No part of this publication may be reproduced or transmitted in any form or by any means without permission from ICMA. This report is intended for general information only and is not intended to be nor should it be relied upon as being legal, financial, investment tax, regulatory, business or other professional advice. Users of this report should seek appropriate independent advice before entering into any kind of specific transaction. While the information contained in this report is taken from sources believed to be reliable, neither ICMA nor the author represents or warrants that it is accurate, suitable or complete and neither ICMA nor the author shall have any liability arising from or relating to the use of this report and its contents.
Executive Summary
In June 2025, the European Repo and Collateral Council (ERCC) of the International Capital Market Association (ICMA) conducted the 49th in its series of semi-annual surveys of the repo market in Europe.
The survey asked a sample of financial institutions in Europe, among other things, for the value and breakdown of their repo contracts that had been transacted before close of business on June 11, 2025, and were outstanding after that date. Replies were received from 59 entities, mainly banks.
Data were also reported separately by the principal automatic inter-dealer repo trading systems (ATS), automated dealer-to-customer repo trading platforms, central counterparties (CCP) and tri-party repo agents in Europe, giving the size and composition of almost all electronic repo trading, central clearing and tri-party repo collateral management in the region.
Total repo business
The total value, at close of business on June 11, 2025, of repos and reverse repos still outstanding on the books of the 59 entities who participated in the latest survey rebounded to a record high of EUR 12,435.9 billion from an updated EUR 10,956.7 billion in 2024. This represented a jump of +13.5% since the previous survey and +11.9% year-on-year.
The driver seems to have been the financial market volatility and economic uncertainty triggered by the shock of hikes and threats of hikes in trade tariffs by the US Administration, which increased demand for precautionary liquidity but also encouraged investors to seek shelter in the money market, not least in repo.
The demand for repo was manifest in sharp increases in both gross repo and reverse repo positions, resulting in a recovery in the survey samples’ net reverse repo position. The recovery in the net reverse repo position is contrary to what might have been expected following the transition in monetary policy from QE to QT.
Trading analysis
The decline in the share of outstanding repos traded on automatic trading systems (ATS) from the recent peak in June 2023 was reversed by a recovery in the share of the euro and continued growth in the trading of peripheral eurozone government securities. The growth in the share of voice-brokered repo accelerated, suggesting more inter-dealer activity in the OTC market.
The value of tri-party repo positions held by the survey sample reached an all-time high of EUR 1,267.3 billion, but their share of the survey fell back, as other market segments grew faster.
Within tri-party repo, the share of GC financing positions reported by the survey sample was unchanged but its share of ATS-traded repo and of tri-party repo positions declined. GC financing remains dominated by Eurex GC Pooling, but growth over the last year was boosted by the vigorous revival of LCH’s €GCPlus service.
Geographical analysis
The largest share of the repo positions held by the survey sample continued to have been executed cross-border into and out of the eurozone and this segment continues on an uptrend. The share of anonymous (CCP-cleared) positions remained steady. There was an unexpected and unexplained jump in the share of positions between non-eurozone counterparties. Positions with APAC counterparties continued to retreat, probably reflecting the declining importance of yen and JGB repos.
Clearing analysis
CCP-cleared positions held by the survey sample benefited from the recovery in ATS-traded repo.
Cash currency analysis
The share of the euro in the survey recovered, as the trend growth in the share of the US dollar abated and the share of the yen fell back. There was also a recovery in the share of APAC currencies other than the yen.
Collateral analysis
US Treasuries remained the largest collateral holding of the survey sample, followed by Italian government securities and, notwithstanding a decline in share, UK gilts. The growth of US Treasuries, and Italian and Spanish government securities, offset an acceleration in the declining share of German government securities and a retreat by French government securities, as fiscal expectations diverged. Consequently, the share of core eurozone government securities continued to contract, while the share of peripheral government securities continued to expand. Overall, the share of government securities as a percentage of all European issues was little changed, at 86.2%.
There was a further but sharper contraction in JGBs.
Securities issued by the EU and held as repo collateral by the survey sample grew slightly to 0.5% of collateral holdings. It is somewhat larger in tri-party repo, at 5.8%. The small size of holdings by the survey sample suggests limited interdealer trading of EU securities.
In tri-party repo, covered bonds remained the next largest component of tri-party repo.
Also in tri-party repo, there was a generally modest but widespread relaxation in average haircuts, especially, for agencies, corporate bonds and collateralised securities. In the case of government and financial corporate bonds, this was a reversal of the change seen in December. In contrast, haircuts deepened for equities and convertible bonds, and, to a lesser extent, for covered bonds and MBS. These latter changes may have reflected high corporate valuations and growing concerns over real estate, respectively.
Contract analysis
Repos that have been guaranteed or indemnified (including the various forms of “sponsored” repo) continued to grow, reaching 5.7% of the outstanding positions of the survey sample.
Repo rate analysis
The share of floating-rate repo --- which started to expand in 2020 in response to central bank interest rate hikes but fell back as rates were cut --- recovered in June 2025, reflecting expectations that the end of the recent cycle of official rate cuts was approaching in the face of growing concerns over long-term inflationary expectations.
Maturity analysis
There was a shift back into short dates (one month or less remaining to maturity) in line with established seasonal patterns across all segments. This shift was especially pronounced into short-dates with more than one day remaining to maturity and out of positions with one to three months remaining to maturity and forward positions. As a consequence, the weighted-average residual term-to-maturity of outstanding repos on the books of the survey sample shortened to 28-64 days from 29-65 days in December.
The survey sample has historically provided significant maturity transformation to the rest of the market. The degree of this transformation was accentuated in the latest survey as the survey sample became a net cash borrower at a residual maturity of one day and a net lender across all other terms.
Product analysis
The share of securities lending executed on repo desks fell back to 8.1% of the combined securities financing positions on repo desks.
Concentration analysis
The share of the survey sample taken by the top 10 respondents in the survey recovered in June, at the expense of respondents with the smallest 80% of repo books. However, this translated into only a modest firming of the index measuring the degree of competition among the survey sample.
Chapter 1: The Survey
On June 11, 2025, the European Repo and Collateral Council (ERCC) of the International Capital Market Association (ICMA) conducted the 49th in its series of semi-annual surveys of the repo market in Europe. The first of these surveys took place in June 2001 and the series now charts an unrivalled history of the development of the core segment of the European repo market over two and a half decades, during which time, the market has burgeoned and matured into an efficient and essential component of the regional financial system, while coping with several episodes of severe economic and financial turbulence.
The survey was carried out and the results analysed on behalf of ICMA by the author, Richard Comotto, under the guidance of the ERCC Committee.
1.1 What the survey asked
The survey asked financial institutions operating in Europe, who are members of ICMA, for the starting value of the cash side of repos and reverse repos that were still outstanding at close of business on Wednesday, June 11, 2025. The survey therefore measures the stock, or outstanding balance, of transactions that have not matured or been terminated by the survey date. Except for two questions (see below), it does not seek to measure the flow of transactions, or turnover, over the period between two successive survey dates.
The survey covers all types of true repo, which means agreements in which collateral is sold and repurchased, in other words, where collateralisation is by the transfer of legal title to the collateral rather than by the creation and attachment of a security interest, such as a pledge. Repo can take the form of repurchase transactions, reverse repurchase transactions, buy/sell-backs and sell/buy-backs. The survey does not cover synthetic structures.
The survey asked respondents to divide their data into repo (cash borrowing by the respondent) and reverse repo (cash lending by the respondent), as well as to break down the resulting positions by:
• location of their counterparty
• market segment
• cash currency
• type of contract
• type of repo rate
• remaining term-to-maturity
• method of collateral management
• origin of collateral and
• some other categories.
In addition, participants were asked to report the outstanding value and composition of any securities lending and borrowing conducted from their repo desks.
Since 2017, the survey has asked for the number of new transactions and the value of turnover since the previous survey (these are the only questions in the survey which measure turnover). Since 2019, it has also asked for the numbers and types of legal agreements under which entities can transact repos.
Since June 2023, questions have been included in the survey about sponsored, guaranteed and indemnified repo.
An extract of the accompanying Guidance Notes for survey respondents is reproduced in Appendix A.
As well as reports sent by respondents, data have also been contributed directly, since 2003, by the principal automatic dealer-to-dealer (D2D) repo trading systems (ATS) and by the main tri-party repo agents in Europe.12 The latter have also reported tri-party securities lending since 2016. Data is now also provided separately by the two principal automated dealer-to-customer (D2C) repo trading systems in Europe.3 These direct sources of data cover almost the entire populations of electronic trading and tri-party management providers in Europe, against which the size, composition and changes in electronic trading and tri-party repos executed by the survey sample can be placed in context. Members of the Wholesale Market Brokers’ Association (WMBA) contributed data on voice-broking directly between 2002 and 2017.
1.2 The response to the survey
The latest survey was completed by 59 entities belonging to 51 financial groups. Two respondents (a subsidiary located in the Netherlands and a party headquartered in Sweden) dropped out of the survey. No new participants joined, nor did any former participants rejoin. There was therefore a net reduction of two in the size of the survey sample between December 2024 and June 2025.
