Credit analysis

Credit analysis in ABF answers one question: will you get paid back? That question breaks down into three parts. Can the originator stay in business long enough to service the loans? Will the borrowers on those loans perform? Does the structure protect you if things go wrong?

You need to assess all three. A strong collateral pool from a weak originator is a problem. Great structure can’t save you from a fundamentally flawed asset class. And excellent originator and collateral quality can still leave you exposed if the structure is poorly designed.

The three pillars

Credit analysis for ABF rests on three pillars, and you should evaluate them in this order:

Originator credit assessment: Can this company survive a downturn? Will they be around to service the loans and honor their obligations?
Collateral credit analysis: What will the underlying borrowers actually do? How much will you lose?
Structural credit analysis: If losses exceed expectations, does the structure protect your position?

Order matters. Originator risk can override collateral quality. If the servicer goes bankrupt and servicing transfers chaotically, even a pristine portfolio will suffer. Start with the originator, then move to collateral, then structure.

Originator credit assessment

Even in “non-recourse” deals, originator credit matters. Here’s why:

Servicing continuity risk. If the originator services the loans (as most do), their operational continuity directly affects collections. A distressed servicer may cut staff, lose experienced collectors, or fail to pursue delinquent accounts appropriately.

Repurchase and indemnity obligations. Most facilities require the originator to repurchase loans that breach representations and warranties. If the originator is insolvent, those repurchase obligations are worthless.

Data integrity. Your entire credit analysis relies on data the originator provides. A struggling originator may have let systems deteriorate or may be tempted to manipulate reporting.

Key metrics to evaluate

Track these originator-level metrics:

Metric	Typical Threshold	What It Tells You
Tangible Net Worth	$5M-25M+ (varies by facility size)	Cushion to absorb losses, honor repurchase obligations
Liquidity	3-6 months operating expenses	Can they fund operations during a market dislocation?
Leverage (Debt/Equity)	<3-4x	How much financial stress can they absorb?
Profitability	Positive unit economics	Is the business model sustainable?

Underwriting quality assessment

The best predictor of future portfolio performance is historical underwriting discipline. Look for:

Guideline exception rates. What percentage of loans fall outside stated underwriting criteria? Exception rates above 10-15% suggest guidelines aren’t really guidelines. Ask for exception data by approval authority.

Vintage consistency. Does performance vary dramatically by origination period? Inconsistent vintages often indicate underwriting drift or policy changes you should understand.

Credit box changes. Has the originator expanded their credit box recently? Loans originated under a new, looser policy have no performance history. Treat them with extra scrutiny.

Worked example: originator scorecard

Here’s how you might score an originator across key dimensions:

Dimension	Score (1-5)	Notes
Financial Condition	3	TNW of $12M vs. $10M covenant; liquidity tight at 4 months
Underwriting Quality	4	Exception rate 8%; consistent vintages; no recent credit box changes
Operational Capability	4	Experienced team; adequate systems; clean audit
Management	3	CEO has 10-year track record but CFO is new
Overall	3.5	Acceptable but monitor financial condition closely

A score of 3 or above is generally acceptable for most facilities. Below 3 requires either enhanced structural protection or a pass on the deal.

Collateral credit analysis

Collateral analysis answers: what will borrowers actually do, and how much will you lose when some of them don’t pay?

Building base case assumptions

Don’t pick assumptions out of thin air. Derive them from data.

CDR (Constant Default Rate) comes from historical delinquency and roll rates. If 2% of current loans roll to 60+ DPD monthly, and 80% of 60+ DPD loans eventually default, you can back into an annualized CDR.

CPR (Constant Prepayment Rate) comes from actual prepayment speeds. Calculate monthly prepayment rate by cohort, annualize it, and understand what’s driving it (refinancing, payoffs, partial prepays).

Severity (loss given default) comes from realized loss data. If the originator has workout history, use it. If not, benchmark against public ABS in the same asset class. Consumer unsecured typically sees 80-90% severity; auto is 40-50%; real estate secured can be 20-40%.

Example: A consumer unsecured portfolio shows 4.5% annualized default rate based on roll rate analysis, prepayment speed of 25% CPR, and historical severity of 85%. Your base case cumulative net loss is approximately 3.8% of original balance over the portfolio’s life.

