Robert Lindley

+41 61 280 8292

[email protected]

Reducing foreign exchange settlement risk 1

Much progress has been made in reducing settlement risk in foreign exchange markets,
particularly through use of CLS Bank. However, the remaining exposures are
sometimes still significantly large and not always well managed, creating the potential
for systemic risk. To address this problem, it is particularly important that prudential
regulators promote effective management of the risk by market participants.

JEL classification: G15, G18, G2, G21, G28, G32.

Foreign exchange settlement risk has proved to be a persistent and

problematic issue in financial markets. Despite much discussion and even a
significant amount of action, the size and nature of the risk mean that it could
still disrupt the stability of global financial markets.
This special feature examines the results of a survey that took place in
April 2006 to assess the degree of risk. The survey was carried out for the
Committee on Payment and Settlement Systems (CPSS) by 27 central banks
and involved 109 institutions (both banks and non-banks) that were selected to
cover 80% of the foreign exchange (FX) market in 15 currency areas
(CPSS (2008)). This feature first sets out the background to the survey, and
then summarises the surveys key findings. Next it explains why there is still a
problem with FX settlement risk, and finally it suggests that there are two key
actions which need to be taken if the problem is to be addressed effectively.

The nature of FX settlement risk

Trading in financial markets typically requires settlement delivery of the asset
by the seller and payment for it by the buyer. The market for foreign exchange
is no different, except that settlement involves two payments ie the exchange
of one currency for another. Although FX settlement is often regarded as a
routine activity that is less interesting than the trading itself, it deserves close
attention because of the risk that can be involved, namely the risk that one
party to an FX trade pays the currency it has sold but fails to receive the
currency it has bought. The risk arises because, using the traditional method of

1
The author thanks Jimmy Shek and Marcus Jellinghaus for their technical assistance. The
views expressed in this feature are those of the author and do not necessarily reflect those of
the BIS, the CPSS or the central banks involved in the survey.

BIS Quarterly Review, September 2008 53

settling trades, there is no mechanism to ensure that you pay only if you are
paid (a mechanism called payment versus payment or PVP). Both
counterparties to the trade therefore commit themselves to paying away the
currency they are selling before they are certain that they will receive the
currency they are buying. Moreover, the traditional settlement process can be a
relatively slow one, meaning that the counterparties can be exposed to this risk
for a significant period, often more than a day (see Box 1 for more on how the
risk arises).
FX settlement risk (sometimes also known as Herstatt risk 2) is therefore FX settlement risk
is a significant
primarily a counterparty risk. It is equivalent to the risk of making an unsecured
counterparty risk
loan to the counterparty: you have paid money to the counterparty with no
guarantee that you will be paid back. As such it involves both principal risk (you
may not get paid at all, so you may lose the full value of the trade) and liquidity
risk (in this context, the risk that, even if the counterparty does pay you, the
payment comes at the wrong time and/or in the wrong currency, leaving you
without the currency you need when you need it). 3 Given the size of the FX
market estimated to involve daily turnover equivalent to $3.2 trillion in April
2007 the potential risk is significant. 4
Because of this, in 1996 the G10 central banks launched a comprehensive Central banks have
a strategy to
strategy to contain FX settlement risk. At the time, the duration and size of FX
reduce it
settlement exposures tended to be underestimated by banks, while their risk
management measures were often inadequate. Indeed, the scale of exposures
arising from settling FX trades was such that the failure of a single participant
in the FX market could have caused systemic risk to materialise ie it could
have caused the failure of other participants (CPSS (1996)). The strategy to
address the problem involved three tracks: action by individual banks to control
their FX settlement exposures; action by industry groups to provide risk-
reducing services for settling FX trades; and action by central banks to induce
private sector progress on the previous two tracks. 5

2
The collapse of Bankhaus Herstatt was one of the earliest cases where FX settlement risk
crystallised. The bank, a medium-sized bank that was active in the FX markets, was closed by
the German authorities on 26 June 1974. Some of its FX counterparties had already paid
Deutsche marks to the bank but had not yet received the US dollars that they were buying in
exchange. For more about this and other cases where settlement problems have arisen, see
CPSS (1996) and Galati (2002).

