EMIR and SFTR data quality report 2021

EMIR and SFTR data

quality report 2021

1 April 2022 ׀ ESMA74-427-607

EMIR and SFTR data quality report 2021 2

EMIR and SFTR data quality report 2021    3 
EMIR and SFTR data quality report 2021 
 
Table of contents 
 Table of contents............................................................................................................................ 3 
 Executive summary ....................................................................................................................... 4 
 ESMA’s Strategic Priorities on EMIR and SFTR Data Quality ................................................... 7 
 EMIR and SFTR Data Quality Frameworks .................................................................................. 9 
1  EMIR DQAP ............................................................................................................................. 9 
2  SFTR DQEF ........................................................................................................................... 10 
3  Cooperation with data users .................................................................................................. 11 
 Recent developments impacting EMIR and SFTR data quality............................................... 14 
1  EMIR - TR data ingestion review ........................................................................................... 14 
2  EMIR regulatory access filtering review................................................................................. 15 
3  EMIR TAR-TSR review .......................................................................................................... 16 
4  SFTR – Implementation of new XML schemas ..................................................................... 18 
5  SFTR – UnaVista wind-down ................................................................................................. 18 
 EMIR reporting trends and selected data quality metrics ....................................................... 20 
1  Data reporting – Key trends ................................................................................................... 20 
2  Data completeness, timeliness and availability ..................................................................... 22 
3  Data integrity – Adherence to format and content rules ........................................................ 26 
4  Data integrity – Reconciliation ............................................................................................... 27 
 SFTR reporting trends and selected data quality metrics ....................................................... 30 
1  Data reporting – Key trends ................................................................................................... 30 
2  Data completeness, timeliness, and availability .................................................................... 32 
3  Data integrity – Adherence to format and content rules ........................................................ 34 
4  Data integrity – Reconciliation ............................................................................................... 35 
 Methodological Annex ................................................................................................................. 39 
 List of abbreviations .................................................................................................................... 42 
   

EMIR and SFTR data quality report 2021 4

2. Executive summary

This is the second edition of report on data quality under the European Market Infrastructure Regulation

(EMIR) and under the Securities Financing Transactions Regulation (SFTR). The objective of the report

is to provide a holistic view of state of play of both reporting regimes as regards the quality of the

reported data and the actions that the national competent authorities (NCAs’) and the European

Securities and Markets Authority (ESMA) are taking to improve the quality of the data. EMIR and SFTR

data play a pivotal role in the fulfilment of NCAs’ and ESMA’s supervisory mandates. As such, the data

is used extensively for those purposes

. Key highlights of the report are as follows:

ESMA strategic priorities on EMIR and SFTR data quality

As regards ESMA’s supervision of TRs, key areas of focus are:

i. Timely and complete reporting of regulatory information to the users of TR data,

ii. Accuracy and confidentiality of data reported by counterparties to and stored by TRs, and

iii. Accuracy of regulatory reports submitted to the users of TR data.

As regards the reporting by counterparties, the key common areas of priority for NCAs and ESMA are:

i. Completeness and accuracy of the reported information, in particular with regards to the

reporting of valuation and collateral data,

ii. Timely submission of the reports, and

iii. Consistency of reported information reflected in the reconciliation of data submitted by the two

counterparties of the same derivative.

Counterparties are strongly encouraged to use the regulatory data in their own internal risk and

compliance management processes. In doing so, counterparties incentives to report accurate data will

be further aligned.

ESMA, with the cooperation of the NCAs, is and will continue to monitor progress in those areas. ESMA

and the NCAs will take actions with the objective to achieve improvement in areas where insufficient

quality of the data is identified.

Recent developments impacting EMIR and SFTR data quality

Regarding EMIR, ESMA carried out three supervisory projects focusing on i) the ingestion and

processing of data by TRs, ii) application of EMIR access filtering rules for provision of data to NCAs

according to their mandates and iii) the assessment of consistency of two key regulatory reports – the

trade activity and trade state reports. In all three cases, ESMA found that TRs broadly follow regulatory

and supervisory expectations. In certain instances, ESMA found some shortcomings in the quality of

the reports provided to regulators

and expects that TRs take appropriate remediation steps. As regards

SFTR, TRs and reporting counterparties implemented first SFTR XML schema update since the start

of reporting in July 2020. The update aimed at removing technical shortcoming that could decrease

quality of the information available to the regulators. ESMA also monitored and coordinated with

NCAs/TRs all relevant aspect of the wind-down of UnaVista repository services under SFTR as well as

the associated porting of SFTR data to other TRs

See for example the Trends Risks and Vulnerabilities report (https://www.esma.europa.eu/sites/default/files/library/esma50-

165-1842_trv2-2021.pdf) and the Annual Derivatives Statistical Report (

https://www.esma.europa.eu/sites/default/files/library/esma50-165-2001_emir_asr_derivatives_2021.pdf)

Such as under-/over-reporting of data in the case NCA access filtering and report accuracy issues in the case of the trade

activity and trade state reports.

EMIR and SFTR Data Quality Frameworks

Data quality under EMIR and SFTR relies on an efficient supervision of the reporting counterparties by

the NCAs and of the TRs by ESMA. Having in mind these complementary supervisory responsibilities,

NCAs and ESMA have established the following dedicated frameworks to coordinate the joint efforts

on ensuring high data quality: the Data Quality Action Plan (DQAP) under EMIR and the Data Quality

Engagement Framework (DQEF) under SFTR.

Under EMIR DQAP the NCAs analysed and followed up with selected supervised entities on the results

of over 30 tests related to different data quality aspects such as completeness, accuracy, or timeliness.

In a thematic review focused on the reporting of valuations and collateral, misreporting of valuations

was significantly reduced, as compared to the previous year, by around 50% of the targeted entities.

Similarly, in a dedicated exercise on timeliness of reporting, most of the targeted entities improved their

reporting practices and eliminated or significantly reduced late reports.

SFTR DQEF was launched in 2021 and focused on the timeliness of reporting, rejections, and pairing.

While the follow-up on this first iteration of DQEF is still ongoing. Overall, it can be noted that use of the

ISO20022 XML end-to-end reporting has brought important benefits in terms of the quality and

accessibility of the data from the very beginning of the reporting. Furthermore, some positive trends in

key metrics such as rejections and reconciliation can already be observed. Recent developments

impacting EMIR and SFTR data quality

Regarding EMIR, ESMA carried out three supervisory projects focusing on i) the ingestion and

processing of data by TRs, ii) application of EMIR access filtering rules for provision of data to NCAs

according to their mandates and iii) the assessment of consistency of two key regulatory reports – the

trade activity and trade state reports. In all three cases, ESMA found that TRs broadly follow regulatory

and supervisory expectations. In certain instances, ESMA found some shortcomings in the quality of

the reports provided to regulators and expects that TRs take appropriate remediation steps. As regards

SFTR, TRs and reporting counterparties implemented first SFTR XML schema update since the start

of reporting in July 2020. The update aimed at removing technical shortcoming that could decrease

quality of the information available to the regulators. ESMA also monitored and coordinated with

NCAs/TRs all relevant aspect of closely the wind-down of UnaVista repository services under SFTR as

well as the associated porting of SFTR data to other TRs.

EMIR reporting trends and selected data quality metrics

Brexit has had an important impact on the EU supervisory data reporting landscape as volumes of

reported derivatives fell by approximately 50%. In terms of data reporting volumes, equities and futures

contracts continue to be the most prominent asset class and contract type respectively. While less than

10% of reported derivatives tend to be reported late by the counterparties, more than 20% do not receive

updated valuation on a daily basis as required by EMIR. Non-reporting dropped sharply due to Brexit

and is now less than 5%. The sharp drop has been driven by the end of the reporting obligation of UK

counterparties and the more limited dual-side reporting which does not allow to detect potential non-

reporting issues. TR rejections continue to be low at around 2%. Furthermore, only 1% of records in TR

regulatory reports seem to not comply with the applicable validation rules. Volumes of duplicated

reporting

are negligible. As regards reconciliation, pairing rate continues to be relatively low at 60%

while there is on average 5% difference between the number of open derivatives reported between a

pair of counterparties. Lastly, in some instances, TRs disagree on the number of derivatives they

reconcile against each other. This may be an indication of further enhancements required for the inter-

TR reconciliation process. While much attention has already been put to timely reporting, reporting of

Reported derivatives are considered duplicated where two or more records have been reported with the same combination of

reporting counterparty ID, ID of the other counterparty and trade ID fields. In contrast, double-sided reporting under EMIR is not

duplicative since two records on the same derivative should be reported always from the perspective of the respective reporting

counterparty.

EMIR and SFTR data quality report 2021 6

valuations and reconciliation, clearly much more improvements are needed and those area s will

continue to be point of focus of ESMA and NCAs going forward.

SFTR reporting trends and selected data quality metrics

Similar to EMIR, data reporting volumes dropped approximately by 50% following Brexit. In terms of

number of open transactions, securities lending and borrowing is the largest SFT type reported with

around 70% share at the end of 2021. Credit institutions report most open SFTs (around 50%) while

credit institutions share has been increasing (to around 30% at the end of 2021). After merely 1.5 years

of reporting, SFTR exhibits comparable results to EMIR across all data quality metrics. Around 10% of

SFTs are reported late (after T+1). On the contrary, rejections have been low (around 2%) and

duplicated reporting does not pose major issues. As regards reconciliation, pairing rate has been only

around 60%. Reconciliation rate of loan and collateral data has been low but increasing to around 40%

and 30% respectively. Similar to EMIR, TRs do not agree on the number of records they reconcile

against each other, which may be an indication of issues in the inter-TR reconciliation process.

Timeliness of reporting, adherence to format and content rules (via rejections) and reconciliation

(pairing) has been the point of focus of ESMA and NCAs during 2021. While progress has been made,

some areas (particularly reconciliation) need to remain areas of focus also in the future.

3. ESMA’s Strategic Priorities

on EMIR and SFTR Data

Quality

Summary: As regards ESMA’s supervision of TRs, key areas of focus are: i) timely and complete

reporting of regulatory information to the users of TR data, ii) accuracy and confidentiality of data

reported by counterparties to and stored by TRs, and iii) accuracy of regulatory reports submitted to the

users of TR data.

As regards the reporting by counterparties, the key common areas of priority for NCAs and ESMA are:

i) Completeness and accuracy of the reported information, in particular with regards to the reporting of

valuation and collateral data, ii) timely submission of the reports, and iii) consistency of reported

information reflected in the reconciliation of data submitted by the two counterparties of the same

derivative. Counterparties are strongly encouraged to use the regulatory data in their own internal risk

and compliance management processes. In doing so, counterparties incentives to report accurate data

will be further aligned.

ESMA, with the cooperation of the NCAs, is and will continue to monitor progress in those areas. ESMA

and the NCAs will take actions with the objective to achieve improvement in areas where insufficient

quality of the data is identified.

TR supervisory objectives: ESMA is the direct

supervisor of TRs under EMIR and SFTR. ESMA

sets its supervisory priorities on an annual basis

and publishes them in ESMA’s annual work

programme.

ESMA is a data-driven and risk-based supervisor.

Thus, it sets its priorities based on risks it

observes which may negatively impact quality of

the reported data. The most prominent risks are

then included in the list of its annual priorities in

the form of a specific project or a supervisory

review. Besides one-off projects, ESMA also

performs a variety of monitoring activities

on an

ongoing basis.

Even though ESMA’s priorities may evolve from

one year to another, there are common themes

that remain present over time.

Those themes are:

1. Timely and complete reporting of

regulatory reports to the users of TR data;

For example: monitoring timeliness and completeness of TR

daily regulatory report submissions.

See Subsection 5.1. on EMIR TR data ingestion review

carried out through 2020 and 2021.

2. accuracy and confidentiality of data

reported by counterparties to and stored

by TRs

; and

3. accuracy of regulatory reports submitted

to the users of TR data

ESMA expects that the TRs pay utmost attention

to the above-mentioned aspects and that they

have processes, systems and controls in place to

monitor and timely identify any issues.

Counterparty reporting supervisory objectives:

NCAs are responsible for the supervision of the

reporting by the counterparties while ESMA

coordinates some key common initiatives in the

context of its supervisory convergence mandate

While common priorities are also set annually,

there are key areas of permanent focus by NCAs

and ESMA:

1. Completeness and accuracy of the

reported information, in particular with

See Subsection 5.2. on EMIR regulatory access filtering

review and Subsection 5.3. EMIR TAR-TSR consistency both

carried out in 2021.

