Enhancing Repo Market Transparency: The EU Securities Financing Transactions Regulation

Abstract

The introduction of the Securities Financing Transactions Regulation into EU law provides a unique opportunity to obtain an in-depth understanding of repo markets. Based on the transaction-level data reported under the regulation, this article contributes to the literature with key facts about the euro area repo market. We start by providing the regulatory background, as well as highlighting some of its advantages for financial stability analysis. We then go on to present three sets of findings that are highly relevant to financial stability and focus on the dimensions of the different market segments, counterparties, and collateral, including haircut practices. Finally, we outline how the data reported under the regulation can support the policy work of central banks and supervisory authorities and contribute to the existing literature on repo markets. We demonstrate that these data can be used to make several important contributions to enhancing our understanding of the repo market from a financial stability perspective, ultimately assisting international efforts to increase repo market resilience.

I. INTRODUCTION

Repo markets play a key role in the financial system by providing a range of economic functions. These include the provision of funding, investment of cash, and collateral transformation, as well as facilitating hedging and liquidity in underlying markets. The ability to raise cash against liquid assets during times of stress can be particularly crucial to prevent forced selling in core bond markets, underscoring the importance of repo market functioning at such times. However, they can also pose significant risks to the financial system. The global financial crisis highlighted major vulnerabilities in parts of the repo market related to issues around transparency, excessive leverage, over-reliance on short-term funding, collateral quality, and haircut procyclicality.¹ Even though the repo market is a secured market, dynamics have been heterogenous across jurisdictions and different segments.² Parts of it can be subject to market freezes, which can have severe consequences for financial stability. Vulnerabilities emanating from the repo market can also spill over to other markets, impacting funding and market liquidity more broadly.³

Significant regulatory efforts have been made to enhance the resilience of the global financial system in both the traditional banking industry and the so-called ‘shadow-banking’ sector.⁴ A clear view emerged that enhanced data collection on securities financing transactions was needed for authorities to obtain more timely and comprehensive insights into these markets. In 2013 the Financial Stability Board (FSB) issued a set of recommendations aimed at addressing risks in relation to Securities Financing Transactions (SFTs).⁵ Some of these measures required national and regional authorities to enhance their data collection in these markets, in order to improve their ability to monitor key financial stability risks and develop robust policy responses. In 2016 the Securities Financing Transactions Regulation (SFTR) was introduced in the EU, requiring all entities based in the EU to report their SFTs. The resulting data collection led to the development of the ECB Securities and Financing Transaction Data Store (SFTDS), a subset of the SFTR, which covers all SFT transactions conducted in the euro area, transactions conducted in euro or involving euro-denominated collateral in the EU, and also transactions of supervised euro area entities conducted outside the EU. While the SFTDS includes all types of SFTs, we focus on the repo market, which is the largest segment.

In this article we aim to show that the SFTDS can greatly enhance our understanding of repo markets from a financial stability perspective, supporting efforts to increase its resilience. We start by showing how the data reported under the SFTR greatly increases repo market transparency and will help authorities to obtain more timely and comprehensive insights into risks. By providing a comprehensive picture of daily activity for a broad range of market segments and counterparties, it will support market monitoring and the development of policy responses on key financial stability issues. We then highlight the potential of the SFTDS by applying our analyses to three broad key dimensions: (i) market segments (for example, jurisdiction, currency, central clearing); (ii) counterparties; (iii) collateral (for example, asset composition and haircuts). Finally, we highlight areas in which the dataset can support policy work by central banks and supervisory authorities.

Our first set of results examines the role of market segments; we look at jurisdiction, currency, the presence of central clearing, and the distinction between general and specific collateral.⁶ We show that there is an important cross-border and cross-currency dimension to the dataset, which involves a large volume of trades by the foreign branches and subsidiaries of euro area supervised entities as well as trades conducted in the euro area by counterparties that have a non-euro area parent.⁷ This indicates that there is significant scope for international transmission of shocks originating both inside and outside the euro area. Further analysis on how this contributes to the interconnectedness of the global financial system will be useful.

We then focus on euro-denominated trades and distinguish between whether a trade is centrally cleared or non-centrally cleared, and whether or not it is backed by general or specific collateral. We restrict our examination to euro-denominated trades as this is the market captured most comprehensively by the dataset. It is therefore the most representative market, and helps us to focus our analysis. From a financial stability perspective, the use of central clearing is relevant in at least two important ways. First, it can facilitate the intermediation capacity of dealer banks by allowing exposures to be netted, which reduces the capital allocated to market-making activities. Second, it can reduce the credit risk exposure for market participants by having a central clearing counterparty (CCP). This provides for more efficient offsetting of losses and mutualization in the event of a default.⁸ A feature of our analysis is that when we focus on outstanding amounts, we find that the share of central clearing is significantly lower than typically reported for daily transactions.⁹ Finally, we look at the role of general or specific collateral and find that the euro area repo market is primarily driven by trades that involve specific collateral. A significant share of this is potentially due to the demand for specific types of collateral known as specialness. This may indicate that the demand for collateral rather than funding is the main driver of the euro-denominated repo market.¹⁰