Static pool analysis

Static pool analysis tracks the performance of loans originated in a specific period. You’re looking for:

Loss curve timing. Are losses front-loaded (hitting in months 6-18) or back-loaded (hitting in months 18-36)? This affects how quickly your enhancement erodes and whether revolving structures have time to build up reserves.

Curve shape consistency. Do all vintages follow roughly the same curve shape, just scaled up or down? Or do some vintages behave fundamentally differently? Inconsistent shapes suggest underwriting or macroeconomic changes you need to understand.

Plateau levels. Where do cumulative losses plateau? For consumer unsecured, you might see 5-8% CNL at plateau for prime borrowers, 10-15% for near-prime, and 20%+ for subprime.

Stratification that matters

Run stratifications on the fields that drive credit performance:

Credit score/tier distribution. What percentage is prime (720+), near-prime (660-719), subprime (<660)? How does performance vary by tier?

Geographic concentration. Are you overweight in any single state? California and Texas concentration over 15% requires additional scrutiny of regional economic sensitivity.

Loan size and term distribution. Larger loans and longer terms typically carry more risk. A consumer portfolio with 5-year terms will behave differently than one with 3-year terms.

Benchmarking

Compare your portfolio against external data:

Public ABS performance. Pull 10-D filings from EDGAR for rated deals in your asset class. These show monthly performance data you can benchmark against.

Rating agency studies. Moody’s, S&P, and Fitch publish default and transition studies that show historical performance by rating category and asset class.

Comparable private deals. If you’ve done similar deals before, how does this portfolio compare? If colleagues or competitors have shared data, use it (appropriately).

Red flags

Watch for these warning signs in collateral analysis:

Untested credit boxes: No performance history for a meaningful segment of the portfolio
Rapid underwriting changes: Guidelines loosened materially in the past 6-12 months
Concentration spikes: Single obligor, geography, or product type exceeding normal thresholds
Data quality issues: Missing fields, inconsistent formatting, unexplained outliers

Structural credit analysis

Structure transforms the raw credit risk of the collateral into positions with different risk profiles. Your analysis determines whether the structure adequately protects your position.

How structure transforms risk

Credit enhancement absorbs losses before they reach senior tranches. A deal with 10% subordination can withstand 10% cumulative losses before the senior note takes any loss.

Waterfall mechanics determine cash flow priority. In a sequential structure, senior gets paid principal first; in pro-rata, all tranches share. Sequential is safer for seniors, pro-rata returns cash faster to junior.

Triggers divert cash flow or accelerate paydown when performance deteriorates. A deal that traps excess spread when delinquencies exceed 3% has an automatic de-risking mechanism.

Evaluating structural protections

For any position you’re evaluating, understand these structural features:

OC (Overcollateralization). What’s the target OC? What’s the floor? How much cushion exists between current OC and the floor? If target is 108% and floor is 105%, you have 3 points of cushion before cash starts getting trapped.

Reserve accounts. Is there a spread account or cash reserve? How is it funded? When can it be released?

Subordination. For rated deals, what’s below your position? How much loss would it take to reach your tranche?

Structural risks to stress

Model these risks explicitly:

Cash flow timing mismatch. If assets pay monthly but liabilities pay quarterly, where does the cash sit between payment dates? Is there liquidity to cover shortfalls?

Basis risk. If assets are fixed-rate and liabilities are floating (or vice versa), or if they’re on different indices (SOFR vs. Prime), you have basis risk. Size it.

Extension risk. If prepayments slow, WAL extends. Can you live with the deal extending 6-12 months beyond expected maturity?

Rating agency methodologies

Rating agencies take a different approach than you do as an investor. Understanding this helps you interpret their analysis and size enhancement appropriately.

How agencies approach credit

Rating agencies stress portfolios to failure levels consistent with their rating categories. For AAA, they apply stresses consistent with Great Depression-era defaults. For BBB, the stress is more moderate but still well above base case.

S&P uses an anchor approach: start with an asset-class-specific anchor rating based on collateral quality, then adjust for structural features (positive) and concentrations/weaknesses (negative).

Moody’s uses an expected loss model: calculate expected loss under stressed assumptions, compare to target loss rates by rating category, size enhancement to achieve coverage.

