CAS 2004 Spring Meeting Handouts
These handouts are available in their original Power Point Presentation format. If you do not have Power Point on your computer, you will need to download the free Power Point Viewer, which is a 2.8 MB file.
Some of the handouts are Portable Document Format (PDF) files. The files were created using Adobe Acrobat. To view the documents, you need the Acrobat Reader. For more information, read the Adobe Frequently Asked Questions. To download the free Acrobat Reader, click the icon below.
If a session handout is not online, unfortunately it was not submitted by the speaker and is not available.
The Truth About Loss Reserve Adequacy
A perennial, loss reserves adequacy, has been the subject of intense renewed interest in the past year. This panel will explore loss reserve adequacy from various perspectives addressing, among other things:
- The state of actuarial practice as it relates to loss reserving, both the art and the science;
- Implications from a regulatory, investor, and legal perspective;
- Implications for the education of the actuary; and
- The need for improved communications with users of actuarial estimates of reserves.
This session will also consider what might be done to address this issue in a business like manner.
C. K. "Stan" Khury, Principal, Bass & Khury
Chuck Bryan, President, CAB Consulting, LLC
Joseph P. Dailey, Partner, Dailey & Selznick
Mary Miller, Actuary, Ohio Department of Insurance
Professionalism Standards in Practice
How long has it been since you knocked the dust off of your Actuarial Standards of Practice binder or read the CAS Code of Professional Conduct? This session, which will be facilitated by members of the CAS Committee on Professionalism Education, will discuss several case studies involving professional dilemmas. These case studies are intended to lead to lively and educational audience interaction. We guarantee that you will walk away with an increased awareness of professionalism and its impact on the reputation of the actuarial community.
Richard J. Currie, Vice President and Actuary, American Re-Insurance Company
John T. Gleba, Consulting Actuary, Madison Consulting Group, Inc.
Roosevelt C. Mosley, Actuary, Pinnacle Actuarial Resources, Inc.
Acting in support of the International Association of Insurance Supervisors (IAIS), the Insurance Regulation Committee of the International Actuarial Association (IAA) formed the Insurer Solvency Assessment Working Party in early 2002 to prepare a report on insurer solvency assessment. This report represents the culmination of that mandate and is meant to help develop a global framework for insurer solvency assessment and determine insurer capital requirements.
In the course of its mandate, the Working Party presented its work to various insurance supervisory and actuarial meetings. Feedback from these presentations has been both positive and constructive. This session will discuss the critical components of the Working Party's recommendations. Audience participation and discussion is strongly encouraged.
Stuart Wason, Director, Mercer Oliver Wyman, Inc.
Allan Brender, Senior Director, Actuarial Division, Office of the Superintendent of Financial Institutions Canada
Glenn G. Meyers, Chief of Actuarial Research and Assistant Vice President, ISO, Inc.
Fair Value Accounting- Can It Work?
Many practicing actuaries are relatively uninformed about fair value accounting standards and the repercussions they will have on their profession and their employers. This panel will present a brief history of fair value accounting standards and the rationale behind its implementation. Panelists will share their thoughts on whether this new concept makes sense for the insurance industry.
Ralph S. Blanchard, III, Second Vice President & Actuary, Travelers Property Casualty Insurance Company
Mark W. Littmann, Principal, PricewaterhouseCoopers LLP
Stephen P. Lowe, Principal & Consulting Actuary, Tillinghast-Towers Perrin
Actuaries in Nontraditional Roles
Casualty actuaries are continuing to expand into nontraditional areas of practice, including those related to risk management, financial management, underwriting, claims, and public policy.
The panelists will discuss their nontraditional positions, including their day to day activities and the obstacles they face being an actuary in a non-actuarial role.