Of the 59 respondents in the latest survey, 43 were headquartered across 15 European countries, including members of the EU (35), Norway (1), Switzerland (1) and the UK (6). The EU respondents were headquartered across 12 of the 27 member states (there were again no respondents in the survey from Finland or Sweden, and none from the former Accession States). 31 EU respondents were headquartered across 10 of the 20 countries of the eurozone (which added Croatia in January). Other survey respondents were headquartered in Japan (5) and North America (12). 19 respondents were branches or subsidiaries of foreign parents, most of which (16) continued to be located in the UK.
Many respondents provided data for their entire European repo business. Others made separate returns for one or more (but not necessarily all) of their European offices. Respondents were asked to report for both their UK and EU offices, where they have divided their European business post-Brexit. A list of the respondents that have participated in the ICMA’s repo surveys is contained in Appendix B.
1.3 The next survey
The next survey is scheduled to take place at the close of business on Wednesday, December 10, 2025. Forms will be sent out at least two weeks in advance.
Any financial institution wishing to join the next survey will be able to download copies of the questionnaire and accompanying Guidance Notes from ICMA’s website.
Entities who participate in the survey will receive a confidential list of their rankings across the main survey categories.
The data received in the survey are used for no other purpose than to inform the survey report. Individual returns are seen only by the author. Only aggregated data is published and ICMA is not permitted to disclose data reported by individual respondents.
Questions about the survey should be sent by e-mail to reposurvey@icmagroup.org
1 The reporting ATS were BrokerTec (CME), two eRepo (formerly TP Repo) platforms, two Eurex platforms, MTS Repo (Euronext) and SIX (which is not an ATS but has been included for convenience). Only Dealerweb did not report.
2 The reporting tri-party agents were Bank of New York Mellon, Clearstream Banking Luxembourg, Euroclear Bank, JP Morgan and SIS, who together account for the bulk of tri-party business in European repo. Agents not reporting included Citibank and Euroclear UKI (Crest), but the latter provided an estimate of the total tri-party repo positions managed by their DBV (Delivery-By-Value) service.
3 Tradeweb has provided data since 2020 and GLMX, who joined the survey in December 2023, from 2022.
Chapter 2: Analysis of Survey Results
The aggregate results of the latest two surveys (December 2024 and June 2025) and of the surveys in each June in the three previous years (2022-2024) are set out in Appendix C. The full results of all previous surveys can be found on ICMA’s website
Total repo business (Q1)
The total value, at close of business on June 11, 2025, of repos and reverse repos outstanding on the books of the 59 entities who participated in the latest survey rebounded to a record high of EUR 12,435.9 billion after having fallen back to a revised EUR 10,956.7 billion in December 2024.4 This represented a jump in the aggregate size of the repo books of the survey sample of +13.5% since the previous survey and +11.9% year-on-year (from EUR 11,114.3 billion in June 2024), compared with a revised contraction of -1.4% and growth of +0.5%, respectively, in the December 2024 survey. The growth in the repo books of the survey sample was against the decelerating trend which started in June 2023 (see Figure 2.1).
Figure 2.1 – Outstanding value of total business by the survey sample
The survey sample as a whole has been a net lender of cash to (and therefore a net borrower of collateral from) the rest of the repo market continuously since 2012, when central banks started to offer longer-term liquidity to sustain the European financial markets in the aftermath of the eurozone sovereign debt crisis (see Table 2.1).
From 2016 until the December 2022 survey, the net cash lending position of the survey sample trended up, reflecting abundant central bank liquidity, including that provided after the “dash for cash” at the start of the Covid-19 pandemic lock-down (see Figure 2.2). By December 2022, net cash lending by the survey sample had reached the equivalent of 5.9% of the total outstanding value of the survey (EUR 617.3 billion).
In June 2023, there was a sharp contraction in the net reverse repo position of the survey sample to 2.6% of the total outstanding value of the survey (EUR 325.9 billion). This reversal was driven by the redeployment of the balance sheets of major dealers from Europe to the US and Asia, in pursuit of more remunerative trading opportunities. It may also have been reinforced by the pivot by central banks to quantitative tightening (QT) and the consequent flow of collateral securities back into and cash out of the market.
However, the contraction was temporary. In the December 2023 survey, net cash lending by the survey sample recovered to 4.8% of the survey, as the drop in the share of reverse repo and the recovery in repo in June were retraced. On the other hand, in terms of absolute size, the net reverse repo position of the survey sample was not fully restored in December 2023, only reaching EUR 598.1 billion, compared with EUR 617.3 billion in December 2022 (prior to the contraction).
In the June 2024 survey, the sample’s net reverse repo position began to shrink again, albeit more gradually than in June 2023, falling to EUR 546.7 billion (4.9% of the survey total). In the December 2024 survey, the net reverse repo position of the survey sample fell to a corrected 3.9% and seemed to be following a downward trend. The shift away from collateral borrowing and securities-driven repo that is implied by the contracting net reverse repo position is what would have been expected in H2 2024, given the switch to QT, albeit that there were many other factors at work in both directions.
In the latest survey, there were sharp increases in both gross repo and reverse repo positions, probably reflecting the market volatility and macro-economic uncertainty triggered by the cycle of trade tariff announcements by the US, culminating on 4 April (“Liberation Day”), when the VIX index hit a 5-year high. The shock to the market not only boosted demand for precautionary liquidity but also prompted investors to seek shelter in the money market, where attractive yields were also available. Repo benefited directly and indirectly through vehicles such as money market funds and ultra-short bond ETFs. The enhanced demand for precautionary liquidity resulted in a rebound in the survey samples’ net reverse repo position.
Figure 2.2 – Total repo versus reverse repo positions of the survey sample
ICMA survey methodology
The survey measures the value of outstanding transactions at close of business on the survey date. While the measurement of the flow of new repo between two dates is useful for some business and market analyses, the stock of transactions outstanding on one date was adopted because it gauges risk exposure and open interest in the market.
Note that outstanding value understates the share of shorter-term repos, given that such transactions run off faster between surveys than longer-term repos. Because repos traded on automatic trading systems (ATS) and cleared on a central counterparty (CCP) are typically very short-term, the consequence is that the share of outstanding balances that have been electronically-traded and centrally-cleared is smaller than their share of turnover. This can be seen by comparing published aggregate SFTR data on new and outstanding repos.
It also needs to be remembered that changes in outstanding balances can reflect cumulative changes in turnover or variations in the tenors of new transactions or both.
Another important feature of the survey methodology is that it recognises repos from their transaction dates (when they are executed by the two parties and contracts are formed), rather than from their value or purchase dates (when cash and collateral are first due to be exchanged). This transaction-date basis means that the outstanding value measured by the survey includes forward repos, as well as unsettled new non-forward transactions. The latter include one-day repos that are transacted on or before the survey date, but that are not due to be settled until the business day after the survey date (tom/next repo) or on the following business day (spot/next repo). This gives greater weight to one-day repo than would measurement on a value-date basis.
The values measured by the survey are not adjusted for the reporting of the same transaction by two respondents. However, a study by the author (see the report of the December 2012 survey) suggested that inflation due to this problem of double-counting may not be very significant. Interestingly, a trade repository in Europe has estimated that two-sided reporting has been less than 30% under the EU Securities Financing Transactions Regulation (SFTR) and less than 15% under UK SFTR, which is consistent with the author’s estimate of double-counting.
The survey does not measure the value of repos transacted with central banks as part of their monetary policy operations, but it should include their investment operations in the repo market with survey respondents.
Growth in market size
In order to accurately gauge the growth of the European repo market (or at least that segment represented by the survey sample), it is not valid to simply compare survey totals. Some changes may represent the entry or exit of respondents into and out of the survey, mergers between respondents or the reorganisation of repo books across respondent banking groups. To offset the impact of changes in the structure and composition of the survey sample, comparisons are also made of the aggregate outstanding positions reported by a sub-sample of those entities who have participated continuously over several surveys.
In the June 2025 survey, the books of the sub-set of 58 survey respondents who took part in the latest three surveys grew by +13.5% since December 2024, the same rate as the full survey sample, and by +12.3% yearon-year, compared with +11.9% in the full survey sample. These similarities in the rates of change reflected the minor changes in the survey sample.
Between December 2024 and June 2025, 41 of the 59 entities who responded to the latest survey expanded their repo books (up significantly from 32 out of 61 in June). The repo books of another 15 respondents contracted over the same period (down from 27 respondents in the previous two surveys). This expansion of repo books was unprecedented in its breadth.
The scale of change across the repo books of the survey sample was also significant. The median position-weighted change in all repo books continued to increase, rising to +10.1% from +4.1%. Moreover, the unweighted-mean change for the respondents who expanded their repo books jumped to +58.8% from a revised +4.0% in December. For those who contracted their books, the change decreased to -12.3% from a revised -17.6% in the previous survey. The weighted-mean change across all repo books also jumped, to +17.1% from a revised +3.8%.