Our second set of results examines the counterparty dimension for euro-denominated trades. This is crucial not just for evaluating risks but also for understanding what economic functions repo markets perform. A stylized way of classifying how counterparties use repo markets from a supply and demand perspective is: (i) funding demand (repos); (ii) cash investment (reverse repos); (iii) collateral demand (reverse repos); (iv) collateral investment (repos); and (v) market-making (repos and reverse repos). For centrally cleared trades we find that in terms of gross positions foreign banks are the largest sector, followed by euro area commercial banks and then euro area investment banks.¹¹ Foreign banks (defined as banks where the parent’s headquarters are located outside the euro area) are the largest cash lenders (collateral borrowers) in net terms, while euro area commercial banks are the largest net cash borrowers (collateral lenders). For non-centrally cleared trades, we find that euro area investment banks are the largest sector in terms of gross position, followed by euro area commercial banks and then foreign banks. In terms of net positions, euro area investment banks tend to be cash lenders, while commercial banks tend to be cash borrowers. We also see that foreign banks’ activity is much less than for centrally cleared trades.

We find that non-banks make up a sizable share of repos in non-centrally cleared transactions. Investment funds (IFs) have the largest gross positions, while the largest sector in net terms is insurance corporations and pension funds (ICPFs) as a net borrower. This may suggest that IFs generally use the repo market for collateral transformation (ie they borrow cash against lower quality collateral and reinvest to acquire higher quality collateral). ICPFs on the other hand may potentially use repos to access cash buffers (for example, to meet cash margin calls) or simply to generate a return by lending collateral that commands a specialness premium.¹²

Our third set of results examines the role of collateral in the non-centrally cleared segment with regards to asset composition and haircuts. We deliberately focus on the non-centrally cleared segment, as this is where financial stability risks are expected to be most pronounced, for example due to differences in counterparty and collateral composition and risk management practices. To further tease out the financial stability risks, we differentiate between banks and non-banks. In terms of asset composition, we find that while the large majority of trades are backed by sovereign collateral, trades backed by non-sovereign collateral can play a sizeable role. We also find that entities in the non-bank sector tend to pledge less liquid assets as collateral in their borrowing operations relative to the collateral they receive against cash, while the reverse is the case for banks. However, aggregate figures may mask important heterogeneity at entity or sector level, given the different uses of the repo market.

We then look at the role of haircuts, focusing on the distribution from the cash borrower perspective. Haircut practices serve as a main risk mitigation tool for non-centrally cleared trades backed by non-sovereign collateral. We find that the share of positive haircuts for non-government collateral is typically quite low, irrespective of whether the cash-borrower is a bank or a non-bank. This further underscores the need to make progress on implementing regulatory rules aimed at setting minimum haircut standards for the non-bank sector in non-centrally cleared markets.¹³

In a final step, we comprehensively highlight how SFTR data reporting can support the policy work of central banks and supervisory authorities. Based on our analysis, we believe that the dataset can make important contributions not only to financial stability but also more broadly to other areas of importance for central banks. First, from a financial stability risk monitoring and supervisory perspective, the SFTDS greatly increases transparency and supports authorities in monitoring important risks related to interconnectedness, concentration, liquidity provision, collateral demand, liquidity/maturity mismatches, leverage, and various collateral metrics. Second, the dataset will play an important role in current international policy initiatives to enhance global financial stability resilience. More specifically, it could be used to inform the development of new policy recommendations aimed at non-bank entities and activities by providing an ex ante impact assessment for new policies, for example those targeting repo markets directly or non-bank entities obtaining leverage via repos. Third, the dataset can help central banks conduct a more broad-based evaluation of how the repo market functions for monetary policy purposes, which may be of particular relevance in the current period of monetary policy normalization, when intermediation capacity may become more constrained. Finally, with its granular information on counterparties’ activity in repo markets and by covering all parts of the euro area repo market, SFTR data can support policy discussions on the impact of post-crisis regulatory reforms on repo markets, complementing existing work in this area.¹⁴

The article proceeds as follows. Section II provides key details on the institutional background of the SFTDS and a description of the data, including some of its key comparative advantages over other data sources. Section III presents high-level findings for three key dimensions deemed crucial from a financial stability perspective. Section IV presents possible implications for policy discussions. Section V concludes.

II. DESCRIPTION OF THE SFTDS

1. Institutional background

In response to the global financial crisis the FSB undertook several initiatives to support global financial markets, including addressing financial stability risks associated with SFTs. In August 2013, it published a comprehensive report setting out policy recommendations.¹⁵ Importantly, the FSB recommended, among other things, that its member institutions collect granular transaction-level data to increase the transparency of SFT markets. It also initiated a global SFT data reporting initiative which will aggregate SFT market data to be reported by member jurisdictions on a global scale.¹⁶

As a follow-up to the FSB recommendations, the EU issued the SFTR which entered into force in 2016. At EU level, the work was also under way regarding the assessment, identification, and monitoring of entities and risks posed by ‘shadow banking’.¹⁷ A key contribution included proposals for potential improvements to the monitoring and supervision of SFT markets.¹⁸ The regulation required all EU SFT market participants (excluding central banks and certain non-financial corporations) to report their SFT trades daily from mid-June 2020 onwards (after a phase-in period). The information comprises comprehensive details on amounts, prices/rates, and collateral, and is reported to dedicated trade repositories which make the data available to relevant authorities via the TRACE portal, a data hub operated by the European Securities and Markets Authority (ESMA).

