Liquidity Risk: An Insurance Perspective

1
Liquidity Risk An Insurance Perspective

• Jack R. Buchmiller
• Capital Markets Bureau
• New York State Insurance Department
• for
• Professional Risk Managers International
  Associations
• Best Practices in Liquidity Risk Management
• Ottawa April 27, 2009
• Montreal April 28, 2009

2
Safe Harbor Statement

• When wind water rise high enough, no harbor is
  safe!
• -- JRB, at Citigroup Smith Barney, 9/11/03

3
Overview of New York State Insurance
DepartmentOne of the most experienced U.S.
financial regulators

• 1851 New York State Banking Department (NYSBD)
• 1859 New York State Insurance Department
  (NYSID)
• 1863 Office of the Comptroller of the Currency
  (OCC)
• 1871 National Association of Insurance
  Commissioners (NAIC)
• 1913 Federal Reserve System (Fed)
• 1930 Bank for International Settlements (BIS)
• 1932 -- Federal Housing Finance Board (FHFB)
• 1933 Federal Deposit Insurance Corporation
  (FDIC)
• 1934 Securities and Exchange Commission (SEC)
• 1938 National Association of Securities Dealers
  (NASD)
• 1970 -- National Credit Union Administration
  (1934 - Bureau of Federal Credit Unions)
• 1974 Commodities Futures Trading Commission
  (CFTC)
• 1974 ERISA Pension Benefit Guaranty
  Corporation (PBGC)
• 1989 Office of Thrift Supervision (FHLB Board
  1932)
• 1992 -- Office of Federal Housing Enterprise
  Oversight (OFHEO)
• 2008 -- Federal Housing Finance Agency (FHFA
  OFHEO FHFB)
• 2009 Systemic Risk Regulator?

4
Overview of NYSIDNYSID-regulated Assets Under
Management 3.6 trillion (excluding public
NFP pension plans)
5
Federal Government in the Insurance Sector

• 1933 - Federal Deposit Insurance Corporation
  (FDIC)
• 1934 Federal Housing Administration (FHA)
• 1935 Social Security Administration (SSA)
• 1934-1989 - Federal Savings Loan Insurance
  Corporation (FSLIC)
• 1938 - Federal Crop Insurance Corporation (FCIC)
• 1953 Small Business Administration (SBA)
• 1957 Civil nuclear power excess of 10 bln
  loss, (Price-Anderson Act of 1954)
• 1965 Medicare Medicaid (SSA)
• 1968 - National Flood Insurance Program (NFIP)
• 1970 - Securities Investor Protection Corporation
  (SIPC)
• 1970 GNMA first insured MBS issued
• 1970 - National Credit Union Share Insurance Fund
  (NCUSIF, by NCUA)
• 1971 - Overseas Private Investment Corporation
  (OPIC)
• 1974 - Pension Benefit Guaranty Corporation
  (PBGC)
• 1981 FNMA-insured MBS issuance begins (split
  from GNMA began in 1968)
• 1996 Federal Agricultural Mortgage Corp.
  (Farmer Mac est. 1971)
• 2002 - Terrorism Risk Insurance Act (TRIA U.S.
  Treasury Dept.)

6
GAOs High-Risk List of Federal Financial
ProgramsGovernment Accountability Offices
High-Risk series was first published in 1990.
http//www.gao.gov/new.items/d09271.pdf

• Pension Benefit Guaranty Corp. 1990 - 1995
• Resolution Trust Corp. 1990 1995
• Farm Loan Programs 1990 2001
• Medicare 1990 present
• Bank Insurance Fund (FDIC) 1991 1995
• HUD Single-family mortgage insurance, et al 1994
  2007
• Pension Benefit Guaranty Corp. 2003 present
• Medicaid 2003 present
• National Flood Insurance Program 2006
  present
• Outdated U.S. Financial Regulatory
  System January 2009

7
Liquidity(But I know it when I see it.)
8
Defining Liquidity

• Liquidity is the ability to fund increases in
  assets and meet obligations as they come due.
• From Sound Practices for Managing Liquidity in
  Banking Organisations, BCBS, February 2000.
  (http//www.bis.org/publ/bcbs69.htm)
• increases? That sounds like a directional
  bias! ( or decreases in liabilities )?
• Market liquidity ability to be easily
  converted through buying or selling without
  causing a significant movement in the price and
  with minimum loss of value.
• -- Wikipedia (www.wikipedia.org/wiki/Liquidity)
• Funding liquidity risk is the risk that the firm
  will not be able to efficiently meet both
  expected and unexpected current and future cash
  flow and collateral needs without affecting
  either daily operations or the financial
  condition of the firm. It differs from the market
  liquidity risk, which is the risk that a firm
  cannot easily offset or eliminate a position
  without significantly affecting the market price
  because of inadequate market depth or market
  disruption. However, in many cases, the same
  factors may trigger both types of liquidity
  risk. emphasis added
• -- The Management of Liquidity Risk in Financial
  Groups - The Joint Forum, May 2006
• Or, are funding market liquidity risks two
  degrees of the same problem?

