Methodology & Data Quality

This page documents the data sources, definitions, and quality considerations behind every chart and table in the Auto Loan and Auto Lease ABS Data Explorers. It is written for practitioners who need to evaluate whether the numbers meet their standard of care before citing them.

Sections tagged Auto Loan or Auto Lease apply to that explorer only. Everything else applies to both.

1. Data Source

Delinquency, repossession and credit-score figures are derived from SEC ABS-EE asset-level filings submitted to EDGAR under Regulation AB II, which securitization trusts are required to file monthly. Loss figures come from the trusts’ monthly Form 10-D servicer reports (section 4). We fetch, parse, deduplicate and aggregate these filings to produce the figures shown. Typical latency from a new SEC filing to the explorer is under 24 hours.

Auto Loan coverage Auto Loan

Approximately 100 fields per loan per reporting period.

• Issuers covered: 24 active auto-loan securitizers (Ally, Ameri­Credit, Bridgecrest, BMW, CarMax, Capital One Prime, Carvana (P), Carvana (N), Exeter, Fifth Third, Ford, GM Financial, Honda, Hyundai, Mercedes‑Benz, Nissan, Santander (SDART), Santander (DRIVE), Toyota, Volkswagen, World Omni, and CarMax Select, Exeter Select, World Omni Select). Carvana’s single shelf is shown as two programs — P (prime) and N (non-prime) — which carry distinct collateral and report on different days of the month; blending them would distort both. Retired trusts (California Republic, USAA) remain available for historical access.
• Time coverage: Issuance vintages begin in 2017, though individual loans in seasoned pools can extend back to ~2010.

Auto Lease coverage Auto Lease

Approximately 70 fields per lease per reporting period.

• Issuers covered (SEC-registered, public ABS-EE filers): BMW Vehicle Lease (BMWLT), Ford Credit (FCALT), GM Financial (GMALT), Mercedes-Benz (MBALT), Nissan (NALT), Porsche (PILOT), Volkswagen (VWALT), World Omni (WOALT).
• Excluded: Toyota Lease (TMCLT), American Honda (AHALT), and Hyundai (HALST) — no ongoing public ABS-EE (Rule 144A private placements; Hyundai filed a single public deal in 2021).

2. Active Pool & Delinquency

Both explorers report count-weighted, mutually exclusive DPD buckets: the 30–59 bucket, the 60+ bucket and the current (<30) bucket sum to the total active count and do not overlap.

• 60+ DPD rate: active assets 60 or more days past due, divided by total active assets.
• 30–59 DPD rate: active assets 30 to 59 days past due, same denominator.
• 30+ DPD rate: the two combined.
• Current rate: active assets less than 30 days past due.

Bucket boundaries

The site cuts every issuer at 30 days (30–59, 60+). Four servicers report their own investor tables the same way (Capital One, Hyundai, Toyota, Fifth Third); the other twenty-two report 31–60 / 61–90 / 91+, so for them the site’s buckets run a few percent above the figures in their monthly reports. The difference is the one-day band, not the data.

Credit scores are as reported

Issuers report the score they underwrote on. The captive finance companies generally use FICO^® Auto scores, which run to 900 and sit some 10–30 points above a classic FICO^® for the same borrower; the subprime and bank issuers report classic scores (300–850). Toyota, Honda and Fifth Third name the model in their filings; the rest say only “bureau” or “FICO”. Score bands are therefore comparable within a family of issuers and only approximately across it; commercial-bureau scores and no-score markers are excluded.

A note on weighting

Delinquency is count-weighted: every active loan or lease counts once, regardless of size, so the rate answers what share of borrowers are behind. Loss is balance-weighted: it is a ratio of dollars, so it answers what share of the pool’s money was lost. Both are ratios of sums rather than averages of deal-level rates, so a small deal with an extreme rate cannot distort the issuer figure.

The two conventions are standard for their respective metrics, but they are not the same question — on a book where large loans behave differently from small ones, delinquency and loss will diverge for reasons of construction alone. Each chart names its denominator on the y-axis.

What counts as “active”

Loan and lease filings report the end of an asset’s life differently, so the two explorers apply slightly different tests:

• Auto Loan a loan is active while it carries an outstanding balance (above a nominal $10).
• Auto Lease a lease is active while it carries a balance and has not been reported as terminated. Lease filings mark ended leases explicitly, and issuers differ in how they mark active ones, so the test relies on the termination status rather than on any single issuer’s convention.

