International Capital Market Association European Repo Market Survey. Number 34 - Conducted December 2017 Published March 2018

Transcription

1 International Capital Market Association European Repo Market Survey Number 34 - Conducted December 2017 Published March

2 Disclaimer This report has been compiled by Richard Comotto, Senior Visiting Fellow, ICMA Centre at Reading University. International Capital Market Association (ICMA), Zurich, All rights reserved. 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. 2 European Repo Market Survey December 2017

3 Contents Executive Summary 4 Chapter 1: The Survey What the survey asked The next survey 6 Chapter 2: Analysis Of Survey Results 7 Total repo business (Q1) 7 Trading analysis (Q1.1) 9 Geographical analysis (Q1.1) 10 Clearing and settlement analysis (Q1.2 and Q1.8) 11 Cash currency analysis (Q1.3 and Q1.4) 12 Collateral analysis (Q1.9) 13 Contract analysis (Q1.5) 18 Repo rate analysis (Q1.6) 18 Maturity analysis (Q1.7) 19 Product analysis (Q2) 23 Concentration analysis 23 Chapter 3: Conclusion 25 About The Author 26 Appendix A: Survey Guidance Notes 27 For further help and information 30 Appendix B: Survey Participants 31 Appendix C: Summary Of Survey Results 35 Appendix D: The ICMA European Repo And Collateral Council 41 European Repo Market Survey December

4 Executive Summary In December 2017, the European Repo and Collateral Council (ERCC) of the International Capital Market Association (ICMA) conducted the 34 th in its series of semi-annual surveys of the repo market in Europe. The latest survey asked a sample of financial institutions in Europe for the value and breakdown of their repo contracts that were still outstanding at close of business on December 6, Replies were received from 64 offices of 60 financial groups, mainly banks. Returns were also made directly by the principal automatic repo trading systems (ATS) and tri-party repo agents in Europe. Total repo business The total value of the repo contracts outstanding on the books of the 64 institutions who participated in the latest survey was EUR 7,250 billion, compared with the EUR 6,455 billion in June This represents an increase in the headline number since the last survey of 12.3% and 28.2% year-on-year. Using a constant sample of banks, it is estimated that the market grew 12.2% since June and 19.4% year-on-year. But this growth was not broadly based across the survey sample. Trading analysis The uptrend in the share of directly-negotiated repo observed since 2012 continued between June and December 2017, while the share of electronic business transacted over automatic repo trading systems (ATS) continued to contract. Moreover, growth in the absolute size of the business reported directly by the principal ATS operating in Europe was muted, although this reflected very mixed fortunes among the ATS. The share of tri-party repo in the survey recovered but remained well below historic levels. Geographical analysis Cross-border business with counterparties in non-eurozone countries continued to gain share but the biggest gain was in domestic business, the trend decline in which may have bottomed out. These gains were largely reflected in a further fall in the share of anonymous (ie CCP-cleared) trading. The same pattern was seen in electronic trading but reversed in tri-party. Clearing and settlement analysis The value of outstanding repo business managed by the four tri-party agents who contributed directly to the survey fell. The share and absolute value of GC financing touched new lows. But the share of transactions negotiated directly or via voice-brokers that were subsequently registered with a CCP increased again. Cash currency analysis The biggest change in the currency composition of the survey was an exceptional increase in the share of Danish and Swedish currencies. Activity in the pound sterling and Japanese yen was buoyant. There was a jump in the share of the Swiss franc in the business reported directly by ATS, reflecting the participation for the first time of SIX Repo in the survey. Collateral analysis The share of government bonds within the pool of EU-originated fixed-income collateral reported in the survey fell back, driven mainly by retreats by French, German and Italian government securities and rapid growth in non-government securities issued in Denmark and Sweden. US Treasuries lost share. ATS reported significant increases in the shares of German, Spanish and UK government securities but a very sharp fall in Italian government securities. In tri-party repo, most eurozone government securities lost ground to US Treasuries, JGBs, eurobonds and other OECD securities. 4 European Repo Market Survey December 2017

5 Maturity analysis Overall, the market as represented by the survey sample, continues to be a net cash borrower in open repos and repos with one week or less remaining to maturity and a net lender in longer maturities. The share of short-dated positions dropped back in December, a typical end-year fluctuation. However, the weighted average term to maturity barely changed. The smaller share of short dates reflected falls in the share of repos with one day and eight days to one month remaining. In contrast, open repos increased share, as did repos with one to three months remaining and forward repos. The change in the share of repos with one to three months remaining is likely to have been driven by demand for high quality liquid assets (HQLA) to meet regulatory reporting requirements at the end of the year. The increase in the share of forwards could reflect efforts by the market to manage end-year positions in advance in order to avoid seasonal pressures. In contrast to the main survey, there was a distinct lengthening of maturities in electronic trading reported directly by ATS. European Repo Market Survey December

