Settlement risk within the forex market is beginning to rise once again. According to BIS, the reason is
attributable to the rise in Non CLS settled currency trading as a percentage of the entire forex market.
Indeed the size of the non settled forex market has increased from 50% to 60% of the Market since 2013; the
total value of that 60% is estimated by BIS to be $8.9Tn. Much of this growth is due to the rise of the
Chinese Renminbi, and to a lesser extent the Russian Rouble. The risk of settlement failure is therefore
growing with the rise in these and other minor, non CLS, currencies and can be assumed, will continue to
grow for the foreseeable future.
One way in which settlement, or Herstatt, risk could be removed is using real time settlement. This is a
capability being introduced for cash settlement within some countries, such as the UK with Open Banking, but
that has not spread to many of the CLS settling currencies, and certainly not for Forex payments.
Indeed, the growth of FX transaction volumes does not look like it is slowing and the demand for real-time
settlement is not disappearing anytime soon either. In the near future, we may see some sporadic attempts,
but the FX market in general is still some way off. The payment technology is driving forward but the market
does not seem ready to adopt it until the path seems clearer and more joined up rather than fragmented by a
range of alternative solutions. Perhaps it will take the group of global FX banks to move in a unified
direction to sway the market. The alternatives being developed and promoted are growing, leading to a wider
spread of use and less cohesion among the market player
We have SWIFT, CLS and ISO 20022 from the established market infrastructure all making advances, but we also
have new entrants such as Google, Apple, Amazon and PayPal all offering a new model of clearing currencies
for smaller amounts and individuals.
We are seeing DLT and the Central Banks also looking at new models for settlement. The utilisation of tokens
could be a new way for CLS to settle, it may also be something to be used by the likes of Visa and
MasterCard, thus becoming an even bigger disruption to the market. Indeed, the use of Tokens could be the
way to mitigate, although not eliminate, Herstatt risk more than ever and could reduce the demand and need
for real time payment v payment, certainly from a settlement risk point of view.
These alternatives for the future will not only require the market to move in a coherent way, but in order to
get ahead of these challenges, market participants need to start reviewing their internal operating models
and controls, embracing and leveraging new technologies; collaborating with others to ensure a consistent
Ensuring the technologies support a process that can adapt and adopt altering market models will be a key to
success; being able to control the activity and position a necessity.
One thing is certain; we are in for an interesting period for FX settlements.
Assessing the extent and nature of the FX settlement risk problem
The Bank for International Settlements’ (BIS) Triennial Central Bank Survey is a key bellwether of
the state of the global foreign exchange market. We explore the findings of the latest survey with
Takeshi Shirakami, Deputy Head of Secretariat of the BIS-hosted Committee on Payments and Market
Infrastructures (CPMI), to examine what can be done about the increasing FX settlement risk, including
the role of industry supervisors in tackling it.
What does the latest Triennial survey data tell us about FX market activity and how big a threat increased
FX settlement risk is becoming?
The results of the 2019 BIS Triennial survey tell us that the FX market continues to grow, as does FX
The survey data tell us that the share of total FX settlement obligations subject to settlement risk grew
from 50% to 60% between 2013 and 20191. In April 2019, average daily global FX trading of $6.6
trillion translated into gross payment obligations worth $18.7 trillion, after taking into account the
number of payments for each instrument. Bilateral netting reduced payment obligations to $15.2 trillion.
About $6.3 trillion was settled on a payment versus payment (PvP) basis using CLSSettlement or a similar
settlement system. This left an estimated daily $8.9 trillion worth of FX payments at risk. FX settlement
risk remains significant. Moreover, once materialised, it could have a significant second-round effect on
the global financial system. Losses could be large since FX contracts involve the exchange of principal
amounts of currencies traded. Such incidents would likely lead to overly precautionary behaviour among FX
market participants withholding outgoing payments before receiving incoming ones where PvP settlement is not
available, which could create a severe stress scenario for broader FX markets. This is exactly what happened
in the summer of 1974 when the Herstatt Bank collapsed.
Generally speaking why are FX settlement risks on the rise?
There are several possible explanations for this in my view. First, changes in the FX market structure might
have contributed to the higher share of payments without PvP protection. For example, some types of
transactions may require additional operational steps and arrangements to be settled on a PvP basis, which
may discourage the use of PvP. Second, some FX market participants may not have access to PvP services or
may not have enough incentives to use the service offered. Third, trading in currencies that are not
eligible for PvP systems has grown. The CPMI considers that digging into these potential explanations is
important but would require more granular information on the structure of FX markets.
There has been a contraction in the number of correspondent banking channels. What impact is this having on
FX settlement risks?
The contraction in the number of correspondent banking channels and FX settlement risks are separate issues.
One thing we do know is that correspondent banking and FX markets are highly concentrated and the key
players are more or less the same. The contraction in the number of correspondent banking channels could,
although does not have to, lead to a rising share of FX payments via the same settlement bank, i.e. “on-us
settlements.”2 With on-us settlements, both legs of settlement may or may not be final and
irrevocable, i.e. may not have PvP protection. Here again, it would be useful to get more granular
information about how on-us settlement operates in practice, how robust it is and how the decline in the
number of correspondent banking channels contributes to concentration in settlement bank services. Apart
from its potential impact on FX settlement risk, the decline in correspondent banking channels is itself an
issue of significant concern. Addressing this issue is one of the key focus points in the on-going G20 work
for improving cross-border payments. As part of this broader G20 work, the CPMI published a report in July
this year, which set out 19 “building blocks” of a global roadmap for enhancing cross-border
payments.3 Some of these aim to address the issues related to correspondent banking such as
consistent and comprehensive application of the rules for anti-money laundering/combating the financing of
terrorism (AML/CFT) and promotion of safer payment corridors.
The on-going work on cross-border payments is quite relevant also to reducing FX settlement risk. While its
scope is broader and the project is multifaceted, it includes a building block on facilitating increased
adoption of PvP to improve cross-border payments (Building Block 9). Going forward, the CPMI, together with
the Basel Committee on Banking Supervision (BCBS) and relevant authorities, will be encouraging the adoption
the 2013 BCBS Supervisory Guidance on managing FX settlement risk,4 and we will be encouraging FX
committees to support the Global FX Code principles. The CPMI will also take stock of PvP arrangements,
analyse drivers for non-PvP settlement and develop proposals for increased adoption of PvP.5
Shadow payments systems which are outside the control of lawmakers and regulators are also on the rise. How
much are they contributing to the problem?
Often shadow payment arrangements settle transactions by moving money between customers’ accounts on
their books. Consequently, users are just as exposed to the liabilities of the entity that is not
regulated as a deposit taking institution both before and after settlement as they are during the settlement
In my view, the concern with shadow payment systems is much broader than just FX settlement risk. They often
lack basic risk management, legal certainty, consumer protection, financial integrity or cyber security.
This highlights the urgency of improving cross-border payments so that market participants do not turn to
shadow payments out of necessity.
Can the challenge to get countries to make their payments systems interoperable be overcome and if so what
impact would that have on reducing risks?
The harmonisation of clearing and settlement procedures and messaging standards do not necessarily affect FX
settlement risk. However, they do have an impact on the speed, efficiency and cost of cross-border payments.
That is why enhancing data and market practices, including the adoption of a harmonised ISO 20022 version
for message formats, is one of the focus areas of the G20’s work for enhancing cross-border payments.
Legal issues are involved in connecting payment systems in different jurisdictions. How much of a difficulty
and obstruction to providing solutions to the settlement risk problem do these represent?
PvP settlement systems often involve connecting payment systems in different jurisdictions, which gives rise
to certain legal issues. For instance, differences in settlement finality rules may lead to a scenario where
a payment is regarded as final in one jurisdiction but not in the other. Moreover, it is more likely that
cross-border systems face issues arising from a conflict of laws, e.g. ambiguity as to which jurisdiction’s
laws apply. Closing such legal gaps sometimes requires new legislation or even treaties, and consequently
will require a considerable amount of time and resources to address.6
How much could better regional monitoring of the FX market help to detect and deal with FX settlement
More frequent data collection on FX settlement risk would help us develop a deeper and timelier understanding
of the market practices and associated risks as the FX market structure and back office arrangements
continue to evolve. In this regard, we welcome on-going dialogue among members of the Global FX Committee
(GFXC7) and its regional FX committees to enhance their data collection. The regional FX
committees collect data on their respective FX markets more frequently than the BIS Triennial survey, so I
think it would be helpful if these regional surveys included questions on FX settlement risk.
Are some regions of the world likely to see the threat of FX settlement risks increasing more than others
over the next few years?
One of the reasons for the relative increase in FX settlement risk is the growth in trading in currencies
that are not eligible for settlement in a PvP system. Increased activity in emerging market (EM) currencies
could lead to an increase in FX settlement risk unless the availability of PvP is widened to include these
currencies. However, just because the risk is related to EM currencies does not mean that EMs bear the risk;
it is the counterparties to a trade, wherever they are based, that bear the risk and must manage it.
That said, I would argue that once the threat of FX settlement risk has materialised in a stressed market
scenario, jurisdictions whose currencies are not eligible for PvP settlement may be more susceptible to
precautionary payments freeze and broader FX market seizure.
How can local and regional banks be encouraged to do more to combat FX settlement risks, for example by
trying to make their FX liquidity management a less complex and more joined up process?
An individual bank can evaluate its FX settlement policies and procedures against the 2013 BCBS Supervisory
Guidance on managing FX settlement risk. Such action by individual banks should be complemented by
collective industry-wide dialogue and cooperation, such as the work of the GFXC.
In recent years, an increasing number of central bank wholesale payment systems, i.e. real-time gross
settlement systems (RTGS) have extended their operating hours, or are considering doing so. Greater overlaps
in operating hours between central bank payments systems from different time zones help to reduce, though
not eliminate, the duration of payments exposures between the final payment of one leg and receipt of the
other leg. Extending operating hours requires close coordination within and across local jurisdictions.
What issues and considerations have the fast growing digital asset and cryptocurrency markets thrown up with
respect to FX settlement risk?
