Portfolio Loss Distribution



Similar documents
THE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek

Causal, Explanatory Forecasting. Analysis. Regression Analysis. Simple Linear Regression. Which is Independent? Forecasting

How To Evaluate A Dia Fund Suffcency

Analysis of Premium Liabilities for Australian Lines of Business

Vasicek s Model of Distribution of Losses in a Large, Homogeneous Portfolio

Optimal Bidding Strategies for Generation Companies in a Day-Ahead Electricity Market with Risk Management Taken into Account

Measuring portfolio loss using approximation methods

How To Calculate The Accountng Perod Of Nequalty

Risk Management and Financial Institutions

Credit Limit Optimization (CLO) for Credit Cards

Stress test for measuring insurance risks in non-life insurance

THE METHOD OF LEAST SQUARES THE METHOD OF LEAST SQUARES

NPAR TESTS. One-Sample Chi-Square Test. Cell Specification. Observed Frequencies 1O i 6. Expected Frequencies 1EXP i 6

Lecture 14: Implementing CAPM

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ).

Discount Rate for Workout Recoveries: An Empirical Study*

Underwriting Risk. Glenn Meyers. Insurance Services Office, Inc.

Reporting Forms ARF 113.0A, ARF 113.0B, ARF 113.0C and ARF 113.0D FIRB Corporate (including SME Corporate), Sovereign and Bank Instruction Guide

Risk Model of Long-Term Production Scheduling in Open Pit Gold Mining

Forecasting the Direction and Strength of Stock Market Movement

Robust Design of Public Storage Warehouses. Yeming (Yale) Gong EMLYON Business School

L10: Linear discriminants analysis

Risk-based Fatigue Estimate of Deep Water Risers -- Course Project for EM388F: Fracture Mechanics, Spring 2008

Can Auto Liability Insurance Purchases Signal Risk Attitude?

Fixed income risk attribution

The OC Curve of Attribute Acceptance Plans

An Alternative Way to Measure Private Equity Performance

The impact of hard discount control mechanism on the discount volatility of UK closed-end funds

Kiel Institute for World Economics Duesternbrooker Weg Kiel (Germany) Kiel Working Paper No. 1120

Copulas. Modeling dependencies in Financial Risk Management. BMI Master Thesis

1. Measuring association using correlation and regression

On the computation of the capital multiplier in the Fortis Credit Economic Capital model

arxiv: v1 [q-fin.pm] 6 Sep 2011

Method for assessment of companies' credit rating (AJPES S.BON model) Short description of the methodology

Efficient Project Portfolio as a tool for Enterprise Risk Management

Characterization of Assembly. Variation Analysis Methods. A Thesis. Presented to the. Department of Mechanical Engineering. Brigham Young University

Outline. Investment Opportunity Set with Many Assets. Portfolio Selection with Multiple Risky Securities. Professor Lasse H.

CHAPTER 14 MORE ABOUT REGRESSION

CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK. Sample Stability Protocol

The impact of bank capital requirements on bank risk: an econometric puzzle and a proposed solution

SDN: Systemic Risks due to Dynamic Load Balancing

ENTERPRISE RISK MANAGEMENT IN INSURANCE GROUPS: MEASURING RISK CONCENTRATION AND DEFAULT RISK

Applied Research Laboratory. Decision Theory and Receiver Design

A Simplified Framework for Return Accountability

Face Verification Problem. Face Recognition Problem. Application: Access Control. Biometric Authentication. Face Verification (1:1 matching)

Chapter 4 ECONOMIC DISPATCH AND UNIT COMMITMENT

The Choice of Direct Dealing or Electronic Brokerage in Foreign Exchange Trading

Abstract. 260 Business Intelligence Journal July IDENTIFICATION OF DEMAND THROUGH STATISTICAL DISTRIBUTION MODELING FOR IMPROVED DEMAND FORECASTING

