Implementing Basel II: Is the Game Worth the Candle

1
Implementing Basel II Is the Game Worth the
Candle?

• Richard J. Herring
• Director of the Lauder Institute
• Co-Director, The Wharton Financial Institutions
  Center

The Future of Banking Regulation London School
of Economics April 7-8, 2005
2
Basel II Attempts to Reconcile a Number of
Irreconcilable Objectives

• Increasing the safety of the banking system
  without changing overall level of capital in
  banking system
• Increasing risk-sensitivity of capital
  requirements without exacerbating pro-cyclicality
  of lending
• Providing incentives for adoption of more
  sophisticated techniques while maintaining a
  level playing field
• Recognizing responsibilities of host country
  supervisors without multiplying compliance costs

3
Concern that Banks Found Ways to Undermine Basel I

• Increase exposure to risk without increasing
  risk-adjusted asset denominator
• Shifting allocation from AAA to BB
• Loan sales
• Securitization
• Increasing exposures to non-credit risks
• Increase numerator without increasing economic
  capital
• Gains trading
• Evergreening

4
Basel I Lacked Incentives for Banks to Adopt
Advances in Risk Management

• Economic capital became industry standard for
  measuring and managing risk
• Amount of equity capital a bank requires to
  provide a given level of protection against
  default to its creditors against unexpected loss
• Operationally expressed as a target rating for
  its long term debt
• VaR (or RAROC) revolution in trading rooms
  extended to loan portfolio
• Better risk taking decisions
• Better distribution of risk

5
Despite Lack of Incentives, Basel 1 Saw
Tremendous Advances in Risk Management

• Why?
• Perhaps facilitated by low costs of compliance
  under Basel I
• Began in North America soon after
  Prompt-Corrective Action Least-Cost Resolution
  reforms increased market discipline
• Banks responded to market demands not regulatory
  commands
• Presumption that banks must be rewarded for
  improving risk management complicates Basel II

6
Internal Models Approach to Market Risk

• 1996 Amendment on Market Risk permitted
  qualifying banks to rely on supervised use of
  their own internal models to determine capital
  charge
• Diminished incentives for regulatory capital
  arbitrage because capital charge reflect banks
  own estimate of risk
• Accommodated financial innovations readily
• Provided an incentive for banks to improve their
  risk management processes and procedures to
  qualify for the internal models approach
• Reduced compliance costs since regulated in the
  same way managed

7
Boundary between Credit Risk and Market Risk
Increasingly Blurred

• Traders increasingly deal with less liquid, less
  creditworthy instruments.
• The market for bank loans is becoming more
  liquid.
• The market in credit derivatives is booming.
• External stakeholders press to consolidate
  accountability for risk management.
• Why not extend the internal models approach to
  credit risk?

8
The holy grail of risk management probability
density function of losses
9
Problems with full model approach

• Conceptual differences in measuring credit losses
• Data limitations in modeling credit risk
• Estimation problems
• Model validation problems
• Inability to deal with low frequency, high
  severity hazards the main source of systemic
  risk
• And so Basel II developed a very complex internal
  ratings based approach and a regulatory model

10
Basel II Aims to

• Eliminate incentives for regulatory capital
  arbitrage by getting risk weights right, even at
  the cost of enormous complexity
• Provide banks with incentives to enhance risk
  measurement and management capabilities
• Extend risk assessment to operational and
  interest rate risks

11
What wont change

• The definition of Tier 1 and Tier 2 capital
• Tier 2 can be no more that 100 of
• Tier 1
• Minimum ratio of capital to risk-weighted assets
  remains 8
• Focus on accounting data, not market values

12
Lack of Attention to the Numerator Undermines
Logic of the Approach

• Enormous attention to refining risk weights to
  replicate economic capital as closely as possible
• E.g. debate last year over including expected
  loss
• Then judge adequacy in comparison with
  numerator that is emphatically not an
  institutions capacity to bear unexpected loss
• Includes debt, hybrid instruments, some reserves,
  and a number of idiosyncratic, country-specific
  items
• Based on accounting values, not market values

13
Pillar 1 is responsible for most of the
complexity and compliance costs

• Capital charges for credit risk
• Standardized Approach
• Internal Ratings Based Approaches
• Foundation IRB
• Advanced IRB
• For most banks, capital charges for credit risk
  are likely to decline

14
But what the Basel Committee giveth, it taketh
away (on average)

• New capital charge for operational risk
  calibrated to offset the reduction in the capital
  charge for credit risk on average
• Capital charges for operational risk
• Basic Indicator Approach
• Standardized Approach
• Advanced Measurement Approaches
• Overall adjustments through single scaling
  factor

15
A New Regulatory Burden for Some Specialized
Institutions

• A particular burden where specialist banks
  compete with nonbanks
• Only 10 of the top 30 asset managers are banks
• 5 of the top 9 transfer agents in the US are
  banks
• May lead to exits from some lines of business
• May lead to acquisitions of specialists by
  diversified institutions

16
Compliance Costs are likely to be Heavier than
Intended

• Costs to banks
• Costs to supervisors
• Costs to economic stability

17
Costs to banks

• Developing, validating and maintaining regulatory
  models as well as internal models
• Complex international negotiations will
  inevitably lag innovations in risk management
• Already lags behind best practice
• Leading banks will need to run separate systems
  for regulatory capital and economic capital
• Inevitably will reduce resources available for
  modeling economic capital
• Costs of regulatory-induced diversions from
  preferred strategy if regulatory capital
  requirements were consistent with economic
  capital models
• Eg., too much credit for residential mortgages,
  too little emphasis on diversification, etc.

18
Costs to Banks (contd)

• For international banks, dealing with multiple
  regulators, multiple regulatory models,
  validation processes, supervisory procedures and
  disclosure requirements
• In principle, both PD and LGD will vary
  differences in legal infrastructure across
  countries
• Prospect of different reports to home and host
• Incentives create competitive disadvantages for
  banks who adopt less sophisticated approaches to
  credit and operational risk
• Level playing field objective rests uneasily
  alongside incentives for banks to qualify for the
  most advanced approaches

19
Costs to Supervisors

• Staffing to monitor and evaluate models
• Risk that a black box model or even a good
  model fit to bad data (or an insufficient span of
  data) may lead to disaster
• Monitoring compliance with multiple requirements
  and preventing cherry-picking
• Dealing with home/host issues for international
  banks under all 3 Pillars
• If home country, evaluating models used at
  foreign branches
• If host country, sharing meaningfully in
  oversight
• Particularly difficult if branch is systemically
  important in host country
• Nightmare case small in home country,
  systemically important in host
• Scandinavian exceptionalism?

20
Costs to Supervisors (contd)

• Assuming greater responsibility for outcomes,
  with certification of models
• Relaxation of market discipline can increase
  burdens on supervisory authorities

21
Costs to the economy

• Intensification of business cycles with capital
  requirements likely to rise precisely as capital
  resources fall
• Misallocation of resources to the extent that
  binding capital requirements diverge from
  economic capital
• Reliance on officially-sanctioned regulatory
  models weakens corporate governance and market
  discipline increases moral hazard
• May increase likelihood of herd behavior
• Deadweight costs of compliance that do not
  produce higher levels of safety

22
Are there sufficient, offsetting benefits?