3
Thus although the main concern is with outright default by the counterparty (eg because of
insolvency), even technical fails that are corrected on a subsequent day (eg when there are
temporary operational difficulties) have the potential to cause liquidity problems.

4
For estimates of the size of the FX market, see BIS (2007).

5
For more information about the 1996 strategy, see CPSS (1996).

54 BIS Quarterly Review, September 2008

Box 1: How FX settlement risk arises
An example of how settlement risk arises when using traditional correspondent banking

In this example, Bank A has a spot trade with Bank B in which it is selling yen for US dollars. The trade is
executed on Day V2 for settlement on Day V (value day).
After the trade has been struck, Bank A sends an instruction to its correspondent in Japan
(Bank Ja), asking the latter to send the yen to Bank Bs correspondent there (Bank Jb) on Day V.
Bank Ja executes this instruction sometime during Day V by debiting the account that Bank A holds
with it and sending the yen to Bank Jb via the relevant payment system. After Bank Jb has received
the funds, it credits them to Bank Bs account and informs Bank B that they have arrived.
In parallel, Bank B settles its side of the trade by a similar process in which it instructs its
correspondent in the United States (Bank Ub) to send US dollars to Bank As correspondent there.
Settlement risk arises because each counterparty may pay the currency it is selling but not
receive the currency it is buying. The underlying cause is the lack of any link between the two
payment processes (in yen and dollars) to ensure that one payment takes place only if the other also
does.
o Looking at the trade from Bank As point of view, its exposure to settlement risk starts when
it can no longer be certain that it can cancel its instruction to pay Bank B. This depends
primarily on any agreement between Banks A and Ja about cancellation. In the absence of a
specific agreement, Bank A cannot be certain whether it can cancel or not and so its
exposure begins immediately it has sent the payment instruction to Bank Ja, which is likely to
be on Day V1 or even V2. Even if there is a specific agreement, Bank Ja may need some
time to process a cancellation request by Bank A, so the exposure may start at least several
hours before the yen payment system opens on Day V. The effective cancellation deadline
may therefore be very early on V or even on V1 in Japanese local time, which, if Bank A is
located in (say) Europe, will be even earlier in Bank As local time because of time zone
differences.
o Bank As exposure ends when Bank Ua credits its account with the dollars received from
Bank Ub. Bank Ua may not receive the funds until just before the close of the relevant
payment system, and it may be some time after that that the funds are credited to Bank As
account. This could be relatively late on Day V in US local time, and even later on Day V or
even on Day V+1 in the local time of Bank A. Bank As actual exposure to this trade could
therefore last more than 24 hours.
Bank B also faces settlement risk. Its exposure period will differ from that of Bank A to the
extent that Banks B, Ub and Jb have different arrangements compared to those of Banks A, Ja and
Ua, and the relevant US and Japanese payment systems have different opening hours. Time zone
differences are also important. In this trade, time zones work against Bank A because it is selling a
currency that settles in an early time zone (so it is committed to selling its currency relatively early)
and buying one that settles in a late time zone (so it will receive the currency it is buying relatively
late), which extends the duration of its exposure. Conversely, the time zone difference works in Bank
Bs favour. However, it is important to note that the problem does not arise solely because of time
zone differences.