Trading analysis (Q1.1)
Figure 2.3 – Trading analysis of the survey sample
Direct (excluding triparty)
Voice-brokered
ATS
Triparty
The decline in the share of outstanding repo traded on automatic trading systems (ATS) from its recent peak in June 2023 was reversed.5 The growth in the share of voice-brokered repo accelerated, suggesting more interdealer activity in the OTC market. The counterparts were relapses in directly-transacted business, both tri-party and bilaterally-managed, which are likely to have mainly been D2C.
Tri-party repo
The value of tri-party repo positions held by the survey sample reached an all-time high of EUR 1,267.3 billion, up +11.5% from a revised EUR 1,136.8 billion in December, but the share of the survey fell back to 10.2% from a revised 10.4% in December, as other market segments outpaced tri-party repo.6
Gross cash borrowing by the survey sample as a share of the sum of gross tri-party borrowing and lending rebounded to 76.3% from 73.1%. Net cash borrowed through tri-party repo by the survey sample recovered to the equivalent of 5.4% of the total survey size from 4.8% in December. Increased tri-party net borrowing therefore helped fund the increased net reverse repo position of the survey sample.
Tri-party repo also grew rapidly outside the survey sample. The value of outstanding tri-party repo reported directly to ICMA by the two ICSDs (International Central Securities Depositories) and SIS (in Switzerland) --- as opposed to the positions reported by the survey sample --- recovered sharply in H1 2025, rising by +15.3%, compared with +1.2% in the previous semester, to reach EUR 937.7 billion from EUR 813.0 billion in December.7 The rate of growth in tri-
5 Automatic trading systems (ATS) are electronic trading venues that operate central limit order books (CLOB), which automatically or semi-automatically match orders entered into the book. ATS in Europe are interdealer markets. They include CME BrokerTec, Dealerweb, TP ICAP eRepo, Eurex Repo and Euronext MTS, all but one of whom report data directly to ICMA..
6 The actual share of tri-party repo in the survey sample is thought to be higher than the reported share, given that some respondents, who are known to use triparty services, did not report their tri-party business.
7 The ICSDs are Clearstream Banking Luxembourg and Euroclear Bank.
party repo managed by global custodians was somewhat slower at +9.4%. Nevertheless, the total of all tri-party repo positions reported directly to ICMA by tri-party agents reached a record level of EUR 1,176.0 billion.8
Table 2.2 – Comparative trading analysis
GC financing9
The value of GC financing positions reported by the survey sample grew by +14.2% over H1 2025, but its share of the repo books of the survey sample was unchanged at 1.4%. The value of GC financing and in its share of electronic trading reported directly to ICMA by the principal ATS in Europe fell to 8.9% from 10.3%. The share of triparty positions reported directly to ICMA by the ICSDs and SIS fell slightly to 27.8% from 28.1%. On the other hand, the share of tri-party positions reported by the survey sample that were from GC financing increased to 13.9% from 13.6% in December.
The provision of GC financing reported directly by the ICSDs and SIS became significantly less concentrated over the last two surveys with the rapid growth of LCH’s GC financing service, notably its €GCPlus platform (to be renamed GCPlus). This accounted for 10.6% of the outstanding value of all reported GC financing, up from 3.6% in December and 0.7% in June 2024 (see Figure 2.4). Growth was therefore +203.6% in H1 2025 over H2 2024 and +1249.4% over H1 2024. The boost to €GCPlus/GCPlus seems to have been spurred by the merger of its default fund with that of the main RepoClear service, which will have reduced margining.
Figure 2.4 – Monthly cleared nominal turnover on LCH SA €GC Plus
Source: LCH SA. Note: single-counted, nominal value of collateral
8 The value of outstanding tri-party repo reported by the survey sample (EUR 1,267.3 billion) is larger than that reported by the tri-party agents (EUR 1,176.0). This difference is likely to reflect a number of factors, including the use of US tri-party repo by the survey sample and DBV tri-party repo in the UK, neither of which are reported directly to ICMA (the latter was GBP
Table 2.3 –
Automatic trading systems (ATS)
The share of outstanding repo transactions that had been transacted by the survey sample across ATS recovered to 26.5% from 26.0% in December. In addition, the value of outstanding ATS-traded repos reported directly to ICMA by the principal platforms in Europe reached an all-time high of EUR 1,862.9 billion, compared with EUR 1,575.6 billion in December, growth of +18.2%, faster than the growth rate of the survey sample.
The last survey suggested that the contraction in the share of ATS repo over H2 2024 may have reflected the increase in US dollar repo trading by the survey sample, as this would have bypassed Europe ATS, given their limited US dollar trading (0.2%). The recovery of European ATS in H1 2025 is consistent with that hypothesis, given the deceleration in the growth of the share of US Treasuries and the contracting share of US dollars on the books of the survey sample (see the section below on the currency and collateral composition of the latest survey).
The latest reports made directly to ICMA by European ATS showed continued strong growth in the trading of Italian government securities (for which trading is concentrated on ATS). Their share expanded to 46.8% of outstanding ATS-traded repos, from 42.9% in December (+29.0%). There was also vigorous growth in ATStraded repo positions in government securities issued in Belgium (by +17.0% to a share of 4.5%), France (+18.6% to 13.5%) and Spain (+20.8% to 7.2%). There was no growth in the share of German government securities (for which the share remained 12.4%) and a fall in the share of UK gilts (by -6.8% to 8.0%), which was mirrored in a contraction in CCP-clearing by LCH Ltd. Among smaller issuers, there were jumps in the share of Portuguese and Polish government securities (respectively, by +139.1% to 1.7% and +86.5% to 0.2%). One of the drivers in the case of Portuguese government securities was likely to have been its ratings upgrade to A in January.
Non-government securities accounted for 0.1% of the positions reported by ATS (the most recent high was 0.8% in June 2019).
A further insight into ATS-traded repo can be gleaned from the monthly turnover figures published by BrokerTec and MTS, the two largest repo ATS in Europe. These continued to show a changing picture in interdealer electronic trading (see Figure 2.5).
At BrokerTec, activity continues to fluctuate but around a rising trend, making up some ground after the stepdown in activity in Q3 2024. Growth in average monthly turnover on BrokerTec was +1.8% semester-on-semester and +4.1% year-on-year. Growth in MTS continued on a robust upward trend, +6.7% semester-on-semester and 24.7% year-on-year, respectively.
BrokerTec may have continued to suffer from lacklustre trading in German and UK government securities, in which it has the largest market shares, but also from intensified competition from both established and new competitors. MTS, on the other hand, will have continued to benefit from its dominant position in Italian government bond repo, and its foothold in the Polish market.
At Eurex, which is the third major repo ATS in Europe, monthly term-adjusted average daily turnover on Eurex Repo Special and GC (ERSGC) surged to its highest level since the all-time peak reached in October 2023 (see Figure 2.6a). Eurex Repo GC Pooling (GCP) spiked in January, as it recovered from the usual year-end hiatus, then fell back in February, before overtaking ERSGC to reach a record high in June.
Monthly term-adjusted average daily turnover on ERSGC averaged EUR 177.9 billion over H1 2025, up +9.2% from a revised EUR 162.9 billion over the previous half-year, while GCP averaged EUR 192.9 billion per day, up +29.1% from an updated EUR 149.4 billion.
BrokerTec
2.6a – Monthly term-adjusted turnover on Eurex repo trading systems
Sources: Eurex, author’s calculations
On an estimated unadjusted basis, the surge in GCP can be seen to have been driven by a dramatic extension in the term-to-maturity of transactions, reminiscent of what was seen in 2023 (see Figure 2.6b). The term-adjustment multiplier increased, particularly in Q2 2025, to 9.2 in June from 5.1 in December. The increased term-to-maturity of positions at GCP likely reflected the supranational institutions and official agencies in its user base, given that these entities prefer longer-than-average investment tenors.
Figure 2.6b –
Sources: Eurex, STOXX, author’s calculations
Turnover on MTS --- which publishes monthly turnover data that are both unadjusted and term-adjusted --- also showed a recovery in unadjusted activity in January and a strong surge in term-adjusted activity over the next three months, that was largely driven by longer terms, to reach a record high in April (see Figure 2.7). The ratio of term-adjusted to unadjusted turnover at MTS peaked at 3.2 but the trend remained around 2.5.
Figure
Figure 2.7 – Average daily term-adjusted and unadjusted turnover on Euronext MTS
Sources: Euronext
ATS-traded repo, as reported directly to ICMA by the ATS, remained overwhelmingly CCP-cleared, although the share of CCP-clearing fell back to 95.6% from 96.3%. However, excluding SIS (which is not CCP-cleared) increases the share of CCP-cleared repos to 98.7%, although this was down from the 99.9% in December.