Robert C. Morand, Partner, DW Simpson & Company
Todd R. Bault, Analyst-Institutional Research, Sanford C. Bernstein & Co., Inc
Eric Lemieux, Principal, The Black Diamond Group, LLC
John V. Mulhall, Insurance Practice Consultant, McKinsey & Company
How does a senior actuary at a property/casualty insurance company gain and sustain the attention of the senior management team on forecast and reserve results without bogging down communications with technicalities and jargon? The panelists will demonstrate how actuarial reserving and pricing analyses can be interconnected with the core concerns of the CEO, President, CFO, and Chief Underwriting Officer.
One presentation will describe a coordinated planning process that is driven by a CEO's desired ROE for the upcoming underwriting year. The discussion will illustrate that by starting with the reserve loss ratio indications and pricing assumptions, the actuaries and underwriters can work together towards a common ROE goal.
In addition, a reserving presentation will show the uses of an expected loss emergence monitoring process to communicate loss reserve indications to senior management. Discussion will include how the senior actuarial team "educated" the company managers to interpret and rely on reserve estimates and, further, how this information should be relayed to the parent company.
Vincent M. Senia, VP & Chief Reserving Actuary, American Re-Insurance Company
Roger A. Atkinson, III, Actuary, Employers Reinsurance Corporation
Dr. John C. Burville
ARIA Prize Paper:
"Fraud Classification Using Principal Component Analysis of RIDITs"
This paper introduces to the statistical and insurance literature a mathematical technique for an a priori classification of objects when no training sample exists for which the exact correct group membership is known. The article also provides an example of the empirical application of the methodology to fraud detection for bodily injury claims in automobile insurance.
With this technique, principal component analysis of RIDIT scores (PRIDIT), an insurance fraud detector can reduce uncertainty and increase the chances of targeting the appropriate claims so that an organization will be more likely to allocate investigative resources efficiently to uncover insurance fraud. In addition, other (exogenous) empirical models can be validated relative to the PRIDIT-derived weights for optimal ranking of fraud/nonfraud claims and/or profiling. The technique at once gives measures of the individual fraud indicator variables' worth and a measure of individual claim file suspicion level for the entire claim file that can be used to cogently direct further fraud investigation resources. Moreover, the technique does so at a lower cost than utilizing human insurance investigators, or insurance adjusters, but with similar outcomes. More generally, this technique is applicable to other commonly encountered managerial settings in which a large number of assignment decisions are made subjectively based on "clues," which may change dramatically over time. This article explores the application of these techniques to injury insurance claims for automobile bodily injury in detail.
Louise A. Francis, Consulting Principal, Francis Analytics & Actuarial Data Mining
Mark Alpert, University of Texas
Patrick L. Brockett , University of Texas
Richard A. Derrig, Automobile Insurers Bureau of Massachusetts
Linda L. Golden, University of Texas
Arnold Levine, Tulane University
Building Communication Skills Through Improvisation
Using fundamental exercises from the theater genre of improvisation, attendees will interact with this session's facilitator to examine business communications. The facilitator will guide attendees with suggestions to improve their communication and negotiation skills. In small groups, attendees will participate in improvisational exercises that emphasize the importance of listening in communication.
- Learn to communicate and negotiate more effectively with colleagues and clients;
- Gain a better understanding of both the input and output of communication;
- Examine their confidence in communicating and negotiating; and
- Improve their communication and negotiation skills.
No experience with improvisation is required.
Robert C. Morand, Partner, D.W. Simpson and Company
California Workers Compensation
Despite the passage of reform legislation effective January 1, 2004, California workers' compensation costs are still more than double the rest of the nation. The estimated reduction in loss costs due to reform ranges from 15% to 18%. The Department of Insurance estimates the reduction at the higher end of the range and the Workers' Compensation Insurance Rating Bureau estimates the reduction at the lower end of the range. In any event, insurers are not rushing back into the state to write new business and most employers are still not seeing adequate rate relief.
The Governor set a March 2004 deadline for the legislature to propose additional, meaningful cost saving measures. Otherwise, there could be an initiative on the ballot in November and the voters will weigh in on the subject.
