At the end of December 2016 the European Repo market experienced an unprecedented level of volatility and dislocation, raising worries over the stability of the whole European Money market. During the trading day of December 28th the repo-rate for specials of several ‘core’ government bonds (including Germany, France, Belgium, Netherland and Austria) tightened to record-low levels, with transactions on some ISINs being carried out at levels as tight as -20% on annual basis.

Given the importance that the Repo market has in the smooth functioning of money market and in monetary policy transmission, the discussion over the causes that led to such break down and its potential implications is currently on the table of both regulators and policy makers. The International Capital Market Association (ICMA) published a report in February 2017 analysing the dynamics of what happened during the last months of 2016 and Benoît Cœuré, member of the Executive Board of the ECB, gave a speech in Paris as recently as April 3rd discussing the potential distortions of ECB’s Asset Purchase Program (APP) with respect to the alignment between short term rates and government bond yields. It appears then of great importance, especially going forward, to have an overview of what repos are and how they significantly affect the broad financial market. The attempt of this article will be to provide such tools.

Repurchase Agreements – or ‘repos’ as they are commonly known – can be defined as an agreement whereby a party sells securities (or some other asset) to a counterparty at a fixed price (‘First leg’ of the transaction) and simultaneously agrees to repurchase such securities (or assets) at an agreed future date and price, the latter taking into account a return on the use of the sale proceeds during the term of the repo (‘Second Leg’ – Fig.1).

As such repos are essentially a form of secured loan where the collateral is not simply pledged to the lender, but sold and repurchased at maturity. As a consequence between the sale and repurchase dates

  • The borrower (repo seller) gets the use of the cash proceeds of the sale of the assets.
  • The lender (repo buyer) gets legal title to the securities in exchange for the cash he has paid. In case of borrower’s default he can liquidate the securities held as collateral in order to recover all (or some) of the cash paid. Moreover, being legally the owner, he can re-use the collateral by selling it outright, pledging it as collateral in another transaction or even repoing it to a second counterparty (in this case he will act as a repo seller). Nonetheless, as the borrower remains exposed to market risk, the repo buyer is required to transfer any payment arising from the securities (i.e. coupons or dividends) to the other party, through the so-called ‘manufactured payments’.

While the borrower is said to have entered a ‘Repo’, on the other side of the transaction the repo buyer is said to have transacted a ‘Reverse Repo’, since he will buy and re-sell the securities involved in the deal.

Depending on its legal structure a repo transaction can be qualified either as a ‘classic repo’ or a Buy/Sell-back. In the latter case the two legs are considered legally distinct trades and this will have specific consequences in terms of collateral management and margining. Although both models are practically used (in Italy for example it is more common the buy/sell-back framework), the focus of what follows will be on classic repos.

Fig.1 – A stylized repo transaction

The are four main reasons for which either one of the two parties may wish to enter a repo transaction:

  1. Safe Investment: one party (the repo buyer) can invest excess cash and earn interest against the safety of the asset provided as collateral.
  2. Cheap Borrowing: the counterparty (repo seller) can borrow cash in order to finance a long position in an asset.
  3. Yield Enhancement: one party (repo seller) can earn incremental income by lending out some securities within his portfolio and by investing the cash proceeds.
  4. Short Covering: the counterparty (repo buyer) may need to borrow a specific security in order to cover a short position taken with an outright sale in the cash market.

According to the initial purpose of the transaction it is possible to distinguish between Cash-Driven repos, where the borrowing party identifies a specific amount of cash that it needs against the collateral available, and Securities-Driven repos, where a repo buyer identifies a specific security that it needs to borrow. Cash-Driven repos are extensively used in money market interbank transactions as a form of secured funding and, given that the primary focus is not on the nature of the securities involved in the transaction, they can be identified with the so called General Collateral (GC) repos (this is typically the case for repos initiated as a form of safe investment or cheap borrowing). On the other side, as repo buyers look into the repo market for specific securities, Securities-Driven repos are generally executed against a precise form of collateral, which is said to be ‘special’ (yield enhancement and short covering goals represent usually the main interests around Securities-driven repos).