9
Measuring Liquidity

• Liquid assets typically short-dated or actively
  traded, and high-quality names and/or ratings
• Overnight, 7, 30 days T-bills, CP Treasurys,
  G-7 sovereign
• (formerly, Agencies, ABS, ABCP)
• Other liquidity unused committed credit
  lines, shelf capacity, unencumbered assets,
  securitization, sale/leaseback, etc.
• But You cant solve a leverage problem by
  borrowing more money, not even from the central
  bank. -- Anon.
• Risky ? Illiquid e.g., SP 500 basket
  versus SLuGS
• (Non-marketable State Local Government Series
  of U.S. Treasurys)
• Only cash is both risk-free and liquid (esp
  deposits/reserves at central bank).
• ALM buckets amount of matching/mismatching
  assets liabilities coming due or possibly
  coming due (first/worst call/put) within given
  time periods.
• Demand, 0-5 days, 5 30 days, , 2 5 years,
  etc.

10
Liquidity SolvencyYou can be solvent but
illiquid or liquid but insolvent.

• Assets gt Liabilities a balance sheet test
  typically regulatory accounting dependent.
• Statutory or GAAP?
• Book value, MTM, or combination of both?
• Ability to pay when due/failure to pay the
  classic bankruptcy event
• Collateral calls count!
• Economic solvency would embrace both
• Net present value of expected cashflows,
  risk-adjusted and discounted at risk-free rate,
  is the economic version of balance sheet
  solvency.
• Risk-adjusted -- analysis of distribution
  probability of possible outcomes.
• No deficit ALM buckets (include asset sales,
  borrowing) is the pay when due constraint.
• Historical cost or book value accounting looks
  at cashflow at t 0, updated annually, but not
  forward-looking (i.e., doesnt consider
  cashflows).

11
The Problems With MTM AccountingMark-to-market
is (or should be) a discipline, not a religion.

• MTM is fickle
• Amazon.com was 106.69 on 12/10/99 and 5.97 on
  9/28/01, even though operating loss halved.
• Subprime traded at par or premium to par
  until it suddenly didnt.
• MTM discounts for illiquidity, a characteristic
  of markets not of securities (or CF).
• Is price information or data?
• Market-implied ratings or default probability
  (Merton Model)
• But for enterprise value, Bloomberg on January
  1st or the 10-K on March 17th?
• GAAP is inconsistent not all assets or
  liabilities are MTM (for example, insurance).
• Insolvency is impossible under full MTM as even
  in bankruptcy, share price is still a few cents.
  
• Own credit discounting of liabilities could
  keep Assets gt Liabilities
• FAS-159 permits encourages earnings
  management
• The objective is to improve financial reporting
  by providing entities with the opportunity to
  mitigate volatility in reported earnings caused
  by measuring related assets and liabilities
  differently without having to apply complex hedge
  accounting provisions. May be applied
  instrument by instrument -- pages 1 2
  (emphasis added and just whose fault are those
  complex accounting rules?)
• FAS-157 159 negate the notion of capital or
  surplus
• GAAP accounting has a single-minded focus on EPS
  solvency regulation is about balance sheets and
  ability to pay when due (future cashflows).
• It reverses financial theory practice by having
  price determine (re-determine) cashflows
• Banks risk-adjust and PV their ABS/RMBS, but then
  give them a 2nd valuation haircut using ABX.
• Which sets a new price for the next bank, who
  then must MTM again, creating a spiral.

