BANKS AND THEIR REGULATION I

1
BANKS AND THEIR REGULATION I 1. What is a
bank? Why are banks as they are (asset and
liability structure)? Historical
approach Theoretical approach 2.
Regulation Motives for regulation of financial
institutions of banks Methods of regulation 3.
What has changed? Implications for banks and
their traditional functions regulation
2
1. What is a bank? The simplest representation
is Assets reserves plus loans (illiquid,
non-marketable) Liabilities demand deposits
plus capital Functions Central to the payments
mechanism Liquidity provision - to both
depositors and borrowers (overdraft
facilities) Main source of credit to small
borrowers Key characteristic is maturity
transformation
3
Historical development (Iberian peninsula in 15th
century - Rajan) Safe-keeping function (cash
inconvenient, unsafe) Why not just that? Risk of
theft by the banks owner, or by the state, so
lending may be safer than storage. This is
because of so-called external incompleteness of
contracts a safe-keeping contract can be written
in a straightforward way, but the external
framework of law and its enforcement is
inadequate to ensuring that the contract will be
upheld. Willingness-in-principle to lend to
depositors, but with no guarantee. This reflects
so-called internal incompleteness of contracts,
which means here that the exact conditions under
which depositors will be able to get loans are
too complex to be precisely defined (ie precisely
defined enough to allow legal recourse by the
would-be borrower after a refusal to lend). Banks
will lose reputation, however, if they make
unreasonable refusals to lend. Main point. We may
think it odd (see below) that maturity-transformin
g banks exist at all Rajan has gives us a
historical explanation, but this will make us ask
whether such banks will endure, since the types
of incompleteness of contracts that gave rise to
them are now less important.
4
Modern evolution External incompleteness is much
less important, so that banks no longer lend
because this is the safer option. Given that, why
do banks still have mismatched maturities?
Borrowing short and lending long carries,
potentially, two risks. The first is that, if
long lending is at fixed interest, banks are
exposed to an unexpected rise in the short rate
of interest, which could lead to them paying a
higher interest rate than they are earning (see
the Savings and Loans crisis, next week). The
second is that banks are exposed to runs (see
below). Since, in fact, there is nothing to stop
banks from charging a variable interest rate on
loans, the second risk is the key
one. Traditional banks have advantages, though,
in providing (illiquid) credit. Most important,
they act as information processors, via their
role in accepting deposits. They assemble a good
deal of private information about would-be
borrowers who have a history of depositing with
them, and even more information about those who
have already borrowed. Information asymmetry is
thereby reduced and the ability to separate good
risks from bad gives the banks an advantage over
other potential lenders. Rajan uses the term
relationship credit in this context. It is fair
to ask whether the vital information asymmetries
are as important now as they once were.
5
The Diamond-Dybvig model This models the
maturity-transforming bank in a two-period
framework. Loans mature after two periods, and
are worth little if called in after one
period Individual depositors may or may not need
liquidity after one period (think of random
medical bills) Bank must decide what fraction of
assets to hold in the form of reserves Results Un
less the bank holds 100 reserves, it is possible
that it will fail for sure (be unable to meet the
period 1 needs of depositors). With less than
100 reserves there are, in any case, two
equilibria (1) the no-run equilibrium in which
only depositors needing liquidity withdraw so
that, with high probability, the bank survives
through to period 2 (2) the run equilibrium, in
which all depositors withdraw and the bank
fails. To make this work, the model assumes that
a depositor who has not tried to withdraw in the
run equilibrium gets nothing, whereas one who has
tried will get something (or a chance of
something). So if others run, you want to run,
and if they dont, you dont (given a small cost
of withdrawal) Prisoners Dilemma. This model
explains runs on solvent banks and provides a
strong formal justification for having a lender
of last resort (LOL)
6
2. Regulation Motives microprudential and
macroprudential Microprudential Any financial
transaction, by definition, entails the exchange
of cash for a promise. Think of bank deposits,
share purchases, insurance premia and pension
contributions. Therefore risk of fraud or, less
extreme, of a failure to respect contractual
conditions, is intrinsic, and a case for
regulation of all financial institutions not
just banks - can be derived from this.
