Financial Stability 1/2002
5 Risk in payment systems

5.1 International developments
5.2 The Norwegian settlement system
Box:
- The liquidity trend in banks
Since the late 1980s, the increased turnover and exposure associated with payment
settlements has led to a stronger focus on risk and to major changes in the
organisation of payment systems, both nationally and internationally.
5.1 International developments
Settlement systems in most countries used to be based on a single daily net
settlement with no mechanisms for limiting settlement risk. In a net settlement,
banks’ positions in relation to one another are netted and their liquidity
requirements for settling the position are thus low. However, the consequences
of a bank being unable to meet its obligations, so that the settlement is not
executed, could be substantial and unpredictable. The positions of the other
settlement participants could be affected, and at worst they might not be able
to meet their own commitments in the payment settlement. Since it was usual
in the past to credit customers’ accounts before settlement actually took
place, this could lead to considerable losses for banks (see box).
Routines for crediting customers
The transfer of a payment from Customer 1 to Customer 2 via their respective
banks will involve the following operations:
- Bank A debits Customer 1's account and sends a payment advice through
the payment system instructing Bank B to credit Customer 2's account.
- The settlement bank debits Bank A’s account and credit’s Bank
B’s account.
- Bank K credits Customer 2’s account.
Customer 1=> Bank A => Settlement bank => Bank B => Customer
2
The order in which these operations are carried out has a strong bearing
on the risk in the settlement system. If banks practice ”advance crediting”,
Bank B will credit Customer 2 before it has received settlement from the settlement
bank. This involves a risk for Bank B, as Bank A may become insolvent before
settlement takes place in the settlement bank. Without advance crediting,
Bank B will not assume an obligation in relation to Customer 2; nor will it
run a risk of loss in the event that Bank A becomes insolvent.
In order to reduce settlement risk, standards were introduced for net settlements.
The most important of these is the Lamfalussy standard (BIS 1990)2).
Net settlements must satisfy six recommendations concerning legal, organisational,
liquidity and operational conditions in order to be regarded as protected net
settlements according to this standard. In the Trundle Report (BIS 1999 and
2001)3) these recommendations have been made
more specific and expanded to the ten BIS Core Principles for Systemically Important
Payment Systems. The Trundle recommendations encompass both gross and net settlements
and also contain recommendations concerning the central bank’s role in
the payment system and responsibility for the implementation of the core principles.
2 BIS 1990 Report of the Committee on Interbank
Netting Schemes of the Central Banks of the Group of Ten Countries
3 BIS 2001 Core principles for systemically
important payment systems
Central to both Lamfalussy and Trundle recommendations is the requirement that
it must be possible to make a settlement even if the bank with the largest net
debit position lacks cover. The principle is that settlement execution should
be assured by the other participants supplying liquidity on behalf of the illiquid
participant (collective guarantee). The amount of liquidity individual participants
have to supply will tend to depend on the size of the bilateral credit lines
they have given the bank that cannot meet its commitments. It must be assumed
that in principle a bank will not provide lines of credit that cause it to have
problems meeting its own commitments. However, the possibility of more than
one participant failing cannot be excluded. In such cases, even protected net
settlements may involve too great a risk. In consequence it is recommended that
real time gross settlement (RTGS) be used for large-value payments.
With RTGS, transactions with cover are accepted for settlement immediately,
and the beneficiary’s bank normally only receives notification that payment
has taken place after the settlement has been executed. In contrast to protected
net settlements, real time gross settlements therefore do not involve credit
risk. The disadvantage of RTGS, however, is that no netting takes place, and
that banks therefore have to have cover for every single transaction. This increases
their liquidity needs compared with net settlement, and may lead to transactions
remaining in a queue waiting for cover. This situation has prompted the development
of hybrid solutions which draw on the advantages of both net and gross settlement.
Hybrid solutions may develop from gross settlement systems through the construction
of gridlock busters. A gridlock buster is used if a liquidity deficiency leads
to queues developing due to inadequate cover. Banks’ positions will then
be netted against one another and settled if the banks have cover for their
net positions. In this way payments can be settled continuously with a minimum
of liquidity. A hybrid solution may develop from net settlement systems if it
is made possible to accept parts of payment transactions that are due to be
netted for settlement if the banks involved have insufficient liquidity. The
remaining transactions are settled as more liquidity becomes available to the
banks.
5.2 The Norwegian settlement system
In Norway the majority of payment transactions are settled with final effect
in Norges Bank’s Settlement System (NBO). Large transactions (over NOK
100 million) and earmarked transactions have been settled since March 1999 in
an RTGS system, while small and medium-sized transactions go into NICS retail
netting and NICS-SWIFT settlements, respectively. Both are net settlements based
on multilateral netting among participants.