As a consequence of their different nature, repo rates on specials are lower than GC repo rates and the difference between the two represents the cost for the repo buyer to borrow the ‘special’ collateral. This form of opportunity cost can be thought to be the same as the cost of borrowing the same security in the Securities Lending market.

Concerning the type of securities used as collateral in repo transactions, although in principle these can be any high-rated and liquid asset such as Covered bonds, ABS, MBS, Certificates of Deposits and Commercial Paper, 80% of the EU-originated repos are traded against Government bonds. This makes the case for a high level of interconnection between the repo market and government bond yields, which in turn are used to price a broad array of financial products.

Repos usually take short term maturities, but their feature may vary across countries. In the US the vast majority of repos are conducted on an overnight basis (with the position being rolled over each business day), while in the EU ‘term repos’ with maturities longer than one day represent more than 80% of the overall market.

Among the most common players in the repo market we find:

  • Securities Dealers: they have been among the earliest users of repos, driving their development throughout the 1980s. They originally entered the market to fund long positions and increase their leverage at lower costs. As they face sudden demand for some securities they may well be required to borrow them in the repo market in order to fulfil their settlement obligations (most-liquid settlement dates for securities financing transactions are usually one day shorter than standard settlement dates for the underlying, special repos usually settle at t+2).
  • Highly Leveraged Investors: this category includes hedge funds and other ‘alternative investment’ companies. As their high level of leverage acts as a constraint on their unsecured borrowing activities the repo market can provide a relatively cheap funding alternative.
  • Prime Brokers: They enter the repo market on the behalf of hedge funds and provide them with securities services (including repo financing, securities lending, execution and clearing).
  • Repo Conduits: these are Special Purpose Vehicles (SPVs) that finance themselves by issuing Asset Backed Commercial Paper and invest in reverse repos (this is an example of a re-use of collateral by a repo buyer).
  • Risk-averse cash investors: they include Money Market Funds and Agent Lenders, who are looking for low-credit and low-liquidity risk investments.
  • Central Banks: they use repos as a tool for conducting monetary policy (an example are ECB’s Main Refinancing Operations), observing expectations of future short term interest rates and providing emergency liquidity.

Because of its double advantage of providing at the same time a cheap source of funding and a safe investment tool, the market for Repurchase Agreements has rapidly increased in size, becoming a critical contributor to the efficient functioning of global capital markets. They proved to be a resilient source of interbank funding during the financial crisis of 2008, partially easing the pressure on central banks’ balance sheets, and their secured nature has been rewarded by regulators, who have set lower capital requirements for repo exposures (thus providing a clear incentive towards their usage). The size of the European market stood at approximately EUR 5,379 billions as of June 2016 (Fig. 2) and its daily volumes are estimated to exceed EUR 300 billion.

Fig.2 – data source

Even though they are rationally considered as a low-risk investment, repos are far from being riskless. They are subject to credit risk in the measure that the borrowing party fails to meet its obligations (counterparty credit risk) or the issuer of collateral defaults, and they come with a significant exposure to liquidity risk, that is the risk for the repo buyer of not being able to liquidate all of its collateral at a fair market price (this explains most of the reliance on government bonds as collateral for repo transactions). Moreover, repurchase agreements parties are exposed to both operational and legal risk when it comes to collateral management and default issues, as it will be briefly explained.

As counterparties of a repo trade want to ensure that the value of the collateral is equal to the cash advanced (this can prove to be difficult by market volatility and valuation issues, which together contribute to liquidity risk) the repo buyer will apply either an ‘haircut’ or an ‘initial margin’ to the borrowing party. The former is a percentage reduction applied to the given market value of the collateral in order to determine the maximum borrowable amount, that is (assuming a bond as a collateral):

 The initial margin instead can be thought to as the collateral face value needed to ensure a pre-determined borrowable amount given the observed market price, that is:

After the inception of the trade and during the whole life of the transaction it will be in the interest of both parties to monitor the value of the collateral to make sure this is always in line with the required initial margins. Given the difficulty and the operational risk embedded in the timely computing and delivery of the so-called ‘variation margins’ (i.e. the difference between the registered market price and the original market price of the collateral, taking into account accrued interest) the two parties of a repo can decide to outsource this activities to a third, neutral institution, called ‘tri-party’ agent. This will act as a common custodian of the collateral, managing all the issues related to, among others, variation margins and manufactured payments (‘tri-party repo’). In Europe, the major tri-party agents are Clearstream, Euroclear, JP Morgan, BNY Mellon and SIS.