6 Chapter 1: The Survey On December 6, 2017, the European Repo and Collateral Council (ERCC) of the International Capital Market Association (ICMA) conducted the 34 th in its series of semi-annual surveys of the repo market in Europe. The survey was managed and the results analysed on behalf of ICMA by the author, who is based at the ICMA Centre at Reading University in England, under the guidance of the ERCC Steering Committee ( ERCC Committee ). 1.1 What the survey asked The survey asked financial institutions operating in a number of European financial centres for the value of the cash side of repo and reverse repo contracts still outstanding at close of business on Wednesday, December 6, The questionnaire also asked these institutions to analyse their business in terms of the currency, the type of counterparty, contract and repo rate, the remaining term to maturity, the method of settlement and the origin of the collateral. In addition, institutions were asked about securities lending and borrowing conducted on their repo desks. The detailed results of the survey are set out in Appendix C. An extract of the accompanying Guidance Notes is reproduced in Appendix A. Separate returns were made directly by the principal automatic repo trading systems (ATS) and tri-party repo agents in Europe. On this occasion, there was no return was made by the London-based European Venues and Intermediaries Association (EVIA), formerly the Wholesale Market Brokers Association (WMBA), because of an insufficient number of respondents. 1.2 The response to the survey The latest survey was completed by 64 offices of 60 financial groups. This is the same number as in the June 2017 survey. However, three institutions dropped out of the survey in December and three rejoined. 48 of the participants were headquartered across 16 European countries, including Norway (1) and Switzerland (2), as well as in Australia (1), Japan (4) and North America (9). 22 respondents were foreign affiliates, most of which (18) were located in the UK. 45 participants were based across 14 of the 28 member states of the EU (there were no institutions in the survey from Finland and Sweden, and only one from a former Accession State). 39 participants were based across 12 of the 19 countries of the eurozone. Many institutions provided data for their entire European repo business. Others provided separate returns for one or more (but not necessarily all) of their European offices. A list of the institutions 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 close of business on Wednesday, June 6, Any financial institution wishing to participate in the next survey will be able to download copies of the questionnaire and accompanying Guidance Notes from ICMA s web site. The latest forms will be published shortly before the next survey at the following website: Questions about the survey should be sent by to reposurvey@icmagroup.org. Institutions who participate in a survey receive, in confidence, a list of their rankings in the various categories of the survey. 6 European Repo Market Survey December 2017

7 Chapter 2: Analysis Of Survey Results The aggregate results of the latest two surveys and of the surveys in each December in the four previous years ( ) are set out in Appendix C. The full results of all previous surveys can be found at Total repo business (Q1) The total value, at close of business on December 6, 2017, of repos and reverse repos outstanding on the books of the 64 institutions which participated in the latest survey was a record EUR 7,249.6 billion, a rise in the headline number of 28.2% year-on-year and 12.3% since the June 2017 survey. Table 2.1 Total repo business from 2001 to 2017 survey total repo reverse repo 2017 December 7, % 52.2% 2017 June 6, % 51.5% 2016 December 5, % 51.9% 2016 June 5, % 52.0% 2015 December 5, % 52.5% 2015 June 5, % 52.0% 2014 December 5, % 51.2% 2014 June 5, % 51.4% 2013 December 5, % 50.8% 2013 June 6, % 50.2% 2012 December 5, % 51.9% 2012 June 5, % 51.3% 2011 December 6, % 49.7% 2011 June 6, % 49.3% 2010 December 5, % 49.0% 2010 June 6, % 46.5% 2009 December 5, % 50.0% 2009 June 4, % 47.8% 2008 December 4, % 50.1% 2008 June 6, % 51.2% 2007 December 6, % 50.6% 2007 June 6, % 49.2% 2006 December 6, % 49.3% 2006 June 6, % 48.3% 2005 December 5, % 45.4% 2005 June 5, % 47.6% 2004 December 5, % 49.9% 2004 June 4, % 49.4% 2003 December 3, % 48.7% 2003 June 4, % 50.0% 2002 December 3, % 49.0% 2002 June 3, % 50.0% 2001 December 2, % 49.6% 2001 June 1, % 50.4% European Repo Market Survey December

