After examining the main features of the repo market as well as some peculiar characteristics of the European credit repo market ( we would like to walk you through what happens when a repo fails (and why this might happen) and the recent reform of the CSDR, i.e. the Central Securities Depository Regulation. This reform will be put in place via RTS (regulatory Technical Standards) through which this reform has been ruled will start applying in September 2020. However, market participants are far from being happy with the new rules and the International Capital Market Association (ICMA) has published several papers (on which this article is largely based) outlining the risk of implementing this reform, particularly in relation to the mandatory buy-in feature and the shorter period before a buy-in must be executed.

A Preamble

In order to understand why the CSDR reform and the new mandatory buy-ins regimes are creating across headaches, especially on the sell-side, it is important to highlight some features of the market making activity in cash bonds.

Some financial intermediaries act as market makers on a specific security (bond) either because of a letter of liquidity or because of a “letter of market making guarantee” or because of an unwritten but clear incentive to do so. Let’s see both in details.

With a letter of guarantee (also called letter of liquidity) the market makers commit to provide quotes for an amount of time such that the liquidity on the secondary market of the security is ensured. The letters of liquidity are most common for government bonds but they are not really popular in the credit space.

Instead, what happens in the credit space is that the banks that took part in the origination and issuance of the bond provide liquidity in the secondary market not because of a formal agreement (i.e., a binding contract) but because not doing so would cause market players to complain with the issuer of the bond and this in turn may increase the possibility that the financial institution will not be receive the origination mandate by the issuing corporate again – in other words, not stand ready to provide a quote to investors that want to sell can cause a reputational damage to the financial institution that could translates into an economic damage . However, because of the much higher number of bonds traded in the credit market and the much lower frequency at which trades happen, this “unofficial” market making is usually limited to the “bid” – in other words, banks that have participated in the issuance of the bond should be ready to buy them back if investors want to sell, but they are not expected to provide the market with an “ask” price. This is due to the limited liquidity of the market that may prevent market makers to find the bond that the investor wants to purchase, particularly when it is an “off-the-run” security issued a long time ago. In addition to this it should be noted that, compared to the government bond secondary markets, financial institutions that were involved in the syndication process actually have an informational advantage over other investment banks that did not take part in the issuance as they actually know who bought the bond in the first place.

Clearly the fact that an investment bank is not always expected to provide an “ask” price does not imply that this is always the case – however the ability to provide the market with a selling price critically depends on the possibility for the bank to borrow that specific bond via a repo. Nonetheless, because of its nature, the repo market does not have any “official” quote. Rather, the repo desk of the financial institution, which is a separated desk, provides the bond trading desk with quotes. It is important to appreciate that because of the nature of the repo market, this cannot be done in real time and thus the market maker have to rely on quotes that may be outdated or, by the time that the trade is done, the bond may actually be no longer available.

To put it differently, the trader has to make a price without knowing if he will be able to borrow the bond and the price at which he will be able to do so. For this reason, the quotes of the market makers are “no guaranteed delivery” meaning that, at that price, the market maker is not assuring the counterparty that he will be able to deliver the bond in T+2. “Guaranteed delivery” trades must be specifically asked by the client, the market maker will usually not accept to enter into one of those and, even if he does, he will charge a relatively high premium.    

Settlement efficiency

For the reasons outlined in the paragraph before, settlement efficiency (i.e. the ability to settle transactions on their intended settlement date) is a critical factor when considering the overall efficiency of the credit repo market, not least since the primary reason for borrowing and lending specifics is to ensure that market-makers, and other participants, are able to settle their trades.

Corporate bond failures as a percentage of overall transactions around 3%, and only slightly higher than the rate for sovereign bonds. When it’s not the economic rational, the reputational and relationship incentives to settle trades seem to work well.

But, if this is the case, what cause a repo to fail? More often than not these are caused by settlement inefficiencies, rather than a deliberate intention of counterparties not to borrow against their short-sales. Main cause of incidents of prolonged fails in the credit repo market is the trend towards smaller trade sizes in the underlying market together with the economics of securities lending, i.e. settlement and administration costs.

In addition to this, the majority of market participants have a policy of minimum size repo and lending trades for corporate bonds, plus the trend toward more rapid dealer turn-over causes a further challenge to the economics of lending bonds, where the average duration of borrow may only be a few days, rather than a few weeks whilst some lenders request minimum borrowing terms of one-week.

