(AGENPARL) – ven 02 dicembre 2022 by Marc Glowka, Alexander Müller, Livia Polo Friz, Sara Testi,
Massimo Valentini and Stefano Vespucci
The views expressed in the papers are those of the authors and do not necessarily reect
Via Nazionale, 91 – 00184 Rome – Italy
Deutsche Bundesbank.
European Central Bank.
by Marc Glowka,* Alexander Müller,* Livia Polo Friz,** Sara Testi,**
Massimo Valentini*** and Stefano Vespucci**
Keywords:
TABLE OF CONTENTS
LYTICS
Box 1
RESULTI
Environmental interdependencies
IQUIDITY
Intraday liquidity indicators
Liquidity saving mechanisms
Cost recovery
Box 2
Additional analysis of simulated operational failure
Operational outages of participants
RTICIP
around the globe. Moreover, although mainly developed for regulatory
importanza sistemica (SPIS), deve valutare costantemente la resilienza di
rischi, l’Eurosistema utilizza un approccio che combina elementi qualitativi e
analyses are, in particular, used for risk assessment in accordance with the
(tiered participation arrangements). They are, however, also instrumental in
The Eurosystem’s quantitative compliance toolkit ranges from individual
Regulation of the European Central Bank (EU) No 795/2014 of 3 July 2014 on oversight requirements for
Regulation (EU) 2017/2094 of the European Central Bank of 3 November 2017 amending Regulation (EU) No
795/2014 on oversight requirements for systemically important payment systems (ECB/2017/32) (OJ L 299,
fruitful exchange with other central banks around the world on the analytics
view of the future challenges and changes to be faced in the payment systems
the functioning and stability of the nancial system, it is of paramount importance
that their risks are efciently managed. They therefore have to comply with
regulatory standards. In 2012, the CPMI-IOSCO’s PFMIs
established new
international standards for dening and assessing the robustness of FMIs in terms
of risks and efciency. The principles, drawn up after the outbreak of the 2008
nancial crisis, are designed to ensure a robust infrastructure to support global
the efcient management of payment system-specic risks, as well as sound
and information security, custody and general business risks. For credit and
and
Committee on Payment and
Regulation of the European Central Bank (EU) No 795/2014 of 3 July 2014 on oversight requirements for systemically
Guideline (EU) 2015/510 of the European Central Bank of 19 December 2014 on the implementation of the
LYTICS
to be supported by the analysis of granular data about the system’s activity
interdependencies, Principle 7 on liquidity risk, Principle 17 on operational
time and in central bank money, with immediate nality. It is used for payments connected with
accessed through different channels, depending on the participants’ needs. These include direct and
indirect participation, addressable bankidentier codes (BICs) and multi-addressee access. Overall,
(EEA), sending payments on their own behalf or on behalf of their customers to around 44,000
to €493.4 trillion in 2013, mainly because of a change in the statistical framework, resulting in some
transactions being excluded from the calculations.
In 2017 the yearly total trafc stood at €432.8 trillion. It has followed a rising trend ever since. In
2013. With the ending of the period for migration to single euro payment area (SEPA) instruments,
belonging to banking communities located in different countries (see Chart A, right-hand panel).
Cross-border trafc has been following a positive trend for the last ten years, both in value and in
allows institutions around the world to exchange euro payments (see also Section 7).
(left-hand panel: x-axis: year; y-axis: yearly total, EUR billions (left-hand side), yearly total, number of payments (right-hand side);
Note: Each colour represents the continent in which the originator bank is located.
RESULTI
FMIs are, by their very nature, interconnected. On the one hand, interconnectedness
positively contributes to reducing the costs and risks associated with transactions,
hand, extensive interlinkages among FMIs may lead to negative effects with the
In order to ensure that the benets of these interconnections outweigh the risks,
PFMI Principle 3 requires an FMI to “have a sound risk management framework
for comprehensively managing legal, credit, liquidity, operational, and other
risks.” In particular, an “FMI should regularly review the material risks it bears
providers, and service providers) as a result of interdependencies and develop
FMIs may be interconnected in several ways. Linkages can arise from direct
from indirect relationships among FMIs through the
common participation of a nancial institution, and from environmental factors,
such as dependence on a common messaging service provider (e.g.SWIFT)
or a third-party IT system service provider. The 2008 CPMI report on the
system-based, institutional and environmental interdependencies respectively.
interconnections and the associated risks, as well as to evaluate the appropriateness
System-based interdependencies arise “from direct cross-system relationships
among two or more systems where the performance of one system relies on the
liquidity and operational risks. Liquidity risks arise when transactions in one
Operational
positions in central bank money. Principle 9 of the PFMIs requires FMIs
commercial bank money but with signicant risk mitigation measures. For
are connected
To analyse vertical system-based interdependencies, the
overview of the number of FMIs active in the system, which are referred to
and geographical area, as well as their trafc. At the end of 2020, 78 ASs
payments classied as AS transactions are included in the calculation.