Auto Loan Some issuers report placeholder days-past-due values for loans that have already been charged off; those loans are excluded from the delinquency buckets.

3. Reconciliation Against Issuer Filings Auto Lease

We validate the lease active-pool definition against issuers’ own monthly investor reports (Form 10-D). Checked against a sample deal’s report, the end-of-period pool count and securitization value reproduce exactly under our definition.

Where the lease numerator diverges from issuer reports

Even with matched denominators, our 30+ DPD count typically runs 10–20% above the corresponding figure in monthly investor reports. Two structural reasons:

• Payment extensions. When a lessee receives a payment extension during a collection period, the issuer treats the lease as cured for reporting purposes. Our ABS-EE snapshot still carries the lease at its pre-extension DPD until the following month’s filing.
• Charge-off-pending exclusions. Issuers typically remove leases flagged for default or repossession from the delinquency table and report them separately. Our snapshot still includes those leases in the 60+ bucket while they retain a balance.

Trends and relative comparisons across issuers remain meaningful. For deal-level absolute counts, consult the issuer’s Form 10-D.

4. Loss Metrics

Loss figures come from monthly 10-D servicer distribution reports (EDGAR), not from the ABS-EE loan tape. The servicer states each deal’s losses directly, which is the authoritative pool-level figure. The Losses tab offers two views of that data.

Cumulative Net Loss, by vintage or series

• Cumulative Net Loss Rate: cumulative net losses (charge‑offs net of recoveries, per the servicer) ÷ original pool balance, indexed by months since issuance. Note this denominator differs from the annualized measure below, which divides by the average outstanding balance.
• Months since issuance count from a deal’s first servicer report, whatever period that report covers: the first report is month 1 and each later report is dated by its collection month, so a missing report leaves a gap rather than shifting what follows. One rule for every issuer, so deals line up on the same axis. (Some servicers number their own reports differently; those numbers are not used.)
• View: selectable — By Vintage (pool‑weighted across deals of the same issuance year) or By Series (individual deals).

Older deals on the loss curves

• 2010–2016 deals: the trust’s own monthly servicer reports (Form 10‑D), read and numbered exactly as for later deals.
• Earlier deals, whose trusts stopped filing monthly reports after about a year: the sponsor’s static‑pool tables in its later prospectuses, which report each prior pool’s cumulative net loss month by month. The curve is that loss ÷ the pool’s initial balance, taken from the latest prospectus that printed it; where a deal also has servicer reports, the two agree. Months are counted from the pool’s cut‑off date, as the tables print them. AmeriCredit published its 2001–2010 tables on its own website; those are read from the archived copies.
• Not shown: Ally’s revolving trusts and whole‑portfolio vintages (neither is a single pool’s loss curve), and CarMax’s pre‑2010 tables, which print percentages without a stated base.
• A deal’s cumulative net loss occasionally falls from one month to the next; every month is shown as reported.
• Earlier issuers (listed separately in the issuer picker, with the years their data spans) no longer issue public auto ABS; their deals are shown by deal age only, from the same two sources.
• All originations (Ally, BMW): the lender’s whole auto-loan book by year of origination, as its prospectuses print it — securitized or not, so not comparable with a securitization’s pool.

Annualized Net Loss, by calendar month

CNL is cumulative and indexed by loan age, so it cannot be placed on a calendar axis: averaging it across an issuer’s deals in a given month produces a figure driven by pool age mix rather than credit. The annualized net loss rate is a point-in-time flow, so it can be aggregated per issuer over calendar time the way the delinquency rate is.

• Formula: 12 × net losses for the period ÷ average outstanding pool balance (the mean of beginning and ending balance), summed across an issuer’s deals in each collection month — the month the figures describe (a September servicer report covers August collections). Each month stands alone: the rate is not smoothed or averaged across months.
• Basis: the All pools / Seasoned (>6 mo) toggle mirrors the Delinquency tab. A young deal books almost no losses, so an issuer writing new deals looks better than it is on the all-pools basis.
• Below zero is a net gain, not a negative loss. The 10-D line item is literally “Principal Net Loss / (Gain)”: a month whose recoveries on previously charged-off receivables exceed its gross charge-offs books a gain, which the filings print in accounting parentheses. Prime shelves cross below zero routinely.

