Sovereign CDS Basis: Why Arbitrage Fails in Stress

Key Takeaways

• The sovereign CDS basis (CDS spread minus bond spread) should theoretically be zero; in practice funding costs, counterparty risk, and delivery optionality keep it persistently non-zero.
• During Italian stress in mid-2011, repo haircuts on Italian bonds jumped from 3–4% to 5–6%, making long-bond-plus-CDS carry uneconomic and driving the basis deeply negative for months.
• Investors treat basis trades as free lunches; they are carry trades against fat-tailed funding risk, basis positions get margin-called exactly when spreads widen most.
• The 2014 ISDA Credit Derivatives Definitions changed how sovereign CDS handles restructuring; pre- and post-2014 contracts are not fungible and cannot be netted as hedges.

What It Is

A CDS pays the buyer if the reference sovereign defaults. A government bond pays a yield above a risk-free benchmark partly as compensation for the same default risk. If you buy the bond and buy CDS protection, you have a nearly risk-free position. Its residual return should match the risk-free rate plus a small basis.

The basis is defined as:

basis = CDS spread - (bond yield - risk-free rate)

A positive basis means CDS trades wide relative to bonds. A negative basis means CDS trades tight relative to bonds, so the bond offers a higher spread than the CDS premium.

The Intuition

Textbook arbitrage says the basis should be zero. In reality it is not. Three frictions explain why:

• Funding cost. To hold the bond, the trader repos it and pays a haircut plus funding spread. The CDS is an unfunded instrument with a smaller cash outlay. When funding is expensive or haircuts jump, buying the bond and selling CDS becomes uneconomic. Basis turns negative.
• Counterparty risk. CDS protection is only as good as the seller. In 2008 and 2011, perceived dealer-balance-sheet risk made CDS buyers discount the protection, widening basis.
• Deliverability and cheapest-to-deliver. CDS settles against an auction of deliverable bonds. If the cheapest deliverable is much cheaper than the reference bond, CDS underprices bond default risk.

In sovereign crises, two forces pull in opposite directions. Flight-to-safety buyers crush bond yields tighter than the true credit deserves, driving basis positive. Funding stress and forced deleveraging prevent arbitrageurs from pressing the trade, so the basis persists.

How It Works

A negative-basis trade (bond spread wider than CDS):

1. Buy the sovereign bond at yield Y.
2. Fund it via repo at rate R plus a haircut.
3. Buy CDS protection on the same sovereign at premium C.
4. Lock in a return of roughly Y - R - C, which should be positive and stable if arbitrage works.

If the sovereign defaults, the CDS pays and the bond loss is offset. If the sovereign survives, the trade earns the basis.

A positive-basis trade (CDS spread wider than bond):

1. Short the bond (hard for sovereigns, often done via repo reverse).
2. Sell CDS protection.
3. Earn the CDS premium net of repo borrow cost.

Positive basis trades require shorting sovereign bonds, which is operationally harder than longing them. That asymmetry is part of why positive basis can persist.

Academic work (Bai-Collin-Dufresne 2011, ECB WP 1271) shows that during the 2008 crisis and the 2010-2012 eurozone stress, basis became substantially negative and stayed that way for months. Short-selling frictions explain persistent positive basis; funding frictions explain persistent negative basis.

Worked Example

Greece, 2010-2012. Greek sovereign CDS spreads jumped above 1,000 basis points in the May 2010 flight-to-safety episode. Greek bond yields also spiked but the cash bond became illiquid as dealers reduced balance sheets. For a stretch in 2011, the CDS priced a lower loss-given-default than the bond implied, giving a positive basis. Other peripheral sovereigns showed a different pattern. Italian bonds around mid-July 2011 saw repo haircuts jump from 3-4% to 5-6%, which made the carry on a long-bond-plus-CDS basis trade uneconomic. Basis went deeply negative and stayed there.

On 9 March 2012 the Greek PSI triggered a CDS credit event. The ISDA auction settled Greek CDS at a recovery of 21.5%, paying protection buyers 78.5 cents per dollar of notional. Basis trades positioned for the credit event resolved, but many basis positions in other peripherals continued to bleed through 2012 as funding remained stressed.

Counter-example for arbitrage limits. Through 2014-2019, with ECB QE compressing peripheral yields, the sovereign CDS basis in most eurozone names narrowed back toward zero. The basis is a mean-reverting series in calm markets and a wildly divergent series in stress.