14 central counterparties (CCPs), two large-value payment systems (LVPSs)
from these connections might spread. The AS daily activity represents a
at 24.6% on average, with a peak at 35.4% in 2014 (see Chart 1, right-hand
panel). In 2020 AS trafc was on average €322 billion per day. Compared
with 2019, the AS trafc showed a decrease of 6%, mainly attributable to
back, the peak in trafc was registered in 2012 (€763 billion) and the lowest
value in 2018 (€303 billion). The steep decrease in 2013 was mainly due to
whereas the more gradual reduction
from 2015 to 2017 was attributable to the migration of central securities
dedicated to the prefunding of positions in the ACHs that process instant
use of tailored indicators, the liquidity patterns arising from their activity
CLS system, as these payments are time-specic and can be very liquidity
Chart 2 shows that in 2020 the average daily value of CLS pay-ins (i.e. the
Horizontal system-based interdependencies also consist of direct relationships
horizontal system-based interdependencies, namely T2S, the platform for
European Central Bank website
of securities transactions. As regards TIPS, participants can pre-fund their
and these two platforms, the horizontal system-based interdependency brings
not only liquidity risk, but also operational risk. While liquidity risk arises
operate, their operational dependency stems from the fact that many of the
events on their respective business days can affect smooth liquidity shifts
Chart 3, left-hand panel). For the remainder of 2017 and until the rst quarter
of 2020, the share of liquidity transfers hovered around 6.4%, while after the
start of the COVID-19 pandemic it reached its peak (7.7% of total TARGET2
trafc value) in July 2020 and then decreased towards the end of the year
(7.0% in December 2020). Exceptionally, in 2020 the liquidity transfers from
compared with 6.5% for 2019. This was a signicant share of the TARGET2
trafc and corresponded to a year-on-year growth of 9.5%, which was mostly
observed in the months following the outbreak of the COVID-19 pandemic.
hand panel of Chart 3 shows the cumulative liquidity transferred
the balance of the transit account rises sharply from €0 to about €75 billion
on average and remains relatively stable throughout the NTS period until the
In general, the efciency of the use of liquidity
transit account’s mean balance. However, in the rst few months of the
COVID-19 pandemic the maximum balance ranges were wider, reaching
occurring through a common nancial institution. In other words, this happens
tools, seen as being most suited to capturing this type of interconnection.
In general, indirect interconnections through a common participant may
lead to contagion effects. This could arise in two ways. First, in the event
of a problem experienced by a nancial institution that has a simultaneous
knock-on effect on more than one FMI. Second, in the event of a problem
experienced by an FMI which then spreads to another FMI through a nancial
institution. This could happen for example when a nancial institution fails to
cover its liquidity requirements vis-à-vis one FMI because the institution has
extremely low. RPSs have the highest number of interconnections (1,131),
participants; they are followed by SSSs (677) and CCPs (672). This result is
driven by the higher number of these AS categories, as shown in Figure 1.
By AS type, the most interconnected are LVPSs, each with an average of 52
connections, closely followed by CCPs with 48 and SSSs with 34. However,
0 (no connectivity) and 1 (maximum connectivity), is very low and stands
at 0.04. This suggests that the risk of contagion effects is low overall and is
are displayed in red, while the nodes coloured differently denote ASs by their type. Only relationships above a daily average of €0.1 billion
and involving ASs with at least 0.1% share of overall AS trafc are displayed.
Environmental interdependencies are indirect relationships that arise from
IQUIDITY
participants, hence the systems, or their participants, may be exposed to
future. PFMI Principle 7 requires an FMI to “effectively measure, monitor,
are available on the participants’ accounts and, irrespective of a possible
the payment recipient and, in turn, result in the recipient not being able to
operator has developed and regularly monitors a wide range of liquidity data
and indicators that look, among others, at the intraday dimensions, both at
have been searching for ways to mitigate the high liquidity requirements.
and liquidity management tools to support effective liquidity usage, reduce
plausible scenarios to assist the system operator in managing liquidity risk. The
IQUIDITY
ITORI
collateral that participants post with their national central banks (NCBs).
increased more than ten times. Chart 5 shows the overall liquidity available
signicantly rose at the time of the sovereign debt crisis in 2011 and 2012,
amid measures taken by the Eurosystem to accommodate banks’ liquidity
demand. Then, the launch of the public sector purchase programme (PSPP)
in March 2015 brought a new surge in liquidity levels. After a period of small
uctuations, liquidity further rose due to the additional stimulus provided
by the Eurosystem in the context of the COVID-19 pandemic through the
pandemic emergency purchase programme (PEPP), standing at €3.2 trillion
Over the same period, the overall value of the available
Eurosystem-wide, almost real-time database to be able to monitor liquidity
Byusing payment-level information, it is possible to calculate the payment
ow and the liquidity available at very granular time intervals. This means that
Through intraday analysis of liquidity available on participants’ accounts, the
make use of the credit line when they do not hold sufcient liquidity on their
has always remained at relatively low levels. During 2020, it decreased even
One important
aspect to monitor for a payment system operator is how specic events or changes
For instance, if a technical problem were to occur towards the end of the day, it would be much easier to handle it in
March 2020, when the COVID-19 pandemic broke out. Payment patterns in
is smoothed out by the end of the morning. An alternative to intraday patterns
is to aggregate the information in a single indicator such as the value-weighted
calculated as the ratio of the actual time payments have spent in the queue
immediately), while a value of one indicates the maximum delay (i.e.payments
as potential free-riding behaviour or the contagion effects of a liquidity shortage.
