Synthetic CDOs: From BISTRO to Single-Tranche Bespoke Deals
Book: Structured Finance and Insurance: The ART of Managing Capital and Risk Author: Christopher L. Culp Publisher: Wiley Finance, 2006 ISBN: 978-0-471-70631-1
Chapter 18 is the final chapter we’re covering in this batch from Part 3. Synthetic CDOs. If cash CDOs are securitization applied to debt, synthetic CDOs skip the securitization entirely and use credit derivatives instead. No assets actually change hands.
SCDOs emerged in the late 1990s and quickly took over a huge chunk of the CDO market. By 2004, the market looked completely different from just a few years earlier. Culp takes us through both generations of products.
First-Generation SCDOs
Fully Funded Balance Sheet SCDOs
In a cash CDO, you sell loans for cash. In a synthetic CDO, you keep the loans but buy credit protection using a credit default swap.
Here’s the structure: The originator (usually a bank) defines a reference portfolio. An SPE issues notes for the full portfolio amount. The proceeds get invested in high-quality collateral (Treasuries or repos) held by a trustee. This collateral is pledged to the originator as a performance guarantee on a CDS that the SPE has sold to the originator.
So the SPE is selling credit protection. The originator pays CDS premiums. If defaults happen on the reference portfolio, the collateral gets liquidated to cover CDS payments. Note holders get whatever is left.
The collateral serves double duty. It backs both the notes and the CDS. A swap dealer typically handles the ALM (asset-liability management) through an asset swap, converting interest on the collateral into scheduled P&I on the notes.
Fully Funded Arbitrage SCDOs
Same basic idea, but the SPE works with a collateral manager to construct the reference portfolio as a yield enhancement opportunity. The counterparties selling CDS protection to the SPE don’t have to be the originators. They could be secondary market purchasers or just CDS dealers. The equity tranche doesn’t need to be retained by anyone inside the structure because moral hazard is managed differently with derivatives than with true asset sales.
Partially Funded SCDOs and BISTRO
Here’s where things got really interesting. Few synthetic CDOs were ever fully funded. Selling the entire capital structure was too hard, too expensive, and too time-consuming.
JPMorgan’s BISTRO (Broad Index Secured Trust Offering) in 1997 pioneered the partially funded structure. The idea: fund only part of the reference portfolio with securities. Make the funded piece high enough quality that the most senior notes are AAA. Then call the unfunded remainder the “super-senior” piece.
For a portfolio of investment-grade assets, the AAA slice of the loss distribution usually attaches at the 15-20% cumulative loss point. That means 80-85% of the portfolio is super-senior. And because even the funded AAA tranche has to be wiped out before the super-senior piece takes any losses, the super-senior piece is essentially “better than AAA.”
The price for super-senior credit protection is dirt cheap. Investment-grade default rates are around 25-30 basis points. Super-senior protection generally costs below 15 basis points.
The BankAustria Example
Amadeus Funding, December 1998. BankAustria was merging with Creditanstalt and wanted to securitize $1.2 billion in ABSs.
Structure:
- Junior first-loss tranche: $24 million (BankAustria retained this)
- AA senior tranche: $48 million excess of $24 million (five-year maturity, 3M-LIBOR+87.5)
- Super-senior: $1.128 billion excess of $72 million (hedged with a CDS from Citibank)
Result: BankAustria got $1.2 billion in credit protection with a 2% deductible for under 100 basis points all-in. Estimated regulatory capital reduction: from $120 million to $26 million.
Regulatory Capital Mechanics
For a bank that synthetically securitizes $500 million in loans:
CDS collateralized with Treasuries: Capital charge is just the retained equity (1% = $5 million). Treasuries have 0% risk weight. Total: $5 million. But investors get a low base rate.
CDS with an OECD bank as fronting protection provider: The OECD bank writes credit protection to the originator, then buys identical protection from the SPE. Capital charge is the equity ($5 million) plus 8% of $500 million at 20% risk weight ($8 million) = $13 million total. But the SPE can hold higher-yielding AAA assets as collateral because it’s pledging to the OECD bank, not the originator. Better base rate for investors.
The fronting bank structure isn’t as good for capital as the Treasury-backed structure, but it’s better for note holders. The OECD bank also provides external discipline on collateral quality because it’ll apply its own haircuts and policies.
Why Synthetic Structures Became So Popular
Culp lists several advantages:
Operational simplicity. No need to deal with restrictions on loan sales, consent requirements from individual borrowers, or differences in local jurisdictions for true sale treatment.
