Search Presentations
STAY TUNED! If you are anticipating additional search filters by attribute and level to align with the CAS Capability Model, it is coming later this Summer. As the CAS begins to code recorded sessions by specific attributes and levels (starting with the 2023 Annual Meeting), these will be tagged in the CAS database of presentations going forward and should be searchable.
But you may use the Capability Model now to help you identify topics. For example, if you want to move up one level under the content area “Functional Expertise,” you may search topics in the particular functional area to expand your knowledge.
Recorded content is searchable by Capability Model attribute and level in the CAS Online Library.
Refine Your Search
The presentation materials are offered in connection with CAS professional education offerings. © 2022 Casualty Actuarial Society. All Rights Reserved. The presentation materials may contain copyrighted content the use of which has not been specifically authorized by the copyright owner. You are permitted to view and print the materials for personal/professional noncommercial research purposes. Except for the foregoing, you agree not to reproduce, distribute, modify, create derivative works, or commercially exploit the presentation materials without prior written permission from CAS. Please direct any copyright permission inquiries regarding use of the presentation materials to acs@casact.org.
STAY TUNED! If you are anticipating additional search filters by attribute and level to align with the CAS Capability Model, it is coming later this Summer. As the CAS begins to code recorded sessions by specific attributes and levels (starting with the 2023 Annual Meeting), these will be tagged in the CAS database of presentations going forward and should be searchable.
But you may use the Capability Model now to help you identify topics. For example, if you want to move up one level under the content area “Functional Expertise,” you may search topics in the particular functional area to expand your knowledge.
Recorded content is searchable by Capability Model attribute and level in the CAS Online Library.
Premium Optimization
Recently, there is quite a lot of interest and concerns around the topic of optimization. Actuaries develop a rate which reflects the expected cost of a risk. We then incorporate business decisions to select a price that may deviate from the rate. Optimization is a tool which provides a more rigorous framework in selecting the difference between the rate and the price. In addition there are several optimization approaches that could be used to support this decision. Join us as we discuss the various approaches and demystify their application within a regulated marketplace.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Serhat Guven
Predicting Homeowners Losses: Which Characteristics Matter?
This roundtable session is focused on the many challenges to writing profitable homeowners business, with discussion topics including by-peril rating, emerging variables and data sources, use of catastrophe models, and other hot industry topics.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Klayton Southwood
The Emerging Technologies of UBI
Available technology and sources of telematics data are rapidly changing. This discussion will center on the new methods of proactively and passively collecting and utilizing telematics data to enhance insurance products.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Robin Harbage
Shark Tanking Ratemaking and Product Design
Recent reporting suggests accelerators and venture capitalists are awakening to the possibilities new data sources and business models for our industry. This roundtable will analyze recent innovations in areas such as usage-based insurance (UBI), social media, wearables and wellness, connected homes and businesses, and peer-to-peer/micro-insurance, and debate which hold the greatest potential for our industry and profession.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Jim Weiss
Predicting Modeling Pitfalls: Challenges in Application 
There are many success stories featuring Predictive Models, but there are also numerous cautionary tales in which the results of well-constructed models were misunderstood and misapplied. After some preliminary discussion to define what is or is not a Predictive Model, this session will feature several short and entertaining stories illustrating common pitfalls in applying Predictive Models. Participants will be also be invited to share their own hard-learned lessons.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Ira Robbin
Using Predictive Modeling Beyond Ratemaking
We will explore some ways insurance companies are using predictive modeling beyond ratemaking. Some examples include in the product management, underwriting, and agency/sales management spaces. If you are using predictive modeling in interesting ways beyond rate making, and would like to talk about it, or explore how to do it, please come join us for this engaging discussion.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Margaret Brinkmann
Usage-Based Insurance (UBI)
A collaborative discussion of the various issues around usage-based auto insurance.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Ryan Morrison
How Can We Leverage the Capability of Machine Learning to Help us Build Better GLMs?
In this session we will have an interactive discussion on how we could maximize the value of machine learning in building a better GLM (and other traditional statistical models).