Fitch applies criteria-based floors and caps: certain structural features trigger automatic rating caps or require minimum enhancement levels regardless of collateral quality.

Your base case vs. agency stress

The difference between your base case and the agency’s stress case is significant:

Scenario	Typical Assumption
Your base case	Historical performance continues; 4% CNL
Agency BBB stress	Moderate recession; 8% CNL (2x base)
Agency AAA stress	Severe downturn; 12-16% CNL (3-4x base)

This delta explains why rated deals require so much more enhancement than your internal analysis might suggest. The agency is sizing to survive scenarios you’re not explicitly forecasting.

Expected loss calculation

Expected loss is the foundation of credit analysis. At its simplest:

EL = PD x LGD x EAD

Where:

PD = Probability of Default
LGD = Loss Given Default (severity)
EAD = Exposure at Default

For an amortizing portfolio, you need to account for declining balances over time.

Worked example: expected loss calculation

Consider a consumer unsecured portfolio:

Original balance: $100 million
Term: 36 months
Expected CDR: 5% annualized
Expected severity: 85%
Expected prepayment: 20% CPR

Step 1: Project pool balance over time

Using standard amortization plus prepayment assumptions, the pool balance declines. By month 18, roughly $55M remains; by month 36, the pool is essentially amortized.

Step 2: Calculate defaults by period

Apply 5% CDR to the outstanding balance each period. Total defaults over the life: approximately $9.5M.

Step 3: Apply severity

$9.5M defaults x 85% severity = $8.1M expected loss, or 8.1% of original balance.

Step 4: Sensitivity analysis

Scenario	CDR	Severity	Expected Loss
Base case	5.0%	85%	8.1%
CDR +50%	7.5%	85%	11.8%
Severity +10pts	5.0%	95%	9.0%
Combined stress	7.5%	95%	13.2%

Your enhancement needs to cover losses under your stress scenario, not just base case.

Credit enhancement sizing

Credit enhancement is the cushion that protects senior positions from losses. Sizing it correctly is the core output of structural credit analysis.

What enhancement covers

Enhancement serves three purposes:

Loss absorption: Absorb expected losses under stressed scenarios
Liquidity: Cover cash flow timing mismatches during disruption
Structural requirements: Meet rating agency or investor minimums

Sizing methodology

Work backward from your target outcome:

Determine the rating or risk level you need to achieve
Find the loss coverage required for that level
Size enhancement to provide that coverage

Rule of thumb for sizing:

A position: base case loss + buffer (1.25-1.5x)
BBB: 2-2.5x base case loss
AAA: 3-4x base case loss

Forms of enhancement

Enhancement comes in several forms, often layered:

Form	How It Works	Typical Sizing
Overcollateralization	Pool balance exceeds note balance	2-10%
Subordination	Junior tranches absorb losses first	5-25%
Reserve account	Cash set aside for losses	1-3%
Excess spread	Interest collected > interest paid	2-5% annually
Third-party guarantee	Insurance or guarantee wraps senior	Varies

Worked example: sizing enhancement for BBB

Using our consumer unsecured example:

Base case CNL: 8.1%
BBB multiple: 2.5x
Required loss coverage: 20.3%

To achieve BBB risk profile, you need 20.3% enhancement. This might come from:

15% subordination (junior notes)
3% OC target
2.3% annualized excess spread (trapped if performance triggers are hit)

If actual losses hit 15%, the subordination absorbs it. If losses hit 20%, subordination is exhausted and OC erosion begins, but BBB notes remain whole.

Stress scenario design

Stress testing isn’t prediction. You’re testing resilience: how bad can things get before your position takes losses?

Standard scenarios

Run at least four scenarios:

Base case: Current performance continues. Use historical CDR, CPR, and severity derived from actual data.

Moderate stress: Recession-like conditions. CDR increases 50-75%, CPR slows 20-30% (borrowers can’t refinance), severity increases 5-10 points.

Severe stress: 2008-level downturn. CDR doubles or more, CPR slows significantly, severity increases materially. For consumer credit, assume unemployment hits 10%+.

Break-even: Solve for the loss level that exactly exhausts your enhancement. This tells you how much cushion you have.

