Category: General actuarial

Actuarial Professionalism Considerations for Generative AI

Mary Pat Campbell

Link: https://www.actuary.org/sites/default/files/2024-09/professionalism-paper-generative-ai.pdf

Graphic:

Excerpt:

This paper describes the use and professionalism considerations for actuaries using
generative artificial intelligence (GenAI) to provide actuarial services. GenAI generates text,
quantitative, or image content based on training data, typically using a large language model
(LLM). Examples of GenAI deployments include Open AI GPT, Google Gemini, Claude,
and Meta. GenAI transforms information acquired from training data into entirely new
content. In contrast, predictive AI models analyze historical quantitative data to forecast
future outcomes, functioning like traditional predictive statistical models.


Actuaries have a wide range of understanding of AI. We assume the reader is broadly
familiar with AI and AI model capabilities, but not necessarily a designer or expert user. In
this paper, the terms “GenAI,” “AI,” “AI model(s),” and “AI tool(s)” are used interchangeably.
This paper covers the professionalism fundamentals of using GenAI and only briefly
discusses designing, building, and customizing GenAI systems. This paper focuses on
actuaries using GenAI to support actuarial conclusions, not on minor incidental use of AI
that duplicates the function of tools such as plug-ins, co-pilots, spreadsheets, internet search
engines, or writing aids.


GenAI is a recent development, but the actuarial professionalism framework helps actuaries
use GenAI appropriately: the Code of Professional Conduct, the Qualification Standards
for Actuaries Issuing Statements of Actuarial Opinion in the United States (USQS), and the
actuarial standards of practice (ASOPs). Although ASOP No. 23, Data Quality; No. 41,
Actuarial Communications; and No. 56, Modeling, were developed before GenAI was widely
available, each applies in situations when GenAI may now be used. The following discussion
comments on these topics, focusing extensively on the application of ASOP No. 56, which
provides guidance for actuaries when they are designing, developing, selecting, modifying,
using, reviewing, or evaluating models. GenAI is a model; thus ASOP No. 56 applies.


The paper explores use cases and addresses conventional applications, including quantitative
and qualitative analysis, as of mid-2024, rather than anticipating novel uses or combinations
of applications. AI tools change quickly, so the paper focuses on principles rather than
the technology. The scope of this paper does not include explaining how AI models are
structured or function, nor does it offer specific guidelines on AI tools or use by the actuary
in professional settings. Given the rapid rate of change within this space, the paper makes no
predictions about the rapidly evolving technology, nor does it speculate on future challenges
to professionalism.

Author(s): Committee on Professional Responsibility of the American Academy of Actuaries

Committee on Professional
Responsibility
Geoffrey C. Sandler, Chairperson
Brian Donovan
Richard Goehring
Laura Maxwell
Shawn Parks
Matthew Wininger
Kathleen Wong
Yukki Yeung
Paul Zeisler
Melissa Zrelack

Artificial Intelligence Task Force
Prem Boinpally
Laura Maxwell
Shawn Parks
Fei Wang
Matt Wininger
Kathy Wong
Yukki Yeung

Publication Date: September 2024

Publication Site: American Academy of Actuaries

Actuarial ChatBots

Mary Pat Campbell

Link: https://riskviews.wordpress.com/actuarial-chatbots/

Graphic:

Excerpt:

Here are several examples of ChatBots and other AI applications for actuaries to try.

Answers that you might get from a general AI LLM such as ChatGPT may or may not correctly represent the latest thinking in actuarial science. These chatBots make an effort to educate the LLM with actuarial or other pertinent literature so that you can get better informed answers.

But, you need to be a critical user. Please be careful with the responses that you get from these ChatBots and let us know if you find any issues. This is still early days for the use of AI in actuarial practice and we need to learn from our experiences and move forward.

Note from meep: there are multiple Apps/Bots linked from the main site.

Author(s): David Ingram

Publication Date: accessed 28 Aug 2024

Publication Site: Risk Views

Regulatory Capital Adequacy for Life Insurance Companies

Mary Pat Campbell

Link: https://www.soa.org/4a194f/globalassets/assets/files/resources/research-report/2023/erm-191-reg-capital-with-final-visuals.pdf

Graphic:

Excerpt:

The purpose of this paper is to introduce the concept of capital and key related terms, as well as to compare and contrast four key regulatory capital regimes. Not only is each regime’s methodology explained with key terms defined and formulas provided, but illustrative applications of each approach are provided via an example with a baseline scenario. Comparison among these capital regimes is also provided using this same model with two alternative scenarios.

The four regulatory required capital approaches discussed in this paper are National Association of Insurance Commissioners’ (NAIC) Risk-Based Capital (RBC; the United States), Life Insurer Capital Adequacy Test (LICAT; Canada), Solvency II (European Union), and the Bermuda Insurance Solvency (BIS) Framework which describes the Bermuda Solvency Capital Requirement (BSCR). These terms may be used interchangeably. These standards apply to a large portion of the global life insurance market and were chosen to give the reader a better understanding of how required capital varies by jurisdiction, and the impact of the measurement method on life insurance company capital.