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Satadru Sengupta, Owen Zhang
Predicting Disease through Data
Actuaries have been using predictive modeling to predict auto and catastrophe losses for years. Can the same techniques be used to predict illness? Please join us as for a brainstorming session on this exciting topic.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Kevin Donnelly, Gerald Frye
CAS Institute Predictive Modeling/Data Science Expert Panel
The Casualty Actuarial Society (CAS) has created The CAS Institute, an organization offering new credentials and specialized professional education for quantitative professionals looking to remain current in their field. The CAS Institute will develop a curriculum for each of its offered specialty areas, initially covering Predictive Analytics and Data Science. Members of the iCAS Expert Panel on Predictive Analytics and Data Science will share information and answer questions regarding the CAS Institute’s initial program during this roundtable discussion.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Roundtable
Panelists:
Louise Francis, James Guszcza, Christopher Monsour, Todd Lehmann
Perspectives on Rate Filing Support for Predictive Models
While the use of predictive analytics to develop rate classification plans has become a widely accepted technique in the insurance industry, they can create additional challenges during the filing process due to their apparent black box nature. While some insurance departments may only require indicated and proposed rating factors, others may require substantially more support including such items as confidence intervals, lift charts, or comparisons to other insurance companies. This session will provide different points of view on filing support for predictive models, including perspectives from those submitting rate filings (insurance company or consultant) and those reviewing the rate filings (insurance department or independent third party reviewer). In this session, we will discuss modeling topics as they relate to filing support such as volatile results and thin data, extrapolating and smoothing indicated relativities, and the use of actuarial judgment. We will also provide perspectives on how to best submit rate filings that rely on predictive modeling techniques beyond GLMs.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Marcus Yamashiro
Panelists:
Paul Anderson, Ryan Purdy, Eric Krafcheck
Discussion of the Proposed Actuarial Standard of Practice on Property-Casualty Ratemaking
The Actuarial Standards Board, its Casualty Committee and Ratemaking Task Force continue the
development of an ASOP on P&C Ratemaking. The ASOP was initially intended to replace the
Consideration section of the CAS Statement of Principles. An exposure draft ASOP is slated for release for review and comment prior to the CAS Annual Meeting.
Join members of the groups involved in the development of this important document for practitioners in ratemaking in discuss the contents of current version. We will highlight the questions the ASB has posed in the transmittal letter and key areas of change from the initial draft.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Benjamin Williams
Panelists:
David Otto, Christopher Carlson
What Makes a Good Florida Rate Filing?
This is your chance to learn about the specific issues and challenges of getting your rate filing approved in Florida. The session is also a chance for the Department staff to share the issues that delay the approval of your rate filing.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Benjamin Williams
Panelists:
Kayne Smith
A Divergence of Opinions: A Follow up to the General Session Session on Price Optimization
This session is a chance to get answers to the questions that weren't answered during the General Session Session.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Howard Eagelfeld, James Lynch, Brent Petzoldt
Ask a Regulator
During this session, each member of a panel of regulators will field questions about current issues, the rate regulation process in their state, and other concerns raised by the audience. Audience suggested "Hot Button" issues and concerns will be discussed. The differences and the similarities in regulatory approach should become apparent.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Kris DeFrain, Thomas Hess, Charles Angell, Kayne Smith
Professionalism in Predictive Modeling and Pricing: Episode VII - The Price Awakens
During the past decade, predictive modeling has increasingly become part of many actuaries' job descriptions. Whether building predictive models, inserting them as inputs to the pricing decision-making and implementation process with business partners, or supporting them in rate filings, actuaries are often asked for their opinions on issues that relate to the ethics of the profession. This session has been given at several previous RPM seminars and the interactive format will remain the same, although the scenarios have been refreshed. Scenarios that touch on common issues around predictive modeling will be presented, including how those issues relate to the Code of Conduct, Statement of Principles, and Standard of Practice. Audience participation is highly encouraged. May the Professionalism be with you.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Nathan Hubbell, Brent Petzoldt
Who Wants To Be An Actuary?
Come refresh your knowledge of actuarial professionalism in this interactive session! This trivia game will cover a number of topics, helping you determine what you know and what you might want to review. Who said professionalism can’t be fun? This session may count as professional continuing education.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Michael Chen, Rebecca Williams, Emilee Kuhn, Fanny Paz-Prizant
Lights! Camera! Professionalism!
Come enjoy the acting of some fine fellow actuaries who will take on several skits involving professional dilemmas. The skits are new for 2016, but as always they will lead to some lively and educational audience discussions. You will walk away from this session with a better understanding of the ASOPs, codes of conduct and how to apply them when you face your own professional ethical dilemmas. This session may provide attendees with Professionalism Continuing Education credits.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Michael Chen, Kathryn Walker, Emilee Kuhn, Fanny Paz-Prizant
Linear Statistical Assumptions for Insurance Segmentation and Ranking Models