This session is designed to deliver up-to-date information on the state of reform, expected benefits, and early indications of how recently enacted reforms may be measured.
David M. Bellusci, Senior Vice President and Chief Actuary, Workers' Compensation Insurance Rating Bureau
Alex Swedlow, Executive Vice President, California Workers Compensation Institute
Lawrence White, Workers Compensation Insurance Policy Advisor, California Department of Insurance
In recent years, medical professional liability insurance premium and claims costs have soared. Stories about physicians relocating from individual states or leaving the practice of medicine altogether are widely reported. Most tort reform efforts at both federal and individual state levels are based on the 1975 California MICRA reform. The centerpieces of MICRA are arbitration and a $250,000 cap on non-economic damages.
This session will begin with a historical perspective on the current crisis along with discussions about tort reform efforts on both a federal and specific state level. The session will continue with a presentation of a recent actuarial analysis estimating the impact of implementing non-economic damage caps in the state of Pennsylvania.
Leon R. Gottlieb, Principal, Mercer Oliver Wyman, Inc.
William Bell, General Counsel, Florida Hospital Association
Richard S. Biondi, Principal and Consulting Actuary, Milliman USA
CAS 2003 Membership Survey
Every five years the CAS conducts a major survey of its members. The results of this survey provide the CAS leadership with valuable input to help shape the short and long-term direction of the Society. Over 52 % of the membership (1,934 members) completed the survey during July 2003. The Membership Survey Task Force has prepared a report on the survey for the CAS Board of Directors. Come to this session to hear what your fellow CAS members thought about a wide variety of issues that are important to the CAS - and see what the CAS leadership is doing in response to the recommendations in the report. The presentation will be both informative and entertaining. Prizes will be awarded to audience members that demonstrate the most knowledge of CAS demographics and attitudes.
Joanne S. Spalla, Senior Vice President and Corporate Actuary, Converium Consulting
Correlation: An Update
Effective Enterprise Risk Management (ERM) requires the aggregation of risk information and quantification of the total risk exposure a firm faces. In order to properly model this composite risk, ERM practitioners need theory and techniques for estimating and modeling correlations and dependencies among all risk sources: hazard, financial, operational and strategic. Recognizing this need, the CAS formed a working party to begin laying the theoretical and practical framework for addressing these problems.
This session will report on the working party's progress to date, including discussions of specific models developed by the working party to address the many sources of insurer risk. The session will also address parameterization issues and show how to implement these models in an insurance company setting.
Glenn G. Meyers, Chief of Actuarial Research and Assistant Vice President, Insurance Services Office
Youngju Lee, Vice President - Quantitative Investment Research, Allianz Hedge Fund Partners
Stephen J. Mildenhall, Senior Vice President, Aon Re Services
Data Quality and Standards
The quality of data used in loss reserving and other analysis has always been an important issue among actuaries, but has never been as hot an issue as it is today. For Sarbanes-Oxley 302 and 404 certifications, publicly traded insurance companies will have to demonstrate that they have stringent controls surrounding data quality to the extent that such data impacts the company's financial statements.
In addition, perhaps in response to certain solvency concerns, the quality of data used in loss reserving analyses has also become a heightened focus of many regulators. In response, the NAIC is expected to modify the 2004 annual statement instructions to require additional testing procedures be performed by an independent certified public accountant on the data used by the opining actuary.
Further, the Actuarial Standard of Practice #23 - Data Quality is in the process of being updated by the Actuarial Standards Board for several purposes, one of which is to clarify the actuary's role and responsibilities as it relates to data quality.
The panel will discuss each of these issues and provide insights into what is expected of actuaries in the future.
Marc F Oberholtzer, Principal Consultant, PricewaterhouseCoopers LLP
Mark W. Littmann, Principal, PricewaterhouseCoopers LLP
John McCauley, Partner, PricewaterhouseCoopers LLP
Estimated Reserve Ranges -
How Do We Determine "Reasonableness"?