BIS Quarterly Review, September 2008 55

Settlement methods
The 2006 survey found that much progress has been made since 1996, Progress has been
made
particularly on the provision of risk-reducing services by industry groups, the
second track of the strategy. Most significant was CLS Bank (CLS), 6 which
started operating in 2002. CLS provides a PVP service that almost completely
eliminates the principal risk associated with settling FX trades. (Box 2 provides
a simple example of how CLS works. For more detail, see CPSS (2008) and
Galati (2002).) Although there are seasonal fluctuations, use of CLS has grown
steadily (Graph 1) and the service is now a well established and critical part of
the global financial infrastructure.
Indeed, the 2006 survey showed that CLS has become the primary
settlement method for FX trades, with 55% of trades being settled this way
(Graph 2). A further 8% was settled by bilateral netting, where two market
participants agree that the settlement obligations resulting from all the trades
between them due to settle on a given day will be netted against each other so
that only the smaller netted amount in each currency needs to be
settled. 7 Various other methods accounted for another 5%. However, the key but risk remains
survey finding was that 32% of trades were still settled by traditional
correspondent banking the major source of FX settlement risk.
This compares to a previous survey in 1997, before CLS was available,
when 85% of FX trades were settled by traditional correspondent banking with
the remainder settled by other methods including netting. However, although it

Growth of CLS
Daily value settled, in billions of US dollars (15-day moving average)1

5,000

4,000

3,000

2,000

1,000

0
Jan 03 Jul 03 Jan 04 Jul 04 Jan 05 Jul 05 Jan 06 Jul 06 Jan 07 Jul 07 Jan 08 Jul 08
1
The value settled is a multiple of the value traded, depending on the number of currency legs that trades
have ie a spot or forward deal has two currency legs, one for each currency, while a swap has four, two
for the spot trade and two for the forward trade. The vertical line indicates when the survey took place.

Graph 1

6
The name CLS Bank is derived from Continuous Linked Settlement, the brand name of the
service provided.

7
The 8% refers to the size of the reduction achieved. The smaller netted amount will then be
settled by another method, typically traditional correspondent banking. (In the survey results,
the 32% share of traditional correspondent banking includes any netted amounts settled this
way.)

56 BIS Quarterly Review, September 2008

Box 2: How CLS works a simplified example

CLS Bank (CLS) is a limited purpose bank for settling FX, based in New York with its main operations in
London. It is owned by 69 financial institutions which are significant players in the FX market. It currently
settles trades in 17 currencies, three in North America (Canadian dollar, Mexican peso and US dollar), two
in Africa and the Middle East (Israeli shekel and South African rand), six in Europe (Danish krone, euro,
Norwegian krone, Swedish krona, Swiss franc and pound sterling) and six in the Asia-Pacific region
(Australian dollar, Hong Kong dollar, Japanese yen, Korean won, New Zealand dollar and Singapore
dollar).
The simple example below, which uses the same yen/US dollar trade as in the previous box, is
designed to show the essence of the CLS mechanism in the case of a single trade. In reality, CLS
settles a large number of trades between multiple counterparties and has complex risk control
mechanisms to enable it to do this safely.

CLS removes principal risk by using PVP you get paid only if you pay. On settlement day,
each counterparty to the trade pays to CLS the currency it is selling eg by using a correspondent
bank, as with the example in the previous box. However, unlike the previous example, CLS pays out
the bought currency only if the sold currency is received. In effect, CLS acts as a trusted third party
in the settlement process. (However, note that CLS is not a central counterparty in the example
shown, the trade remains between Banks A and B.)

CLS could have been designed so that, if one of the counterparties fails, CLS simply returns
the principal amount to the surviving counterparty in the example, it could return the US dollars to
Bank B. However, in practice CLS has committed standby lines of credit with major banks in each of
the currencies it settles. In this case, Bank B was buying yen, so CLS will swap the US dollars for
yen with its yen liquidity provider in Tokyo, and then give the yen to Bank B. In this way, CLS not
only removes principal risk but also reduces liquidity risk. However, the standby liquidity facilities
cannot completely remove liquidity risk. The main underlying reason for this is that the liquidity
facilities are finite while there is no limit on the total value of the trades that you can attempt to settle
via CLS.

BIS Quarterly Review, September 2008 57

 Breakdown of total settlement obligations according to settlement
method
In per cent

2006 survey 1997 survey

5
8 15

32 55

85

CLS Netting
Other Traditional correspondent banking
Graph 2

is a big reduction from 85%, 32% remains a significant share. Moreover, the
values involved are also significant relative to the size of the institutions
concerned on average, equivalent to approximately 70% of their total capital.