To the extent that these types of payments are unregulated, i.e., shadow payment systems, the concern is
broader than the risks associated with the settlement process. Highly volatile crypto-assets such as Bitcoin
would not be well suited as a settlement asset in the first place in my view. I have some doubt about their
legal soundness in ensuring PvP. More generally, the CPMI is closely following private-sector initiatives in
wholesale digital payments also in relation to their potential use in FX settlement.
As the FX market continues to evolve with new products and new participants utilising new technologies, has
the BIS identified any new types of potential FX settlement risk we should all be concerned about?
New FX products do not necessarily affect FX settlement risk, as long as the payment obligations can still be
settled on a PvP basis. However, new participants may not have access to PvP settlement systems or may not
manage their FX settlement risk as rigorously as traditional market participants are required to
Faster and interoperable data exchange with new technologies may help market participants, especially those
that are active globally, to obtain and aggregate information from a large number of correspondent bank
accounts in near real time, and to monitor and control incoming/outgoing payments and liquidity. Such
innovative services will be useful to individual market participants in addressing FX settlement risk.
However, I don’t think they are a complete substitute for PvP mechanisms. More efficient and timely
monitoring and control by individual market participants will help protect them but could inadvertently lead
to payments freezes in a stressed market. A PvP mechanism by contrast could incentivise early payment
submissions, since payers know for sure their payments will not be processed unless incoming payments are
also processed. It can work as a coordinating, and stabilising, device during periods of heightened
The views expressed are those of the interviewee and do not necessarily reflect those of the Bank for
International Settlements, the Committee on Payments and Market Infrastructures or its member
For further information on the challenges of linking payment systems see Bech, M, U Faruqui and T
Shirakami “Payments without borders”, BIS Quarterly Review, March 2020, available at: https://www.bis.org/publ/qtrpdf/r_qt2003h.htm.
Members of the GFXC include central bank-sponsored foreign exchange committees and similar
structures in various regions.
The FX Global Code and settlement risk
The FX Global Code is still relatively young. So how well does it address settlement risk as the FX
market grows and changes? To find out we spoke to Guy Debelle, Deputy Governor of the Reserve Bank of
Australia, Chair of the Global Foreign Exchange Committee and the Code’s author.
The GFXC’s June meeting spent some time discussing settlement risk. Although reiterating the
importance of keeping settlement risk to a minimum would it be fair to say that the Committee seemed unclear
as to its scope and extent?
That’s what we’re trying to establish. We’re working with CLS, which is analysing
settlement data provided by its members and by means of our six-monthly FX surveys we’re aiming to get
a better, more up-to-date picture of settlement risk in the global market. As far as the G10 currencies are
concerned, what we do know is that internalisation of settlement by the major banks goes some way to
explaining the extent of trades not settled through CLSSettlement. The GFXC and the Bank for International
Settlements’ Committee on Payments and Market Infrastructures (CPMI) are trying to get a handle on
this. Among the emerging market currencies, turnover is increasing but some, such as China for example, have
their own onshore settlement systems. So we still have some research to do on the quality the scale of
global FX settlement risk.
Principles 35 and 50 of the FX Global Code make clear respectively that, “Market Participants should
take prudent measures to manage and reduce their Settlement Risks” and that they “should measure
and monitor their Settlement Risk and seek to mitigate that risk when possible.” Shouldn’t this
be sufficient to ensure settlement risk is properly recognised and addressed?
I think it’s important that we remind people every once in a while, to pay attention to it because,
particularly in the front office, it can be a bit “out of sight, out of mind.” Obviously it can
be more front of mind for the back office.
Institutions who have signed up to the Code, should be and I think are, assessing how it should apply to all
aspects of it of their business, which would include the settlement risk principles. So, as I say, it’s
just reminding people that settlement is an important part of the process because if there are issues with
the settlement leg then that can be pretty problematic.
Some have suggested the Code needs to be strengthened as far as settlement risk is concerned, would you
agree with them?
We are going to come up, at our next meeting, with some strengthening of some of that language. We’ve
now discussed settlement risk at two GFXC meetings and also at various local foreign exchange committee
meetings. For example, I was on a New York call a few weeks ago and it occupied a good deal of the
discussion. Giving it that sort of visibility, has, I think, been helpful. I don’t think the sort of
thing we’re going to propose is going to be that contentious, it’s just reminding people to pay
attention to it.
Could you envisage a time when all currencies would be settled through CLS on payment versus payment
The GFXC is solution agnostic, the solution doesn’t have to be CLS. That said, CLS is there, and
handles a decently large chunk of the settlement flow. But if others have got technology solutions that work
even for a small part of the market, then great, we encourage that. Any solution that reduces settlement
risk is definitely to be encouraged.
Exploring the remit, Toolsets and future role of CLS in reducing FX settlement
CLS already settles a vast volume of FX transactions. We spoke to its CEO, Marc Bayle de Jessé
about its role and how it can help to address the rising tide of settlement risk that falls outside its
CLS’s settlement service, CLSSettlement, went live in September 2002 in the wake of a number of banking
crises that exposed FX settlement risk as a significant concern. These included the collapse of Germany’s
Bankhaus Herstatt in 1974, Drexel Burnham Lambert’s failure in 1990, BCCI’s a year later and
Barings in 1995. Today, CLS is owned by its members, which include over 70 of the world’s largest
financial institutions. Its wider ecosystem extends to over 25,000 third-party participants including banks,
funds, non-bank financial institutions and multinational corporations. CLS settles 18 currencies on a
payment-versus-payment (PvP) basis, totalling nearly USD6.0 trillion of payment instructions per day, yet
the funding required to settle this amount is determined on a multilaterally netted basis, reducing the
amount of liquidity required for settlement by approximately 96 percent.
Promoting PvP adoption
Covid-19 was the latest challenge for CLS, as for the rest of the global FX market, and as Bayle de Jessé
explains, “The emergence of Covid-19 resulted in extreme market volatility and increased levels of FX
trading. This reinforced the importance of resilient financial market infrastructures during times of crisis
and the need to mitigate FX settlement risk. Despite record-breaking levels of activity, CLS settled on time
and with no issues or delays. Furthermore, we expect additional volatility at the end of the year with the
results of the US election and Brexit, which is likely to lead to increased FX trading activity.”
Although CLS is settling ever-increasing volumes of FX trades, Bayle de Jessé is very aware that the
proportion of trades in the global FX market settling without PvP protection is increasing, leading to an
overall increase in settlement risk. The ‘Bank for International Settlements (BIS) Triennial Central
Bank Survey’ released in December 2019 highlighted that the proportion of trades settled with PvP
protection has decreased from 50% in 2013 to 40% in 2019. More recently, in recognition of the
importance of PvP settlement, a plan published by the Financial Stability Board, ‘Enhancing
Cross-border Payments: Stage 3 roadmap’, highlighted the need to facilitate increased adoption of PvP
settlement in order to enhance cross-border payments.
Bayle de Jessé suggests some of the reasons behind the rise in settlement risk, “The decline in
PvP protection is explained by two factors. The first is the fact that a significant percentage of trades in
CLS-eligible currencies are settling without PvP protection, and the second is the growth in trading of
currencies not currently eligible for settlement in CLS, which are largely settled without PvP protection.”
CLS believes that now is the time to address increasing FX settlement risk and is working closely with key
regulatory stakeholders and market participants to promote awareness of the rise in FX settlement risk and
to support further PvP adoption.
“Our continued focus is to mitigate FX settlement risk which involves encouraging broader
participation in CLSSettlement across the global FX market. To do this we will continue to work with
regulators, FX committees and market participants to better understand and raise awareness of the
issue and look at ways to address it.”
CLS’s strategy has three short-term objectives: 1) educate key stakeholders about growing FX settlement
risk; 2) strengthen the FX Global Code; and 3) advocate for better, high-quality data collection via the
semi-annual turnover surveys of regional Foreign Exchange Committees (FXCs).
“We are encouraged by the engagement we have had with the market so far,” says Bayle de Jessé.
“We have made real progress in creating awareness, making sure that the whole ecosystem understands
the situation and is ready to explore what can be done to address it.”
“We are also encouraged by the Global Foreign Exchange’s (GFXC) leadership in efforts to enhance
the FX Global Code. We believe that Principles 35 and 50, which relate to settlement risk, should be
strengthened to better promote the use of PvP settlement mechanisms. The GFXC stated it would engage its
member FXCs to consider ways to leverage the semi-annual turnover surveys to obtain a better understanding
of trends in settlement activity. Any new solution that mitigates FX settlement risk will greatly benefit
from efforts to obtain an improved picture of current FX settlement practices.”
Leveraging the BIS Triennial Surveys, CLS has also conducted its own analysis and concluded that the total
volume of CLS-eligible currencies equates to USD5.34 trillion. CLS settles 31 percent of eligible FX
transactions in those currencies. This means that 69 percent of transactions in CLS-eligible currencies are
not settling in CLS.
CLS’s ability to address risks brought about by those trades not settling in CLS, such as adding new
currencies to CLSSettlement, is limited because few remaining currencies can meet CLS’s currency
onboarding standards which (as a systemically important financial market infrastructure) derive from the
Principles for Financial Market Infrastructures (PFMI), other applicable regulations and CLS’s own
standards. PFMI Principle 1, legal basis, and Principle 8, settlement finality, have presented the largest
obstacles to onboarding new currencies to CLS’s settlement system. In 2019, CLS and Banco Central de
Chile announced efforts to onboard the Chilean peso to CLSSettlement. “We are working with Banco
Central de Chile, but it is an extended effort and involves encouraging broader direct participation in CLS
from both local banks and CLS members across the global FX market,” says Bayle de Jessé.
An alternative PvP solution for non-CLS currencies
One approach could be for CLS to offer “alternative solutions” that would provide PvP protection.
CLS is consulting market participants about any currency pairs particularly exposed to settlement risk.
While an alternative PvP solution will take time to develop, Bayle de Jessé is optimistic that with
the help of new processes and technology, CLS will be able to deliver PvP in currencies and currency pairs
it does not currently cover. Meanwhile, CLSNet provides a bilateral netting calculation service that six of
the top ten global banks are already using to enhance post-trade processes. “CLSNet aims to increase
the levels of payment netting calculations for trades not settling in CLSSettlement by introducing
standardization and automation for the entire FX market,” says Bayle de Jessé. “It does
this by calculating the bilateral positions of the participants connected to the service.