Automated trading systems
ICMA receives position data directly from both of the principal automated repo trading systems operating in the dealer-to-customer (D2C) space in Europe. These are GLMX and Tradeweb. The two platforms account for most of automated D2C trading in Europe.10 Their combined growth over H1 2025 was +20.3% in turnover semesteron-semester and +18.8% in outstanding value. This compared with +25.9% and 18.8%, respectively, over H2 2024 and +37.4% and +25.1%, respectively, over H1 2024. Turnover therefore continues to decelerate, but outstanding balances have recovered, suggesting a lengthening of average terms. Moreover, the rates of growth of the automated trading systems suggest that automation is continuing to make rapid inroads into the D2C market.
On Tradeweb’s European platform, strong growth in outstanding balances continued in H1 2025, to reach USD 143.2 billion in June from USD 100.6 billion in December (+31.6% since December and +58.0% year-on-year). Growth in turnover was robust over Q1 2025 (+12.4%) and, despite slowing in Q2 2025, averaged +39.7% semester-on-semester and +42.3% year-on-year (see Figure 2.8).
Figure 2.8 – Monthly turnover and outstanding value in European repo on Tradeweb
Source: Tradeweb
On GLMX, the rate of growth in turnover in Europe and in outstanding balances, having slowed in Q4 2024, accelerated in H1 2025 (see Figure 2.9). However, although turnover grew by +20.3% over H1 2025, to reach an average of USD 632 billion over Q2 2025, the exceptional growth in turnover seen in Q3 2024 meant that growth over H1 2025 could not match the +25.9% seen over H2 2024. On the other hand, outstanding balances grew +20.7% compared with +18.8% in the previous semester.
2.9 – Monthly turnover and outstanding value in European repo on GLMX
Source: GLMX
Figure
Geographical analysis (Q1.1)
Table 2.4 – Geographical analysis of the survey sample
Cross-border and anonymous (CCP-cleared) business recovered in June 2025, at the expense of domestic positions. Cross-border business into and out of the eurozone extended its upward trend to reach an all-time high of 41.7% and remains by far the largest geographical component of the survey (see Figure 2.10).
Figure 2.10 – Geographical analysis of the survey sample
Table 2.5 – Geographical comparisons in June 2025 (December 2024)
Domestic X-border to (other) eurozone
X-border to non-eurozone Anonymous
In the position data reported directly to ICMA by the ATS, there was dramatic build-up in the share of outstanding balances between non-eurozone counterparties to a record 10.1% from 3.7% (see Figure 2.11). This dampened the recovery in the share of business within the eurozone, which rose to 42.2% from 38.7% (driven by Belgian, French and peripheral country government securities).
Figure 2.11 – Outstanding value of ATS business by location of counterparties as reported by the ATS
The share of business with APAC counterparties continued to fall in H1 2025, touching 2.9% of the repo books of the survey sample, compared with 3.5% in December 2024, 4.2% in June 2024 and a peak of 6.8% in December 2022. This decline likely reflected the diminishing share of repos in Japanese yen and JGBs.
Clearing analysis (Q1.2 and Q1.8)
The outstanding value of anonymous (CCP-cleared) repo trading by the survey sample, excluding GC financing, grew strongly, by +14.9% to EUR 1,850.4 billion in June 2025 from a revised EUR 1,610.2 billion in December.
GC financing by the survey sample recovered to EUR 180.6 billion from a revised EUR 157.0 billion, but its share of the survey sample was little changed, at 1.4%.11
Overall, anonymous positions were little changed at 17.0% of the survey sample compared with a revised 16.8% in December. The buoyancy of CCP-clearing reflected the recovery in trading on ATS, demonstrating the intimate link between CCP and ATS.
Turnover reported by the CCP over H1 2025 was steadier than in the previous semester (see Figure 2.12). Data from LCH RepoClear showed growth in the monthly nominal value of CCP-cleared collateral decelerating to +4.3% from +5.9% (see Figure 2.12). However, this was the net result of a sharp deceleration on LCH Ltd (gilt repo) to -5.9% from +22.5% in H2 2024 and an acceleration to +6.4% from +3.0% on SA (euro-denominated repo).
2.12 – Monthly cleared nominal turnover on LCH RepoClear 2020-25
Source: LCH. Note: double-counted
Comparing H1 2025 with H2 2024, there was a recovery in the outstanding nominal value of collateral balances of repos that were cleared by LCH SA. These balances --- calculated using the ICMA survey methodology --grew by +7.6%, compared with -4.0% in the previous semester (see Figure 2.13). However, balances at LCH Ltd continued to fall, by -5.3% over H1 2025, compared with -5.1% over H2 2024. The overall changes at LCH were +5.9% compared with -4.1%. Given the relative under-performance of turnover compared to balances, these data imply average terms of CCP-cleared repo lengthened over H1 2025.
2.13 – Daily outstanding nominal value of cleared repos on LCH RepoClear 2018-2025
Source: LCH. Note: double-counted
Figure
Figure
At Eurex Clearing AG (ECAG) --- which is the CCP for Eurex Specials and GC Repo (ERSGC) and GC Pooling (GCP) --- and for which outstanding data using the ICMA methodology have been available since 2021, it can be seen that the outstanding nominal value of CCP-cleared collateral traded on both platforms accelerated sharply in H1 2025, by +17.2% and +16.8%, respectively, compared with +7.1% and +8.3% over H2 2024 (see Figure 2.14). This was reflected not just in an increase in turnover but, as elsewhere, in a significant lengthening of average term-to-maturity.
2.14 – Daily outstanding values of cleared repos on ECAG 2021-25
Source: ECAG. Note: double-counted, stacked area chart
Figure 2.15 shows that the share of outstanding cleared balances at LCH fell back to 83.1% of the combined outstanding value at LCH and ECAG at end-June 2025, compared with a revised 83.9% in December. These numbers exclude the share of Euronext Clearing (formerly CC&G), which largely clears Italian government repo, as data are not published for that CCP. However, a significant share of its clearing activity is thought to be passed to LCH SA via a link between the two CCPs. There are also smaller repo CCP-clearing operations in Poland and Spain.
2.15 – Daily outstanding value of cleared repos on LCH RepoClear vs ECAG 2021-25 (EUR trillion, doublecounted: calculated using same methodology as the ICMA survey)
Sources: LCH and Eurex
Post-trade clearing
While the bulk of CCP-clearing is of repos transacted on ATS, some trading is executed directly between parties and then registered with a CCP post-trade (see Figure 2.16). Having peaked at 14.1% in December 2020, the share of this post-trade clearing by the survey sample has declined since 2021 but was steady at 3.9% in June compared with December.
Figure 2.16 – Post-trade CCP-clearing by the survey sample
The share of the euro in the survey recovered as the trend growth in the share of the US dollar abated and the share of the yen fell back further. There was also a recovery in the share of APAC currencies other than the yen (see Figure 2.17).
Table 2.6 – Cash currency analysis
The shares of the euro and pound sterling seem to be trending sideways. It remains to be seen if the dollar resumes its upward trajectory.
Figure 2.17 – Currency analysis of the survey sample
Tri-party repo, as reported directly to ICMA by the ICSDs and SIS, saw the share of the euro recover strongly but, in contrast to other market segments, also the share of the yen, largely at the expense of the dollar (see Table 2.7).
Cross-currency tri-party repo, reported directly by the ICSDs and SIS, was unchanged at 22.0% (but its reporting may be erratic).
Table 2.7 – Currency comparison in June 2025 (December 2024)
1.4% (0.8%)
(0.5%)
(0.3%)
0.8% (0.7%)
1.1% (1.5%) 2.2% (0.1%) 0.8% (0.6%)
1.8% (1.6%)
Sources: Clearstream, Euroclear, SIS
(22.0%)
Collateral analysis (Q1.9)
US Treasuries continued to be the largest collateral holding of the survey sample (16.5%), followed by Italian government securities (14.4%) and UK gilts (11.4%).
Figure 2.18a – Collateral analysis of the survey sample
Other eurozone
DKK, SEK
Other EU
EU institutions UK US Japan
Other OECD IFI
Others
Despite the slight reduction in the share of the US dollar in the outstanding cash positions of the survey sample, the acceleration in the growth of the share of US Treasuries, to touch an all-time high of 16.5% from a revised 15.6% in December, widened the gap between the two US assets (see Figure 2.18b).
Figure 2.18b – Collateral analysis of the survey sample12
There was another jump in the share of Italian collateral to 15.4% from 14.6% in December, specifically, Italian government bonds, which grew to 14.4% from 13.5%, buoyed by an improving budgetary situation against the background of a positive ratings outlook and fiscal concerns about core eurozone governments.
There was a surge in the trading of German government debt in March in response to the agreement by the 12 The drop in the share of “other OECD” securities in December 2015 largely reflected the carving out of US and Japanese securities.
government to relax the “debt brake” in order to increase public expenditure in support of defence, environmental and infrastructure spending, but this had run out of steam by the time of the survey.
The share of Spanish collateral bounced back, to reach a record 6.0% from a revised 4.8% in December, with Spanish government bonds increasing their share to 5.4% from 3.9%.
Table 2.8 – Collateral analysis of the survey sample
Note: some December 2024 numbers have been revised.