8 Figure 2.1 Total business 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1, It is important to remember that the survey measures the value of outstanding transactions at close of business on the survey date. Measuring the stock of transactions at one date, rather than the flow between two dates, permits deeper analysis but is difficult to reconcile with the flow numbers published by some other sources. As the survey is a snapshot of the market, it can miss peaks and troughs in business between survey dates, especially of very shortterm transactions. In the latest survey, respondents were asked to also report turnover since the last survey in June. Those who responded accounted for about 42% of total outstanding business. For this sub-sample, turnover over the intervening six months was about 20 times the outstanding total on December 6, The values measured by the survey are gross figures, which mean that they have not been adjusted for the double counting of the same transactions between pairs of survey participants. However, a study (see the report of the December 2012 survey) suggested that the problem of double-counting was not very significant. The survey does not measure the value of repos transacted with central banks as part of official monetary policy operations, which continue to be very substantial. In order to accurately gauge the growth of the European repo market (or at least that segment represented by the institutions who have participated in the survey), it is not valid to simply compare headline numbers. Some of the changes will represent the entry and exit of institutions into and out of the survey, mergers between banks and the reorganization of repo books within banks. The latest survey has been affected by the return of a major participant and the arrival of a significant new participant. To overcome the problem caused by changes in the sample of survey participants, comparisons are made of the aggregate outstanding contracts reported by a sub-sample of institutions which have participated continuously in several surveys. Out of the 64 institutions in the present survey, 56 have participated in all of the last three surveys. Overall, the aggregate value of outstanding repos and reverse repos transacted by that constant sample of 56 institutions grew by 19.4% year-on-year and by 12.2% between the December and June 2017 surveys. The change for the 60 institutions which participated in the last two surveys was +12.2%. The substantial growth in the survey occurred despite the fact that the repo books of 34 of the 64 institutions in the latest survey contracted or did not change, compared with 24 out of 64 repo books which contracted or did not change in the June 2017 survey. The narrower focus of growth is reflected in a jump in the unweighted mean percentage change to +32.6% from +10.8% but a sharp drop in the median percentage change to +0.2% from +5.0%. The position of the survey sample as a net lender to the rest of the market increased. 8 European Repo Market Survey December 2017

9 Figure 2.2 Total repo versus reverse repo business Repo Reverse 4,000 3,500 3,000 2,500 2,000 1,500 1, Trading analysis (Q1.1) Table 2.2 Trading analysis December 2017 June 2017 December 2016 share users share users share users direct 63.3% % % 65 of which tri-party 8.6% % % 44 voice-brokers 12.2% % % 48 ATS 24.5% % % 51 The uptrend in the share of directly-negotiated repo observed since 2012 continued between June and December 2017 and continues to be at the relative expense of electronic business transacted over automatic repo trading systems (ATS), the share of which fell back to 24.5% from 26.6%. Moreover, growth in the absolute size of the business reported directly by the principal ATS operating in Europe was muted, reaching EUR 1,123 billion from EUR 1,102 billion in June (+1.9%), despite the addition of another ATS to the sample (SIX Repo). However, the fall in share and the modest growth in absolute size reflected very mixed fortunes among the ATS. The share of tri-party repo in the survey recovered to 8.6% from 7.8% in June but remained well below historic levels (which averaged 9.9% between 2004 and 2016, and ranged from 7.9% to 12.0%). Tri-party repo continued to be a source of net funding for the rest of the repo market but less than in June (61.4% of tri-party business by survey respondents was repo, ie cash borrowing, compared with 67.4% in June). ATS (which is an interdealer market) continued to be used for net lending by the survey sample. Direct business (which encompasses many customers) switched to being a small net borrowing from the survey sample. Table 2.3 Numbers of participants reporting particular types of business ATS anonymous ATS voice-brokers tri-party repos total European Repo Market Survey December

10 Figure Trading analysis 24.5% ATS 54.7% Direct bilateral 12.2% Voice-brokered 8.6% Direct tri-party Geographical analysis (Q1.1) Table 2.4 Geographical analysis December 2017 June 2017 December 2016 share users share users share users domestic 25.9% 23.7% 23.2% cross-border to (other) eurozone cross-border to (other) non-eurozone 16.5% 16.9% 17.5% 40.5% 39.7% 38.6% anonymous 17.1% % % 42 Cross-border business with counterparties in non-eurozone countries continued to gain share, albeit more slowly, reaching 40.5% from 39.7%. But the biggest gain was in domestic business, which jumped to 25.9% from 23.7%, suggesting the trend decline in this segment may at least be bottoming out. These gains were largely reflected in a further fall in the share of anonymous (ie CCP-cleared) trading to 17.1% from 19.7% in June. Domestic business also increased and cross-border business with non-eurozone counterparties also decreased in the electronic market (as reported directly by the ATS). The pattern was reversed in tri-party business (as reported directly by the tri-party agents), who also saw a fall in the share of cross-border business with counterparties inside the eurozone. The share of anonymous trading reported directly by ATS fell back to 95.3% from an all-time high of 99.3% in June, but this was largely due to the introduction of SIX Repo into the survey (SIX is not connected to a CCP). Table 2.5 Geographical comparisons in December 2017 (June 2017) main survey ATS tri-party EVIA (WMBA) domestic 25.9% (23.7%) 32.6% (29.3%) 29.3% (30.9%) cross-border 57.0% (56.6%) 63.3% (68.7%) 70.7% (69.1%) anonymous 17.1% (19.7%) 10 European Repo Market Survey December 2017