“The service enables improvements to intraday liquidity, greater operational efficiency and increased
risk mitigation for non-CLS-settled currencies, many of which are in emerging markets and are growing faster
than those in current CLS jurisdictions. CLSNet may be a starting point for a solution that includes PvP
settlement. Should CLSNet serve as the initial building block, there could be a series of tailored PvP
settlement systems with groups of two or three countries with strong currency and economic ties.”
Bayle de Jessé concedes that although settlement risk is increasing and CLS is keen to do what it can
to address this, the process will take time. “Our community of third-party participants is now well
over 25,000, which demonstrates how FX market participants are becoming increasingly aware of the need to
mitigate settlement risk.”
Third-party growth is a priority for CLS and it has seen nearly 6% growth in third-party settled values in
the last year. He concludes, “We have been raising awareness of FX settlement risk among the
wider FX community – asset managers, regional banks and corporates – ensuring these participants
are aware of the exposures they currently face and how CLS could help to mitigate that
“Our continued focus is to mitigate settlement risk which involves encouraging broader participation in
CLSSettlement across the global FX market. To do this we will continue to work with regulators, FXCs and
market participants to better understand and raise awareness of the issue and look at ways to address it.”
How better data can help to manage FX settlement risk
By Sam Romilly, FX Global Market Management, SWIFT
The growing amount of FX settlement risk is of concern to all in the financial industry and the nature of the
challenge in its size, and complexity, is significant. The industry settles over $18 trillion each
day1 resulting from a variety of FX instruments, each with their own settlement periods, with
over 1,000 currency pair combinations, traded by 1,000s of financial firms across over 100 different
platforms, spread over 200 jurisdictions across different time-zones. This makes clear the need for up to
date, and accurate, data on the FX post trade process. The good news is that information on a
significant percentage of the daily $6.6 trillion2 of value traded can be derived from the FX
confirmations sent over SWIFT. Over 6,000 financial firms send FX confirmations between themselves, and to
the 2,000 corporates and investment managers indirectly connected to SWIFT via the FX post trade platforms.
Data extracted from the million FX confirmations sent each day is now available as a dataset to complement
data already available from CLS.
The following description of the confirmation process illustrates how this dataset can be an invaluable tool
to better understand FX settlement risk. The FX Global Code advises that trades need to be confirmed
whether settled gross, or netted. Each party should also match the confirmation sent to the one received. If
there is a mis-match then one of the parties will cancel their original incorrect confirmation, and send a
replacement confirmation. This process continues until both parties have correct matches. If the
matched confirmations are to be netted, then there should be a procedure to confirm the bilateral net
amounts in each currency, at a predetermined cut-off point. The SWIFT FX dataset adjusts accordingly
for the cancellations, and so effectively contains details of all matched confirmations. As such, an
accurate simulation of calculations for netting, and counterparty settlement exposure, is now possible.
Hypothetically, this dataset, if it were to operate in real time, could be the basis of a solution to
monitor counterparty exposure values, and even calculate bilateral netting amounts at the relevant cut-off
This dataset is only available to SWIFT users for their own data, and the market level data provided is under
strict controls to guarantee confidentiality and anonymity. It contains details of confirmations on a trade
and value date basis, and can be provided on a daily basis. Data extracted from these confirmations
is unique, consistent and unbiased. The granularity of the SWIFT FX dataset enables analysis at the level of
currency pair, instrument type, region, business segment of the trading parties, and by value, and number of
transactions, each day. A SWIFT user with access to this data can easily find out their daily counterparty
exposure across all their branches, and across all the branches of their counterparty. They can also see
potential reduction in risk if the counterparty were to become a CLS third party, and can use it to help
select the counterparties, and currency pairs, they wish to bi-laterally net.
The Figure 1 chart gives some high level views of the FX currencies confirmed over SWIFT. We estimate that
around 90% of the FX confirmations sent over SWIFT settle outside of CLSSettlement, as trades that settle in
CLSSettlement are usually not confirmed over SWIFT. It may be a surprise then to see that 79% of the
confirmations sent over SWIFT are to confirm trades for CLS currency pairs. However, there are several quite
valid reasons why such trades do not settle in CLSSettlement. For instance, one of the parties to the trade
may not be a CLS member, or a CLS third party. Another reason is that the FX instrument could be a
Non-Delivery Forward or an Option. Finally, if the FX trade is a give-up between a hedge fund and prime
broker, or if it requires a same day settlement, then it is unlikely to be instructed to CLSSettlement.
NETTING DATA RESULTS
Where the data becomes even more interesting and revealing is in the insights it can bring around bi-lateral
netting. The use of netting to reduce settlement risk is an acknowledged best practise, and CLS have a
centralised netting service - CLSNet - open to any financial firm via their preferred SWIFT connectivity
channel. BIS estimate that use of bi-lateral netting today already achieves up to 20% reduction of
The following overview illustrates how the SWIFT FX dataset can help any bank to have a better understanding
of the underlying components of their FX settlement risk, and to quantify the benefits possible from
increasing their use of netting, and of CLSSettlement.
We first created a settlement risk dataset based on (i) FX confirmations sent for all currency pairs during a
24 hour period in August 2020, (ii) for the instrument types spot and swap (the opening leg only), and (iii)
excluding confirmations sent between branches of the same firm.
For illustration purposes we calculated3 the total gross of the Buys and Sells per currency pair
to give the gross settlement value of all the confirmations in this dataset. It is possible to now
derive the actual maximum amount of FX settlement exposure per counterparty of a typical top tier bank.
It is also possible to calculate the amount of settlement risk removed if a trading counterparty were
to become a CLS third party. We calculated the total net across all branches, currencies and
counterparties, which showed it was possible to achieve a 60% reduction in the gross settlement amount. In
terms of the BIS report totals this means that a theoretical reduction from $10 trillion to $4 trillion
settlement risk could be achieved with bi-lateral netting. Obviously, this is just a theoretical
maximum but, when looking at the data across top tier banks at a branch level, we still see impressive gains
possible from netting. However, the dataset also shows that a branch of a top tier bank can have as
many as 500 counterparties, dealing in over 300 currency pairs. The number of combinations this entails
would appear to make it impractical to net everything especially given that banks continue to rely on e-mail
to agree the net amount.
We therefore calculated, for a typical top tier bank, the ideal number of currency pairs, and netting
counterparties, and made the following findings. First, if such a bank were to net the top 20 currency
pairs, across all counterparties, then they could achieve a reduction of settlement risk of 53%, and 15,000
payments reduce to around 3,000. Second, if they were to net with the top 20 counterparties, across all
currency pairs, then there is a reduction of settlement risk of 60%, and 13,000 payments reduce to just 900.
Third, if they were to net the top 20 currency pairs, with the top 20 counterparties then this would cover
70% of all trades, reduce settlement risk by 50%, and 12,000 payments reduce to only 350. So, it is clear
that netting not only has the potential to reduce settlement risk significantly, but can bring many
operational and cost benefits due to the substantial reduction in the number of payments.
The above type of analysis shows how SWIFT data can help make the case for further investments to reduce FX
settlement risk. The SWIFT FX dataset is available to any SWIFT user, and we are ready to undertake
customised investigations via our professional services teams.
BIS Quarterly Review, December 2019 https://www.bis.org/publ/qtrpdf/r_qt1912.htm
BIS Triennial Bank survey https://www.bis.org/statistics/rpfx19.htm
All calculations and estimates in this article are research in progress by the author that are
published to elicit comments and to encourage debate.
Considering alternative ways to settle FX transactions Finding a role for FX
FX settlement risk is on the increase, stemming from long-running issues in the industry coupled with
new factors, including the ongoing Covid-19 crisis. We spoke to Arjun Jayaram, CEO and Founder of Baton
Systems, about the role of fintechs and emerging technologies in helping address settlement risk in the
The Bank for International Settlement (BIS) has warned the markets that FX settlement risk is on the rise
again. What do you believe is driving this increase?
There are a number of factors behind this rise. Firstly, the number of trades has increased dramatically in
FX over the past couple of years. We’re now seeing much tighter bid/ask spreads, which means that
sales and trading yields for banks have fallen. For banks, the only way to recoup that revenue is by
increasing trading volumes. So there are more FX trades being conducted, but more trading means more
The second trend is the recent growth in trading exotic or emerging markets currencies. With the majority of
emerging market or frontier currencies, CLSSettlement is not an option. This is a true risk that has
increased tremendously because there is no alternative, there has been no viable settlement venue for doing
a payment versus payment (PvP) settlement of these emerging currencies.
Volumes in some of the ineligible CLS currencies have grown markedly in recent years. Specifically, CNH, TRY
and RUB are all very actively traded. What’s also interesting to note is that the major participants
in these currencies include banks that are not always thought of as tier 1, global institutions.
Regulators require banks to hold additional capital to protect against this increased risk of unsettled
trades. In turn, this puts downward pressure on the Return on Capital metric, which is so closely
watched by the banks and their shareholders.
Why do you believe an alternative approach may be the answer?
If we look at the lifecycle of a trade, risk is being introduced into the ecosystem from the point a trade is
made and it evolves from that point. As the BIS report highlights, many of these trades are settled in a
non-PvP manner, but that is actually just looking at the final leg of the risk.
So while PvP is an important part of the counterparty settlement risk, there’s actually risk being
built up from the point the trade is made all the way to settlement. We believe that the settlement risk
creates an interesting challenge because different systems do not talk to each other and so there is a lack
of visibility into the trade lifecycle between two banks. That is why at Baton Systems we always look at the
entire lifecycle of the trade.
How can emerging technologies, such as Distributed Ledger or Machine Learning, be harnessed to improve the
situation in reducing FX settlement risk?
This boils down to providing both visibility and control. These two elements are crucial to resolve
settlement risk and some of the associated challenges around liquidity.
With newer technologies such as Distributed Ledger, we are able to see the full lifecycle of the trade. So we
now know if and when a trade is going to be netted and at what point in time in the future it’s
actually going to be settled. We can also check if the client has enough funds for the settlement so they
don’t run into liquidity issues. Then finally, we also enable settlement in a PvP manner.