Because the two tabs read different filings, their as-of dates can differ. Both are dated by the month the figures describe, but servicer reports are filed around the middle of the following month, so a shelf’s newest loss month can sit one month behind its newest delinquency month. The loss chart states its as-of month in the subtitle. A shelf whose 10-D figures fail our internal reconciliation check is withheld from the annualized series rather than published unverified.

Lease shelves: credit losses and residual value Auto Lease

A lease pool can lose money two ways: a lessee defaults (a credit loss), or a returned vehicle sells for less than the residual value written into the lease (a residual loss). Servicers report the two separately, and so do we.

• Net credit loss follows the loan definitions above — cumulative by months since issuance, annualized by collection month, reconciled and gated the same way — as a percentage of the securitization value (the present value of the remaining lease payments plus the securitized residual), which is the pool measure a lease trust reports. When recoveries and sale proceeds on defaulted leases exceed their value, the result is a net credit gain, shown as a negative loss on every lease shelf.
• Cumulative residual value gain (loss) is the servicer’s cumulative gain or loss on vehicles returned at the end of their leases, as a percentage of the original securitization value; a gain is positive. It is cumulative by months since issuance and is the larger of the two risks on a prime lease pool.
• Residual gain (loss) on vehicles sold, the time-series view of the Residuals tab, is the servicer’s gain or loss on the leased vehicles it sold in a month as a percentage of the residual value it measures them against — the margin on the cars that came back. It is summed across a shelf’s deals before dividing, so each deal weighs by what it sold. Servicers differ in which vehicles they count and which value they use: BMW measures against the ALG residual, Ford against the residual portion of the securitization value, the others against the securitization value; Volkswagen counts scheduled terminations only, World Omni scheduled and early terminations, Mercedes-Benz all liquidated leases, Nissan auctioned vehicles. Months with less than $500,000 of residual value sold are not shown.
• The two are not added into one “total loss”: only some servicers report such a figure, and their definitions differ.
• All Issuers on the lease pages adds every lease shelf’s servicer reports together, by balance: delinquency and annualized credit net loss, shown for the months every shelf has reported. Residual results and turn-in are not added across shelves, because each servicer defines them its own way.
• Turn-in rate is the share of leases reaching their scheduled end whose vehicle comes back to the trust rather than being bought by the lessee or a dealer — the driver of residual results, since only a returned vehicle is sold at auction. Ford and GM Financial print the returned and retained units behind it, and we compute the rate from them; Volkswagen, World Omni and Porsche print their own rate, which we show as printed (Volkswagen counts a dealer purchase at lease end as turned in). BMW, Mercedes-Benz and Nissan do not report one. A month in which fewer than 25 leases in a deal reached their scheduled end is not shown, where the report prints that count (all but Porsche). The shelf line combines its deals, each weighted by its leases reaching scheduled end that month. Because the definitions differ, compare a shelf’s trend rather than levels across shelves. A rate of zero before any lease in a deal has reached its scheduled end is shown as no reading.
• Verification. Residual and turn-in figures are the servicers’ own, from the monthly 10-D report. Each day the newest reports are re-read and compared with what we store, and each month’s cumulative residual is checked against the prior month plus the month’s result. Where an issuer’s definition can be rebuilt from the loan-level filings (GM Financial, Ford), we rebuild it as an independent check.

5. Repossessions (First-Time Repossession Rate) Auto Loan

The monthly rate at which loans are repossessed for the first time, from the repossession flag in the loan-level filings.

• Event: a loan’s first repossession — the first month its repossession flag is set. A loan counts once, ever. If the vehicle is later redeemed and the loan reinstated, the repossession still counts: it happened.
• Numerator: first repossessions in the month among loans reported at the start of it, whatever their status then — including repossessions of loans already charged off, which some servicers carry out routinely after charge-off.
• Denominator: loans outstanding at the start of the month: a balance above $10, not yet repossessed and not charged off. Loans already paid off or charged off do not count as outstanding.
• Not a repossession: a flag on a loan that was current at the start of the month and paid in full during it. That is a payoff recorded in the wrong field. A borrower who is current but surrenders the vehicle is still counted, as is a delinquent borrower who redeems it by paying the loan off.
• First month in our data: a flag already set when a deal first appears in our records is not counted — that repossession happened before our history of the deal begins.
• Universe: loans with a consumer credit score of 300–900, as for delinquency; loans without a score, and commercial-scored loans, are excluded.
• Filters: the credit-score range and new/used filters apply together, on 10-point score bands.