See Section 0.

EMIR and SFTR data quality report 2021 8

regards to the reporting of valuation and

collateral data;

2. timeliness of the reports; and

3. consistency of reporting reflected in the

reconciliation of data reported by the two

counterparties of the same derivative.

Reporting counterparties are expected to have

processes, systems and controls in place to

ensure completeness, accuracy and timeliness of

the reported information. Furthermore, they are

expected to actively engage in detecting and

resolving any identified report rejections,

reconciliation breaks and other data quality

issues in the already reported data.

Counterparties are strongly encouraged to use

the regulatory data in their own internal risk and

compliance management processes. In doing so,

counterparties will have the appropriate

incentives to report accurate data and will be in

apposition to better exploit the benefits of

consistent data reporting.

What we aim to achieve: ESMA, with the

cooperation of the NCAs, is and will continue to

monitor progress in those areas. ESMA and the

NCAs will take actions with the objective to

achieve improvement in areas where insufficient

quality of the data is identified.

4. EMIR and SFTR Data

Quality Frameworks

Summary: Data quality under EMIR and SFTR relies on an efficient supervision of the reporting

counterparties by the NCAs and of the TRs by ESMA. Having in mind these complementary supervisory

responsibilities, NCAs and ESMA have established the following dedicated frameworks to coordinate

the joint efforts on ensuring high data quality: the Data Quality Action Plan (DQAP) under EMIR and the

Data Quality Engagement Framework (DQEF) under SFTR.

Under EMIR DQAP the NCAs analysed and followed up with selected supervised entities on the results

of over 30 tests related to different data quality aspects such as completeness, accuracy or timeliness.

In a thematic review focused on the reporting of valuations and collateral, misreporting of valuations

was significantly reduced, as compared to the previous year, by around 50% of the targeted entities.

Similarly, in a dedicated exercise on timeliness of reporting, most of the targeted entities improved their

reporting practices and eliminated or significantly reduced late reports.

SFTR DQEF was launched in 2021 and focused on the timeliness of reporting, rejections and pairing.

While the follow-up on this first iteration of DQEF is still ongoing, overall, it can be noted that use of the

ISO20022 XML end-to-end reporting has brought important benefits in terms of the quality and

accessibility of the data from the very beginning of the reporting. Furthermore, some positive trends in

key metrics such as rejections and reconciliation can already be observed.

4.1 EMIR DQAP

EMIR Data Quality Action Plan (DQAP) the

DQAP is a major project that NCAs and ESMA

jointly launched in September 2014. It aims at

improving the quality and usability of data that is

reported by counterparties and made available by

the TRs.

The DQAP encompasses activities related to the

policy work, NCAs’ supervision of the reporting

counterparties and ESMA’s supervision of the

TRs, to address the potential issues in all areas

that are key for the quality of the final data,

notably: (i) the comprehensive, detailed, and

precise specification of the reporting

requirements; (ii) the complete and correct

reporting by the counterparties to the TRs; and

(iii) the provision of complete and accurate data

by the TRs to the authorities.

Data Quality Review (DQR): The DQR is

currently the main common exercise performed in

the context of the DQAP with regards to the

supervision of the reporting by the counterparties.

Under the DQR, each NCA, applying a commonly

agreed methodology, performs a quantitative

assessment of the quality of data reported by

selected counterparties in their Member State

and follow up with the relevant entities on the

identified issues.

NCAs provide subsequently to ESMA information

on the results of the DQR and on the follow-up

supervisory actions. Based on this feedback

ESMA prepares a summary report that is

subsequently shared with the NCAs. High-level

outcomes are also provided to the Board of

Supervisors as part of the annual update on the

execution of the DQAP.

DQR 2021: Similarly to the previous year, the

2021 DQR contained a series of over 30 data

quality tests grouped into three broad areas: (i)

analysis of pairing and matching of the reports,

(ii) analysis of completeness, accuracy,

timeliness, and rejections of reports made by

significant reporting entities, and (iii) thematic

review: analysis of reporting of valuation and

collateral data.

EMIR and SFTR data quality report 2021 10

In 2021, 19 NCAs participated in the DQR. Given

that the DQR analysis is based on limited

samples of counterparties

and that each

counterparty may face different reporting issues,

the results of the data quality checks are not

representative for the full EMIR dataset and vary

across the different tests and between the

participating Member States. However, based on

the samples considered in the DQR, overall, a

slight improvement has been noted in some of the

analysed areas as compared with the 2020 DQR,

notably the pairing and matching rates as well as

the consistency of number of reported derivatives

with the entities’ internal records.

Thematic review: Furthermore, it is worth

mentioning the outcomes of the thematic review,

focused on the reporting of valuations and

margins, which is a key information for the

monitoring of systemic risks. In addition to

analysing the aggregate results for the samples

from each Member State, ESMA has looked also

into the evolution of one basic measure, notably

the number of trades with empty/zero valuations,

at entity level for all the entities that were selected

for the thematic review based on this measure in

the 2020 DQR. The comparison of the statistics

computed for the purpose of 2020 DQR and 2021

DQR, revealed that approximately 50% of the

entities have significantly reduced the

misreporting of valuations

and further 20%

recorded some reduction in the number of trades

impacted by this data quality issue.

These outcomes confirm the conclusions from

the previous year that adequate supervisory

pressure and close monitoring of the

implementation of remedial actions are needed to

ensure a material long-term impact on the

improvement of the quality of data by all relevant

entities. They also show that targeted actions

directed at the main misreporting entities having

the highest impact on a given data quality aspect

constitute an efficient approach to resolve the

data quality problems.

Framework for provision of information on data

quality issues to NCAs and the follow-up with

supervised entities: In line with these findings

NCAs and ESMA have also established a

common framework, applicable whenever a

significant problem impacting the data quality of

the EMIR data at EU level is identified. The

framework sets up a procedure for an efficient

Each Member State selects 5 entities per each of the three

areas of analyses

resolution of such most significant data quality

issues, specifying, among others:

• the responsibilities of NCAs and ESMA,

• the timelines for the exchange of

information between NCAs and ESMA,

• the format and minimum content of the

statistics to be shared by ESMA,

• the criteria to decide which reporting

should be addressed,

• the feedback information to be provided

by the NCAs to ESMA,

• the steps to ensure that the data quality

problem has been mitigated (incl.

reassessment of the data).

The important feature of the framework is that

the follow-up is focused on a limited subset of

entities with the highest share of incorrect

reports in the total number of impacted reports at

EU level, thus ensuring the most efficient use of

the NCAs resources. The framework was first

launched in practice in 2021 for the follow-up on

the timeliness of reporting and resulted in a

material improvement in the reporting practices

of the targeted entities. (see Subsection 4.3 for

more details).

What have we achieved: EMIR DQAP is a

comprehensive data quality framework based on

the harmonised data assessment methodology.

Implementation of EMIR DQAP by the NCAs

allowed them to detect and follow up on several

data quality issues with their supervised entities.

Furthermore, ESMA shared with relevant NCAs

information on entities with highest numbers of

late reports at EU level. The targeted follow-up

with those entities resulted in an improvement of

their reporting practices.

4.2 SFTR DQEF

SFTR Data Quality Engagement Framework

(DQEF) The reporting under SFTR started in July

2020, following a three-month delay due to the

Covid-19 pandemic. SFTR, as a new transaction-

level reporting regime, has required dedicated

efforts for its supervision and data quality over the

year 2021.

Compared to EMIR, SFTR is ISO20022 XML

end-to-end reporting regime and this has brought

important benefits in terms of the quality and

Reduction of the number of trades with empty/missing

valuation by halve or more was considered as a significant

reduction

EMIR and SFTR data quality report 2021 11

accessibility of the data from the very beginning

of the reporting.

To ensure a consistent and efficient approach to

data quality assurance and supervision, ESMA is

systematically developing and implementing

various convergence tools across all relevant

reporting regimens and data systems including

the SFTR DQEF.

The SFTR DQEF was agreed in 2021 and

leverages on the EMIR DQAP setting out the

SFTs data quality work to be undertaken jointly

by NCAs and ESMA. It defines the necessary

coordinated procedures to verify, communicate

and prioritise the data quality findings detected in

the SFT data submitted by the reporting

counterparties and to subsequently apply the

relevant corrective measures leveraging on the

agreed best practices to foster the SFTs data

quality and to enforce the supervisory actions.

DQEF 2021: 2021 has been the first year of data

quality assessments performed by NCAs and

ESMA on SFTR data. Some national authorities

indicated that due to the novelty of the SFTR

reporting regime and to the complexity of the

activities to be implemented, they were not able

to contribute to the first data quality exercises. To

allow NCAs to focus on building their systems as

well as to facilitate the deployment of adequate

resources for their supervisory activities and

engagement with entities, ESMA has centrally

performed, on behalf of the NCAs, a targeted set

of checks and reported the detected data quality

issues to the NCAs. In turn, during the first year

of activities, the NCAs focused mostly on the

remedial actions and follow-ups with the

counterparties and entities responsible for

reporting under their direct supervision.

The implementation and performance of the data

quality checks is based on an incremental

approach. The 2021 data quality exercise was

performed in two separate rounds of tests on

weekly datasets (June and November) and had a

very targeted nature focusing on the aspects,

which under other reporting regimes have proven

to be the cornerstones of data quality, namely: (i)

timeliness of reporting, (ii) rejected reports due to

incorrectness or inaccuracy of SFT records

according to the validation rules (iii) unsuccessful

paired status of the reported records.

Considering the importance of the SFTR regime

on the one hand and its relative complexity on the

other, it is crucial to perform regular data quality

assessments and expand in the near future the

existing data quality activities with a view to

ensure the usability of the data for monitoring of

financial stability risks.

What have we achieved: The activities related to

the data issues of November 2021 cycle are not

yet completely terminated as the participating

NCAs are finalising the outputs of their

interactions with the supervised entities and

communicate them to ESMA. Therefore, it is

premature to draw conclusions on the overall

outcome of their remedial actions in this report. It

is important to highlight that the trends of the data

quality findings of the two cycles are also

reflected in the relevant outcomes of the SFTR

data quality analyses that are covered in Section

7. In some areas such as rejections and

reconciliation rates, these indicators already

begin to show positive trend.

4.3 Cooperation with data users

EMIR Timeliness Analysis: ESMA performed an

analysis of the timeliness of reports under EMIR

based on a time series constructed for several

dates across 5 months and using both Trade

State Reports (TSR) and Trade Activity Reports

(TAR). Thanks to the interactions with the NCAs,

the analysis was further enhanced to eliminate

certain false positives or legacy trades.

In line with the criteria specified in the framework

for provision of data by ESMA to NCAs (incorrect

reports by a given entity exceeding the 1% of all

incorrect reports in the EU), ESMA has identified

for the follow-up 15 counterparties from 7

jurisdictions. Furthermore, as envisaged in the

framework, all relevant NCAs have provided

ESMA with the feedback on the follow-up with the

entities explaining the reasons for late reporting

and indicating whether the issue has been

resolved.

In order to assess the actual impact of the

exercise, ESMA staff have rerun the timeliness

analysis for other dates in July, i.e., after the

NCAs finalised the follow-up with the entities.

No improvement was observed in the TSR, which

was an expected outcome given the particularity

of late reporting, i.e., once a given derivative is

reported late, it is not possible to ‘correct’ the time

of its initial submission. Therefore, such

derivative until further life-cycle event is reported,

will continue to appear as reported late in the

TSR. However, the approached entities have

enhanced their reporting practices with regards to

EMIR and SFTR data quality report 2021 12

timeliness, as an improvement has been

observed in the TAR submitted after the follow-

up. In particular, among the 10 entities which

were identified for the follow-up based on the

number of late reports in the TAR data, 4 entities

had no more late reports in the TAR of July and a

reduction in the number of late reports has been

observed also for the remaining 6 entities.

This example showed that targeted ad-hoc

exercises directed at the most relevant entities

are very efficient to reduce the most significant

data quality issues. ESMA plans to continue

engaging with NCAs in this way on a broader

number of issues.

Abnormal values: In August 2020 ESMA

implemented a new data quality process to

identify abnormal values on the numerical fields

reported by the counterparties under EMIR

regime. The data were then shared with the

NCAs to support them in their supervisory

activities. Notably, once the data are shared,

NCAs can verify if the detected outliers are due

to data quality issues in the reporting of the

counterparties under their supervision and follow

up accordingly, if needed.