Essentially, we create three pillars to support the process. One is to help manage that risk during the
entire lifecycle of the trade by providing a distributed ledger where all parties are able to see the same
system of record, populated with the same data. The second pillar is our ability to provide clients with
visibility of their funding sources and obligations, to better manage liquidity. Then the third part is
enabling PvP between the counterparties when it comes to the settlement, which eliminates risk between
What has been holding back FX institutions from considering more widespread adoption of newer technologies
or working closer with fintechs to find an alternative solution?
These trades are done in different trading venues which don’t tend to talk to each other, so one of the
first things we have to do is source these trades from those venues. Yet even when we have that data, it
doesn’t mean that the internal systems from the two counterparties that did the trade necessarily
So it’s common to see heterogeneous data sources on the trade venues and heterogeneous sources within
the systems of a bank. The next problem is that the collaboration tools being used by banks can be very old.
Even today, we are still seeing settlements being managed by phone, email and sometimes, believe it or not,
even by fax. It’s rare but we still occasionally see a system which is just a fax sent by one bank to
the other to say what they owe. More common is trying to conduct this by email which is still extremely
prevalent. In terms of operational management, this obviously creates a huge problem, particularly in cases
where they are not in agreement.
The last issue is the settlement itself, particularly if one party has to pay before they receive and there’s
a pre-funding requirement. This is where technologies such as a Distributed Ledger would be very helpful,
and one of Baton’s capabilities is to integrate this technology with the trading venues and the bank
ledgers. We don’t require them to make changes in their existing systems to benefit from our DLT. So
it is a seamless solution to a problem that exists today.
Are there any constraints on the number of currencies that you’re able to offer on the platform?
None. We use commercial bank money and not central bank money for the settlement process, so it is far easier
for us to add new currencies on the platform when required. We built the platform to be truly currency and
even asset-class agnostic – we also support securities in addition to currencies.
Currency coverage is one of the key constraints on existing PvP processes. CLSSettlement is a very efficient
system and it has reduced a considerable amount of risk in the system. So it is a fantastic venue for
settling between its 70+ members but it only settles 18 currencies. If you are not a settlement member or
want to settle transactions where either leg falls outside of those 18 currencies, then you’re out of
luck. CLSSettlement is also a batch-based system, so it only settles once a day currently. In contrast,
Baton is now able to provide on demand settlements 24 hours a day. Instead of settling just once a day, you
can settle every hour using our platform on any currency pair.
Could the industry be working more collaboratively together and with regulators, central banks or with
fintechs to address this issue?
Yes, absolutely. There are some practical steps which the central banks can take to help. One thing that
could really help is for these institutions to be more open by extending the settlement window to nearly 24
hours a day – and there would be huge demand for that. Membership criteria is also a huge issue. This
is very restrictive, with many of the most important financial institutions not eligible to use certain
central banks as a venue for settlement. For example, not even LCH can open an account with the US Federal
Reserve because it is not domiciled in the United States.
Having said this, the central banks are certainly thinking creatively, with the Bank of England and the
Federal Reserve having been notably active. In fact, the Bank of England is going one step further by having
allowed fintech companies, such as Baton, to have access to central banks and facilitate settlement between
Looking ahead, what impact do you think the Covid-19 pandemic will have on settlement risk in the FX
It’s a constantly evolving crisis, but we do see a few trends emerging. March and April saw volumes
shoot up significantly in FX. At the same time, we expected the stimulus measures taken by the central banks
to further increase FX liquidity. Instead, what there was actually a shortage of liquidity due to a huge
demand for US dollars.
The other interesting thing for the banks is the reduction in interest rates, which have dropped effectively
to zero. That has had a profound impact on banks’ ability to generate net interest income, so they
increasingly need to look at new venues or new markets to make money. The focus on these new markets
will, generally, result in the more active movement of assets. This in turn has the potential to create more
settlement risk between members. This is one of the most important macro trends that we see developing in
Following the initial volatility in March and April, we saw significant demand for our product. We believe
this was a response to the crisis having brought into sharp focus for many in the market the full extent of
these liquidity problems, settlement risk and operational inefficiencies. This is where the banks are now
spending a lot more of their time and resources.
We also see an increasing shift to cloud-based technologies. This software is always up to date and it’s
collaborative, allowing a fast and cost-effective way for the banks to improve their settlement risk
efficiently. We see this as an almost inevitable and irreversible trend across the industry.
Why settle for FX Settlement Risk?
By Maryanne Morrow, CEO and Founder 9th Gear Technologies
Settlement risk is unnecessary!
In 1974, settlement risk came to the forefront when the German bank, Bankhaus Herstatt, failed to make its
corresponding dollar payments after receiving Deutsche Marks from their counterparties. This created a
cascading effect causing more banks to stop making payments and the international payments system to
While the industry has made strides to mitigate risk, settlement risk remains a top concern and has increased
with the growth in emerging market currencies. Over the last decade, the industry has been focused on
improving electronic pricing engines and minimizing latency; now is the time to modernize how we exchange
9th Gear has re-imagined the conventional FX process from a “trade then fund” to a “fund
then trade” model which is near real-time and mitigates settlement risk. We leverage smart contracts
and a private permissioned distributed ledger to digitally transform the FX market. To facilitate dealer
participation, we created an intraday lending facility which acts as an alternative source of credit,
allowing liquidity providers to facilitate T0 transactions by borrowing in real-time.
Settlement risk has several sub-components: principal risk, replacement cost risk, liquidity risk, governance
risk and legal risk. The 9th Gear offering, through its pre-funding model and immutable ledger addresses all
but the legal risk, as proper documentation and dealing contracts must still occur.
Pre-funding transactions further negates the need to seek credit approval for transactions removing the need
for dealers to make 3rd party payments on their client’s behalf; two time consuming, low productivity
Today, retail FinTech offerings with innovative payment efficiency and speed lead the institutional space.
Furthermore, traditional trade then fund models open the industry up to failed payments which can cause
issues affecting liquidity and leading to high overdraft charges. Most FX spot transactions are traded T0
and settle on T+2. Should a participant fail to deliver their side of the transaction, the receiving
institution will likely be overdrawn as it depends on this payment. Reconciliation is operationally
intensive and request for good value is time consuming. Pre-funding transactions removes this risk.
9th Gear leverages an immutable distributed ledger along with atomic swap functionality to affect payment vs
payment protection on trades. Atomic swaps utilize smart contacts which ensure that either both legs
of a currency exchange happen or neither leg happens. Using the distributed ledger as a digital
representation of pre-funded assets held at the custodian, we swap digital assets atomically, affecting
change in ownership which is immediately reflected at the custodian giving the participants immediate use of
The concept of programable money opens the possibilities of removing
many of the frictions which exist today. Believing in the continued digitization of assets and keeping
the future in mind, our offering will be interoperable with Central Bank Digitial Currencies (CBDCs) as they
are introduced. The industry’s migration to a completely digital 24/7 instantaneous means of
exchange and settlement will take time. The underlying digital infrastructure and governance will need
to be in place along with the broad adoption of digital assets as an accepted means of exchange, be it a
cryptocurrency, stable coin or CBDCs. Many of the current compliance, governance and regulatory reporting
requirements could be built directly into the ledger mitigating the burdens within the banking community.
We believe the 9th Gear offering is the first step in the drive to modernize the way we exchange value from
the current environment to one where all currencies are replaced with CBDCs allowing for 24hr instantaneous
exchange of value.
Settlement risk is unnecessary!
Symbiont Smart contracts and collateral management
Beyond settlement risk mitigation for those CLS qualifying currencies there’s a world of exposure
that remains in the FX market. Forwards, swaps and options, counterparties that currently rely on prime
brokers for credit support and emerging market currencies – all these require reliance on credit
judgement, tenors and collateralisation that can leave much to chance. These areas of risk are
being addressed by a blockchain based platform called Assembly from New York based FinTech Symbiont. We
spoke to the firm’s FICC Business Development and Strategy Lead Joe Ziccarelli, about its
ground-breaking work to mitigate counterparty credit and intraday FX exposure
Could you explain how your technology and solutions address the problem of settlement risk in FX?
Our solution specifically addresses the risk that exists between trade date and settlement date. If there is
a credit event, you are unlikely to ever reach settlement for those contracts, because you would have
started some close out proceedings with the counterparty. So, the question is, how protected are you
with your variation margin based on current market practice? We’re bringing a modern solution to
the underlying systemic risk associated with variation margin processes that exist in the market today.
How have you been working with Vanguard and other market participants in the area of collateral
Vanguard selected the Symbiont technology after conducting extensive research and determining it was the best
suited for minimising counterparty risk in this market with a decentralized application that would promote
widespread adoptions while eliminating the need for intermediaries that add cost and expense.
While their initial use case focused on modernizing the variation margin process to reduce counterparty risk
for their own investors, the greater goal is to improve the broader market structure by limiting this
systemic risk and enabling new hedging structures to emerge for all.
Could you expand on how this technology ecosystem works for Vanguard and others?
FX market participants include asset managers, banks, broker dealers, corporates, all of whom benefit from
reducing counterparty credit risk. Prior to engaging in normal trading activity, counterparties pre-agree
the terms in the governing ISDA agreement and, importantly, the credit support annex. These define how
collateral is to be calculated and moved. Symbiont’s Assembly platform programmatically reflects these
terms in smart contracts, effectively assuming the role of calculation agent based on the inputs, timings,
etc. agreed by the counterparties.
Valuations can then occur throughout the day; counterparties can see collateral movements intraday based on
minimum threshold amounts. In this day and age, why would you wait on overnight batch processes, use a pad
and pencil or Excel to solve a mathematical equation, when you could automate that with code that executes
systemically with no human intervention? That’s what smart contracts do and what the Symbiont
assembly platform provides.
Given the vast reach of the global foreign exchange market, what is your vision of what Symbiont’s
technology can bring to the market as a whole?
Firstly, it is intended for the whole market: bank-to-client, bank-to-bank, whatever the market demands. Once
you’re on the platform you can apply the same logic to reducing your counterparty credit exposure to
everyone with whom you trade. It’s not limited, for instance, in the way that settlement risk
mitigator CLS is limited to a defined set of currencies.
In fact, the more volatile or exotic the currencies, the greater the benefit our solution delivers, because
that’s where you have the greatest intra-day mark-to-market moves, leading to the greatest intra-day
counterparty credit risks. Although size of a portfolio is another huge factor regardless of the make-up of
the underlying currencies.
