Transcript Banking crisis resolution - lessons learnt

Banking Crisis Resolution:
Lessons Learnt
Franklin Allen
Banking Crisis Resolution:
Theory and Policy
Norges Bank, June 16-17, 2005
Basic Issue
When and how should central banks and
other government entities intervene in
times of financial crisis?
2
Benchmark Case
• Intervention is not necessary when
– financial markets and
– financial contracts
are perfect and complete.
3
Deviations
• When we have a deviation from this ideal
state we may (but may not) have
– a financial crisis
– a market or contracting failure that may justify
ex ante or ex post intervention
4
Key Issues
• What failures caused the crisis?
• Can you trust market prices in the sense
that they reflect future cash flows?
5
Factors that contribute to crises
1. Positive asset price bubbles
2. Negative asset price bubbles
3. Macro factors such as currency regimes
or inconsistent government policies
4. Moral hazard problems
5. Panics
6. Risky investments and growth
6
1. Positive asset price bubbles
Simple theory suggests
Asset price = Discounted PV of cash flows
Often this does not seem to hold
e.g. Norway and Japan
7
1. Positive bubbles (cont.)
What is the market failure that leads to
asset price bubbles?
• Irrational agents
– In this case shouldn’t rely on markets
– This was the rationale for much of the
financial repression that occurred from 19451970’s
8
1. Positive bubbles (cont.)
Other (rational) explanations
• Agency problems
– Investing with borrowed money creates an
incentive for the borrower to take risks if the
lender cannot assess the risk of the
investment
9
1. Positive bubbles (cont.)
• Asymmetric information
– Heterogeneous beliefs
– Lack of common knowledge
• These factors interact with credit
expansion
10
1. Positive bubbles (cont.)
• Whatever their cause, positive asset price
bubbles sow the seeds of future financial
crises
– Collateral values are inflated
– Distorted asset prices lead to inefficient
investment and future NPLs
11
1. Positive bubbles (cont.)
• Important factor in Norway and particularly
Japan for subsequent banking crisis
– Avoid if possible
– If an asset price bubble does cause a banking
crisis it is important to solve the banking crisis
and move on as quickly as possible
12
1. Positive bubbles (cont.)
• Scandinavian countries did things well
– Banks’ balance sheets were corrected quickly
– Costs from doing this were not high (Norwegian
government arguably made a profit!)
– Part of the reason they were able to do this is that the
banks collapsed quickly
13
1. Positive bubbles (cont.)
• Japan was less fortunate in that its banks
were stronger and kept going after the
asset price collapse
– Debt overhang arguably caused immense
problems for the Japanese economy
– Traditional Keynesian stimulation policies
caused Japan to go from one of the least
indebted countries to being the most indebted
14
Sign of Economic Recovery
Real GDP growth
(year on year, %)
8
6
4
2
0
-2
-4
90
91
92
93
94
95
96
97
98
99
00
01
02
03
04
CY
15
1. Positive bubbles (cont.)
• Great reluctance to spend public money on
cleaning up problems in the financial sector but
no reluctance to spend enormous sums on
public works programs
• Current levels of government debt will cause
significant problems going forward
• Would focused intervention in the financial
sector to eliminate debt overhangs have worked
better?
16
1. Positive bubbles (cont.)
• A comparison of the Scandinavian and
Japanese experience after asset price
bubbles underlines the importance of
prompt and effective crisis resolution
• Who should be punished?
17
2. Negative asset price bubbles
• Often in banking and other financial crises
asset prices fall to very low levels and
after a short while rebound quite
substantially, e.g. Asian crises
• These low asset prices can cause great
damage to the financial infrastructure
• E.g. LTCM
18
2. Negative bubbles (cont.)
• How do such “negative bubbles” occur?
• Consider the case with no central bank to
provide liquidity then the market for long
term high return assets must provide
incentives for participants to hold liquidity
19
2. Negative bubbles (cont.)
• Suppose no participants held liquidity then
during a crisis when the long assets are sold
there would be nobody able to buy them and
their price would fall to zero
• This cannot be an equilibrium because
somebody could hold some cash or liquidity and
buy a large amount of assets cheaply – this is
the incentive to hold liquidity
20
2. Negative bubbles (cont.)
• In equilibrium some market participants
hold a small amount of cash to purchase
assets in crisis times
– There is an opportunity cost of holding liquid
assets in non-crisis times
– When the crisis occurs the price falls to “fire
sale” levels and they (just) recoup their
opportunity cost of holding liquid assets by
purchasing at this low price
21
2. Negative bubbles (cont.)
• There is “cash in the market pricing” and
asset prices do not reflect future payoffs of
the asset
• Instead the asset prices in the crisis state
provide incentives to hold liquidity
• Not an efficient solution – efficiency
involves complete markets ex ante
22
2. Negative bubbles (cont.)
• This is why banks and other financial
institutions may be solvent but illiquid
• Current prices of the assets are low but
this does not reflect their future earning
power or the ability of a bank to cover its
future liabilities
23
2. Negative bubbles (cont.)
• What if there’s a central bank?