STATISTICAL DATA ANALYSIS IN EXCEL

Débats économiques et financiers N 1

Chapter 7: Answers to Questions and Problems

Mean Molecular Weight

A Model of Private Equity Fund Compensation

CHAPTER 5 RELATIONSHIPS BETWEEN QUANTITATIVE VARIABLES

1 De nitions and Censoring

CEIOPS-DOC-42/09. (former CP 49) October 2009

THE USE OF RISK ADJUSTED CAPITAL TO SUPPORT BUSINESS DECISION-MAKING

Macro Factors and Volatility of Treasury Bond Returns

What is Portfolio Diversification? What a CAIA Member Should Know

Outline. CAPM: Introduction. The Capital Asset Pricing Model (CAPM) Professor Lasse H. Pedersen. Key questions: Answer: CAPM

An Analysis of Pricing Methods for Baskets Options

Fragility Based Rehabilitation Decision Analysis

Construction Rules for Morningstar Canada Target Dividend Index SM

ECONOMICS OF PLANT ENERGY SAVINGS PROJECTS IN A CHANGING MARKET Douglas C White Emerson Process Management

Transition Matrix Models of Consumer Credit Ratings

Statistical Methods to Develop Rating Models

Marginal Benefit Incidence Analysis Using a Single Cross-section of Data. Mohamed Ihsan Ajwad and Quentin Wodon 1. World Bank.

Measurement of Farm Credit Risk: SUR Model and Simulation Approach

Imperial College London

Linear Circuits Analysis. Superposition, Thevenin /Norton Equivalent circuits

Economic Interpretation of Regression. Theory and Applications

Trackng Corporate Bond Ndces

RELIABILITY, RISK AND AVAILABILITY ANLYSIS OF A CONTAINER GANTRY CRANE ABSTRACT

Most investors focus on the management

INVESTIGATION OF VEHICULAR USERS FAIRNESS IN CDMA-HDR NETWORKS

Staff Paper. Farm Savings Accounts: Examining Income Variability, Eligibility, and Benefits. Brent Gloy, Eddy LaDue, and Charles Cuykendall

Enabling P2P One-view Multi-party Video Conferencing

Forecasting and Stress Testing Credit Card Default using Dynamic Models

Lecture 5,6 Linear Methods for Classification. Summary

PAS: A Packet Accounting System to Limit the Effects of DoS & DDoS. Debish Fesehaye & Klara Naherstedt University of Illinois-Urbana Champaign

Calibration and Linear Regression Analysis: A Self-Guided Tutorial

Prediction of Disability Frequencies in Life Insurance

Planning for Marketing Campaigns

Part 1: quick summary 5. Part 2: understanding the basics of ANOVA 8

How To Find The Dsablty Frequency Of A Clam

HOUSEHOLDS DEBT BURDEN: AN ANALYSIS BASED ON MICROECONOMIC DATA*

Online Appendix Supplemental Material for Market Microstructure Invariance: Empirical Hypotheses

Risk Measurement and Management of Operational Risk in Insurance Companies from an Enterprise Perspective

Optimal maintenance of a production-inventory system with continuous repair times and idle periods

Exhaustive Regression. An Exploration of Regression-Based Data Mining Techniques Using Super Computation

The Proper Use of Risk Measures in Portfolio Theory

Evaluating credit risk models: A critique and a new proposal

Traffic State Estimation in the Traffic Management Center of Berlin

Valuing Customer Portfolios under Risk-Return-Aspects: A Model-based Approach and its Application in the Financial Services Industry

DEFINING %COMPLETE IN MICROSOFT PROJECT

Rate-Based Daily Arrival Process Models with Application to Call Centers

Portfolio Performance Manipulation and Manipulation-Proof Performance Measures

Analysis of the provisions for claims outstanding for non-life insurance based on the run-off triangles

Bank Credit Conditions and their Influence on Productivity Growth: Company-level Evidence

Power-of-Two Policies for Single- Warehouse Multi-Retailer Inventory Systems with Order Frequency Discounts

Transcription:

Portfolo Loss Dstrbuton

Rsky assets n loan ortfolo hghly llqud assets hold-to-maturty n the bank s balance sheet Outstandngs The orton of the bank asset that has already been extended to borrowers. Commtment A commtment s an amount the bank has commtted to lend. Should the borrower encounter fnancal dffcultes, t would draw on ths commtted lne of credt.

