CSDR: The next headache in bond market liquidity


Graham Bippart
Published on:

A rule many thought had died silently in the legislative process is about to be resuscitated, and bond market pros say it will be devastating to bond market liquidity.


Mifid II is finally off the European Commission’s plate, so commissioners are turning their attention to a regulation some had hoped would have been quietly forgotten.

In particular, one article in the level-one text is riling lawyers and lobbyists representing the bond market.


Andy Hill, ICMA

“It’s an all-round destroyer of market liquidity,” says Andy Hill, a senior director in the International Capital Market Association’s (ICMA) market practice and regulatory policy group.

The Commission has signalled that it expects to endorse the European Securities and Markets Authority's (ESMA) regulatory technical standards for the Central Securities Depositories Regulation (CSDR) in the second quarter, according to a second industry lobbyist.

It hasn’t gotten much attention.

After all, it regards the European financial system’s plumbing – a topic most of humanity ignores until its failures manifest in ugly ways – but it does pose a vital threat to repo markets and the bond markets more broadly, experts say.

The crux of the regulation is its insistence that settlement fails be cured, mandatorily, within a certain time period. That doesn’t sound particularly important on the face of it, and perhaps even like a good thing, if looked at from the perspective that it is meant to ensure that settlement fails don’t disadvantage failed-to counterparties in a transaction.

However, crucially, the level-one text reverses the order of payments in the difference of a securities’ price from when the original transaction takes place.

Take a transaction in which a buyer buys a bond at $100, but for which the settlement fails – perhaps because the seller doesn’t have the security, or due to a back-office mistake. The buyer initiates a buy-in, but the bond’s price is now $101.

Normally, the failed-to buyer would elect a third-party who would sell the bond to them at $101, plus a premium – since it’s a distraction from the third party’s normal business activities to do this. The original seller of the bond – the ‘failing party’ – would then send $1 plus the premium to the failed-to party.

If the price of the security declines between the settlement fail and the buy-in, the original buyer pays the original seller the difference in price (minus the premium), since they got the bond cheaper during the buy-in process. Ignoring the premium, the original economics of the transaction are the same for both parties as they would have been if there was no settlement fail.

Epic fail

While the level one text of the regulation reverses the direction of those payments if the bond price drops after the fail, the level two text corrects that mistake, apparently a drafting error, to ensure the seller compensates the buyer if the bond price is higher at the buy-in. But, if the bond price falls, there is no payment from the buyer to the seller; meaning the seller has essentially sold a put to the buyer with a strike at the original trade price.

Add to this the fact that the rule makes buy-ins mandatory, rather than an optionality as they are now, would result in increased costs all around.

Most settlement fails aren’t bought in, says ICMA’s Hill. Among the reasons: sometimes they simply can’t be, as some securities are very illiquid; sometimes buy-in agents can’t be found.

Buy-ins can also be distortive if they signal the presence of a distressed buyer to the market, providing a disincentive to instigate them.

And, in cases where there’s a chain of settlement fails – ie when a party buys a security and then sells it before they even know the original purchase will fail – the rule would require every failed settlement in the chain be bought in, and within a certain time limit. Normally, settlement failure chains are cured by a single buy-in at the end of the chain. But the level two text doesn't explicitly provide for a pass-on of settlement fails. That would result in a big increase in buy-ins, and – because of the premium, not to mention the opportunity cost – increased costs to trading operations.

One senior banker told us ‘this is bigger than all the other post-crisis rules put together, when it comes to liquidity – and we thought we’d got rid of it’. 
 - Andy Hill, ICMA

And, in the case of a security's price falling below the original transaction price, it creates more losers. If trader B buys a security from trader A and sells it to trader C for a profit, and A fails, trader C would initiate a buy-in to B, and B to A. If the price of the security is lower than that of the original transaction, both traders A and B would lose money, even though B wasn't the cause of the fail. The mechanism, as the regulation is written, only works if the price is higher at every point in the chain. 

“The regulation is aimed at increasing settlement rates,” says the second lobbyist. “But settlement rates aren’t low.”

Settlement fails in government bonds are about 1% of trades, Hill says. Corporate bond trade fails are about 2%. And most settlement fails are cured with four or five days of the original settlement date, meaning they’d likely be out of scope of the mandatory buy-in rule.

However, as Hill points out, while the probability of failing is low, the cost of a mandatory buy-in as CSDR requires them would be “massive”.

Among the effects of the rule: banks will ask counterparties to put up margin against potential settlement failures or stop offering certain bonds all together. 

An impact study from ICMA said some 37% of respondents said they’d stop offering illiquid corporate bonds unless counterparties pre-positioned with them for a potential fail. Bid-ask spreads will increase – in the liquid corporate market by some 155%, according to ICMA. Repo markets could all but evaporate in certain segments: 60% of respondents said they’d stop offering term repo for illiquid corporate bonds. And 12% said they’d cease offering term repo for liquid corporates.

“The message we get from our members is that in a mandatory buy-in regime, you cannot afford to risk lending any security that could go special and you might not get back,” says Hill.

‘Special’ in the repo market designates an asset subject to exceptional specific demand compared with similar assets, causing potential buyers to compete for it by offering cheap cash in exchange, according to ICMA’s website.

'Easy fix'

For all the complexity of the regulation’s impact, getting rid of it would be physically easy. All legislators would need to do is cross out a few lines in Article 7 of the level-one regulation, says Hill.

And there will be a review of the rule in 2019 before it even goes into effect, which will be 2020 if the regulation is endorsed this quarter.

Hill believes that regulators, including ESMA, understand the impacts CSDR could have now, and are sympathetic.

However, the second industry lobbyist says it is far from certain that the necessary changes will be made. In fact, he thinks it could easily be passed with much of Article 7 intact.

“It’s unrealistic to think that a regulation which has not entered into force will be completely reviewed,” the lobbyist says. There won’t be any data available for regulators and legislators to review in 2019, he adds.

Hill adds: “One senior banker told us ‘this is bigger than all the other post-crisis rules put together, when it comes to liquidity – and we thought we’d got rid of it’.”

And people have been complaining about liquidity in recent years…