Cash CDOs: Balance Sheet, Arbitrage, and How They Really Work
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 17 gets into what many consider the most important innovation in structured credit: the collateralized debt obligation. CDOs are asset-backed securities where the collateral is a pool of debt instruments. We’re deep in Part 3 now, and this is where all the concepts from previous chapters come together.
Culp acknowledges upfront that a few chapters can’t possibly do justice to CDOs. He’s right. But he provides a solid framework for understanding how they fit into the broader themes of the book.
The CDO Universe
Moody’s estimated rated CDO volume at over $90 billion in 2004, across more than 200 deals. The first rated CDO was Drexel Burnham Lambert’s Long Run Bond deal in 1988, developed by none other than Michael Milken.
CDOs can be categorized several ways:
By collateral type:
- Collateralized loan obligations (CLOs): backed by loans
- Collateralized bond obligations (CBOs): backed by bonds
- Structured finance CDOs (resecuritizations): backed by other ABSs, other CDOs, mortgage-backed securities, trust preferred stock
Starting in 1997, CDOs began including structured debt and ABSs in their collateral pools. This was the beginning of resecuritizations, which would become a huge part of the market.
By economic motivation:
- Balance sheet CDOs: the asset owner wants to get rid of the assets for risk management or funding reasons
- Arbitrage CDOs: a collateral manager selects assets to try to generate trading profits
In 2004, arbitrage CDOs accounted for 92% of all activity.
Balance Sheet CDOs
A bank has a loan portfolio. It wants to reduce credit risk and free up regulatory capital. So it selects 100+ loans (usually $1 billion and up), sells them to an SPE, and the SPE issues securities backed by those loans.
The Rose Funding transaction by NatWest in 1996 is the model. A $5 billion structure based on 200 loans, about 15-20% of NatWest’s loan portfolio. Classes included a senior revolver and senior fixed note with investors in 17 countries. Rose Funding reportedly freed up about $400 million in regulatory capital for NatWest.
Key features of balance sheet CDOs:
No independent collateral manager. The originator selects the assets, possibly with help from a structuring agent.
The originator retains the residual/equity tranche. This is crucial. It works like a deductible. Without it, investors would worry that the bank cherry-picked its portfolio and dumped the worst loans into the CDO. The equity retention forces the bank to keep skin in the game.
Customized credit risk transfer. The senior-sub structure works like synthetic credit reinsurance. Different tranches serve different loss layers. Different investor groups can buy exposure to exactly the risk they want.
Funded credit protection. Unlike insurance or plain vanilla credit derivatives, the money is already in the structure. If the underlying loans default, cash is available immediately. No risk of counterparty nonpayment.
Monetization of assets. The bank gets cash for its loans. Balance sheet shrinks. Debt capacity increases. Risk and funding objectives are integrated.
Reduced adverse selection costs. Complex, opaque credit assets get placed into a self-contained structure that’s transparent and credit-enhanced. This can reduce the “lemons discount” those assets would otherwise carry.
The Regulatory Capital Arbitrage
This one is really important. Under the original Basel Accord, banks had to hold 8% capital against face value of commercial loans. Consider a $500 million loan portfolio earning LIBOR+135 with funding cost of LIBOR+35. Net interest income is $5 million. Capital charge is $40 million. ROE: 12.5%.
Now securitize it. The bank retains 2% equity ($10 million). Under Basel, the capital charge is 100% of the retained liability. So the capital charge drops to $10 million. If senior SPE expenses are 35 bps, the excess spread is 65 bps, or $3.25 million. New ROE: 32.5%.
Even though the bank earns less net income and has a higher percentage capital charge, the ROE is much higher because the equity at risk is dramatically lower. This was a powerful motivation for CDO activity.
Rating Considerations
Rating agencies focus on:
Collateral requirements: Credit information on each asset, concentration and diversity, industry diversity, proportion of high-risk assets, weighted average rating factor (WARF). S&P and Fitch tend to run individual obligors through their own models. Moody’s relies more on industry diversity scoring.
Coverage tests: Two main ones.
The overcollateralization (O/C) test requires that the ratio of collateral principal to total principal on a given tranche and all more senior tranches stays above a minimum. For Class B notes in a three-class CDO, the O/C ratio equals par value of the portfolio divided by principal on Class A plus Class B notes.
The interest coverage (I/C) test requires that scheduled interest on collateral assets (plus any allocated credit enhancement) divided by scheduled interest on a given tranche and all senior tranches stays above a threshold.
Both tests are extremely sensitive to how credit enhancements are allocated in the waterfalls. A good structuring agent is the best way to make sure these tests work on an ongoing basis.
Arbitrage CDOs
Different beast entirely. A collateral manager selects and buys debt assets, often from the open market. The goal is yield arbitrage. The collateral manager tries to buy assets whose collective yield exceeds the cost of the CDO liabilities plus fees.
The key metric is the funding gap: the difference between asset portfolio yield and CDO liability yield (plus senior costs). The bigger the funding gap, the higher the return on equity.
Key differences from balance sheet CDOs:
The residual tranche doesn’t have to be retained by anyone inside the structure. In an arbitrage CDO, moral hazard is managed through other credit enhancements like holdback or mandatory CCA deposits rather than equity retention.
Most arbitrage CDOs are cash flow CDOs (income from scheduled P&I on the collateral) rather than market value CDOs (income from asset liquidation). Market value CDOs were disappearing even when Culp wrote the book.
Portfolios can be static (fixed at inception), lightly managed (manager can sell before defaults but otherwise passive), or actively managed (manager has broad trading discretion). Active management has implications:
- Ramp-up period: Time to acquire the original assets. More complex portfolios take longer.
- Reinvestment period: Principal on existing collateral buys new collateral instead of paying down notes. Securities are often issued with maturities a year or so beyond this period.
- Amortization period: After reinvestment, income gets diverted into cash reserves for eventual bullet repayment of principal.
Multiple maturities in the liabilities can extend the effective reinvestment horizon. If half the securities mature in three years and half in five, only half the portfolio shifts to amortization mode in year two.
Active management affects fees (performance-based compensation for the collateral manager, sometimes from the equity tranche), accounting (QSPE exemption from FIN46R depends on passivity), and disclosure requirements.
Typical arbitrage CDOs hold 50+ securities totaling $100 million and up. Collateral managers specialize by asset type (high-yield debt, Brady bonds, emerging market debt, sovereign debt). They’re forever chasing yields and trying to maintain that funding gap.
Cash CDOs as Whole Capital Structure Products
Culp makes an important conceptual point. Cash CDOs are called “whole capital structure” products because the SPE’s entire capital structure gets sold to third-party investors (except maybe the equity tranche retained by the originator in balance sheet deals).
The SPE has created financial capital representing an exhaustive set of claims on all the capital assets on its balance sheet. This ties directly back to Chapter 1’s framework about the relationship between real assets and financial capital.
My Take
CDOs get a bad reputation because of what happened in 2007-2008. And there were genuine problems with how they were structured, rated, and sold during the housing bubble. But the underlying technology is sound.
Balance sheet CDOs solve a real problem: banks need to manage credit risk and optimize capital allocation. Arbitrage CDOs serve a genuine function too, though they require more scrutiny because the motivations are purely profit-driven.
What Culp does well here is showing that CDOs aren’t some exotic mystery. They’re just securitization applied to debt portfolios. The same principles of tranching, subordination, credit enhancement, and liquidity support that we’ve been building up through Part 3 all apply here.
The next chapter covers synthetic CDOs, where things get even more interesting.