Margin Reform

Adaptiv European
UGM 2016
BCBS – IOSCO Margin Reform
Jonathan Berryman, Risk Strategy
7th September, 2016
Agenda
01
Regulatory framework
02
Initial Margin Requirements
03
ISDA Standard Initial Margin Model (SIMM)
04
Practical implications
2
1.
Regulatory framework
Global Regulatory Framework
• FSB established in April 2009
• Pittsburgh Summit of G20 Heads of State, September 24-25,
2009
– Key objectives to strengthen the international financial system
“Developing internationally agreed rules to improve both the quantity and
quality of bank capital and to discourage leverage…together with
strengthened liquidity risk requirements”
 “Improving over the counter derivatives markets: All standardized OTC
derivatives contracts should be…..cleared through central counterparties.
Non centrally cleared contracts should be subject to higher capital
requirements”
 June 2012, FSB extended the requirement for non-centrally cleared contracts
to include bilateral margining and certain risk management requirements

4
Extending CCP Margin to Non-cleared OTC Derivatives
Price
Trade
Default
Closeout
Initial Margin
should cover this
Variation Margin
should cover this
Time
Source: ISDA
5
FSB objectives implementation
•Final standards published March 2015
(bcbs317)
• Universal two-way exchange of IM
• Two ways of calculating IM
• Segregation and limited or no re-hypothecation of
posted collateral
• Minimum Eligible Collateral Standard
• EUR/USD 50 million threshold IM amount
• Phased-in approach starting from Sept. 2016
• Existing Trades Grandfathered
• Physically settled FX forwards and swaps to be
exempted
6
Implementation
EU Draft RTS
ESAs-2016-21
(Mar 2016)
OSFI
Guideline
E-22
US Joint
Agencies
Japan FSA
7
Phase-in Dates
• Variation Margin
Compliance date
AANA for EU
1 Sep 2016
> €3.0 tn
1 Mar 2017
> € 8 billion
• Initial Margin
Compliance date
AANA for EU
1 Sep 2016
> € 3.0 tn
1 Sep 2017
> € 2.25 tn
1 Sep 2018
> € 1.5 tn
1 Sep 2019
> € 0.75 tn
1 Sep 2020
> € 8 billion
EU has announced a delay until next year
for 1st Sep 2016 IM implementation
 Best guess February 2017
 Switzerland, HK, Singapore and Australia
followed suit
AANA = aggregate month-end average notional amount of non-cleared OTC
derivatives
8
2.
Initial Margin
Requirements
Initial Margin Requirements
• Choice to Two Treatments:
• Standardised Approach (per Haircuts Below), or
• Internal Model
SCHEDULE BASED
APPROACH
• Standardised approach allows for limited
Netting:
Net IM = (0.4 + 0.6 x NGR) x Gross IM,
where NGR = ratio of net to gross replacement cost
• According to an ISDA 2012 QIS, the average
amount of IM needed for in scope
counterparties is $23Bn
• Standardised approaches thought to produce
MUCH higher numbers
10
Summary of IM Model Requirements
Area
Requirement
Horizon
• Based on one-tailed 99% confidence interval over 10-day liquidation horizon
Observation Period
• Models must be based on an equally weighted historical observation period of at least one year and
not more than five years
Calibration
• Models must incorporate a period of significant financial stress for each broad asset class
• Data within each of the identified periods shall be equally weighted for calibration purposes
Review
• Independent Review, Pre Approval
• Must review and revise data used to calibrate the IM model at annually
• Must review the model no less frequently than annually
Netting
• May net within asset class (e.g. currency/rates, equity, credit and commodities), but not across
asset classes
Modelling Approach
• Model shall use risk factors sufficient to measure all material price risks inherent in the transactions
for which initial margin is being calculated
• The model shall include all material risks arising from the nonlinear price characteristics
Governance/
Documentation
• Periodic review of model in light of developments in financial markets and modelling technologies,
and enhancement of the model as appropriate to ensure that it continues to meet the requirements
for approval
11
3.