So you've decided its time to calculate a range of liability estimates before your management sets reserves, what do our guidelines and standards of practice say about which parts of that range constitutes a reasonable estimate? This session will briefly review relevant guidelines and standards and examine different types of methods for calculating ranges, but will focus on using a statistical approach for determining what constitutes a reasonable reserve estimate.
Mark R. Shapland, Actuary, Milliman USA
Lee Van Slyke, President, Capital Management Technology
How to get the biggest bang for your reinsurance buck
It is well understood that reinsurance pricing starts with the use of experience and exposure rating models and the reconciliation between the two. However, what else can be done to ensure that reinsurance prices are properly set regardless of the state of the underwriting cycle?
This panel will discuss the importance of data in establishing the proper price. In order to get the best rate for the reinsurance product, the data requirements for ceding companies are greater than they were a few years ago. When the data is lacking, there is greater uncertainty about the reinsurance product, leading actuaries to be conservative. For example, price-monitoring information is very important. The better the information a ceding company can provide to support significant rate increases, the better off they will be. Another example of the importance of data is that many companies have made significant changes in their underwriting practices that are difficult to reflect in the estimate of loss costs if sufficient data is not provided to measure them.
This panel will discuss a variety of important issues that can assist the role of the Reinsurance Pricing actuary.
Karen Pachyn, Senior Vice President and Chief Actuary, GE Reinsurance Corporation
Michele P. Bernal, VP & Actuary, American Re-Insurance Company
Daniel Carberry, Vice President, Benfield Inc.
Loss Reserve Discounting
This session will provide the perspective of actuaries in three areas where discounting is common practice: Captives, Self-insurers, and Canadian Insurance Companies. The panel will also provide the latest scoop on where international accounting practices are headed, particularly in the European Community.
- Captives and Self-insurers: Reserve discounting remains relatively common in some business segments, such as medical malpractice mutuals. The panelists will discuss some of the more common issues that self-insurers and captives consider when deciding whether or not to discount their loss reserves. They will also present the regulatory, practical, and technical considerations that are commonly encountered by self-insurers and captives.
- Canadian Accounting: Beginning in 2003, Canadian insurance companies are discounting loss reserves as standard practice. We will discuss the reasoning for this policy change, as well as implementation issues that have surfaced.
- International Accounting: The International Accounting Standards Board stated an intent to implement discounted loss reserves as part of their insurance accounting standards project, with implementation possible by 2008, and these standards will be mandatory in the European Union. Is some form of discounting (and possibly fair value) inevitable in the United States?
Following their presentations, the panel will invite attendees to join in a general discussion of the technical, practical, and political challenges faced when discounting loss reserves.
David J. Oakden, Consulting Actuary, Towers Perrin
Ralph S. Blanchard, III, Second Vice President & Actuary, Travelers Property Casualty Insurance Company
Joseph A. Herbers, Principal & Consulting Actuary, Pinnacle Actuarial Resources, Inc.
Pricing actuaries attempt to determine rating plans that match the rates to the risk posed by the insured. Historically, personal lines auto insurers have used car-years as the exposure unit and used variables such as garage location, usage, and annual mileage as proxies for the amount, place, time, and type of driving.
Recent technological advancements allow actual driving patterns to be tracked and transmitted. Presumably, insurers could use this information to refine their rating plans and determine more equitable rates for all consumers. In addition to reducing current inequities, many feel a more accurate rating plan would provide economic incentives that would reduce the amount of driving, especially "risky driving." A reduction in miles driven would not only benefit the environment, but would also reduce the number of accident-related injuries and deaths.
This panel will describe the basic concept of Pay-As-You-Drive (PAYD) insurance, discuss the benefits of PAYD, and highlight some of the barriers to implementation.