Assessing the remaining exposures

Given that traditional correspondent banking remains a significant method of
settling FX trades, the key issue is whether the resulting exposures pose an
unacceptable degree of risk. To assess this, the survey asked about the
duration and size of survey institutions total exposures (ie to all their
counterparties) and largest bilateral exposures (ie to a single counterparty). It
also asked how these exposures were managed.

Total exposures

The survey showed that the duration and size of total FX settlement exposures Total exposures can
be large and long-
can still be significant (Graph 3). Given that, as noted above, FX settlement lasting
risk is the risk of paying without being paid, an institutions exposure starts
when it becomes irrevocably committed to paying away one of the currencies it
is selling. As the graph shows, on average this is at about 06:00 on the day
before settlement. As it becomes committed to paying more currencies, its
exposure increases. Then at some point, the institution will start to receive the
currencies it is buying, causing its exposure to decrease. For a period, its
overall exposure may fluctuate as it becomes committed to paying some
currencies and receives others. On average, the peak exposure (X) is reached
at around 16:00 on settlement day, and the exposure ends when the last
currency is received, on average at around 08:00 on the day after settlement. 8

8
Box 1 explains this process in more detail.

58 BIS Quarterly Review, September 2008

 Exposure profile of an average survey institution (single days
trades)
Shown as a percentage of trades settled by traditional correspondent banking on Day V

V1 V V+1
75
X
60

45

30

15

0
24h 18h 12h 6h 0h 6h 12h 18h 24h 30h 36h 42h 48h
X = time of peak exposure. Graph 3

On average, therefore, an institutions exposure to trades due to settle on

a particular day actually starts on the day before settlement and continues until
the day after settlement ie the duration is more than 24 hours. This means
that an institution using traditional correspondent banking to settle its trades
including typically always has some FX settlement exposure, overnight as well as
overnight
intraday. In addition, it means that, for at least part of the day, an institution is
exposed to more than one days trades. Graph 4 shows average exposure
during the day allowing for this simultaneous exposure to trades settling on
multiple days. 9
During the period the exposure lasts, the size of an institutions total
exposure to all its counterparties varies, as the graphs show, but, on average,
peaks at an amount equal to about 70% of the value settled by traditional
correspondent banking allowing for one days trades (ie point X on Graph 3) or
at about 80% allowing for simultaneous exposure to multiple days trades
(point Y on Graph 4). 10 Moreover, on the latter, multiple day basis, the
exposure is never less than about 50% even during the night.
Translating these percentages into values for the survey participants
overall, the aggregate amount at risk never falls below $0.5 trillion and peaks
at about $1.1 trillion. 11

9
Note that the survey results were daily averages for the survey period. Graph 3 thus shows
the exposure profile for the trades settling on one average day (Day V in the graph), while
Graph 4 is created by superimposing that exposure profile with identical profiles for trades
due to settle on earlier and later average days. In reality, an institutions profile for each day
would vary according to the value and type of trades due to settle that day.

10
The maximum exposures are less than 100% of the value settled primarily because of time
zone differences, which mean that (a) some currency pairs generate no exposure (the bought
currency is received in an eastern time zone before the sold currency is irrevocably paid away
in a western time zone) and (b) the exposure period generated by one currency pair does not
always overlap with that of another currency pair (the exposure period for a trade in two
eastern currencies may not overlap with that for a trade in two western ones).