ESMA is responsible for managing the SFTR reporting regime. This includes defining the technical reporting standards, issuing guidelines, controlling reporting processes, supervising trade repositories (TRs), and improving data quality. The large volume of the data (multiple tables are collected with dozens of reportable fields of daily information for all market participants), its complexity, quality concerns, and confidentiality issues have so far constrained the use of SFTR data for policy analysis.

2. Data description

The SFTDS captures all SFT trades for the euro area, as well as transactions by supervised euro area entities conducted outside the EU. In addition, it covers non-euro area trades in the EU that are euro-denominated or involve euro-denominated collateral. It is jointly managed by the ECB and seven national central banks (NCBs). The SFTDS helps the ECB, NCBs, and the ESRB meet their responsibilities, and also facilitates collaboration among participating institutions. It collects data by accessing an ESMA data hub, where the data reported from TRs is stored. The ECB processes the data, enriches them, and makes them available for users. ¹⁹

The SFTDS covers three main types of SFTs: repos, securities lending, and margin lending. This article focuses on repos, as these represent the largest share of securitized transactions and where, arguably, the data quality is most reliable. The data available to the ECB cover approximately 200,000 unique outstanding and 13,000 unique new transactions per day after data processing,²⁰ with detailed information on counterparties, loans, and collateral.

The dataset greatly increases the transparency of repo markets, which are important for the euro area from a financial stability perspective. More concretely, the data can provide snapshots and trends over time of: (i) cyclical and structural changes in funding conditions (for example, rates, haircuts); (ii) the role of counterparties; (iii) the degree of maturity, liquidity, and currency mismatches (iv) leverage; (v) collateral metrics (for example, composition, transformation, and re-use); (vi) interconnectedness (cross-border and cross-sector); (vii) concentration risks (for example, collateral type, counterparty); (viii) issues around reporting dates (for example, window dressing behaviour²¹).

The SFTDS complements other widely used data sources on the euro area repo market such as the ECB’s money market statistical reporting (MMSR) dataset²² and the International Capital Market Association (ICMA) semi-annual European Repo Survey.²³ It allows a more complete and system-wide perspective on the repo market across multiple dimensions. For example, compared to the MMSR, the SFTDS provides several key advantages for euro area authorities in monitoring and addressing financial stability risks. First, it provides a system-wide perspective of the entire euro area repo market by including a broad set of counterparties. This allows for a novel view of the entire euro area repo market structure and linkages, with risks in the non-bank sector of particular interest. Second, it includes trades denominated not only in euro, but also in other currencies. This allows to shed light on the market, given that previous euro area stress episodes have included an important channel of US dollar funding constraints.²⁴ Third, it includes trades of any maturity, which allows for a more detailed view on the term structure of repos. It therefore also provides a complete picture of the corresponding risk exposures of trades. Fourth, for centrally cleared trades, the MMSR only reports the CCP as the counterparty for end users, whereas the SFTDS allows for matching of the end users of these trades. Fifth, it also includes intragroup trades, which can be an important component of overall repo activity as well as the international dimension of repo activity. However, it is important to note that the SFTDS is currently constrained by a limited sample period, with data available only starting from June 2020, which might impact the robustness of analysis. In particular, the period has been influenced by the Covid-19 pandemic and subsequent rapid tightening of monetary policy conditions which might affect the overall generalizability of certain findings.²⁵

A key characteristic of the SFTR reporting regime is that it requires EU entities to which it applies to also report their repo activity outside the EU. For example, some major European banks play an active role in the US repo market, so this activity is captured. Similarly, the disruption in the UK gilt episode last year largely involved liability-driven investment (LDI) funds domiciled and supervised in Ireland conducting their repo activity with UK counterparties, which is also captured in the data.²⁶ This provides a unique opportunity to monitor cross-border risks and to determine whether regulatory or supervisory action is required in the EU to limit the potential for spill overs to other jurisdictions.

III. FINANCIAL STABILITY ANALYSIS OF THE EURO AREA REPO MARKET

In this section we illustrate the potential of the SFTDS for conducting analyses on the repo market that are relevant from a financial stability perspective. We focus our analyses along three broad dimensions: (i) market segments (for example, jurisdiction, currency, central clearing); (ii) counterparties; (iii) collateral metrics (for example, asset quality and haircuts).

