Credit Enhancement

Subordination creates tranches with loss priority: senior tranches protected by subordinate tranches. Sizing based on expected loss and tail risk analysis.

Example CLO structure: $400M assets backing $300M AAA (75%), $50M AA (12.5%), $30M A (7.5%), $20M BBB (5%) = $400M total. Subordination to AAA = 25% ($100M of BB/equity below AAA). AAA can withstand 25% collateral loss before experiencing principal impairment.

Rating agency subordination requirements vary by asset class: CLO AAA: 30-40% subordination required (reflecting ~12-15% base case default rate on B-rated loans, 40-50% recovery, stress scenarios). Prime RMBS AAA: 2-5% subordination (low default rates on prime mortgages). Subprime RMBS AAA: 10-25% subordination (higher default expectations, though 2005-2007 vintage proved insufficient). Credit card ABS AAA: 10-15% subordination plus excess spread.

Subordination erodes as losses occur. If $40M of CLO assets default with 50% recovery ($20M loss), the $20M equity tranche absorbs $20M loss and is wiped out. Remaining subordination to AAA now 20% instead of 25%.

Excess spread = (Portfolio Yield) - (Bond Coupons + Servicing Fees). Acts as ongoing credit enhancement by absorbing current-period losses.

Example: Credit card ABS with $1B receivables yielding 18%, bondholders receiving 4%, servicer getting 2%. Gross excess spread = 18% - 6% = 12%. This 12% absorbs charge-offs before subordination is impacted. If annual charge-offs run 8%, net excess spread = 4% flows to equity. If charge-offs spike to 15%, excess spread turns negative and subordination begins eroding.

Excess spread is cumulative protection over life of deal. Deal with 10% annual excess spread over 5 years provides 50% cumulative enhancement (simplistically). This is why credit card ABS can have lower subordination than CLOs—ongoing excess spread provides continuous replenishment.

Excess spread accounts often required: Rather than paying excess spread to equity monthly, many deals require accumulating it in reserve account up to target level (5-10% of bonds). This converts flow protection into stock protection, creating dual enhancement.

Critical difference from subordination: Excess spread is dynamic and can turn negative. Subordination is static until losses hit it. Both are essential—excess spread for expected losses, subordination for unexpected losses.

Reserve accounts hold cash to cover temporary shortfalls in interest or principal payments. Funded at closing or built up from excess spread.

Types of reserve accounts: (1) Spread account—accumulates excess spread up to target (3-10% of bonds). Covers temporary charge-off spikes. (2) Liquidity reserve—covers timing mismatches between collections and payments. Typical in auto ABS (1-2% of bonds). (3) Cash collateral account—required by rating agencies for AAA rating in some deals. (4) Prefunded reserve—established at closing from sale proceeds.

Reserve account mechanics: Account releases cash to cover shortfalls in interest payments, credit losses exceeding excess spread, or fee payments. If losses persist, account depletes and subordination takes over. Account typically replenishes from future excess spread if available.

Example: Credit card ABS has $50M spread account (5% of $1B bonds). Month 1: charge-offs $12M, excess spread $10M. Account covers $2M shortfall, balance drops to $48M. Month 2: charge-offs $8M, excess spread $10M. $2M flows back to account, balance rises to $50M.

Reserve accounts failed in 2008 when losses exceeded expectations for sustained periods. Many RMBS deals depleted reserves within 12-18 months, then subordination eroded rapidly. Reserve sizing is critical—too small and they provide false comfort; too large and they're capital inefficient.

Over-collateralization = (Collateral Value) > (Outstanding Debt). Creates asset cushion protecting bondholders.

Two types of OC: (1) Hard OC—actual excess collateral value over debt at all times. Example: $420M assets backing $400M debt = 5% hard OC. (2) Soft OC / OC tests—covenant requiring minimum asset-to-debt ratio. If violated, triggers cash flow diversion or deleveraging. CLOs use OC tests extensively.

OC test mechanics in CLOs: AAA OC test requires collateral ≥ 125% of AAA par (typically). If collateral falls to 123% due to defaults, test fails and excess cash diverts to pay down AAA until 125% restored. This is deleveraging enhancement—deal self-heals by reducing debt to restore ratios.

OC vs subordination distinction: OC measures asset cushion. Subordination measures loss priority. Example: Deal has $400M assets, $300M AAA, $100M BB. Subordination to AAA = 25%. OC = $100M / $300M = 33%. Both protect AAA but via different mechanisms. OC absorbs asset depreciation; subordination absorbs realized losses.

Hard OC requirement in auto ABS: Many auto ABS require maintaining hard OC of 1-3% throughout life of deal. If OC falls below minimum (due to charge-offs exceeding excess spread), triggers early amortization to protect senior bonds.

External credit support from guarantors, insurers, or letters of credit. Less common post-2008 due to monoline insurer failures.

Historical forms: (1) Monoline insurance—MBIA, Ambac, FGIC guaranteed bond payments. Prevalent pre-2008. Collapsed during crisis when insurers couldn't meet obligations. (2) Bank letters of credit—liquidity support from banks. Also failed when banks pulled support in 2008. (3) Government guarantees—GSE guarantees (Fannie/Freddie) for conforming mortgages. These held up through 2008 due to government backing.

Current limited use: Post-2008, market relies primarily on structural enhancement (subordination, OC, excess spread) rather than third-party guarantees. Rating agencies assign minimal value to third-party enhancement unless counterparty is sovereign-backed or has fortress balance sheet.

Counterparty risk: Third-party enhancement is only as good as guarantor's credit. When MBIA and Ambac were downgraded from AAA to junk in 2008-2009, securities they guaranteed were also downgraded, even though underlying collateral performance hadn't changed. This demonstrated the fragility of guarantee-based enhancement.

Current preference: Structural enhancement creates true risk transfer and doesn't depend on any single counterparty's solvency. This is why CLOs (structural enhancement) performed better through 2008 than RMBS (often relied on monoline insurance).

Multiple enhancement layers absorb losses in sequence, creating defense in depth:

First-loss protection: Excess spread absorbs ongoing expected losses in real-time. Example: 8% excess spread covers 8% annual charge-offs without touching other enhancement.

Second-loss protection: Reserve accounts cover temporary spikes above excess spread. Example: charge-offs spike to 12% for 3 months, reserve account covers 4% differential while excess spread covers base 8%.

Third-loss protection: Equity or first-loss tranche absorbs realized losses after excess spread and reserves exhausted. Example: sustained 12% charge-offs for 12 months consumes $40M equity completely.

Fourth-loss protection: Subordinated debt tranches (BB, B) absorb next layer of losses. Example: after equity wiped out, next $30M of losses hit BB tranche.

Senior protection: AAA/AA/A tranches only impacted after all subordinate enhancement exhausted. Example: only after $70M of losses ($40M equity + $30M BB) do AA tranches experience impairment.

This layering explains rating distribution: AAA requires all layers intact and stressed to 99.9th percentile. BBB requires first 2-3 layers and stressed to 95th percentile. BB requires only excess spread and equity, stressed to 80th percentile.

Effective analysis requires understanding which layers are intact, depleted, or under stress, and how much additional losses each can absorb before next layer is breached.