Design flexibility. CDS documentation is easier than loan documentation. You can combine different loans and exposures freely. Single-name CDS markets are liquid and growing.
Selective risk transfer. Cash CDOs transfer the whole asset. Synthetic CDOs let you peel away just the credit component. If you have some advantage in bearing currency or political risk on your loans, you can keep those while transferring just the credit risk.
Lower cost of capital. The super-senior piece gives access to better-than-AAA financing. The credit market has no pricing category above AAA, but the SCDO structure effectively creates one. The “AAAA” rate is the price of super-senior protection.
Easier documentation. CDS master agreements are far more standardized than loan documentation. The 2000-2002 credit events actually accelerated standardization rather than killing the market.
Second-Generation SCDOs
Single-Tranche SCDOs (Bespoke CDOs)
By 2004, these accounted for virtually all new SCDO issuance. The process is reversed: an investor approaches an issuer and says “I want exposure to a specific mezzanine tranche of a credit portfolio.”
The issuer then constructs a reference portfolio (50-100 investment-grade credits with high industry diversity), synthesizes the exposure using CDSs, and issues a single class of securities corresponding to the tranche the investor wants. The issuer retains the other tranches and dynamically hedges them.
This is huge because it eliminates the need to find investors for every tranche of the deal. The time from first inquiry to closing can be very short. When you have to place multiple classes, documentation gets complicated, underwriting takes forever, and deals drag. With a single-tranche SCDO, you just need one willing investor.
The note issued is essentially a credit-linked note. And it doesn’t even need an SPE. The originating bank can issue it directly.
CDO Squared
A CDO of CDOs. You take pieces of existing CDOs (especially those hard-to-sell mezzanine and equity tranches), repackage them, and issue new securities. This lets collateral managers clear out their inventory of unsold junior exposures.
In Europe, CDO-squared deals were populated almost entirely with single-tranche SCDO securities. A CDO manager buys securities from 100 or so STSCDO transactions, holds them as collateral, and issues new securities against the pool.
ST-CDO-squared is when the CDO-squared itself is also a single-tranche deal. This gives enormous flexibility to tailor attachment points across multiple reference portfolios. But it also gets very complex to analyze. The effective subordination depends on the attachment point and rating of the underlying STSCDOs, the correlation between the CDSs that underlie the combined pool, and the overlap in reference names across constituent deals.
Master CDOs
A CDO backed by a pool of ABSs plus a basket of STSCDOs. This emerged in Europe because tight spreads on mezzanine ABS issues were squeezing funding gaps. Adding higher-yielding STSCDO tranches to the ABS pool widened the gap and made equity tranches more attractive.
This created a feedback loop: demand for master CDOs drove demand for STSCDOs, which banks could quickly construct using single-name CDSs.
The Insurance Question
Culp ends with a fascinating discussion about whether synthetic CDOs are creeping into insurance territory.
Credit derivatives clearly aren’t insurance. No insurable interest required. Events of default may not correspond to economic losses. And cash CDOs are safe too since they’re actual asset sales.
But synthetic resecuritizations push the boundary. When you write CDSs on securitized products issued by bankruptcy-remote SPEs, the events of default have to be defined differently. Market participants have defined them as: the reference asset is irrevocably written down, the reference asset is downgraded, or the securitized product is restructured.
As one expert quoted by Culp observed, these events “move significantly closer to ’loss’” compared to standard CDS triggers. The line between derivatives and insurance gets thinner.
The lesson: despite increasing convergence between insurance and capital markets, regulation hasn’t kept pace. Users of these products need to navigate carefully.
My Take
Reading this chapter in 2026, you can’t help thinking about what happened two years after the book was published. The synthetic CDO market, especially CDO-squareds and structures backed by subprime mortgage securities, was at the center of the 2007-2008 financial crisis.
But here’s the thing. Culp’s analysis is actually quite clear-eyed about the risks. He talks about the complexity of ST-CDO-squareds and the difficulty of analyzing effective subordination levels. He discusses the thin line between derivatives and insurance. He notes that the market was evolving faster than regulation.
The technology itself isn’t the problem. Synthetic CDOs are genuinely useful tools for credit risk management and capital optimization. The problems came from misuse: applying these structures to subprime mortgages with faulty loss assumptions, assigning AAA ratings to tranches that didn’t deserve them, and creating chains of resecuritization so long that nobody could trace the underlying risks.
Culp provided the analytical framework to understand these structures. Whether the market participants applied that framework honestly is a different question.