1 CAS Casualty Loss Reserve Seminar Fair Value Accounting E. Daniel Thomas, PricewaterhouseCoopers LLP Christina Gwilliam, Towers Perrin Moderator Don.

1

CAS Casualty Loss Reserve Seminar

Fair Value Accounting

E. Daniel Thomas, PricewaterhouseCoopers LLP

Christina Gwilliam, Towers Perrin

Moderator

Don Rainey, Milliman

2

Overview

• Brief history of fair value

• PwC paper

• Towers Perrin paper

• IFRS 4 update

3

Brief history of Fair Value

• In the beginning, accounting was simple:

– Assets valued at historical cost – reliable, but not relevant

• Early 1980s -- Savings & Loans Crisis, banks were in trouble:– Sell assets where market value over book value– Keep assets where market value under book value

• Result was hidden insolvency

• Proposed solution – Fair Value accounting

4

What is Fair Value?

“The price at which an asset or liability could be exchanged in a transaction between knowledgeable unrelated willing parties.”

Fair value hierarchy exists:1. Market value, if sufficient market exists; else2. Market value of similar items with active markets, adjusted

for differences; else3. Present value estimate of future cash flows, including an

adjustment for risk.

5

Brief history of Fair Value

• 1990s – Capital markets became increasingly global

• Differing accounting standards are impediment to free flowing capital across borders, with perception of lack of transparency

• International Accounting Standards Board (IASB)

• The European Union wants consistency within its borders

• European Union has decided to move toward the IASB’s approach to accounting (i.e. Fair Value)

6

Why should I care?

• You clearly should if you have a European parent

• But SEC is in favor of eventual consistency (“convergence”)– SEC controls U.S. GAAP for publicly owned companies– FASB also in favor of convergence

• FASB and IASB agreed not to introduce anything that furthers the gap to some joint work on the issue– FASB has unofficially agreed to join with the IASB in a

limited joint project on insurance contracts

7

Practical applications

• Discussion at IASB has at times been theoretical and academic

• Life insurance industry wanted a “real world” test of fair value– IAA and ACLI teamed up to conduct such tests

• What about P/C Insurance? – In 2003, CAS sponsored research into practical applications

• PwC and Towers Perrin contributed to this research

8

PwC Paper

9

Outline of PwC Research

• CAS Project Objectives & Scope Exclusions

• Data

• Modeling Approaches

• Findings

• Significant Issues

10

CAS project – objectives

• Evaluate impact of the application of fair value principles on U.S. insurance company loss and LAE reserves

• Identify significant issues associated with the usefulness of fair values in insurance company financial statements

11

Research specifications

• Use publicly available data (e.g. Schedule P)

• 3 lines of business:– Personal Auto Liability (shorter tail)– Workers’ Compensation (long tail – stable)– Medical Malpractice, Claims Made (long tail – volatile)

• Measure impact of:– Discounting– Market risk load

12

CAS project – scope exclusions

• Credit risk

• Adequacy of booked reserves

• Correlation adjustments across lines of business

• Impact of fair value on other balance sheet items

13

Fair Value measurement

• Active trading markets for loss reserves do not exist

• Fair value measurement was based on models using market concepts:- Undiscounted estimate of future payments;- Discounted for time value of money, plus- Margin for risk and uncertainty (“Market Value Margin”

or “MVM”).

14

PwC modeling approaches

• Discount Factor Models– Duration– Matched to yield curve

• MVM Models– Development method – standard deviation– Stochastic simulation – standard deviation– Stochastic simulation – percentile distribution– Return on Capital

15

PwC modeling calibration approach

• Calibrated to Cost of Capital Method at year-end 2002

• Straight average of three companies (1 large, 1 medium, 1 small)

• All lines assume a 10% target rate of return on capital

ClassDevelopment Method

SD MultipleStochastic Method

SD MultipleStochastic Method

Percentile

Personal Auto Liability 1.2 2.0 90%Workers Compensation 1.0 1.5 92%Medical Malpractice 1.5 2.3 95%

16

Discount factor models

Average discount factor by year for Personal Auto Liability

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

7.0%

8.0%

9.0%

10.0%

1998 1999 2000 2001 2002

Year

Dis

cou

nt

fact

or

Duration

Matched

17

Findings – discount modeling

• Well-defined approaches are available.

