Presentation title · Orange 30 November 2015 R238 G116 29 Dark grey R63 G69 B72 Finance, not physics! 19 One asset class/investor The market The observer Managing ESG uncertainty

26/11/2015

1

ESG Limitations and Developments Sam Worthington

30 November 2015

Agenda • Models

• Calibration

• Assumptions

• Judgement

30 November 2015

Colour palette for PowerPoint presentations

Dark blue R17 G52 B88

Gold R217 G171 B22

Mid blue R64 G150 B184

Secondary colour palette

Primary colour palette

Light grey R63 G69 B72

Pea green R121 G163 B42

Forest green R0 G132 B82

Bottle green R17 G179 B162

Cyan R0 G156 B200

Light blue R124 G179 B225

Violet R128 G118 B207

Purple R143 G70 B147

Fuscia R233 G69 B140

Red R200 G30 B69

Orange R238 G116 29

Dark grey R63 G69 B72

ESGs are a key component of internal

models

30 November 2015 3

Business

decisions

Capital

and risk

projection

Balance

sheet

ESG

26/11/2015

2


Dark blue

R17 G52 B88

Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Real World

towerswatson.com © 2014 Towers Watson. All

rights reserved.

ESG

Valuation & Pricing

Focuses on realism

Calibration to historical data

Uses: ORSA, ALM, SAA

Risk Neutral

Market consistent

Calibration to today’s market prices

Uses: MCEV, Pricing

4

Real World and Risk Neutral ESGs



Gold R217 G171 B22






Forest green

R0 G132 B82


Cyan R0 G156 B200





Red R200 G30 B69

Orange

R238 G116 29


Optimal asset allocation is highly sensitive to

model limitations

30 November 2015 5

Calibration Assumptions

Model

Judgement

Modern portfolio theory

H. Markowitz 1952

“When genius

failed: the rise and

fall of LTCM”

1998

“We are seeing 25𝜎

moves several days in

a row”

GS&Co 2007

“Galloping Gerty”

Tacoma Narrows

Bridge 1940

“Wobbly bridge”

Millennium Bridge

2000



Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


The modeller’s challenge

30 November 2015 6

The real world

Model features

Possible

(R&D)

Required

(Validation)

26/11/2015

3


Dark blue

R17 G52 B88

Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Developing a fit for purpose model

30 November 2015 7

Costs of complexity

Development, calibration, documentation, IT costs, key man risks

Opportunity costs of slow run times and slow evolution

Increased frequency and severity of modelling mistakes

Too complicated

Model features

Efficiency “Just right” Too simple

Limitations of ESGS Models

30 November 2015



Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Credit modelling

30 November 2015 9

Sovereign Corporate Securitized /

Collateralized

Financials

industrials

Utilities

Covered

Securitized

Banking

Financial

services

insurance

Asset backed Pfandbriefe

Non-

Pfandbriefe

Covered

Mortgage

Backed

Commercial

Mortgage

Backed

Collateralized

Mortgage

Obligation

Quasi & Foreign Government

Municipals Preferred

Securities

Coverage and risk attribution granularity of credit

26/11/2015

4


Dark blue

R17 G52 B88

Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Capture tail risk

30 November 2015 10

“We've just had a 1 in 1,000 year event happen 3 times last Week”

Inflation

IR

Equity

Expected Levels

Dynamic Interactions



Gold R217 G171 B22






Forest green

R0 G132 B82


Cyan R0 G156 B200





Red R200 G30 B69

Orange

R238 G116 29


Tail Dependency: asymmetric, symmetric - or

none?

30 November 2015 11

Upper and lower tails

Only lower tail

Trend or randomness?

Limitations of ESGs Calibration

30 November 2015

26/11/2015

5


Dark blue

R17 G52 B88

Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Multi-year real world ESG calibration

30 November 2015 13

1 year calibration 5 year calibration Multi-year calibration

Demand for calibration that is consistent over 1 year and multi-year



Gold R217 G171 B22






Forest green

R0 G132 B82


Cyan R0 G156 B200





Red R200 G30 B69

Orange

R238 G116 29


One year real world ESG calibration

30 November 2015 14

• Companies use a wide range of assumptions for 1 year risk distributions

The Towers

Watson Internal

Model

Calibration

Survey shows a

wide spread of

assumptions for

one year

market risk

• These assumptions affect both capital requirements and the relative

attractiveness of different asset classes

Limitations of ESGs Assumptions

30 November 2015

26/11/2015

6


Dark blue

R17 G52 B88

Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Categorising cycles in markets

30 November 2015 16

Market-making cycle

Days/weeks/Months

Economic cycle

3-7 years

Risk premium accrual

100+ years

Regimes

10-20 years



Gold R217 G171 B22






Forest green

R0 G132 B82


Cyan R0 G156 B200





Red R200 G30 B69

Orange

R238 G116 29


Understanding regimes (1)

30 November 2015 17

(Sources: Robert Schiller, Bloomberg)

Typical model

Regime

thinking?

Naïve

extrapolation

US Treasury bond yields



Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Understanding regimes (2)

30 November 2015 18

(Source: Dimson, Marsh and Staunton)

% o

f glo

bal e

quity m

ark

et cap

Unlucky?

Lucky?

26/11/2015

7


Dark blue

R17 G52 B88

Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Finance, not physics!

30 November 2015 19

The observer One asset class/investor The market

Managing ESG uncertainty Judgement

30 November 2015



Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Dealing with model risk: stresses

30 November 2015 21

Stress testing Reverse stress testing

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

40%

50%

-25% -15% -5% 5% 15% 25%

Equity Return

Gilt Return

95% Events and Contours of Surplus

ProbabilityLocus for all95% events

VaR/ReverseStress event

VaR Loss Locus

Judgement

26/11/2015

8


Dark blue

R17 G52 B88

Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


Dealing with model risk: scenario testing

30 November 2015 22

Rank Risk

1 Depression

2 Sovereign default

3 Hyperinflation

4 Banking crisis

5 Currency crisis

6 Climate change

7 Political crisis

8 Insurance crisis

9 Protectionism

10 Euro break-up

11 Resource scarcity

12 Major war

13 End of fiat money

14 Infrastructure failure

15 Killer pandemic

Scenarios Extreme risks

-2000

-1500

-1000

-500

0

500

1000

0 1 2 3 4 5

Projected accounting surplus/deficit (£m)

Year

Judgement



Gold R217 G171 B22






Forest green

R0 G132 B82


Cyan R0 G156 B200





Red R200 G30 B69

Orange

R238 G116 29


More needed from your ESG

30 November 2015 23

Ensure your model captures latest developments

Multiple models can help ensure modelling is “fit for purpose”

Incorporating short-term economic conditions

Dealing with long-term regimes

Better models

Better assumptions

Better interpretation

Supplement model output with model-independent risk analysis

Adopt a framework that controls model risk

Judgement



Gold R217 G171 B22








Cyan R0 G156 B200





Red R200 G30 B69

Orange R238 G116 29


30 November 2015 24

Expressions of individual views by members of the Institute and Faculty

of Actuaries and its staff are encouraged.

The views expressed in this presentation are those of the presenter.

Questions Comments

Documents

Presentation title · Orange 30 November 2015 R238 G116 29 Dark grey R63 G69 B72 Finance, not physics! 19 One asset class/investor The market The observer Managing ESG uncertainty