Comments on Systemic Risk and Regulation

1
Comments on Systemic Risk and Regulation

• Charles W. Calomiris
• NBER Conference on Risks of Financial
  Institutions
• October 22-23, 2004

2
Basic argument of the paper

• Evidence suggests that risk transfer from banks
  to insurance companies has been occurring.
• Risk transfer can be motivated by capital
  arbitrage, if insurance company capital
  regulation is more favorable than banks.
• Risk transfer can reduce system-wide capital
  relative to asset risk.
• Risk transfer, thus, can make the system more
  fragile.

3
Missing pieces

• Key question is whether minimum regulatory
  capital requirements for banks and insurance
  companies are binding constraints.
• Generally, they are not, empirically.
• Stockholders risk preferences
• Debt market discipline (uninsured debts, GICs)
• Policy holders

4
Missing pieces (Contd)

• That was especially true historically, when
  deposit insurance was lacking market discipline
  set the risk of deposits, capital levels, and
  leverage (Calomiris-Mason 2003, Calomiris-Wilson
  2004). Best to conceive of equilibrium as
  three-party game (Calomiris-Kahn 1991).
• Definition of systemic risk and review of
  historical banking crises mixes liquidity
  problems (pre-WWI in US due to unit banking),
  with solvency problems (1930s). Capital
  constraints did not arise from the Depression.
  Minimum capital ratio constraints are a
  phenomenon of 1980s. Under market-disciplined
  systems, capital ratios were much larger than
  today, and capital varied with risk.

5
Missing pieces (Contd)

• Barth, Caprio, Levine (2004), Calomiris-Powell
  (2001), many others, find that market discipline
  is the only source of discipline in most
  countries banking systems today.
• Basel has been ineffectual (Japan) or irrelevant,
  and not even the main binding regulatory
  constraint in the US.
• Key questions are (1) What are real regulatory
  capital constraints, if any, and (2) Do they
  bind?
• In most countries, book capital constraints dont
  matter.
• Risk of systemic risk from capital arbitrage is
  dependent on collapse of market discipline.
  State-contingent problem (1980s).

6
Related question

• Is credit risk transfer to insurance companies
  different from credit risk transfer to
  non-regulated entities via securitization of
  risk?
• Banks retain concentrated credit risk from
  securitization, and so asset risk relative to
  capital of financial institutions goes up.
• Is this a problem, too?

7
Credit card securitization example(Calomiris-Maso
n 2004)
8
Regression evidence that securitizers are on a
market margin

• RegCapRat ManCapRat
• Adj R2 0.084 0.249
• PropSec -0.006 -0.041
• (CT)/TA -0.105 -0.083
• SDPastDue 0.871 0.879

9
Securitization and arbitrage

• Apparently, securitization has been an effective
  means to reduce capital, but this capital amount
  has been sufficient from the standpoint of active
  market discipline.
• Credit card receivables risks, according to the
  market, sometimes require less capital than is
  mandated by regulation.

10
Does market discipline bind generally?

• Look at risk of assets for various regulated and
  unregulated financial institutions and see
  whether book capital / assets and market capital
  / assets vary with asset risk at levels above
  minimum requirements (which are zero for IBs and
  FCs.
• Calomiris-Mason (1998 unpublished) shows that, in
  fact, market discipline is binding for all
  classes of intermediaries and capital is
  substantially above regulatory minimum.

11
Book equity/assets and asset risk, on average,
for various classes of financial intermediaries,
1996

• N EQ/TA SIGVA
• bhc 378 0.09138 0.25553
• inv bk 55 0.29358 0.91031
• lic 51 0.12796 0.27275
• pc ins 108 0.28110 0.52845
• fico 95 0.25442 0.96147

12
Externalities

• Is this still a potential problem from the
  standpoint of systemic risk? Is there an
  externality?
• Clearly, externalities come from liquidity
  crises, forced sales of assets. (Possible abuse
  of safety net is a separate, but important issue
  of concern, too.)
• So current market choices in excess of minimum
  requirements, or sometimes below them, may not be
  adequate.
• Von Peter (2004) model capital adequacy is
  important for preventing bad feedback
• loanloss? credit cut?AsPrDec?loanloss

13
Von Peter (2004)
14
Von Peter (2004)
15
Von Peter (2004)
16
Externalities (Contd)

• Is this still a potential problem from the
  standpoint of systemic risk? Is there an
  externality?
• If we think these are important, then perhaps
  increase prudential regulatory requirements.
• And political economy Reduce risk of forbearing,
  which is disastrous for incentives and macro
  stability.
• But book capital relative to risky assets by
  itself is not the way to go. It is crude,
  inaccurate, and not credible.
• Market discipline can be incorporated in regs
• Standbys in place of reserve requirements or
  capital requirements
• Basing book capital requirements on interest
  rates paid on loans.
• Requiring minimum amount of credibly uninsured
  debt (that cannot run) as part of capital.

17
Ex post interventions

• Ex ante regulation and ex post intervention are
  alternative policies. Growth vs. stability
  tradeoff?
• Very little work has compared various approaches
  to mitigating systemic consequences of capital
  shocks (inadequacy). (Calomiris, Klingebiel,
  Laeven 2004)
• Forbearance
• Loss sharing
• Subsidized purchase of preferred stock (RFC)
• Asset management companies
• Subsidized sales of banks
• Debt forgiveness (redenomination)
• Ex ante and ex post both work better in good
  institutional environments.