Erudite Analytics Offerings - Society of Risk Management ...

Erudite Analytics Offerings - Society of Risk Management ...

EVIDENCE-BASED RISK MANAGEMENT D&O and Cyber Threats Casualty Risk SRMC Benchmarking ANDREW D. BANASIEWICZ, PH.D. Drowning in Data and Still Guessing Copyright Andrew D Banasiewicz. All rights reserved. 2 A Word About Myself: A Combination of Practice & Research Two decades of industry, research & predictive analytical experience Four books, numerous articles and white papers Numerous domestic and international conference presentations & keynotes, including RIMS and Global Risk Forum Copyright Andrew D Banasiewicz. All rights reserved. 3 Erudite Analytics Offerings

Executive Risk Structured Solutions Owned event datasets Proprietary algorithms D&O | Cyber Casualty Risk WC | GL Evidence Based Risk Managemen t Ad Hoc Solutions Explore | Predict Custom Analyses Benchmarking

On-demand studies Copyright Andrew D Banasiewicz. All rights reserved. 4 Structured Solutions D&O | CYBER THREATS 5 Conceptual Backdrop: The Meta Model of Executive Risk Copyright Andrew D Banasiewicz. All rights reserved. 6 Focusing on Securities Class Actions (SCA): Data & Analyses Input: Cause Key Events: IPO; M&A Restatements Accounting Accruals

Key Ratios from Balance Sheet & Income Statement Downside Stock Price Volatility Firmographics Input: Outcome End Result Data Amalgamation, Processing & Analyses Company-Specific Likelihood & Severity Estimates SCA Filings & Settlements Copyright Andrew D Banasiewicz. All rights reserved. 7 SCA Likelihood Estimation

Input data sample: Confusion Matrix: Actual vs. Predicted 441 verified public co SCA filings (last 3 years); A sample of 912 of non-SCA public companies; All available firmographic, financial & stock movement data Predicted SCA Filings No Actual SCA Filings Yes Total Total No Yes Count 792

120 912 % 87% 13% 100% Count 210 231 441 100% % 48% 52%

Count 1,002 351 1,353 % 74% 26% 100% SCA Overall avg. No SCA Regression Trees Copyright Andrew D Banasiewicz. All rights reserved. 8

SCA Likelihood Scoring: All Public US Companies Each public US company was scored with the earlier-derived several hundred regression trees, the results of which were averaged to arrive at company-specific SCA likelihood, leading to the aggregate distribution (left) and classificatory typology (right). Aggregate SCA Score Distribution Mean = .304 Percentiles Std. Dev. = .186 N = 11,019 N Mean Median Std. Deviation Minimum Maximum 10 20 25 30 40 50 60 70

75 80 90 95 99 11,019 0.304 0.283 0.186 0.005 0.913 0.086 0.130 0.146 0.172 0.228 0.283 0.327 0.385 0.422 0.459 0.563 0.665 0.816 Classificatory

Typology Low Moderate Elevated High Objective assessment of their own unique exposure enables companies to make more rational risk transfer (i.e., D&O coverage) choices for instance, Low SCA likelihood suggests higher SIR Copyright Andrew D Banasiewicz. All rights reserved. 9 SCA Severity Estimation Multi-Attribute Indexing Key Highlights: Estimated Loss Ranges for US Exchanges Traded Companies Based on hand-collected sample of 460 individual, verified SCA settlements; Explicitly takes into accounts industry sector differences; Explicitly takes into account regulatory

(PSLRA of 96; SOX) and judicial (US Supreme Court) acts and rulings; Adjusts for market cap differences; Nets out the expected levels of stocks downside variability (loss causation); Non-linear market cap vs. exposure estimation (e.g., 3x market cap does not automatically mean 3x higher loss); Metric No. Companies Mean Median 10 20 25 30 40 50 60 70 75 80 90 95 Percentiles

Company-specific estimation; Expected Loss: Low Expected Loss: High 7,972 $87,211,511 $5,682,518 $200,193 $673,016 $999,476 $1,421,357 $2,752,702 $5,682,518 $11,891,924 $25,434,429 $36,417,244 $55,620,211 $153,910,910 $348,942,557 7,972 $147,920,751 $10,096,950 $447,818 $1,222,868 $1,775,219 $2,457,549 $4,746,858

$10,096,950 $20,621,626 $43,243,010 $61,833,249 $96,342,381 $260,261,239 $581,832,316 Expected loss expressed as pseudo credible intervals; Copyright Andrew D Banasiewicz. All rights reserved. 10 Company-Specific Exposure Reports: Focusing on What Matters Copyright Andrew D Banasiewicz. All rights reserved. 11 Structured Solutions CASUALTY RISKS: WC | GL 12 Meeting Obligations in Fiscally Sound Manner

Typically, casualty claim management is highly reactive, as claim handlers take action in response to claim development The available data makes it possible for claim handlers to act proactively, which will result in lower claim costs Organizations have legal duties when casualty accidents WC/GL arise, but organizations also need to manage their expenses responsibly. Making sound use of the available data, organizations can accomplish both objectives. Copyright Andrew D Banasiewicz. All rights reserved. 13 Managing the Total Cost of Casualty Risk Claim Management Loss Prevention Incidence Analytics Risk Financing

Newly-Open Claims Residual Claims (1 year or less since opening) (1+ years since opening) Which individual claims are likely to develop adversely? How are aggregate claim carrying costs trending? What factors inflate shortterm claim costs? What factors inflate longterm claim costs? Predictive Claim Scoring Claim Carrying Cost Index Financial Analytics Period Scorecard & Cumulative Knowledge Base Copyright Andrew D Banasiewicz. All rights reserved.