Asset-class-specific considerations

Different asset classes respond differently to macro stress:

Consumer unsecured: Highly correlated with unemployment. A 5-point increase in unemployment might increase CDR 50-100%.

Auto: Moderate unemployment sensitivity, but severity is driven by used car prices. In 2008, severity spiked as used car prices fell.

Real estate secured: Home price appreciation (HPA) and vacancy rates drive losses. A 20% HPA decline can double severity in non-QM portfolios.

Equipment: Industry-specific cyclicality. Construction equipment performs very differently from medical equipment in a downturn.

Scenario matrix: presenting results

Present stress results in a matrix that shows outcomes across scenarios:

Scenario	CDR	Severity	CNL	Coverage Ratio	Outcome
Base	5.0%	85%	8.1%	2.5x	Comfortable
Moderate	7.5%	90%	12.5%	1.6x	Adequate
Severe	10.0%	95%	17.0%	1.2x	Marginal
Break-even	12.5%	95%	20.3%	1.0x	Enhancement exhausted

This matrix tells you that under severe stress (not quite as bad as 2008), your BBB position still has a 1.2x coverage ratio. That’s acceptable, but not comfortable.

Watch list indicators

Early warning matters. By the time losses materialize, the damage is done. Track leading indicators that signal deterioration before it shows up in loss numbers.

Early warning metrics

Delinquency trend acceleration. Absolute delinquency level matters less than the rate of change. If 30+ DPD moved from 2.0% to 2.2% to 2.5% to 3.0% over four months, that acceleration is a red flag.

Roll rate deterioration. What percentage of 30-day delinquent loans roll to 60-day? If this rate increases from 40% to 50%, defaults are about to increase even if current DQ levels look stable.

Prepayment slowdown. Declining CPR often signals borrower distress. Healthy borrowers refinance; distressed borrowers can’t. A material prepayment slowdown may be the first sign of credit deterioration.

Originator warning signs

TNW declining toward covenant minimum
Liquidity dropping below 6 months of operating expenses
Origination volume ramping faster than infrastructure can support
Key management departures, especially in credit or servicing
Audit qualifications or accounting changes

Structural warning signs

Trigger thresholds approaching (within 50 basis points)
OC cushion compressing (less than 100 basis points to floor)
Covenant headroom shrinking across multiple tests
Excess spread declining due to yield compression or loss creep

Escalation framework

Level	Indicators	Actions
Yellow	Single metric approaching threshold; trend concerning but not alarming	Enhanced monitoring; monthly data review; informal servicer check-in
Orange	Multiple warning signs; trend clearly negative; approaching trigger	Monthly review calls; formal servicer report; re-run stress scenarios; update IC
Red	Trigger breached or imminent; originator stress evident; structural protection eroding	Workout preparation; legal review of remedies; prepare for potential wind-down

What to do when indicators flash

Immediate actions (first 48 hours):

Verify the data is accurate (not a reporting error)
Contact the servicer for explanation
Pull the latest tape and run updated analytics

Medium-term (first week):

Re-run all stress scenarios with updated assumptions
Assess your exposure at various outcomes
Prepare memo for internal escalation

Escalation protocol:

Yellow: Inform portfolio manager; note in monthly report
Orange: Schedule review with credit committee; prepare action plan
Red: Emergency credit committee; legal on standby; document all communications

Putting it together

Credit analysis culminates in a recommendation. For a new deal, that’s an IC memo. For ongoing monitoring, it’s a surveillance report with a watch list status.

Summary credit view template

Your credit analysis should produce a concise summary:

Originator Assessment: [1-5 rating]

Key strengths: [specific points]
Key concerns: [specific points]
Recommendation: [acceptable / acceptable with conditions / pass]

Collateral Quality Assessment: [vs. benchmark]

Base case CNL: X% vs. benchmark of Y%
Key drivers: [credit tier, geography, seasoning]
Red flags: [if any]

Structural Adequacy Assessment:

Enhancement: X% vs. required Y% for target rating
Trigger headroom: X basis points to first trigger
Key structural risks: [if any]

Key Risks and Mitigants:

Recommendation: [Approve / Approve with conditions / Decline]

Minimum analysis before IC

Before presenting a deal to investment committee, complete this checklist:

Missing any of these creates risk that IC will ask questions you can’t answer. Complete the checklist before you present.