11
The size of the range of the average institutions position in percentage terms (ie 50 to 80%)
is different from the range of all survey institutions aggregate value (ie $0.5 trillion to
$1.1 trillion) because the exposure profile of the average institution is expressed in its local

BIS Quarterly Review, September 2008 59

 Exposure profile of an average survey institution (multiple days
trades)
Shown as a percentage of the average daily value of the trades settled by traditional
correspondent banking

Y 80

60

40

20

0
0h 6h 12h 18h 24h
Y = time of peak exposure. Graph 4

An alternative way to judge the size of the total exposures is to scale them
by the institutions capital, rather than by the value of the settled transactions
themselves. By this measure, an institutions total exposure peaks at 47% and
57% of its total capital on average (single day and multiple day, respectively).
In other words, if such exposures were to be shown on an institutions balance
sheet (which in practice they are not), they would be a significant item.
Institutions exposures to FX settlement risk vary for many reasons. For a
given institution, exposure can vary substantially from day to day depending on
the value and currency composition of the trades. And comparing institutions,
the internal procedures of each institution and its correspondents also have a
significant effect, particularly on the time at which an institutions settlement
exposure in a currency starts. 12 Not surprisingly, therefore, there was very
wide variation about the averages just mentioned, with some institutions having
negligible exposures while others had exposures as large as six times the size
of their capital.

Bilateral exposures

As noted above, FX settlement risk arises because of the possibility that an

individual counterparty will fail to pay. Thus although an institutions aggregate
exposure to all its counterparties (its total exposure) is interesting in order to
get an idea of the overall scale of the potential problem, more relevant from the
point of view of assessing risk are an institutions settlement exposures to its
individual counterparties (its bilateral exposures).
Unfortunately, the survey data do not include direct information about the
size of bilateral exposures. Nor was it possible to come up with robust point

time, which has to be translated into a standardised time (eg GMT) when aggregating across
institutions.

12
That is, there is variation in the cancellation deadlines, the point at which the institution can
no longer cancel the instruction to pay the currency it is selling (Box 1). If an institution and its
correspondent bank improve their procedures, they may be able to move back the time at
which the exposure starts.

60 BIS Quarterly Review, September 2008

 estimates of what those exposures might be. However, in most cases it was
possible to produce a robust estimate of the range within which an institutions
largest bilateral exposure was likely to lie. 13 The results are shown in Graph 5
for the 81 institutions in the survey for which sufficient data were available.
Thus, for example, the largest bilateral exposure of Institution 1, on the left
of the graph, is estimated to have been, on average, somewhere between
about 70 and 190% of its capital. However, that is an extreme case. For most
institutions the range was much lower for a majority it was entirely under the
Exposures to single 10% level. Nevertheless, making some additional assumptions about where
counterparties can
within the possible range the actual exposure was most likely to be, more than
also be significant
one in four of the institutions may have had an exposure to a single
counterparty greater than 5% of capital, with one in eight being over 10%. 14
Moreover, these are estimates for an average day; on a peak day, the
exposures may have been substantially higher. And in order to get a complete
picture of an institution's counterparty exposure, this FX settlement exposure
needs to be added to other types of exposure it has to the same counterparty
(eg as a result of interbank lending). Given that it would normally be regarded
as prudent for a bank to keep its exposure to a single counterparty to no more
than a rather small percentage of its capital, the estimates suggest that many
institutions continue to have significant bilateral FX settlement exposures which
they need to control prudently.

Largest bilateral exposure (daily average range)1

Range of largest bilateral exposure, as a percentage of total capital

All appropriate control criteria met 175

Some appropriate control criteria not met
150

125

100

75

50

25

0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
1
Exposure to multiple days trades. Institutions ranked by upper end of range. Graph 5

13
The survey had data on the aggregate value of an institutions settlement obligations to its
five largest and 10 largest trading counterparties and on the breakdown of this value between
the various settlement methods. Taking the portion of this aggregate value that was settled by
traditional correspondent banking, the ranges were based on estimates of how much or how
little of the portion could be accounted for by a single counterparty. More information about
the method used to calculate the ranges is given in Annex 3 of CPSS (2008).

14
These calculations used additional data provided by CLS about the relative sizes of
institutions five largest counterparties, where largest was judged by trades settled using
CLS, and assumed that the same relative sizes applied to trades settled using traditional
correspondent banking.