All of these approaches are similar in that they identify key risks for which capital should be held (e.g., asset default and market risks, insurance risks, etc.). However, they differ in significant ways too, including their defined risk taxonomy and risk diversification / aggregation methodologies, as well as required minimum capital thresholds and corresponding implications. Another key difference is that the US’s RBC methodology is largely factor-based, while the other methodologies are model-based approaches. For the model-based approaches, Solvency II and BIS allow for the use of internal models when certain conditions are satisfied. Another difference is that the RBC methodology is largely derived using book values, while the others use economic-based measurements.

As mentioned above, this paper provides a model that calculates the capital requirements for each jurisdiction. The model is used to compare regulatory solvency capital using identical portfolios for both assets and liabilities. For simplicity, we have assumed that all liabilities originated in the same jurisdiction as the calculation. As the objective of the model is to illustrate required capital calculation methodology differences, a number of modeling simplifications were employed and detailed later in the paper. The model considers two products – term insurance and payout annuities, approximately equally weighted in terms of reserves. The assets consist of two non-callable bonds of differing durations, mortgages, real estate, and equities. Two alternative scenarios have been considered, one where the company invests in riskier assets than assumed in the base case and one where the liability mix is more heavily weighted to annuities as compared to the base case.

Author(s): Ben Leiser, FSA, MAAA; Janine Bender, ASA, MAAA; Brian Kaul

Publication Date: July 2023

Publication Site: Society of Actuaries

Antique Insurance and Actuarial Books

Mary Pat Campbell

Link: https://thetermguy.ca/books.html

Graphic:

Excerpt:

The Term Guy’s hobby is collecting antique insurance books.  Here we’ve scanned many of our out of copyright books for your enjoyment and perhaps research purposes. Stay tuned, more books coming as I have time to scan them!

We have extracted table data from many of these books and made the information available as excel spreadsheets. In the download of spreadsheets we have also included a high def image of each of the pages containing the tables. The image filename for each page and the excel spreadsheet have the same name, i.e. image0001.jpg.xlxs contains table data from image0001.jpg. You may download and use the data unrestricted, but we would ask that you consider giving us a link from your website so that others can find this information as well.

Author(s):

Publication Date:

Publication Site:

Statement from Matthew Edwards

Mary Pat Campbell

Link: https://www.linkedin.com/posts/matthew-edwards-84a082145_actuaries-ifoa-activity-7089144129007300608-0zr2/?utm_source=share&utm_medium=member_desktop

Excerpt:

Attention #actuaries: with voting underway for the #ifoa Council elections, I share in this short video some thoughts and concerns about the Institute and Faculty of Actuaries. I touch on current problems, what I learned from my time with the Continuous Mortality Investigation and the COVID-19 Actuaries Response Group, and I stumble into the elephant in the room – the recent media coverage of the IFoA (with implications of procedural subterfuge).

Whoever you decide to vote for, please do take the time to engage with the election process, consider what each candidate can contribute to move the profession forwards, and whether you want an active council or a passive council: make your vote count for the sake of your profession.

Author(s): Matthew Edwards

Publication Date: 26 July 2023

Publication Site: LinkedIn

Public Sector Pensions in the United States

Mary Pat Campbell

Link: https://eh.net/encyclopedia/public-sector-pensions-in-the-united-states/

Excerpt:

Although employer-provided retirement plans are a relatively recent phenomenon in the private sector, dating from the late nineteenth century, public sector plans go back much further in history. From the Roman Empire to the rise of the early-modern nation state, rulers and legislatures have provided pensions for the workers who administered public programs. Military pensions, in particular, have a long history, and they have often been used as a key element to attract, retain, and motivate military personnel. In the United States, pensions for disabled and retired military personnel predate the signing of the U.S. Constitution.

Like military pensions, pensions for loyal civil servants date back centuries. Prior to the nineteenth century, however, these pensions were typically handed out on a case-by-case basis; except for the military, there were few if any retirement plans or systems with well-defined rules for qualification, contributions, funding, and so forth. Most European countries maintained some type of formal pension system for their public sector workers by the late nineteenth century. Although a few U.S. municipalities offered plans prior to 1900, most public sector workers were not offered pensions until the first decades of the twentieth century. Teachers, firefighters, and police officers were typically the first non-military workers to receive a retirement plan as part of their compensation.

By 1930, pension coverage in the public sector was relatively widespread in the United States, with all federal workers being covered by a pension and an increasing share of state and local employees included in pension plans. In contrast, pension coverage in the private sector during the first three decades of the twentieth century remained very low, perhaps as low as 10 to 12 percent of the labor force (Clark, Craig, and Wilson 2003). Even today, pension coverage is much higher in the public sector than it is in the private sector. Over 90 percent of public sector workers are covered by an employer-provided pension plan, whereas only about half of the private sector work force is covered (Employee Benefit Research Institute 1997).