14 Predictive Claim Scoring --For Illustrative Purposes Only-An analytic method for estimating the expected future cost of individual workers comp claims, and for identifying the most promising risk reduction actions and strategies. Copyright Andrew D Banasiewicz. All rights reserved. 15 Claim Carrying Cost Index (C3I) C3I enables claim managers to make meaningful cross-time assessment of cost development trends and root causes of the ever-changing (i.e., hard to compare) mix of longer duration (1 year+ since inception) claims. Copyright Andrew D Banasiewicz. All rights reserved. 16 Ad Hoc Solutions SRMC BENCHMARKING

17 Imagine If as SRMC Member You Could Benchmark your clients coverage adequacy in terms of: Cost / Unit of Coverage Limit / Unit of Asset Base Key Endorsements High Endorsement A Avg. Endorsement B Endorsement C Endorsement n Low While also keeping track of changes over time Benchmark your clients WC loss experience in terms of: WC Cost / Revenue

High WC Cost Over Time $$ Avg. $ Low T1 T2 T3 Other Ideas Tn Copyright Andrew D Banasiewicz. All rights reserved. 18

How Would It Work? 1. We would, jointly, develop a standard data submission form 2. Once finalized, the form would be translated into an electronic data capture format (e.g., Survey Monkey) a. The basic organizational structure of data would be Client_ID and Year; all other variables would be treated as attributes b. There is usually a fairly small ($300/year for SurveyMonkey) cost for data capture services that cost would be covered by SRMC 3. Once the electronic data submission system was operational, you would complete the data capture form for each one of your clients a. optionally, we could consider going up to 5 years back for instant longitudinal tracking 4. I would build a searchable database, develop agreed upon benchmarking metrics, and maintain it on ongoing basis 5. There would be no charge to SRMC members for basic benchmarking (to be defined); there would be a charge for any deeper dive / nonstandard work [email protected] www.eruditeanalytics.com 19

Recently Viewed Presentations

  • hdhr.org

    hdhr.org

    HT in Slovenia in the last decade - risks and benefits Damir Franić MD PhD, Slovenia Prof. Helena Meden-Vrtovec MD PhD, Slovenia Ivan Verdenik PhD, Slovenia VI Hrvaško-slovenski
  • BENEFITS AND COMPENSATION - Guilford County Schools

    BENEFITS AND COMPENSATION - Guilford County Schools

    A 6 month break is required for all retirees returning to state employment. During the 6 month break the employee shall not work in any capacity in a organization participating in the Teachers' and State Employees' Retirement System, including part-time,...
  • Chapter 1 Introduction to Science Ch 1 The

    Chapter 1 Introduction to Science Ch 1 The

    Introduction to Science Chapter 1 Ch 1 The Nature of Science Table of Contents Section 1 The Nature of Science Section 2 The Way Science Works Section 3 Organizing Data
  • Rational Decision Making Model - Siue

    Rational Decision Making Model - Siue

    rational decision making model 1. recognize, define problem opportunity rational decision making model gather "relevant" info diagnose cause(s) assumptions! be aware of them understand them problems with defining problem task instructions perceptions "experience" "mental sets ...
  • Ion Sources with Muons, Inc  ICIS 2013. Improving

    Ion Sources with Muons, Inc ICIS 2013. Improving

    Dear Slava, It is very exciting that you will be honoring Vadim on his 70th anniversary. I had hoped to find some old pictures of Vadim to share with you, but after coming up empty, I now appreciate more the...
  • Icebreaker

    Icebreaker

    According to Echevarria, Vogt, and Short (2000), many ELLs need to learn a variety of strategies that fall into three categories: metacognitive, cognitive, and social/affective. They referenced Chamot and O'Malley's CALLA Handbook as a great resource for teaching English language...
  • Breeds of Swine - Schs Ag program

    Breeds of Swine - Schs Ag program

    Breeds of Swine Northeastern Agricultural Education Mr. Caffee Pig Facts Gestation Length 3m 3w 3d Breeding/Market 10years/5-6mon Confinement Ear Marking Protein Feed Landrace Originated in Denmark Drooped ears Known for their maternal instincts White Long Flatter-topped than other breeds Berkshire...
  • Module 3 - Substance Abuse and Mental Health Services ...

    Module 3 - Substance Abuse and Mental Health Services ...

    Mental health disorders are not something people "have" or "don't have"—mental health is a continuum, with mental health disorders at one end. It's better if we think of mental health disorders as existing on a spectrum or along a range:...