Moreover, payment patterns deviations that may indicate abnormal situations,
granular level. Chart 9 shows the payment proles derived using a cluster analysis
The chart shows
that participants can be clustered by relatively homogenous intraday payment
The majority of participants (41%) initiate the bulk of their transactions within the
early birds”), whereas only a few participants (16%) send a signicant share of
valuable insights into the effects of a system’s or participant’s operational outages
during the day. Furthermore, by analysing the deviations in intraday patterns, early
Besides using the system’s liquidity management tools to reduce the high
liquidity costs associated with an RTGS system, participants can actively
manage their own payment ows. One way to do this is by synchronising
outgoing and incoming payments, thus making more efcient use of
on three sources of funds: incoming payments, their account balance, and
(Chart 10). Since June 2010, participants have, on average, funded 73.0%
of their payments using the liquidity available on their account balance,
whereas incoming payments have constituted the second source of funding,
covering, on average, 18.8% of the payment outows. Intraday credit has
been used to provide liquidity for 8.1% of outgoing payments. The usage
of the different payment sources has responded to the upsurges in liquidity
increased over time, whereas the use of incoming payments and intraday
Another way to measure how liquidity is used in an RTGS system is to
look at its velocity. Velocity is the value of payments made for each unit of
Note: The chart covers the period from June 2008 to December 2020 at a monthly frequency.
The liquidity used encapsulates the central
central bank reserves resulting from the Eurosystem’s PSPP start in 2015
decreased from 5.1 in March 2015 to 2.99 in December 2020, with most
liquidity used and the liquidity available has steadily increased since then,
Turnover
Note: The chart covers the period from June 2008 to December 2020 at a monthly frequency.
IQUIDITY
only
daily number of liquidity reservations for highly urgent payments uctuated
around 100 and for urgent payments around 50 (see Chart 12). In addition
reservations and that use is limited could be linked to banks’ internal systems
for payment submission, which, de facto, apply their own reservations and
priorities. Furthermore, the cost of actively managing reservations is too high
for smaller banks. An analysis carried out in 2014 quantied the impact of
revealed that reserved liquidity reduced
The “highly urgent” priority is assigned by the system depending on their type; participants can assign the “urgent”
share of highly urgent payments in value (mostly AS trafc) decreased, while
participant (bilateral) or all other participants (multilateral). While these limits
their effect on system performance is ambiguous and is split into rst and
traps on a small number of accounts, given that they redirect the liquidity and
Since 2018 just 10 to 15 participants have used bilateral and multilateral limits
(see Chart 13). Nevertheless, as can be the case with priorities and reservations,
of counterparties, while others use limits only vis-à-vis very few counterparties.
Since 2013, the total bilateral and multilateral limit values have shown little daily
change but have seen notable structural changes. Compared with the liquidity
With limits, the
increases with the introduction of limits. However, given that limits delay
the processing of a payment to just one counterparty, the liquidity can be
show that the initially negative effects of the limits primarily affect those
participants who tend to process their own payments late. These delays
correspond to the desired effect of limits and are therefore not to be
assessed negatively. The outcome suggests a positive outcome overall from
Timed payments give participants the possibility of establishing points in
these are referred to as the “earliest” and “latest” debit times. On the
one hand, the timing of payments gives participants the opportunity to
manage their intraday liquidity.
On the other hand, a high share of timed
not signicantly affected. The number of payments tied to a latest debit time
time was the least used option (about 1,200 payments per day). Overtime,
the usage of this feature has been mainly inuenced by changes in the
The liquidity pooling services allow participants to consolidate the liquidity
has an interest in monitoring the use of liquidity pooling services. At the
aggregated liquidity functionality. Overall, 30 accounts belonged to these
groups. Twenty-four groups with a total of 85 accounts used the consolidated
information functionality. Due to the availability of centralised internal
members. This might explain the relatively low use of the liquidity pooling
whose use and impact on the smooth processing of payments is regularly
monitored by the operator. The current high liquidity levels, changes in
of such features by participants over time. Moreover, participants may
rely on alternative internal liquidity management systems. Thus, while
of the available liquidity management features may regain relevance in
the future, should liquidity levels decrease. Liquidity management features
may also have a high impact on the system’s liquidity use, especially in
times of stress. Their constant monitoring is therefore necessary to meet the
IQUIDITY
advanced LSMs to support the efcient use of liquidity. These consist of
of the performance of LSMs in the system and their ability to effectively
were altered by removing one or more LSMs, were analysed. Four different
scenarios were selected, starting with a scenario close to a plain RTGS
system, i.e. where all LSMs were deactivated (Scenario 1), and then moving
to scenarios where LSMs were progressively reintroduced. First, bilateral
and multiple queue optimisation separately (Scenario 3 and Scenario 4
respectively). This made it possible to test the effectiveness of the individual
the period from 2014 to 2019 in order to account for different liquidity
times are reduced. This holds true, in particular, for the bilateral and
queuing time of customer and interbank payments in the different scenarios.
The average daily queuing time at system level increases in all scenarios,
in particular in the pure RTGS system scenario (Scenario1). The impact is
consistent with the severity of the constraints imposed. For example, in 2017
the median of the daily queuing time in Scenario 1 increased to 5.7 minutes
as compared with 3 minutes in the benchmark simulation, while in Scenario
2 it stood at 4.2 minutes. In line with the previous results, improvements
This may be explained by two factors. First, LSMs process a small amount
extremely high due to the accumulation of excess liquidity, especially over
the last few years of the analysis. The importance of LSMs is expected to
and 75th percentiles, whereas the whiskers mark the minimum and maximum observations within the 25th/75th percentile ± 1.5 times
the interquartile range. Outside values are excluded.