How the loan-level count compares with the servicer’s

Seven servicers print the number of vehicles repossessed each month in their 10-D reports. We compare it with the loan-level count every day; a shelf whose ratio moves is flagged for review.

Shelf	Loan-level ÷ servicer	Reading
Capital One, Ford, Hyundai, Toyota	≈ 1.0	Agree within a few percent; single months can differ by a month of timing.
Honda	≈ 1.10	The difference matches the share of repossessions later reversed, which Honda’s report excludes and this site counts.
Santander (SDART, DRIVE)	≈ 0.86	The loan-level file flags fewer repossessions than the servicer reports, steadily since 2017; the cause cannot be determined from public data.

Known breaks in the repossession flag

• Exeter, from November 2022: the loan-level file largely stops setting the flag — first-time repossessions fall from 3,826 in October 2022 to 66 in November while charge-offs without a repossession rise in step. Exeter’s series ends at October 2022.
• Ally, January–July 2020: the loan-level file flagged loan payoffs as repossessions (about 12,000 a month against about 450 normally). Within that window no flag on a loan paid in full that month is counted. Corrected, Ally’s 2020 rate is consistent with the 1.53% whole-portfolio repossession rate Ally published for 2020 (Ally Auto Receivables Trust 2022-1 prospectus).

A daily check flags any shelf whose repossession rate suddenly collapses or spikes, so a new break of this kind is caught in the month it appears.

6. Recoveries on Defaulted Loans Auto Loan

How much of what a borrower owed is recovered after a loan defaults — from the sale of the repossessed vehicle and from payments collected afterward. Loss severity is 100% minus the recovery rate.

• Default: the month a loan is reported as charged off in the loan-level filings and stays charged off. A charge-off the borrower cures — the loan is reinstated and payments resume, or it is paid off — is not a default; if the loan defaults again, it enters at the later charge-off. A servicer re-booking the remaining balance after the sale is not a cure: that is still one default, at the first charge-off.
• Balance at default (denominator): the loan’s balance at the end of the month before it defaulted — the same filing field for every issuer.
• Recovered (numerator): all cash collected on the loan in the 12 monthly reports starting with the month of default: vehicle sale proceeds, insurance and later payments from the borrower, net of repossession, auction and collection costs.
• Cohort: each point covers the loans that defaulted over four consecutive quarters, recalculated every quarter. Every dollar in the numerator belongs to a loan in the denominator.

Why cohorts, not “flow” ratios

Issuers often quote a recovery rate as this period’s recoveries divided by this period’s charge-offs. The two come from different loans — most recoveries arrive months after the charge-off they belong to — so the ratio moves with default volumes as much as with vehicle values. When charge-offs fell sharply in 2020–21, one issuer’s flow ratio exceeded 100%. Following the same loans removes that distortion, and is how rating agencies treat recovery: as a property of a group of defaulted loans.

Why 12 months, and the early read

Twelve months captures 88–96% of what is recovered within 24 months at every issuer shown. A 12-month figure can be published about 15 months after the defaults it covers. For a timelier signal the page also offers the 6-month rate, which runs lower by construction; its newest points — loans not yet 12 months past default — are drawn dashed as an early read.

Recent figures can be revised when later filings show a charge-off was cured. In backtests, 12-month figures moved by 0.3 points or less after first publication, and the 6-month early read by up to 0.8 points, always upward.

How each figure is checked

Every figure is measured loan by loan from issuers’ asset-level filings (Form ABS-EE) and checked, deal by deal and month by month, against the trust’s own servicer report (Form 10-D). An issuer is shown only when the number of loans charged off and the recovery cash tie to its servicer reports. Issuers record the same cash in different fields, and some define the balance at default differently in their servicer reports; the table records how each is handled.