The data quality analysis identifies irregular

numerical values (e.g., too high or too low)

reported for EMIR fields such as value of the

contract, margins, notional, fixed-rate legs of the

contract, price rate and quantity

. The focus on

these values is driven by the impossibility to

automatically detect and reject outliers through

the validation rules and the need for specific soft

checks.

The abnormal value analyses were shared with

the NCA monthly from August 2020 to August

2021. This section summarises the main findings.

Firstly, the abnormal values that ESMA identified

represent a small proportion of the outstanding

open derivatives. However, given the high impact

on data quality of the derivatives with abnormal

values, it is important to perform such data quality

checks on an ongoing basis to swiftly detect any

abnormal values and to inform the NCAs

responsible for the supervision of the

counterparties reporting such values. For

instance, a recurrent example concerns the

reported value of the contract above 100 billion

EUR, which may impact significantly results of

analyses based on EMIR data.

ESMA uses various internal methodologies to identify and

treat abnormal values. For example, for the purposes of

economic analysis, ESMA is using a statistical approach

(identifying outliers in notional values) elaborated in detail

here:

From August 2020 to August 2021, ESMA

notified NCAs on 84 occasions, reporting 3493

potential outliers to fifteen jurisdictions. ESMA

then received 19 responses from NCAs clarifying

the reasons for the outlier or committing to

contact the counterparty responsible for the

misreporting. Chart 1 provides an overview of the

notified potential abnormal values detected from

August 2020 to August 2021 per jurisdiction.

Chart 2 shows the EMIR fields with most outliers

identified in the same period.

Chart 1

Cooperation with data users

Number of abnormal values, per country

Source: ESMA data & calculations

Chart 2

Cooperation with data users

Number of abnormal values by EMIR field

Source: ESMA data & calculations

EMIR and SFTR log of data quality issues: Since

2016, ESMA has established a structured

framework for the users of data to be able to

https://www.esma.europa.eu/system/files_force/library/esma

50-165-639_esma-rae_asr-

derivatives_2018.pdf?download=1

EMIR and SFTR data quality report 2021 13

report any encountered issues and, in turn, to

receive feedback. ESMA receives issues

pertaining to TRs (for example, incorrect

generation of regulatory reports) as well as

counterparties (for example, implausible

notional/collateral values). When a counterparty

reporting issue is identified, the issue is

channelled to the responsible NCA. TR issues

are addressed by ESMA.

Chart 3 shows the breakdown of all reported

issues since the inception of the log by their

status, i.e., TR issues closed and open, and

counterparty reporting issues. In 2021, ESMA

processed around 30 issues overall. Most TR

issues are also being closed during the same

year. However, ESMA prioritizes its follow-ups

based on urgency and impact, thus not all the

issues may be addressed during the same year.

What have we achieved: Data quality issues can

be of structural nature, but significant issues can

also appear from one day to another. Thus, it is

important for ESMA and NCAs to maintain agility

to be able to react to issues that were not planned

to be addressed through structural projects such

as the EMIR DQAP and SFTR DQEF.

Through ad-hoc sharing of DQ issues (such

timeliness reporting, abnormal values, and any

other data quality issues) and their immediate

prioritisation, identified problems are being

address in an agile fashion.

Chart 3

Cooperation with data users

Data quality issues reported by NCAs\CBs

Note: Number of issues reported data quality issues since

inception and broken down by status.

Source: ESMA data & calculations

5. Recent developments

impacting EMIR and SFTR

data quality

Summary: Regarding EMIR, ESMA carried out three supervisory projects focusing on i) the ingestion

and processing of data by TRs, ii) application of EMIR access filtering rules for provision of data to

NCAs according to their mandates and iii) the assessment of consistency of two key regulatory reports

– the trade activity and trade state reports. In all three cases, ESMA found that TRs broadly follow

regulatory and supervisory expectations. In certain instances, ESMA found some shortcomings in the

quality of the reports provided to regulators

and expects that TRs take appropriate remediation steps.

As regards SFTR, TRs and reporting counterparties implemented first SFTR XML schema update since

the start of reporting in July 2020. The update aimed at removing technical shortcoming that could

decrease quality of the information available to the regulators. ESMA also monitored and coordinated

with NCAs/TRs all relevant aspect of the wind-down of UnaVista repository services under SFTR as

well as the associated porting of SFTR data to other TRs

5.1 EMIR - TR data ingestion review

In 2020, ESMA identified a need to verify whether

and to what extent EMIR data quality issues arise

during the data ingestion processes

of TRs. A

thematic review was initiated to assess the data

ingestion processes of three out of four registered

TRs using a data-driven supervisory approach.

The sample consisted of trade activity reports for

two consecutive dates (25-26 November 2020)

from eleven major financial counterparties.

ESMA engaged with the French (AMF), Dutch

(AFM) and German (BaFin) authorities to obtain

proprietary trade activity data from the selected

counterparties. Without the collaboration of these

NCAs and the effort made by the counterparties

to extract, prepare, and submit the data to ESMA,

it would have not been possible to fully assess

the data ingestion processes of TRs.

A methodology and algorithms were developed to

assess and identify concrete data integrity issues

stemming from TRs’ data ingestion processes. In

broad terms, it consisted of comparing data

reported by counterparties with data stored by

Such as under-/over-reporting of data in the case NCA access filtering and report accuracy issues in the case of the trade

activity and trade state reports.

Data ingestion refers to the part of TR data processing after it the data is received from the reporting participants and before it

is loaded to TR databases. TRs typically queue, validate and perform other pre-processing tasks on the incoming data before it

is loaded TR databases.

TRs in their internal databases before any

subsequent data transformation, aggregation,

filtering, or report generation processes were

performed. The end-to-end process was also

verified by comparing data reported by

counterparties with data received through

TRACE.

Over 20 million records were processed and

analysed both from a data completeness (paring)

and accuracy (matching) perspective. While in

some cases the ingestion process remains

inherently complex and issues were detected, our

analysis showed a robust EMIR data ingestion

process and a good level of data integrity for all

the three TRs included in our sample. This

implies that the information stored in the TRs’

internal database matches the information

reported by the counterparties.

Chart 4 summarises the results from the data

completeness analysis. Perfect or nearly perfect

pairing rates (>99%) were obtained for eight out

of eleven TR-counterparty sets. This implies that

EMIR and SFTR data quality report 2021 15

all or almost all records in the counterparty

dataset were successfully identified in the TR

dataset for these cases.

Chart 4

EMIR – TR Data Ingestion Review

Comparative analysis of CP and TR data

Note: Trade repositories and counterparties in the sample

are anonymized.

Source: Trade Repositories & ESMA calculations

On the contrary, lower completeness rates were

observed for three out of eleven TR-counterparty

sets. The main reason for these discrepancies is

not directly attributed to deficiencies with the TR’s

data ingestion process but rather to the

submission of out-of-scope data by these

counterparties for this review (i.e., records with

reporting timestamp not being part of the subset

defined in the request for information).

What have we achieved: In terms of data

accuracy, ESMA discovered a range of

discrepancies between the information submitted

by counterparties and how it had been stored in

the TRs’ internal databases. Most were of non-

critical nature and could be explained by the way

the TRs’ have implemented their internal IT

systems. For example, rounding errors of decimal

values, date/time formats and other

misalignments which did not have a critical

ESMA receives all derivatives data reported under EMIR.

impact on data quality when TRs generate

outbound reports for regulatory authorities.

However, a few critical issues caused by

inappropriate modification of counterparty data

were detected. ESMA is liaising with the affected

TRs to rectify these issues which can have an

adverse impact on data quality.

The outcome of this review will mainly be used as

input to ESMA’s data quality risk assessment, by

eliminating risks that could arise from the data

ingestion process and focusing the supervisory

efforts on other EMIR data reporting process.

It is also worth mentioning that this analysis was

carried out under the current reporting

framework. The entry into force of EMIR Refit

could bring significant changes that could

adversely impact the TR’s data ingestion

processes. Going forward, ESMA will continue to

monitor incidents and complaints that are linked

to TR’s data ingestion processes to ensure that

those processes are adequate and resilient to

regulatory changes.

5.2 EMIR regulatory access filtering

review

When providing data to the authorities, TRs need

to apply filtering rules to make the data available

to the authorities based on their respective

mandates. This is important to avoid that an

authority receives data which it is not entitled to

and to ensure that each authority receives all the

data that is necessary to fulfil its mandates.

Mandates of authorities are set out in the Article

81(3) of EMIR and further developed in technical

standards. In order to verify that TRs are

providing data according to the regulatory

requirements, ESMA has cooperated with four

NCAs: CBoI (IE), CNB (CZ), CSSF (LU) and

MFSA (MT). These authorities shared with ESMA

the regulatory reports submitted to them by the

TRs. In parallel, ESMA applied the expected

filtering rules applicable to each authority to its

own reports received from the TRs

. Then

ESMA compared its filtered report with those that

were actually received by the NCAs.

Through this assessment, ESMA confirmed that

TRs seem to broadly follow the regulatory

EMIR and SFTR data quality report 2021 16

requirements with regards to provision of data to

the authorities.

ESMA also identified some shortcomings at TRs

which led to either underreporting or

overreporting of EMIR data to NCAs. Such issues

may affect the data completeness and thus the

ability of the NCAs to effectively supervise all

entities under their mandates.

The project identified the following key findings:

• On average, 4.5% of expected derivative

reports are not provided by TRs

(underreporting). This may prevent NCAs

from fulfilling their supervisory mandate

due to the missing information

• On average, 1.7% more derivative reports

than expected are provided by TRs

(overreporting). This poses a

confidentiality issue as the NCAs receiving

these additional derivative reports are not

entitled to have access to them.

Chart 5

EMIR ESMA regulatory access filtering review

Derivatives overreported to NCAs

Note: The tables show number of open derivatives for one

reference date that were present in the NCA report but

shouldn’t. The first table shows the total number of

overreported records. The second table shows the number

of overreported records as a percentage of the total

number of records that should have been included in each

NCA report (this number is not shown in the table). The

total number of records that should have been included in

each NCA report also serves as a weight to calculate

weighted averages by NCAs and TRs.

TR2 confirmed to ESMA that the problem is related to a

one-off issue with the generation of the regulatory report in

question. The issue has been now remediated.

Source: Trade Repositories & ESMA calculations

As shown in Chart 5, TR2, exhibiting the worst results,

confirmed that the problem is related to a one-off issue with

Chart 6

EMIR ESMA regulatory access filtering review

Derivatives underreported to NCAs

Note: The tables show number of open derivatives for one

reference date that weren’t present in the NCA report, but

should have been. The first table shows the total number

of underreported records. The second table shows the

number of underreported records as a percentage of total

number of records that should have been included in each

NCA report (this number is not shown in the table). The

total number of records that should have been included in

each NCA report also serves as a weight to calculate

weighted averages by NCAs and TRs.

Source: Trade Repositories & ESMA calculations

What have we achieved: With this project, ESMA

has verified that NCAs are broadly receiving data

in accordance with their mandates as specified in

EMIR.

As outlined above, ESMA identified several

shortcomings in the provision of data to

authorities. ESMA will follow up on those issues

with TRs and ensure that they are appropriately

remediated.

5.3 EMIR TAR-TSR review

The incorrect incorporation of the information

contained in the Trade Activity Reports (TAR) into

the Trade State Reports (TSR) is one of the key

TR data quality issues identified by the NCAs

during the execution of the EMIR DQR. ESMA

has been conducting in 2021 a dedicated project

for the comparison of the TAR and TSR data, in

particular, for the assessment of the

completeness and accuracy of the TSR as an

aggregation of individual TARs.

the generation of the regulatory report in question. The issue

has been now remediated.