Common Mistakes

1. Treating CDS basis as a free lunch. It is a carry trade against fat-tailed funding risk. The 2008 and 2011 experience shows basis positions held via short-dated repo get margin-called precisely when spreads widen, forcing unwinds at the worst time. Lehman Brothers' failure triggered forced CDS unwind across the street. Funding lines must be term-matched, not rolled daily.

2. Ignoring the CDS definitions reboot. The 2014 ISDA Credit Derivatives Definitions changed how sovereign CDS handles restructuring, redenomination, and government intervention. Pre-2014 contracts and post-2014 contracts are not fungible. Read the confirm.

3. Assuming cheapest-to-deliver is irrelevant for sovereigns. Sovereigns often have many bond series with different maturities, coupons, and jurisdictions. In a restructuring, the CDS auction selects deliverables that can trade at different prices than the reference bond. This basis to deliverables is a hidden risk in the trade.

4. Forgetting that CDS counterparty risk is real. In the run-up to 2008 and again during Italian stress in 2011, banks that had sold large sovereign CDS protection saw their own CDS spreads widen. If your protection is with a weak dealer, your CDS is not risk-free.

5. Backtesting only during calm periods. A pre-2007 backtest of peripheral-Europe CDS basis would show a small, stable, mean-reverting series. It would not warn you about 2011. The basis is highly non-stationary at regime breaks, and historical Sharpe ratios from calm-period samples overstate the live return.

Frequently Asked Questions

Q: What is the sovereign CDS basis in simple terms? It is the difference between what it costs to buy insurance against a country defaulting (the CDS spread) and the extra yield that country's bonds pay over a risk-free rate. If both priced default risk identically, the basis would be zero. In practice, funding costs and market frictions keep it non-zero and occasionally extreme.

Q: How does the sovereign CDS basis affect investment decisions? A negative basis (bonds cheap relative to CDS) can signal a long-bond, long-protection trade. A positive basis signals CDS is expensive relative to bonds. Both are carry trades against funding liquidity risk, positions put on via short-dated repo can be margin-called precisely when spreads widen most, locking in losses at the worst time.

Q: What is a real-world example of the sovereign CDS basis? During the 2010–2012 eurozone crisis, Italian and Spanish bond repo haircuts jumped sharply, making funded bond positions uneconomic relative to CDS. Basis turned deeply negative and stayed there for months. When Greece triggered its March 2012 CDS credit event, the ISDA auction settled at a recovery of 21.5 cents, paying protection buyers 78.5 cents per dollar of notional.

Q: How can investors use knowledge of the sovereign CDS basis? Use it as a measure of funding stress, not just credit risk. A rapidly widening negative basis in peripheral European or EM bonds signals that dealers are reducing balance sheet and market liquidity is deteriorating. Track repo haircut moves alongside the basis, they move together and give early warning before the wider credit market reacts.

Q: How is the sovereign CDS basis different from a corporate CDS basis? Sovereign bonds often have jurisdiction, governing-law, and delivery option complexities absent in corporate bonds. A sovereign CDS auction delivers bonds across multiple maturities with potentially different restructuring outcomes. These cheapest-to-deliver options create basis components unique to sovereign markets that corporate CDS trades do not face.

Sources

1. Fontana, A. and Scheicher, M. (2010). "An analysis of euro area sovereign CDS and their relation with government bonds." ECB Working Paper 1271. https://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp1271.pdf
2. Gyntelberg, J., Hoerdahl, P., Ters, K. and Urban, J. (2017). "Arbitrage costs and the persistent non-zero CDS-bond basis." BIS Working Paper 631. https://www.bis.org/publ/work631.pdf
3. Longstaff, F., Pan, J., Pedersen, L. and Singleton, K. (2011). "Sovereign CDS and Bond Pricing Dynamics in the Euro-area." NBER Working Paper 17586. https://www.nber.org/system/files/working_papers/w17586/w17586.pdf
4. Bai, J. and Collin-Dufresne, P. (2011). "The CDS-Bond Basis During the Financial Crisis of 2007-2009." http://lamfin.arizona.edu/fixi/creditmod/BaiCollinDufresne11.pdf

Disclaimer

This article is educational content only and is not financial advice. Nothing here is a recommendation to buy, sell, or hold any security. Consult a licensed advisor before making investment decisions.