11 Figure Geographical analysis 17.1% Anonymous ATS 25.9% Domestic 40.5% To non eurozone 16.5% To eurozone Clearing and settlement analysis (Q1.2 and Q1.8) The value of outstanding repo business managed by the four tri-party agents who contributed directly to the survey fell to EUR billion from EUR billion in June. Although one agent dropped out of the survey in December, this did not make a significant difference. The share of GC financing (mainly through Eurex Repo s Euro GC Pooling but also LCH s EuroGC Plus) touched a new low of 0.9% from 1.1% in June and, as a share of directly-reported electronic business, dropped further, to 3.5% from 5.6% in June. The absolute value of outstanding GC financing reported directly by ATS fell sharply to EUR 39.0 billion from EUR 61.5 billion. The share of transactions negotiated directly or via voice-brokers that were subsequently registered with a CCP increased again, to 9.3% from 8.3%. Figure 2.5 Evolution of business cleared across CCP 35.0% 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% ATS + post-trade registration ATS only (ex. GC financing) post-trade registration only 0.0% 6/07 12/07 6/08 12/08 06/09 12/09 06/10 12/10 6/11 12/11 06/12 12/12 6/13 12/13 6/14 12/14 6/15 12/15 6/16 12/16 6/17 12/17 European Repo Market Survey December

12 Cash currency analysis (Q1.3 and Q1.4) Table 2.6 Cash currency analysis December 2017 June 2017 December 2016 EUR 60.9% 61.9% 61.8% GBP 12.3% 12.2% 11.3% USD 14.7% 16.9% 18.8% DKK, SEK 5.9% 2.7% 2.1% JPY 4.5% 4.5% 4.8% CHF 0.0% 0.0% 0.1% other APAC 0.4% 0.7% 0.5% etc 1.2% 1.0% 0.7% cross-currency 1.4% 2.4% 1.8% The biggest change in the currency composition of the survey was the jump in the share of Danish and Swedish currencies to a record 5.9% from 2.7%. However, this growth was exceptional. Activity in the pound sterling and Japanese yen was buoyant. The jump in the share of the Swiss franc in the business reported directly by ATS (to 1.4% from zero) reflects the participation of SIX Repo in the survey for the first time. However, there was no increase in CHF repo reported by the survey sample, possibly reflecting the largely domestic nature of this market. Figure 2.6 Currency analysis 5.9% DKK, SEK 4.5% JPY 0.0% CHF 0.4% other APAC 1.2% other 14.7% USD 60.9% EUR 12.3% GDP 12 European Repo Market Survey December 2017

13 Table 2.7 Currency comparison in December 2017 main survey ATS tri-party EVIA (WMBA) EUR 60.9% 89.6% 49.8% GBP 12.3% 5.4% 5.9% USD 14.7% 0.0% 40.3% DKK, SEK 5.9% 0.3% 0.9% JPY 4.5% 0.0% 2.2% CHF 0.0% 4.6% 0.2% other APAC 0.4% etc 1.2% 0.1% 0.7% cross-currency 1.4% 11.9% Collateral analysis (Q1.9) Table 2.8 Collateral analysis December 2017 June 2017 December 2016 Germany 19.6% 20.0% 20.9% Italy 11.7% 12.0% 10.8% France 13.2% 13.7% 12.1% Belgium 3.0% 2.7% 3.0% Spain 5.5% 5.0% 5.2% other eurozone 4.2% 4.4% 5.0% UK 13.7% 13.5% 11.8% DKK, SEK 6.3% 2.7% 2.5% international financial institutions 1.0% 1.2% 1.9% US Treasuries 4.6% 6.5% 7.9% other US 1.2% 2.2% 2.2% former Accession 0.4% 0.4% 0.3% Japan government 3.3% 3.4% 2.8% other Japan 1.1% 1.0% 1.5% other OECD ex APAC 3.7% 4.3% 4.1% other APAC OECD 0.8% 0.4% 0.5% eurobonds 1.8% 1.8% 2.6% other fixed income 3.4% 4.6% 6.0% equity 0.2% 0.2% 0.1% European Repo Market Survey December