TR/NCA NCA1 NCA2 NCA3 NCA4 TR total

TR1

240,952 285 80 - 241,317

TR2 3,303 100,668 1,001 296,466 401,438

TR3 12,766 308,597 35,938 47 357,348

TR4 27,577 35,640 23,044 N/A 86,261

NCA total 284,598 445,190 60,063 296,513 1,086,364

TR/NCA NCA1 NCA2 NCA3 NCA4

TR weighted

average

TR1

7.1% 0.0% 0.0% 0.0% 2.5%

TR2 35.3% 96.0% 30.3% 10.2% 13.3%

TR3 1.1% 8.0% 4.4% 0.0% 5.4%

TR4 1.7% 2.2% 1.3% N/A 1.7%

NCA weighted

average

4.6% 6.2% 1.2% 4.9% 4.5%

TR/NCA NCA1 NCA2 NCA3 NCA4 TR total

TR1

9,936 9,545 9,996 10,252 39,725

TR2 4 2 1 562 569

TR3 34,003 34,785 1,145 41,340 111,277

TR4 22,199 189,195 25,831 237,226

NCA total 66,142 233,527 36,973 52,154 388,797

TR/NCA NCA1 NCA2 NCA3 NCA4

TR weighted

average

TR1

0.3% 0.6% 0.4% 0.4% 0.4%

TR2 0.1% 0.0% 0.0% 0.0% 0.0%

TR3 2.9% 1.0% 0.4% 4.7% 2.0%

TR4 1.3% 10.7% 1.5% N/A 4.6%

NCA weighted

average

1.1% 3.4% 0.9% 0.9% 1.7%

EMIR and SFTR data quality report 2021 17

The project methodology is based on the

comparison of two consecutive EMIR TSRs. The

information collected in the first TSR is

dynamically updated with the successive TARs

received in the period between them. This results

in a calculated TSR which is compared with the

second TSR submitted by the TR. By comparing

the two files, it is possible to detect quality

problems related to the incorrect incorporation of

the information into the TSR.

The analysis was carried out during 4

consecutive weeks in November 2021, the

information presented in this report represents

the average of the individual results obtained in

each of these weeks. Considering all TRs, more

than 300 million TAR and 80 million of TSR

records were processed for each week of

analysis.

The metrics obtained allow the identification of

two types of data quality issues: first,

completeness issues related to the incorrect

presence or absence of records in the TSR;

second, accuracy issues related to the incorrect

update of a selected group of TSR fields. ESMA

is currently checking and following up with the

TRs on the results presented below.

Completeness of the TSR: Chart 7 shows the

percentage of missing records in the TSR over

the total volume of records (i.e., derivatives

present in the internally calculated TSR and not

found in the TSR received from the TRs). These

are records that were erroneously deleted or not

included in the TSR by the TRs. Although in

general terms the results can be considered

positive, there are certain divergences in the

figures obtained for each TR as one of the TRs

presents results close to 2%. These, together

with the high volumes of operations that

constitute TSR makes a further analysis of the

root causes of this issue necessary. It is equally

relevant to note that some TRs show very positive

results, with error levels close to 0%.

Analogously, as shown in Chart 8, data on

redundant records have also been obtained (i.e.

derivatives present in the TSR received from the

TRs but not in the internally calculated TSR).

These are records that were erroneously

included or not removed from the TSR by the

TRs. The results again diverge when comparing

the different TRs, although it is important to

highlight that one of the entities presents very

positive results with virtually no errors in this

metric. As in the previous test, the results require

further analysis to understand the possible

explanations for these findings.

Chart 7

Completeness of the TSR

Missing records in the TSR

Note: Trade Repositories are anonymized and presented

as TR1, TR2, TR3 & TR4. The repositories included in this

chart are DDRIE, KDPW, UNAVISTA B.V. & REGIS, the

anonymized aliases do not correspond to the order of the

TRs named.

Source: Trade Repositories & ESMA calculations

Chart 8

Completeness of the TSR

Redundant records in the TSR

Note: Trade Repositories are anonymized and presented

as TR1, TR2, TR3 & TR4. The repositories included in this

chart are DDRIE, KDPW, UNAVISTA B.V. & REGIS, the

anonymized aliases do not correspond to the order of the

TRs named.

Source: Trade Repositories & ESMA calculations

Accuracy of the TSR: The second part of the

project consisted in comparing a number of key

fields (36 different fields were selected) of the

TSR to verify whether the information contained

in them has been correctly updated or not. For

this purpose, the records of the internally

calculated TSR and the TSR provided by the TRs

were compared. The results are shown in Chart

EMIR and SFTR data quality report 2021 18

In general terms, it can be observed that all TRs

are close to 100%, which implies that a vast

majority of the fields in the report are correctly

updated. However, in this first cycle of analysis a

selection of 36 fields of the TSR has been used,

so the results may vary if the number of fields

analysed is extended in future iterations.

Chart 9

Completeness of the TSR

Equality of the fields analysed

Note: Trade Repositories are anonymized and presented

as TR1, TR2, TR3 & TR4. The repositories included in this

chart are DDRIE, KDPW, UNAVISTA B.V. & REGIS, the

anonymized aliases do not correspond to the order of the

TRs named.

Source: Trade Repositories & ESMA calculations

Consideration will be given to broadening the

scope of fields included in the accuracy test to

verify whether the problems identified can be

extrapolated to the rest of the fields in the TSR.

Overall, for the two types of analyses included in

the test (completeness and accuracy), the

observed values differ between the different TRs

and the metrics obtained do not show significant

issues in terms of the correct generation of the

two key regulatory reports.

What have we achieved: It has been found that,

in general terms, there is a correct configuration

of the TSR as a result of the messages reported

in TAR. This further enhances the usefulness of

both reports for subsequent analysis by the

NCAs. On the other hand, the project has allowed

the implementation of a verification framework for

this process that can be replicated and increased

in the future. Finally, the project has allowed the

detection of certain data quality issues that, once

solved, will increase the overall quality of the

EMIR reporting framework.

See esma74-362-

1941_consultation_paper_guidelines_on_portability_emir_sft

r.pdf (europa.eu)

5.4 SFTR – Implementation of new

XML schemas

Leveraging on the experience with the

implementation of EMIR, SFTR reporting regime

relied since the beginning on the end-to-end

reporting in a standardised ISO 20022 XML

schema. Such design of the reporting framework

allowed to mitigate many data quality issues and

improve the usability of the data from the start.

What have we achieved: Since the beginning of

reporting under SFTR in July 2020, ESMA has

identified or has been made aware of some

limitations and inconsistencies in the XML

schemas used in SFTR. ESMA collected and

thoroughly analysed all the identified issues and,

basing on this assessment, prepared an updated

version of the schemas. The amendments to the

schemas are aimed to ensure that there are no

technical limitations to the accuracy of the reports

submitted by the counterparties to the TRs or the

reports provided by the TRs to the authorities.

The go-live of the updated schemas took place

on 31 January 2022.

5.5 SFTR – UnaVista wind-down

In August 2021, UnaVista initiated the wind-down

process of its SFTR TR as a result of a decision

to not continue to provide these services.

Although the Guidelines on data transfer between

trade repositories under SFTR were still under

consultation at the time

, the four SFTR TRs

began the implementation of the porting

infrastructure for SFTs. Leveraging on the

already existing portability framework and

infrastructure under EMIR and based on the

guidance provided by ESMA, TRs were able to

quickly adapt the existing porting infrastructure

and implement the necessary functionalities to

enable porting of SFT data.

UnaVista started the porting out process of

outstanding SFTs in November 2021. By late

January 2022, all outstanding SFTs were ported

out to the new TRs. The data transfer process of

the remaining SFT data is to be finalised by

March 2022.

EMIR and SFTR data quality report 2021 19

What have we achieved: Through continuous

monitoring, consequent follow-ups with the

involved TRs, and quick resolution of the

encountered issues, ESMA ensured that the

wind-down activities did not lead to any

interruptions in the continuity of the provision of

regulatory reports to all data users.

6. EMIR reporting trends and

selected data quality

metrics

Summary: Brexit has had an important impact on the EU supervisory data reporting landscape as

volumes of reported derivatives fell by approximately 50%. In terms of data reporting volumes, equities

and futures contracts continue to be the most prominent asset class and contract type respectively.

While less than 10% of reported derivatives tend to be reported late by the counterparties, more than

20% do not receive updated valuation on a daily basis as required by EMIR. Non-reporting dropped

sharply due to Brexit and is now less than 5%. The sharp drop has been driven by the end of the

reporting obligation of UK counterparties and the more limited dual-side reporting which does not allow

to detect potential non-reporting issues. TR rejections continue to be low at around 2%. Furthermore,

only 1% of records in TR regulatory reports seem to not comply with the applicable validation rules.

Volumes of duplicated reporting

are negligible. As regards reconciliation, pairing rate continues to be

relatively low at 60% while there is on average 5% difference between the number of open derivatives

reported between a pair of counterparties. Lastly, in some instances, TRs disagree on the number of

derivatives they reconcile against each other. This may be an indication of further enhancements

required for the inter-TR reconciliation process. While much attention has already been put to timely

reporting , reporting of valuations and reconciliation, clearly much more improvements are needed and

those area s will continue to be point of focus of ESMA and NCAs going forward.

6.1 Data reporting – Key trends

EMIR reporting trends can be viewed through

various dimensions of the data, such as the

number of life-cycle events, contract types, or by

asset class.

Submissions are the life-cycle event reports

received by the TRs representing the conclusion,

modification, and termination of a derivative (as

specified in the EMIR action type field)

throughout its life. When market volatility is high,

it may affect reporting counterparties’ trading

behaviour, leading to an increase in traded

volumes, which in turn will result in an increase in

the number of reported submissions.

Key trends: Most notable in Chart 10 is the

decrease in reported volumes post-Brexit, as UK

counterparties ceased to report under EMIR.

There has been a significant uptick in the

volumes reported by REGIS under EMIR since

Reported derivatives are considered duplicated where two or more records have been reported with the same combination of

reporting counterparty ID, ID of the other counterparty and trade ID fields. In contrast, double-sided reporting under EMIR is not

duplicative since two records on the same derivative should be reported always from the perspective of the respective reporting

counterparty.

the dissolution of CME and ICE Trade

Repositories.

Chart 10

EMIR reporting key trends

Monthly submissions per TR

Note: Total number of submissions per month and TR.

DDRIE is former DDRL and UNAVISTA B.V. is former

UNAVISTA LTD.

Source: Trade Repositories & ESMA calculations

EMIR and SFTR data quality report 2021 21

Chart 11 includes a breakdown of the total

number of submitted EMIR reports by action type.

The action type refers to the type of derivative life-

cycle event reported in the submission by the

reporting counterparty. In order of reported

frequency, the most common submissions are

valuation updates, position components, new

derivatives, modifications, compressions, early

terminations, corrections and errors. Valuation

updates remain the most frequent action type

submitted by the counterparties.

Chart 11

EMIR reporting key trends

Valuations are the most common submitted

action type

Note: Total number of submissions per month and action

type.

Source: Trade Repositories & ESMA calculations

Chart 12 shows that the share of each asset

class has remained relatively stable during the

year 2021. Notably there was a shift post Brexit

with equites representing a higher portion

amongst other asset classes, while commodity

and emission allowances as well as credit occupy

a lesser portion than previously.

Chart 13 provides a breakdown of EMIR

reporting by contract type: despite a relative

reduction after the end of 2020, futures remained

in 2021 the most reported EMIR contract type (on

average 40% of total reporting), followed by

financial contracts for difference (30%), options

(25%) and other types (5%).

Chart 12

EMIR reporting key trends

Equities are the most common EMIR asset

class

Note: Total number of daily submissions per month and

asset class. As a percentage of total.

Source: Trade Repositories & ESMA calculations

Chart 13

EMIR reporting key trends

Futures are the most common EMIR contract

type

Note: Total number of daily submissions per month and

contact type. As a percentage of total.

Source: Trade Repositories & ESMA calculations

EMIR and SFTR data quality report 2021 22

6.2 Data completeness, timeliness and

availability

EMIR Timeliness Analysis: Counterparties are

required to report newly concluded derivative

contracts by the end of the following working day

to a TR of their choice. To assess the timeliness

of reporting by the counterparties, ESMA

considers the difference between the “Execution

timestamp”, reflecting the date and time of a

derivative contract’s conclusion, and the

“Reporting timestamp”, reflecting the date and

time of reporting to the TR.

A derivative is considered “reported on time”, if it

is reported by the working day following the day

on which the contract was concluded, at the

latest. A derivative is considered “late reported”,

if it is reported later than the working day following

the day on which the contract was executed. A

derivative is considered “early reported”, if it is

reported earlier than the date specified in the

“Execution timestamp” field.

A derivative concluded on a Friday or Saturday

and reported on the consecutive Monday is

subject to a “weekend effect” which is accounted

for in the calculation and correctly classified as

“on time”.