Chester J. Szczepanski, Chief Actuary, Pennsylvania Insurance Department
Robin A. Harbage, General Manager, Direct Product, Progressive Insurance Company
Todd Litman, Director, Victoria Transport Policy Institute
The NAIC Risk-Based Capital (RBC) formula implemented in the early 1990s has been deemed an improvement over prior minimum capital standards. The RBC formula provides a ranking process for property casualty insurers with regards to regulatory intervention. But in recent years, the formula has been criticized because the process does not detect troubled companies early enough, despite the fact that the formula was not intended to be used as an "end-all" solution to capital adequacy standards.
Several committees of both the American Academy of Actuaries and the National Association of Insurance Commissioners have been working jointly to test the feasibility of a "trend test" solution, similar to that used for life insurers. The goal is to spot troubled insurers sooner. In essence, this effort is an attempt to evaluate and possibly improve upon the early warning responsiveness of the current RBC formula with regards to company impairment. The panel will discuss the specific research initiative to review actual failed insurers in the interest of gleaning aspects of risk perhaps not captured early enough in the current formula.
Robert F. Wolf, Principal, Mercer Oliver Wyman
G. Chris Nyce, Manager, KPMG LLP
Patents - Can We Share?
Intellectual capital is the main value actuaries bring to their respective employers. Historically, the Casualty Actuarial Society has operated in a very collaborative fashion and actuaries have freely shared information in various publications in an effort to improve the Society as a whole. In landmark case in 1998, the U.S. Court of Appeals Federal Circuit reinterpreted the U.S. patent laws to include business processes as patentable subject matter. As a result of that decision, there has been a proliferation of patents within the P&C insurance industry.
This panel will
- debate the pros and cons of this movement for the P&C industry,
- discuss how companies' actions may need to change, and
- comment on the implications of patents on CAS education.
Donald F. Mango, Director of Research and Development, GE ERC
Tom Bakos, President, Tom Bakos Consulting Inc.
Phil Hargrove, Vice President, Intellectual Asset Management, GE Employers Reinsurance Corporation
Mark Nowotarski, President, Markets, Patents & Alliances L.L.C.
Personal Auto Insurance-Surprised the Car Makes a Difference?
As the mix in personal vehicle types on the road has become more diverse in size and the distribution more equal in numbers, personal auto insurers have re-examined the rating for personal auto insurance. The actual vehicle driven, long a staple in the rating of physical damage insurance, has become an examined element in the rating of first party medical coverages (PIP and medical payments) and third party bodily injury and property damage liability. However, not all insurers have reacted in the same way. What are some of the differences in the rating structures and what are the facts and issues that may have lead different insurers to draw differing conclusions?
Patrick Woods, Assistant Vice President, ISO, Inc.
Daniel Charbonneau, Actuary, Allstate Insurance Company
Thomas Rau, Director, Personal Lines Pricing Support, Nationwide Insurance Company
Presenting Dynamic Financial Analysis Results to Decision Makers
This session will summarize the work to date of the CAS Research Working Party on Executive Level Decision Making Using DFA. The panel of working party members will begin by reviewing their survey of past DFA presentations and will highlight slides or presentation techniques that work well when presenting DFA studies. A PowerPoint template for DFA presentations will be introduced followed by sample presentations that illustrate how the template can be applied. The panel will share their insights gained from their survey of DFA presentations and their own experiences.
Mark R. Shapland, Actuary, Milliman USA
Michael R. Larsen, Working Party Chair, Property Consultant, The Hartford
Raju Bohra, Vice President - Client Modeling, American Re-Insurance Company
Patrick Crowe, VP and Actuary, Market Research, Kentucky Farm Bureau
Aleksey Popelyukhin, VP, Information Systems, Commercial Risk Re
In the United States, pricing actuaries have developed sophisticated methods for identifying the loss costs associated with a particular risk group. Companies try to balance the indicated loss costs against competitive data and current rates to determine an appropriate rate action.
In other markets in the world, pricing actuaries are now successfully integrating loss cost and retention/conversion analysis to systematically determine optimized pricing decisions. The panel will address the fundamentals of integrated optimized pricing, hurdles for implementation, and the relative benefits of the approach. Finally, the panel will discuss how the concepts can be applied in the United States given the various state regulations.