1. Market segments

We focus on five key dimensions: (i) jurisdiction; (ii) currency; (iii) internal or external group activity; (iv) central clearing; (v) the distinction between general and specific collateral. Initially, we examine the first three of these and then go on to consider the last two, restricting ourselves to the euro-denominated repo market.

A key feature of the SFTDS is the inclusion of trades by euro area entities via their foreign branches based outside the euro area. This is especially key for measuring interconnectedness and the risk from international shocks. Cetorelli and Goldberg (2011) showed that global banks played a significant role in transmitting the 2007–09 crisis to emerging economies.²⁷ A key component of this transmission is the internal capital markets of these global banks. Ivashina and others found that European banks’ reliance on wholesale US dollar funding was a key amplifier in the reduction in their dollar lending in the euro area during the sovereign debt crisis.²⁸ Information on the jurisdiction of banks’ international activities is important and should be considered when measuring systemic vulnerabilities.²⁹

Figure 1 reports the outstanding volumes with a breakdown of repo activity by currency, jurisdiction, and intragroup activity. Figure 1(a) shows that euro-denominated transactions make up the largest part of the euro area repo market, followed by those denominated in dollars. A sizable part of these transactions is due to intragroup activity. Figure 1(b) shows the cross-border dimension of repo activity with a breakdown by currency denomination. To capture the cross-border dimension, we distinguish between trades where the counterparty activity is defined as ‘euro area’ or ‘foreign’. ‘Euro area’ captures the activity of counterparties where both the branch and parent group are based/headquartered in the euro area. ‘Foreign’ captures the activity of entities where either the branch or parent group is based/headquartered outside the euro area. Column 1 shows the outstanding volumes of trades involving only ‘euro area’ counterparties. These transactions are mostly euro-denominated. Columns 2 and 3 show the activity when a ‘euro area’ counterparty engages in a transaction with a ‘foreign’ counterparty. When a ‘euro area’ counterparty borrows cash from a ‘foreign’ counterparty (column 2), over half of those transactions are euro-denominated. When a ‘euro area’ counterparty lends cash to a ‘foreign’ counterparty (column 3), less than half of those transactions are euro-denominated. Lastly, column 4 shows the activity involving only ‘foreign’ counterparties.³⁰ Here, we see that roughly half of the transactions are denominated in a currency other than the euro. The above analysis shows that the euro area repo market involves a significant degree of cross-border activity and may thus be subject to a significant transmission channel for shocks originating outside the euro area.

Abstracting from the dimensions of currency and jurisdiction and eliminating intragroup trades, two important dimensions remain for defining market segments. One is the role of central clearing. The second is whether trades are secured by general or specific collateral. From these two dimensions, four core segments can be derived: (i) non-centrally cleared with general collateral (cash demand driven); (ii) centrally cleared with general collateral (cash demand driven); (iii) non-centrally cleared with specific collateral (collateral demand driven); (iv) centrally cleared with specific collateral (collateral demand driven).

With central clearing there are a few important aspects from a financial stability perspective. The first is the ability of financial intermediaries to net their repo and reverse repo transactions, which can help alleviate balance sheet constraints. The second is the reduction of counterparty risk because of the use of a CCP.³¹ Clearing a transaction via a CCP allows the entities involved to mutualize their risk.³² This said, a number of studies have analysed to what extent the CCP-cleared short term funding market is subject to other types of risks. For example, this market is found to be subject to liquidity risk,³³ while other studies have discussed the possibility of a CCP default, for example by providing evidence that investors may price in such risk in times of crisis.³⁴

In light of this, the share of central clearing in terms of volumes can be an important component in understanding dealers’ capacity to supply liquidity during periods of stress, as well as mitigating their risks from counterparty default. The distinction between general and specific collateral may be useful in indicating whether a trade is driven primarily by collateral demand or funding demand.³⁵ This consideration is relevant when assessing whether the repo market can serve as a source of funding in times of stress for market participants who may need to monetize assets to meet margin calls that require cash collateral. Specific collateral should not be understood as synonymous with special collateral. Specialness simply means the repo rate is below the general collateral rate. However, not all transactions that agree on a specific piece of collateral have this feature. Special collateral is therefore a subset of specific collateral.³⁶

Figure 2 shows the results for both these dimensions, with our analysis focusing on outstanding amounts. We examine euro-denominated trades as this is the market most comprehensively captured by the dataset and therefore where it is most representative. This is also for practical purposes, as it is the largest segment and can help focus our analysis, given the sizeable degree of heterogeneity. Most figures reported for central clearing in the euro area repo market focus on daily transactions, which typically report the share of central clearing at around 70 per cent.³⁷ When looking at the stock of repos, however, we find that the share of central clearing for outstanding trades is only around 45 per cent—significantly lower than what is reported in the daily transaction data.³⁸ This difference in terms of counterparty risk is important, as outstanding exposure is the relevant factor for determining dealer capacity.