On the contrary, public-, national- and bank

holidays (i.e. “calendar effect”) are not accounted

for in the calculation, nor is the conversion of

Coordinated Universal Time (UTC) to local time

made. These approximations simplify and speed

up the calculation but could give rise to some

degree of inaccuracy (i.e., records wrongly

classified as “Late Reporting” due to UTC vs local

time differences, or due to calendar effect)

impacting the overall results. Despite these

methodology limitations, the analysis depicts a

fair representation of reporting behaviour by the

reporting entities.

Chart 14 shows the results obtained from the

analysis of daily Trade Activity Reports (TAR) for

2021. The proportion of late reporting remained

on average below 10% while early reporting is

negligible or non-existent.

For example, Maltese MFSA has engaged with two most

problematic counterparties in their jurisdiction and is seeking

remediation of the issue.

Chart 14

EMIR data completeness, timeliness and availability

EMIR TAR - Timeliness Analysis (daily)

Note: A derivative executed at time T and reported at T+1

at latest, is considered “On Time”. A derivative executed at

time T and reported after T+1 is considered “Late

Reporting”. A derivative executed at time T and reported

before T is considered “Early Reporting”. A derivative

subject to “weekend effect” is classified as “On time”.

Source: Trade Repositories, and ESMA calculations using

daily trade activity reports for 2021.

A few spikes in late reporting observed around

New Year, Easter and other public holidays

should be considered as merely a calendar effect

impacting the accuracy of the results.

Chart 15 shows the aggregated results split by

the jurisdiction of reporting counterparties. While

in most jurisdictions the occurrence of late

reporting remains limited and low, there are a few

jurisdictions like Lithuania, Malta, Latvia, Greece,

Austria, amongst others, where reporting entities

tend to systematically report late.

It is often one or a few reporting entities that

drives the build-up of late reporting which is why

it is relevant for NCAs to intervene appropriately

as targeted effort from their side can often rectify

the problem quickly and bring down a significant

portion of late reporting within its jurisdiction

EMIR and SFTR data quality report 2021 23

Chart 15

EMIR data completeness, timeliness and availability

EMIR TAR - Timeliness Analysis by

jurisdiction

Note: A derivative executed at time T and reported at T+1

at latest, is considered “On Time”. A derivative executed at

time T and reported after T+1 is considered “Late

Reporting”. A derivative executed at time T and reported

before T is considered “Early Reporting”. A derivative

subject to “weekend effect” is classified as “On time”.

Source: Trade Repositories, GLEIF and ESMA

calculations using daily trade activity reports for 2021.

To complement and expand the previous

analysis, ESMA applied the same calculation

using a different dataset, namely the Trade State

Report (TSR), which contains the latest state of

all outstanding derivative contracts.

Chart 16 shows the results obtained from the

analysis of the weekly TSR data for 2021. The

proportion of late reporting is significantly higher

compared to the results obtained using TAR data.

This was somewhat expected as TSR data

reflects the cumulative effect of daily reporting

events.

Once a given derivative contract is executed and

reported late, it is not possible to ‘correct’ the time

of its initial submission. Therefore, such

derivative contract will continue to appear as

reported late in the TSR (at least until another

lifecycle event is reported for the same contract).

Moreover, the share of late reports in the TSR

may also be partially caused by batches of back-

dated reporting of previously non-reported

trades.

However, monitoring of the evolution of

timeliness reporting using TSR data can help to

identify persistent patterns caused by the

frequent as well as the sporadic misreporting

events that would otherwise be difficult to identify.

Chart 16

EMIR data completeness, timeliness and availability

EMIR TSR - Timeliness Analysis (weekly)

Note: A derivative executed at time T and reported at T+1

at latest, is considered “On Time”. A derivative executed at

time T and reported after T+1 is considered “Late

Reporting”. A derivative executed at time T and reported

before T is considered “Early Reporting”. A derivative

subject to “weekend effect” is classified as “On time –

weekend effect”.

Source: Trade Repositories, and ESMA calculations using

weekly trade state reports for 2021.

Chart 17 shows the aggregated results split by

the jurisdiction of reporting counterparties. Also,

here the effect of late reporting is more

pronounced compared to the results obtained

using TAR data (see Chart 15). Thirteen out of

thirty jurisdictions have a late reporting rate

above 20% and a few are even above 40%.

Although the late reporting is perceived as a

marginal issue when looking at TAR data, it

becomes more relevant when analysing the

cumulative effect of such misreporting behaviour

in the TSR data.

Identifying reporting entities who consistently

report late newly executed derivatives and

rectifying such behaviour remains important for

EMIR and SFTR data quality report 2021 24

achieving an accurate picture of the derivatives’

market at all times and thus for the efficient

surveillance of the systemic risk.

Timeliness of reporting by counterparties has

been the focus of NCAs during 2021 (see

Subsection 4.3 Cooperation with data users).

Chart 17

EMIR data completeness, timeliness, and availability

EMIR TSR - Timeliness Analysis by

jurisdiction

Note: A derivative executed at time T and reported at T+1

at latest, is considered “On Time”. A derivative executed at

time T and reported after T+1 is considered “Late

Reporting”. A derivative executed at time T and reported

before T is considered “Early Reporting”. A derivative

subject to “weekend effect” is classified as “On time –

weekend effect”.

Source: Trade Repositories, GLEIF and ESMA

calculations using the trade state report of 2021/21/31.

Non-reporting of valuations by counterparties:

EMIR requires that financial and non-financial

counterparties above the clearing threshold

report daily the valuation and collateral data

relating to their open derivatives.

To assess the timeliness of reporting of the

valuation updates, ESMA computes the

difference in number of days between the

reference date of a Trade State Report (TSR) and

the “Valuation timestamp” of a record, which

reflects the date and time of a valuation update.

While a stricter reading of EMIR would mean that

valuations that are older than one working day

are outdated, for the purposes of this analysis it

is considered that valuation updates older than

15 calendar days are outdated and should have

been subject to new valuation updates. Four

distinct buckets are used to group each record

that is in scope for this analysis and measure how

frequent valuation updates occur.

Chart 18 shows that around 80% of open

derivatives have received valuation updates that

are not older than 15 days. It also shows that

around 20% of open derivatives subject to daily

valuation have not received updates for several

days, months and even years.

Chart 18

EMIR data completeness, timeliness, and availability

EMIR TSR - Valuation updates

Note: The analysis uses all open derivatives from Trade

State Reports (TSR) with action type = “N” (new) or “V”

(valuation) and clearing threshold = “Y”. The difference in

number of days is computed between the reference date

of the TSR and the valuation timestamp of a record. Each

record is grouped into one of the four buckets to measure

the magnitude of number of outstanding trades that have

or have not received a valuation update between a certain

number of days.

Source: Trade Repositories and ESMA calculations

A significant number of open derivatives with

outdated valuation timestamps could indicate

misreporting practices by counterparties and/or

open derivatives that have not been properly

terminated (i.e., “dead” trades).

Chart 19 shows the aggregated results split by

the jurisdiction of reporting counterparties. While

in most jurisdictions the stock of open derivatives

is frequently receiving valuation updates, in a few

other jurisdictions there is a significant portion of

open derivatives that are not being updated. The

problem appears to be mostly prominent in

Bulgaria, Estonia, Latvia and Malta based on the

EMIR and SFTR data quality report 2021 25

results obtained from the trade state report of 31

December 2021

Considering the importance of the valuation data

for economic and financial risk analysis, NCAs

and ESMA will continue focusing on the

completeness and timeliness of valuation

reporting going forward.

Timeliness of valuations reporting by

counterparties has been the focus of NCAs

during 2021 during the EMIR DQR (see

Subsection 4.1 EMIR DQAP).

Chart 19

EMIR data completeness, timeliness, and availability

EMIR TSR - Valuation updates by

jurisdiction

Note: The analysis uses all open derivatives from Trade

State Reports (TSR) with action type = “N” (new) or “V”

(valuation) and clearing threshold = “Y”. The difference in

number of days is computed between the reference date

of the TSR and the valuation timestamp of a record. Each

record is grouped into one of the four buckets to measure

the magnitude of number of outstanding trades that have

or have not received a valuation update between a certain

number of days.

Source: Trade Repositories, GLEIF and ESMA

calculations using the trade state report from 2021/12/31.

For example Maltese MFSA confirmed that it engaged with

the most problematic counterparties in their jurisdiction and

they are now actively working on the remediation of the issue.

Non-reporting of derivatives: Chart 20 shows the

scale of potential non-reporting. It is not possible

to estimate the non-reporting fully, however, due

to the double-sided reporting obligation ESMA

has estimated the potential scale of the problem

by identifying derivatives where only a report in

one direction was submitted and a report from the

other direction is expected, i.e., the other

counterparty is in the EEA, and it has a reporting

obligation. Prior to Brexit, the number of

potentially non-reported derivatives stood at

around 3.5 million (approximately 5% of open

reconcilable derivatives at the end of 2020). After

Brexit, with the end of reporting obligation of UK

counterparties the number has sharply dropped

to around 0.5 million open derivatives (around 1%

of open reconcilable derivatives).

Importantly, the drop in number of non-reported

derivatives is driven by the cease of reporting

obligation in the UK. While the reporting

obligation of UK counterparties may have

stopped, the EU side of the trade still has its

reporting obligation, and it is expected to report.

Thus, with Brexit, potential non-reporting has

become somewhat harder to detect.

Chart 20

EMIR data completeness, timeliness, and availability

Potential non-reporting sharply dropped

after Brexit

Note: Number of potentially nonreported open derivatives

by date. The estimate is based on the number of reports

where only one leg of the derivative was reported and the

second leg is expected, i.e., the second counterparty is

EEA, and it has a reporting obligation (legal entity).

Source: Trade Repositories & ESMA calculations

Chart 21 shows a breakdown of open derivatives

where the report of the second counterparty to a

derivative is missing grouped by the jurisdiction

of the reporting counterparty. The largest

jurisdictions in terms of open derivatives also

tend to have the highest number of potentially

non-reported derivatives.

EMIR and SFTR data quality report 2021 26

Chart 21

EMIR data completeness, timeliness, and availability

Top 5 European jurisdictions take above

75% of all potentially non-reported

derivatives

Note: The designation “OTHER” includes countries with

less than 2% values. Those are, in descending order, DK,

EE, NO, FI, LV, AT, BE, LI, HU, CZ, MT, PT, LT, BG, GR,

HR, SK, RO, SC, IS, SI, GB, PA, KY, GI and JE.

Source: Trade Repositories & ESMA calculations

6.3 Data integrity – Adherence to

format and content rules

As shown in the Chart 22, the rejection statistics

provided by TRs pointed to a transitory uptick in

the rejection rates from April to August. A further

breakdown of the data indicates that the elevation

in the rejection rate affected TR2 in April and TR4

in June to July, as illustrated in the Chart 23.

Chart 22

EMIR data Integrity – Adherence to format and content

Total volume and rejections as % of total

Source: Trade Repositories & ESMA calculations

Chart 23

EMIR data Integrity – Adherence to format and content

Rejection rate by TR

Note: Rejection % of total submissions received by each

TR. The repositories included in this chart are ICE, CME,

DDRIE, DDRL, KDPW, UNAVSITA B.V., UNAVISTA LTD

& REGIS, the anonymized aliases do not correspond to the

order of the TRs named.

Source: Trade Repositories & ESMA calculations

Duplicated reporting: Unique derivative is

identified based on three EMIR fields, i.e.,

reporting counterparty ID, ID of the other

counterparty and trade ID. To avoid undue

double-counting when using the data for

economic/financial analysis, it is essential that

counterparties report in a way that the

uniqueness of each derivative is respected. TRs

are expected to verify the uniqueness of reported

new derivatives and to reject those that have

been reported with the same triplet of IDs in the

past. TRs are, however, unable to identify

duplicates when a counterparty reports to two

different TRs.

To assess the volumes of duplicate reports,

ESMA performed analysis of the uniqueness of

the derivatives in the Trade State Report.

Chart 24 shows a percentage of duplicates at TR

level (“intra-TR”). While from time to time, the

number of duplicates can increase, the overall

number is less than one percent. Thus, at TR

level, duplicated reporting does not pose

significant issues.

EMIR and SFTR data quality report 2021 27

Chart 24

EMIR data Integrity – Adherence to format and content

Intra TR duplicate records as % of total

reported volume

Note: Duplicated open derivatives as % of all open

derivatives. A unique derivative is defined at the level of

three EMIR fields: reporting counterparty ID, ID of the other

counterparty, and trade ID. Intra-TR duplicates are

detected in individual TR reports submitted to ESMA.

