BANKS AND THEIR REGULATION II: Liquidity crises and solvency crises

1
BANKS AND THEIR REGULATION II Liquidity crises
and solvency crises 1. Liquidity crises and
intervention (Brimmer) Liquidity crises are a
fact of life Contagion argument that Central Bank
must intervene. There are severe economic costs
if liquidity provision to enterprises is
disrupted and/or if solvent but illiquid
providers of relationship credit are forced to
close. We look at some examples of liquidity
crises, first some from the US Penn Central
Railroad (1970) Penn Central became insolvent and
the Federal government refused a rescue Default
on Penns 3-month commercial paper (CP) was thus
inevitable This led to contagion risk in the CP
market there was a likelihood that other
enterprises would find it difficult to roll
over their CP borrowing (why? note relevance
here of imperfect information). Enterprises would
therefore turn to the banks predictable rise in
bank credit. No excess reserves at the time, so
the Fed intervened to promise extra reserves if
needed. This intervention was successful in all
respects.
2
The 1987 stock market crash The Dow Jones index
of US equity prices fell by 25 in two days.
Many speculators had bought securities (or the
Dow) forward on margin. Such forward contracts
do not have to be paid for, but brokers who
exercise the deals on behalf of their clients
will demand some cash up front. If the forward
price falls, then there is an expected a loss on
the contract, and the broker will demand more
cash to cover that expected loss. This demand for
more cash is known as a margin call. The client
has to get the cash from somewhere, and this
leads to extra demand for bank credit. (technical
details in appendix) As in the Penn Central case,
failure of the Fed to expand reserves to
facilitate bank credit expansion could have led
to a liquidity crisis and a freezing of the
payments mechanism. Therefore the Fed made extra
reserves available. So these two examples have in
common (a) an anticipated rise in the demand for
bank credit, (b) the risk that, in the absence of
extra Central Bank lending (at a non-penal rate),
the market for interbank lending could dry up,
(c) this could threaten the operation of the
payments mechanism, which requires banks to be
willing to lend freely to one another on an
overnight basis.
3
The failure of the hedge fund Long Term Capital
Management (LTCM) 1998 See Krugman/Obstfeld 6th
ed. pp 653-4 LTCM was a fund that arbitraged, on
a huge scale, in the US Treasury Bond market
(between old and newly issued 30-year bonds, as
it happens, but the details dont matter). Assets
were 1,300 billion and Capital was 4.8 billion,
equal to only 0.35 of assets. The 1998 default
on Russian government bonds led to a crisis for
LTCM (mechanism unclear, but perhaps the
historical bond yield regularities on which they
were relying were upset). The New York Fed
organized a rescue when LTCMs capital had fallen
to 0.6 billion. LTCMs creditors were induced to
put up 3.6 billion in exchange for (a) 90 of
profits, (b) control. Why rescue? (a) banks
directly exposed to LTCM, (b) risk of contagion
via asset price falls if LTCM had been forced
into an asset fire sale. Bailout? No, a
bail-in of the banks. No public funds
committed. NY Fed played a coordinating role.
4
Northern Rock , UK 2007 Key elements Northern
Rock (NR) is a UK Building Society and its
solvency is not in question. Like all Building
Societies its assets consist of mortgage loans.
What makes it unusual is that 75 of its
liabilities are in the form of commercial paper
(CP), and only about 25 are deposits. So, as in
example (4) from topic 6, NR is vulnerable to a
fall in demand for CP. Why has such a fall
occurred? The origins lie in the US market for
so-called sub-prime mortgages. Fall in US
housing market raises expected default losses for
holders of mortgage debt, so the value of such
debt falls. The market values of complex and
hard-to-value derivatives based on such mortgage
debt also fall, quite possibly by more than would
be warranted by the fundamental values of these
assets. Some markets may freeze, meaning that
it is impossible to sell the assets
concerned. The effects spill over from the US
into other financial markets, partly for reasons
of confidence, but partly because financial
markets are in fact international.
5
So we have a sharp fall in the supply of
interbank credit and a sharp rise in its price
(plenty of evidence of this, starting on August
9th 2007. Consequence for NR? Market fears that
it would be unable to refinance its maturing CP
were self-fulfilling. The critical point was
reached on Sept 14th,just before NR was going to
be obliged to issue a public profits warning.
Crucially the normal Lender of Last Resort
facility, which can be operated anonymously by
the Bank of England, would not help because
NRs assets are not eligible as collateral. So,
without a change in policy, the Bank could (it
has claimed) do nothing in the period between
foreseeing the problem (Aug 9th) and the crisis
proper (Sep 14th). In Governor Mervyn Kings
words the fact that he could anticipate that the
train might hit the buffers was not a reason to
blow it up beforehand. The sequence of events
then was (a) the Bank created a special LLR
facility and had to do it publicly (b) this
precipitated a run on NR by its depositors (c) to
stem this, the Chancellor ot the Exchequer
extended the deposit guarantee (from 90 and
limited to 100 and unlimited) and in principle
to all deposit-taking institutions.