In terms of underlying collateral we find that the euro area repo market is primarily driven by trades which involve specific collateral. This could suggest the euro area repo market is mostly collateral-driven, consistent with the period of asset scarcity in government bond markets.³⁹ While this feature is well documented in the literature, it is important to note that a trade that is identified as specific does not necessarily imply specialness, as explained above. We would therefore caution that any interpretation implying general collateral serves general funding demand and specific collateral serves specific demand is far from clear cut. Rather, it is likely that a sizeable portion of trades involving specific collateral serve more than one economic function, implying that the provision of funding may not be contingent on the demand for collateral alone.

2. Counterparties

Our second application of the data examines the role of counterparties and their potential different uses of the repo market. A stylized way of classifying the role of counterparties from a supply and demand perspective is: (i) funding demand (repos); (ii) cash investment (reverse repos); (iii) collateral demand (reverse repos); and (iv) collateral investment (repos). However, for some economic functions counterparties may engage in more than one of these—market-making or collateral transformation, for example. When analysing counterparties in repo markets it is important to distinguish between these different drivers of demand and supply.

On the demand side, repos can enable investors to finance leveraged trading strategies, which can in principle contribute to price discovery and market liquidity. However, several recent episodes have shown how vulnerabilities from such strategies can spill over into other markets. Avalos, Ehlers, and Eren found that increased demand for funding from leveraged non-bank financial institutions via US Treasury repos appears to have compounded the strains of the temporary factors in repo markets in September 2019.⁴⁰ In March 2020 US hedge funds using repos to fund the basis trade caused significant stress in US Treasury markets.⁴¹ In September 2022, LDI funds, domiciled in the euro area and trading with UK counterparties, used repos to leverage their exposure to gilts, which created solvency and liquidity issues when the value of these bonds fell as yields increased.⁴²

On the supply side, the repo market is an important factor in market participants being willing to temporarily monetize liquid assets to meet margin calls that require cash, helping to reduce one-sided selling pressures in a ‘dash for cash’. Key to this are the capacity and incentives dealers have to provide liquidity, which are affected by balance sheet constraints, risk tolerance, and how they discriminate between market participants. It is important to look at dealers’ net positions, as these may indicate the degree to which they are performing a market-making role and whether there are unrealized netting benefits which could be captured with central clearing, improving their intermediation capacity. Understanding these aspects is essential to address risks related to the supply of funding. Breckenfelder and Hoerova found a dramatic decrease in bank cash lending to mutual funds in the repo market in March 2020, which affected the liquidity positions of these funds during the crisis.⁴³ In the UK repo market, Hüser and others found a significant increase in volumes traded in the cleared segment of the market and interpreted this as a preference for dealers and banks to transact in the cleared rather than the non-cleared segment.⁴⁴ They also found that the amount of funding to non-banks decreased, suggesting a reluctance among dealers to take on risk.

Figure 3 shows the volumes for each sector broken down by the three major currencies in the data, namely the euro, dollar, and sterling. For euro-denominated transactions, we see that banks account for the largest share. Foreign banks and investment banks tend to lend more cash than they borrow, whereas commercial banks borrow more cash than they lend. Regarding non-banks, IFs account for the largest share in euro-denominated transactions.

For dollar-denominated transactions, banks again account for the largest share of volume. The dominant type are investment banks, which tend to borrow slightly more cash than they lend. However, foreign banks and commercial banks are also active participants in dollar-denominated transactions. In terms of non-banks, investment funds again account for the largest share, however, other financial institutions (OFIs) also engage in a considerable amount of cash lending.

Sterling-denominated transactions differ strongly between borrowing and lending. In terms of borrowing, investment funds are the largest sector. Regarding lending however, foreign banks and commercial banks account for the largest share in volumes, whereas investment funds are much smaller.

Figure 4(a) dives deeper into the euro-denominated transactions and shows the volumes for each sector broken down by repos and reverse repos for both centrally cleared and non-centrally cleared trades. Figure 4(b) shows the net positions for each sector (ie the difference between cash borrowing and lending). For centrally cleared trades we see that foreign banks are the largest sector in terms of gross position, followed by euro area commercial banks and then euro area investment banks.⁴⁵ Foreign banks are the largest net cash lender (collateral borrower) and euro area commercial banks are the largest net cash borrower (collateral lender). As far as the potential for netting gains from centrally cleared trades is concerned, this should in principle be examined at entity level. However, we find that euro area investment banks as a sector appear to take full advantage of netting gains, as their net positions are very low compared to their total gross positions. This does not rule out the possibility that a significant share of banks in the other sectors may also be taking advantage.

Non-centrally cleared trades involve a broader range of counterparty sectors, especially non-banks. In terms of gross position we find that euro area investment banks are the largest sector, followed by euro area commercial banks and then foreign banks. For net positions, euro area investment banks are marginally cash lenders, while commercial banks are marginally cash borrowers. We also see that foreign banks are much less active in non-centrally cleared than centrally cleared trades.