Source: Trade Repositories & ESMA calculations

Chart 25 shows a percentage of duplicates

across TRs (“inter-TR”). Similarly, to the results

shown above, the duplicated records do not

seem to pose significant issues even at the inter-

TR level.

Chart 25

EMIR data Integrity – Adherence to format and content

Inter TR duplicate records as % of total

reported volume

Note: Duplicated open derivatives as % of all open

derivatives. A unique derivative is defined at the level of

three EMIR fields: reporting counterparty ID, ID of the other

counterparty, and trade ID. Inter-TR duplicates are

detected across all TRs reporting to ESMA.

Source: Trade Repositories & ESMA calculations

Revalidation: When a counterparty submits

reports to a TR, the latter needs to validate

whether the incoming data is in line with the

regulatory reporting requirements. For this

purpose, TRs have implemented the ESMA’s

validation rules against which they check the

incoming data. TRs are expected to reject reports

that are not adhering to the validation rules.

Since the introduction of the validation rules in

December 2014, ESMA regularly performs a

revalidation of the data made available by the

TRs with a view to assess whether TRs have

implemented the validation requirements

correctly. In ESMA's analysis, a randomly

selected data sample extracted from one daily

Trade Activity Report per month is used. Each

data point is checked against the ESMA

validation rules in force at the time of the

verification performed by the TRs. Following the

identification of an issue, e.g., a specific field that

causes unduly rejections, ESMA engages with

the relevant TR to remediate the issue at hand.

Chart 26 shows the number of records analysed

and the percentage of errors. On average, each

iteration of the analysis processed 8 million

records. The proportion of records containing

errors remained low and stable, fluctuating

around 1% during 2021. ESMA will continue to

perform the revalidation analysis for monitoring

the evolution and liaise with TRs when material

issues are detected.

Chart 26

EMIR data Integrity – Adherence to format and content

Number of failures/breaking records as %

of total

Note: The analysis uses a randomly selected data sample

(~15%) extracted from one daily Trade Activity Report per

month. Each data point is checked against the current

ESMA validation rules.

Source: Trade Repositories and ESMA calculations

6.4 Data integrity – Reconciliation

Reconciliation: Under EMIR, both counterparties

to the derivative are required to report their side,

to the extent that they are subject to the reporting

obligation. TRs are then required to reconcile the

EMIR and SFTR data quality report 2021 28

data. The TRs provide the results to the reporting

participants (so that any reconciliation breaks can

be addressed) and to the authorities (so that they

can monitor reporting of the counterparties in

their jurisdiction).

Chart 27 shows the results of pairing

. The

pairing rate stood at around 60% at the end of

2021. The rate has remained relatively stable

throughout the year. Considering that pairing is

performed by comparing three EMIR fields only,

its current level is not satisfactory. Unfortunately,

there are several reasons for lack of pairing such

as lack of agreement on the trade ID between

counterparties, under- and overreporting, wrong

identification of the other counterparty or lack of

agreement on the number of reports that should

be submitted in relation to a given derivative.

Chart 27

EMIR data integrity – Reconciliation

Reconciliation: Pairing has stabilised at

around 60%

Note: Pairing is performed based on three fields: Reporting

counterparty ID, ID of the other counterparty and trade ID.

Pairing rate is calculated by paired derivatives by sum of

paired and unpaired (excluding non-EEA derivatives).

Source: Trade Repositories & ESMA calculations

Chart 28 provides some potential further insights

as to the reasons for lack of pairing. The chart

depicts the net difference between the number of

derivatives reported by the two counterparties.

Prior to Brexit, the difference stood at around 8

million (10% of all open derivatives). After Brexit,

the number has dropped as UK counterparties no

longer have a reporting obligation (and the

number of derivatives with expected two legs of a

trade has dropped). Throughout 2021, the

number was around 2 million representing

around 5% of all open derivatives where 2 reports

Pairing is performed on the basis of three fields: Reporting

counterparty ID, ID of the other counterparty and trade ID.

are expected, i.e., excluding derivatives where

the other counterparty is non-EEA.

Chart 28

EMIR data integrity – Reconciliation

Difference in number of records reported

by the two sides

Note: The metric is calculated by taking a difference

between the number of derivatives reported by leg 1 and

leg 2.

Source: Trade Repositories & ESMA calculations

Chart 29 shows a breakdown of open derivatives

where the report of the second counterparty to a

derivative is missing grouped by the jurisdiction

of the counterparty. Top 3 jurisdictions represent

nearly 70% of all issues.

Chart 29

EMIR data integrity – Reconciliation

Difference in number of records reported

by the two sides by jurisdiction

Note: The designation “OTHER” includes countries with

less than 2% values. Those are, in descending order, ES,

DK, LU, SI, PL, SE, IE, BE, LV, FI, LI, EE, AT, HU, PT, CZ,

SK, MT, BG, LT, GR, HR, RO, IS and GB.

Source: Trade Repositories & ESMA calculations

Given that EMIR reports are characterized by a

large volume of data, reconciliation could be for

TRs a complex process. To assess the reliability

EMIR and SFTR data quality report 2021 29

of reconciliation statistics, ESMA started in 2021

to collect and analyse periodic information

requested from TRs in the context of its

supervisory activities. As part of this information,

TRs must provide the number of UTIs paired and

reconciled internally (when both sides of the trade

are reported to the same TR) or with other TRs

(when the two sides of the trade are reported to

two different TRs).

While reconciliation has been the focus of NCAs

during 2021 during the EMIR DQR (see

Subsection 4.1 EMIR DQAP), the results do not

yet show sufficient progress. Reconciliation thus

needs to remain a point of focus going forward.

Chart 30 and Chart 31 provide the preliminary

results of an analysis of the discrepancies in the

inter-TR reconciliation statistics: the number of

UTIs reported by each TR as paired/reconciled

versus the other TRs has been compared with the

number of UTIs reported by the other TRs as

paired/reconciled versus each TR.

Chart 30

EMIR data integrity – Reconciliation

Discrepancies in number of UTIs paired

versus Other TRs

Note: The charts are based on information provided

monthly by TRs to ESMA in the context of periodic

information (Item 36 – Reconciliation Statistics). The

figures in the charts refer to outstanding trades and are

computed as the monthly average for the reference period

June 2021 – December 2021

Source: Trade Repositories & ESMA calculations

The main finding of this analysis is that, although

pairing statistics do not seem to be affected by

significant inter-TR discrepancies, there are

relevant divergences in the number of UTIs

reported as reconciled among the TRs. On top of

the complexity of the reconciliation process,

discrepancies in the inter-TR statistics could be

caused also by the different times at which each

TR submits its data to reconciliation.

ESMA will continue to collect such information

and to investigate potential TR-specific issues.

Chart 31

EMIR data integrity – Reconciliation

Discrepancies in number of UTIs

reconciled versus Other TRs

Note: The charts are based on information provided

monthly by TRs to ESMA in the context of periodic

information (Item 36 – Reconciliation Statistics). The figure

in the charts refers to outstanding trades and are

computed as the monthly average for the reference period

June 2021 – December 2021.

Source: Trade Repositories & ESMA calculations

EMIR and SFTR data quality report 2021 30

7. SFTR reporting trends and

selected data quality

metrics

Summary: Similar to EMIR, data reporting volumes dropped approximately by 50% following Brexit. In

terms of number of open transactions, securities lending and borrowing is the largest SFT type reported

with around 70% share at the end of 2021. Credit institutions report most open SFTs (around 50%)

while credit institutions share has been increasing (to around 30% at the end of 2021). After merely 1.5

years of reporting, SFTR exhibits comparable results to EMIR across all data quality metrics. Around

10% of SFTs are reported late (after T+1). On the contrary, rejections have been low (around 2%) and

duplicated reporting does not pose major issues. As regards reconciliation, pairing rate has been only

around 60%. Reconciliation rate of loan and collateral data has been low but increasing to around 40%

and 30% respectively. Similar to EMIR, TRs do not agree on the number of records they reconcile

against each other, which may be an indication of issues in the inter-TR reconciliation process.

Timeliness of reporting , adherence to format and content rules (via rejections) and reconciliation

(pairing) has been the point of focus of ESMA and NCAs during 2021. While progress has been made,

some areas (particularly reconciliation) need to remain areas of focus also in the future.

7.1 Data reporting – Key trends

SFTR reporting started in July 2020 and was

followed by a phased-in period - concluded in

January 2021 – during which reporting

requirements were gradually extended to

different types of counterparties. Given a more

extensive availability of data reported under

SFTR in 2021, this section of the report contains

more elaborated analyses than in the previous

edition. Since SFTR data have a similar structure

to EMIR, reporting trends can be analysed from

similar perspectives (i.e., types and volumes of

life-cycle events, open contracts and reporting

counterparties).

TRs are required to submit daily to NCAs and

ESMA a set of 4 reports providing a thorough

overview on SFTR reporting activity:

1) the Trade Activity Report (TAR), in which

TRs provide all the life-cycle events reported

by SFT counterparties on the reference

date.

2) the Trade State Report (TSR), which

provides a snapshot of all the outstanding

SFTs at the reference date (i.e.,

incorporating all the reported life-cycle

events and applying them to the respective

SFT records).

3) the Rejection Report, which contains

statistics – at file and SFTs level – on the

acceptance/rejection of the reports received

by the TRs on the reference date (see

Subsection 7.3 Data integrity – Adherence to

format and content).

4) The Reconciliation Report, which provides

information – at a counterparties-pair level –

on the reconciliation activity performed by

TRs on expired and outstanding trades at

the reference date (see Subsection 7.4 Data

integrity – Reconciliation).

Key trends: Due to the removal of reporting

requirements for UK counterparties after Brexit,

SFTR reporting volumes significantly decreased

in 2021, falling to an average of 49 million life-

cycle events reported per month.

As shown in Chart 32, there were 4 TRs

providing SFTR reporting services in 2021:

DDRIE (DDRL before Brexit), which received the

largest share of submissions (on average 41

million records per month in 2021), followed by

UnaVista (5 million), Regis-TR (3 million) and

KDPW (0.1 million). It is worth mentioning that the

volumes reported by UnaVista in the TAR started

to decrease in December 2021 as a

consequence of the voluntary withdrawal of the

TR from the provision of reporting services for

SFTR, which is being finalised in the first half of

EMIR and SFTR data quality report 2021 31

2022 (see Subsection 5.5 SFTR – UnaVista

wind-down).

Chart 32

SFTR data reporting key trends

SFTs reported by Trade Repository

Note: Until Brexit, DDRL was the TR (UK entity part of the

DTCC group) authorized by ESMA for EMIR and SFTR

reporting services. After Brexit, ESMA authorization of

DDRL was withdrawn and granted to DDRIE (DTCC-entity

based in EU)

Source: Trade Repositories & ESMA calculations

Chart 33 provides the figures of SFTR reporting

volumes broken down by action type. Likewise for

EMIR, the action type field refers to the type of

life-cycle event reported for the SFT. The most

reported action types are the ones referring to

contract modification and valuation updates.

Chart 33

SFTR data reporting key trends

SFTs reported by action type

Note: Total number of submissions per month and action

type.

Source: Trade Repositories & ESMA calculations

Through the periodic monitoring of SFTR reports,

ESMA detected a relevant data quality issue

inflating the number of outstanding SFTs caused

by one counterparty failing to report maturity date

on its transactions. As shown in Chart 34 the

number of open SFTs has been constantly

increasing until November 2021, when – as a

result of an action taken by the relevant

competent authority – the counterparty corrected

the data reporting errors by submitting

backloaded early terminations.

Chart 34

SFTR data reporting key trends

Inflation in open SFTs reports caused by

one reporting counterparty

Note: Extract of the TSR submitted by the TR contracted

by one SFT counterparty missing to report early

termination life-cycle events.

Open SFTs, are transactions that have not matured, or

which have not been the subject of reports with action

types ‘Error’, ‘Termination/Early termination’, or ‘Position

component’.

Source: Trade Repositories & ESMA calculations

According to the TSR submitted in 2021, it results

that Securities and Commodities Lending or

Borrowing (SLEB) is the most common type of

outstanding SFTs as shown in Chart 35. The

notable drop in the percentage of SLEB at the

end of 2021 is attributable to the data quality

issue mentioned in the previous paragraph.

Chart 35

SFTR data reporting key trends

Open SFTs by type

Note: Total number of open SFTs per month and SFT type,

as percentage of the total.