12
The Problems With MTM Accounting
(cont)Mark-to-market is a discipline, not a
religion.

• Neither statutory nor GAAP account for future
  not-yet-incurred losses -- e.g., catastrophe
  risks, options, etc. (Is the answer ERM?)
• Own credit discounting is financial economic
  nonsense, and it ignores contract.
• Lehman Bros. set the global all-time record for
  net income in 2008 because under FAS-157 own
  credit their 600 billion of debt was discounted
  to virtually worthless in bankruptcy and becomes
  GAAP income (except that bankrupts switch to cash
  accounting).
• Bankruptcy (U.S.) proof of claim is made at
  face-value, not market (as are CDS)
• (And why is it that liabilities seem to be
  off-balance sheet more often than assets?)
• No valuation method cost, amortized book, MTM,
  etc., is perfect.
• Horses for courses.
• Problems with economic value are its difficulty
  and the subjectivity of estimating and
  risk-adjusting future cashflows.
• If financial accounting is the art of concealing,
  then financial analysis is the art of revealing.
• Economic valuation of asset liabilities
  cashflows is for the good times, too
• If youd modeled long-term cashflows for
  drkoop.com would you have paid 9.00/share at the
  IPO, or 36.88 on 7/6/99? (Last seen
  0.0001/share on 5/12/05)

13
Regulatory Problems With MTM Accounting There
is no such thing as capital yet we regulators
obsess over it.

• There are assets and there are liabilities
  capital is an accounting conceit (surplus
  states it better A gt L a surplus).
• Assets liabilities usually have very separate
  dynamics. (If they didnt, then you wouldnt
  need intermediaries, financial institutions.)
  For example, PC insurance
• Cost of float and its duration versus asset
  yield duration.
• PC duration is not a function of yield curves!
  
• It is time, not price, taking a random walk.
• Catastrophe risk tenor mismatch annual contract
  on 1-in-100 year probabilities.
• Off-balance sheet assets liabilities have
  cashflows, too. (Contingent CF?)
• We are rooted in historical-cost or book-value
  accounting, yet solvency is a matter of future
  cashflows, including principal flows.
• Accounting is, properly, expected value but we
  must regulate to some worse- or worst-case
  probability.
• VAR trailing 52 weeks of tick data versus 100
  years of available data
• Risk-adjusting future cashflows is subjective
  both art science.

14
Theory / Principles of Liquidity It works in
reality but does it work in theory?

• There are none! (I think thats both interesting
  and important.)
• No Markowitz MPT to tell us how to optimize
  liquidity
• No Treynor/Sharpe CAPM to find efficient
  liquidity frontiers
• No Black-Scholes-Merton to adjust our liquidity
  delta
• We know how to talk about it but we dont know
  how to think about it.
• Liquiditys price is observable (yield curve
  CDS/cash bond basis spreads on/off-the-run
  Treasurys benchmark v. MTN) but you cant buy
  it, you either have it or you dont.
• While theres no Theory of Liquidity can we
  apply MPT/CAPM models by considering liabilities
  as assets held short -- and adding dollops of
  behavioral finance and actuarial science.
• Correlations dont go to one in a crisis
  liquidity/asset preferences, the efficient
  frontier, regime switches.

15
Liquidity in 2007-2009Collateral is the
Contagion!

• Collateralization of derivatives, especially CDS,
  and dependence on secured funding (as noted by
  BIS) created at the wholesale level the run on
  the bank effect that deposit insurance (mostly)
  cured at retail.
• What made interest rate, FX, and other
  derivatives so easy is what made CDS so tricky
• Generic market risks v. idiosyncratic (name)
  risk beta v. alpha
• Credit risk is among the jumpiest (like
  liquidity) and least mean-reverting
• All derivatives (incl. CDS) convert market risk
  into counterparty risk and
• collateral coverts counterparty risk into
  liquidity risk!
• What makes sense on a micro-economic level,
  bilateral collateral, creates systemic risk at
  the macro- level.
• (Game theory / Nash equilibrium)

16
Collateral is the Contagion! (cont)

• Mark-to-market must always exceed expected loss
  first it prices expected cashflow, then adds-on
  for riskiness of the cashflows and their PV
  (risk-free rates fluctuate), liquidity risk,
  counterparty risk. In times of stress, an
  uncertainty premium too.
• Posting collateral is a cash transaction,
  identical to pre-paying the claim (loss) except
  you also over-pay by adding counterparty, market,
  liquidity, and other risk premia.
• Your collateral gets haircut, too.
• Markets cannot collateralize, or
  over-collateralize, themselves.
• Collateralizing systemic risk with beta
  assets?
• Can the shorts meet the longs before the
  music stops?
• Definition of risk changes from variance (Std
  Dev) of returns to loss of capital, but variance
  also jumps
• Not enough risk-free assets to go around
• A regime switch of the efficient frontier?
• If an asset class or sectors expected return and
  variance change, then it falls off or on to the
  efficient frontier
• Multiple regimes or a continuum of regimes?
  (1929 v. 2008 v. 1974 v. 1987 )
• Dr. Werner uncertainty Heisenberg meets Dr.
  Kurt incompleteness Gödel.