• Can’t the central bank provide liquidity through
open market operations rather than have private
participants holding it?
• In many cases the answer is yes but in
segmented markets with asymmetric information
the answer is no – LTCM again
24
2. Negative bubbles (cont.)
• Implications for crisis resolution:
If liquidity is scarce and open market
operations cannot supply it in times of
need don’t trust market prices
Lender of last resort function may be one
way to solve this problem
25
2. Negative bubbles (cont.)
• When prices cannot be trusted then
distinctions between illiquid, solvent and
insolvent become very difficult
• If there will be substantial spillovers to the
real economy of a disruption it is better to
save the institution in the short run and
find out find out more information
26
2. Negative bubbles (cont.)
• The IASB and FASB have suggested that
a move towards mark-to-market
accounting is desirable but banks and
insurance companies have resisted
• If negative price bubbles are a problem
then cost based accounting will be better
27
3. Macro driven crises
• As Patrick Honohan pointed out many of
the crises that we observe appear to be
driven by macro factors such as
unsustainable currency regimes or
inconsistent macro policies or widespread
systemic fraud
• Many of the emerging country crises
discussed yesterday were of this type
28
3. Macro driven crises (cont.)
• Here problem is to solve underlying macro
problem as well as solve the banking problem
• Banking crises are more of a symptom than a
cause
• Resolution issue: restore the functioning of the
financial sector as soon as possible to minimize
spillovers into the real economy
29
4. Moral hazard problems
• One of the factors that we hear a great deal
about is the problem of moral hazard and crises
• The basic idea is that if there are government
guarantees or IMF guarantees then banks or
borrowers or somebody else who benefits from
the guarantees will have an incentive to take
risks – they obtain the upside if the risks pay off
while the guarantor bears the downside
30
4. Moral hazard problems (cont.)
• The incentive to take risks leads to a crisis
eventually
• There appear to be a number of examples
where this kind of moral hazard has been
important, e.g. US S&L debacle in the
1980s
31
4. Moral hazard crises (cont.)
• If moral hazard is the cause of the crisis
then it’s important this be taken account of
when designing the crisis resolution
• Important to ensure shareholders and
subordinated debt holders receive nothing
and top management should be replaced
and their rewards should be limited
32
5. Panics
• Much of the theory of crises has focused
on crises as panics (Kindleberger 1978,
Diamond and Dybvig 1983)
– If people believe there will be a crisis then it
can be self-fulfilling
– If they believe there will be no crisis then their
beliefs will again be fulfilled
33
5. Panics (cont.)
• In cases where panics are the important
driving force then policies should be
designed to rule out the bad equilibria, e.g.
deposit insurance
• Important empirical issue of how many
crises are panics rather than being
fundamental driven
34
5. Panics (cont.)
• Resolution issue: Design policies to try to
eliminate multiple equilibria
35
6. Risky investment crises and
growth
• There is some empirical evidence that countries
that have occasional crises grow faster than
countries without crises (Ranciere, Tornell and
Westermann 2003)
• Countries which have financial systems that
allow investment in large amount of high return –
high risk assets may have faster growth and
more crises, e.g. late 19th century US
36
6. Risky investment crises and
growth (cont.)
• Crises driven by good risk taking are
relatively little analyzed and understood
• Here issue is not to punish people as in
moral hazard crises but rather to help
share risk so that more can be borne and
higher returns earned
37
6. Risky investment crises and
growth (cont.)
• Much more analysis of this type of crisis
and its implications for crisis resolution is
needed
38
Problems Going Forward
•
•
•
•
Bankruptcy of a large multinational
financial conglomerate may lead to a
systemic crisis (Herring 2003)
Bankhaus Herstatt
Drexel Burnham Lambert
BCCI
Barings
39
Problems Going Forward (cont.)
• Next crisis in an advanced economy may well be
quite different than previously
• E.g. 1 Hedge fund melt down
• E.g. 2 Problems over Taiwan lead to large sales
of US Treasuries
• E.g. 3 Global imbalances lead to large exchange
rate movements
40