ABS CDO

An ABS CDO (or “CDO-squared”) is a securitization backed by tranches of other asset-backed securities—typically mortgage-backed securities or consumer loan ABSs—rather than by direct loans or bonds. By layering tranches from multiple secondary securities, ABS CDOs compounded leverage and correlation risk, making them especially vulnerable to deterioration in the underlying asset pools.

The leverage chain: borrower → pool → ABS → CDO

An ABS CDO exemplifies the multiplication of leverage inherent in structured finance. The chain begins with mortgages or consumer loans made by banks or mortgage brokers. These loans are then pooled into mortgage-backed securities (MBSs) or asset-backed securities (ABSs), which are themselves tranched: a senior tranche (absorbing little loss), a mezzanine tranche (absorbing moderate loss), and a subordinated or equity tranche (absorbing first losses).

An ABS CDO then takes mezzanine and junior tranches from multiple MBS and ABS pools and uses them as collateral for a new, second-level securitization. The CDO is also tranched—senior, mezzanine, and equity—creating a waterfall of losses. A decline in housing prices or consumer credit quality thus flows through multiple layers: borrowers default on the original loans, losses cascade through the MBS/ABS tranches, triggering defaults on the ABS CDO’s collateral, which then triggers losses on the CDO’s own tranches.

This structure magnified the housing bubble. During 2005–2007, as house prices rose and mortgage credit spreads compressed, the demand for AAA-rated securities seemed insatiable. Investment banks, desperate for collateral and fees, created ABS CDOs to supply it. A pool of subprime mortgages with 5–10% expected loss rates was sliced into tranches; the junior pieces were pooled into an ABS CDO whose senior tranche received a AAA rating and sold at a tight spread over Treasuries. The originating loan originators, lenders, and ABS issuers had no skin in the game—all risk was passed to the CDO buyer.

Correlation concentration and the false safety premise

The critical flaw was correlation assumption. An ABS CDO’s senior tranche is safe if the losses from the underlying ABS tranches are independent—if a loss in one mortgage pool does not predict losses in another. Rating agencies and pricing models implicitly assumed this independence, based on historical data in which mortgage pools in different geographies and vintages did not all default simultaneously.

But this assumption was catastrophically wrong. During the 2008 housing collapse, default rates surged in all mortgage pools at once. Borrowers across the country became underwater on their homes simultaneously; subprime delinquencies became highly correlated. The ABS CDO model assumed that if pool A suffered losses, pool B would be relatively safer—so the CDO’s junior tranches would absorb most of pool A’s loss, leaving the senior tranche intact. In reality, when the housing market seized, pools A, B, and C all experienced synchronized defaults, and senior ABS CDO tranches themselves suffered large losses.

The rating agencies, which had assigned AAA ratings to senior ABS CDO tranches, failed to stress-test for this scenario. Models relied on realized correlation from the past 30 years, when housing prices only rose. No model incorporated the possibility of a nationwide housing collapse, making the ratings worse than worthless—they were dangerously misleading.

Double and triple leverage

An ABS CDO investor bought risk that was already embedded in a secondary security. If an investor had purchased an MBS mezzanine tranche directly, they would absorb losses starting after the mortgage pool’s equity [tranche](/tranche/] lost its capital. An ABS CDO investor in a comparable tranche of the CDO bore risk after both the original ABS equity tranche AND the ABS CDO’s equity tranche were wiped out—but because correlation meant multiple ABS tranches defaulted together, the CDO’s own junior pieces evaporated quickly, leaving the “senior” CDO tranches exposed.

This double leverage made ABS CDOs a magnified bet on housing. Dealers, conscious of this, often hedged senior ABS CDO tranches through credit default swaps on the underlying pools, assuming this would protect them. But during the crisis, CDS spreads widened so sharply that the hedging was insufficient, and counterparties themselves became stressed, unable to pay out CDS claims.

Valuation and sales model

Dealers priced ABS CDOs using models that estimated the probability distribution of losses in the underlying ABS pools. They calibrated these models to historical credit spreads on the MBS and ABS tranches, then solved for the spread that compensated the CDO’s tranches for their expected losses. The models output a probability that the CDO’s senior tranche would suffer any loss at all; if that probability was very low (say, 0.01%), the senior tranche was assigned a AAA rating.

The model outputs were then typically backed by confidence intervals or stress tests—calculations showing loss rates under higher correlation or higher default assumptions. However, the stress tests were often far too narrow. A dealer might assume correlation increased from 20% to 40%, recompute the loss distribution, and find the senior tranche still AAA-rated. In reality, during the crisis, correlation spiked to 80%+ and loss rates were catastrophic, rendering the senior tranches nearly worthless.

Cascade of losses post-2008

When housing prices started to decline in 2006 and mortgage delinquencies spiked in 2007, ABS CDO investors and dealers realized their mistake. The underlying mortgage pools were performing far worse than historical norms suggested they should. The ABS tranches backed by subprime mortgages, which had been rated A or BBB, experienced losses. Those losses flowed through to the ABS CDOs. Senior ABS CDO tranches, previously thought to be nearly risk-free, were suddenly seen as junk.

By 2008, most ABS CDOs were marked far below par, and the secondary market for them virtually halted. Dealers holding inventory suffered enormous losses. Investors who had bought senior ABS CDO tranches as AAA-safe assets received ratings downgrades and realized losses of 50–100%. The reputational damage to rating agencies was severe; their models were revealed to be inadequate, and their conflicts of interest—agencies were paid by issuers—were exposed.

The broader lesson: leverage and correlation

ABS CDOs encapsulate a core lesson from the financial crisis: leverage compounded without corresponding risk reduction is fragile. Each layer of securitization promised to isolate and distribute risk, but it actually concentrated it through correlation. The illusion of diversification—that senior CDO tranches were safe because losses would be absorbed by junior pieces—collapsed when the underlying assets all deteriorated together.

Post-crisis, ABS CDO issuance fell to near zero. Regulatory capital-adequacy rules imposed punitive charges on dealer holdings of structured credit. Investors became skeptical of model-driven ratings and demanded much tighter spreads (higher compensation) for any complex securitization. The secondary market became hostile: buyers knew that dealers holding ABS CDO inventory were in distress and used that leverage to demand fire-sale pricing.

Today, ABS CDOs are rare. Most securitization of mortgages is done at the primary level (government-backed MBSs or plain commercial MBSs), and secondary-level securitization is confined to niche structures with clear, transparent collateral pools and robust credit enhancement. The appetite for double-tranched complexity has not returned.

See also