Structured Finance Case Studies — 2008 RMBS Collapse, CLO COVID Stress & Office CMBS Crisis
When does structured finance collapse and when does it hold? The full 2008 subprime RMBS/CDO collapse; why CLO AAA tranches survived COVID in 2020; the structural roots of the 2023 US office CMBS stress; and Korea's real estate PF ABS crisis.
30-Second Summary
Why the same structured finance technology collapses in some crises and holds in others.
2008 Global Loss
$2.7T+
RMBS/CDO collapse
2020 CLO AAA Loss
$0
Survived — waterfall worked
2023 Office CMBS DQ
8.1%
vs 2.5% pre-COVID
Korea PF Balance
130조
as of 2022
Estimated Loss by Case ($B)
2020 CLO AAA: no principal loss. 2008 RMBS figure per IMF estimate.
The answer converges to three factors: ① genuine quality of the underlying assets, ② accuracy of correlation assumptions, and ③ degree of information asymmetry. This chapter answers through four real-world cases.
2008 Subprime RMBS/CDO Collapse
What happens when you stamp AAA on the riskiest assets — anatomy of a $2.7T loss
Background: 2004–2007 Subprime Mortgage Boom
Subprime Mortgage Issuance
$1.3T
2006 peak
CDO Issuance (cumulative)
$520B
2004–2007
CDO Leverage
10~15×
vs equity
- →NINJA loans (No Income, No Job, No Assets): mortgages approved without verification. The logic: 'Rising home prices will take care of everything.'
- →Teaser rate structure: low initial rates (2–3%) for two years followed by sharp ARM resets. Borrowers who couldn't refinance defaulted automatically.
- →Rating agency model failure: Gaussian Copula model assumed home prices across the U.S. would never decline simultaneously — a scenario never stress-tested.
- →Originator incentive misalignment: Originate-to-Distribute model — mortgages originated and immediately sold, leaving no default risk on the issuer's balance sheet.
The Collapse: Step-by-Step Contagion
4 Structural Lessons
Correlation Assumption Failure
Models assigned near-zero probability to nationwide simultaneous home price declines. When every underlying asset deteriorated at once under stress, the entire tranche structure was rendered meaningless.
Information Asymmetry
Mortgage originators knew the true credit quality of borrowers. Final CDO investors could not know the real quality of the hundreds of mortgages inside their structured product. Originate-to-Distribute institutionalized the information imbalance.
Incentive Misalignment
Rating agencies were paid by issuers. Higher ratings generated higher fee revenue. The conflict of interest between the watchdog and the watched contaminated the entire system.
Excessive Leverage and Opacity
CDO-squared (CDOs of CDOs) pushed leverage to extremes. Even the final investor couldn't identify what was inside. Opacity amplified panic when the underlying assets began failing.
2020 COVID — Why CLO AAA Tranches Survived
How AAA tranches held principal even as leveraged loan prices crashed to 60 cents
Background: The March 2020 Market Shock
- →March 2020: COVID-19 pandemic → economic shutdown → fear of corporate earnings collapse.
- →Leveraged loan average price: 100 cents (January) → 60 cents (March). A 40% crash in just eight weeks.
- →CLO equity tranches: mark-to-market -40 to -70%. BB tranches: -25 to -30%.
- →CLO AAA tranches: only 1–2bps spread widening; zero principal loss.
2020 CLO Tranche Price Performance (Jan = 100)
Approximate based on market data. Source: JP Morgan CLO Research.
3 Reasons CLO AAA Survived
Equity Buffer — The Shock Absorber
A typical CLO equity tranche represents 8–12% of total assets. This means AAA principal is unaffected until asset values fall more than 8–12%. When leveraged loan prices dropped to 60 cents in March 2020 (a 40% decline), the equity + BB + BBB tranches absorbed the entire loss cascade. For losses to reach the AAA tranche, portfolio value would need to decline approximately 70%+ — implying individual corporate default rates of 30–40%, far beyond any plausible scenario.
OC Trigger Activation
CLO indentures contain built-in OC Tests: if the ratio of portfolio par value to tranche par value falls below a threshold, interest payments to junior tranches are suspended and redirected to repay AAA principal early. In 2020, many CLOs activated OC triggers, halting equity distributions and accelerating AAA principal repayment. This structural safety mechanism was absent from the 2008 RMBS CDOs.
Portfolio Diversification — Correlation in Practice
A typical CLO holds 150–250+ leveraged loans across different issuers. The 2020 COVID shock concentrated in airlines, leisure, and energy, but technology, healthcare, and consumer staples held relatively firm. Sector diversification kept overall portfolio default rates far below worst-case scenarios. The 2020 U.S. leveraged loan default rate was approximately 3–4% — versus the 30–40% required to impair AAA tranches.