BIS Quarterly Review, September 2008 61

Control of exposures

However, judged according to three specific criteria, there was a mixed picture
about whether the exposures were in practice controlled appropriately. The
three criteria were whether the institution (1) had established clear senior-level
responsibility for managing the exposures, (2) had appropriate daily
management procedures (including the use of the same counterparty limits as
were applied to other types of similar exposures) and (3) measured the risk in a
way that did not lead to underestimation. 15 Although most institutions in the
survey met the first two criteria ie they had established clear senior-level
responsibility and many had appropriate daily management procedures there
was still a significant minority (8% and 23%, respectively) that did not.
Moreover, most (73%) surveyed institutions failed to meet the third criterion
ie they measured their exposures in a way that at least to some extent
underestimated the amounts they had at risk. 16 Indeed, judged overall by
these criteria, 66% of the surveyed institutions did not appropriately control Also the exposures
are not always
their FX settlement exposures ie only 34% met all three criteria. And as
appropriately
Graph 5 shows, among the institutions with the highest bilateral exposures, the controlled
percentage is even lower. For example, of the 10 institutions with the highest
exposure, only one was judged to control its exposures appropriately.

Evaluation of the risk

Overall, the survey shows that the situation of individual institutions varies
considerably. There are some institutions both large and small that use
PVP services such as CLS as much as they can given the limitations that exist
(these limitations are that some trades, including trades in non-CLS currencies
and many same day trades, are ineligible for CLS settlement and that CLS
cannot be used to settle trades with counterparties that are not themselves
CLS users). Some of these institutions also appropriately control any
exposures that result from the remaining trades that are settled using
traditional correspondent banking ie they meet the three criteria discussed
above. They therefore do all that they can to reduce risk. However, at the other
end of the range are institutions that make little or no use of PVP settlement
and have significant exposures that are not always well controlled.
The lack of appropriate control is clearly an issue. Financial institutions
naturally take many types of risks and this is generally acceptable as long as
those risks are well managed ie understood, properly measured and subject
to appropriate controls, such as counterparty limits. From this perspective, the One view is that the
risk is acceptable
problem is the lack of appropriate management rather than the size of the as long as it is well
exposures themselves. There is therefore a choice of solutions. One is for such managed

15
The three criteria were formulated as objectives. The means by which the objectives were met
were not assessed.

16
Most institutions did not attempt to measure their exposure exactly (as in Graphs 4 and 5) but
instead used an approximation method. For example, a common method would be to assume
that the exposure existed only on the settlement day. For institutions whose exposures could
last for more than one day, this could lead to underestimation of the true position.

62 BIS Quarterly Review, September 2008

 institutions to use PVP services such as CLS so that the exposures are
avoided. But it is also acceptable for them to continue to use traditional
correspondent banking and incur the exposures provided they manage those
exposures in an appropriate way.
However, from a different perspective, FX settlement exposures can be
seen as intrinsically undesirable, even when they are well managed, because
of their possible effects during financial crises. If there is increased market
uncertainty about the financial strength of a counterparty, for example
institutions may prudently decide to reduce their trading limits to that
counterparty in order to reduce settlement risk. And, in doing so, they may
deprive the counterparty of the market access it needs and thus inadvertently
Another view is that cause it to fail. In contrast, if it was possible to make settlement risk-free,
settlement should
institutions could prudently continue to trade, even in uncertain circumstances.
be risk-free
In economic terms, the argument is that the private costs to market participants
of removing the risks are outweighed by the social benefits of risk-free
settlement.
It is true that, in practice, settlement of any transaction including FX
trades is rarely, if ever, completely risk-free. This is because even though
principal risk can usually be removed by good system design, some liquidity
risk typically remains, as is the case with CLS (as noted in Box 2, the reason
for this is that, even with the principal amount of the trade being protected in
the event of a counterparty failure, CLS cannot fully guarantee that you will
receive that amount in the currency you were trying to buy). So the ideal state
of risk-free settlement can never be fully achieved. Nevertheless, from this
perspective, the risk should be reduced as far as possible. Accordingly, the
survey results are of more concern because even well managed FX settlement
exposures are not ideal and it would be better if PVP services such as CLS
were always used.