14 Figure Collateral analysis (main survey) 4.4% Japan 0.4% Accession countries (outside eurozone) 4.5% other OECD US 5.9% 6.3% DKK, SEK 1.0% IFI 6.7% others 3.0% Belgium 13.2% France 19.6% Germany 13.7% UK 4.2% other eurozone 5.5% Spain 11.7% Italy The share of government bonds within the pool of EU-originated fixed-income collateral reported in the survey fell back to 85.7% from the 13-year high of 87.6% reached in June. This change was driven mainly by retreats in French, German and Italian government securities and rapid growth in non-government securities issued in Denmark and Sweden (2.9% from 1.4%). On the other hand, Swedish and UK government securities increased share (respectively, to 2.9% from 0.9% and to 12.1% from 11.8%). US Treasuries lost share (4.6% from 6.5%). ATS reported significant increases in the shares of German, Spanish and UK government securities (26.9% from 23.6%, 7.8% from 6.7% and 6.3% from 4.8%, respectively) but a very sharp fall in Italian government securities (32.6% from 39.5%), which was reflected in the fall in the share of ATS overall. In tri-party repo, as reported directly by tri-party agents, most eurozone securities lost ground to US Treasuries (2.7% from 0.8%), JGBs (3.9% from 0.1%), European Eurobonds (12.8% from 7.0%), other Eurobonds (9.3% from 6.0%) and other OECD securities (5.5% from 1.4%). The share of government securities in directly-reported tri-party repo retreated to 46.1% from a record 47.7% in June. Table 2.9 Tri-party repo collateral analysed by credit rating December 2017 June 2017 December 2016 AAA 23.2% 24.0% 22.2% AA 27.2% 27.3% 25.5% A 10.2% 9.6% 9.0% BBB 13.5% 12.5% 14.0% below BBB- 6.0% 7.1% 3.6% A1/P1 2.9% 2.0% 1.4% A2/P2 0.5% 0.7% 0.4% Non-Prime 0.0% 0.0% 0.0% unrated 16.4% 16.8% 23.8% 14 European Repo Market Survey December 2017

15 Figure Collateral analysis (tri-party agents) by change in credit rating % % 10.0% 0.0% -10.0% -20.0% -30.0% AAA AA A BBB subbbb A1/P1 A2/P2 NP unrated Table 2.10 Tri-party repo collateral analysed by type of asset December 2017 June 2017 December 2016 government securities 46.1% 47.7% 42.0% public agencies / sub-national governments 11.0% 11.1% 14.8% supranational agencies 2.8% 3.6% 0.9% corporate bonds 17.6% 15.6% 16.3% covered bonds 6.4% 5.9% 8.5% residential mortgagebacked commercial mortgagebacked 1.3% 1.0% 1.1% 0.1% 0.1% 0.0% other asset-backed 1.2% 0.9% 0.7% CDO, CLN, CLO, etc 0.7% 0.4% 0.4% convertible bonds 1.1% 1.1% 0.9% equity 11.1% 11.7% 14.1% other 0.5% 0.8% 0.2% European Repo Market Survey December

16 Figure 2.9 Historic collateral analysis (tri-party agents) by credit rating 60.0% 50.0% AAA AA 40.0% 30.0% 20.0% A BBB subbbb A1/P1 A2/P2 NP 10.0% unrated 0.0% Figure Collateral analysis (tri-party agents) by type of asset 11.1% Equity 5.0% other 6.4% Covered 17.6% Corporate 46.1% Government 2.8% Supranational 11.0% Public & sub-national The gradual recovery in the use of corporate bonds in tri-party repo continues, but usage of covered bonds is still subdued. The share of government bonds remains at historic highs. 16 European Repo Market Survey December 2017

17 Figure 2.11 Historic collateral analysis (tri-party agents) by type of asset 60.0% Government 50.0% Public & sub 40.0% Supranational 30.0% Corporate Covered 20.0% M/ABS 10.0% 0.0% Equity Other Table 2.11 Tri-party repo collateral haircuts analysed by type of asset (weighted average haircuts) December 2017 June 2017 December 2016 government securities 2.2% 2.4% 2.5% public agencies / sub-national governments 2.3% 2.9% 2.5% supranational agencies 2.9% 2.7% 2.5% corporate bonds (financial) corporate bonds (non-financial) 6.0% 6.4% 7.9% covered bonds 3.3% 3.9% 4.6% residential mortgage-backed commercial mortgage-backed 4.5% 5.7% 6.9% other asset-backed 3.9% 5.0% 4.9% CDO, CLN, CLO, etc 3.3% 5.4% 3.2% convertible bonds 10.8% 8.9% 11.1% equity 6.8% 7.1% 6.9% other 5.1% 5.1% 3.7% European Repo Market Survey December

18 Contract analysis (Q1.5) Figure Contract analysis 13.8% Documented sell/buy-back 0.2% Undocumented sell/buy-back 86.0% Repurchase agreements Table 2.12 Contract comparison in December 2017 (June 2017) main survey ATS tri-party repurchase agreements 86.0% (85.7%) 64.0% (57.9%) 100.0% (100.0%) documented sell/buybacks undocumented sell/buybacks 13.8% (12.9%) 36.0% (42.1%) 0.2% (1.4%) There was little evidence of the switch by MTS from sell/buy-backs to repurchase transactions in July 2017, perhaps because the survey does not include many domestically-orientated Italian banks (cross-border business in Italy was already in repurchase transactions). Repo rate analysis (Q1.6) There was a small increase in the share of floating-rate repos. Figure Repo rate analysis 13.2% Floating rate 6.2% Open 80.6% Fixed rate 18 European Repo Market Survey December 2017