Macroprudential The Diamond-Dybvig model has
bank runs in equilibrium. It is not clear,
however, what might trigger a switch from the
no-run equilibrium to the run equilibrium. One
reasonable candidate is the failure of another
bank. This is the idea of contagion. A
distinction between illiquidity and insolvency is
needed here. Bank A might become illiquid, ie
unable to satisfy desired withdrawals, and this
could lead to failure and contagion. The risk of
such a liquidity crisis justifies the creation of
a lender of last resort (LOL), as after the US
bank failures of the Great Depression.
7
Is LOL enough? Not if there are fears of
insolvency. For in that case, the last depositor
to try to withdraw during a run will get nothing,
even with LOL hence fear of insolvency, too,
can produce runs and contagion. This leads to the
case for deposit insurance, under which in its
simplest form bank deposits are 100 guaranteed
by the Central Bank. Then even fears of
insolvency will not lead to bank runs. Deposit
insurance and moral hazard Full deposit insurance
of course means that depositors do not need to
worry at all about the solvency or otherwise of
their bank. Does this matter? Well-capitalized
banks. Consider a bank for which the risk of
failure is zero. As the owner of such a bank, the
expected return on a given loan is your expected
(gross) return. Assuming that you are not
risk-averse, you will rationally make loans for
which the expected return is at least as great as
the cost of borrowing. Even if you are
risk-averse this result may hold approximately,
in the case where you make a large number of
small loans whose risks are idiosyncratic, i.e.
independent of one another. This outcome is
socially efficient.
8
Poorly-capitalized banks. Consider a bank with a
high risk of failure, i.e. low or negative
capital. Now losses on new lending will be borne
by the deposit insurer, while gains if large
enough will restore solvency and will accrue to
the banks owner. It is clear in this case, that
the owner may make loans whose gross expected
return falls short of the cost of borrowing (and
may even be negative). This is known as gambling
for redemption and will raise the cost to the
deposit insurer when the bank finally
fails. Methods of regulation addressing the
moral hazard problem Capital adequacy
constraints The so-called Basel ratios banks
must hold 8 of risk-weighted assets in the form
of capital (i.e net worth). An old example from
Rajan is 14th century Barcelona, where a bank had
to have deposited 1000 marks of silver with the
authorities to be allowed to display a tapestry
bearing the shield of the city. Subordinated
debt Force banks to issue some debt that is
junior to deposits. The market interest rate on
this debt can act as a signal to regulators
9
3. What has changed? Technology fast, reliable
communication of information about, for instance,
transactions, and credit records of potential
borrowers Information availability role of
private information of banks is reduced more
creditworthiness information is public, or
potentially so Property rights environment
greater security of property rights and easier
enforcement of (more complex) contracts. Implicat
ions for (a) banks and their traditional
functions, (b) regulation (a) It is possible to
envisage an institutional structure in which at
the centre of the payments mechanism would be
money-market mutual funds instead of banks. MMMFs
hold mainly liquid Treasury Bills and other safe,
liquid, short-dated assets. Their liabilities
would be like bank deposits (i.e. usable for the
purposes of writing cheques and making debit card
transactions). This would make the payments
mechanism automatically immune to the dangers of
bank runs. Another technological advance, the
possibility of what is called fully-collateralize
d real-time settlement can further protect the
payments system by eliminating the lag between an
instruction to transfer funds and the transfer
itself (think of debit cards).
10
What institutions would handle long-term credit
to households and firms? This could be done by
specialist institutions that would issue not
deposit liabilities, but tradable units instead
(as do UK unit trusts and investment trusts).
(b) On regulation, Rajan argues that if the
institutions at the centre of the payments
mechanism, i.e the MMMFs, are immune to the
threat of runs, then both LOL and deposit
insurance become redundant. However, efficient
functioning of the economy still requires that
firms can get access to short-term credit and, as
we see next week, it is possible for systemic
liquidity crises to occur. Hence Rajans
conclusion Once the payments system is
immunised from the idiosyncratic risk of
participants, the only economic reason for
regulatory intervention is if there is a systemic
crisis in institutions specializing in
relationship credit. These institutions may, or
may not, be what have traditionally been called
banks.