A Eurosystem task force composed of operators and overseers carried out
The exercise consisted
of simulations of stress scenarios with liquidity shortages of different severity
of banks, dened as the sum of positive account balances and intraday
credit lines. Such liquidity constraints could therefore affect the ability of
of different level and type that would lead to a decrease in the intraday credit
obtain an indication of the overall efciency of the system, as well as to test
the resilience of liquidity buffers and liquidity management features under
levels seem to be appropriate and supported by the efcient liquidity
drop of 70% of the collateral prices. The exercise was the rst instance in
which such a stress test was run on an RTGS system using real transactions
and participant data within the replicated system functionalities. Thus, the
Eurosystem has been at the forefront of this type of exercise applied to an
RTGS system. In the light of the relatively comforting results and given
repeated the exercise for the time being. Should a change in the Eurosystem
FMIs are exposed to general business risk, and thus operators are expected
to properly identify, monitor and manage such risk. According to PFMI
Principle 15, general business risk includes “any potential impairment of
the FMI’s nancial position (as a business concern) as a consequence of a
decline in its revenues or an increase in its expenses, such that expenses
exceed revenues and result in a loss that must be charged against capital.
(…) Business-related losses also may arise from risks covered by other
principles, for example, legal risk (in the case of legal actions challenging the
FMI’s custody arrangements), investment risk affecting the FMI’s resources,
and operational risk (in the case of fraud, theft, or loss).” Consequently,
system’s costs are higher or the revenues are lower than initially planned, or
if a one-time loss occurs as a materialisation of another risk (i.e. operational
RECOVERY
generated from the participants’ fees. In 2007, based on assumptions of
the future volumes of operations, a pricing scheme was established that
was aimed at recovery of the costs by April 2014. At the same time, a
public-good factor was reected in the platform’s pricing structure and has
thus been taken into account when assessing the cost recovery situation
Following the overall economic slowdown and
assumptions were revised and, as a result, a limited increase in the users’
fees and an extension of the system’s payback period were implemented
reviews the financial performance of the system. Traffic developments
are monitored both in volume and in value terms, and are accessible to
the Eurosystem through interactive dashboards. These dashboards display
traffic at different aggregation levels, namely time evolution, payment
traffic grew in 2011 and 2012 at an annual rate of +3.3% and +3.5%
respectively (see Chart 17, left-hand panel). A drop of -22.2% in 2013
was attributable to the change in the statistical framework (see Box 1).
Thego
live of T2S in 2015 led to a contraction in the turnover growth
rates, which progressively narrowed until positive rates were registered
again in both 2019 and 2020 (+2.0% and +5.6% respectively). In volume
terms, the range of the growth rates was much smaller, with a minimum
affected by the outbreak of the COVID pandemic.
migration and the go-live of T2S, with the consequent migration of CSDs,
be recovered amounted to €42.5 million. Given that the total revenues
generated amounted to €43.5 million, the resulting annual prot was
€1.0 million. At the end of 2020 the loss accumulated since the launch
, No. 505, European Central Bank, Frankfurt
Incorporating a “public good factor” into the pricing
, No. 507, European Central Bank, Frankfurt am
to potentially fraudulent payments. Fraudulent payments are dened as
authorisation of the holder of the funds. These payments are very difcult to
are able to capture the typical payment behaviour (or pattern) of a participant’s
Theidea behind the business transaction pattern monitoring mechanism
Theindicators focus on customer payments, i.e. payments that are done by
banks on behalf of their customers, given that these transactions are more prone
Some of the indicators take into consideration the total daily values and volumes
receiver that appears in the data is also carefully considered. Finally, as failed
emerging from various attributes of a payment, such as all those listed above,
of normality. Instead, it focuses directly on anomalous data and provides an
Operational risk is a major source of risk for FMIs, which is reected in
PFMI Principle 17. An FMI should “identify the plausible sources of
operational risk, both internal and external (…)”, as, “(…) [f]or example,
participants can generate operational risk for FMIs and other participants,
which could result in liquidity or operational problems within the broader
nancial system”. Within this context, an FMI is required to “(…) identify,
monitor, and manage the risks that key participants, other FMIs, and
the business operator of the central bank with which the bank submitting the payment maintains a business relationship
operator has therefore developed a framework to identify and mitigate
the risks related to its key players, the so-called critical participants (see
individual participants. When operational risk materialises in the form of a
the incident’s impact (see Section 6.2). Moreover, it is also in the interests
TIFIC
CRITIC
RTICIP
to the system. Under PFMI Principle 17, the identication should be “(…)
based on the consideration of transaction volumes and values, services
provided to the FMI and other interdependent systems, and, more generally,
the potential impact on other participants and the system as a whole in the
event of a signicant operational problem”. The identication of critical
operational risk management requirements and mitigation measures.
This provides
reasonable assurance that the information security and cyber resilience of
their internal systems are appropriately addressed, that business contingency
and business continuity measures are in place and tested, and that outages
While the PFMIs give a general indication of characteristics that may make a
Given that a denition or explicit economic characteristic of a critical
participant does not exist, the data are, by default, unlabelled, and labels
cannot be back-tested. The goal of the identication is therefore to assign to
done based on indicators or statistics combined with expert knowledge,
due diligence checks, and practicability considerations. Theidentication
namely credit institutions, ASs and third-party service providers. For ASs,
the identication challenge has been resolved by focusing on the type.