Reducing systemic counterparty risk also enables participants to do more transactions at greater size within
their established relationships, and opens the door to new relationships previously not considered, because
of the increased frequency of the collateral movement reducing value-at-risk. This also extends to the
current risks and delays attached to the return of collateral post settlement to the party that posted it.
Lastly, time frames - how long do you think it will be before blockchain is the underpinning technology for
the FX market?
What we’re building is common market infrastructure on a distributed ledger through the use of smart
contracts that has been designed by market participants specifically for market participants. The
distributed ledger acts as the common platform and single source of truth. Based on client readiness, our
intent is to have production trades in the fourth quarter.
You will see a broadening of the activity and participants in 2021 because it’s simply a better, more
automated and secure way to reduce risk through tighter collateral management that the market, industry
groups and Regulators can easily get behind.
SWIFT: gpi and gFIT
By Richard Willsher
SWIFT’s Global Payments Innovation (gpi) messaging has become the norm for cross-border interbank
payments but the innovation continues.
Launched in 2017 to offer a fast, transparent and secure cross-border payment platform, gpi provides
customers with traceability of their payments. The service enables predictable settlement times, transparent
bank fees and FX rates. The key building block of the system is SWIFT’s MT103 message whose “tags”
define the details of payment and the counterparties involved.
UETR and the MT 202
Highly successful as it has proved to be over millions of payments, SWIFT in 2018 added an important new
ingredient, the UETR (Unique End-to-end Transaction Reference). This is effectively a stamp attached to each
individual payment that enabled it to be tracked to wherever it was in the payment process. Moreover, to add
a further traceability and transparency, the UETR can be published to the cloud where, as SWIFT’s
Strategic Relationships - Capital Markets leader Matt Cook explains, “A GUI or API can now interrogate
where a payment is in real-time, the status of that payment, when it’s arrived at an institution, when
it’s been released, or when it’s going to be credited into the beneficiary’s account. This
is a quantum leap.”
Next came the addition of the MT 202. Matt Cook explains that a consultation carried out among its financial
institution users surprised SWIFT as it became clear that MT 202 messages are used to substantiate a large
proportion of intra day funding and in particular FX payments that are made bi-laterally outside of
Therefore, when SWIFT started to trial its new gFIT – gpi for Financial Institutions Transfers –
offering, MT 202 was an essential component. “As of November this year,” says Cook, “entities
on SWIFT will be able to say “Actually we’d like the MT 202 included as part of our gFIT service”
and so for the first time they’ll be able to have the entire payment flow on gpi; completely trackable
and transparent with service level agreements attached.
SWIFT is currently piloting gFIT among a small but significant group of major FX market participants. JP
Morgan is one of these and we wanted to find out how it’s working for them and what the impacts of
gFIT could be, especially for settlement risk. We spoke to Sarah Dunning, J. P. Morgan’s
Commercialization and Communications Lead for Clearing Transformation, Wholesale Payments and her colleague
Matthew Smith, Executive Director, CIB Operations.
“We’ve been delighted to be able to use this data to the benefit of our clients,” says
Sarah Dunning. “We have a large wholesale payments franchise covering so many external financial
institutions clients that will really benefit once MT 202 tracking comes into effect with gFIT. It’s
been really exciting for us to be involved with this programme. It wasn’t very difficult to convince
our internal departments that they would actually benefit from this as well, because seeing the reality of
tracking MT 103s end-to-end, it’s definitely something that is of huge interest such as in FX or
securities settlements. In any treasury management function, the ability to track a MT 202 from
end-to-end, offers enormous potential.”
Matthew Smith agrees. “Having access to the real-time data of where our funds are, incoming and
outgoing, will enable us to manage positions more effectively. If we take less liquid currencies for
example, then it will enable clients to meet their obligations. Some of these currencies don’t have
overdraft facilities where certain countries have regulatory restrictions that prevent them. So, if you do
not have the money in your account, you cannot make these payments. Finding out where the payments are,
giving that transparency, will enable positions to be managed more effectively.
Another useful thing is the data,” Smith continues. “That we can analyse historical data that we’ve
got from gFIT, we can see where have payments been held up, why that has been the case, and how we can we
resolve that for the future.
So, it’s about client education. We can ask them to format their message slightly differently so that
it goes through straight-through processing rather than getting held up along the way. So, it’s
quicker, cleaner, more effective, and therefore reduces friction, which is the goal
In summary, SWIFT is harnessing data within gFIT and MT 202 to control settlement risk, simply by providing
certainty and transparency as to where payments are and what may be holding them up.
How soon will instant FX Settlement be possible?
By Brian Charlick, Principal Consultant at CGI
During the past decade we have witnessed a significant shift in the payments and banking landscape. A drive
to mobile banking, fintechs, younger tech-savvy demographics and new global financial instruments have
driven up payments volumes. This pattern of increased volumes has coincided with ever increasing demands for
payments to be executed and settled in real time. This decade has seen a plethora of developments in
payments and FX settlements, including new entrants, new models and changes to existing models.
Despite all this, the introduction of instant FX settlement and international payments has some serious
obstacles to overcome; both internal and market infrastructure and vitally, how the two intersect.
Issues to be faced
Operational constraints for domestic and international payments have major issues, ranging from regulatory
requirements, operational logistics, risk management and pricing. The regulatory requirements, largest of
all being under AML5 and increasing requirements due under AML6, mean that it is currently impossible to be
compliant while also guaranteeing instant payments. Under AML5 and AML6, the need to monitor the transaction
for potential fraud and check against the sanction lists for not only the payer and payee, but also the
ultimate beneficial owner means exceptions that need to be investigated will increase, possibly up to 5% of
all transactions falling within the scope of the regulation. Given that level, the guarantee of instant
becomes a real problem. Smaller value and crypto currencies will be impacted as they are captured under AML5
Operational logistics such as settlement will need to change, with the current netting model no doubt being
dropped, as this stops real time payments. The systems will need to cope with the increase in payments to be
Pricing tools and processes will need to change. The principle of revaluation of positions at set times in
the day for risk and controls will not work. In place will need to be real–time valuation of
positions, mirroring the valuation and pricing seen by traders for many years.
Risk management will also be impacted as risk positions will require real–time valuation. Limits will
need to be administered instantly, with trades being blocked real time rather than by reference to an
overnight limit update.
A bigger concern is control. The need to monitor and control real time will provide the biggest challenge to
financial service organisations. To be able to see trends and breaks in operation, or risk limits being
broken, will require an enormous effort.
Market infrastructure also has issues to resolve. The model of netting, used by many, will be all but
redundant. Payment instructions will need to be transmitted real time. The role of the correspondent bank
becomes a question too. Using a middle person slows the process. In this case the market needs to have a
payment structure that allows small financial service organisations, as well as non FS firms to enter the
network at a cost–effective rate. SWIFT could still be used as the communication channel, although
others will also seek to provide that service such as Google, Apple, Amazon, Paypal etc. Finally, with the
advent of the tech giants’ payment offerings and developments, the risk of fragmenting the market is
However, a newer challenge is DLT and the Central Bank Digital Currency – how might this be used and to
what impact? Will CLS use it to provide tokens across central banks, fitting within existing models, or will
it be used by firms like Mastercard and become a bigger disruption?
Does that mean real time will not happen?
The market is responding, CLSNow is one starting point, SWIFT gpi and ISO20022 is another. Meanwhile,
technology is beginning to enable Instant Payments, UK Open Banking being one example. It is evident,
however, that there is much still to be done to facilitate real time; as mentioned, across market
infrastructure and internally.
When looking at the potential changes, it is important to realise that in the past we have seen successful
change driven by a series of small, iterative, rapid changes in the models rather than a big bang, certainly
where the market drives the change rather than a regulation being laid down. I believe this will be the case
here. By way of example, we are already starting to see two streams of payments mature with retail consumer
and small business payments going through Travelwise, PayPal, Amazon, etc, while the larger banks and client
organisations are using SWIFT and the traditional model; CBDC could become a part of the process for CLS,
retaining the same market model.
Internally, traders will need to see real–time positions, risks and limits. Hedging will become more
difficult, potentially using algo or analytics to execute, as real time will reduce the ability to foresee
the positions and arising exposure. Risk management will have to become real time and monitoring the limits
for clients will have to change. It will need to be done in real time with limits restricting the trading
ability immediately. Accounting will need to revalue real time in order to provide risk data for monitoring
activity. Settlement systems will have to create immediate settlement instructions that are sent without
manual validation, causing a knock on effect for reconciliation. Finally, and as important, is the need to
create controls in real time; information, I use the word information rather than data, will be required in
real time in order for control to be effective.
On the up side, intelligent automation will enable much of the control to be faster and more prescribed
market intelligence to be created.
The growth of FX transaction volumes doesn’t look like slowing and the demand for real-time settlement
isn’t disappearing anytime soon either. In the near future we may see some sporadic attempts, but the
market in general is still some way off. To get ahead of this challenge, market participants need to review
their internal operating models and controls, embrace and leverage new technologies and collaborate with
others to ensure a consistent industry approach.
GFXD settlement risk, PvP and beyond
Payment versus payment (PvP) is the ideal solution for eradicating settlement risk but where this isn’t
(yet) possible the market needs to consider all options and technologies. We spoke to Andrew Harvey,
Managing Director Europe of the GFMAs Global FX Division.
The Global Financial Markets Associations (GFMAs) Global Foreign Exchange Division (GFXD) was formed in
co-operation with the Association for Financial Markets in Europe (AFME), the Securities Industry and
Financial Markets Association (SIFMA) and the Asia Securities Industry and Financial Markets Association
(ASIFMA). Its members comprise 24 global foreign exchange (FX) market participants , collectively
representing a significant portion of the FX inter-dealer market. It is these members that are most exposed
to settlement risk, particularly that which falls outside of PvP via CLSSettlement. Consequently, over a
number of years, the GFXD has been examining possible solutions and remedies for settlement risk and
contributing to high-level discussions with central banks, policy makers and other market and industry
In 2018, The GFXD’s Recommendations for Settlement Netting1, explained how settlement
netting can reduce settlement risk for those trades that settle outside of CLSSettlement. “In our
view,” Andrew Harvey explains, “the counterparties to trades, should try to use settlement
netting wherever they can. We are very much in agreement with the FX Global Code’s principle 50 which
specifically states that market participants should measure and monitor their settlement risk, seek to
mitigate that risk where possible, and talks about FX settlement netting and encourages that.”