The underlying assets of the 2008 CDO (subprime RMBS) collapsed simultaneously nationwide — correlation converged to 1. The underlying assets of CLOs (leveraged loans) experienced differentiated shocks by sector and company — correlation stayed in the 0.3–0.5 range. The tranche structure was identical. What you put inside determined the outcome.
2023 U.S. Office CMBS Crisis
How remote work rewrote commercial real estate — Brookfield DTLA and Pimco losses dissected
Background: Structural Shift in Office Demand
U.S. Office CMBS DQ Rate
8.1%
2024 (vs 2.5% pre-COVID)
SF & Chicago Vacancy
30%+
2023–2024 high
Office Value Decline
-30~50%
from 2022 peak, select markets
Case 3-A: Brookfield DTLA Fund — $784M CMBS Default
- →Brookfield Asset Management's Los Angeles office fund. Six premium Downtown LA office buildings including Bank of America Plaza and Gas Company Tower.
- →February 2023: $784M CMBS loan maturity extension negotiations fail → default declared. An unusual move for a major institutional sponsor.
- →Strategic default: Brookfield had the financial capacity to extend, but judged office values unrecoverable — chose intentional default over further capital commitment.
- →WeWork bankruptcy impact (November 2023): WeWork was a major tenant in many office buildings; WeWork departures → occupancy collapse → further deterioration of CMBS underlying assets.
- →Pimco: disclosed hundreds of millions in losses from funds holding office CMBS in 2023, notably related to the $1.7B Columbia Property Trust CMBS default.
Office CMBS Crisis Contagion Pathway
| Stage | Description | Impact |
|---|---|---|
| 1 | Structural office demand decline | Hybrid work permanently settled → 80% office utilization → weakened lease renewal leverage |
| 2 | NOI (net operating income) compression | Rising vacancy + rent concessions → NOI down 30–50% → DSCR deterioration |
| 3 | Asset value collapse | Lower NOI × higher cap rates (rate hikes) → double compression pushes values down 40–60% |
| 4 | CMBS maturity hits refinancing wall | Loan maturity with depressed asset values → refinancing impossible → default |
| 5 | Junior tranche losses begin | Special servicer takes over → property disposition → losses absorbed bottom-up through tranches |
| 6 | Market spillover | Office CMBS price declines → mark-to-market losses at insurers and pensions → CRE lending standards tighten |
- •2008 financial crisis: credit recovers → real estate recovers. Demand came back.
- •2020 COVID: economy reopens → offices come back. And they did, partially.
- •2023 office crisis: demand itself permanently changed. Hybrid work is not a temporary trend — it is an official policy adopted by S&P 500 CEOs. Class B and C offices may be unrecoverable without conversion to residential, logistics, or other uses.
Korea Real Estate PF ABS Crisis
How one Legoland default froze a ₩130T market — an in-depth analysis
Understanding Korea Real Estate PF ABS Structure
Project Developer
Develops apartment or mixed-use projects. Needs financing after land purchase and permits secured.
Special Purpose Company (SPC)
Project-specific entity created by developer. SPC issues ABCP or ABS to raise funds.
Securities Firm Credit Enhancement
Securities firms provide purchase commitments (put-backs) or payment guarantees on ABCP. This credit enhancement is what maintains A1–A2 ratings.
ABCP Investors (Short-term)
MMFs, corporate cash, individual investors buy 30–90-day ABCP. Repeated rollovers bridge long-term PF with short-term funding.
Core Structural Vulnerability
ABCP rollover failure → securities firm purchase commitment triggered → rapid short-term debt surge at securities firms → capital pressure → systemic credit freeze. The weakest link was rollover viability, and the Legoland incident broke it.
Legoland Incident (September 2022): The Collapse of a Market Axiom
Crisis Scale and Regulatory Response
Key Scale Indicators
- •PF loan balance: approximately ₩130T (end-2022)
- •ABCP and ABSTB outstanding: approximately ₩40–50T
- •Securities firm PF exposure: top 10 firms combined ₩30T+
- •PF loan delinquency rate: 3.5% in 2024 (vs 0.5% in 2021)
- •Distressed PF projects: hundreds flagged for workout as of 2024
Regulatory Response Package
- •₩20T Bond Market Stabilization Fund: liquidity support through CP, corporate bond, and ABCP purchases
- •PF soft-landing program: FSC-led, separating viable from non-viable projects for targeted support
- •KAMCO (Korea Asset Management Corp) PF normalization fund: acquires and resolves distressed PF assets
- •Bank provisioning expansion: strengthened loan loss reserve requirements for PF-related exposure
- •Securities firm soundness standards: higher capital requirements against PF exposure ratios
Korea PF ABS vs U.S. CMBS: Structural Comparison
| Category | Korea Real Estate PF ABS | U.S. CMBS |
|---|---|---|
| Underlying Asset | Pre-completion development projects (PF) | Stabilized income-producing properties |
| Core Risk | Completion risk + pre-sale absorption risk | NOI sustainability + tenant risk |
| Credit Enhancement | Securities firm purchase/payment guarantees | Tranching structure itself |
| Tenor | Short-term (30–90-day ABCP, repeated rollovers) | 5–10 years (fixed long-term maturity) |
| Market Size | ₩130T+ (2022 PF loan balance) | $800B+ (CMBS outstanding) |
| Crisis Pathway | Securities firm capital pressure → systemic freeze | Asset value decline → institutional losses |
| Oversight | FSC/FSS-centered, disclosure standards limited | SEC disclosure, risk retention rules |
- 1Pre-sale absorption rate: Current vs. break-even absorption rate (BEP) — how much cushion remains?
