A CDO, or collateralized debt obligation, is a security backed by a large pool of debts, such as corporate loans, mortgages, auto loans, credit card balances, or corporate bonds, that is divided into layers of different risk and sold to investors. Instead of owning the underlying loans, an investor buys into a specific layer of the pool and receives payments according to where that layer sits in a fixed order of priority. CDOs became notorious for their role in the 2008 financial crisis, but the structure itself is still widely used today, especially in the form of collateralized loan obligations.
What Sits Inside a CDO
Every CDO begins with a pool of debt. That pool can hold corporate loans, residential and commercial mortgages, auto loans, credit card receivables, corporate bonds, or a mix. Diversification is the whole idea: if borrowers in one sector stop paying, income from other sectors is supposed to cover the shortfall. Pools typically run from several hundred million dollars to well over a billion.
The pool doesn’t stay on the balance sheet of the bank that assembles it. The assets are transferred into a standalone legal entity called a special purpose vehicle, or SPV. The SPV owns the debts and issues the securities backed by them. If the sponsoring bank later fails, its creditors can’t reach the assets inside the SPV. Most CDO SPVs are structured to fit an exemption under the Investment Company Act of 1940, usually by limiting ownership to “qualified purchasers” who hold at least $5 million in investments, so the vehicle doesn’t have to register as an investment company.
How the Tranches Work
Once the pool is in place, the SPV issues securities in layers. Each layer, or tranche, has its own credit rating, its own interest rate, and its own place in line for payments and losses. A typical CDO has three broad tiers.
The senior tranche is first in line for payments and last to absorb losses. Rating agencies frequently assign it a triple-A rating. Across a large sample of collateralized loan obligations, roughly 71% of the total dollar value of issued securities carried a triple-A rating even though the underlying loans averaged a B+ credit quality. The mezzanine tranche sits in the middle, usually rated somewhere between AA and BB, and pays a higher interest rate to compensate for greater risk. The equity tranche is last in line for payments and first to absorb losses. About 98% of CLOs include an unrated equity tranche, and it typically makes up around 9% of the deal’s total value.
This layering is what makes a CDO distinctive. A single pool of B-rated corporate loans can generate a large block of triple-A securities at the top, a smaller slice of mid-grade securities in the middle, and a thin equity cushion at the bottom that soaks up the first losses. Equity investors absorb defaults before anyone above them feels pain, which is why the senior tranche can earn a high rating even though the average loan in the pool is rated much lower.
Attachment and Detachment Points
Each tranche is defined by two numbers. The attachment point is the percentage of total pool losses at which the tranche starts taking hits. The detachment point is the loss level at which the tranche is wiped out entirely. An equity tranche with a 0% attachment and a 3% detachment absorbs the first 3% of losses on the pool. A senior tranche attaching at 22% wouldn’t lose a dollar until cumulative defaults exceed 22% of the collateral. These boundaries are set at the outset and dictate exactly how risk is divided.
Credit Enhancement
The tranche structure itself is a form of credit protection called subordination: losses flow upward through the junior layers before reaching the senior ones. Most CDOs add a second layer of protection called overcollateralization, where the face value of the loans in the pool exceeds the total value of the securities issued against them. If a pool holds $2 billion in loans but only issues $1.6 billion in bonds, that extra $400 million cushion can absorb defaults before any tranche takes a loss. Subordination and overcollateralization together are what allow the senior tranche to carry a high rating when the underlying debts are riskier.
How Investors Get Paid
Cash flows through a CDO in a strict top-down sequence known as the waterfall. Borrowers in the pool make their interest and principal payments, and that money enters the SPV. Administrative expenses and trustee fees come off the top. The senior tranche gets paid in full next, then the mezzanine layers, and finally the equity tranche receives whatever is left. Equity investors earn the highest potential returns precisely because they collect only after everyone else has been paid. This sequence repeats every payment cycle for the life of the deal and is governed by a contract called the indenture agreement.
Coverage Tests
CDOs include built-in tripwires called coverage tests that can override the normal waterfall. The two main tests are the overcollateralization test, which checks whether the collateral pool is still large enough relative to the outstanding debt, and the interest coverage test, which checks whether the pool’s income is enough to cover interest owed to each tranche. When a coverage test fails, cash that would normally flow down to junior investors is redirected upward to pay down senior tranche principal until the test is satisfied again. These mechanisms protect senior investors by forcing the deal to deleverage when the collateral deteriorates.
Common Types of CDOs
Not every CDO is built the same way. The differences come down to what sits inside the pool and how the deal gets its exposure to that debt.
- Cash CDOs. The SPV buys and owns the underlying loans or bonds outright. Income comes from borrowers making their scheduled payments. This is the most straightforward version.
- Synthetic CDOs. The SPV owns no loans. Instead, it uses credit default swaps to take on the economic risk of a reference portfolio. Synthetic CDOs can be created without anyone originating a single new loan, which is how they amplified risk during the financial crisis.
- Hybrid CDOs. These combine cash assets with synthetic exposure in a single deal.
- Collateralized loan obligations, or CLOs. A subset of cash CDOs backed specifically by leveraged corporate loans. CLOs dominate today’s market, with issuance exceeding $400 billion in 2024.