19 Table 2.13 Repo rate comparison in June 2017 (December 2016) main survey ATS tri-party fixed rate 80.6% (80.9%) 85.3% (85.2%) 21.9% (27.5%) floating rate 13.2% (12.7%) 14.7% (14.8%) 7.8% (2.1%) open 6.2% (6.4%) 70.3% (70.4%) Maturity analysis (Q1.7) Table 2.14 Maturity analysis December 2017 June 2017 December 2016 open 7.8% 6.6% 6.4% 1 day 16.9% 18.4% 18.9% 2 days to 1 week 22.6% 22.3% 21.2% 1 week to 1 month 15.9% 20.0% 16.1% >1 month to 3 months 16.3% 12.6% 18.9% >3 months to 6 months 4.4% 4.9% 4.3% >6 months to 12 months 2.5% 3.2% 2.5% >12 months 1.5% 1.5% 1.3% forward-start 12.2% 10.5% 10.5% Figure 2.14 Maturity analysis (main survey) 25.0% 22.6% 20.0% 15.0% 16.9% 15.9% 16.3% 12.2% 10.0% 5.0% 7.8% 4.4% 2.5% 1.5% 0.0% open 1D 1D-1W 1D-1M 1-3M 3-6M 6-12M 12M+ fd-fd European Repo Market Survey December

20 Figure 2.15 Maturity analysis: short dates, longer terms & forwards (main survey) 80.0% 70.0% 60.0% SD + open 50.0% 40.0% 30.0% 20.0% 10.0% 1M+ forward 0.0% Figure 2.16 Maturity analysis: non-forward terms (main survey) 80.0% 70.0% 60.0% SD + open 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% 1-6M 6M European Repo Market Survey December 2017

21 Figure 2.17 Maturity analysis: breakdown of short dates plus open (main survey) 35.0% 30.0% 25.0% 1D 2D-1W 1W-1M open 20.0% 15.0% 10.0% 5.0% 0.0% Overall, the market as represented by the survey sample continues to be a net cash borrower in open repos and repos out to one week and a net lender in longer maturities. The trend in the shares of open repos and short dates (one month or less remaining to maturity) as well as repos with one to six months remained broadly stable (at about 65% and 20%, respectively) but highly seasonal between surveys. In December 2017, the share of short-dated positions dropped back to 55.4% from 60.7% in June in a typical end-year fluctuation (the trend in open repos and repos out to six months has been stable since 2014 at about 65% and 20%, respectively). Despite the relative contraction in short dates, the weighted average term to maturity barely changed, spanning days compared to days in June and days in December 2016 (the lower end of the range assumes that all transactions have the minimum term in each maturity band: the upper end assumes the maximum term). The smaller share of short dates reflected falls in the share of repos with remaining maturities of one day (16.9% from 18.4%) and eight days to one month (15.9% from 20.0%). Repos with three months to one year also lost share (6.9% from 8.1%). In contrast, open repos increased share (7.8% from 6.6%) as did repos with one to three months remaining (16.3% from 12.6%) and forward repos (12.2% from 10.5%). The change in the share of repos with one to three months remaining is likely to have been driven by demand for high quality liquid assets (HQLA) to meet regulatory reporting requirements at the end of the year. The increase in the share of forwards could reflect efforts by the market to manage end-year positions in advance in order to avoid the seasonal pressures. In contrast to the main survey, there was a distinct lengthening of remaining maturities in electronic trading reported directly by ATS, with short dates falling to 93.8% from 97.8% in June, largely driven by a fall in remaining maturities of one day to 83.9% from 95.1%. Within the short dates, repos with two days to one month remaining increased share to 13.1% from 2.7%. The share of repos with one to three months remaining increased to 1.8% from 0.6%. In tri-party repo, as reported directly by tri-party agents the maturity distribution showed little change. Figure 2.18 Maturity analysis (ATS) 90.0% 80.0% 83.9% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% 0.0% 9.3% 3.8% 1.8% 0.6% 0.2% 0.3% open 1D 2-7D 8D-1M 1-3M 3-6M >6M forward European Repo Market Survey December

22 Figure 2.19 Maturity analysis (tri-party agents) 60.0% 50.0% 48.8% 40.0% 30.0% 20.0% 10.0% 7.7% 7.8% 9.2% 10.4% 7.8% 4.4% 3.8% 0.0% open 1D 1D-1W 1W-1M 1-3M 3-6M 6-12M >12M Table Maturity comparison in December 2017 (June 2017) main survey ATS tri-party EVIA (WMBA) open 7.8% (6.6%) 48.8% (48.8%) 1 day 16.9% (18.4%) 83.9% (95.1%) 7.7% (8.1%) 2 days to 1 week 22.6% (22,3%) 9.3% (0.8%) 7.8% (7.9%) 1 week to 1 month 15.9% (20.0%) 3.8% (1.9%) 9.2% (9.0%) >1 month to 3 months 16.3% (12.6%) 1.8% (0.6%) 10.4% (10.4%) >3 months to 6 months 4.4% (4.9%) 0.6% (1.0%) 7.8% (7.7%) >6 months to 12 months 2.5% (3.2%) 0.2% (0.3%) 4.4% (4.9%) >12 months 1.5% (1.5%) 0.3% (0.0%) 3.8% (3.3%) forward-start 12.2% (10.5%) 0.0% (0.3%) 22 European Repo Market Survey December 2017