Adjusted exosure and exected loss Let α be the amount of drawn down or usage gven default. Outstandng + α commtment, Rsky Asset value at later tme H, V H Adjusted exosure s the rsky art of V H. (1 α) commtment, Rskless Exected loss adjusted exosure loss gven default robablty of default * Normally, racttoners treat the uncertan draw-down rate as a known functon of the oblgor s end-of-horzon credt class ratng.

Examle calculaton of exected loss Commtment Outstandng Internal rsk ratng Maturty Tye Unused drawn-down on default (for nternal ratng 3) Adjusted exosure on default EDF for nternal ratng 3 Loss gven default for non-secured asset Exected loss $10,000,000 $3,000,000 3 1 year Non-secured 65% $8,250,000 0.15% 50% $6,188

Unexected loss Unexected loss s the estmated volatlty of the otental loss n value of the asset around ts exected loss. AE EDF LGD 2 2 σ LGD + σ 2 EDF where σ EDF 2 EDF (1- EDF). Assumtons * The random rsk factors contrbutng to an oblgor s default (resultng n EDF) are statstcally ndeendent of the severty of loss (as gven by LGD). * The default rocess s two-state event.

Examle on unexected loss calculaton Adjusted exosure $8,250,000 EDF 0.15% σ EDF 3.87% LGD 50% σ LGD 25% Unexected loss $178,511 * The calculated unexected loss s 2.16% of the adjusted exosure, whle the exected loss s only 0.075%

Comarson between exected loss and unexected loss * The hgher the recovery rate (lower LGD), the lower s the ercentage loss for both EL and. * EL ncreases lnearly wth decreasng credt qualty (wth ncreasng EDF) * ncreases much faster than EL wth ncreasng EDF. Percentage loss er unt of adjusted loss 10% 5% 10% EL EDF

Assets wth varyng terms of maturty * The longer the term to maturty, the greater the varaton n asset value due to changes n credt qualty. * The two-state default rocess aradgm nherently gnores the credt losses assocated wth defaults that occur beyond the analyss horzon. * To mtgate some of the maturty effect, banks commonly adjust a rsky asset s nternal credt class ratng n accordance wth ts terms to maturty.

Portfolo exected loss where EL EL AE LGD EDF EL s the exected loss for the ortfolo, AE s the rsky orton of the termnal value of the th asset to whch the bank s exosed n the event of default. We may wrte EL AE w EL AE where the weghts refer to AE w AE AE AE.

EL AE EL AE AE AE EL AE w EL AE AE w EL EL /AE 1 $10 M 0.5 $1 0.1 2 $4 M 0.2 $0.5 0.125 3 $6 M 0.3 $0.6 0.1 AE $20M 1 w EL AE 0.5 0.1+ 0.2 0.125 + 0.3 0.1 0.105

Portfolo unexected loss ortfolo unexected loss j ρ j w w j j where AE EDF 2 2 2 σ LGD + GD σ E L DF and ρ j s the correlaton of default between asset and asset j. Due to dversfcaton effect, we exect <<.

Rsk contrbuton The rsk contrbuton of a rsky asset to the ortfolo unexected loss s defned to be the ncremental rsk that the exosure of a sngle asset contrbutes to the ortfolo s total rsk. RC and t can be shown that RC j j ρ j.

Undversfable rsk The rsk contrbuton s a measure of the undversfable rsk of an asset n the ortfolo the amount of credt rsk whch cannot be dversfed away by lacng the asset n the ortfolo. RC To ncororate ndustry correlaton, usng ndustry α and j ndustry β RC (1 ) + α ραα β α k β k ρ αβ.