The netting effect in NBO
A net settlement is based on the result of netting a number of interbank
transactions. In consequence, banks require less liquidity than they would
for gross settlement, where they would require cover for each transaction
individually. The netting effect – and hence the advantage of netting
in terms of liquidity – can be calculated by comparing the banks’
net positions with the underlying gross positions. The estimated reduction
in liquidity needs due to netting is found to be just under 80 per cent for
NICS-SWIFT netting, and slightly over 80 per cent for NICS retail netting.
The netting effect (reduction in liquidity needs) obtained for NICS retail
netting is 83%, and for NICS-SWIFT 63%. These figures are based on data for
the first three months of 2002.
Transactions for RTGS settlement account for most of the turnover in NBO, and
the organisation of this settlement is therefore of great importance to the
efficiency and risk of the settlement system. In February 2002, crediting of
customers after settlement was introduced, and thus there is no longer any credit
risk associated with gross settlement. The establishment of an RTGS system in
1999 led to a strong increase in turnover in NBO (Chart 5.1). This implies an
increase in banks’ liquidity requirements. However, their liquidity requirements
are still limited, both because banks coordinate their exchange of transactions
and because a gridlock buster has been established in NBO which automatically
calculates banks’ net positions if queues develop.
Chart 5.1
In February 2002, crediting after settlement was also introduced for transactions
subject to NICS-SWIFT netting. Similar routines for NICS retail netting were
introduced earlier, in June 2000. Banks are thus not exposed to credit risk
when they participate in this netting. Mechanisms have also been established
to make it possible to reverse netting if a participant lacks cover. This satisfies
the BIS requirement of at least one daily settlement even if a participant cannot
fulfil its obligations. Such a solution could demand more liquidity than net
settlements protected by collective guarantees, because each bank must have
sufficient liquidity for its own position. On the other hand, banks normally
reduce bilateral credit lines in periods of market unrest. Banks will therefore
have to provide liquidity for their own positions also in solutions based on
the provision of collective guarantees. Although it differs on several points
from solutions chosen by some other countries, the Norwegian system is organised
in compliance with international recommendations. It will therefore be robust
in periods of market unrest.
Chart 5.2
In line with the BIS recommendations, Norges Bank contributes to assuring the
execution of payment settlements by making liquidity available against collateral.
A particular challenge in this connection is that Norwegian government debt
is relatively limited. As a result Norwegian banks have less government bonds
available for pledging as collateral than banks in some other countries (Chart
5.2). However, Norges Bank accepts more types of security as collateral than
most other central banks. For example, government and private bonds and short-term
paper from the OECD area can be used as collateral in Norges Bank – provided
that the country in question has not renegotiated its debt during the previous
five years, or the private company has a satisfactory rating. Since securities
of this type account for more than 80 per cent of the collateral pledged by
Norwegian banks to Norges Bank, it is reasonable to assume that Norwegian banks
have sufficient securities for use as collateral in payment settlements. In
addition, Norwegian banks have substantial deposits with Norges Bank, and there
are indications that these deposits will increase in the next few years (see
box). This will reduce the risk of problems arising in connection with payment
settlements in Norges Bank due to inadequate liquidity on the part of banks.
The liquidity trend in banks
The execution of payment settlements is contingent on banks’ having
cover, either in the form of deposits with Norges Bank or through a borrowing
facility secured by means of pledged securities. Banks’ deposits are
increased through payments from the government and reduced by payments to
the government. This is because, unlike other economic agents, the government
has its deposits in Norges Bank rather than private banks.
Because of the large government budget surplus, payments to the government
are larger than outgoing payments, which in principle should mean that liquidity
is being withdrawn from the money market. For the past few years, however,
the practice has been to channel the budget surplus to the Petroleum Fund,
and this just about offsets the withdrawal of liquidity. Moreover, some payments
to the government do not entail withdrawal of liquidity from the money market.
This applies to government interest income in Norges Bank, and the transfer
of Norges Bank’s surplus to the government. For the past few years,
these items have varied between NOK 7bn and NOK 12bn. If the government uses
this capital, a corresponding amount is transferred to the money market. In
addition, banks receive interest on their deposits with Norges Bank.
This liquidity can be withdrawn by the government issuing bonds, or by Norges
Bank increasing its use of fixed rate deposits. However, this withdrawal of
liquidity will have no effect on the execution of payment settlements in Norges
Bank, as Norges Bank accepts both fixed rate deposits and bonds as collateral
for loans. In isolation, the factors described above indicate that banks will
have a greater opportunity to obtain liquidity for the payment settlements
over the next few years.
© 2002 Norges Bank webmaster@norges-bank.no