Issuer	Where the vehicle sale cash is recorded	Handling and notes
Santander Drive, Drive Auto	Amounts recovered	Scaled to the servicer report’s net proceeds (trailing four quarters). Loans Santander bought back after default leave the cohort at buyback; the sponsor’s payment is not a recovery.
Exeter	Amounts recovered	The servicer report also deducts recovery fees paid to the servicer; scaled to its net proceeds.
Bridgecrest, Carvana	Amounts recovered, before liquidation expenses	Scaled to the servicer report’s net proceeds. Carvana from defaults in January 2022: its 2021 filings report the written-off balance as recovered on top of the cash.
AmeriCredit, GM Financial	Amounts recovered plus repossession proceeds (which carry repossession expenses as negatives, including those booked before the charge-off)	The two fields together equal the servicer report’s net liquidation proceeds. From defaults in June 2019, when the filings began to carry sale proceeds. The servicer report’s charge-offs run about 4% below the prior-month balance used here.
CarMax, CarMax Select	Amounts recovered	CarMax reinstates some charged-off loans. Its servicer report counts every charge-off and subtracts a loan’s original charge-off when it is reinstated; here a charge-off the borrower cured is not a default.
Volkswagen	Amounts recovered	The servicer report’s balance at default runs about 5% above the filing field; recoveries agree to the cent.
Hyundai	Principal paid in the month of charge-off	Sale proceeds are applied before the charge-off; counted as recovered.
Ford Credit, Capital One	Principal paid in the month of charge-off	Charged off net of the sale; the balance at default is taken before the sale.
BMW	Not reported	Charged off net of the sale; proceeds are inferred as the balance at default less the amount charged off, which ties to the servicer report. Few defaults: shown with its margin and a caution.
Ally	Amounts recovered plus repossession proceeds	The two fields together equal the servicer report’s liquidation proceeds.
Honda, Nissan, Toyota, World Omni, World Omni Select	Amounts recovered	Tie to the servicer reports.

Exclusions

• Sponsor buybacks after default. A payment by the sponsor to repurchase a defaulted loan is not a recovery from the borrower or the vehicle; such loans leave the cohort when bought back.
• Impossible values. A loan whose reported recoveries exceed 150% of what it owed is excluded as a data error.

Precision

Each figure is an exact count of what happened to the loans in the cohort. With few defaults, though, a handful of unusually good or bad sales moves the rate, so a change may reflect chance rather than a shift in recovery values. The chart shades the 95% margin around each line and draws a point hollow and grey where the margin exceeds ±3 percentage points. Issuers with few defaults are shown with that caution rather than left out.

Comparing issuers

Levels differ across issuers for real reasons — credit mix, vehicle age, loan-to-value, how and where vehicles are sold. Trends within an issuer are the cleanest read.

Not shown

• Mercedes-Benz: its loan-level filings omit the vehicle sale cash.
• Fifth Third: its loan-level filings do not yet reconcile to its servicer reports.
• Exeter Select: not yet four quarters of defaults with 12 months of follow-up.
• Lease shelves: lease losses come from residual values on returned vehicles, covered under Residuals and Turn-in.

7. Credit Score Methodology Auto Loan

Analysis uses the consumer credit bureau score reported in SEC ABS-EE filings, filtered to the 300–900 range. Seven of twenty-five issuers explicitly identify this as a FICO score; the remaining eighteen report a generic “bureau score” label. Industry practice strongly favors FICO Auto Score for auto lending decisions, but we cannot independently verify the score model for all issuers. Loans without a reported score (~0–15% by issuer) are excluded from the analysis.

Commercial / dealer bureau scores are excluded because they use a different scoring scale than consumer scores and would contaminate the low-score bands of prime issuers.

Scores are used as reported, on one scale. The site does not convert between score models, so a 640 from one issuer is not necessarily risk-equivalent to a 640 from another (see Known Limitations).

Carvana, 2026 deals: Carvana reports both a VantageScore and a FICO for most loans in its 2026 deals. The FICO is used where both are given, and the VantageScore where it is the only score; both are retained in our data.

8. Vintage Curves vs Time Series

• Vintage Curves plot the metric as a function of months since deal issuance, one line per deal series. Useful for comparing the seasoning behavior of different vintages within the same issuer.
• Time Series aggregates each issuer’s deals into a single calendar-time curve, one line per issuer. Useful for cross-issuer trend comparisons.