Floyd M. Yager, Senior Actuary, Allstate
Peter Orlay, Director, Optimal Decisions Group
Chester J. Szczepanski, Chief Actuary, Pennsylvania Insurance Department
Professionalism and Actuarial Limits of Liability
Recently, actuarial consulting firms are attempting to obtain limits of liability from clients. There are several reasons for this including the plaintiff bar becoming more aware of actuaries, increased frequency of lawsuits against actuaries, the hard market which caused Errors and Omissions premiums to skyrocket, and, of course, the accounting scandals in recent years. Since actuaries charge relatively low fees to estimate billions of dollars of contingent liabilities, are limits of liabilities reasonable? What are some clients doing to address these limits? Will this increase "poor judgment" in the actuarial profession and lead to scandals in the future?
This session will begin with a discussion of professionalism from a legal perspective presented by the General Counsel for the American Academy of Actuaries. Other panelists will then address the current issues of actuarial malpractice and discuss the limits of liability from both a consulting and a client perspective.
Richard Currie, Vice President & Actuary, America Re-Insurance Company
Lauren M Bloom, General Counsel, American Academy of Actuaries
John Gleba, Consulting Actuary, Madison Consulting Group
Amy S. Bouska, Consulting Actuary, Tillinghast-Towers Perrin
A Research Update on Techniques in the Valuation of Insurance Companies
Two papers dealing with the fundamental valuation of insurers have recently been completed as part of a funded research project sponsored by the CAS Committee on Valuation, Finance, and Investments. Research teams from the Czech Republic and the United States authored the studies, and representatives of the research teams will present their findings. Final versions of the papers will be made available to meeting attendees.
This two-part session will compare, contrast, and discuss methods commonly used in the valuation of both insurance and non-insurance companies. Panelists will:
- examine developments in techniques used for valuing property/casualty insurance companies, and
- discuss the approaches used to value non-insurance companies.
Ample time will be allowed for questions, audience participation, and feedback.
Robert F. Wolf, Principal, Mercer Oliver Wyman, Inc.
Wayne E. Blackburn, Principal & Consulting Actuary, Milliman USA
Petr Sosik, University of Economics of Prague
Risk Assessment Database (RAD) is a team of researchers organized in response to a Request for Proposal distributed by the Committee on the Theory of Risk to support research on valuing property-liability risks. In this session, researchers from Risk Assessment Database will present the results of their research, which focused on a specific approach used to value property-liability exposures (CAPM). Several important advancements to the traditional CAPM. were incorporated into the methodology used by the team. Among these are
- the use of sumbeta instead of a single beta model, and
- use of the three factor Fama-French model instead of a single factor model.
Louise A. Francis, Consulting Principal, Francis Analytics & Actuarial Data Mining Inc.
Professor J. David Cummins, Wharton School, University of Pennsylvania
Richard A. Derrig, Automobile Insurers Bureau of Massachusetts
Professor Richard D. Phillips, Georgia State University
Discussion Papers on Generalized Linear Models - Session 1
Severity Distributions for GLMs: Gamma or Lognormal? Evidence from Monte Carlo Simulations
By Luyang Fu, Grange Insurance Company
Richard B. Moncher, Bristol West Insurance Group
Insurance claim costs have been found in numerous studies to be positive and usually positively skewed with variances often proportional to the mean squared. In practice, the gamma and lognormal distributions are the ones with those desired properties most widely used. Most actuarial research in GLMs also report results from normal distributions as a comparison. In this study, we apply Monte Carlo simulation techniques to examine the unbiasedness and stability of the GLM classification relativities assuming gamma, lognormal, and normal distributions. We find that the gamma distribution provides better predictive accuracy and efficiency.
