Eligible Collateral, Haircuts, and the Economics of Margin

When a bank or fund manager posts collateral against a derivatives exposure — whether to a CCP as initial margin, or to a bilateral counterparty as variation margin — the type of collateral matters enormously. Cash, government bonds, agency debt, and equities are all treated differently. Understanding why requires understanding the concept of a haircut and the rationale for eligibility restrictions.

Eligible collateral is collateral that the receiving party — whether a CCP or a bilateral counterparty under a CSA — will accept as security for an obligation. Eligibility criteria are set based on two primary attributes: credit quality and liquidity.

Credit quality. Collateral must have a high probability of retaining its value even in a stressed market environment. A government bond issued by a G10 sovereign is considered the highest quality because the issuer has a very low probability of default and an excellent track record of honouring its obligations. Corporate bonds carry higher credit risk — both because the issuer is more likely to default and because corporate bond prices tend to fall precisely when financial markets are stressed, i.e. when the collateral is most likely to be needed. This correlation between the stressed environment that triggers the need to use collateral and the value of the collateral itself is a key concern in collateral eligibility design.

Liquidity. Even if a piece of collateral retains its credit value, it must be possible to sell it quickly if the holder needs to liquidate it after a counterparty default. On-the-run US Treasuries and UK gilts trade in enormous daily volumes and can be sold in seconds at prices very close to the mid-market. An illiquid structured note, by contrast, might take days or weeks to sell, and the sale price might be far below its nominal value due to the illiquidity discount.

Typical eligibility schedules at major CCPs like LCH include:

• Cash in major currencies (USD, EUR, GBP, JPY, CHF)
• G10 sovereign government bonds (US Treasuries, UK gilts, German Bunds, French OATs, Japanese JGBs, etc.)
• Agency debt (e.g., US agency bonds from Fannie Mae and Freddie Mac) — at a higher haircut than sovereign bonds
• Some high-quality supranational bonds (World Bank, EIB, ESM)

Corporate bonds, equities, structured products, and most non-G10 sovereign bonds are not eligible at major CCPs. This restriction concentrates demand for high-quality liquid assets (HQLAs) and is one reason why government bond markets are so central to the collateral ecosystem.

A haircut is a discount applied to the market value of a security when calculating its collateral value. If a security has a 5% haircut, £100 million face value of that security counts as only £95 million of collateral. The pledging party must therefore post more securities than the nominal margin requirement to cover the gap.

Haircuts are designed to cover two risks:

• Price volatility: the risk that the security's price falls between the time it is received as collateral and the time the CCP or counterparty needs to liquidate it. A security with high price volatility needs a higher haircut to ensure adequate coverage even in a stress event. A 30-year gilt has significantly more duration risk (and therefore price volatility) than a 2-year gilt, so it carries a higher haircut despite being equally creditworthy.
• Liquidity (close-out period): the time required to sell the collateral. The longer the assumed close-out period, the more price volatility can accumulate, and the higher the haircut needed to cover it. LCH's haircuts for SwapClear are calibrated to a 5-business-day close-out period for default management.

Illustrative haircut schedule for sovereign government bonds at a major CCP:

• 0–1 year maturity: 0.5%
• 1–5 year maturity: 2.0%
• 5–10 year maturity: 4.0%
• 10–30 year maturity: 8.0%
• 30+ year maturity: 12.0%

The haircut increases steeply with maturity because longer-dated bonds have higher duration — their price moves more for a given change in interest rates. The September 2022 UK gilt crisis demonstrated this vividly: 30-year gilt prices fell by over 20% in a matter of days, causing the haircuts on long-dated gilts to prove inadequate and triggering emergency intervention by the Bank of England.

Wrong-way risk (WWR) in collateral refers to the situation where the value of the collateral posted is positively correlated with the size of the exposure it is meant to cover — meaning the collateral is least valuable precisely when the exposure is largest.

The most straightforward example: a bank has a large credit exposure to a European sovereign (say, through a CDS sold on that sovereign's debt). Under the CSA, the sovereign or a closely related entity posts that sovereign's own bonds as collateral. If the sovereign's credit deteriorates — causing the CDS exposure to grow (as protection value increases) — the sovereign's bond prices will also fall, reducing the value of the collateral at exactly the moment the exposure is largest. This is specific wrong-way risk.

Most CSAs and CCP eligibility schedules address this with concentration limits and issuer restrictions. A CSA might specify that collateral cannot be issued by the pledging party itself, or by an entity with a high correlation to the pledging party. CCP rules typically restrict posting of bonds issued by entities that are clearing members of the same CCP, to avoid circular dependencies in the default management scenario.

When a pledging party has discretion over which eligible securities to post as collateral (within the eligibility schedule), it will rationally post the cheapest-to-deliver securities — those that meet the eligibility criteria but have the lowest opportunity cost for the pledging party to give up temporarily.

The cheapest-to-deliver is not necessarily the lowest-quality eligible asset. It is the asset that the pledging party values least, relative to its collateral value after haircut. A 10-year gilt that the pledging party needs for other purposes (e.g., it is the cheapest-to-deliver into a gilt futures contract) would not be posted — a different eligible bond that is less useful elsewhere would be preferred.

The substitution option — the right to replace posted collateral with different eligible securities — has economic value. A pledging party that can substitute collateral freely is better placed to manage its overall balance sheet efficiently: it can reclaim specific securities when needed and substitute them with less critical assets. This optionality is captured in the MVA (Margin Valuation Adjustment) component of derivative pricing — the cost of providing margin, accounting for the optionality in collateral selection and substitution over the life of the trade.