9. Scope & Boundary Notes

Auto Loan Issuers with commercial-auto exposure: for issuers that securitize both consumer and commercial auto loans (Ford, Ally), we exclude commercial-bureau-scored loans from our “Below Prime” bucket. Commercial bureau scores use a different scale than consumer FICO (typically 0–100), and including them would mix fleet and small-business loans with subprime consumer loans. For a consumer credit analysis, the stricter exclusion is more defensible.

Auto Loan Score boundary: our “Below Prime” bucket is scores 300–659 inclusive, reflecting 10-point score bucketing. Definitions that treat 660 as below prime will show slightly higher counts; the difference is typically small (∼1–4% for small-volume issuers where a repo happens to land on exactly 660).

Auto Lease Lease shelves show delinquency, credit losses, residual value gain (loss) and turn-in, all from the servicers’ monthly reports; repossessions are not yet published for lease shelves.

10. Known Limitations

• Auto Loan Repossession flag quality varies by issuer: the flag is only as good as each issuer’s file. Known breaks (Exeter from November 2022; Ally, January–July 2020) and how each shelf’s count compares with its servicer’s are set out in section 5.
• Auto Loan Score model ambiguity: 18 of 25 issuers use a generic “bureau” label instead of explicitly naming FICO. A Vantage­Score 640 may not be risk-equivalent to a FICO 640.
• Auto Loan Vehicle valuation varies by issuer: the reported vehicle value may reflect wholesale, retail, dealer purchase price, or another basis depending on the issuer’s servicing system. Cross-issuer LTV comparisons should be treated as directional, not precise.
• Auto Loan Commercial purpose loans: some consumer-bureau-scored loans may have commercial use (e.g., sole proprietors financing work vehicles). The loan tape does not distinguish these, but visible evidence of distortion is minimal.
• Auto Lease Entity-name aliasing: a handful of issuers (notably Ford) file under both their trust name and depositor name (e.g., “Ford Credit Auto Lease Two LLC”); these may appear as separate series in the legend with identical numbers.
• Auto Lease No deal-specific triggers: pool-level performance against deal-specific covenants (delinquency triggers, residual loss triggers) requires prospectus-level data not included in ABS-EE.
• Non-disclosed fields: SEC ABS-EE does not require disclosure of down payment amount, borrower income (only payment-to-income), employment status, housing, or dealer identity. These variables affect default risk but cannot be analyzed from this dataset.

11. Update Schedule & Audit Trail

• Daily pipeline: runs twice a day, every day of the week (6 AM and 6 PM ET). Figures change only when an issuer files new data.
• Build identifier: each generated page carries a build timestamp.
• Changelog: dated changes to definitions, coverage and the pages are listed on the changelog.

12. Credit Card Trusts — What Each Figure Is

The Credit Card page shows each master trust’s monthly figures as its own servicer reports them on Form 10-D, and every stored figure is tied back to the certificate it came from (the daily check cc_10d_tieback re-reads the newest months from the filing and, where the certificate prints the components, recomputes the rate). The trusts do not define the figures alike, so levels are comparable only with the definitions below in mind; trends within a trust are always comparable. Where a certificate prints amounts but no rate, the rate is computed here and says so.

Yield and excess spread are shown on one footing for every trust: yield before losses (finance charges, fees, interchange and, where the trust counts them as finance charge collections, recoveries, over principal receivables, annualized) — the market’s convention, with charge-offs on their own line — and excess spread at the level of the trust: its group figure where it prints one, otherwise each reporting series’ figure weighted by the base of its own yield, leaving out a series in its first month (a partial or long first period distorts it), and blank for a month in which any series’ figure is missing or every series is new. The table says which line each comes from.

Excess spread and the early-amortization trigger. A trust's notes begin to pay down early if its excess spread, averaged over three months, falls below a required level — 0% at every trust today. The Excess Spread chart opens on that 3-month figure, measured where each trust’s trigger tests it. For a trust that issues all its notes from one series (Chase, Citi, Bank of America, Capital One, Synchrony, Wells Fargo) it is that series’ figure, whose base rate averages the coupons of all its notes. For a trust whose series each carry their own trigger (American Express, First National, Comenity, Barclays Dryrock) it is the lowest of its series, leaving out a series in its first month. A series setting money aside to repay its notes on schedule shows a lower figure while it does so; that is part of the test, not a sign of strain. Capital One reports its trigger in dollars only: its percentage is calculated from its reports on its prospectus’s definition (portfolio yield less a base rate of 1.25% plus the weighted coupon of its notes) and matches Capital One’s own published figure. Bank of America’s printed percentage is measured after the losses its Class D certificate absorbs; its trigger is tested in dollars before them. Where a report prints its required level, a level other than 0% is noted under the chart. The Monthly view shows the trust-level figure described above; All Trusts is a monthly market average, since no trigger applies to it.