How this is best done is the crux of the matter. Harvey reiterates the points made in the paper that
counterparties need to agree at the outset, at onboarding, how trades will be netted and adhere to the
agreed process on a consistent basis. This however is not perfect for two reasons in particular. Firstly,
that a settlement of the netted amount still needs to be made and although the amount may be less the
settlement risk is still present. Secondly, Harvey makes the point that in-house operational systems are set
to be as efficient as possible to settle on a net or gross basis. Switching between these, perhaps on an ad
hoc basis, can introduce a measure operational risk.
New technology and interoperability
From a strategic point of view the GFXD would like to see as much flow in the market settled through PvP as
possible simply because it is the tightest possible response to settlement risk. It’s paper on
expanding PvP2 sets out its position. However, recognising that this is not always possible, the
GFXD awaits further analysis from the BIS, CLS and others to learn more about the character and composition
of the residual settlement risk existing in the market.
New technologies and market behaviours may point the ways ahead. For example, extending the time windows of
central bank real-time gross settlement systems or new technologies such as wholesale central bank digital
currencies– all of these have merit the GFXD argues. However, its concern is that there needs to be
interoperability3. As a global market with many participants continually interacting with each
other it is essential that their systems and procedures match. Unless this is enabled, Andrew Harvey says
that new technologies may not be best suited to the market though it is an evolving picture.
He asks, “Is it a case of doing things better today, or of using new technology to do things better
than today, or is it a mix of the two? I think, when you start looking at FX, being by definition
cross-border, you need the two jurisdictions to engage, and then you start looking at the efficiencies of
scale and funding and liquidity provision. It needs to be more than two central banks however. I think that’s
where the CPMI (The Bank for International Settlements’ Committee on Payments and Market
Infrastructure) work on enhancing cross-border payments is well received, and is exactly the right place to
start thinking through some of these scenarios. We’re always happy to engage with them on this and to
provide feedback where helpful.”
Andrew Harvey concludes by reiterating the goal of further de-risking the global FX system through increased
use of PvP. “I think that is where the industry is ultimately looking to move to. I think the
timeframes for achieving that will largely depend on the opportunities that are out there, either through
existing services or new technologies that are being developed.”
As Global Head of Foreign Exchange Operations at Citi, with over 20 years of background at the bank,
Leigh Meyer has an unrivalled perspective of FX operations. We asked him to comment on aspects of
netting settlement risk.
How would you describe the current state of settlement risk in the global FX market?
For the last 18 years CLS has effectively created a multilateral netting process. The success of CLS has
sharpened the market, regulators and central banks appetite to look at the benefits of compression/netting
in the FX market overall. The limitations of expanding CLSSettlement to include further currencies have been
in terms of who can provide the appropriate levels of liquidity, and whether there is critical mass in terms
of that currency in the market to make it cost-effective. Meanwhile, without the current benefit of
something like CLSNet, which is market-wide and effectively provides some consistency and common rules, most
netting is done on a bilateral basis.
So when netting is not taking place through CLS how effectively does the process work?
For example, we would agree with a counterparty what they could net and how we were going to proceed with it.
Now, this is where it becomes a little more risky. This is because you get close to currency cut-offs, and
when you’re agreeing big nets, particularly with prime brokers which are high-volume nets but not
necessarily high notional nets, what is included and what is excluded requires agility and effective
process. If you’re doing multi-thousand item nets, the chances of you not agreeing the net are quite
substantial. But if you’re actually executing over a common platform, it cuts out an element of
process in terms of confirmation. So, in other words, we are good for our price, you take our price. You hit
our bid, we’ll be good for that. In the interbank market this bilateral netting process is
surprisingly effective day-by-day. Right now, our process is more focused on looking at institutional,
corporate and real money clients, and say there’s more risk in that space, particularly where we are
under current circumstances, i.e. COVID adding stress into the operations system, requiring enhanced,
effective and practical controls because of the remote connection with colleagues, approvers and
authorisers. There’s more risk in the current model without these enhanced controls. Ironically,
through payment versus payment, you solve your credit risk, what you don’t actually resolve or remove
is your processing/operational risk.
Could you explain a little more about what you mean by processing/operational risk?
Netting, effectively, undermines much of what has been driven by the banks for years now in terms of
straight-through processing. If you work on the basis that, if I do 10 transactions and I submit 10
confirmations, and you match those, effectively I don’t have to touch them operationally because the
system matches it and proceeds to create the necessary payment messages from a SWIFT standpoint. That’s
great operationally because I’ve effectively eliminated the settlement risk because I know the
economics of that transaction. I know where I’m paying, I know how much I’m paying and I know
what date I’m paying. When you net, you’re effectively saying, “Hold on, I need to take
all those transactions out, compress them, and condense them into a single payment, but at the same time I
am creating operational risk and processing risk by that process.”
In your view then, does CLS remain the best settlement risk solution?
CLS is the market standard. All the new blockchain technologies that are being talked about lack critical
mass, which the market needs to be scalable and able to set standards. In reality, to actually put a
counterparty into CLS might be a little expensive in terms of the upfront cost for the counterparty, but it
gives them a whole suite of agent banks and counterparties to deal with, and of course it gives them CLS’s
full currency suite. This is not to say that new market standards won’t be developed but for the time
being CLS has done a remarkably good job.
Unsettling: the increase of foreign exchange without settlement risk
The foreign exchange (FX) market is global, vast, cross-border, and operates 24 hours per day. Its effective
functioning facilitates international commerce and is a pillar of a sound financial system. A disruption to
the FX market – particularly a materialization of FX settlement risk – could cause panic in
markets around the world.1 Despite these understood risks, the Bank for International Settlements’
(BIS) Quarterly Review (December 2019) suggests that FX settlement risk is on the rise. Further, and of
note, FX settlement risk is growing in currencies not eligible for settlement in CLS.2
CLS believes now is the time to address and reverse the build-up of FX settlement risk. Without immediate
action, FX settlement risk will continue to accumulate and, in parallel, so will the risk to the global
financial system. The regulatory community and industry must join forces to reverse the expansion of FX
settlement risk before it can inflict damage to markets and the economy more broadly. This paper
outlines the history of FX settlement risk and the response by the industry and regulatory community to
date. Specifically, the paper explains:
The history of FX settlement risk, including background on CLS’s origin and current activities
Growing FX settlement risk
Existing obstacles to CLSSettlement currency expansion How to address FX settlement risk
How to address FX settlement risk
History of FX settlement risk
Central banks and regulators have debated FX settlement risk, with varying degrees of intensity, over a long
period of time. On June 26, 1974, German authorities revoked Bankhaus Herstatt’s (Herstatt) license to
conduct banking activities. The close of Herstatt at the end of Germany’s banking day, while New York
markets were still open, resulted in a loss of principal for Herstatt’s counterparties. These
counterparties had already paid Deutsche marks in Frankfurt, but had not yet received dollars when Herstatt’s
New York correspondent bank suspended all outgoing US dollar payments from Herstatt’s account. Even
though Herstatt was not one of Germany’s largest banks, its failure resulted in widespread panic in
the markets, a freezing of interbank lending markets, and tremendous distrust in inter-bank relations. This
episode gave rise to the term “Herstatt Risk” or “settlement risk”. See Figure 1.The
failure of Herstatt was a turning point, and the need to tackle FX settlement risk became a top priority for
the international regulatory community. The 1980s and 1990s saw a flurry of central bank activity, largely
led by the G-10 central banks of the Committee on Payment and Settlement Systems (CPSS). Extensive research,
analysis, and market surveys identified issues and risks raised by cross-border and multi-currency netting
arrangements, as well as existing FX settlement practices. At the same time, several other notable bank
failures emphasized the need for a solution – Drexel Burnham Lambert in 1990, BCCI in 1991, and
Barings Brothers in 1995.
In 1996, CPSS outlined a three-pronged approach for a new partnership between the industry and the central
bank community.3 First, individual banks needed to look within and take steps to apply
appropriate credit controls to their FX settlement exposures. Second, the CPSS called on industry groups to
develop multi-currency settlement and netting arrangements to contribute to the risk-reducing efforts of
individual banks. Lastly, central banks needed to show their support of industry initiatives and cooperate
with these groups to bring about timely, market-wide progress.
Following CPSS’s recommendation, 20 major financial institutions formed a group which, with support
from the central bank community, further refined the linked settlement concept – an arrangement
involving simultaneous PvP exchange of each of the two legs of an FX transaction – that would
eventually lead to the creation of CLS.4 CLS’s FX settlement service (CLSSettlement) went
live in September 2002 with 39 settlement members (many of whom were part of the group of 20 financial
institutions) and seven currencies. See Figure 2. Today, CLS’s membership comprises over 70 of the
world’s largest financial institutions, and CLS is member-owned. Over 25,000 third parties, primarily
buy-side institutions, access CLSSettlement via a number of CLS’s settlement members. CLSSettlement
now settles 18 actively traded currencies, and to carry out these operations CLS has accounts with each of
those 18 central banks.5 Further, these central banks adjusted their operating hours to
accommodate CLS settling in a two-hour settlement window. On average, CLS settles USD6.0 trillion of payment
instructions per day. The funding required to settle this amount is determined on a multilaterally netted
basis, reducing the amount of liquidity required for settlement by approximately 96 percent.The global
financial crisis of 2008 again reminded the world of the importance of mitigating FX settlement risk. While
trading in fixed income, rates, and structured product markets were disrupted or effectively ceased because
of counterparty credit concerns, the FX market continued to function smoothly. Major banks continued to
trade knowing their trades would settle in CLS with the significant risk mitigation provided by
PvP.6 Recognizing this important role in the proper functioning of global FX markets, the United
States’ Financial Stability Oversight Council designated CLS Bank International as a systemically
important financial market utility (i.e., DFMU) in 2012.7 Recent financial market volatility
resulting from the impact of Covid-19 has only reinforced the importance of resilient and well-regulated
financial market infrastructures like CLS. In March 2020, CLS volumes reached record-breaking levels. The
average value of payments settled daily totalled approximately USD7.0 trillion - about 20 percent higher
than normal. CLS processed the added volumes with no issues or delays.