• In general, we observed no significant differences between duration and matching approaches.

• Results can be affected by interest rate fluctuations and shape of the yield curve.

18

MVM by company – personal auto liability

MVM by Company at Year-End 2002

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 2 4 6 8 10 12

Company

MV

M %

ROC MVM

19

MVM by method for one company– personal auto liability

MVM by Method - Company C(M)

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

1998 1999 2000 2001 2002Year-End

MV

M %

Develop SD

Stoch SD

Percentile

ROC

20

MVM by company size - personal auto liability

MVM for All Companies and All years

0%

10%

20%

30%

40%

50%

60%

70%

0 200,000 400,000 600,000 800,000 1,000,000

Reserves ($000)

MV

M (

%)

MVM vs.CompanySize

21

MVM by company– workers’ compensation

MVM by Company at Year-End 2002

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 2 4 6 8 10 12

Company

MV

M %

ROC MVM

22

Findings – MVM modeling

• Indications for MVM varied, sometimes significantly:– by method, for a given company and year-end;– over time, for a given company and MVM method.

• The ranking of MVM’s by method tended to vary over time:– No method consistently was the highest nor the lowest.

• For smaller companies, the MVM tended to be larger (measured as a percentage of the loss reserves)

23

Findings – Fair Value impact on balance sheet (loss reserves)

• Personal auto liability: FV reserves were generally greater than U.S. GAAP reserves

• Workers’ compensation: FV reserves were generally less than or close to the U.S. GAAP reserves

• Medical malpractice claims-made: We did not consider the results of our testing to be meaningful.

• Impact of moving to fair value reserves tended to be greater for smaller companies (i.e., higher MVM charge).

Based on the model calibrations

24

Findings – Fair Value impact on income statement (incurred losses)

• Under FV, prior accident year development may not be benchmarked to zero (due to relative changes in discount and MVM).

• Leveraged impact of reserve changes would likely increase volatility of incurred losses.

25

Significant issues - modeling

• Dealing with real data

• Measures of variation– The constraint to accept booked reserves as mean of

distribution impacts: Expected payment and reporting patterns Variability of experience in relation to expectations

– Variation from the tail/prior accident year bucket– Variation for certain liabilities not amenable to statistical

analysis (e.g. asbestos & environmental)

26

Significant issues - MVM estimation

• Variety of approaches exist, but no single approach universally preferred or accepted.

• Professional guidance may be needed on acceptable methods or calibration procedures for calculating MVM’s to gain industry practice consistency.

• Single industry guideline for MVM calculation unlikely to be appropriate.

• Calibration of MVM models can be challenging and will significantly affect the results.

27

Significant issues - financial statement presentation (1)

• How should accounting standards treat the 3 elements of fair value reserves:– As flowing through the income statement or as a direct

charge to surplus?

• Any presentation separating current and prior accident year contributions may require MVM allocation judgments:– MVM’s are statistically non-additive, so any split by

accident year may require allocation judgment.– Judgments may also be required for disclosures by

business unit or line of business.

28

Significant issues - financial statement presentation (2)

• The required disclosure for prior accident year development may become confusing.

–One-year development of prior reserves would not necessarily be benchmarked to 0.

–Disclosure of the components of one-year development of prior year-end reserves could be quite complicated.

29

Assessment of P/C actuarial methods

Estimating undiscounted reserves: GOOD

Discounting estimated future payments: GOOD

Estimating market value margins:DEVELOPING

Calibration of MVM methods: EMERGING

30

Towers Perrin Paper

31

Background to IFRS 4

• Work of the IASB– 2002: Split Phase I (2005) / Phase II (2008?)– 2004: IFRS 4

• European Union has said IFRS is compulsory for stock exchange listed companies from January 1, 2005

• International Financial Reporting Standards (IFRS) address 3 major questions in the arena of insurance accounting:– What is an insurance contract?– What do you have to disclose?– How do you measure insurance contracts? – Phase II

32

Phase I – IFRS 4

• Product classification

• Catastrophe and equalization provisions prohibited

• Restrictions on changes to accounting policies

• Requirement for liability adequacy testing

• Impairment of reinsurance

• Disclosure

33

Phase II

• Currently undecided if Fair Value will be basis for Phase II (2008)