- Collateralized bond obligations, or CBOs. Built from pools of corporate bonds, typically high-yield.
The pre-crisis market also produced CDO-squared deals, in which the underlying collateral was itself made up of tranches from other CDOs. Stacking structured products on top of structured products made the risk nearly impossible to analyze, and these instruments were among the first to collapse in 2007 and 2008.
The Main Risks
CDOs carry several risks that are easy to underestimate, partly because the structure is designed to look safer than it is.
Default Correlation
The premise of a CDO is diversification: defaults in one part of the pool are offset by performing loans elsewhere. That works when defaults are independent. When borrowers default together because of a common shock like a recession or a housing crash, the math breaks. Research on CDO pricing shows that as correlation among borrowers increases, expected losses shift dramatically from the equity tranche to the senior tranche. In one modeling exercise, expected losses on the most senior layer jumped from 0% at zero correlation to over 43% at high correlation. That is essentially what happened in 2008.
Liquidity Risk
CDO tranches don’t trade on exchanges. The secondary market is thin, and during periods of stress it can vanish entirely. Even in the most liquid corner of the market, standardized index tranches, research from the European Central Bank found that prices include a sizeable liquidity premium driven by wide bid-ask spreads. Bespoke CDOs trade even less. If you need to sell a CDO tranche during a downturn, you may not find a buyer at any reasonable price.
Complexity and Model Risk
CDO valuations depend on statistical models that estimate default probabilities, recovery rates, and correlations among hundreds of underlying borrowers. Small changes in assumptions can produce large swings in a tranche’s estimated value. Many investors in the pre-crisis era did not independently model the deals they bought and instead relied on credit ratings as a shortcut. That reliance proved catastrophic.
Role in the 2008 Financial Crisis
CDOs were not a sideshow in the financial crisis. They were a central transmission mechanism that turned a housing downturn into a global catastrophe.
In the early 2000s, CDO creators became the dominant buyers of the lowest-rated tranches of mortgage-backed securities. By snapping up the BBB slices other investors avoided, CDOs created a new source of demand that kept the mortgage pipeline flowing. That demand required a steady supply of subprime and Alt-A mortgages, and lenders responded by originating riskier and riskier loans. CDO issuance more than doubled every year during this period.
Synthetic CDOs magnified the damage. Because they used credit default swaps rather than actual loans, multiple synthetic CDOs could reference the same pool of mortgages. A single $50 million tranche of mortgage bonds could generate hundreds of millions in synthetic exposure. AIG alone wrote $79 billion in credit default swap protection on super-senior CDO tranches backed primarily by subprime mortgages. When those mortgages defaulted, AIG couldn’t pay its obligations, and the federal government stepped in with a bailout.
The rating agencies compounded the problem. Triple-A ratings on senior CDO tranches encouraged pension funds, insurance companies, and banks worldwide to load up on what they believed were safe investments. When reality caught up, the downgrades were staggering: 91% of triple-A rated tranches from 2007 vintage subprime CDOs were eventually downgraded to junk, and some triple-A tranches lost more than 90% of their value.
Who Can Actually Buy a CDO
CDOs are not available to ordinary retail investors. They are sold through private placements under Rule 144A of the Securities Act of 1933, which limits initial purchases to qualified institutional buyers: institutions that own and invest at least $100 million in securities. In practice that means insurance companies, pension funds, endowments, large banks, and hedge funds. Registered broker-dealers face a lower threshold of $10 million.
The SPV itself is typically structured to require that all investors qualify as “qualified purchasers” under the Investment Company Act, meaning individuals who own at least $5 million in investments or institutions with at least $25 million. For individual accredited investors who might get indirect exposure through a fund, the SEC sets the bar at a net worth above $1 million (excluding a primary residence) or income above $200,000 individually, or $300,000 with a spouse or partner, in each of the prior two years.
How the Rules Changed After 2008
The Dodd-Frank Act of 2010 and the rulemaking that followed reshaped the CDO market in several ways.
Federal law now requires the sponsor of a securitization to retain at least 5% of the credit risk of the assets in the deal and forbids the sponsor from hedging that retained risk away. The idea is simple: if the people creating the deal keep skin in the game, they’ll be more careful about what goes into the pool.
The Volcker Rule generally prohibits banking entities from acquiring or retaining ownership interests in covered funds, which include most CDO vehicles structured under the Section 3(c)(1) or 3(c)(7) exemptions from the Investment Company Act. The rule reaches both direct equity stakes and synthetic positions that mimic equity exposure. Narrow exceptions exist for market-making, underwriting, and risk-mitigating hedging, but the broad effect has been to push banks away from proprietary CDO investing.
The SEC’s Regulation AB II requires issuers of many asset-backed securities to provide standardized loan-level data in a machine-readable format. For residential mortgage-backed securities, the required disclosures include the original loan amount, interest rate, borrower credit score, property type, loan-to-value ratio, and debt-to-income ratio for every loan in the pool. The goal is to give investors enough raw data to run their own analysis instead of relying only on credit ratings, the shortcut that failed so badly the last time around.