The growth of US Treasuries, and Italian and Spanish securities, offset an acceleration in the decline in the share of German government securities (to 8.8% from 10.5%) and a retreat by French government securities (to 10.1% from 10.7%), reflecting the inversion of fiscal concerns abut issuers. Consequently, the share of core eurozone securities contracted to 27.6% from 29.6%, while the share of peripheral securities advanced to 22.5% from 20.4%. Overall, the share of government securities as a percentage of all European issues was little changed, at 86.2% (compared to a revised 86.5% in December).
UK securities maintained their share of collateral held by the survey sample at 12.8% for the third successive survey, but the share of UK gilts fell to 11.4% from 11.8% in December 2024, in line with the trend in core eurozone government securities.
There was a further but sharper contraction in JGBs (to 3.2% from 4.2%).
Securities issued by the EU and held as repo collateral by the survey sample grew slightly, to 0.5% of collateral holdings from 0.4% in December. It was somewhat larger in tri-party repo at 5.8%, although this fell back from 6.2% in December. The small size of holdings by the survey sample implies limited interdealer trading of EU securities. The larger holdings in tri-party repo suggest that positions in EU securities tend not to be financed by dealers. The cash-providers could be official institutions employing liquidity.
Tri-party
collateral
Figure 2.19 – Collateral analysis (selected tri-party agents) by type of asset
MBS, ABS and other securitisations
Equity
Other
In tri-party repo managed by the ICSDs and SIS, as reported directly to ICMA, there were further declines in the value of holdings of corporate bonds and covered bonds, but strong growth in the share of supranational and agencies/subnational issues (see Table 2.10 and Figure 2.20).
Table 2.10 – Tri-party repo collateral analysed by type of asset
Sources: Clearstream, Euroclear, SIS
Figure 2.20 – Historic collateral analysis (selected tri-party agents) by type of asset
In terms of trends, the share of government securities stabilised following the one-off inflow of covered and other credit bonds back into the repo market triggered by the unwinding of TLTRO III and other lending facilities at the ECB. Covered bonds remained the second largest component of tri-party repo collateral but continued to fall back from the high-water mark reached in 2023. Corporate bonds have also continued to trend down.
French non-government securities and UK gilts consolidated their positions as the largest components of tri-party collateral, as reported directly to ICMA by the ICSDs and SIS (at 14.4% and 9.0%, respectively). On the other hand, European eurobonds dropped sharply to 6.2% from 11.5%. Taking account of tri-party collateral holdings at global custodians, non-OECD Latin American securities surged to 4.8% from 0.6% and Danish non-government securities (which includes MBS) expanded to 2.9% from 2.3%.
In the tri-party collateral held by the ICSDs and SIS, US Treasuries fell back to 4.4% of tri-party repo collateral from 5.9% in December. Italian government bonds were little changed at 4.4%. French government securities recovered to 3.0% from 2.5%. JGBs were a little larger at 2.7%, compared to 2.4% in December. German government securities contracted slightly to 3.2% from 3.5%.
Table 2.11 – Triparty repo collateral analysed by issuer --- largest changes (difference in percentage of total reported tri-party balances)
The up-trend in the share of AA collateral in tri-party repo collateral holdings was extended into a fifth survey but was no longer at the expense of the share of AAA-rated securities. Instead, the shares of BBB ratings and high yield issues fell back (see Table 2.12 and Figure 2.21).
There was a generally modest but widespread relaxation in average haircuts on tri-party collateral reported directly to ICMA by the ICSDs and SIS, in particular, for agencies, corporate bonds and collateralised securities. In the case of government and financial corporate bonds, this was a reversal of the change seen in December.
In contrast, average haircuts deepened for convertible bonds, and even more so for equities, but to a lesser extent for covered bonds and MBS. These changes may have reflected concerns over high corporate and real estate valuations, respectively.
Table 2.13 – Weighted-average collateral haircuts (all tri-party agents) analysed by type of asset
In December 2023, the ICMA survey started to ask respondents for the outstanding value of repo business which they had guaranteed or indemnified, including the various forms of “sponsorship”. The share of such repos continued to increase, reaching 5.7% of the survey sample in June 2025 from a revised 4.9% in December. The share of the euro in such repo positions rose to 53.7% from 52.1%, while that of the dollar fell back to 38.1% from 39.8%. It should be noted, however, that it continues to be the case that only a small number of survey respondents have so far reported this type of activity.
The share of the ICMA Global Master Repurchase Agreement (GMRA) in the number of repo master agreements in place among survey respondents was little changed at 85.6%.
Repo rate analysis (Q1.6)
The share of floating-rate repo --- which started to expand in 2020 in response to central bank interest rate hikes --- recovered in June 2025, to 15.7%, up from 15.2% in December but well below the peak of 19.7% reached in December (see Table 2.15 and Figure 2.23). The latest change would seem to have reflected growing expectations that the recent cycle of cuts in central bank interest rates was coming to an end in the face of rising long-term inflation concerns.
Table 2.15 – Repo rate comparison in June 2025 (December 2024)
The balance between gross borrowing and lending by the survey sample using floating-rate repo shifted towards more borrowing.
Maturity analysis (Q1.7)
There was a shift over H1 2025 back into short dates (one month or less remaining to maturity) in line with established seasonal patterns. The shift was especially pronounced into short-dates with more than one day remaining to maturity. Overall, the share of short-dates jumped to 68.4% from 62.5% of the books of the survey sample, although this remained below the share of 70.2% reached in June 2024. The shift into short dates was mainly from positions with one to three months remaining to maturity and from forwards (see Table 2.16 and Figure 2.24). As a consequence, the weighted-average residual term-to-maturity of outstanding repos on the books of the survey sample shortened to 28-64 days from 29-65 days in December.13
Table 2.16 – Maturity analysis of survey sample
One-month to three-month repo positions held by the survey sample remained highly seasonal, peaking in December and troughing in June (see Figure 2.25). Significant activity in this residual-maturity band, and to a lesser extent in the three-month to six-month band, has been driven by collateral swaps. These are exchanges of securities --- often conducted by means of back-to-back repos and reverse repos --- that are used to manage buffers of high-quality liquid assets (HQLA) required under the Liquidity Coverage Ratio (LCR), especially over endyear. There continued to be seasonality (in the opposite direction) in open and short-dated repo.
Figure 2.24 – Maturity analysis of survey sample
Figure 2.25 – Maturity analysis of survey sample: non-forward terms
Repos with one day remaining to maturity have become the dominant tenor of outstanding transactions, replacing two-day to one-week terms, which tended to predominate during the period of QE (see Figure 2.26). Longer short-dates (one-week to one-month) had been trending down over the period of the survey but reversed direction post-QE.
Figure 2.26 – Maturity analysis of survey sample: breakdown of short dates plus open
The survey sample has historically provided significant maturity transformation to the rest of the market. This negative funding gap was accentuated in the latest survey as the survey sample became a net cash borrower at a residual maturity of one day and a net lender across all other terms (see Figure 2.27). The aggregate of mismatches in the maturity bands rebounded to a record 8.9% of the survey total from a revised 3.9% in December.
Figure 2.27 – Maturity analysis of survey sample: maturity transformation profile --- net reverse repo
Note: Each column represents one survey and each cluster of columns represents the change in the share of a particular tenor over surveys going back to December 2016.The red columns represent the latest two surveys.
The skew seen in December 2024 in the average term-to-maturity of ATS-traded repo positions towards a remaining-term-to-maturity of one day relaxed in June, with a sharp drop in one-day positions to a record low of 77.7% from 93.1% in December, while positions with two-to-seven days remaining jumped to 13.3% from 1.5% and remaining short-dated positions increased to 5.7% from 2.7% (see Figure 2.28).
2.28 – Maturity analysis (ATS)
Jun-25
Sources: CME, Eurex, Euronext, SIX, TP ICAP
In tri-party repo, there was a modest shift to short dates from open and, to a lesser extent, longer-term positions (see Figure 2.29).
The ICMA survey measures the securities lending conducted on repo desks as a share of all the securities financing business executed on these desks. The share of securities lending in the latest survey fell back to 8.1% from a peak of 13.7% in December 2024 (see Figure 2.30).
The share of equity loans recovered to 3.2% from 1.0% in December. Fixed-term lending fell back sharply to 59.1% from a revised 71.3%. Domestic lending recovered to 20.9% from a revised 14.4% and cross-border lending within the eurozone rebounded to 43.3% from a revised 26.8%, at the expense of cross-border lending into and out of the eurozone, which fell to 35.7% from a revised 58.8%.
2.30 – Product analysis of the survey sample
The survey sample returned to being a net borrower of securities through securities loans, except in the case of eurozone equity.
Concentration analysis
The share of the survey sample taken by the top 10 respondents in the survey recovered in June, at the expense of respondents with the smallest 80% of repo books (see Table 2.18 and Figure 2.31).