In principle, LVPS, systemically important RPS, CSDs, international CSDs
, Version 15.1, European
(ICSDs) and CCPs are classied as critical. For third-party service providers,
which comprise the SWIFT Service Bureaus (SSB) and Group Hubs (GH),
The identication process is more complex for credit institutions and
requires the combination of two criteria, namely the turnover generated by
a participant and the impact of a simulated technical failure of a participant
indicators and thresholds; participants are deemed critical if at least one
rank participants. In a sequential step, simple hard thresholds are applied.
Participants with a value greater than the threshold full the criterion and
of the indicators and on the judgement of experts with the background
and experience to evaluate risks. The combination of the two criteria,
The classication also incorporates a time dependency element to avoid
frequent reclassications. A participant is categorised as critical only
if the condition of criticality is fullled two years in a row. Similarly, a
in a row. Thisrule is aimed at reducing the volatility in the group of
critical participants by avoiding frequent reclassications arising from
temporary uctuations and giving critical participants the constancy they
in extraordinary circumstances, such as an organisational change in the
The rst criterion measures the criticality of the participants by looking at
the turnover in terms of value they generate in the system, in line with the
approach suggested in the PFMIs. A participant is considered to be critical
Since the identication of critical participants, as laid down
in the PFMIs, focuses on the potential impact of an operational failure of
a participant, the relevant trafc for the computation of the rst criterion
encompasses solely those payments that are actively initiated by the
participant. Payments where the participant is debited but that are initiated
caused by the largest participants failing. Hence, focusing solely on the trafc
share of a participant neglects potential contagion effects in the system. This
contagion is often referred to as a “liquidity sink”: a participant with an
operational failure still receives payments but is unable to send payments,
i.e. the liquidity it receives is no longer available in the system but disappears
into the “liquidity sink”. This lack of liquidity can lead to the system not
incoming liquidity from the failed participant to fund their own payments.
The second criterion addresses the more general denition of criticality as
potential impact of an operational failure on other participants and on the
system. The operational failure of a potential critical participant is simulated
all the payments sent by the participant over an entire day, except for ancillary
system payments debited from its account that are sent by the AS itself and
changes in the participant’s ICL. Thepayments of all other participants remain
unchanged. By assuming that no behavioural reactions and mitigation measures
at participant level take place, the simulation scenarios represent a “worst-case”
scenario. For each participant, independent scenarios for ten business days
are generated and simulated. The simulation results are evaluated in terms of
payments on average in value terms in the period analysed (see Chart 18).
While the threshold check itself is as straightforward as for the rst criterion, the
indicator calculation is more complex given that a large number of simulation
scenarios have to be generated and executed. Toreduce the number of scenarios,
effects, caused by payments not sent due to the operational failure, and
their payments. While the rst-round effect is conceptually identical to the
turnover of the rst criterion, the second-round effect precisely measures
the potential contagion not considered by the rst criterion. The scenario
years, simulation scenarios have been generated and executed for 32 to 37
participants each year, while the nal group of critical credit institutions
after the combination of the two criteria consists of 19 to 21 participants.
Overall, the “simple” indicator of the rst criterion identies the majority
of critical participants, whereas the second criterion captures a lower,
although not negligible, number of critical participants. In addition, the data
generated from the simulation results can be further analysed to increase
understanding of the diversity of contagion channels and the importance of
non-linear effects. One of the main advantages of this analysis is to focus
ADDITIONAL ANALYSIS OF SIMULATED OPERATIONAL FAILURE
analysed at a more granular level by looking at daily simulation results instead of averages. Thismakes
The contagion effect mainly occurs if some participants affected by an outage generate large
second-round effects relative to the rst-round effects they suffer from, i.e. the lack of incoming
payments from a potential critical participant is relatively small compared to the amount
can be considered to be catalysts, given that they spread the initial shock of the operational
outage further through the system. Such participants are not critical in the sense that their
own operational outage generates risks, but the impact on other participants or on the system
Each node represents a participant, and each link is sized to the amount of payments that
forthcoming from the critical participant (CP) in the rst round (shown in blue). At the same time,
in yellow) as compared with participant A and thus acts as a catalyst. From a risk perspective,
the identication of participants amplifying the effect of another participant’s failure makes it
not done as a regular exercise, but rather as a case-by-case analysis if other indicators signal
Source: TAG.
post analysis of the root causes of the incident, of the measures taken
and of the impact on the participants. The latter is assessed by making a
the incident day and on the evolution of rejection rates. This is done in
Then, the pattern of intraday
day for each payment type – typically customer payments, interbank
payments, AS transactions and T2S and TIPS liquidity transfers – and at
interactions with T2S and TIPS in the light of their liquidity interdependency
overall transfers to T2S are included in the chart in the right-hand panel.
have been effective in containing and addressing the situation. The pattern
intervals provides additional insights into pattern deviations at specic points
RTICIP
The absence of evidence and the evidence of absence: an
of the critical participants. This serves to monitor the risk of operational
outages affecting participants and to evaluate the performance of the
algorithm. The algorithm results are cross-checked against the reported
participants – as only they are obliged to report signicant outages to
data for operational outages may stem from the fact that the algorithmic
approach inherently entails uncertainty or the fact that the outages were
not reported. The reason for developing the approach in the rst place
was the absence of comprehensive and reliable information on operational
outage occurrences. Besides the operational incidents communicated, the
CBs use expert judgement, reporting tools, information received, as well
as following up with participants to evaluate potential outages identied.