A Primer on the Exponential Family of Distributions
By David R. Clark, American Re-Insurance Company
Charles A. Thayer, American Re-Insurance Company
Generalized Linear Model (GLM) theory represents a significant advance beyond linear regression theory, specifically in expanding the choice of probability distributions from the Normal to the Natural Exponential Family. This Primer is intended for GLM users seeking a handy reference on the model's distributional assumptions. The Exponential Family of Distributions is introduced, with an emphasis on variance structures that may be suitable for aggregate loss models in property/casualty insurance.
Discussion Papers on Generalized Linear Models - Session 2
Multivariate Spatial Analysis of the Territory Rating Variable
By Serhat Guven, United Services Automobile Association
The average insurer typically utilizes some form of territory ratemaking in its algorithm; thus, in constructing a GLM, one of the major issues revolves around how to reflect location in the statistical solution. The problem arises because there are too many territory categories to directly include in the statistical model. This issue can be resolved by altering the perception of the location dimension from a categorical rating variable to a continuous one.
This paper presents an alternative approach to incorporating the location dimension in the GLM analysis of the rating algorithm. The procedure develops the indicated relativities and boundaries in a statistical multidimensional framework thus removing the distributional effects of other rating variables and measuring the geographic risk alone. Furthermore, the territory procedure is based on the principle of locality, i.e., the expected loss experience at location L is similar to the loss experience around L.
The indicated relativities of each geographic unit are determined by modeling polynomial functions of latitude and longitude in the GLM statistical framework. By expressing the indication in terms of a polynomial the analyst can include location in the statistical model without having to worry about too many additional parameters.
A Practitioner's Guide to Generalized Linear Models
By Duncan Anderson, Watson Wyatt LLP
Sholom Feldblum, Liberty Mutual Group
Claudine Modlin, Watson Wyatt Insurance & Financial Services
Doris Schirmacher, Liberty Mutual Group
Ernesto Schirmacher, Liberty Mutual Group
Neeza Thandi, Liberty Mutual Group
The Practitioner's Guide to Generalized Linear Models is written for the practicing actuary who would like to understand generalized linear models (GLMs) and use them to analyze insurance data. The guide is divided into three sections.
Section 1 provides a foundation for the statistical theory and gives illustrative examples and intuitive explanations which clarify the theory. The intuitive explanations build upon more commonly understood actuarial methods such as linear models and the minimum bias procedures.
Section 2 provides practical insights and realistic model output for each stage of a GLM analysis - including data preparation and preliminary analyses, model selection and iteration, model refinement and model interpretation. This section is designed to boost the actuary's confidence in interpreting GLMs and applying them to solve business problems.
Section 3 discusses other topics of interest relating to GLMs such as retention modeling and scoring algorithms.
Discussion Papers on Generalized Linear Models - Session 3
A Practitioner's Approach to Marine Liability Pricing Using Generalised Linear Models
By Brian Gedalla, Milliman UK
Denise Jackson, Milliman UK
David Sanders, Milliman UK
Marine Liability underwriters - notably those at the Protection and Indemnity (P&I) Clubs - have traditionally used empirical approaches based on individual risk experiences to arrive at their pricing. But P&I is a direct class of insurance and the underwriters have at their disposal significant data volumes. This means that it is more than possible to apply the kind of modeling techniques to P&I (and, for that matter, to other classes in the marine sector) that have become commonplace elsewhere in the General Insurance (Property & Casualty) world. In this paper we note the traditional methods, the data available and indicate how the Generalised Linear Modeling technique can be used to derive rating models that apply to Marine Liability business.
Loss Reserving with GLMs: A Case Study
By Greg Taylor, Taylor Fry Consulting Actuaries and University of Melbourne
Grainne McGuire, Taylor Fry Consulting Actuaries
This paper provides a case study in the application of generalised linear models ("GLMs") to loss reserving. The study is motivated by approaching the exercise from the viewpoint of an actuary with a predisposition to the application of the chain ladder ("CL").
The data set under study is seen to violate the conditions for application of the CL in a number of ways. The difficulties of adjusting the CL to allow for these features of the data are noted (Section 3).