Author(s): Lee A. Craig, North Carolina State University

Publication Date: 16 March 2003, accessed 8 Oct 2022

Publication Site: EH.Net Encyclopedia

Citation: Craig, Lee. “Public Sector Pensions in the United States”. EH.Net Encyclopedia, edited by Robert Whaples. March 16, 2003. URL http://eh.net/encyclopedia/public-sector-pensions-in-the-united-states/

Actuarial Modernization Errors

Mary Pat Campbell

Link: https://www.theactuarymagazine.org/actuarial-modernization-errors/

Excerpt:

Professionalization leads us to an interesting dilemma. Actuarial culture and, for that matter, organizational culture got insurance companies to where they are today. If the culture were not moderately successful, then the company would not still exist. But this is where Prospect theory emerges from the shadows. It is human nature not to want to lose the culture that enabled your success. Many people nonetheless thirst for the gains earned by moving in a new direction. Risk aversion further reinforces the stickiness of culture, especially for risk-averse professions and industries. Drawing from author Tony Robbins, you cannot become who you want to be by staying who you currently are. Our professionalization, coupled with our risk aversion, creates a double whammy. Practices appropriate to prior eras have a propensity to be locked in place. Oh, but it gets worse!

By the nature of transformation and modernization, knowledge and know-how are embedded in the current people, processes and systems. The knowledge and know-how must be migrated from the prior technology to modern technology. Just like your computer’s hard drive gets fragmented, so too do firms’ expertise as people change focus, move jobs or leave companies. The long-dated nature of our promises can severely exacerbate the issue. Human knowledge and know-how are not very compressible, unlike biological seeds and eggs. In a time-consuming defragmenting exercise, information, knowledge and know-how must be painstakingly moved, relearned and adapted for the new system. This transformation requires new practices, further exacerbating the shock to the culture. Oh, but it gets even worse!

The transformation process requires existing teams to change, recombine or communicate in new ways. This means their cultures will potentially clash. Lack of trust and bureaucracy are the most significant frictions to collaboration among networks. The direct evidence of this is when project managers vent that teams x, y and z cannot seem to work together. It is because they do not have a reference system to know how to work together.

Author(s): Bryon Robidoux

Publication Date: September 2022

Publication Site: The Actuary

SOA Diversity Report

Mary Pat Campbell

Link: https://www.soa.org/4a79dc/globalassets/assets/files/static-pages/about/diversity-inclusion/summer-2022-diversity-report.pdf

Graphic:

Excerpt:

The Society of Actuaries (SOA) leadership and staff work closely with the Diversity, Equity, and Inclusion Committee (DEIC) to support the journey to increase diversity in membership and in the actuarial profession, as part of the SOA’s Long-Term Growth Strategy.

We strive for transparency and accountability in our DEI efforts and are committed to sharing our demographic data and long-term goals to support our pledge and responsibility. We have collected member voluntary demographic data since 2015. With this data, we present an infographic for the pathway from aspiring actuaries to members with ASA or FSA designations.

Author(s): Society of Actuaries

Publication Date: Summer 2022

Publication Site: Society of Actuaries

Variations On Approximation – An Exploration in Calculation

Mary Pat Campbell

Link: https://www.soa.org/news-and-publications/newsletters/compact/2014/january/com-2014-iss50/variations-on-approximation–an-exploration-in-calculation/

Graphic:

Excerpt:

Before we get into the different approaches, why should you care about knowing multiple ways to calculate a distribution when we have a perfectly good symbolic formula that tells us the probability exactly?

As we shall soon see, having that formula gives us the illusion that we have the “exact” answer. We actually have to calculate the elements within. If you try calculating the binomial coefficients up front, you will notice they get very large, just as those powers of q get very small. In a system using floating point arithmetic, as Excel does, we may run into trouble with either underflow or overflow. Obviously, I picked a situation that would create just such troubles, by picking a somewhat large number of people and a somewhat low probability of death.

I am making no assumptions as to the specific use of the full distribution being made. It may be that one is attempting to calculate Value at Risk or Conditional Tail Expectation values. It may be that one is constructing stress scenarios. Most of the places where the following approximations fail are areas that are not necessarily of concern to actuaries, in general. In the following I will look at how each approximation behaves, and why one might choose that approach compared to others.

Author(s): Mary Pat Campbell

Publication Date: January 2014

Publication Site: CompAct, SOA

IFRS 17—The Time Is Now

Mary Pat Campbell

Link: https://contingencies.org/ifrs-17-the-time-is-now/

Excerpt:

The more fundamental changes affect the measurement of future services (previously termed as “Reserves”). Many insurance accounting regimes have tried to stabilize their financial statements over the years; therefore, they calculated their reserves based on historic information—locked-in assumptions for insurance parameters as well as historic interest rates. The latter, however, are not in line with the use of market values for the asset side of the balance sheet, which is now perceived as the only fair-value representation for the different stakeholders. Therefore, the measurement of the liabilities in IFRS 17 will always be based on current assumptions.

Due to the compound effect over many projected years, the regular update of assumptions (particularly interest rate or discounting assumptions) can make long-term liabilities much more volatile.

Author(s): Michael Winkler and Sunil Kansal

Publication Date: October 2022

Publication Site: Contingencies