- 2Remaining construction cost and contractor rating: How much cost remains? Is the contractor rated A- or above? (Contractor failure = construction suspension risk)
- 3Nature of credit enhancement: Purchase commitment vs. payment guarantee? Securities firm total PF exposure as a percentage of equity capital?
- 4Rollover risk: Days until next ABCP maturity? Probability of rollover failure under adverse market conditions?
- 5Loan conditions: Current LTV, DSCR, ICR levels
- 6Pre-sale market environment: Local unsold inventory, competing nearby projects, housing price trends
- 7Policy risk: DSR regulations, mortgage lending rules, potential changes to permit conditions
Unified Lessons Across All Four Cases
The three decisive variables that separate structured finance that holds from structured finance that fails.
Underlying Asset Quality — Tranching is Not Magic
2008 CDO: subprime mortgages (NINJA loans) → underlying assets were already impaired. No matter how sophisticated the tranche structure, bad underlying assets destroy everything. 2020 CLO: collateralized corporate leveraged loans → relatively sound underlying assets. The same tranche structure worked. Lesson: before investing in any structured product, verify 'what is inside, and can it actually repay?'
Correlation Assumptions — Do All Assets Fail Simultaneously Under Stress?
2008: 'Nationwide home prices cannot fall simultaneously' → they did. Removing the correlation=1 scenario from the model changed everything. 2023 office: office vacancy rising simultaneously across markets → regional diversification became ineffective. CLO corporate portfolios, by contrast, experienced differentiated sector shocks, making diversification real.
Incentive Alignment & Information Asymmetry — Do Issuers, Investors, and Raters Share the Same Interest?
2008: mortgage originators sold-and-forgot (OTD), rating agencies were paid by issuers, investors had no visibility into the underlying pool. Korea PF: developers, securities firms, and ABCP investors all relied on the assumption that rollovers would continue indefinitely — nobody priced the real risk until Legoland broke the axiom. When information asymmetry and incentive misalignment materialize together, structured products collapse rapidly.
| Case | Asset Quality | Correlation | Incentive Alignment | Outcome |
|---|---|---|---|---|
| 2008 RMBS/CDO | ❌ Poor | ❌ Wrong assumptions | ❌ Misaligned | ❌ $2.7T loss |
| 2020 CLO AAA | ✅ Sound | ✅ Diversification held | ✅ Risk retained | ✅ Principal preserved |
| 2023 Office CMBS | ⚠️ Weakened | ⚠️ Co-deteriorating | ⚠️ Partial | ⚠️ Ongoing |
| Korea PF ABS | ⚠️ Pre-compl | ❌ Rollover-dependent | ❌ Info asymmetry | ❌ Systemic crisis |
FAQ
References
- [1] Financial Crisis Inquiry Commission (2011). The Financial Crisis Inquiry Report. U.S. Government Publishing Office.
- [2] IMF (2008). Global Financial Stability Report: Containing Systemic Risks and Restoring Financial Soundness.
- [3] JP Morgan CLO Research (2020). CLO Performance During COVID-19: Tranching Holds the Line.
- [4] S&P Global Ratings (2021). CLO Ratings Performance: COVID Stress Test Results.
- [5] Moody's Investors Service (2021). Annual Structured Finance Default Study.
- [6] Trepp (2024). U.S. Office CMBS Delinquency Report: Q1 2024.
- [7] Brookfield Asset Management (2023). DTLA Fund Investor Communications.
- [8] 금융위원회 (2022). 부동산 PF 시장 안정화 방안 발표 자료.
- [9] 한국은행 (2023). 금융안정보고서: 부동산 PF 관련 리스크 평가.
- [10] 강원도·감사원 (2023). 레고랜드 사태 관련 감사 결과 보고서.
- [11] BIS (2023). Structured Finance and Financial Stability: Lessons from Recent Stress Events.
Related Market Cases
The actual deal cases behind the events covered in this chapter.
Share this deal