History. Before the trusts filed monthly reports on Form 10-D (2006), they filed the same servicer statements as exhibits to Form 8-K; those are read for the months before each trust’s 10-D series begins, and for months the current reader misses (Bank of America, 2009–2012). Where both read the same filing they agree exactly, and the 10-D figure is the one shown. A measure is carried back only on its present definition: gross yield for Citi (printed yield plus net charge-offs, which matches Citi’s later gross figure exactly), Capital One (printed gross) and Chase from 2006 (printed); excess spread only where the printed figure is already the trust’s own. Older reports that print only dollar amounts give charge-off rates under the same test as above: Chase from 2002 (the pool’s default amount over its balance, at month end before 2007), American Express net from 1996, and Bank of America from 2001 on its collateral certificate, gross — its reports before late 2008 print no recoveries, so All Trusts net charge-offs begin in September 2008. All Trusts counts a trust that is missing a month’s report at its last balance, so the gap counts against the share test rather than shrinking the market. Trusts no longer issuing (Discover, GE Capital (Synchrony), Sears, Providian, Metris, Household / HSBC, HSBC Private Label, Household Affinity, National City, Amex charge cards) are listed under Earlier trusts, with the measures their own reports support; their yield and excess spread are not shown, since their definitions cannot be checked against a current report.

Trust	Payment rate	Charge-offs	Yield	Excess spread
American Express	“Monthly Payment Rate” (basis not printed)	“Annualized Default Rate” (gross) and “…, Net of Recoveries”	Total collections of finance charge receivables, recoveries included, ÷ beginning principal receivables, annualized (the printed “Trust Portfolio Yield” is this less defaults)	“Excess Spread Percentage”: the same Group figure on every series
Bank of America	Total collections (principal and finance charge) ÷ prior-month principal and finance-charge receivables	“Total / Net Charge-Offs as a percentage of Average Principal Receivables Outstanding”	“Total Cash Yield” (finance charges, discount option receivables, interchange and recoveries) as a percentage of the BAseries investor interest	“Excess Available Funds Percentage” of the BAseries
Barclays Dryrock	“Total payment rate”: principal and finance-charge collections, investment earnings excluded (basis not printed)	“Total Gross / Net Losses as a percentage of the Portfolio Principal Balance”	computed here: the trust’s total finance charge collections (fees, interchange and recoveries; investment earnings excluded) ÷ beginning principal receivables plus additions, × 12	weighted across series: each series’ “Excess Spread Percentage” by its allocation amount; a series in its first month is left out
Capital One	“Principal Payment Rate”: principal collections ÷ beginning-of-month principal receivables	“Annualized (Net) Default Rate” ÷ adjusted beginning-of-month principal receivables, additions included	“Annualized Yield” on the same adjusted base (already before losses)	not reported
Chase	“Principal Payment Rate”: principal collections ÷ beginning pool balance	“Gross / Net Losses as a Percentage of Average Pool Balance”, annualized	“Yield – Finance Charge, Fees & Interchange” plus the yield on discount receivables, before net credit losses	“Excess Spread Percentage” of the CHASEseries
Citi	“Total Payment Rate”: principal and finance-charge collections (collateral-certificate basis)	net only: the “Credit Loss Component” of the collateral certificate’s yield	the “Yield Component” of the collateral certificate’s portfolio yield, before the credit loss component	“Surplus Finance Charge Collections” as a percentage
Comenity (WFN)	“Monthly Payment Rate”: total collections (principal and finance charge) ÷ beginning-of-month principal receivables	“Gross / Net Principal Charge-Off Rate”, ÷ end-of-month principal receivables, annualized	“Gross Portfolio Yield”, weighted across series by collateral amount plus principal accumulation account, a series in its first month left out (printed from Nov-2014)	weighted across series: each series’ “Portfolio Adjusted Yield” by its collateral amount plus principal accumulation account; a series in its first month is left out
First National	computed here: total collections ÷ average principal receivables — the certificate prints amounts, not a rate	computed here, gross only: defaulted amount ÷ average principal receivables × 12	computed here: the trust’s finance charge collections (interchange, recoveries, finance charges and fees, discount receivables) ÷ average principal receivables, annualized by days	weighted across series: each publicly reported series’ “Net Yield Percentage” by its collateral amount plus principal accumulation account as of the first day of the month (the base of its yield in the prospectus); a series in its first month is left out, and privately held variable-funding series publish no statement
Synchrony	“Payment Rate”: principal collections ÷ beginning-of-period principal receivables	“Gross / Net Charge-Off Rate”: defaulted amount (less recoveries) ÷ beginning-of-period principal receivables	“Gross Trust Yield”: finance charge collections plus recoveries ÷ beginning principal receivables	“Excess Spread Percentage” = yield less base rate
Wells Fargo	“Principal Payment Rate”: principal collections ÷ beginning-of-period principal receivables (day-weighted in a month with account additions)	“Gross / Net Charge-Off Rate”	trust-wide: “Yield from Finance Charges Collected” plus “Yield from Interchange” (the noteholder series receives all interchange, so its own yield runs higher)	“Excess Available Funds Percentage”