In2020 the algorithm identied a total of 72 days with a potential operational
outage. On average, potential operational outages lasted for 6.6 ten-minute
intervals, i.e. around one hour. There were several shorter potential outages
of less than one hour (26) and a few that were longer and lasted for more
than two hours (6). The longest outage lasted 18 ten-minute intervals or
The results of the quarterly exercises suggest that the algorithm works
very well in most circumstances, and especially in the case of long-lasting
outages. The more consecutive intervals with low payment activity that
outage, although, by construction, identication is difcult when the
other jurisdictions.
In the future, such tools could be employed to monitor
RTICIP
Tiered participation arrangements occur when a participant in an FMI offers
other institutions that are not participants in the FMI itself the possibility
Principle 19, “tiered participation arrangements occur when some rms
(indirect participants) rely on the services provided by other rms (direct
recording facilities”. On the one hand, tiered participation arrangements
may be seen as benecial for the nancial system given that they ensure
wider access to the services of an FMI, fostering the inclusion of smaller
banks that might not be able to afford direct participation or providing
arrangements pose certain risks for the FMI and its functioning. These risks,
including credit, liquidity and operational risks, are particularly relevant
when the degree of tiering in the FMI is high. In other words, the more
signicant (i) the proportion of the total FMI’s trafc originated by indirect
participants, and (ii) the concentration of the trafc originated by indirect
participants in the books of a few direct participants are, the more relevant
Tiered participation arrangements may pose credit, liquidity and
operational risks. Credit and liquidity risks created by tiered participation
and indirect participants. The FMI operator is not expected to play an active
role in managing them. It may, however, apply credit or position limits,
in agreement with the direct participant. In contrast, operational risk is
relevant for the FMI operator, as the disruption of a direct participant may
affect the capacity of all its indirect participants to channel payments to
to many indirect participants, and if the trafc originating by these indirect
arrangements and the related risks. The analysis is conducted on a yearly
generated by tiered participants and the number and distribution of tiered
relationships, both at aggregate and at participant level. PFMI Principle 19
states that “an FMI should identify, monitor and manage the material risks
to the FMI arising from tiered participation arrangements”. Inparticular,
this means that the FMI should gather information about indirect
participation, identify indirect participants responsible for a signicant
proportion of the transactions processed in the FMI, or of the transactions
participation arrangements and take mitigating actions whenever necessary.
focuses mainly on tiering on the sending side given that this is considered
to pose greater risks due to the potential spill-overs to other participants
Tiered participation arrangements can occur on both the debit and the credit
and receiver with its originator and beneciary and excluding intragroup
with the Bank Directory Plus data from SWIFT to identify and exclude
intragroup transactions, i.e. those involving BICs (“tiered” BICs and “direct”
payment is considered to be tiered on the sending side only if the originator
and the sender belong to a different banking group. The same logic applies
on the receiving side (the beneciary and receiver belong to a different
to be tiered if the originator parent and the beneciary parent are the same,
terms over time, whereas it has displayed more variation in volume terms.
receiving sides (see Chart 21). Over the same period, tiering levels in volume
more frequently on the sending side than they ask to receive on their behalf
and, second, that the average tiered payment size is larger on the receiving
than on the sending side. Compared with the value, tiering in volume was
slightly more volatile across both the sending and receiving sides in terms of
historical developments. Monitoring of the aggregate levels of tiering makes
in value terms, whereas customer payments are predominant in volume
Chart 22). On the volume side, customer and interbank payments sent by
Banking groups headquartered in the United States were the top contributors
Out of all the banking
groups that asked direct participants to send payments on their behalf in
value terms (see Chart 23, left-hand panel). From outside the EEA, Switzerland,
Canada, Japan and China also contributed to tiered trafc, but to a lesser
extent. Groups located inside the EEA accounted for 21.0% of the total value.
The most active groups offering tiering services in 2020 were located in
Germany. Of all the banking groups offering to send payments through
were the most significant with 74.2%, followed by France with 13.7%
and Belgium with 6.1% (see Chart 23, right-hand panel). These results
are linked to many factors, including the characteristics of the national
banking systems, namely its size and the presence of banking groups’
headquarters, and the fact that correspondent banking services is a
large number of direct participants in 2020, whereas the majority of groups
used a single access point. A total of 4,768 indirect groups (corresponding
through another institution in 2020 (see Chart 24). The share of indirect
was less than 20%, while more than a third of the indirect groups connected
, European Central Bank, Frankfurt am
Note: The percentages are in terms of total value.