Table 2.18 – Concentration analysis of the survey sample
Figure 2.31 – Concentration analysis
Top 10
Top 11-20
Top 21-30
Remainder
The relative growth of the top two tiers of the survey sample was reflected in a modest increase in the Herfindahl Index (see Table 2.19 and Figure 2.32).
Figure
Repo Securities lending
Table 2.19 – Herfindahl Index for the survey sample14
The higher Herfindahl Index was reflected in the shift to the left in the Gini coefficient curve (see Figure 2.32).
2.32 – Cumulative distribution of market share of the survey sample
Figure
Chapter 3: Conclusion
In the first-half of 2025, the European repo market, as represented by the ICMA survey sample, diverged unexpectedly from the decelerating trend it had been following since 2023 and recorded a burst of growth to reach a record high of EUR 12,436 billion. The drivers seem to have been financial market volatility, macroeconomic uncertainty and geopolitical tensions triggered by the cycle of trade tariff announcements by the US, culminating on 4 April (“Liberation Day”), when the VIX index hit a 5-year high. The shock to the market not only boosted demand for precautionary liquidity but also prompted investors to seek shelter in the money market, where attractive yields were available. Repo benefited directly and indirectly through cash investment vehicles such as money market funds and ultra-short bond ETFs.
The European repo market may also have benefited from global banks bringing balance sheet capacity back from the US market, to where it had been shifted in 2024 in order exploit rallies in US Treasuries and equities, and because of a lack of trading opportunities in Europe. The growth in the aggregate outstanding position of the survey sample was notable for the unprecedented number of participants who expanded their repo books and the unusually large average size of these increases.
The boost to growth in the repo market was accompanied by another divergence in the survey sample’s aggregate net reverse repo position from the deceleration and decline that had been expected as a consequence of the ending of QE and the switch by central banks to QT. The survey sample lent more cash in H1 2025 but borrowed even more.
The survey sample did not make as much use of tri-party repo as the rest of the market and its share of the survey fell back, but the outstanding size of the tri-party repo market (including global custodians), as reported directly to ICMA, reached a record level of EUR 1,176.0 billion.
Within the tri-party market, GC financing almost regained the record level seen in June 2024, accounting for 27.8% of tri-party balances, boosted by more term repo from supranational institutions and official agencies and by the strong revival in LCH’s GC financing facility.
In the latest survey, growth was led, on the cash side, by the euro and, from a smaller base, APAC currencies other than the Japanese yen, mainly at the expense of the US dollar and yen.
The revival of the euro and the continued growth of the share of peripheral eurozone government bonds supported European interdealer automatic trading systems (ATS), reversing the recent decline in their share of the survey sample. Data reported directly to the ICMA, also showed an increase in the size of the outstanding positions executed on ATS by the whole market (not just the survey sample), to a record high of almost EUR 1.9 trillion. The stronger performance of ATS-traded repo supported the share of CCP-clearing. Within the ATS repo market, the relative competitive positions of the platforms continue to be affected by the shift from core to peripheral eurozone bonds and by intensified competition.
Despite the relative slowdown in OTC repo, which includes most dealer-to-customer (D2C) repo, the growth of D2C trading platforms, GLMX and Tradeweb, accelerated and it seems likely that they are continuing to make substantial inroads into the OTC repo market.
On the collateral side, growth continued to be driven by government bonds issued by Italy and other peripheral eurozone countries, in contrast to the contraction in the share of most core eurozone government bonds, reflecting the recent inversion of fiscal expectations. US Treasuries also increased their share and remain the largest component of the repo collateral holdings of the survey sample (followed by Italian government issues and then UK gilts). The bigger share of US Treasuries contrasted with the largely unchanged (although much larger) share of US dollars on the repo books of the survey sample.
The share of UK gilts fell back in H1 2025, and there was a further but sharper contraction in the share of JGB. On the other hand, there were notable jumps in the shares of Polish and Portuguese government bonds, albeit from small bases.
In tri-party repo, there was a generally modest but widespread relaxation in average haircuts, except for equities and convertible bonds and, to a lesser extent, covered bonds and MBS. These changes may have reflected, high corporate valuations and growing concerns over real estate assets, respectively.
There was an unusual change in June in the geography of ATS-trade repo positions. Outstanding repo traded on ATS between non-eurozone counterparties jumped to a record 10.1% from 3.7%. The counterpart was a contraction in the share of domestic and cross-border repo into/out of the eurozone.
Another significant feature of the June survey was a general lengthening of the average remaining term-to-maturity of repo positions. In the books of the survey sample, this was driven by a shift into longer short-dates (more than one day). In ATS-trade repo, the aggregated data reported directly to ICMA by ATS showed a sharp drop in positions with one day remaining to maturity to a record low of 77.7% from 93.4% in December and, as with the survey sample, a shift into longer short-dates. On MTS, there was a sharp extension of the average term over February-April and, on Eurex, right across H1 2025. The longer terms seen on Eurex were, in part, a reflection of the particular composition of its client base, which includes a large number of official institutions, who tend to lend for term.
The lengthening of the average remaining-term-to-maturity was also evident on the automated D2C trading platforms (GLMX and Tradeweb) and on LCH RepoClear in the differences between the growth rates in their positions and turnover. Tri-party repo, as reported directly to ICMA, showed a shift from open positions to all short-dates.
The survey sample continues to provide significant maturity transformation to the rest of the market. The degree of transformation increased in H1 2025 as the survey sample became a net cash borrower at a residual maturity of one day and a net lender across all other terms, and the aggregate mismatch across all maturity bands rebounded to a record level.
About the Author
This report was compiled by Richard Comotto, who is Senior Consultant to the ICMA’s European Repo and Collateral Council (ERCC). He is also author of the ICMA’s Guide to Best Practice in the European Repo Market, Repo FAQs, APAC Repo Market Guides, SFTR Task Force’s Repo Reporting Recommendations and CSDR Cash Penalty Best Practice Recommendations and FAQs, as well as being Course Director of the annual ICMA Professional Repo and Collateral Workshop. In addition, Richard provides technical assistance on behalf of ICMA, IMF, World Bank, Asian Development Bank and other organisations to developing repo markets around the world.
Appendix A: Survey Guidance Notes
The data required by this survey are: the total value of the repos and reverse repos booked by your repo desk that are still outstanding at close of business on Wednesday, June 11, 2025, and various breakdowns of these amounts, as well as the total value of all repos and reverse repos turned over the six months since the previous survey (which was on December 11, 2024).
Branches of your bank in other countries in Europe may be asked to complete separate returns. If your repo transactions are booked at another branch, please forward the survey form to that branch. If branches of your bank in other countries run their own repo books, please copy the survey form to these branches, so that they can also participate in the survey.
Guidance Notes
General guidance
a) Please fill in as much of the form as possible. For each question that you answer, you will receive back your ranking in that category.
b) If your institution does not transact a certain type of repo business, please enter ‘N/A’ in the relevant fields. On the other hand, if your institution does that type of business but is not providing the data requested by the survey, please do not enter anything into the relevant field. If your institution does that type of business but has no transactions outstanding, please enter zero into the relevant field.
c) You only need to give figures to the nearest million. However, if you give figures with decimal points, please use full stops as the symbols for the decimal points, not commas. For nil returns, please use zeros, not dashes or text. Do not use negative signs.
d) Please do not re-format the survey form, ie change its lay-out, and do not leave formulae in the cells of the underlying spreadsheet.
e) Include all varieties of repos, ie repurchase transactions (classic repos and pensions livrées) and sell/buybacks (e.g. simultaneas and PCT). There is a separate question (see question 2) on securities lending and borrowing transactions (including securities lending and borrowing against cash collateral).
f) Exclude repo transactions undertaken with central banks as part of their official money market operations. Other repo transactions with central banks, e.g. as part of their reserve management operations, should be included.
g) Give the value of the cash which is due to be repaid on all repo and reverse repo contracts (not the market value or nominal value of the collateral) that are still outstanding at close of business on Wednesday, June 11, 2025. This means the value of transactions at their repurchase prices.
h) Outstanding” means repos and reverse repos with a repurchase date, or which will roll over, on or after Thursday, June 12, 2025. You should include all open repos and reverse repos that have been rolled over from Wednesday, June 11, 2025, to a later date and all forward repos and reverse repos that are still outstanding as forward contracts at close on Wednesday, June 11, 2025.
i) Give separate totals for (a) repos plus sell/buy-backs and (b) reverse repos plus buy/sell-backs.
j) The survey seeks to measure the value of repos and reverse repos on a transaction date basis, rather than a purchase date basis. This means that you should include all repo and reverse repo contracts that have been agreed before close of business on Wednesday, June 11, 2025, even if their purchase dates are later.
An unavoidable consequence of using the transaction date is that tom/next and spot/next transactions that are rolled over will be counted more than once, eg a tom/next repo transacted on the day before the survey date and rolled over on the survey date will feature twice.
k) Give gross figures, i.e. do not net opposite transactions with the same counterparty. If this is not possible, please indicate that your figures are net.
l) Do not report synthetic repos.
m) You should include intra-group transactions between different legal entities or between foreign branches and the parent company.
