Liquidity and Money Market Operations

1
Liquidity and Money Market Operations by C.A.E.
Goodhart Four main problems have been
highlighted by recent crisis- (1) Stigma
effect of ELA/LOLR Most Central Banks (CBs) have
run a version of a Corridor system, whereby
policy rate is maintained by a variety of market
operations. Misses, and other pressures, limited
by upper band, standing facility or discount
window, and lower band, paying a positive
interest rate on deposits at CB. But use of upper
band, and LOLR, compromised by stigma,
reputational effect.
2
(2) CBs only control one price (interest rate)
can they affect two, or more, points on the yield
curve? No problems arose in providing banks with
as much cash as needed since the turmoil arose.
Overnight money market awash with cash. Problem
was overnight rates consistent with policy rate
also consistent with one month/three month LIBOR
rates at emergency levels. Can CB carry out
Operation Twist, i.e. lend (Term Auction
Facilities) at longer maturities, while
withdrawing funds overnight?
3
(3) Bank liquidity ratios have been continuously
declining. Can/should this be reversed? Commerci
al banking systems have been putting the
management of liquidity onto the shoulders of
CBs. A worse put than the Greenspan put. (4)
CBs saw credit crunch coming, but did
nothing Failure of will, or lack of
instruments? Instruments (i) interest rates
predicated to price stability (ii) CARs
procyclical. Devise a scheme to improve issues
(1), (3) and (4). The preferential access scheme
(PAS).
4

• The Preferential Access Scheme
• Objectives
• to get rid of stigma problem
• to provide an instrument which can be varied over
  time to affect liquidity
• to give banks an incentive, especially in normal
  times, to hold adequate liquidity.

5
Method To have a corridor for each bank
individually which is initially zero (i.e. free
liquidity, apart from opportunity cost of using
collateral) and widens out in a series of
(probably equally sized) steps to the common
corridor of plus, or minus, 100 b.p. (1).
6
Each step of length X would be X of bank js
(sterling retail deposits in UK bank) from a base
(3? 6?) months ago. It is assumed that retail
deposits can be legally identified. Why that
base? Because those are the fragile depositors,
and it protects UK located banks. Why back-date
the base? To generate frictions to prevent
gaming and artificial creation of
deposits. Borrowing, and deposits, could be at
length up to 30 days, at rates shown per tranche.
Not more than 30 days, since the size of tranche
X could be varied by BoE at one month notice,
(though normally left unchanged). Borrowing
would need to be secured by collateral in normal
way.
7
Why does this help? (1) The first, zero spread
tranche is free liquidity, apart from opportunity
cost of using collateral. Banks should normally
borrow to get that advantage. So all UK retail
banks would, almost always, be borrowing from
BoE. The total amount borrowed, and marginal
rate of borrowing, would remain secret, but
should be easier to keep it thus. So the stigma
effect should be much curtailed, if not
eliminated, at least for UK retail
banks. (2) The choice of X is a policy variable
(bounded between 0 and 100). It could be raised
in liquidity shortage periods and lowered during
booms. This would also be a useful signal.
(N.B. if adopted, the current choice of X would
have to be worked out.)
8
(3) The PAS has been drawn symmetrical. That is
simple but not necessary. The length of
borrowing steps could be shorter (or longer) than
length of deposit steps. (4) Why not flood UK
(retail deposit) banks with free liquidity by
making X huge? Indeed possible to do so in a
crisis, but NZ recent experiment showed that risk
averse banks will just sit on CB deposits, if
there is little penalty of doing so. As UK banks
get more liquidity, they must have an increasing
incentive to spread it around.
9
Long-run incentive to hold liquidity The above
deals with objectives (1), removing stigma, and
(2), giving BoE an extra usable instrument, but
not yet (3) an incentive to hold adequate
liquidity. Objective 3 can be tackled by stating
that, once normal times have been restored, or by
end 2008, whichever is sooner, X will be
calculated as a of each banks retail
deposits, interacted with a variable which is a
function of each banks prior assessed liquidity.
My proposed liquidity variable would be the
average of the coverage ratio (say at 1 week, 1
month and 3 months) at two prior dates (say 1 and
2 years previously). The aim would be to allow
banks to run down their liquidity in crises
without completely eliminating their future
access to PAS.
10
It would be important that each banks relevant
coverage ratio (for the appropriate horizons)
should be published and known, and that each
banks access to PAS should also therefore be
public knowledge. This should provide an
incentive for all UK retail banks to keep
adequate liquidity in normal times. Banks
without access to PAS, or those who had exhausted
their PAS, could still access the normal upper
band facility.