measure, especially in the event that one of the direct participants
failure of a direct participant. In this regard, it should be underlined that tiered
operator has developed, over time, dedicated frameworks combining qualitative
and quantitative elements. The latter are used to assess the system’s compliance
with specic regulatory provisions for which risk assessment needs to be supported
by granular data analysis. This is in particular the case when assessing risks arising
from interdependencies (Principle 3 on the comprehensive management of risks),
analysing liquidity risk (Principle 7), monitoring general business risk (Principle
15), identifying the system’s critical participants (Principle 17 on operational risk)
well as more complex exercises using compound indicators and specic tools,
Data analysis supports two dimensions of compliance with the general
general denition of criticality as the potential impact of an operational failure
general indication of characteristics that could make a participant critical,
entire payment chain and comparing the sender and receiver of a payment
of tiering mainly on the sending side, given that this is considered to pose greater
risks due to the potential spill-overs that might arise if liquidity did not reach
the recipients of tiered participants’ payments. In value, tiering stood at 6.6%
payments. The risks associated with tiering are mainly linked to the value of
Although developed with the objective of regulatory compliance, these tools
have been important for monitoring the system and gaining knowledge and
understanding of the system’s activity over time. These indicators and the
related studies offer important insights into trafc patterns, system efciency,
the effectiveness of various system features, liquidity ows, payment patterns,
the behaviour of individual participants and their interconnections. They have
and, in exceptional circumstances, to deepen analysis and monitor the impact
of specic events, such as incidents or functional changes in the system.
developments, its access rules and interlinkages. They have also been key to more
general discussions around European FMIs, such as the recent pan-European
strategy for instant payments and preparation of the T2-T2S consolidation. For
it is therefore of the outmost importance to keep investing in these tools with a
indispensable. To this end, a new analytical environment with enhanced
to support payment-level data analysis by the operator and the overseer of
Liquidity Management Features
account, payments are queued according to their priority. Participants can choose from the
priorities “normal”, “urgent” or “highly urgent”. The default priority is “normal”. The priority
“highly urgent” is only allowed for certain payment types, such as FMI-related payments and
Reservations
cumulated payment amount that a participant is willing to pay to another participant (bilateral)
received payments (that are credits) rst.
Timed payments
transaction will either remain in the queue or be rejected (the so-called latest debit time).
using the service, the participant has to specify a hierarchical structure for the accounts to
be included in the liquidity pooling arrangement. The manager of a group of accounts has
the possibility of either viewing the liquidity position of all group accounts simultaneously,
thereby having the benet of information on the aggregated liquidity positions of the whole
Active queue management
queued payments. They can change the priority, the earliest and/or latest time stamp or the
order in the queue of each queued payment. They may also revoke payments.
Liquidity Saving Mechanisms
Calculates, for each participant, those payments that can be executed in compliance with the
participant’s bilateral and multilateral limit position from all queued payments and all priorities.
Works in a similar way to ALGO1, but is able to deallocate payments (i.e. keep them in
liquidity position, the aim being to turn those negative positions into positive positions. It
can also end in payments being unsuccessful if bilateral/multilateral limits are breached or
positions are not covered.
demand. It consists of two parts, one bilateral and the other multilateral. It can also end in
payments being unsuccessful if bilateral/multilateral limits are breached or positions are not
covered.
partial optimisation with ancillary
transactions in its runs.
liquidity accounts) only.
Note: Algorithms 1-3 are “time triggered”, while algorithms 4 and 5 are “event triggered”.
Ancillary system (AS)
requirements of the location of the infrastructures offering services in euro, as amended from time to time
and published on the ECB’s website, in which payments or nancial instruments are exchanged and/or
Central counterparty (CCP)
Central securities depository (CSD)
entry transfer of securities. In addition to safekeeping and administration of securities, a CSD may also
Credit line
Maximum collateralised overdraft position of the balance on a payment module (PM) account. The PM
account holder can obtain information about changes in the credit line through the information and control
module (ICM). Changes in credit lines will be executed immediately. In the event of a reduction in a credit
line, the change is given “pending” status if the reduction would lead to an uncovered overdraft position. The
change will be executed when the overdraft position is covered by the reduced credit line.
Dedicated transit account
A cash account in the RTGS system and in T2S that is held and used by the system operator concerned to
transit account and the transit account opened within the RTGS system is referred as the T2S dedicated
transit account.
Intraday credit
Credit extended and to be reimbursed within a period of less than one business day; in a credit transfer
accepts and acts on a payment order even though it will not receive the nal funds until the end of the
business day. The credit may take the form of a collateralised overdraft or of a lending operation against
collateral pledge or established under a repurchase agreement.
Intraday liquidity
Funds that may be accessed during the business day, usually to enable nancial institutions to make
payments on an intraday basis.
Infrastructures (PFMI)
An arrangement whereby transfer orders are kept pending by the sending direct participant or by the system
insufcient funds or quantitative limits on the funds.
subsequently takes place through a payment system or through correspondent banking relationships.
provides services for central securities depositories (CSDs) and central banks that make it possible to
basis in central bank money.
(TIPS) service
Technical account
An account used for ancillary system (AS) operations as an intermediary account for the collection of debits/
credits (or debits, as the case may be) of an overall equal amount.
Arjani, N. and Heijmans, R. (2020), “
Diehl, M.and Müller, A.(2014), “
ECB takes steps to ensure pan-European reach of instant payments
”, MIP News, European Central
Garratt, R., Antoine, M. and McAndrews, J. (2014), “
Turnover in Fedwire Funds Has Dropped
Considerably since the Crisis, but it’s Okay
Kaliontzoglou, A. and Müller, A. (2016), “
Implementation aspects of indicators related to payments
D. (2021), “
Ancillary system (AS)
oversight requirements of the location of the infrastructures offering services in euro, as amended
from time to time and published on the ECB’s website, in which payments or nancial instruments
and the relevant Eurosystem central bank.