Solutions
Whichever perspective of settlement risk is held, there seem to be two main
weaknesses with the current situation which need to be addressed.
The first is that the existing risk-reducing services for settling FX trades
are not sufficiently comprehensive. The survey showed that over a third of the
trades subject to settlement risk were between CLS users but involved types of
trades that they currently cannot settle using CLS. As noted above, such trades
include same day trades (where the difficulty is that the CLS settlement
There need to be process takes place too early in the day) and trades in non-CLS currencies. To
new services
reduce settlement risk on these trades, either the CLS service needs to be
modified or a new settlement service introduced.
The second and perhaps more important weakness is the lack of
incentives for individual institutions to take action to better manage FX
settlement risk. Discussions with survey participants suggest that many FX
market participants who have not already taken the necessary action are
unlikely to do so unless they are given stronger incentives or compelled to do
so by regulatory authorities. The problem is that taking action costs the

BIS Quarterly Review, September 2008 63

institutions money. But at the same time, it seems that the risk is not well
understood or is perceived as less serious than equivalent counterparty risks
that arise from other activities. Why this should be so is not completely clear
it is perhaps because the exposures are not very transparent. More
fundamentally, even if individual institutions were fully aware of the risk to
themselves, they would not necessarily take into account the social benefits to
the market as a whole of the reduced systemic risk that would result from using and stronger
incentives to act
safe settlement methods. In any event, there is often a reluctance to spend the
necessary money, suggesting that there is a need for incentives or regulatory
inducements, both of which are lacking at the moment.
As far as use of CLS is concerned, certain market-based incentives that
some had hoped for (such as smaller spreads on FX trades settled through
CLS, recognising the reduced risk involved) have apparently failed to
materialise. And although existing CLS users can point to operational savings
from the standardised and automated procedures for using CLS, these seem to
be outweighed in the minds of many non-CLS users by the size of the fee for
using the CLS service. Incentives for addressing the problem through better
management of the exposures from traditional correspondent banking are
equally lacking. Given this, it is not surprising that many institutions felt that
further improvements to the management of FX settlement risk are unlikely
unless there is a clear regulatory requirement for them. Particularly important including
regulatory
here is action by the banking supervisors. In 2000, the Basel Committee on
requirements
Banking Supervision (BCBS) issued guidance on managing FX settlement risk.
The BCBS and CPSS have recently agreed to work together to review and
update the guidance with the aim of setting a higher standard for how banks
manage FX settlement risk.
When publishing the survey results, the CPSS recommended a series of
actions to bring about further progress in addressing FX settlement risk
(CPSS (2008)). Given the analysis above, two of these actions seem
particularly important. One is that CLS or other industry groups should continue
to develop services to reduce FX settlement risk, particularly services for same
day trades and trades involving additional currencies. The other is that central
banks should work with banking supervisors and other regulators to explore
ways to encourage market participants to manage their settlement risks better.
For example, regulators could require FX settlement risk to be managed and
controlled in the same way as other formal short-term credit extensions of
similar size and duration (eg unsecured overnight interbank loans). 17 Success
in implementing these two actions will be key to determining whether the
potential threat of FX settlement risk to the stability of the global financial
system can finally be removed.

17
Another possibility that is sometimes proposed is to put a capital charge on the exposures.

64 BIS Quarterly Review, September 2008

References
Bank for International Settlements (2007): Triennial central bank survey of
foreign exchange and derivatives market activity in 2007.

Basel Committee on Banking Supervision (2000): Supervisory guidance for

managing settlement risk in foreign exchange transactions.

Committee on Payment and Settlement Systems (1996): Settlement risk in

foreign exchange transactions.

(2008): Progress in reducing foreign exchange settlement risk.

Galati, G (2002): Settlement risk in foreign exchange markets and CLS Bank,
BIS Quarterly Review, December, pp 559.

BIS Quarterly Review, September 2008 65