Growing FX settlement risk
Although the launch of CLS in 2002 reduced the amount of FX settlement risk in the market, a 2008 CPSS report
demonstrated that banks were not mitigating this risk as much as they could and urged banks to do more.8
The need for an industry response was re-emphasized by the Basel Committee on Banking Supervision (BCBS) in
February 2013 via its “Supervisory guidance for managing risks associated with the settlement of
foreign exchange transactions” (commonly referred to as BCBS 241). Following its publication, the BCBS
expected banks and national supervisors to implement BCBS 241 in their jurisdictions while also considering
the size, nature, complexity, and risk profile of banks’ FX activities. Seven years later, more work
is required to implement BCBS 241 into national supervisory practices. The BCBS recognized this shortcoming
in October 2019, and publicly acknowledged the need for further measures to mitigate FX settlement risk.
9 The BIS Quarterly Review (December 2019) concluded that a significant portion of the global FX
market continues to be settled without PvP protection.10 Of the USD18.7 trillion of daily gross
FX payment obligations, USD8.9 trillion of payments (approximately half) are at risk. While the decline in
the proportion of FX transactions settled with PvP protection is partly explained by the growth in
currencies not currently eligible for settlement in CLS, a significant percentage of trades in CLS-eligible
currencies are also settled without PvP protection. Independent analysis reinforces the BIS’s
Existing obstacles to CLSSettlement currency expansion
Conscious of these market evolutions and derived challenges, CLS continually assesses ecosystem systemic risk
mitigation measures it can bring to the market. In 2018, CLS launched CLSClearedFX as the first PvP
settlement service specifically designed for OTC cleared FX derivatives. In Q3 2019, CLS launched CLSNow
– a same-day FX PvP gross settlement service. For the first time, CLS settlement members are able to
mitigate FX settlement risk in the same-day market for Canadian dollar, euro, UK pound sterling, and the US
dollar. Plans are underway to expand to more currencies. Few remaining currencies can meet currency
onboarding standards, which derive from the Committee on Payments and Market Infrastructures (CPMI) and the
International Organization of Securities Commissions’ (IOSCO) Principles for financial market
infrastructures (the PFMI), other applicable regulations, and CLS’s own standards.12
Principle 1 (legal basis) and Principle 8 (settlement finality) of the PFMI have presented the greatest
challenge to onboarding additional currencies to CLSSettlement, in particular matters relating to
availability and enforcement of settlement finality legislation.13 For example, in 2019, CLS and
Banco Central de Chile announced efforts to onboard the Chilean peso. This work is now possible following
changes to Chile’s settlement finality legislation. If successful, the Chilean peso will be the first
CLS-eligible currency from South America. However, many countries seeking PvP protection for FX settlement
may not be able to obtain it under the current regulatory regimes applied to FMIs offering such services.
Addressing FX settlement risk
CLS believes immediate action is required to address the apparent growth of FX settlement risk. Specifically,
CLS and the industry more broadly, with support of the regulatory community, should focus efforts on: 1)
further promotion and adoption of PvP settlement amongst banks and non-banks; and 2) mitigation of growing
FX settlement risk in non-CLS currencies.
1. Promotion and adoption of PvP settlement: The BIS 2019 Triennial Survey data and the
BIS Quarterly Review (December 2019) demonstrate the industry can do more to promote and adopt PvP
settlement solutions. One potential course of action is for banks and non-banks to evaluate existing
operations and identify which transactions are and are not settling via PvP, and for what reason. Following
this type of analysis, these market players would be in a position to consider ways to maximize the use of
PvP settlement solutions. Additionally, relevant industry codes or regulatory guidance could be reviewed and
amended to further promote PvP as a best practice for market participants.14
2. Solutions for non-CLS currencies If FX settlement risk in non-CLS currencies is to
be mitigated, a fundamental consideration is whether a new model is better than the outright risk taken
today by financial market participants in trading these currency pairs. Further, trade-offs and choices in
design elements, which must be different to CLSSettlement, should be considered to achieve a model that can
be implemented and can maximize broad-based risk mitigation.
Preventing further growth of FX settlement risk is not an impossible task, and mitigation of this risk should
be at the forefront of the industry and regulatory agenda globally. While CLS is encouraged by recent
acknowledgments that more work is needed to mitigate growing FX settlement risk, a cooperative effort
between the industry and regulatory community is required to take this work forward and to ensure its
success. Together, the unsettling increase of FX without settlement risk mitigation can be
The Committee on Payment and Settlement Systems (CPSS), which was renamed the Committee on Payments and
Market Infrastructures (CPMI) in 2014, defines FX settlement risk as the risk that one party to an FX
transaction will pay the currency it sold but not receive the currency it bought.
BIS: “BIS Quarterly Review - International banking and financial market developments”,
specifically Bech and Holden: “FX Settlement Risk Remains Significant” (December 2019).
CPSS: “Settlement Risk in Foreign Exchange Transactions” (March 1996). bis.org/cpmi/publ/d17.pdf
PvP ensures the final transfer of a payment in one currency occurs if and only if the final transfer of
a payment in another currency or currencies takes place.
Australian dollar, Canadian dollar, Danish krone, euro, Hong Kong dollar, Hungarian forint, Israeli
shekel, Japanese yen, Korean won, Mexican peso, New Zealand dollar, Norwegian krone, Singapore dollar,
South African rand, Swedish krona, Swiss franc, UK pound sterling and US dollar
Levich: “Why foreign exchange transactions did not freeze up during the global financial
crisis: The role of the CLS Bank” (July 2009). voxeu.org/article/clearinghouse-saved-foreign-exchange-trading-crisis
CLS Bank International is the legal entity operating CLSSettlement.
As a DFMU, CLS must comply with regulations and standards applicable to infrastructures of systemic
importance, which is the Federal Reserve’s Regulation HH. The PFMI applies to all FMIs determined
by national authorities to be systemically important.
Principle 1 requires FMIs to have “…a well-founded, clear, transparent, and enforceable
legal basis for each material aspect of its activities in all relevant jurisdictions”. Principle 8
states “An FMI should provide clear and certain final settlement, at a minimum by the endof the
value date. Where necessary or preferable, an FMI should provide final settlement intraday or in
For example, there may be scope to strengthen Principle 50 of the FX Global Code (relating to FX
settlement risk mitigation) to better emphasize the use of PvP settlement solutions.
The unsettling nature of rising FX Settlement risk
By Marc Bayle de Jessé, CEO, CLS
The Bank for International Settlements’ (BIS) Quarterly Review published in December 2019 concluded
that FX settlement risk is on the rise due to a significant portion of the global FX market being settled
without protection. According to the BIS data, the volume of trades settled with PvP protection decreased
from 50 percent in 2013 to 40 percent in 2019. Of note, the data showed that FX settlement risk is growing
in currencies not eligible for settlement in CLS. In emerging markets, the settlement risk exposure against
the US dollar (USD), and to a lesser extent the euro, equates to approximately the same amount that CLS was
built to address when it was created in 2002.
CLS responded to the BIS data with a white paper calling for the FX industry and regulatory community to
collaborate to address growing settlement risk. The topic was presented at the Global Foreign Exchange
Committee (GFXC) and several local FX committees, and has brought the issue of settlement risk back to the
forefront for regulators and market participants alike.
“The amount of FX settlement risk exposure that remains is a prudential exposure bilaterally
between counterparties, as well as of systemic concern when aggregated”.
Larry Sweet, Senior Vice President, Federal Reserve Bank of New York
As a global financial market infrastructure, CLS has a responsibility to help market participants understand
the impact of settlement risk within the FX ecosystem. To further the debate, I recently chaired a
discussion on this topic with representatives from both the public and private sector. Here we share those
viewpoints as we look to advance the settlement risk discussion.
According to Larry Sweet1, Senior Vice President, Federal Reserve Bank of New York, settlement
risk remains the major source of bilateral exposure and systemic disruption. “The implementation of
PvP initiatives led to significant improvements in the industry. However, over time the private and public
sector have been monitoring this and unfortunately, notwithstanding the progress that has been achieved, the
amount of FX settlement risk exposure that remains is a prudential exposure bilaterally between
counterparties, as well as of systemic concern when aggregated.”
“Emerging markets are even more challenging as clients expect the banks to be flexible with
regards to gross settlement payments.”
William Shek, Head of FX, EM Rates & Commodities, Debt Trading & Financing, ASP HSBC Global
Banking & Markets.
Bill Holmes, Head of Bank and Counterparty Risk, Europe & America, ANZ also highlighted existing
settlement risk issues. “Being full on settlement limits has wider implications for our client
relationships, not just our bank relationships. As a result, we have negotiated individual net settlement
agreements with some of our most active FX trading counterparties, but since they operate on a legal entity
basis some of their subsidiaries and our interbank counterparts still remain exposed to this
Larry supported this assertion by citing anecdotal evidence that suggests firms are not properly managing
their settlement risk exposures. Commenting on the prolonged periods of exposure that result from issuing
payment instructions well before the settlement date, Larry noted, “The risk of paying and not
receiving can begin once you no longer have control of your outgoing payment instructions”. Larry also
observed that despite settlement risk exposure being a pure counterparty credit exposure, firms tend not to
treat it as such and are willing to tolerate a higher exposure to settlement risk than traditional
counterparty risks. This is partly because there typically is no regulatory capital charge placed on
According to Larry, the solution to reducing settlement risk should be managed through a combination of
individual trader awareness of the size/duration of exposures and the ability to measure and control it,
after which it can be reduced. Larry mentioned that various FX committees have recently focused their
attention on settlement risk and are considering individual and collective action to address it. We support
these initiatives as we believe settlement risk is best addressed through market re-education about
settlement risk exposures and a collective and consistent industry response.
William Shek, Head of FX, EM Rates & Commodities, Debt Trading & Financing, HSBC Global Markets, Asia
Pacific agreed with this perspective, noting that in emerging markets client behavior is more likely to
change through an industry-wide approach and a broader understanding and recognition of settlement risk. He
said, “When we talk to clients, many are aware of the benefits of CLSSettlement, but as they normally
receive sufficient settlement limits from banks, they do not have that incentive to change to a new model.
