Presenters - PowerPoint PPT Presentation

1 / 30
About This Presentation
Title:

Presenters

Description:

The Professional Risk Managers' International Association ... Part A: One Step Scaling Factor. Scaling Factor/Buffer is determined for each rating grade ... – PowerPoint PPT presentation

Number of Views:32
Avg rating:3.0/5.0
Slides: 31
Provided by: JohnB1
Category:
Tags: aone | presenters

less

Transcript and Presenter's Notes

Title: Presenters


1
The Professional Risk Managers International
Association
  • Presenters
  • Roger Kumra,
  • Vivian Ozohili,
  • Conn ODonnell
  • iesteering_at_prmia.org

2
A look at the dynamics of rating systems An
overview of the CEBS approaches moving into
the discussion
3
A look at the dynamics of rating systems
4
Dynamics of Rating Systems
  • Models can be split into
  • Point-in-Time (PIT) models
  • where the model is designed so that predicted PD
    is expected to be close to the actual default
    rate, in any given year
  • Could include factors such as Current Months
    Arrears or factors that pick up changes in the
    economic climate.

5
Dynamics of Rating Systems
  • Models can be split into
  • Through-the-cycle (TTC) models
  • where the model is designed so that the predicted
    PD will be close to the average default rate over
    the cycle, in any given year.
  • Could include factors such as LTV at Drawdown
    which would remain static throughout the cycle.

6
Dynamics of Rating Systems
  • Most models usually fall in between the two
    extremes.
  • These are called Hybrid Models
  • The predicted PD follows the cycle only
    partially.

7
Grade Migrations PIT PD Shifts
The recessionary phase increases the likelihood
that capital requirements shoot up as a
consequence of borrowers downgrades. CEBS
Position paper on a countercyclical capital
buffer.
The PD of the portfolio would change over the
cycle as a result of the migration of borrowers
across grades and of the change of grade PDs.
CEBS Position paper on a countercyclical capital
buffer.
8
Simplistic Model Build
  • A hypothetical one factor model with two
    attributes / questions
  • So each obligor / account falls into either
    attribute 1 or attribute 2
  • In a recession, the number of obligors in
    attribute 1 decreases and the number of obligors
    in attribute 2 increases

9
Simplistic Model Build
  • Hypothetical model, built using data from the
    previous example, calibrated to an anchor point
    of 4 and a cyclicality weight of 50
  • PD
  • 0.9 Attribute_1 21.2 Attribute_2

Calibrated to an Anchor Point of 4
10
Result
  • Predicted Default Rates move with the cycle
    through time.
  • Therefore Capital Requirements also move through
    time
  • In benign periods capital requirements are at
    their lowest
  • Coming into a recessionary period we have the
    effect of capital requirements increasing at a
    time when funding is likely to be more scarce
  • Example Below 20 increase in capital
    requirements from minimum to maximum.

11
An overview of the CEBS approaches
12
Pro-Cyclicality Options
  • Two options proposed for the Implementation of
    the Capital buffer
  • Portfolio level option
  • PD Scaling option
  • Time Varying Confidence Interval
  • Rating-grade level (i.e. more granular) option
  • One- Step PD Scaling Factor
  • Two- Step PD Scaling Factor
  • CEBS shares the view that any Capital Buffer
    should be bank-specific

13
Portfolio Level Option
  • Part A PD Scaling option
  • PD Scaling Factor for the portfolio ratio
    between recessionary PD (highest PD) and current
    PD
  • Current Portfolio PD

14
Portfolio Level Option
  • Key Issues -Current PD of portfolio will change
    through the cycle
  • Grade Migration (more pronounced in PIT rating
    systems
  • Change in Grade PD (more pronounced in TTC rating
    systems)

15
ILLUSTRATION - EXAMPLE
16
Evolution of Buffer through the cycle
17
MCR VS. BUFFER
18
MCR VS. BUFFER
19
Portfolio Level Option
  • Part B Time Varying Confidence Interval
  • Buffer/Scaling factor determined using
    Risk-Weight Function with Time-Varying Confidence
    Intervals i.e. Stress Test of the capital
    requirement to reflect economic conditions
  • Capital requirement (K) LGD N(1 R)-0.5
    G(PD) (R / (1 R))0.5 G(0.999) PD x LGD
  • Calculate capital charge in Bad year (economic
    downturn) using above formulae and adjust IRB
    risk weight function upward in good years
    (economic upswing).