Open SFTs, are transactions that have not matured, or

which have not been the subject of reports with action

types ‘Error’, ‘Termination/Early termination’, or ‘Position

component’.

The chart contains all open SFTs including those that were

‘overreported’ as shown in Chart 35.

Source: Trade Repositories & ESMA calculations

Chart 36 shows the distribution of outstanding

SFTs providing a breakdown by type of reporting

counterparty: most of the open SFTs have been

EMIR and SFTR data quality report 2021 32

reported by credit institutions (on average 60% of

the outstanding SFTs), followed by Investment

Firms (20%). It is notable how in January 2021

the distribution of types of counterparties

changed (less Investment firms due to Brexit and

more Other Financial CPs and Non-Financial

CPs after the SFTR phase-in).

Chart 36

SFTR data reporting key trends

Open SFTs by Counterparty Sector

Note: Total number of open SFTs per month and reporting

counterparty sectors, as percentage of the total. Other

Financial CPs are Insurance and Re-insurance firms,

AIFMs, Pension Funds and UCITs.

Open SFTs, are transactions that have not matured, or

which have not been the subject of reports with action

types ‘Error’, ‘Termination/Early termination’, or ‘Position

component’.

The chart contains all open SFTs including those that were

‘overreported’ as shown in Chart 35.

Source: Trade Repositories & ESMA calculations

7.2 Data completeness, timeliness,

and availability

SFTR Timeliness Analysis: The timeliness of

reporting under SFTR was performed using the

same concept and methodology as for EMIR (see

section 6.2).

Chart 37 shows the results obtained from the

analysis of daily Trade Activity Reports (TAR) for

2021. The proportion of late reporting remained

on average below 10% while early reporting is

negligible or non-existent.

Chart 37

SFTR data completeness, timeliness, and availability

SFTR TAR - Timeliness Analysis (daily)

Note: An SFT executed at time T and reported at T+1 at

latest, is considered “On Time”. An SFT executed at time

T and reported after T+1 is considered “Late Reporting”.

An SFT executed at time T and reported before T is

considered “Early Reporting”. An SFT subject to “weekend

effect” is classified as “On time – weekend effect”.

Source: Trade Repositories, GLEIF and ESMA

calculations using daily trade activity reports for 2021

A few spikes in late reporting observed around

New Year, Easter and other public holidays

should be considered as merely a calendar effect

impacting the accuracy of the results. In other

cases, a more persistent upward trend

accompanied by a drop in late reporting is

observed. This could be an indication of

misreporting behaviour by reporting entities.

Chart 38 shows the aggregated results for the full

2021 split by the jurisdiction of the reporting

counterparties. While in most jurisdictions the

occurrence of late reporting remains limited and

low, there are a few jurisdictions like Greece,

France, and Italy where reporting entities have

higher volumes of late reports on an aggregated

level.

EMIR and SFTR data quality report 2021 33

Chart 38

SFTR data completeness, timeliness, and availability

SFTR TAR - Timeliness Analysis by

jurisdiction

Note: An SFT executed at time T and reported at T+1 at

latest, is considered “On Time”. An SFT executed at time

T and reported after T+1 is considered “Late Reporting”.

An SFT executed at time T and reported before T is

considered “Early Reporting”. An SFT subject to “weekend

effect” is classified as “On time – weekend effect”.

Source: Trade Repositories, GLEIF and ESMA

calculations using daily trade activity reports for 2021

It is often one or a few reporting entities that drive

the build-up of late reporting, which is why it is

relevant for NCAs to intervene timely and

appropriately. Therefore, a targeted effort from

NCA’s side can often rectify the problem swiftly

and bring down a significant portion of late

reporting within its jurisdiction.

To complement and expand the previous

analysis, ESMA applied the same calculation

using a different dataset, namely the Trade State

Report (TSR) which contains the latest state of all

outstanding SFTs.

Chart 39 shows the results obtained from the

analysis of daily TSR data for 2021. The

proportion of late reporting is slightly lower

compared to the results obtained using TAR data,

while early reporting, which was non-existent in

TARs, is an apparent issue in this analysis.

Chart 39

SFTR data completeness, timeliness, and availability

SFTR TSR - Timeliness Analysis (daily)

Note: An SFT executed at time T and reported at T+1 at

latest, is considered “On Time”. An SFT executed at time

T and reported after T+1 is considered “Late Reporting”.

An SFT executed at time T and reported before T is

considered “Early Reporting”. An SFT subject to “weekend

effect” is classified as “On time – weekend effect”.

Source: Trade Repositories, GLEIF and ESMA

calculations using daily trade state reports for 2021

Once an SFT is concluded and reported early or

late, it is not possible to ‘correct’ the time of its

initial submission. Therefore, such SFT will

continue to appear as reported early or late in the

TSR.

Consequently, the significant stock of early

reported SFTs visible in the TSR occurred before

2021 as the TARs of 2021 does not show such

reporting behaviour. The issue has been

encapsulated in the TSR, vanishing only once

those SFTs reach maturity or are terminated, as

it can be observed in the sharp drop that occurred

towards the end of 2021. Moreover, the share of

late reports in the TSR may also be partially

caused by batches of back-dated reporting of

previously non-reported trades.

Monitoring of the evolution of timeliness reporting

using TSR data can help to identify patterns

caused by frequent as well as sporadic

misreporting events occurring both in the present

as well as in the past which would otherwise be

difficult to identify.

EMIR and SFTR data quality report 2021 34

Chart 40 shows the aggregated results split by

the jurisdiction of reporting counterparties. The

effect of early reporting is more pronounced

compared to the results obtained using TAR data

(see Chart 38).

Chart 40

SFTR data completeness, timeliness, and availability

SFTR TSR - Timeliness Analysis by

jurisdiction

Note: An SFT executed at time T and reported at T+1 at

latest, is considered “On Time”. An SFT executed at time

T and reported after T+1 is considered “Late Reporting”.

An SFT executed at time T and reported before T is

considered “Early Reporting”. An SFT subject to “weekend

effect” is classified as “On time – weekend effect”.

Source: Trade Repositories, GLEIF and ESMA

calculations using the trade state report of 2021/12/31.

While early reporting is evident in Lithuania,

Belgium, and Austria, late reporting is more

apparent in Greece, Slovenia, and Hungary.

Early reporting becomes more relevant when

analysing the cumulative effect of such

misreporting behaviour using TSR data. In some

jurisdiction above 20% of open SFTs have been

reported with the event date being later than the

reporting timestamp.

Identifying reporting entities who consistently

misreport newly executed SFTs and rectifying

Unique transaction under SFTR is defined on the basis of

three fields: Reporting counterparty ID, ID of the other

such behaviour remains important for achieving

an accurate picture of SFTs’ market at all times

and thus for the efficient surveillance of the

systemic risk.

Timeliness of reporting by counterparties has

been the focus of NCAs during 2021 during the

SFTR DQEF (see Subsection 4.2 SFTR DQEF).

7.3 Data integrity – Adherence to

format and content rules

As the SFTR regime settles in, the rejection rate

for SFT submissions continues a downward

trend. Furthermore, Chart 41 displays the

significant decrease in the volume of rejected

records. This trend has been noticeable mainly

since May-June and has become more

pronounced throughout the rest of the series. The

positive decreasing trend is the result of the

interaction with TRs and the joint work of NCAs

and ESMA in the context of the SFTR DQEF on

the quality of reporting by the counterparties (see

Subsection 4.2 SFTR DQEF).

Chart 41

SFTR data integrity – Adherence to format and content

Total volume and rejection as % of total

Note: Rejection rate is the percentage of records rejected

from the total amount of records received by the Trade

Repositories.

Source: Trade Repositories & ESMA calculations

Duplicated reporting: When validating transaction

messages submitted by reporting participants,

TRs are asked among other things, to validate the

incoming data and ensure that no duplicates

are being reported.

counterparty and trade ID. In the case of collateral data

submissions, master agreement type is also used.

EMIR and SFTR data quality report 2021 35

Having said that, TRs can validate uniqueness of

reporting on data that is reported to them, i.e., a

TR cannot validate whether another TR receives

the same record.

To verify uniqueness of reporting, ESMA

identifies duplicated records on, both, TR level

(duplicates that TRs are expected to identify are

reject) and across TRs (duplicated records which

cannot be identified by TRs, and which are pure

counterparty misreporting issue).

Chart 42 and Chart 43 depict split between

duplicated and unique open SFT transactions on

monthly basis on a TR level (intra-TR) and across

all TRs (inter-TR). While there have been one-off

incidents at one TR where large number of

duplicates have appeared in the regulatory report

(particularly during end of 2020 and early 2021),

the results currently show that only a small

fraction of records is duplicated, and this issue

does not, now, pose any real problems for end

users.

Chart 42

SFTR data integrity – Adherence to format and content

Intra TR duplicate records as % of total

reported volume

Note: Duplicated transactions are identified based on three

fields: Reporting counterparty ID, ID of the other

counterparty and trade ID. In case collateral report, master

agreement type is used as well. The analysis has been

performed on open SFT transactions on a given reference

date.

Source: Trade Repositories & ESMA calculations

Pairing is performed on the basis of three fields which define

uniqueness of an SFT transaction: Reporting counterparty ID,

ID of the other counterparty and trade ID.

Chart 43

SFTR data integrity – Adherence to format and content

Inter TR duplicate records as % of total

reported volume

Note: Note: Duplicated transactions are identified based

on three fields: Reporting counterparty ID, ID of the other

counterparty and trade ID. In case collateral report, master

agreement type is used as well. The analysis has been

performed on open SFT transactions on a given reference

date.

Source: Trade Repositories & ESMA calculations

7.4 Data integrity – Reconciliation

Reconciliation: Reconciliation is one of the key

data quality processes performed by TRs. TRs

provide information on the results of the process

to all key stakeholders, i.e., reporting participants

and NCAs/ESMA. To visualize the results, ESMA

relied on the information provided in the

reconciliation reports for outstanding SFT

transaction. These reports contain, among

others, detailed breakdown of reconciliation

status for each open SFT.

Chart 44 shows the result of pairing

over time.

In the 4

quarter of 2020, the pairing rate stood

at about 45-50%. There has been a positive trend

throughout 2021 and by the end of the year the

pairing rate increased to 60%. Considering that

the SFTR is a relatively young reporting regime,

EMIR and SFTR data quality report 2021 36

these are not entirely disappointing results

(indeed EMIR pairing rate stands at around 60%

as well after 8 years of reporting). Having said

that, ESMA’s intention is to continue to focus on

pairing (and reconciliation more broadly) as part

of its work with the NCAs (see Subsection 4.2 for

more details).

Chart 44

SFTR data integrity – Reconciliation

Paired records

Note: Pairing is performed based on three fields: Reporting

counterparty ID, ID of the other counterparty and trade ID.

The results show pairing rate on open SFTs.

Source: Trade Repositories & ESMA calculations

Chart 45 shows the distribution of pairing rate

across jurisdictions. Clearly, there is a substantial

variation in the results as counterparties tend to

display varying performance levels in different

jurisdictions.

Chart 46 once a transaction is successfully

paired, TRs proceed with an attempt to match the

remaining fields

between the two reported sides

of a transaction. TRs perform matching both on

the loan as well as on the collateral information of

each transaction.

As regards loan matching, through 4

quarter

2020 and until end 2021 there has been a positive

trend as the matching rate increased from around

30% to nearly 50%.

The list of reconcilable fields that TRs use, including, where

applicable, tolerances can be found in the RTS on data quality

Chart 45

SFTR data integrity – Reconciliation

Paired records by country of the reporting

counterparty for the period Q4 2021

Note: Pairing rates by jurisdiction.

Source: Trade Repositories & ESMA calculations

Chart 46

SFTR data integrity – Reconciliation

Reconciled loan components

Note: Reconciliation rate on loan components of open

SFTs.

Source: Trade Repositories & ESMA calculations

EMIR and SFTR data quality report 2021 37

Chart 47 shows loan reconciliation rate by

jurisdiction. Once again, there is variation in the

performance of the results across countries.

Chart 48 shows reconciliation of collateral

components. Considering that collateral

reconciliation is the most complex step in the

entire process it is not surprising to see that

successful reconciliation exhibits the lowest rates

– only around 20% at the end of 2021.

Importantly, like in the previous cases, there has

been an increasing trend throughout the

displayed period.