Guidance on specific questions in the survey form
1.1 Transactions (1.1.1) direct with counterparties or (1.1.2) through voice-brokers should exclude all repos transacted over an ATS (see below). These should be recorded under (1.1.3). (1.1.1) should include repos executed on automated systems such as GLMX or TradeWeb (which offer a request-for-quote (RFQ) trading model). Repos executed on automated systems should also be included in (1.2.2), which measures all electronic trading.
(1.1.2) Transactions through voice-brokers should be broken down in terms of the location of the counterparties, rather than the location of the voice-brokers.
(1.1.3) “ATSs” are automatic or semi-automatic trading systems (e.g. BrokerTec, Eurex’s platforms, MTS, eRepo and SIX Repo) but not voice-assisted electronic systems used by voice-brokers (where voice-brokers record and communicate transactions agreed by telephone or electronic messaging) or automated systems such as GLMX or TradeWeb (which offer a request-for-quote (RFQ) trading model). Nor does use of an ATS include trading assisted by electronic means of structured messages and confirmations such as Bloomberg’s RRRA and similar screens.
Transactions on automated trading systems (RFQ systems) should be included in both (1.1.1) and (1.2.2). Transactions through voice-assisted systems should be included in (1.1.2).
Anonymous transactions through an ATS with a central counterparty (e.g. Euronext Clearing (formerly CC&G), LCH, BME Clearing (formerly MEFFClear) and Eurex Clearing) should be recorded in either (1.1.3.4) or (1.1.3.5). (1.1.3.4) is for GC financing systems. These are ATS that are connected to a CCP and a triparty repo service. Examples include Eurex’s GC Pooling (GCP), LCH SA’s €GCPlus and LCH Ltd’s £GC. They do not include GC basket trading on ATS, in which the seller manually selects the securities to be delivered from a list prescribed by the ATS. This activity may be cleared across a CCP, but does not involve a tri-party service and should be recorded in (1.1.3.5).
(1.2.1) This item includes all the transactions recorded in (1.1.3) plus any transactions executed directly with counterparties and via voice-brokers which are then registered with and cleared through a central counterparty. (1.2.2) Questions (1.1.3.1) to (1.1.3.5) measure repos and reverse repos transacted on automatic or semi-automatic trading systems such as BrokerTec, Eurex’s platforms, MTS, eRepo and Dealerweb, but not voice-assisted electronic systems used by voice-brokers (where voice-brokers record and communicate transactions agreed by telephone or electronic messaging) or automated systems such as BrokerTec Quote, GLMX, MTS BondVision or TradeWeb (which offer a request-for-quote (RFQ) trading model). Business on automated systems should be reported as direct trades in (1.1.1) and included in (1.2.2), which asks for the total value of business transacted on any electronic trading system, whether automatic, semi-automatic or automated, and therefore including automated systems such as GLMX or TradeWeb, which offer a request-for-quote (RFQ) trading model. Electronic trading Is defined in terms of where the contract is executed and so does not include voice-assisted electronic systems used by voice-brokers or trading assisted by electronic means of structured messages and confirmations such as Bloomberg’s RRRA and similar screens.
1.5 “Repurchase transactions” (also known as “classic repos”) include transactions documented under the Global Master Repurchase Agreement (GMRA) 1995, the Global Master Repurchase Agreement (GMRA) 2000 or the Global Master Repurchase Agreement (GMRA) 2011 without reference to the Buy/Sell-Back Annexes, and transactions documented under other master agreements. “Sell/buy-backs” are therefore taken to include all transactions that are not documented. Repurchase transactions are characterised by the immediate payment by the buyer to the seller of a compensatory or manufactured payment upon
receipt by the buyer of a coupon or other income on the collateral held by the buyer. If a coupon or other income is paid on collateral during the term of a sell/buy-back, the buyer does not make an immediate compensatory or manufactured payment to the seller, but reinvests the income until the repurchase date of the sell/buy-back and deducts the resulting amount (including reinvestment income) from the repurchase price that would otherwise be due to be received from the seller. Sell/buy-backs may be quoted in terms of a forward price rather than a repo rate. Where sell/buy-backs are documented (e.g. under the Buy/SellBack Annexes to the GMRA 1995, 2000 or 2011), periodic adjustments to the relative amounts of collateral or cash - which, for a repurchase transaction, would be performed by margin maintenance transfers or payments - are made by adjustment or re-pricing. All open repos are likely to be repurchase transactions.
1.6 “Open” repos, which are reported in (1.7.3), are defined for the purposes of this survey as contracts that have no fixed repurchase date when negotiated but are terminable on demand by either counterparty. Open repos should also be included in fixed-rate repo (1.6.1) unless their repo rates are linked to interest rate indexes which will be refixed during the life of the repos, in which cases, they would be reported as floating-rate repos (1.6.2).
1.7 This section asks for the remaining term to maturity (not the original term to maturity) of repos to be broken down as follows:
(1.7.1.1) 1 day – this means:
• all contracts transacted prior to Wednesday, June 11, 2025, with a repurchase date on Thursday, June 12, 2025;
• overnight, tom/next, spot/next and corporate/next contracts transacted on Wednesday, June 11, 2025.
(1.7.1.2) 2–7 days – this means:
• all contracts transacted prior to Wednesday, June 11, 2025, with a repurchase date on Friday, June 13, 2025, or any day thereafter up to and including Wednesday, June 18, 2025;
• contracts transacted on Wednesday, June 11, 2025, with an original repurchase date on Friday, June 13, 2025, or any day thereafter up to and including Wednesday, June 18, 2025 (irrespective of the purchase date, which will vary).
(1.7.1.3) More than 7 days but no more than 1 month – this means:
• all contracts transacted prior to Wednesday, June 11, 2025, with a repurchase date on Thursday, June 19, 2025, or any day thereafter up to and including Friday, July 11, 2025;
• contracts transacted on Wednesday, June 11, 2025, with an original repurchase date on Thursday, June 19, 2025, or any day thereafter up to and including Friday, July 11, 2025 (irrespective of the purchase date, which will vary).
(1.7.1.4) More than 1 month but no more than 3 months – this means:
• all contracts transacted prior to Wednesday, June 11, 2025, with a repurchase date on Monday, July 14, 2025, or any day thereafter up to and including Thursday, September 11, 2025;
• contracts transacted on Wednesday, June 11, 2025, with an original repurchase date on Monday, July 14, 2025, or any day thereafter up to and including Thursday, September 11, 2025 (irrespective of the purchase date, which will vary).
(1.7.1.5) More than 3 months but no more than 6 months – this means:
• all contracts transacted prior to Wednesday, June 11, 2025, with a repurchase date on Friday, September 12, 2025, or any day thereafter up to and including Thursday, December 11, 2025;
• contracts transacted on Wednesday, June 11, 2025, with an original repurchase date on Friday, September 12, 2025, or any day thereafter up to and including Thursday, December 11, 2025 (irrespective of the purchase date, which will vary).
(1.7.1.6) More than 6 months but no more than 12 months – this means;
• all contracts transacted prior to Wednesday, June 11, 2025, with a repurchase date on Friday, December 12, 2025, or any day thereafter up to and including Wednesday, June 10, 2026;;
• contracts transacted on Wednesday, June 11, 2025, with an original repurchase date on Friday, December 12, 2025, or any day thereafter up to and including Wednesday, June 10, 2026 (irrespective of the purchase date, which will vary).
(1.7.1.7) More than 12 months – this means;
• all contracts transacted prior to Wednesday, June 11, 2025, with a repurchase date on Thursday, June 11, 2026, or any day thereafter;
• contracts transacted on Wednesday, June 11, 2025, with an original repurchase date on or after Thursday, June 11, 2026 (irrespective of the purchase date, which will vary).
(1.7.2) Forward repos are now defined for this survey as contracts with a purchase date of Wednesday, June 18, 2025, or later, in other words, settling on T+5 or later. This definition has been amended to avoid an overlap with corporate/next transactions, which usually settle at T+3 or T+4.
(1.7.3) Open repos in this field should equal open repos in item (1.6.3).
1.8 Please confirm whether the transactions recorded in the questions in (1.6 and 1.7) include your tri-party repo business. Some institutions do not consolidate their tri-party repo transactions with their direct or voice-brokered business because of delays in receiving reports from tri-party agents or the complexity of their tri-party business.
(1.8.1) and (1.8.2) should not include any repos transacted across GC financing systems and recorded in (1.8.3).
1.9 “Eurobonds” (also known as “international bonds”) are defined as securities held outside national central securities depositories (CSD), usually in an ICSD such as Clearstream or Euroclear, or a custodian bank; typically with the ISIN prefix XS; often issued in a currency foreign to the place of issuance; and sold crossborder to investors outside the domestic market of the place of issuance. Eurobonds should be recorded in (1.9.30-33), except for those issues by “official international financial institutions”, which should be recorded in (1.9.20). Eurobond does not mean a bond denominated in euros.