However, I believe that settlement risk should be a higher priority because FX volumes in emerging markets
“We have also seen that the enhanced global regulatory framework is pushing more institutions
to join CLSSettlement.”
Sandra Laielli van Scherpenzeel, Executive Director / Global Head Cash Banks, UBS Switzerland AG,
Corporate & Institutional Clients
Further, according to William, “Emerging markets are even more challenging as clients expect the banks
to be flexible with regards to gross settlement payments; if the banks are not flexible enough, it could
impact market share because of competition with other dealers.”
The positive news is that there is growth in the CLSSettlement community both through an increase in
settlement members as well as through increased third-party access. CLS has seen an annual 6% increase in
third-party activity since 2019.
Sandra Laielli van Scherpenzeel, Executive Director / Global Head Cash Banks, UBS Switzerland AG, Corporate
& Institutional Clients, highlighted the receptiveness of third-party market participants to better
manage settlement risk. “In the last three to five years we have seen an increase in interest among
third-party users to join CLSSettlement, which has led to an uptick in volumes. For example, looking at
institutional firms such as pension funds, we have seen their investment appetite is changing, and they are
looking differently at their returns. As a result, they recognize they need a strong infrastructure which
can mitigate the risk that arises from their FX trading activities.”
Sandra added that regulation is leading greater numbers of third-party market participants to acknowledge the
importance of PvP settlement. “We have also seen that the enhanced global regulatory framework is
pushing more institutions to join CLSSettlement, despite the fact that their own domestic currency is not
yet being settled in CLSSettlement.”
Following on from this discussion, I asked Paul Sassieni, Head of Capital Markets and Treasury Credit Risk at
Northern Trust, for his views on why there has been a 10% growth in asset managers joining CLS over
the past year. He attributed this trend to three key industry drivers: “First, the growth in
global investment. In the past funds tended to focus on domestic markets; however more recently they are
investing internationally, which has created a tremendous amount of transactional FX activity. Second, there
is an increase in third-party FX trading. Previously funds tended to trade FX primarily with their custodian
bank. However, as many clients have moved to a multi-dealer model for FX, funds are not necessarily
transacting their FX business with their custodian bank where settlement risk is eliminated. Hence over the
last few years more trades have become subject to settlement risk. Third is the trend for block trading,
where, in the quest for best execution, asset managers conduct a single block trade for a large number of
funds they manage. While some of those funds may be clients of a custodian bank, other funds may not, which
creates settlement risk.”
“The growth in CLSSettlement is not only being driven by credit risk mitigation, but also the
operational efficiencies CLS provides.”
Paul Sassieni, Head of Capital Markets and Treasury Credit Risk at Northern Trust
Paul and Sandra both agreed that third-party membership of CLS is not only being driven by the need for
settlement risk mitigation, but also the wider benefits that CLSSettlement delivers. Sandra pointed out the
importance of enhanced liquidity and keeping the cost of capital down through the use of services such as
CLSSettlement. Indeed, through multilateral netting, CLSSettlement reduces funding costs by 96% and frees up
the liquidity the global FX market needs to function effectively.
Paul also pointed out the operational efficiency benefits delivered by CLSSettlement. “The growth in
CLSSettlement is not only being driven by credit risk mitigation, but also the operational efficiencies CLS
provides through fewer failed trades and cost efficiencies for clients and custodian banks”. Echoing
this opinion, Sandra remarked, “The need to improve operational efficiencies has created a strong
business case for banks to join CLS”.
On the topic of rising settlement risk in trading non-CLS currencies, Bill said it is the most significant
risk ANZ faces. He believes that settlement risk is increasing due to increasing trading activities in
non-CLS currencies, particularly in Asia. Adding to this, Bill said, “When trading Asian currencies,
you are exposed to a longer time lag between the Asian currency settling and the US dollar
Bill noted particular challenges with the Chinese Yuan Renminbi (CNH). “Looking specifically at CNH, it
is increasingly settling on International Monetary Market dates resulting in tremendous pressure on large
banks, particularly those with prime brokerage franchises where activity can increase five-fold on selected
“When trading Asian currencies, you are exposed to a longer time lag between the Asian
currency settling and the US dollar settling.”
Bill Holmes, Head of Bank and Counterparty Risk, Europe & America, ANZ
William emphasized other settlement risk challenges in the region. He noted that there can be frequent
settlement failures in emerging market currencies, as these the settlement infrastructures for these markets
are not as developed or efficient.. “ Settlement failures and risk could be reduced if payments took
place on a PvP settlement basis.”
This discussion has highlighted many of the themes that we, at CLS, are discussing as we engage the industry
in venues including the GFXC and various local FX committees. It is clear from our stakeholder engagement
that the industry recognizes the issue of rising settlement risk and stands ready to address it. In these
discussions, we have provided views on a potential settlement risk solution for non-CLS currencies that may
involve adjusting the current CLSSettlement model to provide an alternative form of PvP protection. However,
we recognize that any new PvP model can only be achieved through close collaboration between the public and
private sector – a partnership to which we are wholly committed.
1 The views expressed are the individual’s own and do not necessarily represent the views of
the Federal Reserve Bank of New York
THE LAST WORD
By Sam Romilly, FX Global Market Management, SWIFT
SWIFT commissioned this e-Forex magazine supplement because of the size of the outstanding risk, and the
complexity of the problem. Our objective was to bring together some of the top practitioners, solution
providers, and thought leaders to really dissect the problem, to examine what solutions exist today, and to
see what could be done in the future. Addressing FX settlement risk will require major improvement,
and evolution, of the current methods. It could also potentially require radical paradigm shifts to
alternatives such as pre-funding, collateral and new technologies.
This has turned out to be good timing as the FSB has just released their Stage 3 report on enhancing
cross-border payments where one of the ‘Building Blocks’ outlined is titled “Facilitating
increased adoption of PvP” that directly addresses the topic of FX settlement risk. (See Figure
1 below) The FSB will present a consolidated report to the G20 each year to report on progress, and
to review actions and timelines.
Currently CLS removes around $5.5 trillion of settlement risk each day, which leaves, according to BIS
figures, at least twice this amount to settle each day by other means. BIS estimates that the
proportion of trades with PvP protection appears to have fallen from 50% in 2013 to 40% in 2019.
Some underlying reasons for this are: (i) the long-term upward trend in overall volumes, (ii) the increasing
share of prime brokerage, (iii) the increasing use of swaps, and (iv) the trading volumes of non-CLS
currency pairs where the share of emerging market currencies rose to 23% in 2019 from 19% in 2016. The
increasing use of the swap instrument may well be the more significant factor behind this increase in
settlement risk given its take-up by the small regional banks, who are less likely to use CLSSettlement, and
the fact that each swap trade implies four settlement legs.
A common theme across all the articles is the need for more industry data to understand how the 60% of those
FX obligations without PvP protection actually settle. Whilst a significant proportion would be via
the correspondent banking process, other settlement mechanisms exist such as cross account movements for the
“on-us” settlement process, internal account movements for intra-group trades, and back-to-back
movements for the prime broker ‘give-ups’ process.
Another theme is the need for inter-operability between new alternative settlement solutions being
contemplated. On the one hand, central banks are exploring PvP settlement solutions based on new models such
as the ‘synchronisation’ of RTGS cash movements, extended opening times, and use of wholesale
crypto-currencies. On the other hand, FinTech companies are promoting new PvP solutions based around smart
contracts, distributed ledger and use of escrow accounts.
However, in many cases such innovations introduce a dependency on pre-funding which has implications on
liquidity as funds need to be available at trade date rather than at value date. Any ‘same day’,
or even ‘instant’, settlement needs to address this issue of liquidity. CLS has a significant
advantage as multilateral netting reduces pre-funding requirements by up to 96%. In addition, CLS offers a
liquidity management In/Out swaps facility to reduce funding requirements by a further 3%.
This then is the quandary faced by some of these new models. How to align prefunding of FX transactions with
the need for liquidity. There are many innovative ideas such as intra-day borrowing, posting of collateral
etc also discussed elsewhere in this supplement.
These new models are more likely to be complementary to CLSSettlement than alternatives, but
inter-operability will be key to their success. There also still remain significant opportunities for growth
in CLS based settlements by increasing the number of third party clients, as can be seen and quantified, in
the SWIFT FX dataset.
The introduction of gpi for the MT 202 financial transfer instruction, the gFIT service does
not take away settlement risk but it makes a major improvement to the monitoring and
management of it
The crucial step in FX settlement are the actual payment instructions, both for netted FX transactions, and
for FX transactions that will settle gross. Whether settled gross or settled net, each party will send
an outgoing payment, and will expect to receive an incoming payment. This is of course the heart of
the FX settlement risk as if the outgoing payment is sent, but the incoming payment does not arrive, then a
settlement failure occurs. And the amounts involved can be large as verified by our SWIFT FX dataset
where we can see the average EUR:USD gross settlement exposure between any two players for every day in the
The introduction of gpi for the MT 202 financial transfer instruction, the gFIT service, means that FX
parties can now monitor both the progress of the payment they sent for the currency they sold, as well as
the incoming payment they expect for the currency bought. Whilst this does not take away settlement
risk, it makes a major improvement to the monitoring and management of the risk, especially if it were to be
combined with bi-lateral netting solutions. gpi is in the process of transforming the correspondent
banking payment processes.
Looking forward to the future this basic principle of linking the two payment legs together via gpi could
support new PvP based solutions. For example, if a crypto-currency were to become the settlement
mechanism for a FX trade it would still require links to accounts to know when both parties have made their
payments in order to transfer ownership. Similarly, a solution based on an escrow account arrangement
would be able to release funds based upon receipt of gpi confirmation of credits for each payment leg. gpi
users could themselves set-up a market infrastructure to offer PvP services based on the updates for each
payment leg received from the gpi service. The importance of data and of visibility on assets is clear, and
to this end SWIFT shall continue to extend the transaction management and data capabilities of our
FX settlement risk is an industry critical subject and we hope that this e-Forex supplement has helped to
contribute to the debate. Conversations and analysis will certainly continue and SWIFT looks forward
to be part of the dialogue. We also look forward to your comments on this supplement.