In the non-bank sector, where the risks related to the demand for repo may be greater, we find a considerable role for IFs and ICPFs. IFs are the non-bank sector with the largest gross positions but are marginally cash lenders in net terms. They show a high demand for collateral, for example to engage in collateral transformation, where the collateral they pledge in repos is of a lower quality than what they acquire in reverse repos. IFs may also have a demand for funding, for example to acquire even more collateral. However, their main objective is likely to be collateral demand, resulting in a net lending position. ICPFs are the largest non-bank sector in terms of net positions; being cash borrowers; this indicates that they potentially use the market to access cash buffers (for example, to meet cash margin calls) or simply generate a return by lending out collateral that commanding a specialness premium.

We next present an application of the dataset to the LDI episode where we analyse the composition of outstanding transactions over this crisis period. This analysis showcases how the SFTDS can be used to better understand how repo transactions impact the balance sheet of entities. Figure 5 shows the volume of sterling denominated transactions that have gilts underlying, broken down by sector. Figure 5(a) shows cash borrowing. It can be observed that all sectors besides the investment fund sector (which includes LDI funds), which makes up approximately 45 per cent of these transactions, stay relatively constant in terms of volume. The investment fund sector, however, experiences a gradual decline. In particular, cash borrowing by investment funds drops by approximately EUR 40 billion shortly after the start of the episode. Additional analysis has shown that this drop comes with a lag instead of taking effect immediately because of the term structure of the transactions.⁴⁶ The funds are mostly engaged in transactions with tenors between three month and one year which means that these transactions can only gradually expire and thereby remain on the balance sheet for some time. This showcases an important feature of the SFTDS, namely the inclusion of transactions across all tenors. Due to this feature, it is possible to more accurately analyse how the balance sheets of entities evolve over time.

Figure 5(b) shows cash lending. Contrary to cash borrowing, it shows that all sectors are relatively constant in their lending, apart from foreign banks, which experience a drop of approximately EUR 40 billion. Thus, it is likely that the decline we observe during the LDI episode relates to investment funds reducing their borrowing from foreign banks. In fact, the large majority of foreign banks in these transactions are from the UK.

Contrary to the above analysis, the SFTDS can also be used to analyse immediate reactions of entities to certain crisis events. To do this, flow volumes⁴⁷ must be considered instead of outstanding volumes. This is because in flow data every transaction is only recorded the first time it appears. Thus, it reflects the new activity of entities, whereas outstanding data captures activity from a balance sheet perspective. Figure 6 offers an application of the banking turmoil. It shows the flow volume of dollar and euro-denominated lending and borrowing by banks during the stress episode. Figure 6(a) shows dollar denominated transactions, whereas Figure 6(b) shows the euro denominated transactions. However, it can be observed that neither type of transactions exhibits strong immediate reactions to the banking turmoil. A possible explanation could be that this crisis did not lead to extreme stress of euro area banks the way it did for US banks, for example. This does not exclude the possibility of reactions at the individual bank level; however, at the sector level we do not see a strong reaction in the euro area. Structurally, however, it is still interesting to observe that most of the short-term transactions in dollar are conducted with non-banks lending to banks, which can be inferred from the sector breakdowns.

3. Collateral

Our third application of the dataset focuses on collateral and examines differences in asset composition and haircut practices in non-centrally cleared transactions. A key concern in repo markets is counterparty risk—the risk that the provider of collateral defaults and fails to repurchase it. This leaves the cash lender with collateral that may have lost value or be illiquid.⁴⁸ A related risk is wrong-way risk, where the value of the collateral is negatively correlated with the credit risk of the cash borrower, meaning the collateral is expected to lose value at the same time the counterparty is more likely to default. This can occur when the collateral is exposed to similar economic or market factors as the cash borrower. Barbiero, Schepens, and Sigaux have shown that borrowers from the same country as the collateral they pledge pay a premium.⁴⁹ A third risk is excessive collateral re-use, which may lead to the build-up of leverage by creating a chain of interconnected exposures that can amplify market shocks and contagion.⁵⁰

Figure 7 shows the differences in collateral composition by cash borrowing and lending activity, distinguishing between banks and the non-bank sector. We deliberately focus on the non-centrally cleared segment, as again this is where financial stability risks are expected to be more pronounced, and also because centrally cleared transactions are almost exclusively backed by sovereign securities. To tease out further the financial stability risks, we examine the differences in asset composition between banks and non-banks. For banks and non-banks, we find that while the large majority of trades are backed by sovereign collateral (around 70 per cent), there remains a sizeable role for non-sovereign collateral. We also find that the non-bank sector tends to pledge lower quality collateral in its borrowing relative to the collateral it receives against cash, while the reverse is the case for banks. However, aggregate figures may mask important heterogeneity at entity or sub-sector levels.