Central counterparty (CCP)
Central securities depository (CSD)
book-entry transfer of securities. In addition to safekeeping and administration of securities, a CSD
Credit line
Maximum collateralised overdraft position of the balance on a payment module (PM) account. The
PM account holder can obtain information about changes in the credit line through the information
and control module (ICM). Changes in credit lines will be executed immediately. In the event of
a reduction in a credit line, the change is given “pending” status if the reduction would lead to an
uncovered overdraft position. The change will be executed when the overdraft position is covered by
Dedicated transit account
A cash account in the RTGS system and in T2S that is held and used by the system operator
RTGS dedicated transit account and the transit account opened within the RTGS system is referred as
the T2S dedicated transit account.
Intraday credit
Credit extended and to be reimbursed within a period of less than one business day; in a credit
institution if it accepts and acts on a payment order even though it will not receive the nal funds
until the end of the business day. The credit may take the form of a collateralised overdraft or of a
lending operation against collateral pledge or established under a repurchase agreement.
Intraday liquidity
Funds that may be accessed during the business day, usually to enable nancial institutions to make
payments on an intraday basis.
Infrastructures (PFMI)
An arrangement whereby transfer orders are kept pending by the sending direct participant or by the
submission due to insufcient funds or quantitative limits on the funds.
throughout the world. A SWIFT payment message is an instruction to transfer funds; the exchange
Eurosystem provides services for central securities depositories (CSDs) and central banks that make
versus payment (DvP) basis in central bank money.
(TIPS) service
Technical account
An account used for ancillary system (AS) operations as an intermediary account for the collection of
The balance of this account is always zero because debits (or credits, as the case may be) are always
Acknowledgements
suggestions, and Ioana Duca-Radu for the valuable contributions she provided. We are also grateful
in accordance with Article 1(2) of Decision ECB/2010/9 of 29 July 2010 on access to and use of
Marino, Luca Mibelli, Laura Ricciardi and Giovanni M. Sabelli
by Mauro Arcese, Domenico Di Giulio and Vitangelo Lasorella
Green Bonds: the Sovereign Issuers’ Perspective,
by Raffaele Doronzo, Vittorio Siracusa and
central bank money,
by Cristina Mastropasqua, Alessandro Intonti, Michael Jennings, Clara
Mandolini, Massimo Maniero, Stefano Vespucci and Diego Toma
analysis with Bitcoin and other infrastructures,
Proposal for a common categorisation of IT incidents,
Bundesbank, European Central Bank, Federal Reserve Board, Financial Conduct Authority,
Ministero dell’Economia e delle Finanze, Prudential Regulation Authority, U.S. Treasury
Inside the black box: tools for understanding cash circulation,
by Luca Baldo, Elisa Bonifacio,
Marco Brandi, Michelina Lo Russo, Gianluca Maddaloni, Andrea Nobili, Giorgia Rocco,
Gabriele Sene and Massimo Valentini
The impact of the pandemic on the use of payment instruments in Italy,
by Guerino Ardizzi,
Alessandro Gambini, Andrea Nobili, Emanuele Pimpini and Giorgia Rocco
by Paolo Bramini,
Valentini
n. 10
A digital euro: a contribution to the discussion on technical design choices,
by Emanuele Urbinati,
Alessia Belsito, Daniele Cani, Angela Caporrini, Marco Capotosto, Simone Folino, Giuseppe
SSUES
n. 11
From SMP to PEPP: a further look at the risk endogeneity of the Central Bank,
by Marco
n. 12
TLTROs and collateral availability in Italy,
by Annino Agnes, Paola Antilici and Gianluca
Overview of central banks’ in-house credit assessment systems in the euro area, by L
Auria, Markus Bingmer, Carlos Mateo Caicedo Graciano, Clémence Charavel, Sergio Gavilá,
Alessandra Iannamorelli, Aviram Levy, Alfredo Maldonado, Florian Resch, Anna Maria Rossi
n. 14
The strategic allocation and sustainability of central banks’ investment, by
Davide Di Zio,
n. 15
Climate and environmental risks: measuring the exposure of investments, by
Ivan Faiella
Johnny Di Giampaolo
and Davide Nasti
PUBLISHED
‘
RKETS
STRUCTURES
YSTEMS
’
SERIES
n. 16
Mechanism, by
Fabrizio Dinacci and Ann Börestam
n. 17
Daniela Della Gatta
n. 18
and Antonino Fazio
n. 19
Francesco Di Stasio, Carlo Palmers, Faisal Alhijawi, Erol Kaya, Christophe Piccarelle, Stuart
Butler, Jwallant Vasani, Giancarlo Esposito, Alberto Tiberino and Manfredi Caracausi
n. 20
n. 21
SSUES
n. 22
by Paola Antilici, Gianluca Mosconi and Luigi Russo
n. 23
by Guerino
Ardizzi and Massimiliano Cologgi
n. 24
Press news and social media in credit risk assessment: the experience of Banca d’Italia’s
house Credit Assessment System,
by Giulio Gariano and Gianluca Viggiano
The bonre of banknotes,
by Michele Manna
n. 26
Matteo Caruso, Stefano Cossu, Giuseppe Galano, Simone Mancini, Gabriele Marcelli, Piero
Martella, Matteo Nardelli and Ciro Oliviero
n. 27
by Guerino Ardizzi and Alessandra Righi
n. 28
TIPS: a zero-downtime platform powered by automation, by
Gianluca Caricato, Marco
by Marc Glowka, Alexander Müller, Livia Polo Friz, Sara Testi,
Massimo Valentini and Stefano Vespucci