Calculaton of EL, and RC for a two-asset ortfolo ρ EL default correlaton between the two exosures ortfolo exected loss EL EL 1 + EL 2 ortfolo unexected loss RC 1 RC 2 + 2 2 1 2 2 12 rsk contrbuton from Exosure 1 rsk contrbuton from Exosure 2 + ρ RC1 1(1 + ρ2) / RC2 2(2 + ρ1) / RC 1 + RC 2 << 1 + 2

Fttng of loss dstrbuton The two statstcal measures about the credt ortfolo are ortfolo exected loss; ortfolo unexected loss. At the smlest level, the beta dstrbuton may be chosen to ft the ortfolo loss dstrbuton. Reservaton A beta dstrbuton wth only two degrees of freedom s erhas nsuffcent to gve an adequate descrton of the tal events n the loss dstrbuton.

Beta dstrbuton The densty functon of a beta dstrbuton s > > < < Γ Γ + Γ otherwse 0 0 0, 1 0, ) (1 ),, ( 1 1 ) ( ) ( ) ( β α β α β α β α β α x x x x F Mean and varance β α α µ +. 1) ( ) ( 2 2 + + + β α β α αβ σ 1 x f(x, α, β)

Economc Catal If X T s the random varable for loss and z s the ercentage robablty (confdence level), what s the quantty v of mnmum economc catal EC needed to rotect the bank from nsolvency at the tme horzon T such that Pr[ X T v] z. Here, z s the desred debt ratng of the bank, say, 99.97% for an AA ratng.

frequency of loss X T EL EC

Catal multler Gven a desred level of z, what s EC such that Pr[ X EL EC] T z. Let CM (catal multler) be defned by EC CM then Pr X T EL CM z.

Monte Carol smulaton of loss dstrbuton of a ortfolo 1. Estmate default and losses 2. Estmate asset correlaton between oblgors Assgn rsk ratngs to loss facltes and determne ther default robablty + Assgn LGD and σ LGD Determne arwse asset correlaton whenever ossble OR Assgn oblgors to ndustry groungs, then determne ndustry ar correlaton

3. Generate random loss gven default 4. Generate correlated default events Determne stochastc loss gven default + Correlated default events + Decomoston of covarance matrx + Smulate default ont

5. Loss calculaton 6. Loss dstrbuton Calculate faclty loss for each scenaro and obtan ortfolo loss Construct smulated ortfolo loss dstrbuton

Generaton of correlated default events Generate a set of random numbers drawn from a standard normal dstrbuton. Perform a decomoston (Cholesky, SVD or egenvalue) on the asset correlaton matrx to transform the ndeendent set of random numbers (stored n the vector e ) nto a set of correlated asset values (stored n the vector e ). Here, the transformaton matrx s M, where e M e. The covarance matrx and M are related by M T M.

Calculaton of the default ont The default ont threshold, DP, of the th oblgor can be defned as DP N 1 (EDF, 0, 1). The crteron of default for the th oblgor s default f e < ' DP no default f e ' DP.

Generate loss gven default The LGD s a stochastc varable wth an unknown dstrbuton. A tycal examle may be Recovery rate (%) LGD (%) σ LGD (%) secured 65 35 21 unsecured 50 50 28 LGD LGD + σ s s f LGD where f s drawn from a unform dstrbuton whose range s selected so that the resultng LGD has a standard devaton that s consstent wth hstorcal observaton.

Calculaton of loss Summng all the smulated losses from one sngle scenaro Loss Adjusted exosure Oblgors n default LGD Smulated loss dstrbuton The smulated loss dstrbuton s obtaned by reeatng the above rocess suffcently number of tmes.

Features of ortfolo rsk The varablty of default rsk wthn a ortfolo s substantal. The correlaton between default rsks s generally low. The default rsk tself s dynamc and subject to large fluctuatons. Default rsks can be effectvely managed through dversfcaton. Wthn a well-dversfed ortfolo, the loss behavor s characterzed by lower than exected default credt losses for much of the tme, but very large losses whch are ncurred nfrequently.