• Challenging Topics on a Commissioner's Agenda - How Actuaries Can Help 
• A Panel on Social Inflation

• Reflecting Social Inflation in Actuarial Methods
• From Research to Real-World Results
• Individual Claim Reserving with Machine Learning
• A Turnkey Approach to Fix Burning Cost Development
• Correlation of Loss Reserve Estimates
• Credibility as Data Augmentation
• Natural Language Models for Dummies
• Loss Reserving with Credibility
• Can AI find your optimal reserving segmentation?

• State Specific Insurance Laws - What, Where, Why?
• Subsequent Events - CA Wildfires Case Study
• Surviving a Financial Exam
• Updates to Schedule F: What the F is in there?
• Ahead of the Curve: Battling Inflation
• RMAD

• State of the Workers Comp Lube Report
• State of the MPL Market
• Special Considerations in MPL Reserving
• Economic Impact on Workers Compensation 

• Pitfalls of Using New Technology/A.I.
• Machine Learning in Reserving
• Actuaries Working in Tech
• Case Study: Actuary versus Machine
• An AI Approach to Quantifying Social Inflation
• The Discovery of Tech-enabled Claim Instigation

• 2025 Purr-spective on Pet Insurance: Emerging Trends and Reserving Considerations
• Cyber Insurance Market Trends 2025: Insights from the Frontlines of Brokerage, Pricing, and Reserving
• Deep Dive on Personal Lines (Umbrella, Bodily Injury, UM/UIM)
• Mortgage Credit Risk Transfer in Uncertain Economic Times
• Home Sweet Home: Navigating Mortgage and Title Insurance
• The Current State of Flood Insurance 
• Commercial Auto    

• Winning the Communication Battle
• Recent ASOP Changes
• Professionalism Kahoot
• Coffee Klatch
• Unconscious Bias in Actuarial Reserving

• A Behavioral Approach to Loss Reserves
• Constructing Your Business Insurance Puzzle
• AI & Risk Management
• Wildfires 

• Publicly Available Data in Actuarial Analyses
• Antipatterns in Loss Reserving Workflows
• Reserve Allocations: They Matter!
• Industry Reserve and Loss Ratio Trends

• ** WS-1: Maximizing the Value of Your Actuarial Report and Opinion (Full-day)
• ** WS-2: Reserve Modeling Using Bayesian MCMC (Full-Day)
• ** WS-3: Bermuda: A Python Library for Insurance Triangles (Half-Day AM)
• ** WS-4: Linear Models for Reserving (Half-Day AM)

Description

Your actuarial opinion and reports are reviewed and analyzed for appropriateness to different stakeholders, including company management, auditors, regulators, court officials and jurors. Each group has different needs for information, and you will want to be able to communicate effectively to all. This workshop will deepen your expertise on the latest requirements for P&C Statements of Actuarial Opinions with a review of actuarial qualification standards, standards of practice and interactive case studies.

The workshop will offer attendees the opportunity to:

• Gain an understanding of regulatory/company/auditor perspectives and expectations.
• Remain up to date on the latest regulations and standards.
• Delve deeper into subjects such as RMAD, ranges, documentation, data and report writing.
• Focus on communicating effectively to various stakeholder groups.

Speakers

• Susan Gozzo Andrews, FCAS, Property Casualty Actuary, State of Connecticut Insurance Department
• Kathleen Odomirok, FCAS, Principal, EY
• Christopher Schubert, FCAS, Senior Consulting Actuary, Pinnacle Actuarial Resources
• Trevar Withers, ACAS, Director, WTW

Description

This will be a hands-on introduction to building reserve models using Bayesian MCMC with a high-level introduction to some of the modeling concepts.

The workshop will use a case study approach to walk participants through building a Bayesian MCMC model. The workshop will alternate between short presentations of the next task in building a model and hands-on work by each participant to accomplish the next step. In the morning, we will work through a simple modeling case and briefly introduce some Bayesian modeling concepts. In the afternoon, we will introduce additional modeling options and concepts to handle more complex cases, and participants can choose between modeling, with our help, more complex cases: inflation is changing, the claim department changes operations, and/or the business unit changes operations.

The Bayesian MCMC modeling will demonstrate how to combine a regression type formula to simultaneously model the mean and variance of loss payments with credibility weighting. We will give a short explanation of why the different steps work but will not go into an in-depth explanation of the theory. We will point out how actuarial judgement can be integrated into the modeling process. We will go through the process of reading the data set, setting initial modeling assumptions, trying out different model forms, comparing modeling results and graphs to present model results to a client.

The modeling will take place in Rstudio using the brms macro writer to create the Bayesian MCMC models that will run in STAN, and we will use R packages in the tidyverse to manipulate the data and create graphs. We will demonstrate how existing open-source packages can provide diagnostics on Bayesian MCMC results. While we will not assume prior knowledge of Bayesian MCMC modeling, we will assume basic reserving knowledge at the current Exam 5 level as well as some ability to operate in the Rstudio environment and some exposure to regression modeling.

Participants will be able to save the reserving examples they work on during the workshop on their laptops.

Speakers

• Mike Larsen, FCAS, AAA, Retired
• Dr. Thomas Struppeck, FCAS, CERA, Assistant Professor at University of Texas Austin

Description

Bermuda is an open-source Python library for storing, cleaning, and manipulating insurance loss triangles. This workshop will cover Bermuda’s core functionality, focusing on practical applications for working with cumulative and incremental triangles in various formats. Attendees will learn how to efficiently load, transform, and analyze triangle data by integrating Bermuda into their existing actuarial workflows.

Outline:

1. Introduction to Bermuda
   • Purpose and key features of the library
   • The role of Bermuda in insurance triangle analysis 
2. Core Components
   • The Triangle class and its core methods
   • The Cell type and its role in triangle manipulation 
   • Differences between Triangles/Cells and other common triangle data formats
3. Usage Patterns
   • Loading and saving triangles in different formats (JSON, CSV, DataFrame, Binary)
   • Common transformations including summarization, aggregation, and filtering
   • Information-rich triangle visualization methods
   • Bootstrapping, moment matching, and other power-user methods 
4. Integration into Actuarial Workflows
   • Incorporating Bermuda into data pipelines
   • Case studies demonstrating practical use cases
5. Interactive Session
   • Hands-on exercises using real-world datasets
   • Problem-solving and Q&A.

Speakers:

• Nathaniel Haines, Manager, Data Science Research, Ledger Investing
• Kyle Safran, Lead Data Scientist, Ledger Investing

Description

Chain-ladder, Bornhuetter-Ferguson, Cape Cod, and Generalized Linear Models all share a common framework: they are all variants of linear models. To these techniques we may add linear mixed models and the incremental loss ratio method. When we consider loss reserving as a linear modeling problem, we may leverage common data structure and model assessment techniques. This session will demonstrate how to carry this out.

Speakers

• Brian Fannin, ACAS CSPA, Actuary, Akur8
• Michael Henk, FCAS, MAAA, Actuary, Akur8

The opinions expressed by speakers at this event are their own and do not necessarily reflect the opinions of the CAS.