17
Collateral is the Contagion! (cont)

• Of the Three Cs of Credit
• With securitization, we traded Character for
  Correlation
• With subprime, we said Capacity and Character
  are, like, so over-rated
• With counterparty risk, we put it all on
  Collateral
• And that mattered! (Especially as it became a
  liquidity crisis.)
• Correlation and collateral turned out to be
  imperfect substitutes for credit analysis.
• (See penultimate slide)
• Back to Banking 101?
• Collateral doesnt make bad loans good it just
  makes secondary sources of repayment more
  collectable.
• Your primary source of repayment remains
  capacity.
• Character I wouldnt lend to a man I didnt
  trust against all the bonds in Christiandom! --
  JP Morgan, dean of old-school bankers.
• But whose character the SPV, underlying
  borrower, rating agency?

18
A Collateral Is The Contagion Case Study
Berkshire Hathaway (BRK) v. American
International Group (AIG)(Note Based on public
information only.)
19
A Collateral Is The Contagion Case Study
Berkshire Hathaway (BRK) v. American
International Group (AIG)(Note Based on public
information only.)
20
AIG-FP(YTD 9-mos 9/30/08 10-Q )

• CDS losses paid not disclosed (immaterial?!)
• Excluding 2a-7 and other ratings puts
• PV of loss 7.8 12.0 bln, respectively,
  under A B scenarios.
• 30 days-to-loss roll rates 2006 07 vintages
  80, 2005 70, pre-05 60
• Severities 50 - 60
• Scenario B 20 worse delinquency severity than
  A (but lt100)
• FASBs (so-called) Fair Value of Derivative
  Liability 32.3 bln in total
• Of which 30.2 bln, 93.5, on arbitrage
  multi-sector CDO
• Collateral posted 32.8 bln on 9/30/08 (7.1b
  to 39.9b 36 days later, 11/5/08)
• Collateral 32.8 / 32.3 102 of total FV
• Collateral 32.8 / 30.2 109 of FV of
  multi-sector CDO
• Collateral 32.8 / 12.0 273 of PV of claims
  in scenario B
• Collateral 32.8 / 7.8 421 of PV of
  claims in scenario A
• Collateral 8 of cash claims paid (Claims are a
  poor economic metric)
• Source 10-Q, pages 20, 45, 50, 99, 114, 125
• Lending money to post collateral for future
  losses and uncertainty repeats the MTM gt PV
  error (while nationalizing it).

21
AIG-FP10-Q page 114 (PowerPoint misaligned some
column headings)
22
Collateral Was The Contagion

• Alternative Solutions?
• Substitute the Fed as guarantor
• Guaranty last-dollar of loss instead of
  first-dollar MTM
• Fed steps into AIG-FPs CSA as credit support
  provider promising to pay if/when AIG-FP
  doesnt. Neither cash nor securities (i.e.,
  collateral) changes hands.
• Counterparties have the Feds name, the same name
  as on our (USA) currency.
• Complication Counterparties probably need to
  deliver collateral to their counterparties.
• Bankruptcy of AIG-FP and, probably, AIG Inc. as
  guarantor
• The lack of collateralization is why some of our
  monolines are still liquid.
• Consistent with the indemnification nature of
  insurance contracts
• Was Fannie Mae Freddie Macs problem their
  off-balance sheet guarantees (underwriting
  leverage) or their on-balance sheet RMBS (funding
  leverage)?

23
Quantitative Finance and Risk Management Have
Their Limitations!

• In his paper How We Came Up With The Option
  Formula the late Fischer Black wrote
• I have an A.B. in physics, but I didnt
  recognize the equation as a version of the heat
  equation which has well-known solutions.
• The Journal of Portfolio Management, Winter,
  1989, page 5.
• Buchmillers Caveat The operation of the heat
  equation does not depend upon a 20-something
  trader in New York or London picking up the
  phone, updating a spreadsheet, and making a
  market.
• -- at PRMIA-NY, 7/16/02

24
An Additional Warning

• One of Murphys Laws
• Mother Nature favors the hidden flaw.
• Buchmillers Caveat
• but human nature seeks it out.
• -- at PRMIA-Stamford, 12/5/06

25
Contact Information

• Jack R. Buchmiller
• Supervising Risk Management Specialist
• Tel [phone]
• Email jbuchmil_at_ins.state.ny.us
• Capital Markets Bureau - 4F
• New York State Insurance Department
• 25 Beaver Street
• New York, NY 10004
• Website www.ins.state.ny.us