20
Rating-grade level (i.e. more granular) option
  • Part A One Step Scaling Factor
  • Scaling Factor/Buffer is determined for each
    rating grade
  • Current PD Long run average 1- year default
    rate in that Grade at time T. Downturn PD is
    highest PD observed for Grade over a
    predetermined time frame.
  • Approach is non-neutral with respect of rating
    philosophy

21
Rating-grade level (i.e. more granular) option
  • Part B Two Step Scaling Factor- accounts for
    Grade Migration.
  • Introduce when calculating current PD Calculate
    Modified Current PD

22
Rating-grade level (i.e. more granular) option
23
Rating-grade level (i.e. more granular) option
  • MOD PD (100 - 10 - 15) 5 102
    1510 5.45
  • So the modified PD for grade G is 5.45, compared
    with the original PD for Grade G of 5.
  • Similarly Calculate Downturn Modified PD

24
moving into the discussion
25
Summing up
  • PDs produced by rating systems are cyclical
  • Regardless of rating philosophy, we see
    systematic fluctuations in PDs through the cycle
  • This leads to cyclical capital requirements
  • CEBs have proposed a variety of approaches to
    address this
  • These are based around the idea of a capital
    buffer
  • PDs are scaled up the capital buffer is the
    difference between the capital requirement using
    normal PDs and scaled PDs
  • Different approaches focus on the portfolio as a
    whole or on individual grades
  • Key assumptions/limitations of approach
  • Changes in PD are due to cyclical effects
  • Changes in portfolio composition are not
    specifically addressed
  • The impact of increased realised defaults and
    losses on capital are not explicitly addressed

26
Issues to Explore
  • Core Issues
  • Impact on changes in portfolio composition
  • Are these adequately taken into account?
  • Are there incentives to decrease underlying
    portfolio quality? (higher capital requirement
    leads to higher costs less sensitivity to risk)
  • Robustness of scaling factor
  • Do banks have sufficiently long time series?
  • Whats the impact on changes to bank structures
    or mergers how can these be taken into account?
  • Want measures that are robust through the cycle
    does this limit the power of the approach?
  • Adverse incentives
  • Can the approach create adverse incentives? Eg
    holding higher risk assets that have risk not
    captured by the approach

27
Issues to Explore
  • Interpretation and Use
  • Interpretation of results
  • Type of rating system (point in time vs through
    the cycle)
  • What does the capital buffer mean
  • Are there issues in justifying it to the Board?
  • Use of buffer within business
  • Are there Use-Test issues?
  • Should information on the buffer be incorporated
    into pricing? If so, how can this be done?
  • Use of the buffer by external parties
  • Will rating agencies and external analysts expect
    the buffer to be maintained through a down-turn?

28
Issues to Explore
  • Implementation
  • Developing the ideas further
  • Should the term of loans on the book be
    considered? The shorter the term the more
    flexibility the bank has (except maybe for
    bullets refinanced by new loans).
  • Pillar 1 vs Pillar 2
  • What are the advantages/disadvantages of holding
    the buffer under either Pillar 1 or 2?
  • Comparison to Stress Testing
  • Does this approach overlap with stress testing?
  • What additional information does this approach
    provide that stress testing doesnt? And
    vice-versa?

29
Issues to Explore
  • What else should be considered?
  • Its unlikely we have captured everything here.
  • In addition, the approach doesnt look at the
    actual increased defaults and increased LGDs that
    would occur in a downturn. How might these be
    incorporated?
  • What else would you like to discuss on this
    topic?

30

To Join www.PRMIA.org Irish Steering
Committee iesteering_at_prmia.org Member Support
support_at_prmia.org The PRM program
certification_at_prmia.org
Write a Comment
User Comments (0)
About PowerShow.com