The way of dealing with small size trades that used to be in place within repo desks was covering the short with a bigger size repo, i.e. covering a 250k short with a $1mn borrow, calculating the cost of carrying the excess amount on the matched-book (so $750k excess, in this example), and passing on this cost through the rate charged to the trading desk or the client. Similarly, with minimum borrow terms, repo desks used to carry the over borrow for the additional days, and again charge the cost of this back to the trading desk or client. However, given the scarcity and cost of such instruments on the balance sheet, as well as the relative increase in the number of odd-lot shorts that require financing, often repo desks are unable to justify running balances of excess borrow. This has made substantially more difficult to borrow small sizes of any corporate bond.   

It is important to note that in order to minimise the probability of a failure to deliver, prime-brokers do not permit their clients to enter short positions in credit without first obtaining a ‘good locate’ from the prime-broker itself (i.e. the prime-broke is comfortable that it can cover the short on repo on the client’s behalf). A ‘locate’ is a securities lending term, adopted by credit repo traders, which is the equivalent to a request for quote (RFQ) for borrowing. A ‘good’ is an indication from the credit repo trader that it is borrowable and therefore safe to short.

Buy-Ins, The CSDR Reform and the Future

What would have happened before the CSDR reform in the case of a fail-to-delivery? Usually the party that was supposed to receive the bond would have issued, after 15 days, a buy-in notice to the counterparty that was supposed to deliver the bond (usually, the market maker). A buy-in notice is used to communicate the failed counterparty that a guaranteed delivery mandate had been given to a third party agent and that the failed dealer would have had to pay for the cost incurred by the client to receive the bond from the third party agent.

It has to be noted, however, that a failure to deliver from the dealer to the client may originate as a result of another counterparty’s inability to provide the dealer with the specific bond that the dealer has to provide his clients with. Or alternatively, multiple fails can be the result of a single fail by one counterparty that impacts the whole chain of transactions in the underlying bond. Where counterparties are part of the chain, but not the cause of the fail, they have both a failing purchase and a failing sale buy-in (which makes buy-ins transferable along settlement chains), it seems to be an effective and efficient way of reaching the root cause of the fail, even though the failing entity may be completely unknown to the originator of the buy-in notice.

But so, what is a buy-in? Buy-ins are the standard contractual remedy and risk-management tools open to non-defaulting purchasers of bonds in the event that they are failed too, and which allows them to force delivery without changing the original terms or economics of the trade.

European authorities however have decided to partially reform the procedures associated with a failure to deliver, i.e. the so called CSDR reform. In order to do so, the European Commission finally published the regulatory technical standards (RTS) for the mandatory buy-in regime as part of CSDR Settlement Discipline on May 25th 2018.

These RTS encompass aims to change several features of the status quo. Among these, the most important are the reduction in the time frame before the start of the buy-in process, the fact that the buy-in becomes mandatory (with a cash compensation in the event that that it is unsuccessful, so that the buy-in agent is not able to find the bond that the dealer has failed to deliver).

Lowering from 15 to 7 days the time before a mandatory buy-in must be issued seems to create additional risks for almost everybody involved in the transaction, according to surveys conducted by the International Capital Market Association. In fact, market participants’ reactions to the new rules, especially the mandatory buy-ins, were far from enthusiastic, to say the least. In particular, among the key concerns expressed by market players are the fear that mandatory buy-ins would reduce the supply of corporate bonds to the repo and lending markets and the fact that the risk of losing a lender relationship due to the failed return and subsequent buy-in of an odd-lot corporate bond loan was not commensurate with the low revenues generated through lending corporate bonds to the market the risks of buy-ins. In addition to this, the lack of protection that sellers witness with respect to failing repos is something that market-makers need to consider as part of their pricing as well as their willingness to provide offer-side liquidity.

In other words, not only dealers are going to adjust their offer prices to account for the additional risk that the mandatory buy-in regime introduces, but this higher risk is also extremely likely to impact the volumes of the market.

As if this wasn’t enough, the mandatory buy-in regime also creates additional risks for investors. In fact, where a buy-in is not possible, the regulation dictates that the fail must be resolved with a cash compensation remedy. In the event of a settlement fail, purchasers of bonds may find that they are unwittingly forced out of their long positions in return for cash compensation (over which they have little or no control). This uncertainty and effective increase in market exposure could be particularly costly for investors where the bond purchase is hedged or part of a package, including other bonds or derivatives, and which would subsequently need to be unwounded as a result of the cash compensation.

A final remark: as mentioned in the introduction, this article as well as the first part published two weeks ago are largely based on ICMA public available papers. In case you would like to examine from a more technical point of view the topic, we would like to invite you to visit ICMA website at


Leave a Reply

Avatar placeholder

Your email address will not be published. Required fields are marked *