Capital

Capital model

Excess capital

Available capital (free reserves)

Required capital

Regulatory capital

Economic capital

protect policyholders interests

assets

liabilities

business objectives

more realistic basis

internal capital model

risk profile of assets and liabilities

correlation of the risks

desired level of overall credit deterioration

risk measure

features

risk profile

risks being modelled

key outcome to measure success

risk measure e.g. VaR

links outcome to capital needed

given degree of confidence

given time horizon

risk tolerance

required confidence level within chosen risk measure

minimum requirement

hold more to

reduce risk of available capital < minimum

greater security for policyholders

meet requirements of other stakeholders

e.g. debt providers

maintain credit rating

maintain buffer between

actual profitability

dividend stream

Risks

Insurance risk

risk of loss from inherent uncertainty of

occurrence

amount

timing

insurance liabilities

premiums

expenses

Gross underwriting risk

Gross reserving risk

relates to risks yet to be written/earned

Sources

attritional loss frequency

large loss frequency and severity

catastrophes

poor aggregate management of CAT exposures

mis-pricing risk

claims inflation

legislation / court awards

volumes/mix of business

anti-selection risk

underwriting cycle

latent claims

propensity to claim

macroeconomic conditions

capital charge

UW result at chosen risk tolerance

for business written/earned

UW result on a realistic basis

any expected profit should be excluded

divide firm's business

classes

territories/currencies

data required

estimates of future business volumes and mix

firm's BE business plan

claims and expense ratios for each class

historic performance

market experience

adjust for differences

assess variability of claims and expenses

fit statistical distributions

use stress tests

split claims

attritional

large

catastrophes

future latent

claims made during extreme conditions

below reinsurance retention

model in aggregate based on past experience

use mildly-skewed distribution

e.g. lognormal

test against experience

if SD small % of mean use N distribution

if insufficient data use loss ratios

model separately to attritional

allow for reinsurance recoveries

use F-S approach based on past experience

e.g. Poisson for frequency

heavily-skewed distribution for severity

e.g. Pareto

derive or test against revalued historic data

cannot model using past experience (too rare)

natural cats

proprietary model

appropriate to insurer's exposure

allow for demand surge

test results against recent events

man-made cats

bespoke model

treat separately

not modelled based on past experience (heterogeneity)

e.g. severe recession

stochastic model (possibly)

if sufficient past data available

scenario-based approach

allows for subjective view of scenarios that may emerge

bespoke model

determine effects on different classes of business

if large

writing significant volumes of business

writing capital-intensive business e.g. CAT risks

correct price uncertain

lack of data

uncertain claims experience

uncertain mix

uncertain T&Cs

heterogeneity of business

relates to risks already written/earned

i.e. claims yet to be reported/settled

capital charge

ultimate cost of settling claims

at firm's chosen risk tolerance

from business written/earned before modelled period

current BE reserves

approach analogous to that for UW risk

data required

firm's actual BE reserves

initial loss ratios for B-F method

earlier year results

benchmarks

assess variability in firm's claims settling

bootstrapping

Mack method

using stress tests

if large

less (risky) business than previously

medium to long-term business

Net insurance risk

significant if

low levels of reinsurance

uncertain recoveries

nature of the cover e.g. v. high layer

events causing losses v. volatile

high reinsurance exhaustion risk

Aggregate insurance risk

should allow for correlations between

classes

UW and reserving risks

if large

write large amounts of (volatile) business

significant accumulations of risk

may be affected by

underwriting cycle

parameter error

reinsurance of unexpired risk

multi-year policies

management actions

discounting

model should reflect movements in

rates

T&Cs

include margin for variability to

allow for less than full credibility of past data

allow for cost of purchasing reinsurance on LOD basis

to cover future period being modelled

capital charge should suffice for

all risks taken on during modelled period

following UW losses firm will increase rates

unless not shared by market

does not add to insurance risk

unless profits are prematurely anticipated as a result

Market risk

risk of fluctuations

in value of assets

in the income from assets

market movements

Sources

interest rates

market value of investments

value of liabilities (if discounting)