For Insurance predictive modeling applications, we often deal with cases where we try to predict the true outcome of an event, such as occurrence of a claim, identification of a fraud, ultimate development of a claim severity. In such cases, statistically speaking, given the specific target variables, we perform logistic or General Session GLM regression. However, more often than not, we can approach such prediction using a segmentation approach by ranking observations based on the predicted values from the lowest to the highest. One of the most popular segmentation models is a credit score model, of which the better the credit score, the better the risk. In this presentation, we show that for segmentation model or scoring model, the segmentation result is not sensitive to the underlying distribution assumption. For applications with a binary target variable, the linear regression would yield segmentation results that are very close to what Logistic regression yields. Using historical claims data from a couple of Insurance companies, we will show that the segmentation obtained from both methods would be very similar. More importantly, we will prove theoretically that the coefficients of the independent variables obtained from both the methods would have the same sign and that for the case when there is only one independent variable, the ranking of observations based on the predicted values obtained from both the methods is exactly the same. We will also discuss important implications of such finding for the insurance predictive modeling applications.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Luyang Fu, Kranthi Nekkalapu, Debashish Banerjee
CPXR – A Powerful New Method for both Regression Modeling and Regression Model Analysis
In this session, we will introduce (a) a new style of regression models, namely pattern aided regression (PXR) models, and (b) a contrast pattern aided regression (CPXR) method, to build accurate and easy-to-explain PXR models. PXR models were motivated by two observations: (1) Regression modeling applications often involve complex diverse predictor-response relationships, which occur when the optimal regression models (of given popular model types) fitting distinct natural subgroups of the data of an application are highly different. (2) State-of-the-art regression methods are often unable to adequately model such highly diverse predictor-response relationships. To accurately model highly diverse predictor-response relationships, a PXR model uses several pattern and local regression model pairs, which respectively serve as logical and behavioral characterizations of distinct predictor-response relationships, to define a prediction model. In experiments over 50 applications, the PXR models built by CPXR are very accurate in General Session, often outperforming state-of-the-art regression methods (including linear regression, piecewise linear regression, boosting, neural networks, and random forest) by wide margins. Using around seven simple patterns on average and using linear local regression models, those PXR models are easy to interpret; in contrast, other regression models are often blackboxes or contain hundreds of individual regression models, making them hard to interpret. CPXR is especially effective for high dimensional data. The CPXR methodology can also be used for analyzing and improving existing prediction models, correcting their prediction errors, and discovering structure of data with respect to prediction. We will discuss how to use CPXR for risk prediction and classification, including results on (insurance/medical) risk prediction. We will also compare the CPXR methodology with other regression modeling algorithms.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Jun Yan, Guozhu Dong
And the Winner Is: How to Pick a Better Model
You have just finished running some data through a predictive modeling package. Now all you need to do is summarize the results, send them along, and you're done, right? Wrong. At the absolute minimum, the modeler should understand and demonstrate the goodness of fit of the model. In most cases, the modeler should also prove that the constructed model provides lift over the existing rating structure. After all, what good is a new model if it cannot outperform the competition? In this session we will explore, in significant detail, three often overlooked components of the modeling process: measuring goodness-of-fit, assessing lift, and internally validating a predictive model. Model development is usually a major investment. We should make sure our models are performing well to get the best bang for the buck.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Hoi Leung, Hernan Medina, Dan Tevet
GLM - III - The Matrix Reloaded
This session will consider advanced techniques and refinements to the basic GLM which can add material value to the modeling process. It will specifically consider amendments which address some of the purported failings of GLMs in comparison to emerging methods such as machine learning techniques.
The session will include a discussion of:
- An innovative approach to detecting subtle and higher dimensional interactions in an efficient way, potentially eliminating the need to consider alternative, harder-to-implement model forms such as nonlinear models.
- The pros and cons of modeling pure premium directly as opposed to frequency and severity individually
- Innovative ways of modeling bodily injury claims and other miscellaneous refinements.
- Best practices in driver assignment vs. driver averaging approach
- How to capture geographic risk both within and beyond GLMs
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Brent Petzoldt, Evan Petzoldt
GLM- II
GLM I provided the case for using GLMs and some basic GLM theory. GLM II will be a practical session outlining basic modeling strategy. The discussion will cover topics such as overall modeling strategy, selecting an appropriate error structure and link function, simplifying the GLM (i.e., excluding variables, grouping levels, and fitting curves), complicating the GLM (i.e., adding interactions), and validating the final model. The session will also include a discussion on diagnostics that help test the selections made.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Ernesto Schirmacher, Paul Bailey
GLM - I
Do terms such as “link function,” “exponential family,” and “deviance” leave you puzzled? If so, this session will clarify those terms and demystify General Sessionized linear models (GLMs). The session will provide a basic introduction to linear models and GLMs. Targeted at those who have modest experience with statistics or modeling, the session will start with a brief review of traditional linear models, particularly regression, which has been taught and widely applied for decades. Session leaders will explain how GLMs naturally arise as some of the restrictive assumptions of linear regression are relaxed. GLMs can model a wide range of phenomena, including frequencies and severities as well as the probability that a claim is fraudulent or abusive, to name just a few. The session will emphasize intuition and insight in addition to mathematical calculations. Illustrations will be presented using actual insurance data.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Nathan Hubbell
Panelists:
Ernesto Schirmacher
Developing Loss Cost Models: Modeling Exercise or Transformational Business Process?
Many companies view loss cost modeling as a point in time activity that requires a one-off project to update the data each year. In reality, when a company makes the decision to price based on loss cost models this is a fundamental change to one of their core business processes, pricing. Just like the claims process and underwriting process, the pricing process needs to be managed.
Source:
2016 Ratemaking and Product Management Seminar
Type:
Concurrent Session
Moderators:
Kevin Donnelly
Panelists:
Kathryn Walker, Howard Kunst, Drew Lawyer