Delinquency is the share of receivables past due at the end of the collection month for every trust, but the first bucket is 30–59 days at some trusts and 31–60 at others, and Capital One measures it against total rather than principal receivables; the 30+ figure is the sum of the buckets. Charge-offs on the page are the net rate where a trust reports one and the gross rate where that is all it reports (First National), and say so. All Trusts weights each trust’s own figure by its receivables; it is shown only where at least three trusts holding at least 80% of receivables report the figure.

Definitions. Charge-offs are the month's principal charge-offs, net of recoveries unless marked gross, annualized, over principal receivables. Delinquency is balances 30 or more days past due at month end over principal receivables. The payment rate is the month's collections over the prior month's receivables. A trust's own printed figure is shown as printed; where a report prints only dollar amounts, the rate is computed on these definitions, and used only where, in the later months that show both, it matched the trust's own printed rate to within 0.05 of a percentage point in a typical month and within 0.10 of a point in at least 9 months out of 10. Amex's 30+ day delinquency before Jan-2014 is its delinquent dollars over ending principal receivables (its later printed rate divides by total receivables and runs about 0.03 point lower). Amex's net charge-off rate before Oct-2008 is its gross default rate less the share of that month's defaults recovered, as Amex computes its own later net rate. Bank of America's charge-off rates before Nov-2015 are those of its collateral certificate (Series 2001-D, nearly all of the trust's investor interest), which were within a few hundredths of a point of the trust's rate in most months both were printed. Chase's net charge-off rate before Nov-2015 is its pool default amount (gross losses less recoveries), annualized, over the pool's average balance. Capital One's net charge-off rate before Feb-2016 is its gross default rate less the share of that month's defaults recovered. Comenity's 30+ day delinquency before Nov-2015 is its delinquent principal over month-end principal receivables.

Coverage. First National reports a gross default rate only; its charge-off series is gross of recoveries and is left out of the All Trusts charge-off line. Capital One's servicer report does not state excess spread. Citi reports a net default rate only. Amex's delinquency buckets begin Jan-2014 (the 91–120 and 120+ buckets Nov-2015). Chase reports a net loss rate only before Nov-2015. Comenity's default rates begin Sep-2011 and its delinquency buckets Nov-2015; its portfolio yield Nov-2014. Barclays Dryrock's net default rate begins Nov-2015 (gross from Dec-2012). Months in which a trust had no public notes outstanding are blank.

13. Data Downloads & Licence

Every series on these pages can be downloaded as a CSV file from the Downloads page: auto loan delinquency, net loss and repossessions, auto lease delinquency and net loss, the credit card trusts’ monthly figures, and the macro indicators. Each file is written from the same figures the charts draw and opens with comment lines naming its source, definition, the month it runs through and the citation to use. The files are published under CC BY 4.0: free to use, share and adapt with attribution to Consumer Credit Matters.

This methodology page is regenerated with every pipeline run and is always in sync with the data currently served on both explorers.