CSDR – mandatory buy-ins
COGESI meeting, November 26th 2014
Overview
 What is a buy-in?
 Problems executing buy-ins
 CSDR Mandatory buy-ins
 SFTs and partial exemption
 The likely impact of mandatory buy-ins
 Damage limitation: ICMA and industry recommendations
 Raising industry awareness and interaction with regulators
Key take-aways
 CSDR is settlement regulation with major trading impacts, which is in itself a
problem
 Mandatory buy-ins pose a significant threat to European bond market liquidity
 Mandatory buy-ins will discourage lending of securities and fragment the
European repo market
 Mandatory buy-ins will most likely have the counterproductive impact of reducing
settlement efficiency
 While it may no longer be possible to reverse the regulation, it is critical that
users of the secondary bond and repo markets work with ESMA, the EC, the
ECB, local regulatory authorities, local central banks, and DMOs to ensure that is
implemented in a way that causes the least disruption and damage to market
liquidity and efficiency
What is a buy-in?
 A buy-in is a remedy available to the purchaser of a security in the event that the
seller of the purchaser fails to make good delivery on intended settlement date.
 The failed-to purchaser (the ‘disappointed counterparty’) has the right to purchase
the failing securities from another counterparty (normally executed by an
appointed principal intermediary, known as the ‘buy-in agent’), for guaranteed
delivery, to replace the original failing purchase.
 An differential in price between the original failing purchase and the buy-in is
settled between the two counterparties.
 This ensures that both counterparties, from an economic perspective, are in the
same position they would have been had the original trade settled as intended,
 In practice, the failing counterparty will often be disadvantaged since the buy-in
will leave them with a long position that they will need to flatten in the market, at a
market price that will, in almost all instances, be below the buy-in price.
 The less liquid the bond, the greater the buy-in price/market price differential is
likely to be.
 Transactions under ICMA rules allow for a buy-in remedy.
Buy-in flow diagram
A’s sale to B fails
A
B
B issues a buy-in against A
A
Price difference
B
Buy-in
Agent
Market
Buy-in chains and ‘pass-ons’
A’s sale to B fails, causing a chain of fails
A
B
C
D
D issues a buy-in against C, and the buy-in is passed along the chain back to A
A
Pass-on
Px Difference
B
Pass-on
Buy-in
C
D
Px Difference
Px Difference
Market
Buy-in
Agent
Problems with executing buy-ins
 Tend to be in less liquid bonds, which makes them more difficult to find
 Buy-in prices can often be very far from ‘fair market value’, creating market
distortions
 A counterparty being bought-in has market risk until they flatten their position. If
there is a delay in communicating that the buy-in has been executed this will
expose them to unknown market risk
 Bought-in securities still may not settle
 It may be difficult to find buy-in agents
 Buy-ins can cause relationship issues
 Most investors understand that to get market pricing in, and market exposure to,
illiquid securities, they may need to accept that securities will not settle on ISD
CSDR Mandatory Buy-ins
 Regulation on improving securities settlement in the European Union and on
central securities depositories (CSDR) aims to improve safety and efficiency of
securities settlement in Europe
 Article 7, which deals with measures to prevent settlement fails (better known as
‘settlement discipline’), provides for:
CSDs to establish systems to monitor fails
CSDs to provide a penalty mechanism which will serve as a deterrent for
settlement fails
A mandatory buy-in process to be initiated where a transaction is still failing 4 days after
intended settlement date (ISD) – this has scope to be increased to 7 days, depending on
liquidity of the security being bought in
 Securities financing transactions (SFTs) to have a partial exemption from mandatory
buy-ins
Challenges of implementing mandatory buy-ins
 Buy-ins currently occur at the trading level. CSDR provides that this should occur
at the settlement level. How can this disconnect be reconciled?
Can CSDs differentiate between failing transaction types (important for exempt SFTs)?
Can CSDs identify why a trade is failing (i.e. should there even be a buy-in)?
Can CSDs identify fail-chains and know who should be bought-in to settle the chain?
What happens with fail-chains across different CSDs (including those outside of EEA)?
How and when are buy-ins communicated at the trading level (since this increases
market risk)?
 Central clearing counterparties (CCPs) are exempt. How does this impact buy-in
chains?
 What should be the calibration for the extension periods (4-7 days) for different
securities (MiFID II?)
 What is the impact of exempting some SFTs and not all?
SFTs and partial exemption ?
 CSDR Level 1 text suggests that some SFTs may be exempt from mandatory
buy-ins “where the timeframe of these operations is sufficiently short and renders
the buy-in ineffective”. This is interpreted as meaning near-legs of SFTs where the
far-leg falls before the earliest practicable settlement date for a related buy-in.
Thus the near-legs of short-dated repos would be exempt, but not the near-legs of
term repos.
 The text is vague about whether far-legs should be exempt or not. Logically there
is no reason to exempt far-legs, regardless of the duration of the SFT. But in the
event that far-legs are exempt, a GMRA mini-close-out could still provide some
risk mitigation in a mandatory buy-in regime.
Current remedies for failing SFTs (under GMRA/GMSLA)
 The Global Master Repurchase Agreement, which is the standard legal framework
for the European repo markets, and the Global Master Securities Lending
Agreement (GMSLA), provide remedies for the disappointed counterparty in the
event of a failed settlement.
 In the event of default (i.e. a fail), the contract allows for the disappointed
counterparty to initiate a ‘mini-close-out’. This effectively closes the transaction.
 Where the fail is on the near-leg of a repo/loan, the disappointed counterparty can
claim from the failing counterparty any interest that would have been accrued had
the repo settled, up until the close-out date.