investment income levels

exchange rates

market value of overseas investments

size of foreign currency claims

value of repatriated profits

equity risk

property risk

spread risk

non-constant credit spreads

concentration risk

lack of diverse portfolio

mismatching risk

can lead to liquidity risk

systemic risk (contagion)

other risks

inflation risk

if assets held not real

reinvestment risk

if duration of assets < liabilities

economic conditions

severe economic downturn

adverse interest rates

fall in equity market

higher taxes

extreme events

assets valued incorrectly

Correlations are important aspect of prudent management of market risk

Modelling

simple stress tests

standard investment portfolios

sense check more complicated model

consider relationships between risks

ESG

generates values for economic variables

can be very complex

output should be scrutinised for reasonableness

gives joint probability distribution of outcomes

defines form variables can take

and relationship between them

point is chosen reflecting desired confidence level

express chosen confidence level

stress and scenario-type benchmarks

examine scenarios around chosen scenario

to understand drivers of capital charge

examine individual variables

within overall economic scenario

identify those to which balance sheet most sensitive

examine implied correlations within the output

Capital charge

if small % of overall requirement

shorter-tail liabilities

invested in stable assets

well-matched portfolio

Liquidity risk

unable to meet obligations as fall due

timing mismatch between assets and liabilities

Factors to consider

uncertainty of liability cashflows

size

timing

extent of mismatch between A & L

amount of liquid, marketable assets

changes in market conditions

affecting ability to liquidate promptly

failure to forecast CF requirements accurately

unplanned changes in funding sources

process weaknesses

poor credit control

management of disputes

minimum level of free assets required

across UW cycle

impact of a large loss

correlations with

insurance risk

market risk

Modelling

possible stress tests

increase in attritional claims

delay between large loss and making recovery

CAT loss

reduced level of NB

Capital charge

if low

investing in stable assets

significant reinsurance in place

to meet CAT losses

significant stream of future income

to fund short-term cashflow needs

Credit risk

risk of loss due to 3rd party failure

investment credit risk

counterparty credit risk

Factors to consider

probability of default

e.g. from holding non-gov't bonds

financial strength rating of investments

probability of moving from one grade to another

size of loss given default occurs

degree of exposure

spread risk

correlations between asset classes

likelihood of systemic financial risk

cannot be diversified

potential correlations within

financial and economic modelling factors

used within assessment of market risk

Factors to consider

Examples

e.g. balances with intermediaries

e.g. reinsurer default

e.g. premium debtors

probability of default

for each counterparty

by reference to credit ratings

level of risk in different environments

size of loss

correlations between counterparties

hidden costs of adverse credit scenarios

potential non-recoveries

reinsurance/broker disputes

any collateral held

e.g. letters of credit

Modelling

stochastic model may be used

stress tests needed

to check for reasonableness

help calibrate the assumptions

failure of largest reinsurer/intermediary

one notch downgrade of all reinsurers

existing or future disputes relating to reinsurance contracts

default of largest corporate investment

Capital charge

if high

significant reinsurance in place

use of low-grade reinsurers

significant debtors

Operational risk

inadequate or failed

internal processes

people

systems

administration

higher where

recent acquisition

new classes written

business written cross-territory

governance

dominance risk

loss of key personnel

new CEO

Board restructure

compliance

legislative requirements

internal requirements

higher where new class written

fraud

may increase following management change

strategy

inability to

implement appropriate business plans

make decisions

allocate resources

adapt to changes in business environment

higher where entering new market

technology

risk higher where

high quote volumes

reliance on technology

website sales

quote systems

rating engine

claims reporting

less infrastructure

pension scheme provision

requirement to rectify a funding shortfall

risk higher following an acquisition

poor claims controls

claims leakage

poor underwriting

claims handling systems unable to cope

external events

natural disaster

difficulty sourcing skilled labour

Modelling

difficult