As the performance of a repo depends on the buyer’s rights to collateral in case of the seller’s default, it is common practise to minimize legal risks via a pre-agreed written contract. This is usually achieved through an internationally recognized form that goes under the name of Global Master Repurchase Agreement (GMRA). Such documentation, first published in 1992, has the two-fold function of facilitating trading and setting out clearly the rights and obligations of both parties.

Going back to the prologue. At the end of December 2016 the European repo market came under significant pressure, with dealers and fund managers reporting having hard times in borrowing High Quality Liquid Assets (HQLA), especially in the three days prior to the calendar year’s end. On December 28th the market broke-down, with German and French GC repo rates trading at levels below -8% and specials rates tightening to the record-low levels of -15% and -20% in some cases. Repo buyers, those looking for GC and special securities, had then to face the hard choice of paying such high level of (negative) interest rates in order not to fail on their previous transactions (imagine a dealer who had sold a German T-Bill in the cash market rushing to cover its position prior to the settlement date). Alternatively they could have bought the same securities on the cash market at inflated prices (3-months French bonds traded as low as -1.2% YTM on December 28th).

Although the evidence suggests that actual settlement fails did not increase throughout the year’s end, with market participants preferring to pay such inflated costs, the event has nonetheless shown the hidden fragility of the European repo market. For those claiming about the efficiency of financial markets, according to ICMA on December 29th an investor holding dollars could have swapped them into euros receiving a basis of +25% until the year’s end and then entered a reverse repo for the same term against, say, German T-bills collateral, closing in an annualized close-to-risk-free profit of around +17%. The fact that this was not the case is symptomatic of the level of the impairment experienced by the whole financial intermediation system.

At this point of the journey the question is whether such volatility peaks should represent a new normal in today’s money market. So far three main causes have been identified in relation to the events described above, helping to explain this point:

  1. Short Positioning: participants reports show that towards the year’s end the market experienced a sustained increase in the short positioning in European core government bonds, both as result of directional and basis trades, thus boosting the demand for such form of collateral within the repo market (as GC and specials French repo rates where among the most beaten, one could ask whether this was partly due to speculation on local political uncertainty and approaching political elections).
  2. ECB Quantitative Easing: as a consequence of the Public Sector Purchase Program (PSPP) the ECB had bought over EUR 1.25 trillion of government bonds as of December 2016, of which EUR 303 billion in German bunds (its single largest holding representing 30% of the total outstanding). This has surely impacted in a sensible way the overall supply of collateral for repo transactions. On December 8th the ECB decided to implement a securities lending program aimed at providing access to the securities acquired under its Asset Purchase Program (APP) against cash collateral, but set the overall limit at EUR 50 billion, thus addressing the issue in a partial way.
  3. Regulation: as new regulation, such as the one arising from Basell III, became more demanding in terms of liquidity held and overall activity carried out by some key market participants (the Liquidity Coverage Ratio and the Leverage Ratio being an example of such requirements), these were forced to control the size and composition of their Balance Sheets around key reporting dates. This is particularly true for quarter-end and year-end dates, namely for the period involved by the recent market break-down. Since international accounting rules look at the economic nature of financial transactions, the net effect on the BS of a repo seller is actually to inflate its size. This is clearly a deterrent for reporting participants to act as repo seller around key dates, thus sucking up collateral supply. Interestingly the end of the first quarter of 2017 saw a similar ‘flash crash’ in the repo rates for both GC and specials transactions, although not as large as the one observed in 2016 (Fig. 3).
Fig.3 – data source

As repurchase agreements remain a key tool for the smooth functioning of modern financial markets, contributing to the efficiency of capital markets and to informative prices, it is in the interest of all policy makers to ensure they do not go out of track. Since market infrastructural and regulatory reforms are implemented and monetary authorities have still to understand the full implication of their unconventional instruments, it will be interesting to see how they will tackle the considerations posed by the repo market in the near future.

Leave a Reply