 Where the fail is on the far-leg of a repo/loan, the disappointed counterparty can
claim from the failing counterparty any costs arising from replacing the securities.
Economically, this is the equivalent of a buy-in.
 These remedies apply regardless of the term of the repo/loan, or the underlying
security or asset type.
Current remedies for failing repos: pros and cons
Pros:
 Lenders (repo-ers) are not discouraged from lending, since they are not exposed
to any unquantifiable buy-in risk. Given the economics of securities lending, this is
critical to support market liquidity.
 Lenders (repo-ers) are protected in the event of a borrower (reverser) failing to
return securities at the end of a repo, in which case they can issue the equivalent
of a buy-in against the failing counterparty.
 A buy-in would be the wrong remedy for a failing repo start-leg, since the
disappointed counterparty is only looking to borrow the security, not to buy it
Cons:
 It is not possible to pass-on buy-ins in a fails chain that includes the near-leg of a
repo.
 Market-makers who borrow (reverse) securities to facilitate sales are unable to
hedge their buy-in exposure. However, in a market environment where buy-ins
are infrequent, this risk is relatively low. It could, however, become significant
under a mandatory buy-in regime.
CSDR partial exemption: the worst of all scenarios
 Partial-exemption will fracture liquidity in SFTs, creating a two-tier market with
different demand-supply skews:
Lender of securities will only want to lend for short-dates (exempt) to avoid
unquantifiable buy-in risk in the case of failing on their delivery of the near–leg of the
trade
Borrowers (such as market-makers) will have a preference to borrow securities for
term (non-exempt) to hedge their buy-in risk
 If CSDs are responsible for initiating or managing buy-ins, they will need to be
able to differentiate between SFTs and outright trades, near-legs and far-legs of
SFTs, and the term of the SFT (exempt or non-exempt)
Partial exemption: the worst of all scenarios (continued)
 Failing near-legs of exempt SFTs will still break any potential buy-in chain, leading
to multiple buy-ins.
 Repo desks will face greater buy-in risk from managing term mismatches, which
will deter their liquidity provision to the market.
 CCPs may need to separate exempt and non-exempt SFTs for different netting
treatments.
 Dealers and other market participants will need to manage exempt transactions
(short-dated SFTs) separately from non-exempt transactions (cash trades and
term-SFTs).
o NB: it is assumed that far-legs of all SFTs (regardless of term) would be in scope,
but this is not clear from the text. If the intention is to partially exempt the far-legs
of some SFTs, this will cause even more complexity and asymmetry.
Term structure of the European repo markets
25%
20%
15%
repo
reverse repo
10%
5%
0%
Source: ICMA-ERC Repo Survey June 2014
What will be the likely impact of mandatory buy-ins
 Increased risk and cost to market-makers, who will either only show offers in
securities they hold, or will widen spreads to reflect risk
 Increased administrative and legal stress, as well as market risk, as number of
buy-ins to manage increases exponentially
 Market disruption caused by multiple buy-ins, particularly in less liquid securities
 Reduced lending of securities where SFTs are in scope of buy-ins
 A reduction in settlement efficiency and increased fails as lending pool of
securities reduces
 Ever more buy-ins as an unvirtuous circle of settlement inefficiency takes hold
Damage limitation: ICMA and industry recommendations
 There are already initiatives in place to improve settlement efficiency in Europe
(Target2-Securities, ICSD ‘bridge’ enhancement, COGESI workstreams, etc.):
these should be successfully implemented first
 A well designed cash penalty/compensation scheme could improve settlement
efficiency would have a less disruptive impact than buy-ins
 Buy-in should be executed at the trading level, not the settlement level
 Buy-in chains should be initiated at the end of the chain (last counterparty being
failed to), and the buy-in passed along to the initial failing counterparty
 CCPs should be included in buy-in chains
 Extension periods (4-7 days) should be consistent with liquidity calibrations of
MiFID II
 The near-legs of SFTs, as much as possible, should be out of scope of
mandatory buy-ins (with a minimum 3mth term cut-off for exemption)
A solution looking for a problem
Bond Market Settlement Efficiency
100.0%
98.0%
96.0%
94.0%
92.0%
90.0%
88.0%
86.0%
84.0%
82.0%
ISD+0
ISD+1
ISD+2
ISD+3
ISD+4
ISD+5
EM
Source: ICMA-ERC Settlement Efficiency survey, 2014
Gvt
ISD+6
Corp
ISD+7
ISD+8
ISD+9
ISD+10
Later
A solution looking for a problem
Repo Market Settlement Efficiency
100.0%
98.0%
96.0%
94.0%
92.0%
90.0%
88.0%
86.0%
84.0%
ISD+0
ISD+1
ISD+2
ISD+3
ISD+4
ISD+5
EM
Source: ICMA-ERC Settlement Efficiency survey, 2014
Gvt
ISD+6
Corp
ISD+7
ISD+8
ISD+9
ISD+10
Later
Raising industry awareness and interaction with regulators
To date:
 CSDR CP and DP consultation (May 2014)
 ERC discussion and position paper (June 2014)
 AFME workshop and ICMA hosted industry meetings (June/July 2014)
 2-day ICMA-AFME industry workshop to prepare for ESMA roundtable
(September 2014)
 ESMA industry roundtable on settlement discipline (September 2014)
 ESMA meeting with ICMA and ISLA on mandatory buy-ins and SFTs (September
2014)
Ongoing:
 Continue to raise awareness of negative impact of mandatory buy-ins among
market participants and national regulatory bodies
 ESMA consultation on settlement discipline expected later in 2014