Chart 49 shows breakdown of collateral

reconciliation rates by jurisdiction. Consistently

with the findings in the previous charts, there is a

substantial variation in reconciliation rates across

different jurisdictions.

While SFTR reconciliation (namely pairing) has

been the focus of NCAs during 2021 during the

SFTR DQEF (see Subsection 4.2 SFTR DQEF)

and some progress has been made, it is

reconciliation still needs to be point of focus going

forward.

Chart 47

SFTR data integrity – Reconciliation

Reconciled loan components by country of

the reporting counterparty for the period Q4

2021

Note: Loan reconciliation by jurisdiction.

Source: Trade Repositories & ESMA calculations

Chart 48

SFTR data integrity – Reconciliation

Reconciled collateral components

Note: Collateral component reconciliation of open SFTs.

Source: Trade Repositories & ESMA calculations

Chart 49

SFTR data integrity – Reconciliation

Reconciled collateral components by

country of the reporting counterparty for

the period Q4 2021

Note: Collateral reconciliation by jurisdiction.

Source: Trade Repositories & ESMA calculations

EMIR and SFTR data quality report 2021 38

A similar analysis to EMIR inter-TR reconciliation

process discrepancies (see Subsection 6.4 Data

integrity – ) has been carried out under SFTR.

Chart 50 shows that there are significant

divergences in the number of UTIs that a TR

considers that it has paired against other TRs.

Such discrepancies are an indication that the

reconciliation process performed by TRs may not

function appropriately and that information

provided by TRs on the outcomes of the

reconciliation process to the reporting

participants and NCAs may not correctly reflect

the correct state of reconciliation between two

legs of any SFT.

In case of SFTR, the reconciliation is somewhat

more complex than in EMIR ad TRs are required

to reconcile both the Loan and Collateral

component) which results in more evident

discrepancies (see Chart 51).

Chart 50

SFTR data integrity – Reconciliation

Discrepancies in number of UTIs paired

versus Other TRs

Note: The charts are based on information provided

monthly by TRs to ESMA in the context of periodic

information (Item 36 – Reconciliation Statistics). The

figures in the charts refer to outstanding trades and are

computed as the monthly average for the reference period

June 2021 – December 2021

Source: Trade Repositories & ESMA calculations

Chart 51

SFTR data integrity – Reconciliation

Discrepancies in number of UTIs

reconciled versus Other TRs

Note: The charts are based on information provided

monthly by TRs to ESMA in the context of periodic

information (Item 36 – Reconciliation Statistics). The

figures in the charts refer to outstanding trades and are

computed as the monthly average for the reference period

June 2021 – December 2021

Source: Trade Repositories & ESMA calculations

8. Methodological Annex

EMIR reporting trends and data quality metrics

Data reporting – key trends: ESMA monitors key trends in the reporting volumes by performing a

count of all daily submissions and open derivatives for a given reference date by action type, asset

class and contract type.

Data completeness, timeliness, and availability – execution vs. reporting timestamps: ESMA

measures the timeliness of reporting by counterparties by applying the following four assumptions: (1)

derivatives concluded at time T and reported at T+1 at the latest, are considered “On Time”, (2)

derivatives concluded at time T and reported after T+1 are considered “Late Reporting”, (3) derivatives

concluded at time T and reported before T are considered “Early Reporting”, and (4) derivatives

concluded on a Friday or Saturday and reported on the consecutive Monday are subject to a “weekend

effect” which is accounted for in the calculation and correctly classified as “on time”. Submissions with

action type N (New) or P (Position component) reported at transaction level (Level = T) are used for

this analysis. For each submission in the sample, we compute the difference between the “Reporting

Timestamp” and the “Execution Timestamp” expressed in days. Public-, national- and bank holidays

(i.e., “calendar effect”) are not accounted for in the calculation, nor is the conversion of Coordinated

Universal Time (UTC) to local time made. These approximations simplify and speeds up the calculation

but could give rise to some degree of inaccuracy (i.e., records wrongly classified as “Late Reporting”

due to UTC vs local time differences, or due to calendar effect) impacting the overall results.

Data completeness, timeliness, and availability – non-reporting: ESMA estimates the number of

non-reported derivatives by counting a number of open derivatives reported between a counterparty

pair (i.e., EMIR fields ‘Reporting counterparty ID’ and ‘ID of the other counterparty’) in both directions

(i.e., CP1 vs. CP2 and CP2 vs. CP1) and taking a difference in those instances where open derivatives

were reported only in one direction. Non-EEA counterparties and open derivatives with non-LEIs in ID

of the other counterparty are excluded from the calculation. Member State of non-reporting is identified

by the country of LEI in ‘ID of the other counterparty’ using the GLEIF reference data.

Data completeness, timeliness, and availability – non-reporting of valuations by counterparties:

ESMA measures non-reporting of valuations by counterparties by analysing all open derivatives with

action type = N (new) or V (valuation) and clearing threshold = Y. The difference in number of days is

computed between the reference date of the TSR and the valuation timestamp of a record. Each record

is grouped into buckets (0-15 days, 16-30 days, 30-365 days, and >365 days) to measure the frequency

of valuation updates.

Data accuracy – adherence to format and content – revalidation and rejection rates: ESMA

performs a data revalidation process on the daily submissions to detect data quality issues linked to the

validation process of TRs. The analysis uses a randomly selected data sample (~15%) extracted from

one daily submission report per month, per TR. Each data point is checked against the current ESMA

validation rules.

Rejection statistics produced by TRs are aggregated by ESMA and used to monitor how many reports

are being rejected by TRs due to misreporting by CPs.

Data integrity – reconciliation: ESMA performs reconciliation process on open derivatives on a given

reference data by replicating the process applied by the TRs. Firstly, non-EEA open derivatives are

excluded from reconciliation.

Pairing is performed by finding second leg of each derivative by using a unique key (i.e., EMIR fields

‘Reporting counterparty ID’, ‘ID of the other counterparty’, and ‘Trade ID’). The second leg of a derivative

is found by looking CP1-CP2-TradeID vs CP2-CP1-TradeID. Both sides of each derivative are counted

towards the aggregate values.

The difference in the number of reported derivatives is calculated by counting open derivatives reported

between a counterparty pair (i.e., ‘Reporting counterparty ID’ and ‘ID of the other counterparty’) in both

directions (i.e., CP1 vs. CP2 and CP2 vs. CP1) and taking a difference.

The analysis on inter-TR reconciliation issues is based on data submitted by TRs monthly according to

the Guidelines on Periodic Information

. The scope of this analysis includes only information related to

open derivatives. The reference period of the analysis provided in this report is June’2021-

December’2021. The relevant values that are considered are the following:

- The number of UTIs each TR reports to have paired/reconciled with all other TRs (A)

- The number of UTIs all other TRs report to have paired/reconciled with that specific TR (B)

- The difference between the number of UTIs reported by the TR and the other TRs (A – B)

- The percentage difference [(A – B)/A]

Those values have been calculated for each month of the reference period and have been averaged

out.

SFTR reporting trends and data quality metrics

Data reporting – key trends: ESMA monitors key trends in the reporting volumes by performing a

count of all daily submissions and open SFTs for a given reference date by action type, type of SFT

and sector of the reporting counterparties.

Data accuracy – adherence to format and content: Total number of accepted and rejected SFTs is

computed from dedicated TR regulatory reports containing aggregated as well as SFTs level

information on rejected and accepted SFTs submitted to TRs by counterparties.

Duplicated records are identified using counterparty ID, ID of the other counterparty and trade ID. In the

cases of collateral components master agreement type is used as well. The analysis is performed on

the trade state report containing all open SFTs on a given date.

Data completeness, timeliness, and availability – event date vs. reporting timestamps: ESMA

measures the timeliness of reporting by counterparties by applying the following four assumptions: (1)

new SFTs concluded at time T and reported at T+1 at the latest, are considered “On Time”, (2) new

SFTs concluded at time T and reported after T+1 are considered “Late Reporting”, (3) new SFTs

concluded at time T and reported before T are considered “Early Reporting”, and (4) SFTs concluded

on a Friday or Saturday and reported on the consecutive Monday are subject to a “weekend effect”

which is accounted for in the calculation and correctly classified as “on time”. For each submission in

the sample, we compute the difference between the “Reporting Timestamp” and the “Event Date”

expressed in days. Public-, national- and bank holidays (i.e., “calendar effect”) are not accounted for in

the calculation, nor is the conversion of Coordinated Universal Time (UTC) to local time made. These

approximations simplify and speeds up the calculation but could give rise to some degree of inaccuracy

(i.e., records wrongly classified as “Late Reporting” due to UTC vs local time differences, or due to

calendar effect) impacting the overall results.

Data integrity – reconciliation: Pairing and reconciliation flags are calculated using dedicated fields

with respective reconciliation statuses in the trade state report. Only records where both counterparties

The item analysed is “Item 36a – Reconciliation Statistics”. For more information, please consult the “Guidelines on periodic

information” and the Data Reporting templates on the ESMA web page.

have a reporting obligation are considered. Information on the jurisdiction of the two counterparties to

the transaction (reporting counterparty ID and ID of the other counterparty) is obtained from GLEIF.

The analysis on inter-TR reconciliation issues is based on data submitted by TRs on a monthly basis

according to the Guidelines on Periodic Information

. The scope of this analysis includes only

information related to outstanding SFTs. The reference period of the analysis provided in this report is

June’2021-December’2021. The relevant values that are considered are the following:

- The number of UTIs each TR reports to have paired/reconciled with all other TRs (A)

- The number of UTIs all other TRs report to have paired/reconciled with that specific TR (B)

- The difference between the number of UTIs reported by the TR and the other TRs (A – B)

- The percentage difference [(A – B)/A]

Those values have been calculated for each month of the reference period and have been averaged

out.

The item analysed is “Item 36b – Reconciliation Statistics”. For more information, please consult the “Guidelines on periodic

information” and the Data Reporting templates on the ESMA web page.

9. List of abbreviations

AFM

Autoriteit Financiële Markten

AMF

Autorité des Marchés Financiers

BaFin

Bundesanstalt für Finanzdienstleistungsaufsicht

BIS

Bank for International Settlements

CBol

Central Bank of Ireland

CCP

Central Counterparty

CNB

Česká národní banka

CSSF

Commission de Surveillance du Secteur Financier

Credit Derivatives

CDS

Credit Default Swap

CFD

Contract for Difference

Clearing Member

CME

CME Trade Repository Ltd. (CME TR)

Commodity Derivatives

CSD

Central Securities Depositories

Counterparty

Currency Derivatives

DDRIE

DTCC Data Repository (Ireland) Plc

DDRL

DTCC Derivatives Repository Plc

DQAP

Data Quality Action Plan

DQEF

Data Quality Engagement Framework

DQR

Data Quality Review

EEA

European Economic Area

EMIR

European Markets Infrastructure Regulation

Equity Derivatives

ESMA

European Securities and Markets Authority

ETD

Exchange Traded Derivatives

Financial Counterparty

FSB

Financial Stability Board

GLEIF

Global Legal Entity Identifier Foundation

HHI

Herfindahl-Hirschman Index

ICE

ICE Trade Vault Europe Ltd. (ICE TVEL)

IORP

Institutions for Occupational Retirement Provision

IRD

Interest Rate Derivatives

IRS

Interest Rate Swaps

ISDA

International Swaps and Derivatives Association

KDPW

Krajowy Depozyt Papierów Wartosciowych S.A.

LEI

Legal Entity Identifier

MIC

Market Identifier Code

MiFIR

Markets in Financial Instruments Regulation

MFSA

Malta Financial Services Authority

NCA

National Competent Authority

NFC

Non-Financial Counterparty

OTC

Over-the-Counter

REGIS

REGIS-TR

REPO

RTS

Repurchase Agreement

Regulatory Technical Standard

SFT

Securities Financing Transaction

SFTR

Securities Financing Transactions Regulation

SLEB

Securities and Commodities Lending or Borrowing

TAR

Trade Activity Report

Trade Repository

TSR

Trade State Report

UCITS

Undertakings for Collective Investment in Transferable Securities

UNAVISTA B.V

UnaVista TRADEcho B.V. (The Netherlands)

UNAVISTA LTD

Unavista limited

UTC

Coordinated Universal Time

UTI

Unique Trade Identifier

XML

Extensible Markup Language

Countries abbreviated according to ISO standards

Currencies abbreviated according to ISO standards

c6dc753b04500f08d4070e0f17cef281