Figure 8 shows the share of positive haircuts for non-sovereign collateral by different asset type. Haircuts are an important risk mitigation tool since they essentially limit the amount of leverage that can be obtained from repo. With low or even negative haircuts, entities can engage in chains of repo transactions that can lead to excessive levels of leverage. Moreover, haircuts may also be useful for banks in mitigating potential counterparty risks emanating from the non-bank sector. We find that the share of positive haircuts for non-sovereign collateral is typically quite low, for both banks, and non-banks. Positive haircuts are highest for securitized products, which make up a significant share of non-bank cash borrowing. For banks these are mostly driven by financial debt securities and securitized products. This further underscores the need to make progress on implementing regulatory rules aimed at ensuring minimum haircut practices for the non-bank sector in the non-centrally cleared space.⁵¹

Figure 9 dives deeper into haircuts of non-government securities and shows the flow volume of transactions that have a positive haircut and non-government securities underlying over time, while distinguishing between banks and non-banks. In this figure we again focus on the flow volume instead of outstanding in order to capture reactions of the market to stress events. In particular, this shows that the SFTDS can be used to capture haircut dynamics. Three stress events are indicated in the charts, namely the commodity turmoil, the LDI episode, and the banking turmoil.

Figure 9(a) shows volumes where banks are borrowing. We can see that the activity of transactions with positive haircuts and non-government collateral underlying picks up significantly after the commodity turmoil. Particularly, before this event, the share of positive haircuts for these transactions was between 10 per cent and 40 per cent. This was dominated by haircuts between >0 per cent and 2 per cent. After the event, however, we see a pick-up in haircuts in the range 2 per cent to 4 per cent. The share of positive haircuts rises to a range between 60 per cent and 80 per cent. This then decreases, but shortly after the LDI crisis, there is again a pick-up. This seems to be slightly decreasing or staying flat, until the banking turmoil starts and these shares begin picking up again. Figure 9(b) shows the volumes where non-banks are borrowing. While there is also a slight upward trend to be noticed, it is not as clearly linked to stress events as in the case for banks borrowing. In future applications, the SFTDS could be used to study further the dynamics of haircuts in both normal and stress times as well as to provide further insights in the determinants of haircuts.⁵²

IV. FUTURE APPLICATIONS AND IMPLICATIONS FOR POLICY

The above analysis indicates that the SFTDS can make important contributions both to regulatory efforts to enhance the resilience of the financial system, and also more broadly to other areas of importance to central banks, as well as complementing the existing economic literature on repo markets.

First, from a financial stability risk monitoring perspective, the dataset greatly increases transparency and helps authorities obtain more timely and comprehensive visibility into trends in repo markets. It provides an exhaustive picture of repo market activity within the EU as well as activity associated with EU-supervised institutions outside it. In particular, the data can provide snapshots and trends over time of: (i) cyclical and structural changes in funding conditions (for example, rates, haircuts); (ii) the role of counterparties; (iii) the degree of maturity, liquidity, and currency mismatches; (iv) leverage; (v) collateral metrics (for example, composition, transformation, and re-use); (vi) interconnectedness (cross-border and cross-sector); (vii) concentration risks (for example, collateral type, counterparty). By exploring the dynamics of repo markets along these seven dimensions and exploiting the rich, comprehensive, and high-frequency nature of its underlying data, future analysis with the SFTDS could shed further light on the question how heterogeneity in market structure matters for resilience against systemic shocks across different parts of the repo markets. Such future work could, for example, deepen the analysis of the LDI and 2023 banking turmoil episodes presented in this article to provide further analysis on the channels by which these shocks have or have not impacted repo markets. Also future work could study how the reduction of central bank balance sheets affects the structure of repo markets and risk mitigation techniques. Such work could build on and add to the existing literature that has revealed the importance of the underlying market structure for the existence of vulnerabilities.

Second, the dataset will play an important role in current international policy initiatives to enhance global financial stability resilience, of which a key priority is to address vulnerabilities in the non-bank sector. More specifically, it could be used to inform the development of policy recommendations by providing an ex ante impact assessment for policies targeting the build-up of leverage in the non-bank sector. As part of its broader response to address risks related to SFTs, the FSB has issued a comprehensive list of policy recommendations.⁵³ This includes minimum standards and haircut floors on non-centrally cleared SFTs targeting the procyclical build-up of leverage in the non-bank sector. However, progress in implementing these reforms has been slow, and the dataset will be valuable for assessing not only the progress made but also whether enhanced requirements will be necessary. The comprehensive coverage of transactions involving non-banks allows to assess the design and the calibration of the FSB haircuts and the impact of haircut floors on euro area repo markets and leverage used by non-bank financial intermediaries. The SFTDS could also be merged with entity-level information (for example, data on alternative investment funds) to assess the complementary nature of different policy tools (for example, minimum haircuts and entity-level limits) affecting leverage in the non-bank financial sector. A related issue highlighted by the FSB is the role of collateral re-use.⁵⁴ While we have not explored this dimension here, partially due to ongoing data issues, inclusion of this variable in the dataset will be important for monitoring these risks and determining whether policy action is required.