23 Product analysis (Q2) The share of securities lending conducted on repo desks was unchanged at 10.9%. Figure 2.20 Product analysis 10.9% Securities lending 89.1% Repo Concentration analysis Table 2.16 Concentration analysis December 2017 June 2017 December 2016 top % 62.1% 64.6% top % 85.3% 85.8% top % 93.7% 94.0% other 6.2% 6.3% 6.0% Figure Concentration analysis 9.1% Top % Remainder 23.9% Top % Top 10 The concentration of business in the top 20 of the survey sample was reduced, despite the narrow base of the growth in business. European Repo Market Survey December

24 Table 2.17 Herfindahl Index 1 index numbers in survey December June December June December June December June December June December June December June December June December June December June December June December June December June December June December The Herfindahl Index is the sum of the squares of market shares divided by the square of the sum of market shares. The higher the index, the lower the degree of competition. If the index is higher, the more a single institution has a dominant market share and/or the more insignificant the market shares of all the other survey participants. A market in which several institutions have very large market shares can therefore have a relatively low index. 24 European Repo Market Survey December 2017

25 Chapter 3: Conclusion The recovery of the European repo market from the dislocation at the end of 2016 appears to have accelerated over the remainder of This was notwithstanding the pressure placed on the market by the implementation of MiFID II and MiFIR (which significantly delayed many responses to this survey). However, the strong growth in the second half of last year was more narrowly based than in the first half. Some but not all banks appear to have adapted to the impact of easy monetary policy and heavy regulation, and their activity may be catching up with their balance sheet capacity. It now remains to be seen if the recovery of activity will become more general. The pattern of business in the European repo market was largely unchanged, suggesting the focus remains on customers and HQLA in preference to low-margin business with other intermediaries. Thus, the survey sample shows that the market continued to be a net lender of cash and borrower of collateral, largely through direct business crossborder between parties inside and outside the eurozone. Electronic trading (an interdealer market segment) lost more share and tri-party repo (a pure funding market) remains lacklustre in the face of abundant central bank liquidity. There continues to be considerable post-trade registration of repos with CCPs (rather than clearing through ATS). The overall lengthening of maturities reflected the approach of the year-end. On the other hand, some changes have less obvious drivers. This includes a noticeable recovery in domestic business, longer maturities in electronic trading and a contraction (despite the demand for HQLA) in the share of government securities in EU-originated collateral. European Repo Market Survey December

26 About The Author This report was compiled by Richard Comotto, who is a Senior Visiting Fellow at the ICMA Centre at the University of Reading in England. He is also Course Director of the ICMA Professional Repo Market Course and of the ICMA-ISLA GMRA-GMSLA Workshop. He is author of the ICMA s Guide to Best Practice in the European Repo Market and its Repo FAQs. The author acts as an independent consultant providing research, advice and training on the international money, securities and derivatives markets to professional market associations, government agencies, regulatory authorities, international financial institutions, banks, brokers and financial information services. This includes advising technical assistance missions by the IMF, World Bank, other multilateral development banks and other organisations to build repo markets in emerging economies. The author has written a number of books and articles on a range of financial topics, including the foreign exchange and money markets, swaps and electronic trading systems. He takes particular interest in the impact of electronic trading systems on the bond and repo markets. Since the financial crisis, he has been advising the ICMA s European Repo and Collateral Council on regulatory initiatives and has produced a series of papers: in July 2010, a White paper on the operation of the European repo market, the role of short-selling, the problem of settlement failures and the need for reform of the market infrastructure ; in September 2011, Interconnectivity of central and commercial bank money in the clearing and settlement of the European repo market ; in February 2012, Haircuts and Initial Margins in the Repo Market ; in March 2012, Shadow Banking and Repo ; and in Collateral damage: the impact of the Financial Transaction Tax on the European repo market in April He writes on repo market topics on the ICMA Centre blog. The author served for ten years at the Bank of England, within its Foreign Exchange Division and on secondment to the International Monetary Fund in Washington DC. 26 European Repo Market Survey December 2017

27 Appendix A: Survey Guidance Notes The following extract is based on the Guidance Notes issued to participants in conjunction with the survey that took place on Wednesday, December 6, 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, December 6, 2017, 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 June 7, 2017). 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. Please feel free to copy the survey form to other banks, if you discover that they have not received it directly. 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. 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/buy-backs (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, December 6, 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, December 7, You should include all open repos and reverse repos that have been rolled over from Wednesday, December 6, 2017, to a later date and all forward-forward repos and reverse repos that are still outstanding as forward contracts at close on Wednesday, December 6, 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, December 6, 2017, 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) In the case of equity repo, for synthetic structures, please give the value of the cash payment. m) You should include intra-group transactions between different legal entities or between foreign branches and the parent company. European Repo Market Survey December