(1.9.20) “Official international financial institutions, including multilateral development banks” such as:
African Development Bank (AfDB)
Asian Development Bank (AsDB)
Bank for International Settlements (BIS)
Caribbean Development Bank (CDB)
Central American Bank for Economic Integration (CABEI)
Corporacion Andina de Fomento (CAF)
Council of Europe Development Bank
East African Development Bank (EADB)
European Bank for Reconstruction and Development (EBRD)
Inter-American Development Bank Group (IADB)
International Fund for Agricultural Development (IFAD)
Islamic Development Bank (IDB)
Nordic Development Fund (NDF)
Nordic Investment Bank (NIB)
OPEC Fund for International Development (OPEC Fund)
West African Development Bank (BOAD)
World Bank Group (IBRD and IFC)
Securities issued by the EU (but not individual EU members) should now be included in the new question 1.9.37. EU issuers include:
European Commission
European Financial Stability Mechanism (EFSM)
European Financial Stability Facility (EFSF)
European Investment Bank (EIB)
European Stabilisation Mechanism (ESM)
European Union (EU)
(1.9.21) “US Treasury” includes bills, notes and bonds, including floating-rate notes, issued by the US central government but not securities guaranteed by that government, such as Agency securities.
(1.9.23) “Japanese government” includes bills, notes and bonds issued by the Japanese central government but not securities guaranteed by that government.
(1.9.25) “Other OECD countries” are Australia, Canada, Chile, Iceland, Israel, Korea, Mexico, New Zealand, Norway, Switzerland and Turkey.
(1.9.26) “Other non-OECD European, Middle Eastern & African countries” should exclude any EU countries.
(1.9.34) “Equity” includes ordinary shares, preference shares and equity-linked debt such as convertible bonds.
2.1 This question asks for the total gross value of transactions with a transaction date on or after Thursday, December 12, 2024 (the day after the previous survey date), to and including Wednesday, June 11, 2025 (the latest survey date). In other words, it asks for the turnover or flow of business over the six month interval and includes all business transacted since the last survey date, even if it has matured before the survey date. This section is therefore different from the rest of the survey, which asks for the value of business outstanding on the survey date, in other words, the stock of transactions.
2.2 This question asks for the number of individual transactions with a transaction date on or after Thursday, December 12, 2024 (the day after the previous survey date), to and including Wednesday, June 11, 2025 (the latest survey date), even if it has matured before the survey date. In other words, this is the number of tickets written.
3 This question asks for the cash value of any repos in which the survey participant is not a principal but provides a guarantee, indemnity or similar credit support. This support could be through a facility such as DTCC Sponsored Repo, LCH Sponsored Clearing or Eurex ISA Direct, or could be a bilateral arrangement.
4 “Total value of securities loaned and borrowed by your repo desk” includes the lending and borrowing of securities with either cash or securities collateral. Exclude any securities lending and borrowing done by desks other than your repo desk. If your repo desk does not do any securities lending and borrowing, this line will be a nil return.
5.1 “Active” means about once a week or more often.
For further help and information
If, having read the Guidance Notes, you have any further queries, please e-mail the independent survey administrator at reposurvey@icmagroup.org
Appendix B: Survey Participants
List of respondents
ABN Amro Bank
Allied Irish Banks
AXA Bank Europe
Banc Sabadell
Banca d'Intermediazione Mobiliare (IMI)
Banca Monte dei Paschi di Siena
Banco BPI
Banco Santander
UniCredit Bank Austria (Bank Austria)
Bank fuer Arbeit und Wirtschaft und Oesterreichische Postsparkasse (Bawag)
Bank of Ireland
Bank Przemyslowo-Handlowy SA
Landesbank Berlin
Banque de Luxembourg
Banque et Caisse d'Epargne de l'Etat
Barclays Capital
Bayerische Landesbank
BBVA
BHF-Bank
BHF-Bank International
BNP Paribas
Bundesrepublik Deutschland Finanzagentur
Caixabank (including Bankia)
Caixa d'Estalvis de Catalunya
Bankia SA (formerly Caja de Ahorros y Monte de Piedad de Madrid (Caja Madrid))
CA-CIB (formerly Calyon)
Citigroup
Commerzbank
Canadian Imperial Bank of Commerce and Credit (CIBC)
Commonwealth Bank of Australia
Confederación Española de Cajas de Ahorros (CECA)
Credit Suisse Securities (Europe) Ltd
Danske Bank
Daiwa Securities SMBC Europe
Dekabank Deutsche Girozentrale
Deutsche Bank
Deutsche Postbank
Belfius Bank (formerly Dexia)
Banque Internationale Luxembourg (formerly Dexia BIL)
Dexia Kommunal Bank Deutschland
DNB Bank ASA
DZ Bank
EFG Eurobank Ergasias
Erste Bank der Oesterreichischen Sparkassen
Euroclear Bank
European Investment Bank
Hypothekenbank Frankfurt International (formerly Eurohypo Europäische Hypothekenbank)
Fortis Bank
Goldman Sachs
HSBC
HSBC Athens
HSBC France
HSH Nordbank x
Unicredit Bank Germany (Bayerische Hypo-undVereinsbank)
ICBC Standard Bank
ING Bank
Intesa SanPaolo
Jefferies International
JP Morgan
KBC
KfW
Kingdom of Belgium Federal Public Service Debt Agency
Landesbank BadenWürttemberg, Stuttgart
Landesbank HessenThüringen -Girozentrale (Helaba)
Lloyds Bank Commercial Banking
Lloyds Bank Plc
Macquarie Bank
Bank of America Merrill Lynch
Mitsubishi Securities
International
Mizuho International
Morgan Stanley
National Australia Bank
National Bank of Greece
Newedge
Nomura International
Norddeutsche Landesbank Girozentrale
Nordea Markets
Norinchukin Bank
Nova Ljubljanska Banka d.d.
Nykredit Bank A/S
Piraeus Bank
Post Italiane
Rabobank
Royal Bank of Canada
NatWest Markets (formerly Royal Bank of Scotland)
Société Générale
Standard Chartered
Swedbank
Toronto Dominion Bank
UBS
UniCredit Bank AG Milano Branch
Unicredit Bank Spa
Westdeutsche Landesbank Girozentrale
Appendix C: Summary Of Survey Results
Q1 What are the total gross values of cash due to be repaid by you and repaid to you on repo transactions maturing after survey date? (figures in EUR billions)
Of the amounts given in response to question (1) above:
1.1 How much was transacted:
direct with counterparties
• in the same country as you
• cross-border in (other) eurozone countries
•
in non-eurozone countries
through voice-brokers
• in the same country as you
•
in (other) eurozone countries
• cross-border
on ATSs with counterparties
• in the same country as you
• cross-border in (other) eurozone countries
• cross border-border in non-eurozone countries
• anonymously across a GC financing system
• anonymously across a central clearing counterparty but not GC financing
• total through a central clearing counterparty
• transacted across any electronic system
1.2 How much of the cash is denominated in: • EUR
• GBP
• USD
• SEK, DKK
• JPY
• CHF
• other Asian and Pacific
other currencies
1.3 How much is cross-currency?
1.4 How much is:
•
1.5 How much is: •
• open
1.6 How much fixed and floating rate repo is (1.6.1) for value before (survey date) and has a remaining term to maturity of:
• 1 day
• 2 - 7days
• more than 7 days but no more than 1 month
• more than 1 month but no more than 3 months
• more than 3 months but no more than 6 months
more than 12 months
• forward-forward repos
open
1.7 How much is tri-party repo:
• for fixed terms to maturity
• total tri-party repo
1.8 How much is against collateral issued in:
Austria
• by the central government
• by other issuers
Belgium
Denmark •
Finland •
France
•
Germany
Greece
Ireland
Italy •
Luxembourg
Netherlands •
Portugal
•
Spain •
Sweden • by the central government
by other issuers
UK
other issuers
US but settled across EOC/CS
other countries
Bulgaria
• by the central government
• by other issuers
Cyprus
• by the central government
• by other issuers
Czech Republic
• by the central government
• by other issuers
Estonia
• by the central government
• by other issuers
Hungary
• by the central government
• by other issuers
Latvia
• by the central government
• by other issuers
Lithuania
• by the central government
• by other issuers
Malta
• by the central government
• by other issuers
Poland
• by the central government
• by other issuers
Romania
• by the central government
• by other issuers
Slovak Republic
• by the central government
• by other issuers
Slovenia
• by the central government
• by other issuers
Other EU members by central government
Other EU members by other issuers
• by official international financial institutions
Japan
• Japanese government
• Other Japanese issuers
Other Asian & Pacific OECD countries in the form of fixed income securities, except eurobonds
Other OECD countries in the form of fixed income securities, except eurobonds
issued by Asian & Pacific entities
issued by other entities
collateral of unknown origin or type
collateral in tri-party which cannot be attributed to a country or issuer
gross values of repo & reverse repo with APAC
Q2 What is the total value of securities loaned and borrowed by your repo desk: to/from counterparties in the same country as you