Recent stress episodes have shown important non-bank amplification channels due to excessive leverage obtained in the repo market. For example, in March 2020, leveraged hedge funds using repos to fund basis trades caused significant stress in US Treasury markets.⁵⁵ The SFTDS could be used to investigate the presence of basis trade investment strategies in euro area and euro-denominated markets. This is because the dataset also fully covers non-bank activity in these markets. Thus, by combining the repo activity of investment funds with data on futures, inferences can be made on investment strategies and their potential risks. Our analysis of the LDI episode in section III.2 indicates the usefulness of the SFTDS in this respect. The FSB has already undertaken initiatives to address these risks. The Basel Committee on Banking Supervision (BCBS) is also currently exploring whether banks’ risk management of exposures to leveraged non-bank entities could be enhanced.⁵⁶ A key part of this will be examining their counterparty risk-management practices with non-banks in repo markets.

Related to this are policy efforts to ensure that repo markets function during times of stress, which is an important component of mitigating the impact of large margin calls on derivatives that require payment of cash collateral. A key part of this is ensuring market participants have access to liquidity to meet margin calls. However, while the repo market can be a useful component of contingent funding sources, excessive reliance could put a strain on how smoothly it functions. A recent FSB report on liquidity in core government bond markets also examined ways to increase the availability and use of central clearing for government bonds, cash, and repo transactions in particular.⁵⁷ The monitoring of repo market liquidity, in terms of dealer capacity and demand for such liquidity, will be important to support policy efforts aimed at ensuring repo market resilience.

The dataset can also play an important role for central banks in evaluating how the repo market functions more broadly for monetary policy purposes. From a euro area perspective, the consensus has been that repo markets have generally functioned well. However, in the current period of monetary tightening, intermediation capacity could be more constrained, and it is important to examine developments more closely.⁵⁸ Other important issues are the impact of specialness in the repo market and a better understanding of the growing role of the non-bank sector for monetary policy purposes. Using the SFTDS to further examine these issues would complement existing work using other data sources that have been instrumental in shedding light on monetary policy transmission via repo markets. For example, using data from repo trading platforms, Corradin and Maddaloni as well as Arrata and others have assessed the impact of ECB bond purchases on repo rates and the degree of ‘specialness’ of specific bonds.⁵⁹ Similarly, Ballensiefen and others documented segmentation in the money market where repo rates lent by banks with access to the ECB’s deposit facility and secured against QE eligible assets are collateral-driven and disconnected from benchmark rates.⁶⁰ Also, Eisenschmidt, Mab, and Zhang, using the ECB’s proprietary MMSR data, investigated how dealer market power affects the passthrough of monetary policy in repo markets.⁶¹ Exploiting the same dataset, Carrera de Souza and Hudepohl found that the Eurosystem’s purchase programmes during the Covid-19 crisis have put significant downward pressure on interest rates.^{62  63} These analyses are limited to data from specific market segments, as for electronic trading platforms, or to a fixed set of counterparties (for example, banks), as for MMSR. By expanding the universe of market participants and allowing a holistic view of the euro area repo market, the SFTDS can provide a more in-depth understanding of the transmission of monetary policy, especially in the context of non-bank financial intermediaries.

Finally, the SFTDS may be used to support policy discussions on the impact of regulatory reforms on repo markets. For example, the granular information on counterparties’ activity in repo markets can be exploited to understand how regulations have impacted the participation of specific types of counterparties in repo markets. Such analysis would contribute to the already existing literature on the effects of post-crisis reforms on repo markets that has used different data sets. For example, the Committee on Global Financial System has already provided a comprehensive overview of how post-crisis regulatory reform requirements have affected market functioning.⁶⁴ Ranaldo, Schaffner, and Vasios have shown, with data from trading platforms, how the European Market Infrastructure Regulation (EMIR), interacting with the Basel III leverage ratio, induced a significant downward pressure on short-term repo rates as CCPs supplied cash posted as margin to the repo market.⁶⁵ Similarly, Kotidis and Van Horen, using data from the Sterling Money Market Database (SMMD), have shown how the introduction of the leverage ratio affected repo market intermediation, documenting a reduction in repo activity by dealers affected by the regulatory requirement in the UK.⁶⁶ Lastly, Bassi and others, through the use of MMSR data, have illustrated how the reporting of regulatory ratios at quarter and year-ends incentivizes window-dressing behaviour around reporting dates.⁶⁷ The SFTDS can greatly help expand on this literature. The matching of end users of centrally cleared trades allows for a more comprehensive picture of CCP repo market activity. Moreover, the SFTDS will be instrumental in monitoring the impact of future regulatory activity in the non-bank sector.

V. CONCLUSION

In this article we have shown how the data reported under the SFTR greatly increase market transparency and will help authorities obtain more timely and comprehensive insights into trends in repo markets. In particular, the data will help authorities to monitor important risks related to interconnectedness, concentration, liquidity provision, collateral demand, liquidity/maturity mismatches, leverage, and various collateral metrics. It might also play an important role in current international policy initiatives to enhance global financial stability and can assist central banks in evaluating how the repo market function more broadly. The results we have presented highlight how the dataset can shed light on key financial stability issues for the euro area repo market. However, we have only touched on these issues; there is scope for a much more comprehensive analysis to gain deeper insights into these issues.

Footnotes

Author notes