Eric Kolchinsky

1
Hedged Trades Lessons from the Crisis

• Eric Kolchinsky
• Chief Operating Officer
• Moodys Evaluations Inc

September 25, 2008
2008 NYSSCPA Banking Conference
2
MEI

• Moodys Evaluations Inc. (MEI) provides
  evaluated pricing on approximately 1.6 million
  securities daily
• MEI uses customized evaluation approaches based
  on the type of market and the specific risks of
  security
• MEI is a Registered Investment Advisor and was
  formed from the purchase of Mergent Pricing

3
Disclosure
Opinions expressed herein are solely those of
Eric Kolchinsky and do not necessarily reflect
the views of Moodys Evaluations Inc. or Moodys
Corporation.
4
Five Structural Causes of the Crisis

• Ignored Incentives
• Incentive concerns treated as quaint
• Overreliance on Models
• Quantitative models too fragile in light of poor
  incentives
• Increased Financial Instrument Complexity and
  Customization
• Allowed optimization of models and regulations
• Securitization
• Failure of financial supply chain management
• Regulatory environment
• Too many and inconsistently applied
  rules/standards

5
Recent Headlines
Merrill Upped Ante as Boom In Mortgage Bonds
Fizzled Wall Street Journal, April 16 2008
How a Good Subprime Call Came to Hurt Morgan
Stanley Wall Street Journal, November 7, 2007
ACA pain spreads far and wide Creditflux
Newsletter, February 2008
UBS reveals the failures that led to its 37bn
sub-prime writedowns The Times of London,
April 22, 2008
6
Observation

• Many of the writedowns during the recent crisis
  have been on positions which were considered
  hedged

Once hedged, ... the Super Senior positions were
VaR and Stress Testing neutral (i.e., because
they were treated as fully hedged, the Super
Senior positions were netted to zero and
therefore did not utilize VaR and Stress limits)
... In several Market Risk Control reports,
the long and short positions were netted, and the
inventory of Super Seniors was not shown, or was
unclear. For the hedging AMPS trades, the zero
VaR assumption subsequently proved to be
incorrect ... -Shareholder Report on UBS's
Write-Downs, April 18, 2008
7
ABS CDO Basics
Super Senior

• Collateralized Debt Obligations (CDOs) are pools
  of securities (e.g. loans)
• ABS CDOs are CDOs backed primarily by sub-prime
  and other mortgage-related securities
• The Super Senior tranche is the largest and the
  most senior tranche in the capital structure

Pool of securities
Mezzanine
Equity
8
The Negative Basis Trade
L spread/Premium
Protection Fee
Monoline / Protection Seller
CDO
Bank
Credit Protection

• Primary source of monoline counterparty risk for
  banks in structured finance
• Banks would enter into a long risk position on a
  CDO and then buy protection (insurance) on the
  same asset from a monoline or other protection
  seller
• Because it was viewed as hedged, the trade
  required little or no regulatory capital allowing
  the banks to book the entire payment stream as
  profit
• Structuring/origination banks held CDOs on
  balance sheet and bought protection in order to
  increase deal capacity

9
Counterparty Risk Case Study ACA

• ACA was a single-A rated monoline insurer which
  originally specialized in troubled municipal
  credits
• ACA first branched out to initially managing and
  later providing insurance for ABS CDOs
• ACA eventually insured over 20 billion of ABS
  CDOs Super Seniors
• Most banks have written the value of their ACA
  protection on ABS CDOs to zero
• ACA entered a restructuring agreement with
  structured finance counterparties in August 2008

10
Counterparty Risk

• It appears that with respect to these negative
  basis trades, prudent analysis was not performed
• ACA had a higher nominal default probability than
  the asset it was insuring
• Appears that some regulations allowed a minimum
  rating requirement without regard to the default
  probability of the insured instrument
• The correlation of the insured and the insurer
  was ignored
• ACAs business, both with respect to insurance
  and management, was structured finance
• One financial institution entered into 6.7
  billion of trades with ACA in 3Q07, when the
  approximate mark to market loss on the insured
  securities may have been multiple times ACAs
  capital

Based on TABX 40-100 _at_ approx 0.40 in 8/07,
assuming all CDOs were mezzanine
11
Counterparty Risk Lessons Learned

• Already re-learned
• Counterparties are risky
• Need to sum risk across company (e.g. monoline
  risks to SF, ARS, VRDN, GICs, ABCP)
• Needs improvement
• Independent and thorough analysis of the
  counterpartys credit and correlation to the
  insured instrument and to the insuring bank
• Valuation
• Take account of any basis risk between the
  different legs of the trade
• Account for appropriate recoveries and
  correlations

12
Convexity Risk Case Study Morgan Stanley

• MS proprietary trading desk entered into a series
  of trades designed to take advantage of their
  negative view on subprime mortgages
• Short mezzanine levels of the ABX index
• Long Super Senior tranches of ABS CDOs to fund
  the trade
• Very likely, the Super Seniors would have been
  purchased with financing
• The trade was positioned to profit if the
  subprime market deteriorated slightly
• As losses mounted, the trade reversed and MS took
  a 3.7 billion loss related to the trade in 4Q07

13
Convexity Risk
Tail Risk
Tail Risk
Flat Trade
Desired Outcome
Our risk exposure swang from short to flat to
long -- Colm Kelleher MS CFO

• The tail risk stems from the large long position
  in ABS CDOs
• Leverage existed not only in the trade itself,
  but also within its components
• There appears to have been no attempts to hedge
  the tail end of the trade

14
Convexity Risk Case Study II UBS

• UBS retained the Super Senior portions of ABS
  CDOs they originated. UBS hedged these positions
  in several ways, including the AMPS program
• The AMPS hedge explicitly excluded tail risk
  protection

Amplified Mortgage Portfolio ("AMPS") Super
Seniors these were Super Senior positions where
the risk of loss was initially hedged through the
purchase of protection on a proportion of the
nominal position (typically between 2 and 4
though sometimes more). Much of the AMPS
protection has now been exhausted, leaving UBS
exposed to write-downs on losses to the extent
they exceed the protection purchased. As at the
end of 2007, losses on these trades contributed
approximately 63 of total Super Senior losses.
- Shareholder Report on UBS's Write-Downs, April
18, 2008
15
Convexity Risk Lessons Learned

• Hedging trades needs thorough stress testing
• The stress testing must cover not only the likely
  outcome, but also extreme scenarios
• Where the long and the short leg reference
  different assets (basis risk), the hedge
  effectiveness needs to stressed as well
• Valuation
• Each leg must be thoroughly modeled (no
  simplifications)
• Must take account of the entire probability space
  (i.e. the tail risk)
• Must account for any leverage in the trade, as
  well as any liquidity issues with the referenced
  assets

16
Conclusion

• There is no such thing as a perfect hedge
• Hedging strategies allowed excess leverage to
  build on balance sheets
• The flexibility of OTC derivatives increase
  opportunities for optimization with respect to
  any regulatory or internal controls
• Incentives matter fancy models (VaR and Stress
  Testing) are useless if they are ignored because
  the trade is deemed riskless