28 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). (1.1.2) 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 trading systems (e.g. BrokerTec, Eurex Repo and MTS, but not voice-assisted electronic systems used by voice-brokers or automated systems such as Bloomberg). Transactions through voice-assisted systems should be included in (1.1.2). Anonymous transactions through an ATS with a central counterparty (e.g. CC&G, LCH, MEFF and Eurex Clearing) should be recorded in either ( ) or ( ). ( ) is for GC financing systems. These are ATS that are connected to a CCP and a tri-party repo service. Examples include Eurex Euro GC Pooling and LCH s GC Plus. They do not include GC basket trading on ATS in which the seller 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 ( ). 1.2 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. 1.5 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/Sell-Back 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 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. This item should be equal to item (1.8.3). Open repos could be regarded as floating-rate, given that rates may be updated, but this tends to be irregular, so open repos are being treated separately from floating-rate repo (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 day this means: all contracts transacted prior to Wednesday, December 6, 2017, with a repurchase date on Thursday, December 7, 2017; overnight, tom/next, spot/next and corporate/next contracts transacted on Wednesday, December 6, ( ) 2 7 days this means: all contracts transacted prior to Wednesday, December 6, 2017, with a repurchase date on Friday, December 8, 2017, or any day thereafter up to and including Wednesday, December 13, 2017; contracts transacted on Wednesday, December 6, 2017, with an original repurchase date on Friday, December 8, 2017, or any day thereafter up to and including Wednesday, December 13, 2017 (irrespective of the purchase date, which will vary). ( ) More than 7 days but no more than 1 month this means: all contracts transacted prior to Wednesday, December 6, 2017, with a repurchase date on Thursday, December 14, 2017, or any day thereafter up to and including Monday, January 8, 2018; contracts transacted on Wednesday, December 6, 2017, with an original repurchase date on Thursday, December 14, 2017, or any day thereafter up to and including Monday, January 8, 2018 (irrespective of the purchase date, which will vary). 28 European Repo Market Survey December 2017

29 ( ) More than 1 month but no more than 3 months this means: all contracts transacted prior to Wednesday, December 6, 2017, with a repurchase date on Tuesday, January 9, 2018, or any day thereafter up to and including Tuesday, March 6, 2018; contracts transacted on Wednesday, December 6, 2017, with an original repurchase date on Tuesday, January 9, 2018, or any day thereafter up to and including Tuesday, March 6, 2018 (irrespective of the purchase date, which will vary). ( ) More than 3 months but no more than 6 months this means: all contracts transacted prior to Wednesday, December 6, 2017, with a repurchase date on Wednesday, March 7, 2018, or any day thereafter up to and including Wednesday, June 6, 2018; contracts transacted on Wednesday, December 6, 2017, with an original repurchase date on Wednesday, March 7, 2018, or any day thereafter up to and including Wednesday, June 6, 2018 (irrespective of the purchase date, which will vary). ( ) More than 6 months but no more than 12 months this means; all contracts transacted prior to Wednesday, December 6, 2017, with a repurchase date on Thursday, June 7, 2018, or any day thereafter up to and including Thursday, December 6, 2018; contracts transacted on Wednesday, December 6, 2017, with an original repurchase date on Thursday, June 7, 2018, or any day thereafter up to and including Thursday, December 6, 2018 (irrespective of the purchase date, which will vary). ( ) More than 12 months this means; all contracts transacted prior to Wednesday, December 6, 2017, with a repurchase date on Friday, December 7, 2018, or any day thereafter; contracts transacted on Wednesday, December 6, 2017, with an original repurchase date on or after Friday, December 7, 2018 (irrespective of the purchase date, which will vary). (1.7.2) For repos against collateral that includes a transferable security regulated under the EU MiFID and that have been traded or which it is possible to trade on a MiFIR-regulated trading venue (regulated market, multilateral trading facility or organised trading facility), forward-forward repos are defined for the purposes of this survey as contracts with a purchase date of Monday, December 11, 2017, or later. There is therefore an overlap with corporate/next transactions. If the latter cannot be identified separately, it is accepted that they will be recorded as forward-forward repos. It does not matter than many repos may actually be traded for T+1 (ie a purchase date of Thursday, December 7, 2017). For repos transacted in the OTC market or against collateral not regulated under MiFID, the definition of forward-forward may be different. (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 voicebrokered business because of delays in receiving reports from tri-party agents or the complexity of their tri-party business. 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 ( ), 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. 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) East African Development Bank (EADB) European Bank for Reconstruction and Development (EBRD) European Commission (EC)/European Financial Stability Mechanism (EFSM) European Financial Stability Facility (EFSF) European Repo Market Survey December