6
Parallels? NR provides both parallels and
contrasts with the case of LTCM. The main
parallel is an increase in liquidity preference
threatening the ability of the institution to
survive (i.e. to be able to redeem maturing CP
liabilities). There are several contrasts (a)
LTCM was close to insolvency while NRs problem
was illiquidity not insolvency. (b) US banks
held claims on LTCM, so insolvency of LTCM could
have rendered some US banks insolvent (c) Because
of (b) other banks had a direct interest in
saving LTCM this was not true of NR
7
2. Looting some economics of capital inadequacy
and its implications FIs that are close to
insolvency may behave in socially inefficient
ways. (a) Gambling for redemption See last
weeks lecture. It is rational to take unfair
gambles if you will enjoy any gains but not have
to bear any losses. (b) Looting, or bankruptcy
for profit If the capital of an (FI) is low
enough, a situation can arise where the maximized
current period return to the owner exceeds the
maximized expected value of the firm at the start
of the next period. Then bankruptcy for profit,
or looting is rational. Rest of lecture based
on Akerlof/Romer 1993
8
Let A represent a vector of assets/liabilities
chosen by the owner in the current period
(actually A might just as well stand for
actions). Define M(A) as the owners current
period revenue Define V(A) as the value of the
enterprise at the end of the period, including
M Let the maximized value of M be M and the
maximized value of V be V. Suppose, as is
reasonable, that when MM, VM (bankruptcy).
Then we can distinguish three possibilities (a)
V high and bigger than M. Then we have a
healthy FI that does best to remain a going
concern. (b) V low and equal to M Now looting
is optimal take the money and run. Clearly,
for given V, looting is more likely to be
optimal the larger is M. This is favoured by
(i) weak regulation (low public discipline
makes it easier to loot, raising M) (ii) deposit
insurance (no private discipline via depositor
withdrawals)
9
Looting is logically distinct from gambling for
redemption with looting, any future returns are
strictly irrelevant. Looting can induce perverse
behaviour in counterparties, since it is in the
interest of the FI to make loans with a low
expectation of repayment, but which produce high
current period revenues in the form of interest
or origination fees. Examples include (a) fake
developments, (b) junk bonds. Looting is made
easier if ownership of the FI is concentrated or
interlocked. Concentrated ownership lowers the
risk that actions on or over the border of
illegality come to the notice of regulators.
Interlocked ownership allows for the portfolios
of institutions that are being looted to be
stuffed with worthless assets (see
Southmark-Jacinto story).
10
2.1 Looting examples and methods 2.1.1 Chilean
banking system after liberalization in 1979 In
1979 Chile embraced internal and external
financial liberalization, and fixed its peso to
the as a means of inflation control
(international monetarism). However
backwards-indexed wage contracts led to severe
real overvaluation and many insolvencies. Owners
of insolvent firms have a clear incentive to
borrow as much as possible and then abscond with
the money (making the lending bank insolvent as
well). This happened but why did the bank
owners, and their depositors, allow it? Owners.
As part of the liberalisation package, ownership
restrictions on banks had been removed, so that
Chilean conglomerates (grupos) could purchase
banks, from which they could then borrow.
Moreover, the funds needed for buying a bank
could be obtained by a loan from the self-same
bank., an elegant manoeuvre known as the
bicycleta. (See also Goodhart FT
2007). Depositors. The banks were able to
attract deposits from overseas, following
liberalization of international capital flows.
Deposit insurance had been abolished in theory,
but depositors were bailed out in practice.
11
2.2.2 Savings and Loans institutions in the US,
1980s Assets were long term fixed-interest loans
for house purchase. Regulation Q, which placed a
ceiling on the deposit rate which could be paid
by US FIs provided protection to the SLs against
adverse interest rate changes. Liabilities were
federally-guaranteed deposits. Late-1970s
inflation forced up nominal interest rates and
regulation Q had to be repealed (in the face of
an outflow of from the US to the Eurodollar
market, where interest rates could not be
regulated). So a solvency crisis for the
SLs. Federal response was deregulation capital
requirement lowered from 5 to 3 lending-type
and ownership restrictions removed.
12
Now the simple arithmetic of looting was that any
loan that could yield a current return of more
than 3, net of deposit interest, was profitable.
For the bank owner only needed to put up 3 of
any loan in the form of extra capital. No-recourse
development loans Developer puts up, out of
the loan, an origination fee (typically 2.5)
and interest, as well as paying himself or
herself a fee as well. Junk bonds High risk, high
interest rate bonds used in the 1980s to finance
take-overs and management buyouts. The key broker
of these bonds was Michael Milken (of Drexel
Burnham Lambert), subsequently jailed. Akerlof
and Romer argue that SL finance delayed the
collapse of this market. Asset reshuffles Example
is Southmark-Jacinto. Southmark was the parent
company of Jacinto, a SL. Southmark swapped a
bundle of worthless companies for 140 million of
real estate owned by Jacinto as part of the
looting strategy. So, when Jacinto went into
bankruptcy, the burden on the federal government
was 140 million higher than it would otherwise
have been.
13
Appendix Margin Calls I have contracted to buy
some GM shares on Ist March next for 100. I have
done this through a reputable broker, as has my
counterparty (the seller), and no money has to
change hands now. This futures contract is
marketable - you can buy and sell them out
there. So long as the price remains at or above
100, my broker faces no risk. If I walk away, he
just sells the contract on to someone else and
loses nothing (maybe he gains). If the price
should fall to 80, my broker is not so happy.
Hes going to need 100 on 1st March, and I might
disappear before then. So he makes a margin
call, demanding 20 from me right now. I need
to get the money from somewhere, and if I dont
have it to hand, I am going to the bank. Result
a sharp rise in the demand for bank credit, which
the banks are going to want to meet. Nobody is
insolvent here maybe Ill lose my house, though.