– We will engage on what main areas and drivers of transformational impact has been, where transformation has been evident
– We will unpack the panellists role and their aspirations for the actuarial society as pertains to matters of transformation.
MEASURING AND MANAGING SYSTEM RISK FOR PENSION FUNDS – A SOUTH AFRICAN PERSPECTIVE
Qhivi Tiva, Kelebogile Moloko
Pre-recorded Oral Presentation
Relevant practice area: Investments
Suggested audience knowledge level: Intermediate
The current global macroeconomic environment of high interest rates and high inflation poses increased systemic risk to the global financial system.
The topic aims to highlight system risk for pension funds by providing a South African perspective. This is done by sharing some of practical considerations and challenges of managing LDI portfolios in the current market and then looking at system risk in pension portfolios using a principal component analysis.
Furthermore, lessons from UK LDI crisis are shared highlighting the impact of systemic risk in sovereign debt markets.
SVB COLLAPSE AND OTHER BANK FAILURES
Koketso Mano, Parin Gokaldas, Michael Tichareva
Moderator: Lian van Oudheusden
Panel Discussion
Relevant practice area: Banking
Suggested audience knowledge level: Foundational
Silicon Valley Bank (SVB) was founded in 1983 and provided financing for nearly half of United States (US) venture-backed technology and healthcare companies that listed on stock markets in 2022. On 10 March 2023, SVB failed following a run on the bank. SVB’s failure was followed by the closure of Signature Bank and Silvergate Bank, both of which served the cryptocurrency industry, and by the collapse of Credit Suisse, amid liquidity challenges and risk management failures.
The panel discussion will unpack the events leading to the abovementioned bank failures and associated root causes. Focus will be placed on understanding the risk management and risk culture themes underlying the failures and on key lessons that banks and other financial services firms can apply within their own organisations.
APPLICATION OF DEEP REINFORCEMENT LEARNING IN ASSET LIABILITY MANAGEMENT
Takura Asael Wekwete
Presentation with Paper
Relevant practice area: Data Science
Suggested audience knowledge level: Intermediate
Asset Liability Management (ALM) is an essential risk management technique in Quantitative Finance and Actuarial Science. It aims to maximise a risk-taker’s ability to fulfil future liabilities. ALM is especially critical in environments of elevated interest rate changes, as has been experienced globally between 2021 and 2023.
Traditional ALM implementation is still heavily dependent on the judgement of professionals such as Quants, Actuaries or Investment Managers. This over-reliance on human input critically limits ALM performance due to restricted automation, human irrationality and restricted scope for multi-objective optimisation.
This paper addressed these limitations by applying Deep Reinforcement Learning (DRL), which optimises through trial, and error and continuous feedback from the environment. We defined the Reinforcement Learning (RL) components to optimise for ALM. These include the RL Agent, Environment, Actions, States and Reward Functions.
The results demonstrate that implementing DRL provides a superior approach compared to traditional ALM. DRL allows for increased automation, flexibility, and multi-objective optimisation in ALM, reducing the negative impact of human limitations and improving risk management outcomes.
The study shows that DRL can be used both at a given time point and dynamically on a high-frequency basis. DRL achieves similar levels of general duration matching and significantly better dynamic duration matching compared to traditional approaches, with fewer theoretical assumptions and restrictions.
The findings and principles presented in this study apply to various institutional risk-takers, including insurers, banks, pension funds, and asset managers. Overall, the application of DRL in ALM provides a promising avenue for improving risk management outcomes.
PRACTICAL OUTCOMES FOR THE AUDIENCE
– Deep Reinforcement Learning can be used to implement ALM to a similar level of performance compared to traditional approaches but with less reliance on theoretical assumptions.
– Reinforcement Learning ALM demonstrated better robustness to volatile market conditions compared to the traditional asset liability modelling approach due to less reliance on market assumptions.
– Reinforcement Learning ALM is performed within scalable computational frameworks which unlocks faster and more automated Asset Liability Management. This can enable one to practically apply Asset Liability Management on a high-frequency basis at an enterprise level with high-velocity and voluminous data.
– The work was a first step. There remain many potential areas of additional research such as implementation within legacy IT systems, circumventing real-world practical challenges, incorporating different data types and streamlining Reinforcement Learning ALM for ease of use.
TECHNICAL EXPERTS AS MANAGERS AND SERVANT LEADERS
Twané Wessels
Pre-Recorded Oral Presentation with Paper
Relevant practice area: Professional Matters
Suggested audience knowledge level: Foundational
The main aim of this study was to explore how a servant leadership style can assist technical expert managers to become more effective leaders. Secondary objectives of the research included exploring the leadership challenges that technical expert managers experience, the factors that make it harder or easier for a technical expert to be a servant leader and investigating to what extent servant leadership can be relevant. A qualitative design was followed. The sample comprised eight subordinates to technical expert managers and eight technical expert managers, and data were collected through semi-structured interviews.
Leadership challenges identified for technical expert managers included challenges in the interpersonal domain and limited time capacity. The findings of leadership challenges related to the interpersonal domain included difficulty relating to subordinates, egocentric thinking, and limited influence higher up in the hierarchy. There was also evidence of the paradox of power that may inhibit advice-taking by technical expert managers. Servant leadership may encourage delegation to empower subordinates and assist in mitigating the leadership challenges identified for technical expert managers in the interpersonal domain because servant leaders tend to operate from a stronger relationship domain.
Some factors may make it easier for technical experts as managers to assume a servant leadership style. Firstly, emerging technical experts look for a technical expert as a leader for guidance and mentorship. The legacy motive for many technical experts aligns with servant leadership in that it goes beyond just making a difference on a technical level to impacting people. Also, the example of servant leaders creates more servant leaders. Organisational culture and where an organisation is in its growth cycle can support or not support a servant leadership style. Finally, interpersonal and leadership skills improve with experience.
The research concluded that an appropriate balance between transformational and servant leadership might be best for a technical expert. If organisations can improve leadership efficiency, the organisation won’t just be an enjoyable place to work but can also gain a competitive advantage. Leadership awareness and development can start earlier in the technical expert’s education and career with an appropriate balance between theoretical and practice-based learning. Governing bodies of professions can make leadership a module alongside ethics in professionalism courses. Technical experts as managers can further be supported by leadership training, mentorship, coaching, and having an appropriate allocation between technical work and people management in their key performance objectives.
Keywords: Leadership challenges; effective leadership; servant leadership; technical expert; transformational leadership.