uncertain and subjective nature

require considerable judgement

overlap with other risk types

avoid double-counting

identify all material scenarios

specific to firm's business

compile/maintain risk register

brainstorming session with stakeholders

Capital charge

not defined % of other risk charges

stochastic model infrequent

insufficient past experience of extreme failures

use deterministic scenario-based approach

degree of loss from each risk

type of event causing loss

frequency of such a loss

net and gross of mitigating controls

if low

model error

inadequate assessment

Group risk

being part of group structure

ownership structure of the firm

how it is funded by parent company

Forms

reputational risk

group reinsurance risk

concentrated credit risks

reliance on parent for centralised functions

e.g. accounting

mitigants

capital not a good mitigant

ERM policy

Modelling

stochastic model unlikely

scenario-based approach more likely

consider

likelihood

financial consequences

insolvency

downgrading of parent company

Political risk

political change alters

outcome/value of economic action

important for firms operating in emerging markets

may be correlated to insurance risk

regulatory changes may stem from a cat

Modelling

stress testing using a known example

diversification

if no credit total CR =

sum of risk charges

if risks perfectly independent total CR =

sqrt of sum of risk charges squared

Factors affecting

regulatory requirements

model used

internal capital model

standard formula

time available

IMAP

credibility of estimated requirements

may hold margin if low

basis used

run-off

going concern

windup

desired level of ruin probability

i.e. risk tolerance

time horizon of calculation

one-year

ultimate

risk appetite

probability of adverse events/cats

probability of breaching MCR

expected line size to MCR

output of stochastic model/DFA

size of risks

adding a small business to large

may have minimal impact

opportunity cost of capital

start-up costs need to be funded up-front

market perceptions

increasing CR could be perceived badly

achieve desired credit rating

capital required to attract policyholders

desired flexibility

risk selection

reinsurance strategy

level of prudence

investment strategy

ability to withstand changes in

legislation

tax

judicial inflation

business plan

growth increases CR

profitability

profit commission arrangements

can be volatile

will reduce profits

may improve liquidity

if less return commission payable upfront

reliance on retained profits may be limited

new business strain

soft market

dividend policy

removing shortfall

depends on size of deficit

Underwriting risk

reduce business volumes

fixed costs still need to be covered

write less capital-intensive business

shorter-tail classes

fewer liability risks

less volatile direct risks

e.g. smaller commercial properties

less volatile reinsurance business

working layer vs. high layer

regions not cat-exposed

increase diversification

write different business to previous years

write more classes

purchase more reinsurance

lower attachment points

higher upper limits

lower aggregate deductibles

unlimited reinstatements

Reserving risk

purchase run-off reinsurance

e.g. LPT

Market risk

invest in lower risk assets

cash

short-term gilts

exit classes exposed to currency risk

Credit risk

use more highly-rated reinsurers

require reinsurers to post collateral

tighter control of broker balances

Liquidity risk

match liabilities more closely

increase premium income

Operational risk

discuss areas of concern

implement controls to mitigate concerns

Access to further capital

shareholders (proprietary)

unwilling?

may demand high returns

Names (syndicate)

parent company

group risk

financial reinsurance

if permitted by PRA

e.g. financial QS

other finance provider

guarantees from government

implicit

explicit

retained profits

new class of business

product development costs

product design

marketing costs

new business strain

administrative expenses

paying commission

setting up reserves

including solvency margin

Modelling

create model points

reflecting expected business mix

build cashflow model projecting

premiums

claims

expenses

investment return

reinsurance

premiums

recoveries

tax

allow for required solvency margin

discount cashflows to determine reserves

choose appropriate basis

realistic

economic capital

prudent

regulatory capital

Consider

type of model

deterministic

time horizon

lifetime of business

accuracy required

number of simulations

correlations between variables

consistency between assumptions

allowance for any guarantees

future tax position

scenario and sensitivity testing

stochastic

easier to design

quicker to build/run

clearer what stresses/scenarios tested

easier to understand