Enterprise wide Economic Capital Model using a structured

Enterprise wide Economic Capital Model using a structured and integrated modeling platform Patrick Grealy FIA Israel June 2012

How to Develop an Internal Model • Review Key stages of the process • New business including reinsurance • Prior year business run off • Include Opening Balance Sheet items (Assets, UPR) • Additional deliverables from the implementing of the Model • How to integrate this into the Company’s systems and structures and also to use the system on a day to day basis • Look at a fully functioning example

Environment Overview Inflation Natural Catastrophe Group Co 1 FX Changes Legal – Secular Your Risks/ Cedants Your Company Assets Your Reinsurer 1 Interest Rates Group Co 2 Your Reinsurer 2

Additional Deliverables from Internal Model. • Pricing of new business • Reinsurance Optimization & Scenario Testing • Calculate Economic Capital Required (“ECM”) • Investment decision making • Counterparty Credit exposure Assessment • Cash-flow & Treasury Management • Defining Risk Appetite • Capital Strategies e. g. internal reinsurance structures • Acquisition & Divestiture

The Key Stages of Process • Decision to develop a Model • Assess internal/external resources available • Consider the form of the Model 1. Spreadsheet; 2. Programming Language; 3. Toolbox (Specialist programming environment) [Generally provided by Brokers and Consultants] ; and 4. Fully Compiled integrated product [Generally provided by professional software Companies] • Specify timelines and deliverables. • 1 Calendar Year or Multi Year • Full or Partial Capital Model • Start, Develop, Revise, Use

New Business • May have very useful information from Business plan and also current underwriting and pricing Studies • Often the area where most of the modeling is centered • Need assumptions for • Business Volume • Expenses(ALAE And ULAE) • Claims (split by type) • May have a complex reinsurance structure to model • Again look to split data losses into attritional, Large and Catastrophic. (incorporating RMS/EQEcat/AIR etc) • Often the scale of these data items become large • Correct Choice of Catastrophe Model

Reinsurance & Mitigation • Reinsurance Structures can be very complicated particularly where there are inuring reinsurances and also intra group. • Especially when there are non traditional covers in place • Possible that prior years reinsurance details are even more problematic. • Can also factor in Bad Debt both on New business and Old Business. Often with the same counterparty. • Sometimes there are possible and serious exhaustion issues, especially in the case of Run off entities. • Where Catastrophe covers are written, need to ensure that losses are modeled at a granularity and sophistication to capture Event identifiers. • Often a multitude of various reinsurance structures are superimposed on Gross Losses to find “efficient” structures.

Assets • Not usually the main area of study/analysis for non life Actuaries • Can indeed be a significant source of Balance sheet Volatility if there are significant FX exposures and also mismatched terms (e. g. short tail bonds/cash against Workers Comp Liabilities) • Need to deploy the macro economic scenarios created by an Economic Scenario Generator to revalue Assets at future balance sheet dates and Dividend Yield and interest income. • Link/drive the performance of the some of the relevant liability items to these macro economic variables to create realistic liability valuations • FX is often the main area of concern for Non life (re)insurers together with Inflation and also GDP growth and unemployment rates which can drive attritional losses and also claims on Credit / Bond business. • Also consider Index Linked Assets (Cat Bonds) for matched liabilities

Correlation • Generally 3 types to consider. 1. Causal • Certain Events / Drivers Act to affect Asset and Liability values often with many separately modeled items being affected simultaneously • Example would be a inflation driver affecting the loss payments on classes that were subject to inflation 2. Copula Based • Matrix of correlation coefficients that are used when sampling random variables used to generate losses • Often used to generate loss ratios on attritional losses • More complex copulas (Clayton, Gumbel) to create Tail Dependency 3. Environmental / Economic • Economic environment affecting Asset Values • Change in spread and yield curves affecting Bond Yields

Parameters & Integration into Core Function of the Company Actuarial Assumptions (Loss Ratios, Correlations) Macro Economic ESG (FX, GDP, Inflation, Unemployment, Interest) Data from IT System (Premium, Exposures) Calculation Kernel (Internal Capital Model) Catastrophe Event Files (RMS, AIR, Eqecat) Reports Reinsurance Covers