Regression, and particularly GLM regression, is introduced as a structural and rigorous form of data analysis. This enables the investigation and modeling of a number of complex features of the data responsible for the violation of the CL conditions. These include superimposed inflation and changes in the rules governing the payment of claims (Sections 4 to 7).
The development of the analysis is traced in some detail, as is the production of a range of diagnostics and tests used to compare candidate models and validate the final one. The benefits of this approach are discussed in Section 8.
Discussion Papers on Generalized Linear Models - Session 4
Estimating Claim Settlement Values Using GLM
Roosevelt C. Mosley, Jr., Pinnacle Actuarial Resources, Inc.
The goal of this paper is to demonstrate how generalized linear modeling (GLM) can be applied in non-traditional ways in property and casualty insurance. Specifically, we will use a property and casualty closed claims database to aid in estimating ultimate claim settlement amounts, evaluating claim trends, and assisting in improving claims handling procedures. This specific example will be used to demonstrate the potential of the application of GLM to different areas of an insurance company.
A GLM will be developed with data from the Insurance Research Council (IRC) closed claims study. The model will be populated with characteristics of closed automobile claims along with final settlement amounts. Using this data, the paper will examine how GLM can be used to identify:
- Trends in claims severities over time,
- Differences in severities that exist between current ratemaking characteristics (e.g. state, territory), characteristics of the claims and the injured parties, and other factors (e.g. time from reporting to settlement, attorney involvement, use of arbitration), and
- Interactions between these factors.
Diagnostics will also be discussed which can be used to test the validity and robustness of the GLM models that are developed, and several applications of the results of this type of analysis will be presented.
The Case of the Medical Malpractice Crisis: A Classic Who Dunnit
Robert J. Walling III, Pinnacle Actuarial Resources, Inc.
We actuaries, detectives of the first order, are presented with a most intriguing case: numerous, grisly bodies of dead insurance companies and physicians' practices in public view, various signs of intrigue and foul play abound, suspects galore, an abundance of alibis, and an endless supply of opinions on how the culprit(s) must repay their debt to society. This case of the medical malpractice crisis is complicated because there is not only no consensus on "who dunnit?" but not even an agreement on "what happened?" This is the situation we are currently faced with in the medical malpractice insurance industry. There is evidence scattered all over the medical malpractice insurance landscape, but there is no agreement at all on the cause, the culprit, the motives or the appropriate sentence.
Discussion by Michael G. Wacek
Proceedings PapersStephen P. D'Arcy and Michael A. Dyer's "Ratemaking: A Financial Economic Approach"
In their 1997 Proceedings paper D'Arcy and Dyer survey the property-casualty ratemaking landscape from a financial economics perspective. This discussion is devoted entirely to Section 8 of their paper, which deals with a method drawing upon Option Pricing Theory (OPT). The purpose of this discussion is to correct, clarify and extend the work of D'Arcy and Dyer. The author will point out and correct what he sees as the paper's shortcomings, rework and extend the examples, and expand the exposition to allow for a more complete treatment of taxes and claims that vary stochastically.
David Ruhm's "Distribution-Based Pricing Formulas are Not Arbitrage-Free"
Discussion by Michael G. Wacek
David Ruhm is entirely correct that risk load formulas based on transforming probability distributions of contract outcomes cannot guarantee arbitrage-free prices. This is what he illustrates by a clever and entertaining example. But the title of the paper seems to assert that no method of transforming distributions is arbitrage-free. This is not the case, as transforms of the probabilities of the underlying events that generate the outcomes are well known to produce arbitrage-free prices. In fact, Ruhm illustrates this by showing that the Black-Scholes formula arises from such a transform. He also shows that this formula builds in risk adjustments to prices, thus addressing the misapprehension that since the options prices come from a risk-neutral valuation they do not incorporate risk adjustments. To illustrate the application of probability transforms to fundamental events in insurance, this discussion provides an example of using a newly reported transform of underlying frequency and severity distributions to price loss layers.