ISDA Standard Initial
Margin Model (SIMM)
ISDA Standard Initial Margin Model (SIMM)
Industry identified nine key criteria that an Initial Margin model should adhere to:
Non-procyclicality
Ease of replication
Transparency
Quick to calculate
Extensible
Predictability
Costs
Governance
Margin appropriateness
Source: ISDA
13
SIMM Model Outline
• The model defines a hierarchy of products and risks
– Each trade is assigned into a product class, and the risks of that trade are used to calculate the margin
– Each individual risk (or risk factor) is assigned into a risk bucket, and each risk bucket is assigned into a risk class
– Netting and diversification is recognised using a correlation structure at each level (except for product class).
Product
Class
Risk
Class
Risk
Bucket
Risk
Factor
Risk Class
Risk Bucket
Risk Factor
Interest Rate
Currency
Subcurve/Tenor interest rate
Credit (Qualifying)
Sector/quality groupings
Issuer/Seniority/TenorCDS spread
Equity
Sector/region/size groupings
Equity spot price
Commodity
Sector groupings
Commodity spot price
FX
(not used)
FX spot rate (vs base currency)
• For example, a USD interest rate swap would be put into the “RatesFX” product class. A major risk class for
this trade would be “Interest Rate”, and the relevant risk bucket is “USD”, with one risk factor being the USD
5y swap rate against 3m Libor.
Source: ISDA
14
Calibration Period
• Identification of Stress Period
– Based on representative risk factors for each risk class, a one-year contiguous time period that represents
the most volatile time window is chosen as the stress period
– It is determined by sliding a one-year volatility (of the 10-day overlapping return of relevant pseudo index)
window from January 2, 2008 through January 31, 2015
• The time period for calibration is constructed by combining the one-year stress period and the
most recent contiguous three-year period
Source: ISDA
15
Major Model Assumptions
Major Assumption
Justification
Mitigation
Risk factors are grouped into buckets for
correlation structure
SIMM avoids the need for an impractical large correlation
matrix, avoiding large data and positive-definite problems
Backtesting exercise
Same correlation is used for Delta and Vega
Follows FRTB example in avoiding complexity of extra
vega correlation matrix
Backtesting exercise
Risk weights are grouped into buckets, which
involves some averaging within buckets
Keeps number of parameters down to reasonable level and
eases incorporation of additional risk factors
Backtesting exercise
Some risk factors are omitted, such as
skew/smile, inflation volatility, credit
correlation, and equity dividend
SIMM captures material risk, and does not include the
entire universe of risk factors
Backtesting exercise
and specific testing
Gamma can be approximated from vega
Theoretical mathematical analysis, plus numerical testing of
real-world portfolio
Backtesting exercise
and specific testing
Assumes the portfolio is neither small nor
highly concentrated in an unusual
product/risk type
The aim of SIMM is to reduce real-world risk between
market participants and it focuses on the bulk of real-world
portfolios
Immateriality
Source: ISDA
16
Challenges
• Reconciliation of Risk Data Across Industry
• Global Standard
– Treatment of Affiliates
– Risk Factor vs. Asset Class
• Margin Call Timing
– Cross Time Zone Issues
– Computation Times for More Complex Products
• Lack of Global Standard
– Differing Product Scope
– Substituted compliance
17
4.
Practical implications
Practical implications
Collateral and documentation requirements
Multiple and overlapping CSAs
ISDA
Netting
Agreement
Legacy
VM CSA
Margin Reform
VM CSA
Bloomberg
Margin Reform
IM CSA
Other CSAs
IA, Ccy VM
See Jean-Marc’s ACR presentation at 2.40pm for
more details
19
Practical implications
Potential Future Exposure
Trade
Exposure
Initial
Margin
Variation
Margin
20
Questions
21